Neural Circuit and Cognitive Development: Comprehensive Developmental Neuroscience [2 ed.] 0128144114, 9780128144114

Table of contents :
Neural Circuit and Cognitive Development
Copyright
Contributors
Part I. Circuit Development
1. Neural circuits of the mammalian main olfactory bulb
1.1. Introduction
1.2. Synaptic organization of the main olfactory bulb
1.2.1. Organization of sensory inputs
1.2.2. Synaptic microcircuits
1.2.2.1. Glomerular layer microcircuits
1.2.2.2. External plexiform layer microcircuits
1.2.3. Neural computation
1.2.3.1. Contrast enhancement
1.2.3.2. Slow timescale decorrelation
1.2.3.3. Fast timescale synchronization
1.2.3.4. Downstream decoding
1.2.4. Modulation of sensory processing
1.2.4.1. Local circuits and centrifugal innervation
1.2.4.2. Brain state and context
1.3. Plasticity in the main olfactory bulb
1.3.1. Adult neurogenesis
1.3.1.1. Regeneration of sensory input
1.3.1.2. Adult-born interneurons
1.3.2. Circuit and synaptic plasticity
1.4. Concluding remarks
Acknowledgments
References
2. Functional circuit development in the auditory system
2.1. Introduction to auditory system development
2.1.1. A neurobiological approach to studying auditory system development
2.1.2. Overview of auditory circuitry
2.1.3. Basic concepts of cochlear transduction
2.1.4. Scope of this chapter
2.2. Development of peripheral circuits
2.2.1. Development and maturation of cochlear hair cells
2.2.2. Development of the place code
2.2.3. Development of afferent and efferent peripheral circuits
2.2.3.1. Phase 1 - initial wiring of the cochlea
2.2.3.2. Phase 2 - synapse formation and refinement
2.2.3.3. Phase 3 - ear opening and maturation of response to sound
2.2.4. Summary and themes
2.3. Development of brain stem circuits
2.3.1. Functional circuit assembly in the brain stem
2.3.2. Development of fine-scale connectivity in the medial superior olive
2.3.3. Development of fine-scale connectivity to the lateral superior olive
2.3.4. Afferent regulation of cochlear nucleus development
2.3.5. Afferent regulation of third-order brain stem nuclei
2.3.6. Influence of the source and pattern of afferent activity on brain stem circuits
2.3.7. Conclusions
2.4. Development of auditory midbrain and forebrain circuits
2.4.1. Development of thalamocortical subplate circuitry
2.4.2. Development of local cortical circuits
2.4.3. Afferent regulation of higher auditory circuit development
2.4.4. Experience-dependent influences on circuit development
2.4.5. Developmental regulation over reinstating hearing in the deaf
2.4.6. Conclusions and directions for future research
References
3. Development of the Superior Colliculus/Optic Tectum
3.1. Nomenclature
3.2. Functional role
3.3. General anatomical organization of the superior colliculus
3.4. Spatial topographies, multisensory integration, and motor output
3.5. The maturation of the superior colliculus
3.5.1. The neonate
3.5.2. Sensory chronology
3.5.3. The development of multisensory neurons
3.5.4. Superficialedeep (multisensory) layer maturational delay
3.5.5. The development of multisensory integration
3.5.6. The impact of sensory experience on the maturation of multisensory integration
3.6. Summary
Acknowledgments
References
4. Multisensory Circuits
4.1. Overview of the microcircuit in the cerebellar cortex
4.1.1. Cell types and afferent fibers
4.1.2. Generation of neurons that constitute microcircuit on PCs
4.1.3. Compartmentalization of the cerebellum
4.2. Development of CFePC synapses
4.2.1. Multiple innervation of PCs by CFs in early postnatal period
4.2.2. Functional differentiation of multiple CFs
4.2.3. Dendritic translocation of single CFs
4.2.4. Early phase of CF synapse elimination
4.2.5. Late phase of CF synapse elimination
4.3. Development of PFePC synapses
4.3.1. Formation of PFePC synapses
4.3.2. Stabilization and maintenance of PFePC synapses
4.3.3. Developmental elimination of PFePC synapses
4.3.4. Heterosynaptic competition between PF and CF inputs
4.4. Development of inhibitory synapses from basket cells and stellate cells to PCS
4.4.1. Formation of basket Cell-PC synapses
4.4.2. Formation of stellate Cell-PC synapses
4.4.3. Activity-dependent remodeling of inhibitory synapses
4.5. Summary and conclusions
Acknowledgments
References
5. Cerebellar Circuits
5.1. The loose and uncritical use of the term in ways that are so generalized as to be unhelpful and even confusing
5.2. A lack of a universal presence of certain columns within cortical areas, brains, and species is undermining the idea that similar building blocks comprise all cortical circuits
5.2.1. The use of physiological methods to reveal columns
5.2.2. Columnar organization of some afferent and efferent projections
5.2.2.1. Modules of visual cortex
5.2.2.2. Ocular dominance columns/stripes
5.2.2.3. Orientation columns
5.2.3. Gene expression in the cortex in “columnar” fashion
5.2.3.1. A system of interleaving modules in rodent layer II
5.2.3.2. Overlap between columnar entities within the same structures; combining physiological and anatomical definitions
5.3. General concept that the cortical column (even just an arbitrary unit column that includes the full depth of the cortex) has a universal constant number of neurons associated with it
5.3.1. Number of neurons in a cortical column
5.4. Lack of correlation between the absence or presence of particular columns and a specific sensory or cognitive processing network (comparisons across the same brain and across close and more distant species)
5.4.1. Microscopic and macroscopic cell patterning defining cortical modules
5.4.2. Are barrels cortical columns?
5.4.2.1. A system of interleaving modules in rodent layer VI
5.4.2.2. Function of barrels
5.4.2.3. Microcolumns and apical dendritic bundles
5.4.3. Complex relationship relations between minicolumns and dendritic bundles
5.4.3.1. Columns outside the mammalian isocortex
5.4.3.2. Columns in nonmammals
5.4.4. What is the function of a cortical column?
5.4.5. Columns in neuropathology
5.5. What is the correlation between the columnar development of the brain and future columns
5.5.1. Cortical columns during development
5.5.1.1. Ontogenic units/columnsdthe fundamental building blocks in the developing neocortex
5.5.1.2. Sibling neuron circuits in the developing columns
5.5.2. Transient columnar domains during development
5.5.3. The way forward
Acknowledgments
References
6. Dendritic Spines
6.1. Introduction - synaptic plasticity and synaptic learning rules
6.2. Discovery of STDP
6.3. Definition and forms of STDP
6.3.1. Hebbian STDP
6.3.2. Anti-Hebbian STDP
6.4. STDP is part of a broader, multifactor plasticity rule
6.5. Functional properties of STDP
6.5.1. Properties of Hebbian STDP
6.5.2. Properties of anti-Hebbian STDP
6.5.3. STDP and circuit homeostasis
6.5.4. Neuromodulation of STDP
6.6. Cellular mechanisms for STDP
6.6.1. Mechanisms for LTP and LTD components of STDP
6.6.2. Dendritic excitability and STDP
6.7. Is STDP a realistic learning rule in vivo?
6.8. How does STDP contribute to development of neural circuits?
6.9. How does STDP contribute to adult plasticity and learning?
6.10. Summary - role of spike timing in synaptic plasticity
References
7. Cortical Columns
7.1. Introduction
7.2. The mature cortical circuit
7.2.1. Supragranular excitatory neurons
7.2.2. L4 excitatory neurons
7.2.3. Infragranular excitatory neurons
7.2.4. Inhibitory neuron connectivity
7.2.5. Corticothalamic and corticocortical inputs
7.3. Connections during birth and migration
7.4. Thalamocortical innervation
7.5. Developmental critical periods and barrel formation in the somatosensory system
7.5.1. Formation of barrels
7.5.2. Receptive fields of barrel cortex neurons
7.5.3. Critical periods for functional connectivity of ascending somatosensory pathways
7.5.4. Intracortical excitatory plasticity
7.5.5. Critical periods for inhibition in barrel cortex
7.6. Adult plasticity
7.7. Conclusion
List of acronyms and abbreviations
Acknowledgments
References
8. Neonatal Cortical Rhythms
8.1. General principles of motor cortex - history, evolution, and organization
8.1.1. What is the “motor cortex”?
8.1.2. Discovery of motor cortex - historical outlook
8.1.3. Evolution of the motor cortex
8.1.4. Topographic organization of the motor cortex - species-specific organization
8.2. The connectivity of the motor cortexdafferent and efferent projections
8.2.1. Efferent connectivity of the motor cortex
8.2.1.1. Anatomical organization of subcerebral projections
8.2.1.2. Functional organization of subcerebral projections
8.2.1.3. Organization of motor cortex-spinal cord connectivity
8.2.1.4. Species-specific differences of corticospinal connectivity
8.2.1.5. Organization of motor cortex-striatum connectivity
8.2.1.6. Organization of reciprocal connectivity between motor cortex and the thalamus
8.2.2. Afferent connectivity of the motor cortex
8.2.2.1. Organization intracortical motor cortex connections
8.2.2.2. Organization of basal forebrain afferent projections to motor cortex
8.3. Development of motor cortex connections
8.3.1. Specification and differentiation of subcerebral projection neurons
8.3.2. Axon guidance of subcerebral, including corticospinal projections
8.3.3. Development of corticospinal connectivity in the spinal cord
8.3.4. Activity-dependent developmental refinement of corticospinal connectivity
8.3.5. Molecular development of human corticospinal circuitry
8.3.6. Development of intracortical and subcortical afferent connectivity
8.3.6.1. Development of thalamic connectivity
8.3.6.2. Development of intracortical connectivity
8.3.6.3. Development of afferent cholinergic inputs from the basal forebrain
8.4. Technological advances to enable future investigations of motor cortex connectivity and function including in nonhuman primates and humans
8.5. Conclusion
References
9. Spike Timing-Dependent Plasticity
9.1. Introduction - circuit organization
9.1.1. Adult organization - a brief review
9.1.1.1. Subfield features and numbers
9.1.1.2. Zonal specializations. Example - upper and deep CA1
9.1.2. Functional backdrop
9.1.2.1. The spatial navigation system in rodents
9.1.2.2. Developmental milestones in humans
9.2. Circuit development
9.2.1. Early stages
9.2.2. Neurogenesis
9.2.3. Connections - EC
9.2.4. Cell autonomous organization
9.2.5. Neural activity
9.2.6. Maturational events
9.2.7. Coordinated network activity
9.3. Postnatal development of electrophysiological patterns
9.3.1. Postnatal development of single cell electrophysiological properties
9.3.1.1. Entorhinal cortex layer II stellate cells
9.3.1.2. Postnatal development of the dentate gyrus
9.3.1.3. Postnatal development of CA3 pyramidal cells
9.3.1.4. Postnatal development of CA1 pyramidal cells
9.3.1.5. Postnatal development of interneurons
9.3.2. Early developmental patterns of brain activity
9.3.3. Development of major hippocampal rhythms
9.3.3.1. Development of theta oscillations
9.3.3.2. Development of SWR
9.4. Conclusion
References
10. Methods to traces circuits
10.1. Introduction
10.2. General organization of the basal ganglia
10.3. Organization of corticostriatal projections
10.4. The striatum
10.4.1. Physiology
10.4.2. Striatal interneurons
10.5. The effect of direct and indirect striatal output pathways on behavior
10.6. Direct and indirect striatal output pathways
10.7. External segment of the globus pallidus
10.8. The subthalamic nucleus
10.9. Development of the basal ganglia
10.10. Summary
References
11. Visual cortex connections
11.1. Introduction
11.2. Organization of the adult circuitry of the cerebellar cortex
11.2.1. Cajal and the cerebellar circuit
11.3. The modular organization of the cerebellar cortex
11.4. Development of the heterogeneity of Purkinje cells
11.5. Development and refinement of climbing fiber projections
11.6. Development and refinement of mossy fiber projections
11.7. Cerebellar interneurons in the cerebellar circuit
11.7.1. Excitatory interneurons
11.7.1.1. Granule cells
11.7.1.2. Unipolar brush cells (UBCs)
11.7.2. Inhibitory interneurons
11.7.2.1. Granule cell layer interneurons
11.7.2.2. Purkinje cell layer interneurons
11.7.2.3. Molecular layer interneurons
Acknowledgment
References
Part II. Cognitive Development
12. Somatosensory cortex connections
12.1. Introduction
12.2. Frameworks and methods
12.2.1. Conceptual frameworks
12.2.2. Eye-tracking
12.2.3. Electrophysiology
12.2.4. MRI and other imaging methods
12.2.5. Summary
12.3. Overview
References
13. Theories in developmental cognitive neuroscience
13.1. Introduction
13.2. Why do we need theories?
13.3. Frameworks for understanding human functional brain development
13.3.1. Maturational viewpoint
13.3.2. Interactive specialization
13.3.3. Skill learning
13.4. Assumptions underlying the three frameworks
13.4.1. Deterministic vs. probabilistic epigenesis
13.4.2. Static vs. dynamic mapping
13.4.3. Plasticity
13.5. Predictions and evidence
13.6. Functional brain imaging
13.7. Critical or sensitive periods
13.8. Atypical development - from genetics to behavior in developmental cognitive neuroscience
13.9. Interactive specialization - future challenges
13.10. Summary, conclusions, and recommendations
Acknowledgments
References
14. Motor cortex connections
14.1. Introduction
14.2. Postmortem studies and histology
14.2.1. Synaptogenesis and pruning
14.2.2. Myelination
14.2.3. Sex-specific differences
14.2.4. Summary
14.3. Magnetic resonance imaging volume analyses
14.3.1. Gray matter decreases in development
14.3.2. Regional and temporal dynamics
14.3.3. White matter increases in development
14.3.4. Sex differences
14.4. Magnetic resonance imaging brain mapping approaches
14.4.1. Voxel-based strategies
14.4.2. Cortical thickness
14.4.3. White matter
14.4.4. Sex differences
14.4.5. Summary
14.5. Diffusion magnetic resonance imaging
14.5.1. Diffusion tensor imaging theory
14.5.2. Diffusion parameters in development
14.5.3. fiber tractography
14.5.4. Sex differences
14.5.5. Advanced diffusion magnetic resonance imaging techniques
14.5.6. Summary
14.6. Connecting different techniques
14.6.1. Multimodal imaging
14.6.2. Brainebehavior relationships
14.7. Conclusions and future directions
References
15. Motor circuits
15.1. Learning probability distributions
15.2. Learning co-occurrence statistics
15.2.1. Learning co-occurrence statistics in speech - word segmentation
15.2.2. Do infants learn words from co-occurrence statistics?
15.2.3. Learning co-occurrence statistics in the visual domain
15.2.4. Learning co-occurrence statistics in speech - word segmentation
15.2.5. Learning co-occurrence relations between words and referents - cross-situational learning
15.3. Linking individual differences in statistical learning to language development
15.4. Statistical learning in individuals with language delays and disorders
15.5. Scaling statistical learning to real-world challenges
15.6. Conclusions
Acknowledgment
References
16. Prefrontal cortex connections
16.1. Classic theoretical accounts
16.1.1. Piagetian theory
16.1.2. Gestalt theory
16.2. Prenatal development of the visual system
16.2.1. Development of structure in the visual system
16.2.2. Prenatal visual function
16.3. Visual perception in the newborn
16.3.1. Visual organization at birth
16.3.2. Visual behaviors at birth
16.3.3. Faces and objects
16.4. Postnatal visual development
16.4.1. Visual physiology
16.4.2. Critical periods
16.4.3. Development of visual attention
16.4.4. Cortical maturation and oculomotor development
16.4.5. Development of visual memory
16.4.6. Development of visual stability
16.4.7. Object perception
16.4.8. Face perception
16.4.9. Critical period for development of holistic perception
16.5. How infants learn about objects
16.5.1. Learning from targeted visual exploration
16.5.2. Learning from associations between visible and occluded objects
16.5.3. Learning from visual-manual exploration
16.5.4. Hormonal and environmental influences on object perception
16.6. Summary and conclusions
References
17. Corpus callosum/intracortical connections
17.1. The development of visuospatial processing
17.1.1. Anatomical organizations of the primary visual systems
17.1.2. Ventral stream processes
17.1.2.1. Perception of the global and local levels of visual pattern structure
17.1.2.2. Perception of faces
17.1.2.3. Spatial construction
17.1.3. Dorsal stream processes
17.1.3.1. Spatial localization
17.1.3.2. Spatial attention
17.1.3.3. Mental rotation
17.1.4. Trajectories of dorsal and ventral stream development
17.1.5. Neurodevelopmental disorders of visuospatial processing
17.1.5.1. Perinatal stroke
17.1.5.2. Spina bifida
17.1.5.3. Neurogenetic syndromes
17.1.6. Summary and conclusions
References
18. Striatal (Basal Ganglia) Connections
18.1. Introduction
18.2. Different forms of memory
18.2.1. Short- and long-term memory
18.2.2. Declarative and nondeclarative memory
18.2.3. Declarative or explicit memory
18.2.4. Nondeclarative, procedural, or implicit memory
18.2.5. Relations between different forms of memory
18.3. Developmental changes in declarative memory
18.3.1. Episodic memory
18.3.2. Autobiographical memory
18.4. Mechanisms of developmental change
18.4.1. Neural structures and processes
18.4.2. The neural substrate of declarative memory
18.4.3. Development of the neural substrate supporting declarative memory
18.4.4. Functional consequences of development of the temporal-cortical network
18.4.5. Basic cognitive and mnemonic processes
18.4.6. Encoding
18.4.7. Consolidation and storage
18.4.8. Retrieval
18.4.9. Conclusion
References
19. Thalamic Connections
19.1. Introduction
19.1.1. The nature of language
19.2. Speech perception
19.2.1. Prenatal perception of speech and language
19.2.2. Speech perception in neonates
19.2.3. The development of speech perception in infancy
19.2.4. Audiovisual integration in early speech perception
19.3. Speech production
19.3.1. Speech production in infancy
19.3.2. The relationship between speech and motor development
19.3.3. Phonological development
19.4. Socialecognitive foundations of language
19.4.1. Social engagement
19.4.2. Infant-directed talk
19.4.3. Intentional communication
19.4.4. Pragmatic development
19.5. Lexical development
19.5.1. Stages of lexical development
19.5.2. Developmental processes
19.5.3. Neural bases of word learning
19.6. Syntactic development
19.6.1. Developmental stages in syntactic development
19.6.2. Early sentences
19.6.3. Grammatical morphology
19.6.4. Later grammatical development
19.6.5. Neural bases of grammatical development
19.7. Language disorders
19.7.1. Overview of developmental language disorders
19.7.2. Speech perception
19.7.3. Speech production
19.7.4. Socialecognitive foundations of language
19.7.5. Lexical development
19.7.6. Syntactic development
19.7.7. Neural foundations of language disorders
19.8. Conclusions
Acknowledgments
References
20. Hippocampal connections
20.1. Introduction
20.2. Face processing in adults
20.2.1. How adults process faces
20.2.2. Models of face processing
20.2.3. Neural substrates of face processing
20.2.4. Conclusions
20.3. Face processing in the first year of life
20.3.1. How infants learn to see faces
20.3.2. How infants process facial expressions
20.3.3. Neural substrates of face processing in infants
20.3.4. Neural signatures of face processing in infants
20.3.5. Conclusions
20.4. Face processing in toddlers and preschoolers
20.4.1. How young children process faces
20.4.2. How young children process facial expressions
20.4.3. Functional signatures of face processing in young children
20.4.4. Conclusions
20.5. Face processing in school-age children and adolescents
20.5.1. How children and adolescents process faces
20.5.2. How children and adolescents process facial expressions
20.5.3. Neural substrates of processing in children and adolescents
20.5.4. Conclusions
20.6. Impairments and atypical development of face processing
20.6.1. Prosopagnosia
20.6.2. Congenital cataract
20.6.3. Autism spectrum disorder and Williams syndrome
20.6.3.1. How individuals with autism process faces
20.6.3.2. How individuals with autism spectrum disorder process facial expressions
20.6.3.3. Neural substrates of face processing in autism spectrum disorder
20.6.3.4. Face processing in Williams syndrome
20.7. Conclusions
Acknowledgments
References
21. Tectal connections
21.1. Early sensitivity to mental states - prior neural and behavioral evidence
21.2. Early sensitivity to mental states - neuroimaging studies of young children and infants
21.3. Neural correlates of ongoing theory of mind development in childhood
21.3.1. Response selectivity - fine-tuning preferential responses
21.3.2. Reliable spontaneous (uninstructed) responses to movies
21.3.3. Integration and separation of functional networks
21.4. Future directions - open questions and challenges
21.4.1. Neural correlates of structural changes in theory of mind
21.4.2. Discovering reliable neural markers of individual differences in theory of mind
21.4.3. The role of developmental experience - language
21.4.4. The role of developmental experience - culture
21.4.5. The role of family on ToM - shared environment and shared genes
21.5. Conclusion
References
22. Tegmental connections (substantia nigra)
22.1. Introduction
22.2. Clearing up definitional issues
22.3. The development of empathy
22.3.1. Affect sharing and physiological synchrony
22.3.2. Emotion recognition
22.3.3. Emotion understanding
22.3.4. Perspective-taking and theory of mind
22.3.4.1. Neurophysiological approaches to understanding cognitive empathy
22.3.5. Emotion regulation
22.3.6. Motivation to care
22.4. Neurodevelopmental changes in empathic responding
22.4.1. Evidence from event-related potential
22.4.2. Evidence from functional magnetic resonance imaging
22.5. Maladaptive alterations in developmental trajectories of empathy
22.5.1. Conduct problems
22.5.2. Autism spectrum disorder
22.6. Conclusions
List of abbreviations
References
23. Cerebellar connections
23.1. Introduction
23.2. Facets of attention
23.2.1. Attention and self-regulation
23.3. Brain networks
23.3.1. Alerting
23.3.2. Orienting
23.3.3. Executive attention
23.4. Development of brain and behavior
23.4.1. Infancy
23.4.2. Toddlerhood
23.4.3. Childhood
23.5. Individual differences
23.5.1. Temperament
23.5.2. Genes
23.5.3. Environment
23.6. Plasticity of attention networks
23.7. Summary and integration
Acknowledgments
References
24. Hindbrain connections
24.1. The development of cognitive control and its neural basis
24.1.1. Error monitoring
24.1.2. Control instantiation
24.2. The role of cognitive control in decision-making, motivation, and social behavior
24.2.1. Motivation, decision-making, and cognitive control
24.2.2. Cognitive control and social behavior
24.3. Individual differences in cognitive control
24.3.1. Temperament, cognitive control, and psychopathology
24.3.2. Cross-cultural differences in the development of cognitive control
24.4. Chapter summary and future directions
References
25. Introduction to Cognitive Development from a Neuroscience Perspective
25.1. Introduction
25.1.1. Normative developmental trajectories for executive function from infancy to adolescence
25.2. Clinical insights, from infancy to adolescence
25.3. From biological to environmental predictors of individual differences in executive function
25.3.1. Early executive function predicts academic, socio-cognitive and social success at school
25.4. Conclusions
References
26. Theories in Developmental Cognitive Neuroscience
26.1. Introduction
26.2. The anatomy and physiology of stress
26.3. Prenatal stress and neurobehavioral development
26.3.1. Fetal programming
26.3.2. Stress regulation and pregnancy
26.3.2.1 Changes in the maternal hypothalamicepituitaryeadrenocortical and placental axes over the course of pregnancy
26.3.2.2. Fetal adrenal development
26.3.2.3. Fetal brain development and susceptibility to stress and stress hormones
26.3.3. Gestational stress influences the human fetus
26.3.4. Prenatal maternal psychosocial stress and infant and child development
26.3.4.1. Socioemotional development
26.3.4.2. Hypothalamicepituitaryeadrenocortical axis functioning
26.3.4.3. Cognitive development
26.3.5. Prenatal maternal biological stress signals and infant and child development
26.3.5.1. Social/emotional development
26.3.5.2. Hypothalamicepituitaryeadrenocortical axis functioning
26.3.5.3. Cognitive development
26.3.6. Sex differences
26.3.7. Epigenetics
26.3.8. Interactions with the postnatal environment
26.3.9. Is this fetal programming?
26.3.9.1. Summary
26.4. Postnatal stress and neurobehavioral development
26.4.1. Social regulation of the hypothalamicepituitary eadrenocortical axis and the role of caregivers
26.4.2. Early adversity
26.4.2.1. Diurnal cortisol following postnatal stress
26.4.2.2. Effects of early care on cortisol set points and reactivity
26.4.3. Individual differences in sensitivity to experience
26.4.4. Summary
26.5. Future directions
References
27. Structural Brain Development - Birth Through Adolescence
27.1. Introduction
27.1.1. Issues in studying sex differences
27.1.2. Interpreting sex differences
27.2. Psychological sex differences - nature and development
27.2.1. Cognitive skills
27.2.1.1. Spatial skills
27.2.1.2. Mathematical skills
27.2.1.3. Verbal skills
27.2.1.4. Memory
27.2.1.5. Perceptual speed
27.2.2. Noncognitive sex differences
27.2.2.1. Gender identity
27.2.2.2. Sexual orientation
27.2.2.3. Physical and motor skills
27.2.2.4. Activity interests
27.2.2.5. Temperament and personality
27.2.2.6. Social behaviors
27.2.2.7. Psychological disorders
27.3. Explanations for psychological sex differences
27.3.1. Socialization perspectives
27.3.1.1. Socialization of cognitive sex differences
27.3.1.2. Socialization of noncognitive sex differences
27.3.2. Genetic perspectives
27.3.3. Hormone perspectives
27.3.3.1. Evidence for hormone influences on nonhuman sex-typed behavior
27.3.3.2. Early hormone influences on human behavior
27.3.3.3. Adolescent hormone influences on human behavior
27.3.3.4. Circulating hormone influences on human behavior
27.3.3.5. Exogenous hormone influences on human behavior
27.3.4. Integrated perspectives
27.4. Brain sex differences - nature, development, and consequences
27.4.1. Issues in studying the brain
27.4.2. Sex differences in brain structure and their development
27.4.2.1. Brain volume
27.4.2.2. Regional structure volume
27.4.2.3. Gray matter
27.4.2.4. White matter
27.4.2.5. Implications of sex differences in brain structure
27.4.3. Sex differences in brain function (activation)
27.4.3.1. Lateralization
27.4.3.2. Spatial skills
27.4.3.3. Language
27.4.3.4. Emotion-related processing
27.4.3.5. Functional connectivity
27.4.3.6. Development of sex differences in brain function
27.4.3.7. Implications of sex differences in brain function
27.5. Explanations for brain sex differences
27.5.1. Socialization perspectives
27.5.2. Genetic perspectives
27.5.3. Hormone perspectives
27.5.3.1. Prenatal hormone influences on brain sex differences
27.5.3.2. Adolescent hormone influences on brain sex differences
27.5.3.3. Circulating hormone influences on brain sex differences
27.5.3.4. Exogenous hormone influences on brain sex differences
27.6. Conclusions and future directions
Acknowledgments
References
Index

Citation preview

Neural Circuit and Cognitive Development Comprehensive Developmental Neuroscience Second Edition

Senior Editors-in-Chief

John Rubenstein

Department of Psychiatry & Weill Institute for Neurosciences University of California, San Francisco, San Francisco, CA, United States

Pasko Rakic

Department of Neuroscience & Kavli Institute for Neuroscience Yale School of Medicine, New Haven, CT, United States

Editors-in-Chief

Bin Chen

Department of Molecular, Cell & Developmental Biology University of California, Santa Cruz, Santa Cruz, CA, United States

Kenneth Y. Kwan

Michigan Neuroscience Institute & Department of Human Genetics University of Michigan, Ann Arbor, MI, United States

Section Editors

Hongkui Zeng Allen Institute for Brain Science, Seattle, WA, USA

Helen Tager-Flusberg

Department of Psychological and Brain Sciences & Center for Autism Research Excellence Boston University, Boston, MA, USA

Academic Press is an imprint of Elsevier 125 London Wall, London EC2Y 5AS, United Kingdom 525 B Street, Suite 1650, San Diego, CA 92101, United States 50 Hampshire Street, 5th Floor, Cambridge, MA 02139, United States The Boulevard, Langford Lane, Kidlington, Oxford OX5 1GB, United Kingdom Copyright © 2020 Elsevier Inc. All rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopying, recording, or any information storage and retrieval system, without permission in writing from the publisher. Details on how to seek permission, further information about the Publisher’s permissions policies and our arrangements with organizations such as the Copyright Clearance Center and the Copyright Licensing Agency, can be found at our website: www. elsevier.com/permissions. This book and the individual contributions contained in it are protected under copyright by the Publisher (other than as may be noted herein). Notices Knowledge and best practice in this field are constantly changing. As new research and experience broaden our understanding, changes in research methods, professional practices, or medical treatment may become necessary. Practitioners and researchers must always rely on their own experience and knowledge in evaluating and using any information, methods, compounds, or experiments described herein. In using such information or methods they should be mindful of their own safety and the safety of others, including parties for whom they have a professional responsibility. To the fullest extent of the law, neither the Publisher nor the authors, contributors, or editors, assume any liability for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions, or ideas contained in the material herein. Library of Congress Cataloging-in-Publication Data A catalog record for this book is available from the Library of Congress British Library Cataloguing-in-Publication Data A catalogue record for this book is available from the British Library ISBN: 978-0-12-814411-4 For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Nikki Levy Acquisitions Editor: Natalie Farra Editorial Project Manager: Andrae Akeh Production Project Manager: Surya Narayanan Jayachandran Cover Designer: Mark Rogers Cover Image: Shenfeng Qiu Typeset by TNQ Technologies

Contributors Ariel Aguero, University of Notre Dame, Notre Dame, IN, United States Natacha A. Akshoomoff, Center for Human Development, University of California, San Diego, La Jolla, CA, United States; Department of Psychiatry, University of California, San Diego, La Jolla, CA, United States Fabrice Ango, INM, University of Montpellier, CNRS, INSERM, Montpellier, France Patricia J. Bauer, Department of Psychology, Emory University, Atlanta, GA, United States

Elysia Poggi Davis, Department of Psychology, University of Denver, Denver, CO, United States; Department of Psychiatry and Human Behavior, University of California Irvine, Irvine, CA, United States Jean Decety, Department of Psychology, Department of Psychiatry and Behavioral Neuroscience, The University of Chicago, Chicago, IL, United States; The Child Neurosuite, The University of Chicago, Chicago, IL, United States

L. Bayet, American University, Washington, DC, United States

Jenalee R. Doom, Department of Psychology, University of Denver, Denver, CO, United States; Center for Human Growth and Development, University of Michigan, Ann Arbor, MI, United States

Adriene M. Beltz, University of Michigan, Ann Arbor, MI, United States

Jessica A. Dugan, Department of Psychology, Emory University, Atlanta, GA, United States

Sheri A. Berenbaum, The Pennsylvania State University, University Park, PA, United States

Anne Engmann, Department of Stem Cell and Regenerative Biology, and Center for Brain Science, Harvard University, Cambridge, MA, United States

Stefanie C. Bodison, Chan Division of Occupational Science and Occupational Therapy, University of Southern California (USC), Los Angeles, CA, United States; Keck School of Medicine of USC, Department of Pediatrics, Los Angeles, CA, United States

Daniel E. Feldman, Department of Molecular & Cell Biology, Helen Wills Neuroscience Institute, UC Berkeley, Berkeley, CA, United States Kayla H. Finch, Department of Psychological & Brain Sciences, Boston University, Boston, MA, United States

S.D. Burton, Department of Neurobiology and Anatomy, University of Utah, Salt Lake City, UT, United States; Department of Neurobiology, University of Pittsburgh, Pittsburgh, PA, United States

N.A. Fox, University of Maryland, College Park, MD, United States

G.A. Buzzell, University of Maryland, College Park, MD, United States

Charles R. Gerfen, Intramural Research Program, NIMH, Bethesda, MD, United States

Claire E.J. Cheetham, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States

Aryn H. Gittis, Department of Biological Sciences and Center for the Neural Basis of Cognition, Carnegie Mellon University, Pittsburgh, PA, United States

Hughes Claire, Newnham College, Cambridge University, Cambridge, United Kingdom; Centre for Family Research, Cambridge University, Cambridge, United Kingdom John B. Colby, Department of Radiology and Biomedical Imaging, University of California, San Francisco, CA, United States A. Conejero, Mind, Brain and Behavior Research Center (CIMCYC), University of Granada, Granada, Spain

L.V. Goodrich, Harvard Medical School, Boston, MA, United States Megan R. Gunnar, Institute of Child Development, University of Minnesota, Minneapolis, MN, United States Frank Haist, Center for Human Development, University of California, San Diego, La Jolla, CA, United States; Department of Psychiatry, University of California, San Diego, La Jolla, CA, United States

xv

xvi

Contributors

Richard Hawkes, Department of Cell Biology & Anatomy and Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada

Amanda N. Noroña, Department of Psychology, University of Denver, Denver, CO, United States; Department of Psychiatry, University of Colorado Anschutz Medical Campus, Aurora, CO, United States

Bryan M. Hooks, Department of Neurobiology, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States

Abdulkadir Ozkan, Department of Stem Cell and Regenerative Biology, and Center for Brain Science, Harvard University, Cambridge, MA, United States

Mark H. Johnson, Department of Psychology, University of Cambridge, Cambridge, United Kingdom

Michele Pignatelli, RIKEN-MIT Center for Neural Circuit Genetics at the Picower Institute for Learning and Memory, Department of Biology and Department of Brain and Cognitive Sciences, Massachusetts Institute of Technology, Cambridge, MA, United States; Howard Hughes Medical Institute, Massachusetts Institute of Technology, Cambridge, MA, United States

Scott P. Johnson, University of California, Los Angeles, CA, United States Masanobu Kano, Department of Neurophysiology, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan P.O. Kanold, Johns Hopkins University, Baltimore, MD, United States; University of Maryland, College Park, MD, United States Dominic P. Kelly, University of Michigan, Ann Arbor, MI, United States Taehyeon Kim, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States A. Lahat, University of Toronto, Toronto, ON, Canada Jill Lany, University of Liverpool, Liverpool, United Kingdom G. Lepousez, Perception and Memory Unit, Institut Pasteur, Centre National de la Recherche Scientifique, Paris, France P.-M. Lledo, Perception and Memory Unit, Institut Pasteur, Centre National de la Recherche Scientifique, Paris, France Jeffrey D. Macklis, Department of Stem Cell and Regenerative Biology, and Center for Brain Science, Harvard University, Cambridge, MA, United States; Bauer Laboratory, Cambridge, MA, United States Kalina J. Michalska, Department of Psychology, University of California, Riverside, CA, United States Zoltán Molnár, Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, United Kingdom C.A. Nelson, III, Harvard Medical School, Boston, MA, United States; Boston Children’s Hospital, Boston, MA, United States; Harvard Graduate School of Education, Cambridge, MA, United States

Hilary Richardson, Department of Brain and Cognitive Sciences, MIT, Cambridge, MA, United States Kathleen S. Rockland, Department of Anatomy and Neurobiology, Boston University School of Medicine, Boston, MA, United States Benjamin A. Rowland, Department of Neurobiology & Anatomy, Wake Forest School of Medicine, WinstoneSalem, NC, United States M.R. Rueda, Department of Experimental Psychology, University of Granada, Granada, Spain; Mind, Brain and Behavior Research Center (CIMCYC), University of Granada, Granada, Spain Vibhu Sahni, Department of Stem Cell and Regenerative Biology, and Center for Brain Science, Harvard University, Cambridge, MA, United States; Burke Neurological Institute, Weill Cornell Medicine, White Plains, NY, United States; Feil Family Brain and Mind Research Institute, Weill Cornell Medicine, New York, NY, United States Rebecca Saxe, Department of Brain and Cognitive Sciences, MIT, Cambridge, MA, United States Constantino Sotelo, Sorbonne Universités, UPMC Université Paris 06, INSERM, CNRS, Institut de la Vision Paris, France; Instituto de Neurociencias de Alicante, UMH-CSIC, Universidad Miguel Hernández de Elche, Alicante, Spain Elizabeth R. Sowell, Keck School of Medicine of USC, Department of Pediatrics, Los Angeles, CA, United States; Developmental Cognitive Neuroimaging Laboratory, Children’s Hosiptal, Los Angeles, CA, United States

Contributors

xvii

Terrence R. Stanford, Department of Neurobiology & Anatomy, Wake Forest School of Medicine, WinstoneSalem, NC, United States

Helen Tager-Flusberg, Department of Psychological & Brain Sciences, Boston University, Boston, MA, United States

Barry E. Stein, Department of Neurobiology & Anatomy, Wake Forest School of Medicine, WinstoneSalem, NC, United States

Abbie Thompson, Valparaiso University, Valparaiso, IN, United States

Joan Stiles, Department of Cognitive Science, University of California, San Diego, La Jolla, CA, United States; Center for Human Development, University of California, San Diego, La Jolla, CA, United States

M. Wachowiak, Department of Neurobiology and Anatomy, University of Utah, Salt Lake City, UT, United States Masahiko Watanabe, Department of Anatomy, Hokkaido University Graduate School of Medicine, Sapporo, Japan

Chapter 1

Neural circuits of the mammalian main olfactory bulb S.D. Burton1, 3, G. Lepousez2, P.-M. Lledo2 and M. Wachowiak1 1

Department of Neurobiology and Anatomy, University of Utah, Salt Lake City, UT, United States; 2Perception and Memory Unit, Institut Pasteur,

Centre National de la Recherche Scientifique, Paris, France; 3Department of Neurobiology, University of Pittsburgh, Pittsburgh, PA, United States

Chapter outline 1.1. Introduction 1.2. Synaptic organization of the main olfactory bulb 1.2.1. Organization of sensory inputs 1.2.2. Synaptic microcircuits 1.2.2.1. Glomerular layer microcircuits 1.2.2.2. External plexiform layer microcircuits 1.2.3. Neural computation 1.2.3.1. Contrast enhancement 1.2.3.2. Slow timescale decorrelation 1.2.3.3. Fast timescale synchronization 1.2.3.4. Downstream decoding

3 4 4 6 8 9 11 11 12 13 14

1.2.4. Modulation of sensory processing 1.2.4.1. Local circuits and centrifugal innervation 1.2.4.2. Brain state and context 1.3. Plasticity in the main olfactory bulb 1.3.1. Adult neurogenesis 1.3.1.1. Regeneration of sensory input 1.3.1.2. Adult-born interneurons 1.3.2. Circuit and synaptic plasticity 1.4. Concluding remarks Acknowledgments References

14 14 16 17 17 17 18 20 21 21 21

1.1 Introduction Sensory systems are specialized biological devices by which organisms perceive their external sensory space. The mammalian brain harnesses several sophisticated sensory systems that operate according to a specific set of rules to transform sensory information from one dimension to another. For the chemical senses, such as olfaction, this transformation concerns the ways in which chemical information gives rise to specific neuronal responses in a dedicated sensory organ (Ache and Young, 2005). Several factors make the transformation of olfactory stimuli particularly complex and computationally demanding. For example, odorants (i.e., volatile molecules activating the terrestrial main olfactory system) are inherently high-dimensional, and thus cannot easily be classified along a single dimension (such as frequency for auditory stimuli). Further, each natural odor (i.e., olfactory percept) is typically composed of numerous distinct odorants (e.g., coffee comprises >900 distinct volatile organic compounds (Farah, 2012)) that are nevertheless integrated into a single percept (a process called configural or synthetic perception) (Gottfried, 2010). In addition, olfactory perceptual intensity, of which odorant concentration is only one contributing factor, can vary substantially without changes in perceived odor quality (Mainland et al., 2014). Relatedly, navigating toward an odor source in nature requires sensing and integrating information contained across dynamically fluctuating plumes of high- and low-concentration odorant filaments (Baker et al., 2018). The neural architecture responsible for processing olfactory stimuli must thus harbor profound flexibility and computational power. Olfactory systems and chemosensation more generally have evolved from the earliest known life forms to meet crucial needs such as locating potential food sources, detecting dangers such as predators, and mediating social and sexual interactions (Ache and Young, 2005). Despite these highly conserved functions, interest in other sensory modalities has historically dominated neuroscience, in part due to the comparative ease of manipulating lower dimensional sensory stimuli

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00001-9 Copyright © 2020 Elsevier Inc. All rights reserved.

3

4 PART | I Circuit development

FIGURE 1.1 Centripetal and centrifugal projections of the main olfactory bulb. Schematic depiction of the main centripetal and centrifugal projections of the main olfactory bulb (MOB). The MOB processes sensory input received from the main olfactory epithelium, and transmits information to multiple brain regions that collectively form the olfactory cortex (blue). In turn, several regions of the olfactory cortex, in addition to major neuromodulatory centers of the brain (red), densely innervate the MOB to modulate sensory processing.

such as light, and in part due to the now debunked notion that the human sense of smell is poor or unimportant (Shepherd, 2011; McGann, 2017). Nevertheless, neuroscience has made considerable progress in understanding how the brain perceives odors so precisely, propelled in large part by the discovery in 1991 of a multigene family of odorant-binding G-protein-coupled receptors (GPCRs) that revealed the molecular underpinnings of peripheral odorant recognition (Buck and Axel, 1991). This pivotal discovery - awarded the Nobel Prize in Physiology or Medicine in 2004 - together with the recent explosion in advanced molecular techniques for labeling, monitoring, and perturbing distinct neuron types has yielded an increasingly clear picture of how chemical information is processed throughout the main olfactory system. Below, we review how chemical information is encoded and processed at the first central processing station of the main olfactory system, the main olfactory bulb (MOB) (Fig. 1.1). In addition to the MOB, which processes olfactory information detected by sensory neurons in the main olfactory epithelium, a related structure in many mammals called the accessory olfactory bulb processes pheromonal information detected in the peripheral vomeronasal organ (Mohrhardt et al., 2018). Due to space constraints, however, we focus exclusively on the main olfactory system, with a predominant focus on the rodent experimental preparation and MOB. In the further interest of space, recent comprehensive reviews (in addition to key representative studies) are cited where possible to provide direction for more thorough exploration of topics.

1.2 Synaptic organization of the main olfactory bulb The main olfactory system is responsible for encoding sensory information from thousands to millions of different odorants. To accomplish this complex task, sensory information is processed through distinct units. At each of these units, a modified representation of the sensory information is generated. Following a bottomeup approach, we will start our description from the olfactory sensory organ located in the nasal cavity.

1.2.1 Organization of sensory inputs As our knowledge about the neurobiology of olfaction grows, it is becoming increasingly evident that the main olfactory systems of animals in disparate phyla share many strikingly parallel features. In particular, virtually all olfactory systems require odorant interaction with specific receptors expressed on the dendritic cilia of peripheral sensory neurons; this interaction is transduced by an intracellular second messenger signaling cascade into neural activity, which then propagates

Neural circuits of the mammalian main olfactory bulb Chapter | 1

5

along sensory neuron axons to anatomical structures called glomeruli in the first central processing station of the olfactory system (Ache and Young, 2005). If these common features represent adaptive mechanisms that have evolved independently, then their study will likely bring valuable knowledge about the way the nervous system extracts olfactory information from the environment. In mammals, olfaction begins with the activation of peripheral olfactory sensory neurons (OSNs), which line the main olfactory epithelium (Fig. 1.2A,B). Each OSN of the mouse typically expresses a single odorant-binding receptor type, most of which belong to the w1,000 functional GPCR odorant receptor (OR) types first characterized by Buck and Axel. These ORs evolutionarily subdivide into the fish-like Class I ORs and terrestrial-like Class II ORs (Mori and Sakano, 2011; Bear et al., 2016). In addition to these classical ORs, recent research has further uncovered the trace amine-associated receptors (TAARs), a second class of odorant-binding GPCRs that, while few in number (mice express 15 functional TAAR types), interact with volatile and ethologically relevant amines capable of triggering innate behavioral responses (Liberles and Buck, 2006). Each OR and TAAR interacts with a specific subset of odorants and, as with any molecular receptor, these interactions are governed in a concentration-dependent manner according to receptor/ligand binding affinities. Increasing odorant concentrations therefore not only increase activation of receptors highly sensitive to those odorants, but also activate additional receptors less sensitive to those odorants (Mainland et al., 2014). All OSNs expressing the same OR or TAAR project their axons centrally to one (or sometimes a few) glomeruli in each medial and lateral half of the ipsilateral MOB, forming roughly mirror-symmetric glomerular maps within each MOB (Fig. 1.2AeE) (Mori and Sakano, 2011; Liberles, 2015). Each glomerulus is a large spherical neuropil structure (w100 mm diameter in mice) wherein OSNs release glutamate to activate diverse neuron types. Sensory information propagating from the peripheral epithelium to a glomerulus is thereby processed by multiple local circuit interactions before being transmitted to higher brain centers via the MOB projection neurons (Fig. 1.1). The coalescence of OSN axons into a glomerulus is coordinated by a host of guidance cues and molecular interactions, including the odorant-binding receptor itself, and represents one of the most exquisitely specific anatomical substrates in the brain (Mori and Sakano, 2011).

FIGURE 1.2 Glomerular organization of sensory input to the main olfactory bulb. (A) Sagittal whole mount view of the medial glomerulus formed by OSNs expressing the P2 odorant receptor (OR). Dashed line: MOB outline. Arrowhead: glomerulus. (B,C) Magnification of the OSNs (B) and glomerulus (C) in (A). Inset: magnification of two fluorescently labeled OSNs in the main olfactory epithelium. Arrowhead: dendrites. Arrow: axons. (D) Dorsal whole mount view of medial and lateral glomeruli formed by OSNs expressing trace amine-associated receptor (TAAR) 3 (green) and TAAR4 (red). Dashed line: bilateral MOB outlines. Arrowhead and arrow: lateral and medial glomeruli, respectively, of left MOB. (E) Magnification of the boxed region in (D), showing the precise convergence of thousands of axons to neighboring glomeruli. (F) Dorsal whole mount view of glomeruli formed by OSNs expressing Class I ORs (yellow), Class II ORs (red), and TAARs (cyan), forming Domain I, Domain II, and the TAAR Domain of the dorsal MOB. (G) Schematic dorsolateral view of the domain organization of glomeruli in the MOB. Dashed black line: approximate border between the dorsal and ventral MOB. Dashed gray line: accessory olfactory bulb outline. D, dorsal; M, medial; P, posterior. (AeC) Adapted from Mombaerts, P., Wang, F., Dulac, C., Chao, S.K., Nemes, A., Mendelsohn, M., Edmondson, J., Axel, R., 1996. Visualizing an olfactory sensory map. Cell 87, 675e686, with permission; (B inset, DeF) Adapted from Pacifico, R., Dewan, A., Cawley, D., Guo, C., Bozza, T., 2012. An olfactory subsystem that mediates highsensitivity detection of volatile amines. Cell Rep. 2, 76e88, with permission; (G) Adapted from Bear, D.M., Lassance, J.M., Hoekstra, H.E., Datta, S.R., 2016. The evolving neural and genetic architecture of vertebrate olfaction. Curr. Biol. 26, R1039eR1049, with permission.

6 PART | I Circuit development

Each glomerulus is formed by the axonal convergence of w2,000e40,000 OSNs (Fig. 1.2AeE) (Bressel et al., 2016) onto the apical dendrites of w30 projection neurons on average (Schwarz et al., 2018), yielding a convergence ratio of 102e3:1. The convergence of so many OSNs is thought to not only broaden the dynamic range of the net sensory input to each glomerulus, but may also allow each MOB projection neuron to integrate input from numerous OSNs, heightening signalto-noise ratios and ensuring the detection of even faint sensory input (Mainland et al., 2014). OSNs expressing each OR or TAAR are randomly distributed within one of a few dorsoventral zones of the main olfactory epithelium, and this zonal distribution is conserved in the MOB glomerular map (Fig. 1.2F,G). OSNs in the ventral epithelium express Class II ORs and project their axons to glomeruli in the ventral MOB. In dorsal zones of the epithelium, intermixed OSNs express Class I ORs, Class II ORs, and TAARs, and interestingly segregate their axonal projections to glomeruli in the anterodorsal Domain I, posterodorsal Domain II, and mediodorsal TAAR Domain of the MOB, respectively (Mori and Sakano, 2011; Pacifico et al., 2012; Liberles, 2015). Within these domains, glomerular positions are roughly conserved between bilateral MOBs, across animals, and even across species (Soucy et al., 2009). A striking exception to these general properties of the main olfactory system is the recently discovered family of membrane-spanning, 4-pass A odorant-binding receptors (MS4ARs) (Bear et al., 2016; Greer et al., 2016). In contrast to ORs and TAARs, each of the 12 functional MS4AR types in mice is a non-GPCR that transduces odorant binding through distinct and still unknown signaling cascades. Further, while each OR and TAAR type is expressed singularly by OSNs randomly distributed throughout one of the dorsoventral epithelial zones, MS4AR expression is very specifically localized to OSNs within the epithelial recesses (or “cul-de-sacs”), where each OSN further expresses multiple different MS4AR types. Axons of the MS4AR-expressing OSNs selectively terminate in the few dozen necklace glomeruli of the posterior MOB, which anatomically ring the accessory olfactory bulb (Fig. 1.2G). Like TAARs (and like odorant-binding receptors within olfactory subsystems outside of the main olfactory system), MS4AR types are few in number and yet critically involved in driving specific olfactory-guided behaviors (Munger et al., 2009). As each odorant-binding receptor (referred to below as olfactory receptor) responds to a specific set of odorants and is expressed by OSNs projecting to conserved domains and approximate positions in the MOB, the glomerular layer (GL) forms an approximate two-dimensional anatomical representation or map of the receptor repertoire (Wachowiak and Shipley, 2006; Mori and Sakano, 2011). Upon binding in a concentration-dependent manner with often multiple olfactory receptor types, odorants are thus first encoded as sensory information in the main olfactory system by the combinatorial map of OSN activation and glutamate release within MOB glomeruli. These input maps directly reflect the anatomical domain organization of the MOB; for example, distinct acid and ketone odorants activate Class I and Class II ORs and evoke OSN activation and input to glomeruli within Domains I and II, respectively (Bozza et al., 2009). However, whether there exists a topographical relationship between glomerular position and OSN tuning to specific odorant physicochemical properties at scales finer than the level of gross domains remains unclear (Soucy et al., 2009; Ma et al., 2012; Chae et al., 2019). This lack of obvious fine-scale topography contrasts with the direct topographical mapping of stimulus properties to neural space in other sensory systems, but is perhaps unsurprising given the high dimensionality of olfactory stimuli (Cleland, 2010). Irrespective of the degree of topographical organization, however, the precise map of OSN input directly impacts olfactory processing and perception: odorants evoking more similar maps are more difficult to perceptually discriminate (Linster et al., 2001). Beyond spatial patterns of sensory input, olfactory information is also encoded by the temporal pattern of OSN activity, which is heavily sculpted by the active sampling of odorants via repetitive sniffing. Indeed, rather than simple static maps of combinatorial OSN activation and input to glomeruli, odorants trigger bursts of activity in OSNs with onset latencies distributed throughout each sniff (Wachowiak, 2011). Such sniff-driven pacing not only emerges through chemosensation of odorants by olfactory receptors, but further arises through mechanosensory activation of ORs by nasal airflow (Grosmaitre et al., 2007; Chen et al., 2012; Connelly et al., 2015). These chemo- and mechanosensory temporal patterns of OSN activity are then propagated to downstream MOB circuits such that neural activity is broadly distributed throughout the duration of each sniff (Shusterman et al., 2011; Iwata et al., 2017), providing a unique temporal framework for the processing of sensory information (Wachowiak, 2011). Collectively, the spatiotemporal glomerular patterns of sensory input that are driven and shaped by sniffing are then modulated on both intra- and interglomerular scales by local circuits within the MOB, further increasing the coding and consequent perceptual capacity of the main olfactory system.

1.2.2 Synaptic microcircuits Because of its laminar organization and accessible location in rodents, the MOB is an ideal model system for investigating the principles underlying network processing of sensory information. With the application of in vitro slice recordings, together with recent advances in molecular techniques for labeling, monitoring, and perturbing distinct neuron types within

Neural circuits of the mammalian main olfactory bulb Chapter | 1

7

the MOB (Burton, 2017), the complex processing of sensory information by local circuits is beginning to be revealed. Of particular note, and distinct from most other brain regions, dendrodendritic synapses constitute a principal mode of signaling within MOB microcircuits (Schoppa and Urban, 2003; Egger and Urban, 2006). OSN axons project to the surface of the MOB, where they intertwine as olfactory receptorespecific (i.e., homotypic) fascicles within the olfactory nerve layer (ONL) before converging onto individual glomeruli within the deeper GL (Fig. 1.3AeC). Within each glomerulus, OSN axon terminals release glutamate with high probability (Murphy et al., 2004) to excite the apical dendritic tufts of the MOB projection neurons (Najac et al., 2011; Gire et al., 2012; Bourne and Schoppa, 2017). Each mature projection neuron emits a single apical dendrite terminating in a single expansive and elaborate tuft, and therefore receives sensory input within only a single glomerulus (Fig. 1.3C,D). The map of sensory

FIGURE 1.3 Organization of the main olfactory bulb circuit. (A) Nissl-stained coronal section of the mouse main olfactory bulb (MOB) showing the different concentric layers. (B) Magnification and rotation of the boxed region in (A). (C) Laminar distribution of the major MOB neuron types (black: glutamatergic neurons; red: anaxonic GABAergic neurons; blue: axon-bearing GABAergic neurons). Arrangement of layers directly corresponds to the Nissl-stained section in (B). dGC, deep granule cell; EPL, external plexiform layer; EPL-dSAC, external plexiform layer-projecting deep short-axon cell; EPL-IN, external plexiform layer-localized interneuron; ETC, external tufted cell; GCL, granule cell layer; GCL-dSAC, granule cell layer-projecting deep short-axon cell; GL, glomerular layer; GL-dSAC, glomerular layer-projecting deep short-axon cell; IPL, internal plexiform layer; LOT, lateral olfactory tract; MC, mitral cell; MCL, mitral cell layer; ONL, olfactory nerve layer; OSN, olfactory sensory neuron; PGC, periglomerular cell; sGC, superficial granule cell; sSAC, superficial short-axon cell; TC, tufted cell. (D) Somatodendritic and partial axonal reconstructions of MCs (gray), TCs (black), and an ETC (magenta) from acute slice recordings. Each cell emits a single apical dendrite to terminate as a tuft within a spherical glomerulus, while MCs and TCs further emit extensive lateral dendrites within the deep and superficial EPL, respectively. (E) Reconstructions of deep GCs (light red) and superficial GCs (dark red) from acute slice recordings. Deep and superficial GC apical dendrites preferentially display gemmules in the deep and superficial EPL, respectively. Dashed lines: superficial and deep borders of the EPL. (A,B) Images courtesy of the Allen Mouse Brain Coronal Reference Atlas. © Allen Institute for Brain Science [Available from: http://atlas.brain-map.org/atlas?atlas¼1&plate¼100960472#atlas¼1&plate¼100960472&resolution¼8. 21&x¼2400.006223192402&y¼2207.9237059050915&zoom¼-2].

8 PART | I Circuit development

input evoked by each odorant is thus transmitted from OSNs to the projection neurons. The MOB does not simply serve as a relay of peripheral sensory input, however. Each glomerulus is surrounded by hundreds of diverse juxtaglomerular interneurons that modulate the transmission of sensory input to the projection neurons. Further, each projection neuron emits two to five lateral dendrites that extend up to w1 mm throughout the external plexiform layer (EPL) beneath the GL (Fig. 1.3C,D), where they interact with a distinct set of interneurons to recurrently influence their own activity and laterally influence the activity of projection neurons connected to other glomeruli (and potentially other projection neurons connected to the same glomerulus). The projection neurons are thus the backbone of two serial intrabulbar circuits: one targeting apical dendrites in the GL on a predominantly intraglomerular scale and one targeting lateral dendrites in the EPL on a predominantly interglomerular scale (Schoppa and Urban, 2003; Burton, 2017). Both GL and EPL circuits are further heavily modulated by additional intrabulbar local circuits, as well as profuse centrifugal innervation from cortical and neuromodulatory areas. All of these circuits together dynamically transform sensory information received from the OSNs before and as it is transmitted to downstream brain regions. Pioneering investigations by Ramón y Cajal divided the MOB projection neurons into two types based on their laminar differences (Fig. 1.3C,D) (Ramón y Cajal, 1955). Mitral cells (MCs) have somata located in the compact mitral cell layer (MCL) beneath the EPL, extend their lateral dendrites throughout the deep half of the EPL, and extend axon collaterals only throughout the deepest portions of the MOB. In contrast, tufted cells (TCs) have somata located throughout the EPL, extend their lateral dendrites throughout the superficial half of the EPL, and extend axon collaterals within the compact internal plexiform layer (IPL) directly beneath the MCL. These laminar differences, while seemingly subtle, have critical implications for the involvement of MCs and TCs in distinct local circuits, which contribute to distinct patterns of MC and TC activity and sensory processing (Macrides et al., 1985; Geramita et al., 2016). Beyond laminar differences within the MOB, MCs and TCs further project their axons out of the MOB via the lateral olfactory tract in complementary patterns, forming parallel pathways of MOB output. Specifically, MCs project broadly across both anterior and posterior olfactory cortical regions; in contrast, among TCs, only deeply-positioned TCs project to posterior olfactory cortical regions, and there extend only short distances from the lateral olfactory tract, while TCs located more superficially in the EPL only target anterior olfactory cortical regions (Macrides et al., 1985; Mori and Sakano, 2011). These complementary patterns of axonal projections impose important constraints on how sensory-evoked activity is propagated to downstream brain regions. For example, lower odorant concentration response thresholds are observed more anteriorly in olfactory cortex, where TC projections are concentrated (Wilson et al., 2006; Mainland et al., 2014).

1.2.2.1 Glomerular layer microcircuits The GL spans the surface of the MOB and constitutes the first site of integration for sensory information in the main olfactory system. In total, there are w2,000e3,000 glomeruli in the rodent MOB, arranged one-to-a-few glomeruli deep throughout the GL (Fig. 1.3A,B). Juxtaglomerular interneurons surrounding each glomerulus can be subdivided into three main neuron types: (1) periglomerular cells (PGCs), (2) superficial short-axon cells (sSACs), and (3) external tufted cells (ETCs) (Fig. 1.3C). Of these, ETCs and most PGCs have dendrites extending within only a single glomerulusdsimilar to the apical dendrites of M/TCsdwhile sSACs extend sparsely branched dendrites throughout the juxtaglomerular space (Wachowiak and Shipley, 2006; Burton, 2017). PGCs constitute a morphologically, neurochemically, and physiologically heterogeneous population of interneurons. In general, however, PGCs exhibit extremely small somata with no axon and a single small dendritic tuft extending throughout a fraction of a glomerulus. Through dendritic release of GABA onto the apical dendrites of M/TCs, PGCs provide the main source of inhibition in the GL, and mediate both feedforward and recurrent M/TC inhibition (Wachowiak and Shipley, 2006; Burton, 2017). The precise timing, strength, and balance of feedforward versus recurrent M/TC inhibition strongly depend on PGC subtype: non-overlapping populations of PGCs expressing calcium-binding proteins calbindin versus calretinin exhibit strong versus weak sensory-evoked activity and M/TC connectivity, respectively (Najac et al., 2015; Benito et al., 2018). A third PGC subtype synthesizes dopamine in addition to GABA (Kosaka and Kosaka, 2008; Galliano et al., 2018) and thus may modulate M/TC activity (and OSN activitydsee below) on distinct timescales through ionotropic and metabotropic signaling. Collectively, PGC signaling within the glomerulus regulates the strength (and timing) of sensory signals as they propagate from OSNs to M/TCs, and through low sensory input thresholds, can suppress or filter out the transmission of weak sensory inputs to sculpt M/TC odorant tuning (Cleland, 2010). In addition, stronger PGC-mediated feedforward inhibition onto MCs than TCs (Geramita and Urban, 2017), together with stronger OSN input to TCs and greater TC excitability, can differentially modulate MC versus TC activity patterns to support the encoding of complementary information along parallel MC versus TC pathways (Mainland et al., 2014). For example, GL-localized inhibition (mediated

Neural circuits of the mammalian main olfactory bulb Chapter | 1

9

at least in part by PGCs) specifically delays sensory-evoked MC (but not TC) firing in a concentration-dependent manner, enabling MC latency to encode information about odorant concentration while TC activity signals olfactory receptor activation (and thereby odorant identity) (Fukunaga et al., 2014; Mainland et al., 2014). Similar to the subset of combined dopaminergic/GABAergic PGCs, sSACs are also simultaneously dopaminergic and GABAergic, with recent estimates of sSACs forming P75) in an enhanced acoustic environments (EAE) using multitone pips resulted in a decreased representation of the exposed sounds and increased representation of sounds that were not part of the EAE (Eggermont, 2013; Noreña et al., 2006) (Fig. 2.12C). The difference between these two results is likely due to the nature of the presented sounds (single tone vs. multitone pips), the maturational state of the cortex, e.g., state of intracortical inhibition, and the attentional state of the animal during the exposure (Eggermont, 2013). Whereas rearing animals in fixed frequency tone environments induces a region-specific reallocation of territory within the cortical map and likely underlying changes in the balance of excitation and inhibition to single cells, rearing rodents in

50

PART | I Circuit development

FIGURE 2.13 Experience-dependent map reorganization in the developing auditory cortex. Schematics of tonotopic maps reconstructed from microelectrode recordings of tone-evoked action potentials at approximately 100 different points within A1 (indicated by the color of each dot), with example tuning curves shown below. (A) Normal adult rodents display a smooth and orderly tonotopic gradient. A1 neurons are tuned to the same frequency for sounds presented to each ear, yet responses to the contralateral ear are more robust than the ipsilateral ear. (B) Rodents that have been reared in acoustic environments dominated by presentation of a single middle frequency tone (represented by green colors) show an overrepresentation of that tone frequency within the tonotopic map such that recording sites at distant points within the map have similar tuning. Frequency response areas (FRAs) below illustrate tuning preferences. (C) Cats that have been reared from P75 on in acoustic environments enriched by multitone stimuli (frequency range is illustrated by the color gradient from yellow to green) show a reduced presentation of middle frequencies (e.g., yellow and green colors) and an overrepresentation of tone frequencies bordering the exposure region (e.g., red and blue). (D) In rodents that were reared in continuous white noise, the tonotopic map is disrupted, featuring numerous recording sites with abnormally broad frequency tuning (represented by gray dots). (E) Reversibly closing the contralateral ear during critical periods of development is associated with an overrepresentation of low-frequency tones, the loss of contralateral tuning at several points (open circles), and an enhancement of tuning strength to tones delivered to the developmentally unobstructed ipsilateral ear.

broad-spectrum noise is associated with a more global disruption of ACtx tonotopy and an abundance of cells with abnormally broad tuning, as revealed by electrophysiological recordings (Fig. 2.13D). Similar to the extended visual cortex critical periods observed in animals reared in total darkness, the critical period for the effects of single tone exposure is abnormally prolonged in animals reared in white noise. Thus, deprivation of patterned acoustic input postpones the onset of the molecular program that normally limits the critical period, and the effects of singe tone exposure can distort the map many weeks after the critical period would normally have closed (Chang and Merzenich, 2003). That a critical period normally observed in infancy can be delayed into adulthood supports the possibility that progressive development of auditory feature representation in ACtx may be more closely linked to an experiential timeline, rather than a strict chronological timeline. A parallel can also be drawn between the reversible lid suture method used extensively in studies of visual cortex development and the effects of reversible ear canal ligation on ACtx and IC development. Ligating the ear canal temporarily interferes with the transmission of external acoustic signals to the middle ear, and ultimately the brain, particularly at high frequencies. By reopening the ear canal prior to recording from the contralateral AI, researchers have observed a degraded tonotopic map populated by weak high-threshold responses to contralateral tones and the occasional absence of contralateral tuning in higher-frequency regions of the map (Popescu and Polley, 2010). By contrast to the contralateral bias observed in normal animals (Fig. 2.13A), the quality of tuning for stimuli delivered to the ipsilateral ear is often superior to the normally dominant contralateral ear (Fig. 2.13E). Interestingly, ipsilateral inputs were only enhanced when the ear canal was ligated in infant and juvenile animals, but not in adulthood. Collectively, these results indicate that the allocation of representational resources within the ACtx is quite dynamic and can be adaptively reassigned to the inputs that are most prominent during critical periods of early postnatal development.

Functional circuit development in the auditory system Chapter | 2

51

2.4.5 Developmental regulation over reinstating hearing in the deaf Synaptic transmission studies in the brain slice preparation have uncovered some fundamental changes in the intrinsic and synaptic properties of neurons after sensory loss, such as which receptors are expressed. It is likely that neurons all along the auditory processing stream will show intrinsic and synaptic changes and that as such the transmission of activity from the periphery will be dramatically altered. Thus, these studies help us to identify the complications and possibilities associated with reinstating hearing in deaf individuals. Unlike birds and other nonmammalian vertebrates, which can regrow hair cells throughout life, mammals are born with all the cochlear hair cells they will ever have. Nevertheless, hearing is a possibility for profoundly deaf individuals through the use of the cochlear implant (CI), a neural prosthetic device that bypasses the dysfunctional transduction machinery within the cochlea and reinstates afferent signals through direct electrical stimulation of auditory nerve fibers. As of 2012, more than 300,000 individuals have been fitted with cochlear implants, and currently, w45,000 CIs are sold annually. It was discovered early on that the age of surgical implantation plays a crucial role in the quality of hearing experienced by cochlear implant users. While postlingually deaf individuals often recover acceptable hearing and speech recognition whether they are implanted as children or adults, congenitally deaf individuals stand the best chance of experiencing the full benefit of the CI if they undergo the implantation procedure at an early age, typically by the time they are 7 years old (Dorman et al., 2007). Thus, understanding how the processing of peripheral signals is altered in the deaf brain generates important insights to improve the performance of CIs. Through careful study of cochlear implants in a special breed of congenitally deaf cats, researchers have begun to understand how auditory brain areas represent signals delivered through the cochlear implant and the manner by which these representations are shaped through development and experience. While most synaptic studies utilized rodents, the advantage of this cat model is that cats can be fitted with CIs due to their size, plus hearing range overlaps significantly with the human hearing range. As an experimental control for the developmental changes observed in the genetically deaf cats, normally hearing cats are acutely deafened with an ototoxic drug and immediately fit with a cochlear implant. Neural recordings are then made from the ACtx of acutely or congenitally deaf cats at various ages in response to brief electrical pulses delivered by the CI to the auditory nerve (Fig. 2.14A). A comparison of activation patterns across development in acutely deafened cats reveals an exuberant spatial spread of neural activity across the ACtx in young kittens that is culled to a topographically restricted activation area by 3 months of age (Fig. 2.14B). By contrast, activating the implant in congenitally deaf kittens at 1.5 months evokes a weak cortical response. At 3 months postnatal, deaf cats show the exuberant activation patterns comparable with normally hearing kittens at 1.5 months, and these activation areas are not consolidated until early adulthood (Kral et al., 2005). The precise mechanisms of altered activation patterns are unclear and likely involve multiple synaptic changes across the auditory pathway. Nevertheless, the hyperexcitability of the ACtx in congenitally deaf cats at 3 months (Fig. 2.14B) is consistent with and may stem from the diminished synaptic inhibition and augmented synaptic excitation observed in brain slices of rodents that undergo cochlear ablation in infancy (Fig. 2.11). Although the mechanisms governing the recovery of hearing in cochlear implant users remain unclear, one initial clue has been found through analyzing the endbulb of Held synapse in deaf cats that began hearing through chronic use of the cochlear implant at a young age. As described previously, the endbulb synapse fails to develop normally in the absence of cochlear signaling. Strikingly, the endbulb synapse from cats using the cochlear implant for 3 months beginning at an early age was largely indistinguishable from normally hearing cats (Ryugo et al., 2005). Therefore, reinstating afferent activity to the central auditory system in early life can rescue the progressive synaptic degradation observed at several levels of the central auditory system.

FIGURE 2.14 Afferent regulation of auditory cortex recruitment by peripheral inputs. (A) Lateral surface of the cat brain. Gray denotes the total area of auditory cortex in normal cats. Black square denotes area from which neurophysiological signals are measured in panel. Auditory nerve fibers are activated electrically by a cochlear implant (CI). C, caudal; D, dorsal; R, rostral; V, ventral. (B) The areal extent of neural responses in auditory cortex (ACtx) evoked by activation of a stimulating electrode implanted proximal to the auditory nerve fibers in control (left) or acutely deafened (right) cats.

52

PART | I Circuit development

2.4.6 Conclusions and directions for future research While sound-evoked patterns of afferent action potentials can influence both the brain stem and the IC, most evidence for the profound influence of sensory activity on the emergence of mature auditory processing comes from ACtx, possibly due to a more intense focus of study. Eliminating hearing or even perturbing the normal balance of signals between the ears or across various frequency channels can dramatically alter auditory signal processing. This plasticity has been documented at multiple levels of analysis, ranging from the synaptic transmission between two neurons up to the coordinated arrangement of frequency tuning across hundreds of thousands of neurons. Although circuit reorganization at higher levels of the auditory system is striking, additional information on the normal course of development in the IC, MG, and ACtx as well as a deeper mechanistic understanding about the contributions of transient microcircuits such as the SPN to circuit assembly prior to ear opening will greatly aid efforts to understand how the stable networks that underlie normal hearing are constructed. Absent data that bridge the gap between synapses and neural networks in normally developing animals, it is difficult to know whether activity- or experience-dependent modifications derail normal development or introduce an abnormal outcome following the conclusion of normal maturation. While current evidence suggests that ACtx might be more malleable by early sensory experience than brain stem circuits, changes in brain stem and midbrain might have been overlooked. Given the large amount of descending connections from ACtx to IC and brain stem, it is reasonable to imagine that the entire bidirectionally connected system is sculpted by afferent input. Thus, future studies might reveal a much more malleable brain stem and midbrain circuitry than currently appreciated. Moreover, to date, very little is known about the function and assembly of descending pathways. Importantly, common manipulations to study experience-dependent development include deafening or sound exposure, and the interpretation of the results has assumed that animals are passive. However, loss of a sensory modality or listening in challenging conditions can lead to behavioral changes, e.g., paying more attention to some sound features but not others, and this change of behavior might contribute to the observed functional changes. Attentional changes can influence processing across the whole ascending auditory pathway via extensive top-down connection and thus might play a powerful role. However, the role of such influences during development remains to be elucidated. The past decades of auditory developmental research have revealed a great deal about the interplay between intrinsic molecular events and dynamic electrical signaling in functional circuit maturation. The scope of research in the coming years may widen to include other biological factors that help to shape functional circuits and how these circuits operate together to represent relevant information. Although neurons are the central players in functional circuits, they do not operate in isolation. Networks of developing nonsensory glial cells, cells that form the developing vasculature and the molecules that form the extracellular matrix may play key roles in modulating neural signaling and providing a physical substrate for growth. As we have seen in the cochlea, the nonsensory cells that make up Kölliker’s organ may act as the catalyst for electrical signaling throughout the prehearing auditory system, and it is probable that exciting new connections between neural development and the nonneural cells and molecules that support them remain to be discovered. Moreover, at early ages and even in adults, there is an array of interaction between sensory systems, and thus, their developmental trajectories might influence each other. On the other end of the spectrum, it will be important to link the observations made at a neurobiological level to their potential behavioral consequences measured at the level of the whole organism and also causally link behavioral changes during development to circuit changes. Demonstrations of abnormal circuit connectivity or map organization along the auditory pathway are important but take on additional importance when they can be causally associated with changes in perceptual abilities. Studies that bridge the gaps between these various levels of analysis and stations of processing along the auditory pathway as well as the rest of the brain promise to reveal more about the etiology and therapeutic possibilities for treating hearing impairment and may teach us a great deal about the fundamental principles of neural development.

References Bandyopadhyay, S., Shamma, S.A., Kanold, P.O., 2010. Dichotomy of functional organization in the mouse auditory cortex. Nat. Neurosci. 13 (3), 361e368. https://doi.org/10.1038/nn.2490. Barr-Gillespie, P.-G., 2015. Assembly of hair bundles, an amazing problem for cell biology. Mol. Biol. Cell 26 (15), 2727e2732. https://doi.org/10.1091/ mbc.E14-04-0940. Basch, M.L., Brown, R.M., Jen, H.-I., Groves, A.K., 2016. Where hearing starts: the development of the mammalian cochlea. J. Anat. 228 (2), 233e254. https://doi.org/10.1111/joa.12314. Born, D.E., Rubel, E.W., 1988. Afferent influences on brain stem auditory nuclei of the chicken: presynaptic action potentials regulate protein synthesis in nucleus magnocellularis neurons. J. Neurosci. 8 (3), 901e919. https://doi.org/10.1523/JNEUROSCI.08-03-00901.1988.

Functional circuit development in the auditory system Chapter | 2

53

Brainard, M.S., Doupe, A.J., 2013. Translating birdsong: songbirds as a model for basic and applied medical research. Annu. Rev. Neurosci. 36 (1), 489e517. https://doi.org/10.1146/annurev-neuro-060909-152826. Chang, E.F., Merzenich, M.M., 2003. Environmental noise retards auditory cortical development. Science 300 (5618), 498e502. https://doi.org/10.1126/ science.1082163. Clause, A., Kim, G., Sonntag, M., Weisz, C.J.C., Vetter, D.E., R}ubsamen, R., Kandler, K., 2014. The precise temporal pattern of prehearing spontaneous activity is necessary for tonotopic map refinement. Neuron 82 (4), 822e835. https://doi.org/10.1016/j.neuron.2014.04.001. Cramer, K.S., 2005. Eph proteins and the assembly of auditory circuits. Hear. Res. 206 (1e2), 42e51. https://doi.org/10.1016/J.HEARES.2004.11.024. Dallos, P., Zheng, J., Cheatham, M.A., 2006. Prestin and the cochlear amplifier. J. Physiol. 576 (1), 37e42. https://doi.org/10.1113/jphysiol.2006.114652. de Villers-Sidani, E., Chang, E.F., Bao, S., Merzenich, M.M., 2007. Critical period window for spectral tuning defined in the primary auditory cortex (A1) in the rat. J. Neurosci. 27 (1), 180e189. https://doi.org/10.1523/JNEUROSCI.3227-06.2007. Deitch, J.S., Rubel, E.W., 1984. Afferent influences on brain stem auditory nuclei of the chicken: time course and specificity of dendritic atrophy following deafferentation. J. Comp. Neurol. 229 (1), 66e79. https://doi.org/10.1002/cne.902290106. Deng, R., Kao, J.P.Y., Kanold, P.O., 2017. Distinct translaminar glutamatergic circuits to GABAergic interneurons in the neonatal auditory cortex. Cell Rep. 19 (6), 1141e1150. https://doi.org/10.1016/j.celrep.2017.04.044. Dorman, M.F., Sharma, A., Gilley, P., Martin, K., Roland, P., 2007. Central auditory development: evidence from CAEP measurements in children fit with cochlear implants. J. Commun. Disord. 40 (4), 284e294. https://doi.org/10.1016/j.jcomdis.2007.03.007. Druckenbrod, N.R., Goodrich, L.V., 2015. Sequential retraction segregates SGN processes during target selection in the cochlea. J. Neurosci. 35 (49) https://doi.org/10.1523/JNEUROSCI.2236-15.2015. Echteler, S.M., 1992. Developmental segregation in the afferent projections to mammalian auditory hair cells. Proc. Natl. Acad. Sci. U.S.A. 89 (14), 6324e6327. Echteler, S.M., Arjmand, E., Dallos, P., 1989. Developmental alterations in the frequency map of the mammalian cochlea. Nature 341 (6238), 147e149. https://doi.org/10.1038/341147a0. Eggermont, J.J., 2013. On the similarities and differences of non-traumatic sound exposure during the critical period and in adulthood. Front. Syst. Neurosci. 7, 12. https://doi.org/10.3389/fnsys.2013.00012. Frank, M.M., Goodrich, L.V., 2018. Talking back: development of the olivocochlear efferent system. Wiley Interdiscip. Rev. Dev. Biol. 7 (6), e324. https://doi.org/10.1002/wdev.324. Geal-Dor, M., Freeman, S., Li, G., Sohmer, H., 1993. Development of hearing in neonatal rats: air and bone conducted ABR thresholds. Hear. Res. 69 (1e2), 236e242. https://doi.org/10.1016/0378-5955(93)90113-F. Géléoc, G.S.G., Holt, J.R., 2003. Developmental acquisition of sensory transduction in hair cells of the mouse inner ear. Nat. Neurosci. 6 (10), 1019e1020. https://doi.org/10.1038/nn1120. Harris, J.A., Rubel, E.W., 2006. Afferent regulation of neuron number in the cochlear nucleus: cellular and molecular analyses of a critical period. Hear. Res. 216e217, 127e137. https://doi.org/10.1016/J.HEARES.2006.03.016. Hoerder-Suabedissen, A., Hayashi, S., Upton, L., Nolan, Z., Casas-Torremocha, D., Grant, E., et al., 2018. Subset of cortical layer 6b neurons selectively innervates higher order thalamic nuclei in mice. Cerebr. Cortex 28 (5), 1882e1897. https://doi.org/10.1093/cercor/bhy036. Hoffpauir, B.K., Grimes, J.L., Mathers, P.H., Spirou, G.A., 2006. Synaptogenesis of the calyx of held: rapid onset of function and one-to-one morphological innervation. J. Neurosci. 26 (20), 5511e5523. https://doi.org/10.1523/JNEUROSCI.5525-05.2006. Holcomb, P.S., Hoffpauir, B.K., Hoyson, M.C., Jackson, D.R., Deerinck, T.J., Marrs, G.S., et al., 2013. Synaptic inputs compete during rapid formation of the calyx of Held: a new model system for neural development. J. Neurosci. 33 (32), 12954e12969. https://doi.org/10.1523/JNEUROSCI.108713.2013. Huang, L.-C., Thorne, P.R., Housley, G.D., Montgomery, J.M., Reichardt, L.F., Xiang, M., 2007. Spatiotemporal definition of neurite outgrowth, refinement and retraction in the developing mouse cochlea. Development 134 (16), 2925e2933. https://doi.org/10.1242/dev.001925. Hudspeth, A.J., 2008. Making an effort to listen: mechanical amplification in the ear. Neuron 59 (4), 530e545. https://doi.org/10.1016/ j.neuron.2008.07.012. Jackson, H., Hackett, J.T., Rubel, E.W., 1982. Organization and development of brain stem auditory nuclei in the chick: ontogeny of postsynaptic responses. J. Comp. Neurol. 210 (1), 80e86. https://doi.org/10.1002/cne.902100109. Johnson, S.L., Eckrich, T., Kuhn, S., Zampini, V., Franz, C., Ranatunga, K.M., et al., 2011. Position-dependent patterning of spontaneous action potentials in immature cochlear inner hair cells. Nat. Neurosci. 14 (6), 711e717. https://doi.org/10.1038/nn.2803. Kanold, P.O., Luhmann, H.J., 2010. The subplate and early cortical circuits. Annu. Rev. Neurosci. 33 (1), 23e48. https://doi.org/10.1146/annurev-neuro060909-153244. Kawashima, Y., Géléoc, G.S.G., Kurima, K., Labay, V., Lelli, A., Asai, Y., et al., 2011. Mechanotransduction in mouse inner ear hair cells requires transmembrane channel-like genes. J. Clin. Investig. 121 (12), 4796e4809. https://doi.org/10.1172/JCI60405. Kim, G., Kandler, K., 2003. Elimination and strengthening of glycinergic/GABAergic connections during tonotopic map formation. Nat. Neurosci. 6 (3), 282e290. https://doi.org/10.1038/nn1015. Kotak, V.C., Fujisawa, S., Lee, F.A., Karthikeyan, O., Aoki, C., Sanes, D.H., 2005. Hearing loss raises excitability in the auditory cortex. J. Neurosci. 25 (15), 3908e3918. https://doi.org/10.1523/JNEUROSCI.5169-04.2005. Kotak, V.C., Takesian, A.E., Sanes, D.H., 2008. Hearing loss prevents the maturation of GABAergic transmission in the auditory cortex. Cerebr. Cortex 18 (9), 2098e2108. https://doi.org/10.1093/cercor/bhm233.

54

PART | I Circuit development

Koundakjian, E.J., Appler, J.L., Goodrich, L.V., 2007. Auditory neurons make stereotyped wiring decisions before maturation of their targets. J. Neurosci. 27 (51), 14078e14088. https://doi.org/10.1523/JNEUROSCI.3765-07.2007. Kral, A., Tillein, J., Heid, S., Hartmann, R., Klinke, R., 2005. Postnatal cortical development in congenital auditory deprivation. Cerebr. Cortex 15 (5), 552e562. https://doi.org/10.1093/cercor/bhh156. Leake, P.A., Snyder, R.L., Hradek, G.T., 2002. Postnatal refinement of auditory nerve projections to the cochlear nucleus in cats. J. Comp. Neurol. 448 (1), 6e27. https://doi.org/10.1002/cne.10176. Levi-montalcini, R., 1949. The development of the acoustico-vestibular centres in the chick embryo in the absence of the afferent root fibers and of descending fiber tracts. J. Comp. Neurol. 91 (2), 209e241. https://doi.org/10.1002/cne.900910204. Liberman, M.C., 1978. Auditory-nerve response from cats raised in a low-noise chamber. J. Acoust. Soc. Am. 63 (2), 442e455. Mann, Z.F., Kelley, M.W., 2011. Development of tonotopy in the auditory periphery. Hear. Res. 276 (1e2), 2e15. https://doi.org/10.1016/ J.HEARES.2011.01.011. Marcotti, W., 2012. Functional assembly of mammalian cochlear hair cells. Exp. Physiol. 97 (4), 438e451. https://doi.org/10.1113/ expphysiol.2011.059303. Marrs, G.S., Spirou, G.A., 2012. Embryonic assembly of auditory circuits: spiral ganglion and brainstem. J. Physiol. 590 (10), 2391e2408. https://doi.org/ 10.1113/jphysiol.2011.226886. Meng, X., Kao, J.P.Y., Kanold, P.O., 2014. Differential signaling to subplate neurons by spatially specific silent synapses in developing auditory cortex. J. Neurosci. 34 (26), 8855e8864. https://doi.org/10.1523/JNEUROSCI.0233-14.2014. Meng, X., Solarana, K., Bowen, Z., Liu, J., Nagode, D.A., Sheikh, A., Winkowski, D.E., Kao, J.P.Y., Kanold, P.O., 2019. Transient Subgranular Hyperconnectivity to L2/3 and Enhanced Pairwise Correlations During the Critical Period in the Mouse Auditory Cortex. Cereb Cortex. Metherate, R., Hsieh, C.Y., 2003. Regulation of glutamate synapses by nicotinic acetylcholine receptors in auditory cortex. Neurobiol. Learn. Mem. 80 (3), 285e290. Mogdans, J., Knudsen, E.I., 1994. Site of auditory plasticity in the brain stem (VLVp) of the owl revealed by early monaural occlusion. J. Neurophysiol. 72 (6), 2875e2891. https://doi.org/10.1152/jn.1994.72.6.2875. Mrsic-Flogel, T.D., Versnel, H., King, A.J., 2006. Development of contralateral and ipsilateral frequency representations in ferret primary auditory cortex. Eur. J. Neurosci. 23 (3), 780e792. https://doi.org/10.1111/j.1460-9568.2006.04609.x. Noreña, A.J., Gourévitch, B., Aizawa, N., Eggermont, J.J., Eggermont, J.J., 2006. Spectrally enhanced acoustic environment disrupts frequency representation in cat auditory cortex. Nat. Neurosci. 9 (7), 932e939. https://doi.org/10.1038/nn1720. Norton, S.J., Bargones, J.Y., Rubel, E.W., 1991. Development of otoacoustic emissions in gerbil: evidence for micromechanical changes underlying development of the place code. Hear. Res. 51 (1), 73e91. https://doi.org/10.1016/0378-5955(91)90008-W. Nothwang, H.G., Ebbers, L., Schlüter, T., Willaredt, M.A., 2015. The emerging framework of mammalian auditory hindbrain development. Cell Tissue Res. 361 (1), 33e48. https://doi.org/10.1007/s00441-014-2110-7. Oswald, A.-M.M., Reyes, A.D., 2011. Development of inhibitory timescales in auditory cortex. Cerebr. Cortex 21 (6), 1351e1361. https://doi.org/ 10.1093/cercor/bhq214. Petitpré, C., Wu, H., Sharma, A., Tokarska, A., Fontanet, P., Wang, Y., et al., 2018. Neuronal heterogeneity and stereotyped connectivity in the auditory afferent system. Nat. Commun. 9 (1), 3691. https://doi.org/10.1038/s41467-018-06033-3. Popescu, M.V., Polley, D.B., 2010. Monaural deprivation disrupts development of binaural selectivity in auditory midbrain and cortex. Neuron 65 (5), 718e731. https://doi.org/10.1016/j.neuron.2010.02.019. Rothschild, G., Nelken, I., Mizrahi, A., 2010. Functional organization and population dynamics in the mouse primary auditory cortex. Nat. Neurosci. 13 (3), 353e360. https://doi.org/10.1038/nn.2484. Rubenstein, J.L.R., Rakic, P., 2013. Cellular migration and formation of neuronal connections. Cellular Migration and Formation of Neuronal Connections. https://doi.org/10.1016/C2011-0-07731-1. Ryugo, D.K., Kretzmer, E.A., Niparko, J.K., 2005. Restoration of auditory nerve synapses in cats by cochlear implants. Science 310 (5753), 1490e1492. https://doi.org/10.1126/science.1119419. Sanes, D.H., Constantine-Paton, M., 1983. Altered activity patterns during development reduce neural tuning. Science 221 (4616), 1183e1185. Sanes, D.H., Bao, S., 2009. Tuning up the developing auditory CNS. Curr. Opin. Neurobiol. 19 (2), 188e199. https://doi.org/10.1016/j.conb.2009.05.014. Schwander, M., Kachar, B., Müller, U., 2010. Review series: the cell biology of hearing. J. Cell Biol. 190 (1), 9e20. https://doi.org/10.1083/ jcb.201001138. Seidl, A.H., Grothe, B., 2005. Development of sound localization mechanisms in the Mongolian gerbil is shaped by early acoustic experience. J. Neurophysiol. 94 (2), 1028e1036. https://doi.org/10.1152/jn.01143.2004. Shrestha, B.R., Chia, C., Wu, L., Kujawa, S.G., Liberman, M.C., Goodrich, L.V., 2018. Sensory neuron diversity in the inner ear is shaped by activity. Cell 174 (5), 1229e1246.e17. https://doi.org/10.1016/j.cell.2018.07.007. Sorensen, S.A., Rubel, E.W., 2006. The level and integrity of synaptic input regulates dendrite structure. J. Neurosci. 26 (5), 1539e1550. https://doi.org/ 10.1523/JNEUROSCI.3807-05.2006. Sun, S., Babola, T., Pregernig, G., So, K.S., Nguyen, M., Su, S.-S.M., et al., 2018. Hair cell mechanotransduction regulates spontaneous activity and spiral ganglion subtype specification in the auditory system. Cell 174 (5), 1247e1263.e15. https://doi.org/10.1016/j.cell.2018.07.008. Tierney, T.S., Russell, F.A., Moore, D.R., 1997. Susceptibility of developing cochlear nucleus neurons to deafferentation-induced death abruptly ends just before the onset of hearing. J. Comp. Neurol. 378 (2), 295e306. https://doi.org/10.1002/(SICI)1096-9861(19970210)378:23.0.CO;2-R.

Functional circuit development in the auditory system Chapter | 2

55

Tolner, E.A., Sheikh, A., Yukin, A.Y., Kaila, K., Kanold, P.O., 2012. Subplate neurons promote spindle bursts and thalamocortical patterning in the neonatal rat somatosensory cortex. J. Neurosci. 32 (2), 692e702. https://doi.org/10.1523/JNEUROSCI.1538-11.2012. Tritsch, N.X., Rodríguez-Contreras, A., Crins, T.T.H., Wang, H.C., Borst, J.G.G., Bergles, D.E., 2010. Calcium action potentials in hair cells pattern auditory neuron activity before hearing onset. Nat. Neurosci. 13 (9), 1050e1052. https://doi.org/10.1038/nn.2604. Tritsch, N.X., Yi, E., Gale, J.E., Glowatzki, E., Bergles, D.E., 2007. The origin of spontaneous activity in the developing auditory system. Nature 450 (7166), 50e55. https://doi.org/10.1038/nature06233. Tucci, D.L., Rubel, E.W., 1985. Afferent influences on brain stem auditory nuclei of the chicken: effects of conductive and sensorineural hearing loss on N. Magnocellularis. J. Comp. Neurol. 238 (4), 371e381. https://doi.org/10.1002/cne.902380402. Vale, C., Schoorlemmer, J., Sanes, D.H., 2003. Deafness disrupts chloride transporter function and inhibitory synaptic transmission. J. Neurosci. 23 (20), 7516e7524. Viswanathan, S., Sheikh, A., Looger, L.L., Kanold, P.O., 2017. Molecularly defined subplate neurons project both to thalamocortical recipient layers and thalamus. Cerebr. Cortex 27 (10), 4759e4768. https://doi.org/10.1093/cercor/bhw271. Von Bekesy, G., 1960. Experiments in Hearing. McGraw-Hill, New York. Walsh, E.J., McGee, J., 1987. Postnatal development of auditory nerve and cochlear nucleus neuronal responses in kittens. Hear. Res. 28 (1), 97e116. https://doi.org/10.1016/0378-5955(87)90157-2. Wang, H.C., Lin, C.-C., Cheung, R., Zhang-Hooks, Y., Agarwal, A., Ellis-Davies, G., et al., 2015. Spontaneous activity of cochlear hair cells triggered by fluid secretion mechanism in adjacent support cells. Cell 163 (6), 1348e1359. https://doi.org/10.1016/j.cell.2015.10.070. Werthat, F., Alexandrova, O., Grothe, B., Koch, U., 2008. Experience-dependent refinement of the inhibitory axons projecting to the medial superior olive. Dev. Neurobiol. 68 (13), 1454e1462. https://doi.org/10.1002/dneu.20660. Werker, J.F., Yeung, H.H., 2005. Infant speech perception bootstraps word learning. Trends Cogn Sci. 9 (11), 519e527. Wess, J.M., Isaiah, A., Watkins, P.V., Kanold, P.O., 2017. Subplate neurons are the first cortical neurons to respond to sensory stimuli. Proc. Natl. Acad. Sci. 114 (47), 12602e12607. https://doi.org/10.1073/pnas.1710793114. Wu, J.S., Young, E.D., Glowatzki, E., 2016. Maturation of spontaneous firing properties after hearing onset in rat auditory nerve fibers: spontaneous rates, refractoriness, and interfiber correlations. J. Neurosci. 36 (41), 10584e10597. https://doi.org/10.1523/JNEUROSCI.1187-16.2016. Yin, Y., Liberman, L.D., Maison, S.F., Liberman, M.C., 2014. Olivocochlear innervation maintains the normal modiolar-pillar and habenular-cuticular gradients in cochlear synaptic morphology. JARO J. Assoc. Res. Otolaryngol. 15 (4), 571e583. https://doi.org/10.1007/s10162-014-0462-z.

Chapter 3

Development of the superior colliculus/optic tectum Barry E. Stein, Terrence R. Stanford and Benjamin A. Rowland Department of Neurobiology & Anatomy, Wake Forest School of Medicine, WinstoneSalem, NC, United States

Chapter outline 3.1. 3.2. 3.3. 3.4.

Nomenclature Functional role General anatomical organization of the superior colliculus Spatial topographies, multisensory integration, and motor output 3.5. The maturation of the superior colliculus 3.5.1. The neonate 3.5.2. Sensory chronology 3.5.3. The development of multisensory neurons

57 57 58 59 63 63 63 67

3.5.4. Superficialedeep (multisensory) layer maturational delay 3.5.5. The development of multisensory integration 3.5.6. The impact of sensory experience on the maturation of multisensory integration 3.6. Summary Acknowledgments References

67 68 68 73 73 73

3.1 Nomenclature Anatomically, the tectum is that portion of the mesencephalon, or midbrain, sitting between the hindbrain and the forebrain. The name is Latin for roof, and reflects the view of early anatomists that the tectum formed a roof over the fluid filled cerebral aqueduct and the tegmentum. In mammals, it is sometimes referred to as the tectal plate (or quadrigeminal plate), and is composed of two pairs of bumps or colliculi (collis means hill in Latin), one on each side of the midbrain. The more rostral pair is referred to as the superior colliculi and the more caudal pair as the inferior colliculi. However, the reader should be aware that there is some confusion in nomenclature. The term optic tectum (OT) properly refers to the nonmammalian homolog of the superior colliculus (SC), and its name reflects its preeminent visual role in these animals. Thus, discussions of the OT are generally specific to the major central nervous system visual nucleus in birds, reptiles, amphibians, and fish, where much has been learned about its function. But while the mammalian homolog of the OT is called the superior colliculus, projections to and from it are, respectively, called tectopetal (e.g., retinotectal, corticotectal) and tectofugal (e.g., tectoreticular, tectospinal), a confusing nomenclature that is compounded by the tendency of some to refer to the SC as the OT, and others to use colliculo(ar) for tecto(al). The primary subject of the following discussion is the SC. However, the OT and SC have fundamental functional similarities and research on the OT will be referred to below when instructive for understanding SC organization and/or function.

3.2 Functional role While the SC, like its nonmammalian counterpart, is clearly visually dominant, it is not strictly visual, but then, neither is the OT. Furthermore, the role of the SC is not restricted to visuomotor behavior (nor is the OT). Rather, it is a multisensory structure, containing visual, auditory, and somatosensory representations, all of which contribute to its role in initiating orientation and localization behaviors that involve multiple sensory organs. Thus, while the visuomotor role in gaze shifts

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00003-2 Copyright © 2020 Elsevier Inc. All rights reserved.

57

58

PART | I Circuit development

(i.e., movement of the eyes, with or without corresponding movement of the head) is best known (Sparks, 1986), it is also involved in orientation of the ears (Stein and Clamann, 1981) and limbs (Stein and Gaither, 1981; Stuphorn et al., 2000). It is best to think of the SC as involved in the transformation of sensory signals (visual, auditory, and/or somatosensory) into motor commands. Its ability to represent salient stimuli and engage in sensorimotor transduction so that the organism can initiate rapid orientation to the initiating event is a function critical for survival. The rapid maturation of the SC, compared to that of cortex, reflects its importance in minimizing early ecological vulnerabilities as neonates develop greater independence. This is especially evident in altricial species, whose birth at an early stage of maturation makes observation of neural development far easier than in their precocial counterparts. However, in the adult stage of both altricial and precocial species the communication between cortex and SC reaches its apogee, and their complementary and interdependent functions become linked through their reciprocal connections. Before discussing the developmental features that render the SC capable of performing its adult role or how its inherent plasticity is reflected in changes resulting from sensory experience and/or trauma, it is first necessary to describe its overall structure and function in the mature brain. What follows is a discussion of the major features of this remarkable structure.

3.3 General anatomical organization of the superior colliculus The SC is a laminated structure, composed of seven alternating fibrous and cellular layers (Kanaseki and Sprague, 1974) as shown in the illustration from a cat in Fig. 3.1. Operationally, however, it is divided into two broad regions: superficial (IeIII) and deep (IVeVII) layers (see Harting et al., 1973; Edwards, 1980; Stein, 1984). The former is strictly visual. It receives a substantial direct projection from the retina, as well as many visual projections indirectly from a host of other subcortical and cortical structures (Edwards et al., 1979). Its outputs ascend the neuraxis via sensory thalamus and are relayed from there to extrastriate cortex. Its descending projections pass through the deeper layers to target another visual site, the parabigeminal nucleus. In contrast, the deep layers of the SC contain a much more heterogeneous group of neurons; recipients not only of visual inputs (few come directly from the retina, but many are relayed from other brain regions, especially cortex), but also inputs from the auditory, somatosensory, and motor systems (Stein and Meredith, 1993). In keeping with their role in sensorimotor transformation, the deeper layers send very heavy descending projections to regions of the brainstem and spinal cord that control movements of the eyes, ears, head, and limbs, as well as outputs to nonspecific sensory and motor areas of thalamus (Edwards, 1980; Edwards and Henkel, 1978; Edwards et al., 1979; Harting, 1977; Harting et al., 1973, 1980; Moschovakis and Karabelas, 1985; Redgrave et al., 1985, 1986a,b; Rhoades et al., 1987; Weber et al., 1979; Sommer and Wurtz, 1998, 2004a,b). It is the deep layers that are most closely associated with the sensorimotor roles for which the SC is best known, and because they contain neurons responsive to multiple sensory modalities, they will also be referred to below as the multisensory layers.

FIGURE 3.1 A lateral view of the cat brain and superior colliculus (SC). A portion of cortex has been removed to reveal the SC and IC (unlabeled). A coronal section of the SC to the right shows its layers: SAI, stratum album intermediale; SAP, stratum album profundum; SGI, stratum griseum intermediale; SGP, stratum griseum profundum; and SGS, stratum griseum superficiale; SO, stratum opticum; SZ, stratum zonale. From Stein, B.E., Meredith, M.A., 1993. The Merging of the Senses. MIT Press, Cambridge, Mass.

Development of the superior colliculus/optic tectum Chapter | 3

59

3.4 Spatial topographies, multisensory integration, and motor output Visuotopy: Physically, the SC is aligned with the general axis of the brain, so its rostral pole points forward and its caudal pole points backward. Neurons in its superficial visual layers are arranged in a visuotopic (i.e., retinotopic) map of contralateral visual space, and the map in the cat SC (e.g., see Feldon et al., 1970) is generally representative of those in animals having forward facing eyes (see Cynader and Berman, 1972, for a description in monkey). SC neurons having receptive fields in central visual space are located rostral in the structure, and are smallest. Thus, this portion of the map has the greatest spatial resolution. Neurons having receptive fields progressively more peripheral (i.e., temporal) in space are located progressively more caudal in the structure and are largest. This portion of the map has the lowest spatial fidelity. In short, the horizontal meridian from central to peripheral visual space runs roughly rostral-caudal. The vertical meridian is roughly orthogonal to the horizontal meridian, with superior visual space represented medial and inferior space represented lateral. The central visual point (i.e., the fovea in primates and area centralis in carnivores) is the point at which the horizontal and vertical meridians cross. The central representation of visual space is expanded in the SC. Thus, for example, the central 10 degrees of space occupies more than a third of the tissue devoted to the map in the cat SC and more than half the tissue in monkey SC. In some species, such as the cat, neurons located rostral to the vertical meridian extend the map progressively more nasal to represent up to 10 degrees of the ipsilateral hemifield. This nasal representation is essentially nonexistent in monkey (Cynader and Berman, 1972), but is much larger in animals such as the rat, whose eyes are on the sides of the head (e.g., see Siminoff et al., 1966). A retinotopic organization is also characteristic of the OT, although in some nonmammalian species, the vertical and horizontal meridians are not always as well aligned with the brain’s rostral-caudal axis as they are in mammals (e.g., see Gaither and Stein, 1979). The ubiquity of a retinotopy likely represents the ease of using a map to determine the location of a visual event, and of transforming the visual cues it provides into motor coordinates for orientation responses. However, this sensorimotor transduction is primarily a function of the deep (i.e., multisensory) layers of the SC, which also contain a map of visual space. The retinotopy in the deep layers is similar to that of the overlying superficial layer map, though the source of visual afferents differs somewhat. Receptive fields are considerably larger than those in the overlying layers and, as a result, there is lower spatial resolution in this visual map (McIlwain, 1975, 1991; Meredith and Stein, 1990). Yet it has a better representation of the far periphery, and also extends a bit further into the ipsilateral hemifield than does the superficial layer map (Meredith and Stein, 1990). Shifts in receptive field size follow the same trend as those in superficial layers, with central visual fields being smallest and far peripheral receptive fields being largest. But the visual map in the multisensory layers is only one of three overlapping spatiotopic sensory representations in this region of the structure. Somatotopy: The somatosensory representation in the multisensory SC is largely of the cutaneous surface and its maplike representation is called “somatotopic.” Its relationship to the visuotopic representation has also been studied most extensively in the cat (see Stein et al., 1975; Meredith et al., 1991; but see also Drager and Hubel, 1975; Benedetti, 1995; McHaffie et al., 1989). Like the visual representation, it is formed from comparatively large receptive fields that are organized into a map of the body, with a geometric expansion of the representation of the face and head. The face representation is made up of the smallest receptive fields, is located rostral, and roughly overlaps the representation of central visual space. The body and rump are represented progressively more caudal and the limbs extend laterally, so that upper body space is represented medial to coincide with the representation of upper visual space, and the lower body space is represented lateral to coincide with the representation of lower visual space. Given that the SC also receives inputs from pathways carrying information about potentially harmful stimuli, it is not surprising to find that it also has neurons responsive to noxious stimuli (Stein and Dixon, 1978). These neurons have many of the same properties found in structures better known for dealing with nociceptive information (Rhoades et al., 1983; Larson et al., 1987). Although this representation has been studied only in rodents, and is largely restricted to the face and forelimb (McHaffie et al., 1989; see also Auroy et al., 1991) it incorporates the same general topographic features as do the other sensory maps (see Stein and Meredith, 1993). Audiotopy: Unlike the visual and somatotopic maps, which are formed directly via afferents from different regions of the retina or skin (albeit with some geometrical distortions), the spatiotopic nature of the auditory representation must be derived via neural computation based on comparisons of the timing, intensity, and frequency content of the sound signals at the two ears. Although the term “audiotopic” is not in common usage, and as a result seems a bit awkward, it is as appropriate as “retinotopic” and “somatotopic” in this context. Its properties have also been examined closely in the cat, and follow the same general organizational features of the maps that are outlined above. Forward auditory space is represented in the rostral aspect of the structure, temporal auditory space in its caudal aspect, and superior-inferior space is laid

60

PART | I Circuit development

out along its medial-lateral axis (Middlebrooks and Knudsen, 1984); see also Palmer and King (1982); King and Palmer (1983). Multisensory integration: The alignment of SC sensory maps is evident in the overlap among modality-specific receptive fields of multisensory SC and OT neurons. Indeed, the individual maps are largely a reflection of the receptive field properties of multisensory neurons which, in the cat at least, comprise the majority of sensory responsive neurons in the deep layers. Each multisensory neuron has at least two receptive fields, one for each of the sensory modalities to which it responds (e.g., see Stein and Meredith, 1993 for a review). For example, a visual-auditory neuron with a visual receptive field in central visual space will have an auditory receptive field in an overlapping region of central visual space, so that only visual and auditory cues from roughly the same locations will affect that neuron. A key element of SC function derives from the ability of multisensory neurons to integrate the influences of the different sensory modalities, and the nature of such integration hinges strongly on map alignment. While much of the information about multisensory integration comes from studies of single neurons in cat SC, similar observations have also been made in the SC and OT of other species (see, e.g., Hartline et al., 1978; Gaither and Stein, 1979; Stein and Gaither, 1981; Zahar et al., 2009; Wallace et al., 1996; Bell et al., 2001, 2005; Van Wanrooij et al., 2009). Operationally defined, “multisensory integration” is the process by which stimuli from different senses, when combined (i.e., cross-modal), produce a response that differs from those produced by the component stimuli individually. At the level of the single neuron, integration corresponds to a statistically significant difference between the number of impulses evoked by a crossmodal combination of stimuli and the number evoked by the most effective of these stimuli individually (see Stein and Stanford, 2008 for discussion). This difference can be manifested as either a response increase or decrease depending on how the stimuli are configured in space. Cross-modal stimuli that signal a common event (they are in the same place at the same time) impinge upon overlapping modality-specific receptive fields and the integration of their influences yields response enhancement. Conversely, modality-specific cues emanating from disparate locations impact nonoverlapping regions of their respective sensory maps; disparate cues thus fail to produce enhancement and, in some instances, produce response depression (Kadunce et al., 1997). Examples of multisensory integration that yield response enhancement and response depression are shown in Fig. 3.2. Remarkably, convergence of inputs from different senses alone is not sufficient for multisensory integration in the SC. A matching set of converging unisensory inputs from association cortex (primarily from the anterior ectosylvian sulcus, (AES; Fuentes-Santamaria et al., 2009; Alvarado et al., 2009; Fuentes-Santamaria et al., 2008), but inputs from the rostral lateral suprasylvian sulcus (rLS) also play a roledsee Jiang et al., (2001), without which SC neurons respond to unisensory inputs but do not integrate them to yield response enhancement (Wallace and Stein, 1994; Jiang et al., 2001; Alvarado et al., 2007, 2008, 2009). Rather, the neural response to cross-modal stimulation is no greater than that to the best of the stimulus components alone, a finding that is mirrored by an absence of the multisensory advantages that are typical for SCmediated behaviors (Jiang et al., 2002, 2006, 2007; Wilkinson et al., 1996). As discussed below, the requirement for cortical input has implications for the development of SC multisensory integration and the behaviors it supports. Insofar as the magnitude of activity within SC sensory maps corresponds to the physiological salience of an external stimulus, and thus the likelihood that it will generate a motor response, the implications of multisensory integration for

FIGURE 3.2 Multisensory enhancement. Left: A square-wave broadband auditory stimulus (A, first panel) and a moving visual stimulus (ramp labeled V, second panel) evoked unisensory responses from this neuron (illustrated below the stimulus by rasters and peri-stimulus time histograms). Each dot in the raster represents a single impulse and each row a single trial. Trials are ordered from bottom to top. The third panel shows the response to auditory and visual stimulus presentation at the same time and location, which is much more robust than either unisensory response. But, when the auditory stimulus was moved into ipsilateral auditory space (Ai) and out of the receptive field, its combination with the visual stimulus elicited fewer impulses than did the visual stimulus individually. This “response depression” is illustrated within the fourth panel. Right: The mean number of impulses/ trial elicited by each of the four stimulus configurations. Note the difference between multisensory enhancement (VA) and multisensory depression (VAi). From Stein BE, Rowland BA (2011) Organization and plasticity in multisensory integration: early and late experience affects its governing principles. Prog. Brain Res. 191:145e163.

Development of the superior colliculus/optic tectum Chapter | 3

61

behavior are quite clear. Enhanced activity for spatially concordant cues corresponds to an increased likelihood of stimulus detection and an associated motor response to orient to that stimulus. In contrast, the multisensory depression that results from stimuli at competing locations would have an opposing effect (Spence et al., 2004; Calvert et al., 2004; Gillmeister and Eimer, 2007; Stein and Stanford, 2008). In either case, sensory-related activation must be transformed into motor commands, which, like their sensory counterparts, assume the form of topographically organized map. Mototopic representation: Many of the neurons in the deep/multisensory layers are output neurons that project to one or more brainstem and spinal cord regions responsible for moving the eyes, ears, whiskers, head, and limbs (Edwards and Henkel, 1978; Coulter et al., 1979; Weber et al., 1979: Holcombe and Hall, 1981a,b; Harting, 1977; Huerta and Harting, 1982a,b; Grantyn and Grantyn, 1982; Moschovakis et al., 1998; see Sparks, 1986; Sparks and Mays, 1990; Hall and Moschovakis, 2004; Gandhi and Katnani, 2011 for reviews). The term “mototopic” like “audiotopic” is not in common usage, but, once again, seems equally appropriate in this context given the more commonly used terms “visuotopic” and “somatotopic.” By far, the most is known about the motor topography for producing gaze shifts, movements of eyes or eyes and head to place stimuli of interest in the line of site. Accordingly, the motor map for gaze shifts is in register with the visual, auditory, and somatosensory topographies; therefore, the sensory evoked activity represents the distance and direction of a stimulus from the current line of sight and the motor activity at that site represents a command for shifting gaze, the corresponding distance and direction. The SC motor map is two-dimensional, with gaze amplitude (from small to large) represented along its rostro-caudal axis and gaze direction (from upward to downward) represented along its medio-lateral axis (Robinson, 1972; Schiller and Stryker, 1972; Goldberg and Wurtz, 1972a; Wurtz and Goldberg, 1972; du Lac and Knudsen, 1990; Pare et al., 1994). Analogous to sensory receptive fields, constituent neurons of the SC motor map have movement fields, such that they respond most vigorously in association with gaze shifts within a particular range of amplitude and direction and gaze shifts are produced consequent to the activity of their premotor activity. Fig. 3.3 illustrates a typical SC movement field for a neuron recorded from the right SC of a monkey. This neuron discharges most vigorously (peak of the 3D plot) for saccades having a direction of approximately 190 degrees (leftward and slightly downward) and having an amplitude of approximately 7 degrees of visual angle (i.e., deviation of eye from the straight ahead position). Considered in the context of the motor map of the right SC, this neuron would be located toward the rostral pole (small movements) and slightly lateral (downward). Thus, like the SC sensory maps, which constitute “place codes” for stimulus location in sensory space, the SC motor map is a “place code” for movement vector such that the locus of activation determines the amplitude and direction of an impending movement. Maintaining sensory and motor map alignment: The overlap among the sensory and motor maps is unlikely to be serendipitous, as it appears to be the most straightforward way to coordinate cross-modal sensory cues and movements of the various sensory organs toward a salient event. Nevertheless, given the ability to move the different sensory organs independent of one another, retaining the registry of their midbrain maps presents a nontrivial problem. The problem is partially solved by the nonstatic nature of these maps. Studies in a variety of species show that shifting the eyes voluntarily, or by inducing long-term shifts in the optical axis via surgical or prismatic means, produces a corresponding shift in the SC auditory map (Jay and Sparks, 1984; Hartline et al., 1995; Peck et al., 1995; King et al., 1988; Brainard and Knudsen, 1998). A similar shift may also be initiated in the somatosensory map (Groh and Sparks, 1996). These observations suggest that the different senses are linked to an oculocentric coordinate system, though compensatory shifts in the nonvisual maps do not completely compensate for large ocular misalignments (Metzger et al., 2004) Maintaining intermap registry ensures that any visual, auditory, or somatosensory cue (and any combination of them) derived from the same location in space activates neurons in approximately the same SC location. This, in turn, accesses the same point in the motor map to produce coordinated orientation of all the sensory organs toward the location of the initiating event. This not only helps determine the nature of that event, but puts the animal in an appropriate position to evaluate it. Beyond sensory and motor: Orienting toward a stimulus of interestda principal function of SC motor outputdis inextricably linked to mechanisms of spatial attention (see Knudsen, 2011; Krauzlis et al., 2013; Basso and May 2017 for reviews). Both stimulus-driven (bottom-up) and cognitive (top-down) factors are known to influence SC activity in a manner that increases the likelihood that an SC-mediated motor choice is appropriate for a given context. Thus, for example, highly salient stimuli, which are known to attract attention automatically (see Carrasco, 2011 for review), are associated with relatively higher levels of SC activity in corresponding regions of its sensory and motor maps. So-called “saliency maps” represent the relative saliency of regions of space and may be read out to direct movement to the locations where conspicuous stimuli appear. Such conspicuity may be the product of size, frequency, intensity, color, or any other physical feature that distinguishes a stimulus from its surrounding milieu. Accordingly, numerous studies in behaving primates have demonstrated that visually contingent SC activity evolves rapidly to “select” salient visual targets, so-called “oddballs” that stand out when presented within an array alongside numerous nontarget elements (McPeek and Keller,

62

PART | I Circuit development

FIGURE 3.3 Superior colliculus movement field. Neurons in the SC have movement fields and discharge most for movements within particular ranges of amplitude and direction. The 3D polar plot (above) illustrates the movement field of an SC neuron that discharges for relatively small and slightly downward saccadic eye movements. Below left, the premotor activity profiles are shown aligned on the onsets of saccades (blue dotted line) of five different amplitudes (top to bottom) and representing a “slice” through the peak of the movement field along the line of best direction. Representative saccade trajectories are shown to the right of the activity profiles. Note that maximum activity corresponds to a slightly downward saccade having amplitude corresponding to roughly 7 degrees of visual angle. From Stanford & Sparks, unpublished observations.

2002, 2004; Kim and Basso, 2008). Analogously, the response enhancement observed when spatially concordant stimuli from multiple stimulus modalities are integrated by SC neurons is also an example of a saliency scaling that facilitates spatial orienting to potentially important environmental events (see Stein and Stanford, 2008 for review). Importantly though, the SC’s representation of potential goals for orienting extends beyond that which is simply dictated by the physical characteristics of the stimuli themselves. Whereas the superficial layers of the SC may indeed constitute a pure visual saliency map tied to physical stimulus attributes (White et al., 2017a,b), the intermediate and deeper layers, where multiple sensory representations coexist along with the SC motor map, represent something more akin to “priority” (see Fecteau and Munoz, 2006 for review). Priority maps conjoin saliency and internal decision variables reflecting current task objectives and expected outcomes to prioritize regions of space for action (see Awh et al., 2012 for review). Thus, how activity is distributed within a priority map is the product of both bottom-up and top-down attentional mechanisms. Some of the earliest studies of single-neuron activity in the primate SC demonstrated enhanced activity for stimuli that were known in advance to be the goals of a rewarded action as opposed to those that were inconsequential to the task at hand (Goldberg and Wurtz, 1972b). Since those early observations, increasingly sophisticated experimental

Development of the superior colliculus/optic tectum Chapter | 3

63

designs have demonstrated that activity within the SC is the product of both the stimulus and its relevance to behavioral outcome (see Fecteau and Munoz, 2006; Krauzlis et al., 2013; Basso & May 2017 for reviews). Although both bottom-up and top-down attentional mechanisms manifest strongly in SC sensory and motor activity, a critical question concerns the degree to which SC activity plays a causal role in the voluntary allocation of spatial attention, one that is independent of its role in generating overt movement (Kustov and Robinson, 1996; Song et al., 2011; Krauzlis et al., 2013; Sridharan et al., 2017). It is well established that attention can be deployed covertly to peripheral regions of space while maintaining the current line of sight, the key consequence of which is an increase in perceptual sensitivity at these locations. Experiments combining psychophysics and neurophysiology have revealed a plausible neural substrate for how such increased perceptual acuity arises based on activity within cortically based networks. Numerous studies have shown that allocating attention modulates the visual feature selectivity of neurons in extrastriate visual cortex whose receptive fields align with an attended location (Spitzer et al., 1988; Conner et al., 1996, 1997; Treue and Maunsell, 1996; McAdams and Maunsell, 1999; Williford and Maunsell, 2006; see Maunsell, 2015 for review). In a series of seminal studies, Moore and colleagues demonstrated that such response modulation shows a causal dependence on input from a region of prefrontal cortex known as the frontal eye fields (FEF). By artificially manipulating activity in the FEFd suppressing it via pharmacological inactivation or enhancing it with electrical microstimulationdthey induced corresponding changes in extrastriate activity and psychophysical performance in providing evidence that FEF is the source of a top-down signal for the control of spatial attention (Moore and Fallah, 2001, 2004; Moore and Armstrong, 2003; Armstrong et al., 2006; Armstrong and Moore, 2007; Schafer and Moore, 2007; Noudoost et al., 2014; see Noudoost et al., 2010, Clark et al., 2015 for reviews). Recent studies have applied the same experimental logic to determine if activity in the SC is integral to the deployment of covert spatial attention and have reported analogous spatially specific effects on perceptual judgments consequent to experimentally induced enhancement or suppression of activity within circumscribed regions of the SC sensorimotor topography (Cavanaugh and Wurtz, 2004; Muller et al., 2005; Lovejoy and Krauzlis, 2010; Zenon and Krauzlis, 2012). Although, the respective roles of cortical and midbrain networks to attentional control remain a matter of debate (Sridharan et al., 2017), current evidence suggests that SC’s role extends beyond the generation of orienting movements toward attended locations.

3.5 The maturation of the superior colliculus 3.5.1 The neonate In altricial species like the cat, the ability of the SC to use sensory information to initiate overt behaviors is poorly developed at birth. Unlike primates, ungulates, and other precocial species, the newborn carnivore (rodent, lagomorph, marsupial, etc.) is poorly equipped to deal with its sensory environment. Its eyes are closed, there is a vascular network around the lens that impedes the transmission of light (Thorn et al., 1976; Bonds and Freeman, 1978), and it is functionally blind. Similarly, its ear canals are still sealed, and it is deaf. The only sensory inputs that activate SC neurons at this time are tactile. These inputs are already functional prenatally, but are only weakly effective and stay that way for some time after birth (Stein et al., 1973). Its motor capabilities are also poorly developed, and it depends heavily on its mother for warmth and protection, and even to initiate feeding (Larson and Stein, 1984; Rosenblatt, 1971). The immaturity of the newborn cat’s SC makes it a good model for exploring how sensory responses develop, how the different sensory representations become established, and how multisensory integration develops so that SC neurons can use the available sensory information synergistically to optimize SC-mediated behavior.

3.5.2 Sensory chronology As noted above, somatosensory responsiveness in some cat SC neurons precedes birth, and provides the structure with its first source of information about the world. Auditory responses begin to appear in some SC neurons at five days postnatal and neurons develop visual responsiveness last (Stein et al., 1973). Interestingly, visual responses develop in two sequential phases, first appearing from top to bottom in the most superficial (i.e., visual) layers of the structure, and then in similar fashion in neurons in its multisensory (deeper) layers. These phases are widely separated in time, with neurons in the most superficial strata of the superficial layers beginning to respond to visual stimuli at about six postnatal days, and those in the multisensory layers not showing responsiveness to visual stimuli for several weeks (Kao et al., 1994). Retinotectal inputs and the development of a superficial layer visuotopy: Because of the visually dominant role of the SC, the development of its visual inputs (especially those coming directly from the retina) has received a great deal of attention. Once again, the cat has been one of the primary sources of this information, along with the rodent, opossum, and

64

PART | I Circuit development

FIGURE 3.4 Retinotectal projection patterns to the superficial superior colliculus are highly developed in cat before birth. Shown are photographic illustrations of the distribution of a tracer (horseradish peroxidase shown as dark regions) from one eye to both colliculi approximately 17 (A, at E38) and 9 days before birth (B, at E56). Sections are ordered from rostral (bottom) to caudal (top), with the contralateral SC on the left. The calibration bar is 1 mm. Note that the projection has largely withdrawn from the ipsilateral SC and changed its distribution in the contralateral SC by E56. Several days later (C, at E61), an almost adultlike pattern input is apparent. Adapted from Williams, R.W., Chalupa, L.M., 1982. Prenatal development of retinocollicular projections in the cat: an anterograde tracer transport study. J. Neurosci. 2, 604e606.

monkey. Given the functional immaturity of the newborn cat’s SC (there is no visual responsiveness yet), it is surprising to note the advanced state of its retinotectal topography during late embryonic development (Graybiel, 1975; Williams and Chalupa, 1982). Although, at this point, the retinotectal projections appear to be restricted to the superficial layers of the SC, segregation of inputs from the contralateral and ipsilateral eyes is already apparent several days before birth (Fig. 3.4B; embryonic day 56), and by embryonic day 61 (Fig. 3.4C), further refinement leads to an almost adultlike patterning of inputs. This advanced pattern of retinal projections is achieved by sculpting it from a far more widespread projection that is evident at approximately embryonic day 38 (Fig. 3.4A). At that point the projections from the two retinas are intermingled and distributed across the entire rostral-caudal and medial-lateral extent of the SC. This pattern is progressively altered so that by embryonic day 61 (4e7 days before parturition), the pattern of ipsilateral and contralateral retinotectal inputs looks very much like that characteristic of the adult. Although the timing is different, similar developmental processes have been noted in monkey (Rakic, 1977), hamster (Frost et al., 1979), rat (Land and Lund, 1979), and opossum (Cavalcante and Rocha Miranda, 1978). Data from a variety of nonmammalian and mammalian models have suggested that chemoaffinity cues provide the basis for the fundamental topographical order of retinotectal projections, and this coarse representation is subsequently refined by activity-dependent mechanisms (Sperry, 1963; Walter et al., 1987; Fraser, 1992; Drescher et al., 1997; Ruthazer et al., 2003; Ruthazer and Cline, 2004). There is good reason to assume that similar processes are involved in determining the adultlike pattern of sensory afferents to the multisensory layers of the SC, but observations of the development of its retinotectal projections are hampered by weak input, and less attention has been directed at the maturation of nonvisual inputs. Nevertheless, it has been noted that somatosensory tectopetal afferents are already widespread in the multisensory layers of the SC at birth, and they too appear to have an adultlike configuration (see Stein et al., 1982, 1983; McHaffie et al., 1986, 1988). The apparent maturity of the representation of tectopetal sensory afferents (inputs) at birth is in contrast to the significant immaturity in at least one of its tectofugal efferent (output) sensory projectionsdthose to the lateral geniculate nucleus (LGN). This transient tectogeniculate projection is topographic and part of the ascending wing of the SC’s visual projection that eventually reaches cortex via polysynaptic pathways. In adults the tectogeniculate projection is confined to the ventral C layers, but in the neonate it extends across all layers of the LGN and also extends into the medial interlaminar nucleus. It is retained for approximately three postnatal weeks, long after visual function has been initiated, and is eliminated by mechanisms that are not yet fully understood (see Stein et al., 1985). Its immaturity contrasts with the more rapidly maturing tectofugal pathways involved in motor functions (see below).

Development of the superior colliculus/optic tectum Chapter | 3

65

Despite the adultlike pattern of retinotectal projections at birth, the visual system is not yet functional. Visual responsiveness begins in superficial layer SC neurons at 6 days of age. At this time the eyelids are still closed (they open naturally at 7e11 days postnatal), visually responsive neurons are rare, and they are clustered together between ineffective loci (Stein et al., 1973a; Kao et al., 1994). This reflects either random onset of visual activity, via random functional coupling of retinotectal afferents, or their SC target neurons, and/or the random openings in the vascular networks around the lens which begin at about the time of eye-opening (see Freeman and Lai, 1978). Most active sites are located in the middle portion of the structure and in its most superficial aspect. Nevertheless, the general topographic organization established by the afferent projections is apparent at this time (Kao et al., 1994, see also Fig. 3.5). Curiously, the initial visual activity is restricted not only across the horizontal aspect of the SC, but, as noted above, also in its vertical aspect. Neurons in the most superficial portion of the SC, where inputs from retinal W cells terminate, begin responding earliest. Neurons deeper in the superficial layers, where Y cell inputs dominate, develop later and visual

FIGURE 3.5 The maturation of the visuotopic map in the superficial superior colliculus. Data from three age groups is shown. Each contains a diagram of the visual field on the left and a schematic of the dorsal surface of the SC on the right. Circles on the SC represent electrode penetrations (filled ¼ no visual activity, open ¼ visual activity). Correspondence between the electrode penetrations and visual receptive fields are shown by numbers and letters (only the “best area” of a receptive field was mapped). No visual activity was encountered before 6 days postnatal (dpn), but was already represented in a maplike pattern when first encountered. Adjacent electrode penetrations in the SC yielded visual activity at adjacent sites in visual space. By 9e10 dpn most SC locations had become responsive to visual stimulation. From Kao, C.Q., Stein, B.E., Coulter, D.A., 1994, Postnatal Development of Excitatory Synaptic Function in Deep Layers of SC.

66

PART | I Circuit development

FIGURE 3.6 During maturation, superior colliculus visual activity progresses from the top to the bottom of the superficial layers. Vertical lines illustrate electrode penetrations through coronal sections of the SC in each age group. Sections are arranged rostral-caudal. Each region of visual activity is represented by a thick vertical black line spanning the distance between the first and last location of visually active neurons. Note that the visually active spans increase with age so that by 9e10 dpn visual activity is evident throughout the depths of the superficial layers. No deep layer visual activity was present at this time. From Kao, C.Q., Stein, B.E., Coulter, D.A., 1994, Postnatal Development of Excitatory Synaptic Function in Deep Layers of SC.

responsiveness in the subjacent multisensory layers develops last (Kao et al., 1994; Wallace and Stein, 1997). The bases for this particular pattern of functional development remain obscure, but it is occurring at the same time that the number of active loci across the horizontal aspect of the structure increases to yield a continuous retinotopy that underlies its visuotopic organization (Fig. 3.6). These early responsive neurons are functionally immature and have many properties that typify functionally immature neurons elsewhere in the nervous system. Their receptive fields are very large, and require either long duration flashed stationary targets or very slowly moving stimuli for activation. They fatigue readily with repeated stimulation, and require very long interstimulus intervals to respond to sequential stimuli, and have very long latencies (Stein et al., 1973b; Kao et al., 1994). They also lack binocularity and direction selectivity, properties that characterize these neurons when mature and are believed to facilitate orientation to moving targets. These properties develop over a 6e8 week period. It is not clear what, if any, visual behaviors these early responding neurons can support, as clear overt visual function is not observed in the cat until 2e3 weeks after birth when they have already matured considerably (Fox et al., 1978; Stein et al., 1973a; Norton, 1974). Nevertheless, the physiological maturation of the receptive field properties of these superficial layer neurons takes two or more months to reach their adult status. This reflects, in part, the physiological maturation of visual tectopetal afferents; especially those from cortex (e.g., see Stein and Gallagher, 1981). Although sensory experience is critical for the formation of many neural properties in the central nervous system, the fundamental features of SC visuotopy and the responsiveness of its constituent neurons appear to progress independent of visual experience. Superficial and deep (multisensory) layer neurons in dark-reared cats show robust visual responses and adultlike resistance to fatigue with repeated stimulation, albeit their receptive fields remain comparatively large, and their more sophisticated properties develop more gradually. Similarly, in the newborn monkey, whose ocular properties are far more mature than those of the cat, SC neurons already show a well-ordered and continuous superficial layer visuotopy. Although the receptive fields of its neurons are also larger than in the adult, and their visual latencies are longer, the neurons respond robustly to visual stimuli and already have many adultlike response properties (Wallace et al., 1997). The maturational differences in the newborn monkey and cat SC likely reflect both the shorter gestational period in the cat and its higher dependence on corticotectal inputs for construction of its neuronal response properties (Stein, 1984). Development of deep layer sensory topographies: Attention to the maturation of neurons in the deep layers has been directed more toward their functional properties rather than their topographies. Thus, much of what we know about this feature of their development has to be inferred. The somatosensory and auditory receptive fields of neonatal neurons are, like their visual counterparts, extremely large. Neonatal somatosensory receptive fields cover much of the contralateral

Development of the superior colliculus/optic tectum Chapter | 3

67

body and early auditory receptive fields are “omnidirectional,” having receptive fields that encompass all of contralateral auditory space. They gradually shrink in size over the first few months of life, and in the absence of evidence that neurons have receptive fields that are “mislocated,” it is assumed that the resolution of their maps gradually increases as individual receptive fields contract. The functional changes that ensue in these layers as a consequence of development are dealt with in the section dealing with the maturation of multisensory integration. Here examples of receptive field development are also provided.

3.5.3 The development of multisensory neurons Although a great deal of effort has been expended understanding the maturation of the superficial layer visuotopy to understand the maturation of SC-mediated orientation behaviors, as noted earlier, it is actually the deep (multisensory) layer visual responsiveness which is most closely linked to this function. And these layers develop more slowly. Because the maturation of these neurons has been most closely studied in cat, unless otherwise stated, the descriptions in the text below relate to this species. Because neurons in the multisensory layers become responsive to visual cues only after 3 postnatal weeks (Kao et al., 1994; Wallace and Stein, 1997), it is obvious that neither the weak direct retinal projection to these layers (see Beckstead and Frankfurter, 1983; Berson and McIlwain, 1982) nor the relay of visual input from active neurons in the superficial layers (e.g., Grantyn et al., 1984; Moschovakis and Karabelas, 1985; Behan and Appell, 1992) are capable of activating these neurons before this time. Yet somatosensory- and auditory-evoked neuronal activity is already apparent, albeit the responsive neurons also have the large receptive fields and tendency to fatigue that are characteristic of neonatal neurons. This developmental chronology (somatosensory first, auditory second, visual last) parallels the development of sensory-evoked orientation behaviors (Fox et al., 1978; Norton, 1974; Villablanca and Olmstead, 1979). At this time there is a gradual increase in the number of SC neurons that can respond to multiple sensory inputs. Thus, each of the different possible convergence patterns of responsiveness to visual, and/or auditory, and or somatosensory inputs become evident. However, one of the hallmark features of multisensory SC neurons is their ability to use information from the different senses synergistically. This process of integrating cross-modal inputs markedly facilitates SCmediated behavior and begins at about 4 postnatal weeks.

3.5.4 Superficialedeep (multisensory) layer maturational delay As noted earlier, visual responsiveness occurs comparatively early in superficial layer neurons, beginning before the end of the first week, and comparatively late in the multisensory layers, where it begins at 3e4 postnatal weeks. This maturational delay relative to superficial layer visual activity underscores the significant maturational difference between the sensoryspecific properties of superficial layer neurons and their multisensory counterparts in the deep layers. Surely synthesizing information from multiple senses is a more complex process than responding to any one of them individually, and requires greater maturational time. These superficialedeep layer differences also extend to the functional roles to which their information processing contributes. Although both neuronal populations probably share functions related to perception and overt behavior, superficial layer visual neurons are believed to contribute more heavily to the former and deep layer multisensory neurons to the latter. Developing the ability to integrate cross-modal inputs is a multistep process. Neurons first develop responsiveness to a single sensory input, then to at least two different (i.e., cross-modal) inputs, and then can finally develop the ability to integrate the information carried in multiple sensory channels. The maturation of the various possible multisensory convergence patterns follows closely the chronology of unisensory maturation. Thus, somatosensory-auditory neurons are the first multisensory neurons to appear. They become evident 10e12 days postparturition, and visual-nonvisual neurons become evident at approximately 3 postnatal weeks, as soon as visual responsiveness begins in the multisensory layers. But, it still takes many weeks of maturation before the adultlike incidence of the various modality-convergence patterns is reached and the adultlike incidence of neurons capable of multisensory integration is achieved. As might be suspected from their unisensory counterparts, the receptive fields of the initial multisensory neurons are very large. They contract over several months, thereby progressively increasing the spatial resolution of the individual sensory maps to which they contribute and their spatial register with one another (Fig. 3.7). The changes in receptive field size are accompanied by an increase in response vigor and reliability, as well as a decrease in response latency. These functional changes reflect a combination of developmental factors taking place in afferent systems, as well as within the internal circuitry of the SC.

68

PART | I Circuit development

FIGURE 3.7 The maturation of multisensory receptive fields. Shown are exemplar SC receptive fields from visual-auditory (left) and visualsomatosensory (right) neurons, and were mapped in animals aged 22e135dpn. Note that receptive fields become progressively smaller at older ages, increasing their spatial resolution and their register with one another. Adapted from Wallace, M.T., Stein, B.E., 1997. Development of multisensory neurons and multisensory integration in cat superior colliculus. J. Neurosci. 17, 2429e2444. (see also Stein and Rowland, 2011).

3.5.5 The development of multisensory integration The ability to integrate cross-modal inputs is delayed until at least 1 month postnatal (Wallace and Stein, 1997), and before it develops, the cross-modal responses of multisensory neurons look like the responses elicited by one (the more effective) of the component stimuli alone. At this age, however, some neurons respond to cross-modal stimuli with responses that significantly exceed the presumptive unisensory responses, but such neurons are rare at this time. The incidence of these integrative neurons progressively increases with age, but the mature condition is not reached until the animal is several months old, and has had a great deal of sensory experience. The absence of an ability to integrate cross-modal inputs in the neonatal cat’s SC is not due to the general immaturity of these neurons. The macaque monkey is born much later in development than is the cat. Its eye and ears are open at birth and it sees and hears quite well, and unlike the cat, it already has multisensory SC neurons in its SC. But they cannot integrate their cross-modal inputs to produce response enhancement (Wallace and Stein, 2001), and, therefore, respond to these stimuli very differently than do their adult counterparts (Wallace et al., 1996). The likely reason is that they have not yet had the necessary sensory experience. Apparently, postnatal sensory experience is not critical for the appearance of multisensory neurons, but these observations strongly suggest that it is necessary for them to integrate their different sensory inputs. Observations from experiments with human subjects are consistent with this hypothesis (e.g., Neil et al., 2006; Gori et al., 2008; Putzar et al., 2007), though the critical direct observations showing that the newborn cannot yet use cross-sensory cues synergistically, are not yet available. Nevertheless, newborn human infants are capable of engaging in a host of multisensory tasks, the best known of which is cross-modal matching (see Stein et al., 2010 for further discussion).

3.5.6 The impact of sensory experience on the maturation of multisensory integration As noted earlier when discussing multisensory integration, inputs from association cortex (e.g., primarily from the anterior ectosylvian sulcus, or AES, in cat) have been found to be essential for SC multisensory integration. Thus, one might expect that the functional coupling of this input with multisensory SC neurons is a necessary precondition for the development of SC multisensory integration, and this appears to be the case. Anatomically, the corticotectal projections from AES have already grown into the multisensory layers of the SC prior to birth and long before SC neurons are responsive to multiple sensory inputs (McHaffie et al., 1988). However, they are unlikely to be functional at this time. Their critical contribution

Development of the superior colliculus/optic tectum Chapter | 3

69

FIGURE 3.8 The maturation of multisensory integration parallels the functional coupling of AES-SC projections. The development increase in the incidence of SC neurons capable of multisensory integration is paralleled by the effectiveness of AES deactivation in blocking the expression of this capability. Nearly all SC neurons lost this ability during AES deactivation, regardless of age (the number of neurons tested at every time point is shown in parentheses). The small proportion of SC neurons whose multisensory integration capability was unaffected by AES deactivation likely depended on an adjacent area of association cortex (i.e., the rostral lateral suprasylvian cortex, see Jiang et al., 2001). From Wallace, M.T., Stein, B.E., 2000. Onset of cross-modal synthesis in the neonatal superior colliculus is gated by the development of cortical influences. J. Neurophysiol. 83, 3578e3582.

to multisensory integration becomes obvious as soon as a neuron in the SC exhibits this capability. As noted earlier, this begins to first happen for some rare neurons at about 1 month of age. Now deactivating AES eliminates that capacity as effectively as it does in adulthood. The neuron’s responses to the cross-modal stimulus combination are now no longer significantly different from its response to the most effective component stimulus (Wallace and Stein, 2000, see also Fig. 3.8). Although the association cortex component of the SC circuit appears to be critical for multisensory integration, it is not sufficient. Appropriate sensory experience is also a key factor. This early phase of life is one in which the brain is learning about external events and the statistical relationships among stimuli derived from the same event. Cross-modal stimuli that are produced by the same event are temporally and spatially coincident or at least proximate in space and time. These relationships must be learned (e.g., via Hebbian learning rules) and somehow represented in the underlying circuitry responsible for multisensory integration. Given the arbitrary nature of many cross-modal relationships, it is difficult to conceive of an effective scheme of incorporating this information without actual experience. Experience with cross-modal stimuli establishes a principled way of interpreting and interacting with external events so that only a select group of cross-modal stimuli will produce multisensory integration. For example, the brain learns to expect that some visual and auditory cues are linked to the same event, based on their location and timing. This information is used to establish principles for categorizing cross-modal stimuli derived from the same, or different, events. To test the hypothesis that sensory experience plays a key role in the development of multisensory integration capabilities, cats were raised in darkness in order to preclude visual-nonvisual experiences. Normally, the physiology of SC multisensory integration is reached at 3e4 months of age (Wallace and Stein, 1997), and these dark-reared animals were not studied until at least 6 months of age. Visual experience was not essential for the appearance of visually responsive neurons, and such neurons were common in these dark-reared animals, as well as in the newborn monkey (see Wallace and Stein, 2001). Each of the modality-convergence patterns was also well represented among the neurons encountered (Wallace et al., 2004). However, their receptive fields were still quite large and more typical of the neonate than the mature animal (Fig. 3.9). The failure to contract their receptive fields was indicative of their physiological immaturity, and this was most apparent in their inability to integrate cross-modal cues. As shown in Fig. 3.9, their multisensory response to a crossmodal visual-nonvisual stimulus was approximately equal to their response to the visual component stimulus alone. Multisensory responses that approximate the response to the most effective component stimulus are typical of neonatal

70

PART | I Circuit development

FIGURE 3.9 Multisensory neurons have large receptive fields and lack multisensory integration capabilities after dark-rearing. The visual and auditory receptive fields and the multisensory responses of a typical normal SC multisensory neuron are shown at the top. The bar graph to the right shows the characteristic enhanced response to the cross-modal pair of stimuli in spatial and temporal register. In contrast, the neuron below, taken from the SC of a dark-reared animal, has much larger receptive fields, and shows no evidence of multisensory response enhancement to the cross-modal stimulus. A, auditory stimulus; V, visual stimulus; VA, cross-modal stimulus. Adapted from Wallace, M.T., Perrault, T.J., Jr, Hairston, W.D., Stein, B.E., 2004. Visual experience is necessary for the development of multisensory integration1. J. Neurosci. 24, 9580e9584.

neurons that have not yet developed their capacity for multisensory integration and of adult multisensory neurons deprived of the critical input from association cortex. Similar results have been obtained in human subjects who have had their early vision compromised by congenital cataracts. They also show persistent visual-nonvisual integration deficits even many years after the cataracts have been removed (Putzar et al., 2007). But does this abnormal developmental outcome reflect a lack of multisensory experience, or just a lack of visual experience? As discussed above, the SC is appreciated as a visually dominated structure, and there is evidence that the visual modality is crucial for guiding the alignment of cross-modal receptive fields during development (Brainard and Knudsen, 1998). It is plausible that disruptions of normal visual development compromise a variety of SC functions, and that defective multisensory integration in dark-reared animals might reflect lack of visual experience rather than a lack of multisensory experience. Convergent evidence was required. In one manipulation, animals were raised in omnidirectional sound rooms in which loud broadband noise effectively masked experience with patterned auditory (and, crucially, paired auditory-non-auditory) stimuli. Like the dark-reared animals, these “noise-reared” animals developed complements of neurons responsive to visual, auditory, and somatosensory stimulidbut did not integrate spatiotemporally concordant auditory-visual pairs to produce enhanced responses (Xu et al., 2012, 2017). Like the human patients described above, these defects did not quickly remit when animals were moved to normal environments as adults (Xu et al., 2017). The same lack of enhancement was observed in animals that were reared in the dark and given daily exposure to independently appearing visual and auditory stimuli with randomized location and timing (Xu et al., 2012). In a detailed study of these manipulations interrogating trisensory (visual-auditory-somatosensory) neurons, it was found that individual cells could express multiple cross-modal processing rules: neurons would integrate pairings of modalities they had experienced together (e.g., visual-somatosensory in noise-reared animals, auditory-somatosensory in the dark-reared animals) but not those involving restricted modalities (Xu et al., 2015). These observations support the hypothesis that cross-modal experience is a critical factor in the maturation of the capacity to integrate multisensory cues (e.g., see Fig. 3.10), but do not provide information about whether the nature of that early experience determines the integrative principles themselves. To test this possibility, animals were reared from birth to 6 months of age in a dark room in which pairs of visual and auditory stimuli were periodically presented, but these stimuli were always spatially disparate (Wallace and Stein, 2007). The properties of their multisensory SC neurons were then assessed. Once again, visually responsive neurons were common, as were visual-nonvisual neurons. Yet, the majority of visualauditory neurons appeared to have been unaffected by experience with the disparate visual-auditory cues, and had

Development of the superior colliculus/optic tectum Chapter | 3

71

FIGURE 3.10 Multisensory enhancement is evident in the human superior colliculus using fMRI. Coronal (left) and sagittal (right) sections reveal high degrees of enhanced SC multisensory activity (red indicates elevated BOLD responses). Elevated activity is in response to visual-auditory as opposed to the best of these stimuli (i.e., visual) individually. Adapted from Calvert, G.A., Hansen, P.C., Iversen, S.D., Brammer, M.J., 2001. Detection of audiovisual integration sites in humans by application of electrophysiological criteria to the BOLD effect. Neuroimage 14, 427e438.

properties characteristic of neonates. Their receptive fields were very large and they were unable to integrate these crossmodal cues. But, there was a substantial minority of such neurons that did appear to have incorporated the visual-auditory experience. Their receptive fields were somewhat smaller (but still quite large) and were in poor spatial register with one another. Some neurons had visual and auditory receptive fields that were elongated along their horizontal axes and had only small portions overlapping one another. In other cases there was no receptive field overlap, a configuration that is highly unusual in normal animals, but one that was consistent with the early visual-auditory experience of these animals. Most important in the present context was that they could integrate visual-auditory stimuli. But the cross-modal stimuli had to be spatially disparate in order to simultaneously fall within their respective receptive fields. An example of such a neuron is shown in Fig. 3.11. Collectively the data reveal that experience is essential for the development of multisensory integration and that the nature of the experience directs the formation of the underlying neural circuits through which this integration is achieved. Although the specifics of that circuit remain to be fully explored (e.g., see Fuentes-Santamaria et al., 2006, 2008a,b, 2009), cortex is known to play a critical role (e.g., see Wallace and Stein, 1994; Jiang et al., 2001; Stein, 2005; Alvarado et al., 2009), and its ablation early in life precludes the maturation of SC multisensory integration (Jiang et al., 2006, 2007). It appears that early experience is essential for the brain to learn the statistics of cross-modal events, and there is good reason to suspect that this experience exerts its critical impact on the AES-SC projection. This possibility was explored by Rowland et al. (2014), who deactivated AES and adjacent association cortex ipsilaterally during the period in which SC multisensory integration normally develops. This deprived the cortex of crossmodal experience, but did not compromise the responsiveness of SC neurons to these stimuli via many other input channels. Chronic deactivation was accomplished by implanting muscimol-infused pledgets of Elvax (a polymer) over association cortex. The GABAa agonist is slowly released from the polymer, deactivating underlying neurons. When the polymer is depleted or removed, cortical activity returns rapidly, and is once again responsive to environmental stimuli. Behavioral and physiological studies were then done when the animals had reached 1 year of age. Animals appeared normal in their ability to respond to visual stimuli in both visual fields. They also benefited from visual-auditory cues in the ipsilateral hemifield as much as did normal animals. However, they were severely compromised in their multisensory responses to cross-modal stimuli in their contralateral hemifield. Cross-modal stimuli in this hemifield were no more effective in facilitating SC-mediated behavior than was the visual stimulus alone. Thus, as expected, the inability of association cortex to monitor the statistics of visual-auditory events compromised the maturation of the circuit necessary for SC multisensory integration. This was also evident in the inability of multisensory neurons in the ipsilateral SC to integrate visual-auditory information. That these deficits in multisensory integration were apparent long after cortex was once again active (a period far longer than that required for its normal acquisition in early life) could be interpreted as reflecting a “critical” or “sensitive” period for instantiating this process. If the former were the case, this capacity would never develop. This possibility was examined in some of the same animals which were retained for several years. These animals were then retrained in the behavioral task and subsequently tested as before. They appeared to be normal, indicating that they acquired the ability to integrate visualauditory cues later in life. That this acquisition likely involved multisensory SC neurons was indicated by the finding that

72

PART | I Circuit development

FIGURE 3.11 Early experience with spatially disparate visual-auditory stimuli results in atypical requirements for their integration. Shown is an SC neuron from an animal reared from birth to 6 months of age with simultaneous, though spatially disparate, visual-auditory cues. The neuron’s visual and auditory receptive fields were atypical by being nonoverlapping. When visual-auditory stimuli were spatiotemporally coincident and within the visual (left) or auditory (center) receptive fields, the multisensory response was no greater than that evoked by the most effective component stimulus. But, when the two stimuli were disparate in space and simultaneously presented within their respective receptive fields, they elicited a significantly enhanced multisensory response. Thus, the neuron integrated spatially disparate visual-auditory stimuli as normal animals integrate spatially concordant visualauditory stimuli, a seeming “reversal” of the spatial principle. Adapted from Wallace, M.T., Stein, B.E., 2007. Early experience determines how the senses will interact. J. Neurophysiol. 97, 921e926.

the multisensory responses of ipsilateral SC neurons, while not completely normal, did exhibit multisensory response enhancement to spatiotemporally coincident visual-auditory stimuli. This capability of adults to acquire multisensory enhancement capabilities was directly demonstrated in dark-reared and noise-reared animals that were repeatedly exposed in a “sensory training” paradigm to spatiotemporally concordant pairs of visual-auditory cues (Yu et al., 2010, 2013; Xu et al., 2017). Perhaps surprisingly, this sensory training could be accomplished even under anesthesia, suggesting that the acquisition of these capabilities is not dependent on reward or even an alert brain. This adult development was shown to be “site-specific”: only neurons whose receptive fields encroached on the exposure locations developed multisensory enhancement capabilities. Normal levels of multisensory enhancement were observed within a few months of weekly exposure sessions (Yu et al., 2010, 2013; Xu et al., 2017). This rapidity contrasts with the much longer time scale required for this adult development in normal environments; however, the reason for this difference is currently unknown. Motor development: Much less is known about the development of motor than sensory properties of SC neurons. Although there is little doubt that the fundamental properties involved in gaze shifts are not species-specific, most of the available information comes from studies in one preparation, the alert monkey trained to make gaze shifts. For obvious reasons, technical limitations make conducting such studies in neonates quite difficult, and especially of making any direct correlations with the maturation of sensory properties in the primary sensory developmental model, the cat. Nevertheless, for neonatal SC sensory responses to have any direct impact on behavior, their efferents to the brainstem and/or spinal cord must be in place and capable of carrying tectofugal signals.

Development of the superior colliculus/optic tectum Chapter | 3

73

These pathways can be seen exiting the SC and contacting targets involved in eye movements within hours of birth in the cat. The target areas include the central gray overlying the oculomotor nucleus, an area into which oculomotor dendrites project, as well as segments of the pontine and medullary reticular formation that connect to the abducens nucleus (Stein et al., 1982). Tectofugal projections also reach the cervical spinal cord. This region is involved in controlling head and limb movement, and these projections are described in detail in adults by Huerta and Harting (1982a,b). Using electrical stimulation of the SC, this motor pathway has already been demonstrated to have functional capabilities in 2-day-old cats (Stein et al., 1980), days before natural auditory stimuli can activate multisensory layer neurons and weeks before visual stimuli are effective. Such electrical stimulation elicits eye, ear, neck, whisker, and limb movements, albeit with higher threshold and lower reliability than in adults. Furthermore, the motor topography in the SC is already evident. Stimulation of homotopic loci in each SC produces mirror image eye movements. Nevertheless, many SC stimulation sites proved to be ineffective at this stage of maturation. Presumably, this is due to the immaturity of the SC, as direct stimulation of the oculomotor nucleus evoked reliable eye movements. It is not yet known, however, whether tactile stimuli, which can already activate SC neurons in preterm kittens, can initiate SC-mediated movements in neonates. But, if as suggested by some (e.g., see Sparks and Porter, 1983; Sparks, 1986), the SC is organized in motor coordinates, it could indicate that early motor function could influence later auditory and visual organization. It would also be consistent with the idea (e.g., see Hein et al., 1979) that eye movement-generated movement of an image across the retina is necessary for interpreting that visual image, and that during development, the former precedes the latter (though spontaneous or vestibular cues could produce the necessary eye movements in neonatal cats (see Fish and Windle, 1932).

3.6 Summary It is important to acknowledge that this chapter has focused on certain anatomical and physiological aspects of SC development for which functional implications are readily apparent. Thus, for example, the formation of topographic representations and the development of multisensory integrative capabilities have obvious implications for producing behavioral output and behavioral correlates in developing animals. In doing so, however, voluminous bodies of literature on the chronology of synapse formation and the appearance of histochemical markers that parallel SC functional development has been ignored. Likewise, with few exceptions, the focus has been primarily on the SC, but there is also a wealth of comparative data pertaining to the OT from studies of a wide array of nonmammalian species. While the same general principles of development generally apply, maturational differences dictated by the varied and specialized environmental niches that these animals occupy would require a discussion that is beyond the present scope. Lastly, it is important to point out what already may be obviousdthat knowledge of SC development disproportionately favors aspects relevant to its sensory capabilities. This weighting belies the truly integrated sensorimotor function of the SC and highlights a rather substantial gap in our knowledge of how the SC matures into its adult role. We can only hope that future studies will address this imbalance to provide a more complete picture of the SC developmental chronology.

Acknowledgments The research described here was supported in part by NIH grants NS36916 and EY016716 and EY12389.

References Alvarado, J.C., Rowland, B.A., Stanford, T.R., Stein, B.E., 2008. A neural network model of multisensory integration also accounts for unisensory integration in superior colliculus. Brain Res. 1242, 13e23. Alvarado, J.C., Stanford, T.R., Rowland, B.A., Vaughan, J.W., Stein, B.E., 2009. Multisensory integration in the superior colliculus requires synergy among corticocollicular inputs. J. Neurosci. 29, 6580e6592. Alvarado, J.C., Stanford, T.R., Vaughan, J.W., Stein, B.E., 2007. Cortex mediates multisensory but not unisensory integration in superior colliculus. J. Neurosci. 27, 12775e12786. Armstrong, K.M., Fitzgerald, J.K., Moore, T., 2006. Changes in visual receptive fields with microstimulation of frontal cortex. Neuron 50, 791e798. Armstrong, K.M., Moore, T., 2007. Rapid enhancement of visual cortical response discriminability by microstimulation of the frontal eye field. Proc. Natl. Acad. Sci. U. S. A. 104, 9499e9504. Auroy, P., Irthum, B., Woda, A., 1991. Oral nociceptive activity in the rat superior colliculus. Brain Res. 549, 275e284. Awh, E., Belopolsky, A.V., Theeuwes, J., 2012. Top-down versus bottom-up attentional control: a failed theoretical dichotomy. Trends Cogn. Sci. 16, 437e443. Basso, M.A., May, P.J., 2017. Circuits for action and cognition: a view from the superior colliculus. Annu. Rev. Vis. Sci. 3, 197e226.

74

PART | I Circuit development

Beckstead, R.M., Frankfurter, A., 1983. A direct projection from the retina to the intermediate gray layer of the superior colliculus demonstrated by anterograde transport of horseradish peroxidase in monkey, cat and rat. Exp. Brain Res. 52, 261e268. Behan, M., Appell, P.P., 1992. Intrinsic circuitry in the cat superior colliculus: projections from the superficial layers. J. Comp. Neurol. 315, 230e243. Bell, A.H., Corneil, B.D., Meredith, M.A., Munoz, D.P., 2001. The influence of stimulus properties on multisensory processing in the awake primate superior colliculus. Can. J. Exp. Psychol. 55, 123e132. Bell, A.H., Meredith, M.A., Van Opstal, A.J., Munoz, D.P., 2005. Crossmodal integration in the primate superior colliculus underlying the preparation and initiation of saccadic eye movements. J. Neurophysiol. 93, 3659e3673. Benedetti, F., Ferro, I., 1995. The effects of early postnatal modification of body shape on the somatosensory-visual organization in mouse superior colliculus. Eur. J. Neurosci. 7, 412e418. Berson, D.M., McIlwain, J.T., 1982. Retinal Y-cell activation of deep-layer cells in superior colliculus of the cat. J. Neurophysiol. 47, 700e714. Bonds, A.B., Freeman, R.D., 1978. Development of optical quality in the kitten eye. Vis. Res. 18, 391e398. Brainard, M.S., Knudsen, E.I., 1998. Sensitive periods for visual calibration of the auditory space map in the barn owl optic tectum. J. Neurosci. 18 (10), 3929e3942. Calvert, G.A., Hansen, P.C., Iversen, S.D., Brammer, M.J., 2001. Detection of audio-visual integration sites in humans by application of electrophysiological criteria to the BOLD effect. Neuroimage 14, 427e438. Calvert, G.A., Spence, C., Stein, B.E., 2004. The Handbook of Multisensory Processes. MIT Press, Cambridge, MA. Carrasco, M., 2011. Visual attention: the past 25 years. Vis. Res. 51, 1484e1525. Cavalcante, L.A., Rocha-Miranda, C.E., 1978. Postnatal development of retinogeniculate, retinopretectal and retinotectal projections in the opossum. Brain Res. 146, 231e248. Cavanaugh, J., Wurtz, R.H., 2004. Subcortical modulation of attention counters change blindness. J. Neurosci. 24, 11236e11243. Clark, K., Squire, R.F., Merrikhi, Y., Noudoost, B., 2015. Visual attention: linking prefrontal sources to neuronal and behavioral correlates. Prog. Neurobiol. 132, 59e80. Connor, C.E., Gallant, J.L., Preddie, D.C., Van Essen, D.C., 1996. Responses in area V4 depend on the spatial relationship between stimulus and attention. J. Neurophysiol. 75, 1306e1308. Connor, C.E., Preddie, D.C., Gallant, J.L., Van Essen, D.C., 1997. Spatial attention effects in macaque area V4. J. Neurosci. 17, 3201e3214. Coulter, J.D., Bowker, R.M., Wise, S.P., Murray, E.A., Castiglioni, A.J., Westlund, K.N., 1979. Cortical, tectal and medullary descending pathways to the cervical spinal cord. Prog. Brain Res. 50, 263e279. Cynader, M., Berman, N., 1972. Receptive-field organization of monkey superior colliculus. J. Neurophysiol. 35, 187e201. Dräger, U.C., Hubel, D.H., 1975. Responses to visual stimulation and relationahip between visual, auditory, and somatosensory inputs in mouse superior colliculus. J. Neurophysiol. 38, 690e713. Drescher, U., Bonhoeffer, F., Müller, B.K., 1997. The Eph family in retinal axon guidance. Curr. Opin. Neurobiol. 7, 75e80. du Lac, S., Knudsen, E.I., 1990. Neural maps of head movement vector and speed in the optic tectum of the barn owl. J. Neurophysiol. 63, 131e146. Edwards, S.B., 1980. The deep layers of the superior colliculus: their reticular characteristics and structural organization. In: Hobson, J.A., Brazier, M.A. (Eds.), The Reticular Formation Revisited. Raven Press, New York, pp. 193e209. Edwards, S.B., Henkel, C.K., 1978. Superior colliculus connections with the extraocular motor nuclei in the cat. J. Comp. Neurol. 179, 451e467. Edwards, S.B., Ginsburgh, C.L., Henkel, C.K., Stein, B.E., 1979. Sources of subcortical projections to the superior colliculus in the cat. J. Comp. Neurol. 184, 309e329. Fecteau, J.H., Munoz, D.P., 2006. Salience, relevance, and firing: a priority map for target selection. Trends Cogn. Sci. 10, 382e390. Feldon, S., Feldon, P., Kruger, L., 1970. Topography of the retinal projection upon the superior colliculus of the cat. Vis. Res. 10, 135e143. Fish, M.W., Windle, W.F., 1932. The effect of rotatory stimulation on the movements of the head and eyes in newborn and young kittens. J. Comp. Neurol. 54, 103e107. Fox, P., Chow, K.L., Kelly, A.S., 1978. Effects of monocular lid closure on development of receptive-field characteristics of neurons in rabbit superior colliculus. J. Neurophysiol. 41, 1359e1372. Fraser, S.E., 1992. Patterning of retinotectal connections in the vertebrate visual system. Curr. Opin. Neurobiol. 2, 83e87. Freeman, R.D., Lai, C.E., 1978. Development of the optical surfaces of the kitten eye. Vis. Res. 18, 399e407. Frost, D.O., So, K.F., Schneider, G.E., 1979. Postnatal development of retinal projections in Syrian hamsters: a study using autoradioagraphic and anterograde techniques. Neuroscience 4, 1649e1677. Fuentes-Santamaria, V., Alvarado, J.C., McHaffie, J.G., Stein, B.E., 2009. Axon morphologies and convergence patterns of projections from different sensory-specific cortices of the anterior ectosylvian sulcus onto multisensory neurons in the cat superior colliculus. Cerebr. Cortex 19, 2902e2915. Fuentes-Santamaria, V., Alvarado, J.C., Stein, B.E., McHaffie, J.G., 2008. Cortex contacts both output neurons and nitrergic interneurons in the superior colliculus: direct and indirect routes for multisensory integration. Cerebr. Cortex 18, 1640e1652. Fuentes-Santamaria, V., McHaffie, J.G., Stein, B.E., 2008. Maturation of multisensory integration in the superior colliculus: expression if nitric oxide synthase and neurofilament SMI-32. Brain Res. 1242, 45e53. Fuentes-Santamaria, V., Stein, B.E., McHaffie, J.G., 2006. Neurofilament proteins are preferentially expressed in descending output neurons of the cat superior colliculus: a study using SMI-32. Neuroscience 138, 55e68. Gaither, N.S., Stein, B.E., 1979. Reptiles and mammals use similar sensory organizations in the midbrain. Science 205, 595e597. Gandhi, N.J., Katnani, H.A., 2011. Motor functions of the superior colliculus. Annu. Rev. Neurosci. 34, 205e231. Gillmeister, H., Eimer, M., 2007. Tactile enhancement of auditory detection and perceived loudness. Brain Res. 1160, 58e68.

Development of the superior colliculus/optic tectum Chapter | 3

75

Goldberg, M.E., Wurtz, R.H., 1972. Activity of superior colliculus in behaving monkey. I. Visual receptive fields of single neurons. J. Neurophysiol. 35 (4), 542e559. Goldberg, M.E., Wurtz, R.H., 1972. Activity of superior colliculus in behaving monkey. II. Effect of attention on neuronal responses. J. Neurophysiol. 35, 560e574. Gori, M., Del, V.M., Sandini, G., Burr, D.C., 2008. Young children do not integrate visual and haptic form information. Curr. Biol. 18, 694e698. Grantyn, A., Grantyn, R., 1982. Axonal patterns and sites of termination of cat superior colliculus neurons projecting in the tecto-bulbo-spinal tract. Exp. Brain Res. 46, 243e256. Grantyn, R., Ludwig, R., Eberhardt, W., 1984. Neurons of the superficial tectal gray. An intracellular HRP-study on the kitten superior colliculus in vitro. Exp. Brain Res. 55, 172e176. Graybiel, A.M., 1975. Anatomical organization of retinotectal afferents in the cat: an autoradiographic study. Brain Res. 96, 1e23. Groh, J.M., Sparks, D.L., 1996. Saccades to somatosensory targets. III. eye-position-dependent somatosensory activity in primate superior colliculus. J. Neurophysiol. 75, 439e453. Hall, W.C., Moschovakis, A., 2004. The Superior Colliculus: New Approaches for Studying Sensorimotor Integration. CRC Press, Boca Raton, FL. Harting, J.K., 1977. Descending pathways from the superior colliculus: an autoradiographic analysis in the rhesus monkey (Macaca mulatta). J. Comp. Neurol. 173, 583e612. Harting, J.K., Hall, W.C., Diamond, I.T., Martin, G.F., 1973. Anterograde degeneration study of the superior colliculus in Tupaia glis: evidence for a subdivision between superficial and deep layers. J. Comp. Neurol. 148, 361e386. Harting, J.K., Huerta, M.F., Frankfurter, A.J., Strominger, N.L., Royce, G.J., 1980. Ascending pathways from the monkey superior colliculus: an autoradiographic analysis. J. Comp. Neurol. 192, 853e882. Hartline, P.H., Kass, L., Loop, M.S., 1978. Merging of modalities in the optic tectum: infrared and visual integration in rattlesnakes. Science 199, 1225e1229. Hartline, P.H., Vimal, R.L., King, A.J., Kurylo, D.D., Northmore, D.P., 1995. Effects of eye position on auditory localization and neural representation of space in superior colliculus of cats. Exp. Brain Res. 104, 402e408. Hein, A., Vital-Durand, F., Salinger, W., Diamond, R., 1979. Eye movements initiate visual-motor development in the cat. Science 204, 1321e1322. Holcombe, V., Hall, W.C., 1981. Laminar origin of ipsilateral tectpontine pathways. Neuroscience 6, 255e260. Holcombe, V., Hall, W.C., 1981. The laminar origin and distribution of the crossed tectoreticular pathways. J. Neurosci. 1, 1103e1112. Huerta, M.F., Harting, J.K., 1982. Tectal control of spinal cord activity: neuroanatomical demonstration of pathways connecting the superior colliculus with the cervical spinal cord grey. Prog. Brain Res. 57, 293. Huerta, M.F., Harting, J.K., 1982. Projections of the superior colliculus to the supraspinal nucleus and the cervical spinal cord grey of the cat. Brain Res. 242, 326. Jay, M.F., Sparks, D.L., 1984. Auditory receptive fields in primate superior colliculus shift with changes in eye position. Nature 309, 345e347. Jiang, W., Jiang, H., Rowland, B.A., Stein, B.E., 2007. Multisensory orientation behavior is disrupted by neonatal cortical ablation. J. Neurophysiol. 97, 557e562. Jiang, W., Jiang, H., Stein, B.E., 2002. Two corticotectal areas facilitate multisensory orientation behavior. J. Cogn. Neurosci. 14, 1240e1255. Jiang, W., Jiang, H., Stein, B.E., 2006. Neonatal cortical ablation disrupts multisensory development in superior colliculus. J. Neurophysiol. 95, 1380e1396. Jiang, W., Wallace, M.T., Jiang, H., Vaughan, J.W., Stein, B.E., 2001. Two cortical areas mediate multisensory integration in superior colliculus neurons. J. Neurophysiol. 85, 506e522. Kadunce, D.C., Vaughan, J.W., Wallace, M.T., Benedek, G., Stein, B.E., 1997. Mechanisms of within- and cross-modality suppression in the superior colliculus2. J. Neurophysiol. 78, 2834e2847. Kanaseki, T., Sprague, J.M., 1974. Anatomical organization of pretectal nuclei and tectal laminae in the cat. J. Comp. Neurol. 158, 319e337. Kao, C.Q., Stein, B.E., Coulter, D.A., 1994. Postnatal Development of Excitatory Synaptic Function in Deep Layers of SC. Kim, B., Basso, M.A., 2008. Saccade target selection in the superior colliculus: a signal detection theory approach. J. Neurosci. 28, 2991e3007. King, A.J., Palmer, A.R., 1983. Cells responsive to free-field auditory stimuli in Guinea-pig superior colliculus: distribution and response properties. J. Physiol. 342, 361e381. King, A.J., Hutchings, M.E., Moore, D.R., Blakemore, C., 1988. Developmental plasticity in the visual and auditory representations in the mammalian superior colliculus. Nature 332, 73e76. Knudsen, E.I., 2011. Control from below: the role of a midbrain network in spatial attention. Eur. J. Neurosci. 33, 1961e1972. Krauzlis, R.J., Lovejoy, L.P., Zenon, 2013. A Superior colliculus and visual spatial attention. Annu. Rev. Neurosci. 36, 165e182. Kustov, A.A., Robinson, D.L., 1996. Shared neural control of attentional shifts and eye movements. Nature 384, 74e77. Land, P.W., Lund, R.D., 1979. Development of the rat’s uncrossed retinotectal pathway and its relation to plasticity studies. Science 205, 698e700. Larson, M.A., Stein, B.E., 1984. The use of tactile and olfactory cues in neonatal orientation and localization of the nipple. Dev. Psychobiol. 17, 423e436. Larson, M.A., McHaffie, J.G., Stein, B.E., 1987. Response properties of nociceptive and low-threshold mechanoreceptive neurons in the hamster superior colliculus. J. Neurosci. 7, 547e564. Lovejoy, L.P., Krauzlis, R.J., 2010. Inactivation of primate superior colliculus impairs covert selection of signals for perceptual judgments. Nat. Neurosci. 13, 261e266. Maunsell, J.H.R., 2015. Neuronal mechanisms of visual attention. Annu. Rev. Vis. Sci. 1, 373e391.

76

PART | I Circuit development

McAdams, C.J., Maunsell, J.H., 1999. Effects of attention on orientation-tuning functions of single neurons in macaque cortical area V4. J. Neurosci. 19, 431e441. McHaffie, J.G., Ogasawara, K., Stein, B.E., 1986. Trigeminotectal and other trigeminofugal projections in neonatal kittens: an anatomical demonstration with horseradish peroxidase and tritiated leucine. J. Comp. Neurol. 249, 411e427. McHaffie, J.G., Kruger, L., Clemo, H.R., Stein, B.E., 1988. Corticothalamic and corticotectal somatosensory projections from the anterior ectosylvian sulcus (SIV cortex) in neonatal cats: an anatomical demonstration with HRP and 3H-leucine. J. Comp. Neurol. 274, 115e126. McHaffie, J.G., Kao, C.Q., Stein, B.E., 1989. Nociceptive neurons in rat superior colliculus: response properties, topography, and functional implications. J. Neurophysiol. 62, 510e525. McIlwain, J.T., 1975. Visual receptive fields and their images in superior colliculus of the cat. J. Neurophysiol. 38, 219e230. McIlwain, J.T., 1991. Distributed spatial coding in the superior colliculus: a review. Vis. Neurosci. 6, 3e13. McPeek, R.M., Keller, E.L., 2002. Saccade target selection in the superior colliculus during a visual search task. J. Neurophysiol. 88, 2019e2034. McPeek, R.M., Keller, E.L., 2004. Deficits in saccade target selection after inactivation of superior colliculus. Nat. Neurosci. 7, 757e763. Meredith, M.A., Stein, B.E., 1990. The visuotopic component of the multisensory map in the deep laminae of the cat superior colliculus. J. Neurosci. 10, 3727e3742. Meredith, M.A., Clemo, H.R., Stein, B.E., 1991. Somatotopic component of the multisensory map in the deep laminae of the cat superior colliculus. J. Comp. Neurol. 312, 353e370. Metzger, R.R., Mullette-Gillman, O.A., Underhill, A.M., Cohen, Y.E., Groh, J.M., 2004. Auditory saccades from different eye positions in the monkey: implications for coordinate transformations. J. Neurophysiol. 92, 2622e2627. Middlebrooks, J.C., Knudsen, E.I., 1984. A neural code for auditory space in the cat’s superior colliculus. J. Neurosci. 4, 2621e2634. Moore, T., Armstrong, K.M., 2003. Selective gating of visual signals by microstimulation of frontal cortex. Nature 421, 370e373. Moore, T., Fallah, M., 2001. Control of eye movements and spatial attention. Proc. Natl. Acad. Sci. U. S. A. 98, 1273e1276. Moore, T., Fallah, M., 2004. Microstimulation of the frontal eye field and its effects on covert spatial attention. J. Neurophysiol. 91, 152e162. Moschovakis, A.K., Karabelas, A.B., 1985. Observations on the somatodendritic morphology and axonal trajectory of intracellularly HRP-labeled efferent neurons located in the deeper layers of the superior colliculus of the cat. J. Comp. Neurol. 239, 276e308. Moschovakis, A.K., Kitama, T., Dalezios, Y., Petit, J., Brandi, A.M., Grantyn, A.A., 1998. An anatomical substrate for the spatiotemporal transformation. J. Neurosci. 18, 10219e10229. Muller, J.R., Philiastides, M.G., Newsome, W.T., 2005. Microstimulation of the superior colliculus focuses attention without moving the eyes. Proc. Natl. Acad. Sci. U.S.A. 102, 524e529. Neil, P.A., Chee-Ruiter, C., Scheier, C., Lewkowicz, D.J., Shimojo, S., 2006. Development of multisensory spatial integration and perception in humans. Dev. Sci. 9, 454e464. Norton, T.T., 1974. Receptive-field properties of superior colliculus cells and development of visual behavior in kittens. J. Neurophysiol. 37, 674e690. Noudoost, B., Chang, M.H., Steinmetz, N.A., Moore, T., 2010. Top-down control of visual attention. Curr. Opin. Neurobiol. 20, 183e190. Noudoost, B., Clark, K.L., Moore, T., 2014. A distinct contribution of the frontal eye field to the visual representation of saccadic targets. J. Neurosci. 34, 3687e3698. Palmer, A.R., King, A.J., 1982. The representation of auditory space in the mammalian superior colliculus. Nature 299 (5880), 248e249. Pare, M., Crommelinck, M., Guitton, D., 1994. Gaze shifts evoked by stimulation of the superior colliculus in the head-free cat conform to the motor map but also depend on stimulus strength and fixation activity. Exp. Brain Res. 101, 123e139. Peck, C.K., Baro, J.A., Warder, S.M., 1995. Effects of eye position on saccadic eye movements and on the neuronal responses to auditory and visual stimuli in cat superior colliculus. Exp. Brain Res. 103, 227e242. Putzar, L., Goerendt, I., Lange, K., Rosler, F., Roder, B., 2007. Early visual deprivation impairs multisensory interactions in humans. Nat. Neurosci. 10, 1243e1245. Rakic, P., 1977. Prenatal development of the visual system in rhesus monkey. Philos. Trans. R. Soc. Lond. B Biol. Sci. 278, 245e260. Redgrave, P., Odekunle, A., Dean, P., 1985. Tectal cells of origin of predorsal bundle in rat: location, and segregation from ipsilateral descending pathway. Neurosci. Lett. Suppl. 22, S320. Redgrave, P., Mitchell, I.J., Dean, P., 1986. Stereotaxic mapping of efferent descending pathways of the superior colliculus in rat: an anterograde WGAHRP study. Neurosci. Lett. Suppl. 26, S349. Redgrave, P., Odekunle, A., Dean, P., 1986. Tectal cells of origin of predorsal bundle in rat: location and segregation from ipsilateral descending pathway. Exp. Brain Res. 63, 279e293. Rhoades, R.W., Mooney, R.D., Jacquin, M.F., 1983. Complex somatosensory receptive fields of cells in the deep laminae of the hamster’s superior colliculus. J. Neurosci. 3, 1342e1354. Rhoades, R.W., Mooney, R.D., Klein, B.G., Jacuin, M.F., Szczepanik, A.M., Chiaia, N.L., 1987. The structural and functional characteristics of tectospinal neurons in the golden hamster. J. Comp. Neurol. 255, 451e465. Robinson, D.A., 1972. Eye movements evoked by collicular stimulation in the alert monkey. Vis. Res. 12 (11), 1795e1808. Rosenblatt, J.S., 1971. Suckling and home orientation in the kitten: a comparative developmental study. In: Tobach, E., Aronson, L., Shaw, E. (Eds.), The Biopsychology of Development. Academic Press, New York, pp. pp345e410. Rowland, B.A., Jiang, W., Stein, B.E., 2014. Brief cortical deactivation early in life has long-lasting effects on multisensory behavior. J. Neurosci. 34 (21), 7198e7202. https://doi.org/10.1523/JNEUROSCI.3782-13.2014.

Development of the superior colliculus/optic tectum Chapter | 3

77

Ruthazer, E.S., Cline, H.T., 2004. Insights into activity-dependent map formation from the retinotectal system: a middle-of-the-brain perspective. J. Neurobiol. 59, 134e146. Ruthazer, E.S., Akeman, C.J., Cline, H.T., 2003. Control of axon branch dynamics by correlated activity in vivo. Science 301, 66e70. Schafer, R.J., Moore, T., 2007. Attention governs action in the primate frontal eye field. Neuron 56, 541e551. Schiller, P.H., Stryker, M., 1972. Single-unit recording and stimulation in superior colliculus of the alert rhesus monkey. J. Neurophysiol. 35 (6), 915e924. Siminoff, R., Schwassmann, H.O., Kruger, L., 1966. An electrophysiological study of the visual projection to the superior colliculus of the rat. J. Comp. Neurol. 127, 435e444. Sommer, M.A., Wurtz, R.H., 1998. Frontal eye field neurons orthodromically activated from the superior colliculus. J. Neurophysiol. 80, 3331e3335. Sommer, M.A., Wurtz, R.H., 2004. What the brain stem tells the frontal cortex. I. Oculomotor signals sent from superior colliculus to frontal eye field via mediodorsal thalamus. J. Neurophysiol. 91, 1381e1402. Sommer, M.A., Wurtz, R.H., 2004. What the brain stem tells the frontal cortex. II. Role of the SC-MD-FEF pathway in corollary discharge. J. Neurophysiol. 91, 1403e1423. Song, J.H., Rafal, R.D., McPeek, R.M., 2011. Deficits in reach target selection during inactivation of the midbrain superior colliculus. Proc. Natl. Acad. Sci. U.S.A. 108, E1433eE1440. Sparks, D.L., 1986. Translation of sensory signals into commands for control of saccadic eye movements: role of primate superior colliculus. Physiol. Rev. 66, 118e171. Sparks, D.L., Mays, L.E., 1990. Signal transformations required for the generation of saccadic eye movements. Annu. Rev. Neurosci. 13, 309e336. Sparks, D.L., Porter, J.D., 1983. Spatial localization of saccade targets. II. Activity of superior colliculus neurons preceding compensatory saccades. J. Neurophysiol. 49, 64e74. Spence, C., Pavani, F., Driver, J., 2004. Spatial constraints on visual-tactile cross-modal distractor congruency effects. Cognit. Affect Behav. Neurosci. 4, 148e169. Sperry, R.W., 1963. Chemoaffinity in the orderly growth of nerve fiber patterns and connections. Proc. Natl. Acad. Sci. U.S.A. 50, 703e710. Spitzer, H., Desimone, R., Moran, J., 1988. Increased attention enhances both behavioral and neuronal performance. Science 240, 338e340. Sridharan, D., Steinmetz, N.A., Moore, T., Knudsen, E.I., 2017. Does the superior colliculus control perceptual sensitivity or choice bias during attention? Evidence from a multialternative decision framework. J. Neurosci. 37, 480e511. Stein, B.E., 1984. Development of the superior colliculus. Annu. Rev. Neurosci. 7, 95e125. Stein, B.E., 2005. The development of a dialogue between cortex and midbrain to integrate multisensory information. Exp. Brain Res. 166, 305e315. Stein, B.E., Burr, D., Constantinidis, C., Laurienti, P.J., Alex, M.M., Perrault Jr., T.J., Ramachandran, R., Roder, B., Rowland, B.A., Sathian, K., Schroeder, C.E., Shams, L., Stanford, T.R., Wallace, M.T., Yu, L., Lewkowicz, D.J., 2010. Semantic confusion regarding the development of multisensory integration: a practical solution. Eur. J. Neurosci. 31, 1713e1720. Stein, B.E., Clamann, H.P., 1981. Control of pinna movements and sensorimotor register in cat superior colliculus. Brain Behav. Evol. 19, 180e192. Stein, B.E., Clamann, H.P., Goldberg, S.J., 1980. Superior colliculus: control of eye movements in neonatal kittens. Science 210, 78e80. Stein, B.E., Dixon, J.P., 1978. Superior colliculus cells respond to noxious stimuli. Brain Res. 158, 65e73. Stein, B.E., Gaither, N.S., 1981. Sensory representation in reptilian optic tectum: some comparisons with mammals. J. Comp. Neurol. 202, 69e87. Stein, B.E., Gallagher, H.L., 1981. Maturation of cortical control over superior colliculus cells in cat. Brain Res. 223, 429e435. Stein, B.E., Rowland, B.A., 2011. Organization and plasticity in multisensory integration: early and late experience affects its governing principles. Prog. Brain Res. 191, 145e163. Stein, B.E., Labos, E., Kruger, L., 1973. Sequence of changes in properties of neurons of superior colliculus of the kitten during maturation. J. Neurophysiol. 36, 667e679. Stein, B.E., Labos, E., Kruger, L., 1973. Determinants of response latency of neurons in superior colliculus in kittens. J. Neurophysiol. 36, 680e689. Stein, B.E., Magalhaes-Castro, B., Kruger, L., 1975. Superior colliculus: visuotopic-somatotopic overlap. Science 189, 224e226. Stein, B.E., McHaffie, J.G., Harting, J.K., Huerta, M.F., Hashikawa, T., 1985. Transient tectogeniculate projections in neonatal kittens: an autoradiographic study. J. Comp. Neurol. 239, 402e412. Stein, B.E., Meredith, M.A., 1993. The Merging of the Senses. MIT Press, Cambridge, Mass. Stein, B.E., Spencer, R.F., Edwards, S.B., 1982. Efferent projections of the neonatal superior colliculus: extraoculomotor-related brain stem structures. Brain Res. 239, 17e28. Stein, B.E., Spencer, R.F., Edwards, S.B., 1983. Corticotectal and corticothalamic efferent projections of SIV somatosensory cortex in cat. J. Neurophysiol. 50, 896e909. Stein, B.E., Stanford, T.R., 2008. Multisensory integration: current issues from the perspective of the single neuron. Nat. Rev. Neurosci. 9, 255e266. Stuphorn, V., Bauswein, E., Hoffmann, K.P., 2000. Neurons in the primate superior colliculus coding for arm movements in gaze-related coordinates. J. Neurophysiol. 83, 1283e1299. Thorn, F., Gollender, M., Erickson, P., 1976. The development of the kittens visual optics. Vis. Res. 16, 1145e1149. Treue, S., Maunsell, J.H., 1996. Attentional modulation of visual motion processing in cortical areas MT and MST. Nature 382, 539e541. Van Wanrooij, M.M., Bell, A.H., Munoz, D.P., Van Opstal, A.J., 2009. The effect of spatial-temporal audiovisual disparities on saccades in a complex scene. Exp. Brain Res. 198, 425e437. Villablanca, J.R., Olmstead, C.E., 1979. Neurological development of kittens. Dev. Psychobiol. 12, 101e127.

78

PART | I Circuit development

Wallace, M.T., McHaffie, J.G., Stein, B.E., 1997. Visual response properties and visuotopic representation in the newborn monkey superior colliculus. J. Neurophysiol. 78, 2732e2741. Wallace, M.T., Perrault Jr., T.J., Hairston, W.D., Stein, B.E., 2004. Visual experience is necessary for the development of multisensory integration1. J. Neurosci. 24, 9580e9584. Wallace, M.T., Stein, B.E., 2007. Early experience determines how the senses will interact. J. Neurophysiol. 97, 921e926. Wallace, M.T., Stein, B.E., 2000. Onset of cross-modal synthesis in the neonatal superior colliculus is gated by the development of cortical influences. J. Neurophysiol. 83, 3578e3582. Wallace, M.T., Stein, B.E., 1994. Cross-modal synthesis in the midbrain depends on input from cortex. J. Neurophysiol. 71, 429e432. Wallace, M.T., Stein, B.E., 1997. Development of multisensory neurons and multisensory integration in cat superior colliculus. J. Neurosci. 17, 2429e2444. Wallace, M.T., Stein, B.E., 2001. Sensory and multisensory responses in the newborn monkey superior colliculus. J. Neurosci. 21, 8886e8894. Wallace, M.T., Wilkinson, L.K., Stein, B.E., 1996. Representation and integration of multiple sensory inputs in primate superior colliculus. J. Neurophysiol. 76, 1246e1266. Walter, J., Henke-Fahle, S., Bonhoeffer, F., 1987. Avoidance of posterior tectal membranes by temporal retinal axons. Development 101, 909e913. Weber, J.T., Martin, G.F., Behan, M., Huerta, M.F., Harting, J.K., 1979. The precise origin of the tectospinal pathway in three common laboratory animals: a study using the horseradish peroxidase method. Neurosci. Lett. 11, 121e127. White, B.J., Kan, J.Y., Levy, R., Itti, L., Munoz, D.P., 2017. Superior colliculus encodes visual saliency before the primary visual cortex. Proc. Natl. Acad. Sci. U.S.A. 114, 9451e9456. White, B.J., Berg, D.J., Kan, J.Y., Marino, R.A., Itti, L., Munoz, D.P., 2017. Superior colliculus neurons encode a visual saliency map during free viewing of natural dynamic video. Nat. Commun. 8, 14263. Wilkinson, L.K., Meredith, M.A., Stein, B.E., 1996. The role of anterior ectosylvian cortex in cross-modality orientation and approach behavior. Exp. Brain Res. 112, 1e10. Williams, R.W., Chalupa, L.M., 1982. Prenatal development of retinocollicular projections in the cat: an anterograde tracer transport study. J. Neurosci. 2, 604e606. Williford, T., Maunsell, J.H., 2006. Effects of spatial attention on contrast response functions in macaque area V4. J. Neurophysiol. 96, 40e54.22. Wurtz, R.H., Goldberg, M.E., 1972. Activity of superior colliculus in behaving monkey. 3. Cells discharging before eye movements. J. Neurophysiol. 35 (4), 575e586. Xu, J., Yu, L., Rowland, B.A., Stanford, T.R., Stein, B.E., 2012. Incorporating cross-modal statistics in the development and maintenance of multisensory integration. J. Neurosci. 32 (7), 2287e2298. https://doi.org/10.1523/JNEUROSCI.4304-11.2012. Xu, J., Yu, L., Rowland, B.A., Stein, B.E., 2017. The normal environment delays the development of multisensory integration. Sci. Rep. 7 (1), 4772. https://doi.org/10.1038/s41598-017-05118-1. Xu, J., Yu, L., Stanford, T.R., Rowland, B.A., Stein, B.E., 2015. What does a neuron learn from multisensory experience? J. Neurophysiol. 113 (3), 883e889. https://doi.org/10.1152/jn.00284.2014. Yu, L., Rowland, B.A., Stein, B.E., 2010. Initiating the development of multisensory integration by manipulating sensory experience. J. Neurosci. 30 (14), 4904e4913. https://doi.org/10.1523/JNEUROSCI.5575-09.2010. Yu, L., Xu, J., Rowland, B.A., Stein, B.E., 2013. Development of cortical influences on superior colliculus multisensory neurons: effects of dark-rearing. Eur. J. Neurosci. 37, 1594e1601. https://doi.org/10.1111/ejn.12182. Zahar, Y., Reches, A., Gutfreund, Y., 2009. Multisensory enhancement in the optic tectum of the barn owl: spike count and spike timing. J. Neurophysiol. 101, 2380e2394. Zenon, A., Krauzlis, R.J., 2012. Attention deficits without cortical neuronal deficits. Nature 489, 434e437.

Chapter 4

Cerebellar circuits Masanobu Kano1 and Masahiko Watanabe2 1

Department of Neurophysiology, Graduate School of Medicine, The University of Tokyo, Tokyo, Japan; 2Department of Anatomy, Hokkaido

University Graduate School of Medicine, Sapporo, Japan

Chapter outline 4.1. Overview of the microcircuit in the cerebellar cortex 4.1.1. Cell types and afferent fibers 4.1.2. Generation of neurons that constitute microcircuit on PCs 4.1.3. Compartmentalization of the cerebellum 4.2. Development of CFePC synapses 4.2.1. Multiple innervation of PCs by CFs in early postnatal period 4.2.2. Functional differentiation of multiple CFs 4.2.3. Dendritic translocation of single CFs 4.2.4. Early phase of CF synapse elimination 4.2.5. Late phase of CF synapse elimination 4.3. Development of PFePC synapses 4.3.1. Formation of PFePC synapses

79 79 81 81 82 82 82 83 85 86 89 89

4.3.2. Stabilization and maintenance of PFePC synapses 4.3.3. Developmental elimination of PFePC synapses 4.3.4. Heterosynaptic competition between PF and CF inputs 4.4. Development of inhibitory synapses from basket cells and stellate cells to PCS 4.4.1. Formation of basket Cell-PC synapses 4.4.2. Formation of stellate Cell-PC synapses 4.4.3. Activity-dependent remodeling of inhibitory synapses 4.5. Summary and conclusions Acknowledgments References

89 91 91 93 93 94 95 95 96 96

4.1 Overview of the microcircuit in the cerebellar cortex 4.1.1 Cell types and afferent fibers The cerebellum consists of the cortex and the centrally located deep cerebellar nuclei (DCN). The cerebellar cortex exhibits a characteristic trilaminar structure composed of the molecular layer, Purkinje cell (PC) layer, and granular layer. In its mediolateral extent, the cerebellar cortex is divided into three longitudinal regions: vermis (medial cerebellum), paravermis (intermediate cerebellum or pars intermedia), and hemisphere (lateral cerebellum). Each of these regions is folded into lobules. The DCN also have three divisions: the medial (fastigial), interpositus (globose and emboliform), and lateral (dentate) nuclei, each of which is connected topographically with the vermis, paravermis, and hemisphere, respectively. Cerebellar neurons with distinct cytological and neurochemical properties reside in specific layers and sites of the cerebellum. They are connected with each other and also with specific brain regions outside the cerebellum (Fig. 4.1). PCs are the sole output neurons of the cerebellar cortex. The somata of PCs are aligned in the PC layer. PCs extend well-arborized dendrites in the molecular layer and project g-aminobutyric acid (GABA)ergic axons to the DCN and vestibular nuclei. There are two distinct excitatory afferents to the cerebellum, that is, climbing fibers (CFs) and mossy fibers (MFs) (Palay and Chan-Palay, 1974). In adulthood, each PC is innervated by a single CF originating from the inferior olive of the contralateral medulla oblongata (Eccles et al., 1966). Each CF forms hundreds of synapses by twisting around the proximal dendritic compartment. Therefore, activation of CFs causes strong depolarization of PC dendrites, triggers regenerative “Ca2þ spikes” due to activation of voltage-dependent Ca2þ channels (VDCCs) in PC dendrites (Miyakawa et al., 1992), and generates characteristic “complex spikes” in the PC soma (Eccles et al., 1966). Each complex spike is composed of a fast somatic Naþ action potential followed by slow dendritic Ca2þ spikes. In contrast, MFs

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00004-4 Copyright © 2020 Elsevier Inc. All rights reserved.

79

80

PART | I Circuit development

Parallel f.

Stellate ML Basket

Purkinje

PCL

Granule

Pinceau

Lugaro Climbing f.

GL

Golgi UB

Cerebellar cortex Mossy f. To RN, VL /VA

DCN ION From, SC, PN/RF

FIGURE 4.1 Neuronal component and synaptic wiring diagram of the cerebellum. Neurons painted with warm colors (yellow, orange, and pink) represent excitatory neurons, while those with cold colors (blue, light blue, and green) do inhibitory neurons. DCN, deep cerebellar nuclei; GL, granular layer; ION, inferior olivary nucleus; ML, molecular layer; PCL, Purkinje cell layer; UB, unipolar brush cell.

originating from various extracerebellar regions, such as the spinal cord, pontine nuclei, and reticular formation, convey motor and sensory information to the distal dendritic compartment of PC through parallel fibers (PFs), the bifurcated axons of granule cells (GCs) (Ito, 1984). Approximately, 105e106 PFs innervate a given PC, while each PF forms usually one, or occasionally two, synapses onto individual PCs (Napper and Harvey, 1988). Thus, excitation of a single PF depolarizes PC dendrites only weakly, and about 50 GCs should fire synchronously to generate a single Naþ action potential (called “simple spike”) in the PC soma (Barbour, 1993). PFs also excite two types of GABAergic interneurons in the molecular layer: basket cells and stellate cells. Although these two types of interneurons share similar features (Sultan and Bower, 1998), they differ in short-term plasticity (Bao et al., 2010), gene expression (Schilling and Oberdick, 2009), and most importantly, the target of projection. Basket cells innervate the PC soma and their axons surround the axon initial segment (AIS) of PC, while stellate cells innervate PC dendrites (Palay and Chan-Palay, 1974). The specialized conical structure formed by basket cell axons around the AIS is called the pinceau (Ango et al., 2004). Molecules for chemical GABAergic inhibition are highly concentrated at basket cell synapses on PC somata, whereas they are mostly lacking in the pinceau (Iwakura et al., 2012). Instead, the pinceau is enriched with Shaker-type Kþ channels and their scaffolding protein PSD-95 (Iwakura et al., 2012; Laube et al., 1996), and exerts ultrarapid electrical inhibition of PC firing (Blot and Barbour, 2014). In the granular layer, there are several types of interneurons, which do not project to PCs directly. Golgi cells, Lugaro cells, and globular cells are inhibitory interneurons with dual gycinergic/GABAergic phenotype (Ottersen et al., 1988). Golgi cells are distributed throughout the granular layer, have large polygonal somata, and extend dendritic trees radiating into the molecular layer. They receive excitatory inputs from MFs directly in the granular layer and indirectly in the molecular layer via GCs and PFs. Although CFs emit thin branches named Sheibel collaterals that approach Golgi cell somata and dendrites, synaptic contacts between them have not been demonstrated so far (Galliano et al., 2013). Golgi cells, in turn, innervate GC dendrites in the cerebellar glomeruli, thus providing feedforward and feedback inhibition to GCs. Golgi cell also receive inhibitory inputs from basket, stellate, and Lugaro cells. Lugaro cells are fusiform interneurons lying just beneath the PC layer. They receive GABAergic inputs from recurrent axon collaterals of PCs, and project axons to the molecular layer to innervate basket and stellate cells, thus providing feedback inhibition to PCs via molecular layer

Cerebellar circuits Chapter | 4

81

interneurons (Laine and Axelrad, 2002). Globular cells have globular somata located at variable depths in the granular layer, and are thought to be a subtype of Lugaro cells (Laine and Axelrad, 2002). Lugaro and globular cells are neurochemically distinguished from Golgi cells; the former express calretinin, while most of the latter express mGluR2 and neurogranin (Simat et al., 2007). Unipolar brush cells are excitatory interneurons in the granular layer, which are clearly distinct from GCs and populated in the vestibulocerebellum. They are characterized by single short dendrites terminating with a brush of dendrites, which engulf one or two rosettes of glutamatergic and cholinergic MFs. They innervate dendrites of other unipolar brush cells and GCs thus regarded as an intermediate component that amplifies excitatory drives of MFs on to GCs (Mugnaini and Floris, 1994). DCN neurons receive excitatory inputs from collaterals of MFs and CFs, and inhibitory inputs from PC axons. Thus, direct inputs from MFs and CFs and indirect inputs that are highly processed, integrated, and modified in the cerebellar cortex converge at DCN neurons. In the DCN, GABAergic neurons are either local interneurons or projection neurons targeting their axons to the inferior olive (De Zeeuw et al., 1988), while glutamatergic neurons project to other brain regions including the red nucleus and thalamus. Thus, the DCN is the key node of the cerebellum as the source of its output.

4.1.2 Generation of neurons that constitute microcircuit on PCs The bHLH transcription factors Ptf1a and Atoh1 are required for producing GABAergic and glutamatergic neurons, respectively, in the cerebellum. All GABAergic neurons in the cerebellum originate from the Ptf1a domain in the ventricular zone (Hoshino et al., 2005). Of these, projection neurons, that is, PCs and DCN neurons projecting to the inferior olive, are specified within the Ptf1a domain at the onset of cerebellar neurogenesis in early prenatal life (Altman and Bayer, 1997; Miale and Sidman, 1961). GABAergic interneurons also derive from the Ptf1a domain, but the progenitors continue to proliferate in the prospective white matter up to postnatal development (Zhang and Goldman, 1996). Phenotypic specification and diversity of GABAergic interneurons appear to be created by instructive cues provided by the microenvironment of the prospective white matter (Leto et al., 2009). Two subpopulations in the Ptf1a domain that express Olig2 and Gsx1 correspond to PC progenitors and Pax2-positive interneuron progenitors, respectively (Seto et al., 2014). On the other hand, the rhombic lip that expresses Atoh1 generates all progenitors of glutamatergic cerebellar neurons, which then migrate via different pathways (Yamada et al., 2014). Glutamatergic neurons in the DCN migrate to the nuclear transitory zone before descending to the prospective DCN (Fink et al., 2006). GC precursors first migrate to the cerebellar surface and form the external granular layer. There, they continue to proliferate during the postnatal period, and then descend to the internal granular layer (Rakic, 1971). Unipolar brush cells migrate to their destination through developing white matter (Englund et al., 2006).

4.1.3 Compartmentalization of the cerebellum Although the basic cellular composition and wiring diagram are uniform across the cerebellum, longitudinal organization of the cerebellar cortex has been demonstrated using anatomical, physiological, and molecular mapping techniques (Apps and Hawkes, 2009). Longitudinal cerebellar zones have been defined anatomically by cholinesterase labeling in the white matter and topographic projection of CFs and PC axons (Voogd and Ruigrok, 2004). Longitudinal zones are further divided into smaller units called microzones, based on high synchrony of complex spike activity (Llinas and Sasaki, 1989; Sugihara et al., 1993) and Ca2þ spikes (Mukamel et al., 2009; Schultz et al., 2009). Each microzone is w500 mm in width, stable across behavioral states, and has a sharp boundary with the neighboring microzones (Mukamel et al., 2009). This synchrony is based on electrical coupling of nearby olivary neurons through dendrodendritic gap junctions (Llinas et al., 1974; Sotelo et al., 1974) and topographical olivocerebellar projections, that is, from given subregions of the inferior olive to specific longitudinal cortical zones (Sugihara et al., 2001). Elaborate and fine compartmentalization can also be recognized as cerebellar stripes by histochemistry for various molecules expressed in PCs. The best example of “late” stripe markers, which reveal the adult topography (postnatal day 15 onwards), is aldolase C or zebrin II antigen (Hawkes and Leclerc, 1987), while “early” stripe markers of topography, such as calbindin and L7, typically reveal zones and stripes during perinatal development (embryonic day 13 (E13) to postnatal day 5(P5)) (Wassef et al., 1985). These stripes are reproducible between individuals and conserved across species. Using a novel molecular marker phospholipase Cb4 (PLCb4), which is continuously expressed in zebrin IInegative PCs from embryonic stage to adulthood (Nakamura et al., 2004; Watanabe et al., 1998), some stripes in the adult cerebellum have been shown to derive from two or more distinct embryonic clusters (Marzban et al., 2007). Importantly, studies with small tracer injection have shown that the topography of zebrin II expression pattern corresponds

82

PART | I Circuit development

to that of the olivary projection to the cerebellar cortex and further to that of the olivary projection to the DCN and the cerebellar cortical projection to the DCN (Pijpers et al., 2005; Sugihara and Shinoda, 2004, 2007). Furthermore, an in vivo two-photon Ca2þ imaging study has shown a precise structural and functional relationship between alsolase C expression and complex spike activity in PCs (Tsutsumi et al., 2015). These lines of evidence support the idea that the entire cerebellar system is formed by parallel assembly of an olivo-cortico-nuclear microcomplex (Ito, 1984).

4.2 Development of CFePC synapses 4.2.1 Multiple innervation of PCs by CFs in early postnatal period PCs undergo drastic changes in their morphology from late embryonic to early postnatal days (Armengol and Sotelo, 1991). PCs are born at the cerebellar ventricular zone at E10-E13, and migrate to form a multilayer below the molecular layer. At the end of radial migration, PCs exhibit bipolar shapes at E19-P0 (“simple and fusiform cell” by Armengol and Sotelo, 1991). Then, at P1eP3, new stem dendrites emerge from all aspects of the cell bodies, which confer the complex shapes of PCs (“complex-fusiform cell” by Armengol and Sotelo, 1991). From P3 to P6, such stem dendrites disappear by retraction of the long dendritic branches (“regressive-atrophic dendrites” by Armengol and Sotelo, 1991). In the meantime, PCs line up in a monolayer, which is completed by P5. PCs undergo explosive outgrowth of perisomatic protrusions that emerge in all directions from the cell bodies (the stage of “stellate cells” by Armengol and Sotelo, 1991). Up to P10, PCs extend single or double stem dendrites into the molecular layer, which grow and split into many dendritic branches, with simultaneous withdrawal of long somatic processes. The polarity of PC is determined during this developmental period. From P10 to P15, the growth of the dendritic arbor occurs mainly in its lateral domain, whereas from P15 on, the dendritic growth occurs in the vertical plane and the height of the dendritic field reaches the adult level at P30. After reaching the primitive cerebellum around E18, axons of inferior olivary neurons give rise to thick and thin collaterals (Wassef et al., 1992). This stage is called the “creeper stage” (Chedotal and Sotelo, 1993) (Chedotal and Sotelo, 1993) at which the “CFs,” the thick collaterals of olivary axons, creep between immature PCs. At this stage, PCs have just completed their migration and are organized in a multilayer of “simple and complex-fusiform cells” (Armengol and Sotelo, 1991). Initially, each olivocerebellar axon forms about 100 “creeper” CFs (Sugihara, 2005). Then, the CF to PC synapse undergoes three distinct phases of postnatal development, which are described by Ramón y Cajal in his pioneering studies (Cajal, 1911): the “pericellular nest” stage, the “capuchon” stage, and the “dendritic” stage. At the “pericellular nest” stage, CFs surround the cell bodies of PCs. At this stage, PCs have many perisomatic protrusions that emerge in all directions from the cell bodies, and therefore, this stage is called the phase of “stellate cells” (Armengol and Sotelo, 1991). CFs establish contacts with the abundant pseudopodia stemming from the cell bodies and form a plexus on the lower part of the PC soma. Among the 100 “creeper” CFs of each olivocerebellar axon, only around 10 can develop to form “pericellular nests.” Then, the “capuchon” stage is characterized by the displacement of the plexus of CF collaterals to the apical portion of PC somata and main dendrites. Finally, at the “dendritic” stage, CFs undergo translocation to growing PC dendrites and expanding their innervation territories. Earlier electrophysiological studies on juvenile rats in vivo showed that stimulation to the inferior olive after P3 elicits CF-mediated responses in PCs (Crepel, 1971). However, in contrast to the all-or-none nature of CF responses in the adult animals, the responses of juvenile PCs are graded in parallel with increase in stimulus strength (Crepel et al., 1976). This is the first evidence that PCs are innervated by multiple CFs in early postnatal development. Later extensive studies in vivo revealed that both the percentage of PCs innervated by multiple CFs and the average number of CFs innervating individual PCs decrease with postnatal development and that most PCs become singly innervated by CFs (Crepel et al., 1981; Mariani and Changeux, 1981). These results clearly indicate that elimination of redundant CF inputs occurs during postnatal development.

4.2.2 Functional differentiation of multiple CFs Multiple CFs initially form functional synapses on the PC soma at around P3 (Chedotal and Sotelo, 1993; Morando et al., 2001). When recorded from PCs in cerebellar slices at this developmental stage, excitatory postsynaptic currents (EPSCs) elicited by stimulating multiply-innervating CFs are much smaller than those of mature CFs (Bosman et al., 2008; Hashimoto and Kano, 2003). Therefore, CF inputs become stronger, while redundant CFs are eliminated during postnatal development. Changes in the relative synaptic strengths of multiple CFs innervating the same PC have been systematically studied during postnatal development (Hashimoto and Kano, 2003) by recording CF-mediated EPSCs in PCs from cerebellar slices

Cerebellar circuits Chapter | 4

83

of mice aged P2eP21. This systematic study shows that more than five discrete CF-EPSCs with similar amplitudes are recorded in PCs from neonatal mice around P3 (Fig. 4.2A, wP3). In contrast, in the second postnatal week, PCs with multiple CF-EPSCs have one large CF-EPSC and a few small CF-EPSCs (Fig. 4.2A, wP7 and wP12). These results indicate that synaptic strengths of multiply-innervating CFs are relatively uniform in neonatal mice, and one CF is selectively strengthened during postnatal development (Bosman et al., 2008; Hashimoto and Kano, 2003, 2005). Quantitative assessments of the disparity among the amplitudes of multiple CF-EPSCs in individual PCs demonstrate that one CF is selectively strengthened among multiple CFs innervating the same PC from P3 to P7 (Fig. 4.2B) (Hashimoto and Kano, 2003). These electrophysiological data are consistent with the morphological observation that the innervation pattern of CFs over PCs drastically changes during this postnatal period in rats (Sugihara, 2005). At P4, CFs have many creeping terminals in the PC layer and their swellings do not aggregate at particular PC somata (creeper type). Then, from P4 to P7, CFs surround several specific PC somata and form aggregated terminals on them (nest type) (Sugihara, 2005). Recently, Good et al. recorded PC population activity from juvenile mice using in vivo two-photon Ca2þ imaging, and have demonstrated that CF responses from a population of PCs are highly synchronized at w P4 and are massively desynchronized thereafter to w P8 (Good et al., 2017). This desynchronization of CF population responses is thought to reflect the CF network refinement from “creeper type” to “nest type” terminals (Good et al., 2017). There are clear differences in electrophysiological properties between EPSCs elicited by the strongest CF input and those by other weaker inputs. Transient rises of glutamate concentration in the synaptic cleft are significantly higher after stimulation of the strongest CF than the weaker CFs (Hashimoto and Kano, 2003). This is thought to result from the fact that the probability of multivesicular release (i.e., more than one synaptic vesicle released simultaneously to a given postsynaptic site from the corresponding presynaptic release site) is higher for the strongest CF than for the weaker CFs. Further electrophysiological examination suggests that the number of release sites facing a narrow postsynaptic PC region is larger in the strongest CF than in weaker CFs (Hashimoto and Kano, 2003). Hashimoto et al. demonstrated that the P/Q-type VDCC, a high voltage-activated Ca2þ channels constituting >90% of the total Ca2þ current density in PCs (Mintz et al., 1992; Stea et al., 1994) and abundantly distributed in PC dendrites and spines (Kulik et al., 2004; Miyazaki et al., 2012), is crucial for the selective strengthening of a single CF (Hashimoto et al., 2011). In PC-specific P/Q-type VDCC knockout mice, multiple CFs are nonselectively strengthened during the first postnatal week (Hashimoto et al., 2011). Kawamura et al. performed in vivo whole-cell recordings from PCs and have shown that a single CF input closest in time to PC’s spike output is selectively strengthened during the first postnatal week, and that this selective strengthening is impaired in PC-specific P/Q-type VDCC knockout mice (Kawamura et al., 2013). These results strongly suggest that biasing the competition toward a single CF input is mediated by P/Q-type VDCCs in PCs. Selection of the single “winner” CF and subsequent elimination of “loser” CFs are thought to be influenced by factors that strengthen and/or maintain CF synapses. For example, when such factors are abundant, weaker CF synapses may not easily be eliminated. Conversely, when their levels decrease, weaker CFs may be difficult to survive and their elimination would be facilitated. In this context, insulin like growth factor I (IGF-1) strengthens both the strongest and weaker CF synapses and thereby regulates the degree of CF synapse elimination from P8 to P12 (Kakizawa et al., 2003). Sherrard et al. (2009) reported that decrease in brain-derived neurotrophic factor (BDNF) to Trk B signaling might switch CFePC synaptogenesis to elimination of surplus CF synapses. This conclusion is based on the prerequisite that BDNF to Trk B signaling strengthens/maintains CF synapses (but see below about the role of BDNF in the late phase of CF synapse elimination). Moreover, C1ql1, a member of C1q family proteins, derives from CFs and anterogradely strengthens/maintains a single winner CF through acting on brain-specific angiogenesis inhibitor 3 (Bai3) in PCs (Kakegawa et al., 2015; Sigoillot et al., 2015). On the other hand, Uesaka et al. performed a systematic screening of candidates of retrograde signaling molecules from postsynaptic PCs to presynaptic CFs that are involved in CF synapse development or elimination. They show that Semaphorin 3A (Sema3A) secreted from PCs strengthens CF synapses by acting on Plexin A4 (PlxnA4) on CF terminals and thereby opposing CF synapse elimination (Uesaka et al., 2014). In addition, they have recently demonstrated that progranulin, a growth factor implicated in the pathogenesis of certain forms of frontotemporal dementia, functions as a retrograde signaling molecule from PCs and strengthens CF synapses through acting on Sort1 on CF terminals (Uesaka et al., 2018). Progranulin is shown to counteract CF synapse elimination independently of Sema3A. Thus, Sema3A and progranulin are thought to fine-tune the selection of a single “winner” CF and elimination of “loser” CFs.

4.2.3 Dendritic translocation of single CFs Morphological evidence indicates that the site of CF innervation of PC changes from soma to dendrite during early postnatal development, a phenomenon known as “CF translocation” (Altman and Bayer, 1997). The relationship between

84

PART | I Circuit development

Adult

(A)

Parallel fiber

~P12 ~P7 ~P3

Granule cell

Purkinje cell

Mossy fiber

(B) Multiple innervation

Functional differentiation (P3–P7)

Climbing fiber

CF translocation (P9-) Early phase of CF elimination (P7–P11)

Late phase of CF elimination (P12–P17)

Mono innervation

FIGURE 4.2 Postnatal development of CFePC synapses. (A) Diagrams of CFePC synapses at four representative stages of postnatal development in mice. (B) Four distinct phases in postnatal development of CF synapses. Reproduced with permission from European Journal of Neuroscience 34, 1697e1710, with permission.

the selective strengthening of single CFs and CF translocation was investigated by using both electrophysiological and morphological techniques (Hashimoto et al., 2009a). The location of synapses along the somatodendritic domains of PC can be estimated by analyzing the kinetics of quantal EPSCs (qEPSCs) arising from single synaptic vesicles in CF terminals. At P7eP8 when the selective strengthening of a single CF on each PC has just been completed, there is no significant difference in the rise times (i.e., time from the onset to the peak) of qEPSCs for the strongest compared with the weaker CFs. Since the rise time of qEPSC is proportional to the distance from the synaptic site to the somatic recording site (Roth and Hausser, 2001), synapses of the strongest and weaker CFs are thought to be located on the soma at around P7 (Fig. 4.2A, wP7). At P9eP10, the incidence of qEPSC with slow rise time is more frequent for the strongest than for the weaker CFs, suggesting the initiation of CF translocation (Fig. 4.2B). The difference in the distribution of qEPSC rise times for the strongest compared with the weaker CFs becomes larger from P11 to P14. While the incidence of qEPSC with slow rise time becomes more frequent for the strongest CFs with age, the qEPSC rise times for the weaker CFs remain almost unchanged from P9 to P14. These electrophysiological data collectively indicate that (1) synaptic competition among multiple CFs occurs on the soma before P7 (Fig. 4.2A, wP3 and wP7, Fig. 4.2B), (2) only the strongest CF (“winner” CF) starts to translocate to dendrites at P9 and the translocation continues thereafter (Fig. 4.2A, wP12, Fig. 4.2B), and (3) synapses of the weaker CFs (“loser” CFs) remain around the soma (Fig. 4.2A, wP12). Morphological data are consistent with these electrophysiological observations (Hashimoto et al., 2009a). When subsets of CFs are labeled by an anterograde tracer, biotinylated dextran amine (BDA), injected into the inferior olive, pericellular nests with extensive branching of CFs are observed at P7, P9, and P12. At P7, in spite of the presence of immature stem dendrites in PCs, CFs innervate the soma and spare the dendrites (Fig. 4.3A). Dendritic innervation of CFs starts at P9 (Fig. 4.3B) and at P12 and thereafter, the territory of innervation extends progressively along the PC dendrites (Fig. 4.3C, D and E). Hashimoto et al. stained subsets of CFs by injecting a small amount of an anterograde tracer into the inferior olive and immunostained all CF terminals simultaneously with the CF terminal marker vesicular glutamate transporter 2 (VGluT2) (Hashimoto et al., 2011). In PC-specific P/Q-type VDCC knockout mice, dendrites of a PC often have VGluT2-positive CF terminals with the tracer (VGluT2/tracer double positive terminals) and those without the tracer (VGluT2 single positive terminals), indicating that the dendrites of such PCs are innervated by at least two CFs with different neuronal origins in the inferior olive (Hashimoto et al., 2011). This result strongly suggests that the P/Q-type VDCC is crucial for dendritic translocation of a single winner CF in each PC. Carrillo et al. performed two-photon multicolor imaging of CFs in developing mouse cerebellum in vivo (Carrillo et al., 2013). They demonstrate that the motility of CF terminals on the PC

Cerebellar circuits Chapter | 4

85

(E) (D)

(C) (B) (A)

P7

P9

P12

P15

P20

FIGURE 4.3 Developmental profile of CF innervations from perisomatic nest stage to peridendritic stage. (A)e(E) Fluorescent labeling of CFs with BDA (red) and PCs with calbindin antibody (green) at respective postnatal days. Reproduced from Hashimoto K, Ichikawa R, Kitamura K, Watanabe M, and Kano M (2009) Translocation of a “winner” climbing fiber to the Purkinje cell dendrite and subsequent elimination of “losers” from the soma in developing cerebellum. Neuron 63: 106e118, with permission.

soma is much higher than those on PC dendrites, and that the CF which has begun dendritic translocation indeed becomes the “winner” in most of the cases (Carrillo et al., 2013). In mice deficient in the C1ql1 to Bai3 anterograde signaling, not only a “winner” CF but also “loser” CFs undergo translocation to PC dendrite and the extent of dendritic translocation of the winner CF is reduced at P26 and thereafter (Kakegawa et al., 2015; Sigoillot et al., 2015). Similarly, in adult PC-specific progranulin knockout mice, the extent of dendritic translocation of the winner CF is persistently reduced (Uesaka et al., 2018). These results indicate that molecules involved in strengthening/maintenance of CF synapses during early postnatal development are also used for proper extension of dendritic innervation territories of the winner CF throughout life.

4.2.4 Early phase of CF synapse elimination Detailed assessments of the postnatal development of CF innervation in mouse cerebellar slices demonstrate that there is no significant reduction in the average number of CFs per PC from P3 to P6, when functional differentiation of multiple CFs occurs (Hashimoto et al., 2009b). The value then decreases progressively from P6 to around P15 (Hashimoto et al., 2009b). CF synapse elimination therefore does not proceed in parallel with functional differentiation of multiple CFs but starts after the strengthening of single CFs to individual PCs. Crepel et al. have demonstrated that elimination of surplus CFs consists of two distinct phases, the early phase up to around P8 and the late phase from around P9 to P17 in rats (Crepel et al., 1981). The early phase occurs normally in animals with mild X-irradiation to the cerebellum during the early postnatal period, which causes selective loss of GCs and PFs while leaving PCs intact. In marked contrast, the late phase is severely impaired by inhibiting GC production by X-irradiation. This study indicates that the early phase of CF synapse elimination is independent of PFePC synapse formation, whereas the late phase is critically dependent on it. However, since the animal models with “hypogranular” or “agranular” cerebella often have abnormalities of cerebellar development other than GC genesis and PFePC synapse formation, there remains a possibility that CF synapse elimination might be influenced by such developmental defects. Indeed, strong X-irradiation restricted to P5 and P6 causes massive GC loss, severe impairment of PC morphological development, and persistent multiple CF innervation in the adult (Bailly et al., 2018). The analysis of mutant mice deficient in the glutamate receptor d2 subunit (GluRd2 or GluD2) revealed that there are two distinct phases of CF synapse elimination. GluD2 is richly expressed in PCs and its deletion causes impairment of PFePC synapse formation leading to reduction of PFePC synapse number to about half of that in wild type mice. In spite of the severe impairment of PF synapse formation, GluD2 deletion does not significantly affect the laminar structure of the cerebellum and morphology of the PC and its dendritic tree (Kashiwabuchi et al., 1995; Kurihara et al., 1997). In GluD2 knockout mice, the average number of CFs innervating each PC is similar to that of control mice from P5 to P11. However, the value becomes significantly larger than that of control mice from P12 to P14. Thus, CF synapse elimination in mice can be classified into two distinct phases, namely, the “early phase” from P6 to around P11, which is independent of PFePC synapse formation, and the “late phase” from around P12 and thereafter, which requires normal PFePC synapse formation (Fig. 4.2B) (Hashimoto et al., 2009b).

86

PART | I Circuit development

Patterns of CF activity have been reported to influence the early phase of CF synapse elimination. Andjus et al. disrupted the normal activity pattern of CFs in rat at P9eP12 by administration of harmaline, which induced synchronous activation of inferior olivary neurons (Andjus et al., 2003). This treatment caused persistent multiple CF innervations of PCs in rats from P15 to P87. Furthermore, Lorenzetto et al. (2009) demonstrate that PC activity is crucial for CF synapse elimination. They generated transgenic mice that expressed a chloride channel-YFP fusion protein specifically in PCs to suppress their excitabilities (Lorenzetto et al., 2009). In this mouse line, the expression of chloride channel is observed in PCs during the “early phase” at P9, and multiple CF innervation persists up to P90. Therefore, perturbation of PC activity is considered to cause impairment of the “early phase” of CF synapse elimination. Furthermore, in PC-specific P/Q-type VDCC knockout mice, CF synapse elimination from P7 to around P12 is severely impaired, although initial CF to PC synapse formation appears normal at around P4 (Hashimoto et al., 2011). These results clearly indicate that PC activity, activation of P/Q-type VDCC, and Ca2þ elevation in PCs are crucial for the early phase of CF synapse elimination. Molecules that strengthen/maintain CF synapses including IGF-1 (Kakizawa et al., 2003), Sema3A-PlxnA4 (Uesaka et al., 2014), and progranulin-Sort1 (Uesaka et al., 2018) are thought to counteract the force of CF elimination and thereby regulate the extent of elimination of surplus CFs. Indeed, CF synapse elimination is accelerated by blockade or deletion of Sema3A-PlxnA4 and progranulin-Sort1 (Uesaka et al., 2014, 2018). In contrast, C1ql1-Bai3 anterograde signaling is reported to facilitate elimination of weaker CFs after P9, although the same signaling is involved in strengthening/ maintenance of the strongest CF (Kakegawa et al., 2015). As for other molecules involved in the early phase of CF elimination, delayed CF synapse elimination is reported in serotonin 3A (5-HT3A) receptor knockout mice (Oostland et al., 2013). The 5-HT3A receptor appeared to be expressed in GCs, and deletion of this receptor causes accelerated maturation of PC dendrites, abnormal physiological maturation of PFePC synapse, and delayed CF synapse elimination (Oostland et al., 2013). Besides, acceleration of the early phase of CF synapse elimination together with impaired presynaptic maturation of inhibitory synapses is reported in chondroitin sulfate proteoglycan (CSPG)5/neuroglycan C knockout mice (Juttner et al., 2013). Morphological data indicate that CFs that undergo dendritic translocation keep their synapses on the PC soma during the second postnatal week. In contrast, synaptic terminals of the weaker CFs are confined to the soma and the basal part of the primary dendrite. The characteristic pericellular nest consists of somatic synapses originating from collaterals of a single predominant CF and from the other weaker CFs in each PC and thus represents multiple CF innervation of PCs (Hashimoto et al., 2009a). Therefore, CF synapse elimination is thought to be a process of nonselective pruning of perisomatic synapses, which spares dendritic synapses of a single predominant CF and leads to mono-innervation of that CF (Hashimoto et al., 2009a).

4.2.5 Late phase of CF synapse elimination In GluD2 knockout mice, CFs invade into the distal dendrites and form ectopic synapses there (Hashimoto et al., 2001; Hashizume et al., 2013; Ichikawa et al., 2002). These ectopic CF synapses appear around P10 when PF synapse formation and PC dendritic arborization occur most vigorously. The similar type of multiple CF innervation is also found in a mutant mouse deficient in Cbln1 in which PF to PC synapse formation is severely impaired (Hirai et al., 2005). These results indicate that PFs compete for the innervation territory with CFs during development and play a role in restricting CF innervation to proximal dendrites (Fig. 4.4). The details of this phenomenon will be described in Section 4.3.4. Another role of PF synapses is to activate type 1 metabotropic glutamate receptor (mGluR1) and its downstream signaling cascades in PCs to drive the process of CF synapse elimination. It is shown that the mutant mouse deficient in mGluR1 is impaired in CF synapse elimination (Kano et al., 1997; Levenes et al., 1997). Mice deficient in signaling molecules downstream of mGluR1, that is, Gaq, PLCb4, and protein kinase Cg (PKCg), are also impaired in CF synapse elimination (Hashimoto et al., 2000; Ichise et al., 2000; Kano et al., 1995, 1997, 1998). Electrophysiological examination of CF innervation following postnatal development demonstrates that the regression of CF synapse normally occurs during the first and second postnatal weeks in all of the four mouse strains. However, these mice display abnormality in CF synapse elimination during the third postnatal week. These results indicate that the signaling cascade from mGluR1 to PKCg is essential for the late phase of CF synapse elimination (Fig. 4.2B). Importantly, the formation and function of PF to PC synapses are normal in these mutant mice. Therefore, the impaired CF synapse elimination is not caused secondarily by the defect in PF synaptogenesis. The defect in the CF synapse elimination in the mGluR1-deficient mouse is restored in the mGluR1a-rescue mice in which mGluR1a, a variant of mGluR1 generated by alternative splicing (Ferraguti et al., 2008), has been introduced specifically into PCs (Ichise et al., 2000). In contrast, PC-specific expression of mGluR1b, another mGluR1 splice variant, into mGluR1 knockout mice fails to rescue the impaired CF synapse elimination (Ohtani et al., 2014). While mGluR1a has

Cerebellar circuits Chapter | 4

Cbln1 GluD2

87

Myosin Va ?

Parallel fiber Gq

BSRP ?

GLAST

mGluR1a m GluR1a

Strong Climbing fiber

AMPAR

Interneuron

Bergmann glia

4 PLCβ4

Microglia q

PKCγ

Retrograde signaling BDNF

GABAAR

7A ma7 Sema7A

Granule cell

PQ

Maturation

Ca2+

Arc

Gq

Mossy fiber

ItgB1/PlxnC1

AMPAR en endocytosis

AMPAR NMDAR RPγ-2 TARPγ-2

TrkB

Purkinje cell

Weak W Climb Climbing fiber Elimination

FIGURE 4.4 Current diagram for mechanisms underlying the late phase of CF synapse elimination. Neural activity along the pathway of mossy fiber, GC and PF drives mGluR1 to PKCg signaling in PCs. At the downstream of this pathway, Sema7A and BDNF mediate retrograde signaling from PCs to weak CFs through acting on ItgB1/PlxnC and TrkB, respectively. PFePC synapses are stabilized by Cbln1-GluD2 and occupy the postsynaptic sites on the PC distal dendrites, which confine the CF innervation sites to the proximal dendrites. Elevation of intracellular Ca2þ level through Ca2þ influx through P/Q-VDCCs elevates Arc expression, which induces AMPA receptor endocytosis, suppression of CF synaptic efficacy, and eventually facilitates elimination of weak CFs. GABAergic inhibition from molecular layer interneurons suppress Ca2þ influx through P/Q-VDCCs and regulates elimination of weak CFs. Microglia promote maturation of GABAergic inhibition and thereby regulate CF synapse elimination. GLAST in Bergmann glia reduces extracellular glutamate concentration and keeps wrapping of synapses by Bergmann fibers, which ensures strengthening of a single strong CF and elimination of weaker CFs.

a long C-terminal domain that interacts with scaffolding proteins, mGluR1b lacks such long C-terminal domain (Ferraguti et al., 2008). Regression of CF synapses is impaired in mice by PC-specific expression of a PKC inhibitor peptide (De Zeeuw et al., 1998). Furthermore, the distribution of multiply-innervated PCs in the cerebellum of PLCb4 knockout mice exactly matches that of the PCs with predominant expression of PLCb4 in the cerebellum of wild type mice (Kano et al., 1998). These lines of evidence clearly indicate that the signaling from mGluR1 to PKCg in PCs plays a central role in CF synapse elimination. The mGluR1 signaling required for the late phase of synapse elimination is thought to be driven by PF activity, since mGluR1 can readily be activated by PF inputs (Batchelor et al., 1994; Finch and Augustine, 1998; Takechi et al., 1998). Furthermore, chronic blockade of N-methyl-D-aspartate (NMDA) receptors within the cerebellum results in the impairment of CF synapse elimination (Rabacchi et al., 1992), specifically in the late phase (Kakizawa et al., 2000). NMDA receptors are not present at either PF or CF synapses onto PCs during the second and third postnatal weeks, but they are abundantly expressed at MF to GC synapses (Kakizawa et al., 2000). Therefore, the chronic blockade of NMDA receptors within the cerebellum is expected to affect MF to GC transmission. These results suggest that neural activity along MFeGCePFePC pathway and subsequent activation of mGluR1 are prerequisite for the late phase of CF synapse elimination (Fig. 4.4) (Kakizawa et al., 2000). While screening candidate retrograde signaling molecules from PCs to CFs that mediate CF synapse elimination, Uesaka et al. identified molecules that function downstream of mGluR1 to exert CF synapse elimination. MicroRNAmediated knockdown of semaphorin7A (Sema7A) in PCs of neonatal mice causes impairment of the late phase of CF

88

PART | I Circuit development

elimination (Uesaka et al., 2014). Double knockdown of mGluR1 and Sema7A in neonatal PCs impairs CF synapse elimination to the same extent as mGluR1 knockdown alone, indicating that the effect of Sema7A knockdown is occluded by simultaneous mGluR1 knockdown. Knockdown of putative Sema7A receptors, Plexin C1 (PlxnC1) and Integrin B1 (ItgB1), in the inferior olive of neonatal mice causes impairment of CF synapse elimination. Further studies demonstrate that PlxnC1 and ItgB1 facilitate CF synapse elimination by inactivating cofilin and activating focal adhesion kinase, respectively, in CFs (Uesaka and Kano, 2018; Uesaka et al., 2014). Taken together, these results indicate that retrograde Sema7A to PlxnC1/ItgB1 signaling is driven downstream of mGluR1, and facilitates CF synapse elimination after P15 by regulating cofilin and focal adhesion kinase in CFs (Uesaka and Kano, 2018; Uesaka et al., 2014). Following the identification of Sema7A as a downstream molecule of mGluR1, Choo et al. (2017) have recently found that BDNF functions downstream of mGluR1 and promotes the late phase of CF synapse elimination by acting retrogradely onto TrkB in CFs (Choo et al., 2017). Although previous studies show that CF synapse elimination is impaired in mice with global or cerebellum-specific knockout of TrkB (Bosman et al., 2006; Johnson et al., 2007), it remained unknown how and on which cell-type TrkB is activated. Choo et al. (2017) have shown that CF synapse elimination after P15 is impaired in PC-specific BDNF knockout mice, in PCs with microRNA-mediated BDNF knockdown, and in PCs surrounded by CFs in which TrkB is knocked down by injecting lentivirus carrying microRNA against TrkB into the neonatal inferior olive. Knockdown of mGluR1 in PCs of wild type mice and that of PC-specific BDNF knockout mice causes impairment of CF synapse elimination to the same extent, suggesting that mGluR1 and BDNF function along the same signaling pathway. Effect of Sema7A knockdown in PCs is occluded in PC-specific BDNF knockout mice, suggesting that BDNF and Sema7A share a common signaling pathway for the late phase of CF synapse elimination (Choo et al., 2017). Besides the aforementioned mGluR1 and its downstream signaling molecules, GABAergic inhibition to PCs regulates CF synapse elimination (Nakayama et al., 2012). CF synapse elimination after P10 is impaired in mice with deletion of a single allele for the GABA synthesizing enzyme GAD67. GABAergic inhibition from basket cells to PCs is attenuated in the GAD67 heterozygous knockout mice particularly in the second postnatal week. The reduced GABAergic inhibition leads to enhanced Ca2þ transients in PCs following activation of weaker CFs, which may permit the survival of weaker CFs that should otherwise be eliminated (Nakayama et al., 2012). As introduced in the previous section, the P/Q-type VDCC in PC is crucial for the early phase of CF elimination. However, recent studies indicate that the P/Q-type VDCC is also essential for the late phase of CF elimination. The immediate early gene Arc/Arg3.1 is shown to mediate the postsynaptic activity of PCs and facilitate the late phase of CF synapse elimination (Mikuni et al., 2013). Arc expression in PCs was elevated by Ca2þ influx to PCs through P/Q-type VDCC (Mikuni et al., 2013). Moreover, in PC-specific TARPg2 (stargazing) knockout mice in which the amplitudes of EPSCs in PCs are globally scaled down to about 60% of those in wild type mice, the late phase of CF synapse elimination is impaired because of insufficient activation of Arc in PCs (Kawata et al., 2014). Besides aforementioned molecules, a motor protein, myosin Va (Takagishi et al., 2007), a glutamate transporter, GLAST (Miyazaki et al., 2017; Watase et al., 1998) (see Section 4.3.4) and a novel brain-specific receptorlike protein family BSRP (Miyazaki et al., 2006) are also reported to be involved in the late phase of CF synapse elimination. As for other molecules potentially involved in the late phase of CF synapse elimination, null mutant mice deficient in Ca2þ/ calmodulin-dependent protein kinase IV (CaMKIV) are reported to have persistent multiple CF innervation, but it is unclear at what stage of postnatal development the impairment occurs (Ribar et al., 2000). It is also reported that null mutant mice deficient in a-calcium/calmodulin-dependent protein kinase II (CaMKIIa) display multiple CF innervation at P21eP28, but this phenotype disappears in adulthood (Hansel et al., 2006), suggesting that CaMKIIa deficiency delays but does not prevent CF synapse elimination. During synapse elimination in the neuromuscular junction, bulb-shaped tips of retreating motor axons and the axon fragments (axozomes) are engulfed by Schwann cells (Bishop et al., 2004). These axon bulbs, axozomes, and Schwann cell cytoplasm are often positively stained with Lysotracker Red, a marker for lysosomes and late endosomes of living cells, suggesting axonal digestion through autophagy and subsequent heterophagy by Schwann cells (Song et al., 2008). It is also reported that Lysotracker-positive structures surrounding PCs, which are presumed to be within Bergmann glia, are abundant during the second and third postnatal weeks (Song et al., 2008). This result suggests that retreating CF axons might be digested in a manner similar to the retreating motor axons at neuromuscular junction. In the developing retinogeniculate synapses, accumulating evidence indicates that microglia contribute to synapse pruning by engulfment and phagocytosis of redundant synaptic terminals (Schafer et al., 2012). Very recently, Nakayama et al. examined whether and how microglia contribute to CF synapse elimination (Nakayama et al., 2018). In marked contrast to the retinogeniculate synapse, they found no clear morphological evidence for massive engulfment of CFs by microglia. Nevertheless, CF synapse elimination after P10 is severely impaired in mice with genetic deletion of microglia. Notably, GABAergic inhibition on PCs is impaired and CF elimination is restored by enhancement of GABAA receptor

Cerebellar circuits Chapter | 4

89

sensitivity by diazepam. They conclude that microglia primarily promote GABAergic inhibition and secondarily facilitate the mechanism for CF elimination inherent in PCs (Nakayama et al., 2018). A study using organotypic slice culture suggests that CF synapse elimination occurs only during the critical period that depends on the maturation stage of postsynaptic PCs but not on presynaptic olivary neurons (Letellier et al., 2009). The authors cocultured immature or mature medulla containing the inferior olive with naïve or nonnaïve PCs (i.e., PCs that have not undergone synapse elimination or those that have experienced synapse elimination, respectively). Nonnaïve PCs cannot eliminate multiple CFs. These results suggest that CF synapse elimination during the critical period leaves indelible trace in PCs that prevents the elimination process from occurring in the later stage (Letellier et al., 2009). A current overview of the canonical molecular pathways for the late phase of CF synapse elimination is schematically illustrated in Fig. 4.4.

4.3 Development of PFePC synapses 4.3.1 Formation of PFePC synapses PFePC synapses increase in number and mature in the early postnatal period, concomitant with differentiation of GCs and growth of PC dendrites. During the first 10 days of a rodent’s life, production and migration of GCs as well as growth of PC dendrites are slow (Altman and Bayer, 1997). In this period, the proliferative or outer zone of the external granular layer has constant thickness of four to 5 cells, while the depth of the premigratory or inner zone increases progressively (Altman and Bayer, 1997). In the premigratory zone, postmitotic GCs extend future PFs in the transverse plane, and then migrate downwards along Bergmann fibers in the molecular layer (Altman and Bayer, 1997; Rakic, 1971). As a consequence, T-shaped axons of GCs differentiate, and horizontal beams of newly generated PFs pile up on the superficial upper zone of the molecular layer. PC dendrites are immature, particularly, in the upper zone of the molecular layer, where dendrites extend filopodiumlike protrusions, PFePC synapses are few in number, spines are often free of innervation, and coverage by Bergmann glia is incomplete (Kurihara et al., 1997; Yamada et al., 2000). In the next 10 postnatal days, PC dendrites grow dynamically, the bulk of GCs come into existence, and PFePC synapses explosively increase in number (Takacs and Hamori, 1994). Moreover, almost all spines form synaptic contact with PF terminals, and PFePC synapses are equipped with well-developed postsynaptic density and complete coverage by Bergmann glia (Kurihara et al., 1997; Spacek, 1985; Yamada et al., 2000). Analyses using agranular and hypogranular animal models, where PFePC synaptogenesis is hindered by spontaneous gene mutation or postnatal X-ray irradiation, have demonstrated that PF synapse formation in PCs plays a critical role in the elongation, branching, and development of dendritic trees (Sotelo, 2004).

4.3.2 Stabilization and maintenance of PFePC synapses Our understanding of the molecular mechanism of PF synapse formation in PCs has been greatly advanced by molecular identification of GluD2 (Araki et al., 1993; Lomeli et al., 1993) and through phenotypic analyses of mutant mice defective in this gene grid2 (GluD2-knockout mice and spontaneous hotfoot mutant mice (Guastavino et al., 1990; Kashiwabuchi et al., 1995)). Like other ionotropic GluR subunits, GluD2 preserves three transmembrane domains (TM1, TM3, and TM4), reentrant hairpin loop (TM2) surrounding a channel pore, ligand binding domains in the N-terminal region, and protein-protein interaction sites in the C-terminal region (Uemura et al., 2004; Yuzaki, 2004). However, GluD2 does not function as a glutamate-gated ion channel (Kakegawa et al., 2007a, 2007b). In the brain, GluD2 is strongly expressed in PCs (Araki et al., 1993; Lomeli et al., 1993), and selectively localized at PF but not CF synapses (Landsend et al., 1997; Takayama et al., 1995). PCs in GluD2-defective mice display characteristic phenotypes mostly related to PF synapse structure and function, including reduction in the number of PF synapses per PC to about half of the wild type mice (54% in control PCs (Kurihara et al., 1997), emergence of free spines lacking synaptic contact in the distal dendritic domain (37% of the total spines (Ichikawa et al., 2002; Kurihara et al., 1997)), mismatching of pre- and postsynaptic specialization at PF synapses (Guastavino et al., 1990; Takeuchi et al., 2005), impaired long-term depression (LTD) at PF-PC synapses (Kakegawa et al., 2008; Kashiwabuchi et al., 1995; Uemura et al., 2007), impaired motor learning (Kakegawa et al., 2008; Kishimoto et al., 2001), and severe ataxia (Guastavino et al., 1990; Kashiwabuchi et al., 1995). In a drug-inducible, PC-specific GluD2-knockout mouse strain, mismatched PFePC synapses, free spines, and motor discoordination are induced and exacerbated in the adult cerebellum, concomitant with a decrease in GluD2 protein (Takeuchi et al., 2005). Furthermore, expression of the N-terminal domain of GluD2 in human kidney cells induces presynaptic differentiation (Uemura and Mishina, 2008) and its viral transfer to adult GluD2-knockout mice rapidly

90

PART | I Circuit development

restores PF synapse formation and motor coordination (Kakegawa et al., 2009). Furthermore, while injured axons usually do not regenerate in the adult central nervous system, PFePC synapses regenerate rapidly after surgical transection of PFs (Chen and Hillman, 1982). Importantly, such regeneration does not occur at all in GluD2-knockout mice (Ichikawa et al., 2016b). Thus, GluD2 strengthens, maintains, and regenerates the connectivity of PFePC synapses in both the developing and adult cerebella. On the other hand, the last seven amino acids known as the T-site, which binds to various PDZ domain-containing proteins, including postsynaptic density (PSD)-93, PTPMEG, delphilin, nPIST, and synaptic scaffolding molecule, are essential for cerebellar LTD and motor learning (Kakegawa et al., 2008; Uemura et al., 2007). Cbln1 or precerebellin was originally identified as a precursor of the PC-specific peptide cerebellin (Slemmon et al., 1984). However, C-terminal two-thirds of Cbln1 share significant structural similarity with the globular domain of complement C1q chain (Urade et al., 1991), and the full-length Cbln1 is secreted into the culture medium as a hexameric complex (Bao et al., 2005). Thus, Cbln1 now belongs to the C1q/tumor necrosis factor superfamily. Of four members (Cbln1e4), Cbln1 is highly expressed in cerebellar GCs together with Cbln3 (Hirai et al., 2005; Miura et al., 2006), exists as Cbln1 homomeric and Cbln1/3 heteromeric complexes (Iijima et al., 2007; Pang et al., 2000), and selectively accumulates in the synaptic cleft facing PF terminals, but not CF terminals, in PCs (Iijima et al., 2007; Miura et al., 2009). Cbln1-knockout mice show characteristic phenotypes similar to, or even severer than, GluD2-knockout mice (Hirai et al., 2005). In PCs of Cbln1-knockout mice, 78% of the spines are free of innervation and 14% have mismatching in pre- and postsynaptic differentiation (Figs. 4.5). Hence, only 8% of spines in Cbln1-knockout mice establish normal matched synapses with PF terminals. LTD at PFePC synapses and motor coordination are also impaired. When recombinant Cbln1 is applied to the subarachnoid space of adult Cbln1-knockout mice, only a single injection can rapidly restore PFePC synapse structure and function, and cerebellar ataxia (Ito-Ishida et al., 2008). These common phenotypes in GluD2- and Cbln1-knockout mice and their rapid rescue by genetic or molecular supplementation are explained by the fact that

(A)

(B)

f

f

f

m f

(C)

(D) f f f

f

m

ff n

f/m (?)

Scale bar, 200 nm.

FIGURE 4.5 Electron micrographs of free spines (f) and mismatched synapses (m) in Cbln1-knockout mice. Free spines are thoroughly enwrapped by the lamellate processes of Bergmann glia. Arrowheads indicate the portion of the PSD that is not opposed by the presynaptic active zone. Normal matched synapses (n) are very rare in this mutant. In (d), the presence of postsynaptic density with no presyaptic terminal differentiation suggests that the spine with such postsynaptic density is either free or mismatched spine (hence labeled “f/m(?).” Scale bar ¼ 0.2 mm. Reproduced from Watanabe, M., Molecular mechanisms governing competitive synaptic wiring in cerebellar Purkinje cells. Tohoku Journal of Experimental Medicine 214, 2008:175e190, with permission.

Cerebellar circuits Chapter | 4

91

transsynaptic interaction of postsynaptic GluD2 and presynaptic neurexins is mediated by Cbln1. Thus, through this unique interaction, Cbln1 acts as a bidirectional synaptic organizer for both pre- and postsynaptic components at PFePC synapses (Matsuda et al., 2010; Uemura et al., 2010). The strengths of PF synapses, but not CF synapses, in PCs are selectively decreased when postsynaptic mGluR1 or inositol 1,4,5-trisphosphate (IP3) signaling is chronically inhibited, when PF activity is inhibited by suppressing NMDA receptor-mediated inputs to GCs, or when antibody against BDNF is applied in vivo (Furutani et al., 2006). The weakening of PFePC synaptic strength is reversed by in vivo application of BDNF. These results suggest that mGluR1 activation and the following IP3 signaling maintain presynaptic function through BDNF at PFePC synapses. In this regard, it is interesting to note that free spines on PC dendrites emerge in the spontaneous ataxic mutant rigoletto (rig) (also known as waddles; wdl), which is caused by a 19 bp deletion in the exon eight of carbonic anhydrase-related protein Car8 (Hirasawa et al., 2007). Car8 is known to bind to IP3 receptor and reduces its affinity for IP3 (Hirota et al., 2003). In future studies, it is important to investigate how Car8 modulates the mGluR1-IP3 signaling and whether Car8 is involved in the GluD2Cbln1-neurexin interaction, both of which regulate the connectivity of PFePC synapses.

4.3.3 Developmental elimination of PFePC synapses In addition to mono-innervation by CFs, segregation of the PF and CF territories on PC dendrites is another distinguished feature in excitatory synaptic organization of PC. Until P9, CF and PF territories are initially separated, because CFs innervate the soma and PFs are connected to the dendrite. Then, these two territories become overlapped by commencement of dendritic translocation of a single “winner” CF from P9 to P15, when PFs remain to innervate the whole extent of PC dendritic tree (Ichikawa et al., 2016a). From P15 to P20, the territories segregate again by massive elimination of PF synapses from the proximal compartment of PC dendrites. This developmental process of PF synapse elimination is impaired in either of the mutant mouse strains lacking mGluR1, PKCg, or P/Q-type VDCC, leading to sustained overlapping territories of CF and PF innervation at the proximal compartment of PC dendrites (Ichikawa et al., 2016a; Miyazaki et al., 2004, 2012).

4.3.4 Heterosynaptic competition between PF and CF inputs The proximal compartment of PC dendrite is smooth in contour due to low spine density and is innervated by a single CF. On the other hand, the distal compartment is made up of spiny branchlets studded with numerous spines and innervated by PFs (Fig. 4.6, middle). Accumulated experimental evidence indicates that the construction of such excitatory synaptic organization stands on competitive equilibrium between CFs and PFs, whose expansions are promoted by distinct

FIGURE 4.6 Summary diagram of molecular mechanisms for competitive synaptic wiring in PCs. Note that the climbing fiber (CF) and parallel fiber (PF) territories are reversed in mutant mice defective in GluD2/Cbln1 (left) and P/Q-type VDCC (right). With both mechanisms, CF and PF territories are sharply segregated, and CF mono-innervation is established in wild type animals (middle). Reproduced from Watanabe, M., 2008. Molecular mechanisms governing competitive synaptic wiring in cerebellar Purkinje cells. Tohoku Journal of Experimental Medicine 214, 2008:175e190, with permission.

92

PART | I Circuit development

mechanisms. Surgical, pharmacological, and genetic manipulations that shift this equilibrium can alter the organized innervation patterns. Disruption of GluD2 gene in mice not only causes abnormal structure and function of PF synapses but also affects the mode of CF innervation (Ichikawa et al., 2002). In the molecular layer, CF branches are distributed in the inner four-fifth (84% of the molecular layer thickness) in control mice, whereas their distribution almost reaches the pial surface (95%) in GluD2-knockout mice. When the tracer-labeled CFs are followed from the soma to the tips of PC dendrites by serial electron microscopy, CF branches in GluD2-knockout mice extend distally and take over the free spines on the distal compartment of PC dendrites. Such aberrant extension occurs toward not only distal dendrites of the same PCs but also those of the neighboring PCs. The latter type of spine takeover results in multiple innervation of a given PC by CFs of different neuronal origins. This anatomical evidence for multiple CF innervation is consistent with the data from electrophysiological recording combined with Ca2þ imaging. In GluD2-knockout mice, a single strong CF elicits large EPSCs with a fast rise time and large Ca2þ transients over the entire dendritic tree, whereas weak CFs elicit small EPSCs with a slow rise time and small Ca2þ elevation that is confined to distal dendrites (Hashimoto et al., 2001). These findings indicate that GluD2 is essential for restricting CF innervation to the proximal dendritic compartment and thereby preventing multiple CF innervation at the distal dendritic compartment (Fig. 4.4, Fig. 4.6 right). This mechanism is also active in the adult cerebellum. The ablation of GluD2 in adulthood also leads to progressive distal extension of ascending branches of CFs, and they aberrantly innervate distal dendrites of the target and neighboring PCs (Miyazaki et al., 2010). Furthermore, transverse branches of CFs, which are short motile collaterals forming no synapses in wild type animals (Nishiyama et al., 2007), display aberrant mediolateral extension to innervate distal dendrites of neighboring and remote PCs. Consequently, many PCs are connected by single main CF and surplus CFs that innervate small parts of the distal dendrites. Surplus CF-EPSCs with slow rise time and small amplitude also emerge progressively after GluD2 ablation. Therefore, GluD2 is essential to keep CF mono-innervation in the adult cerebellum by suppressing aberrant invasion of CF branches to the territory of PF innervation. In contrast, CF innervation is regressed but PF innervation expands to the proximal compartment, when surgical lesion to olivocerebellar projections is made in adult animals or activities in the cerebellar cortical neurons are blocked with the sodium channel blocker tetrodotoxin or with the AMPA receptor antagonist 2,3-dioxo-6-nitro-1,2,3,4tetrahydrobenzo[f]quinoxaline-7-sulfonamide (Bravin et al., 1995; Cesa et al., 2007; Kakizawa et al., 2005). The latter change often accompanies hyperspiny transformation or impaired elimination of PF synapses at the proximal dendritic compartment (Bravin et al., 1995; Cesa et al., 2007). Similar changes are reproduced in mice with global knockout of the P/Q-type VDCC (Miyazaki et al., 2004). In P/Q-type VDCC knockout mice, hyperspiny transformation is induced at proximal dendrites and somata of PCs, and many of these ectopic spines are innervated by PF terminals. Conversely, the distribution of CFs is regressed to lower portions of the molecular layer, and they innervate spines from somata and basal dendrites. Furthermore, in more than 90% of P/Q-type VDCC knockout PCs, their basal dendrites and somata are innervated by CFs of different neuronal origins. As a result, the proximal somatodendritic compartment in P/ Q-type VDCC-lacking PCs receives chaotic innervation by numerous PFs and multiple CFs (Fig. 4.6, right). Thus, CF activities leading to AMPA receptor activation and subsequent Ca2þ influx through P/Q-type VDCCs are essential for monopolizing the proximal dendritic compartment by a single main CF and for expelling other excitatory inputs from that compartment. The glutamate transporter GLAST is exclusively expressed in Bergman glia, and plays essential roles in competitive processes of PC synaptic wiring. Genetic ablation and pharmacological blockade of GLAST impair the establishment and maintenance of mono-innervation by the single strongest CF and PF synapse elimination from the proximal dendritic compartment of PC (Miyazaki et al., 2017). Without GLAST that keeps the local extracellular glutamate concentration low, the competitive mechanisms for constructing glutamatergic CF and PF synaptic organization may be disturbed. Because of high ambient glutamate level, synapses of predominant inputs may not be strengthened enough to eliminate those of less predominant inputs, while the latter themselves may not be further weakened and tend to survive. Taken altogether, excitatory synaptic wiring in PC is formed and maintained through homosynaptic competition among CFs and heterosynaptic competition between PFs and CFs. GluD2 and Cbln1 fuel heterosynaptic competition in favor of PF innervation, whereas the P/Q-type VDCC, mGluR1 signaling pathway, and GLAST facilitate both heterosynaptic and homosynaptic competitions in favor of mono-innervation by a single main CF. Based on these molecular mechanisms, PCs can establish two distinguished features of their excitatory synaptic wiring, that is, territorial innervations by PF and CF inputs and mono-innervation by a single “winner” CF.

Cerebellar circuits Chapter | 4

93

4.4 Development of inhibitory synapses from basket cells and stellate cells to PCS 4.4.1 Formation of basket Cell-PC synapses Basket cells innervate the AIS of PC and form the characteristic structure “pinceau” in the mature cerebellum (Ito, 1984; Palay and Chan-Palay, 1974). The innervation by basket cell axons of PCs seems to begin when basket cells migrate across the PC layer at the end of the first postnatal week (Ango et al., 2004). Basket cell axons initially form synaptic contacts on the somata of PCs. Then, they appear to move directly to the AIS without searching for other possible targets. Upon reaching the AIS, basket cell axons extend multiple terminal branches and establish the pinceau. Several molecules have been identified that accumulate at the AIS of PC. These include membrane-associated adaptor protein ankyrin-G and one of its binding partner, neurofascin 186 (NF186), a splice variant of neurofascin which belongs to the L1 subgroup of the Ig superfamily (Brummendorf et al., 1998). It is thought that ankyrin-G is stabilized at the AIS partly through its interaction with b4-spectrin tetramers which bind to actin network (Davis et al., 1996). Ankyrins and b-spectrins are known as intracellular adaptor proteins that recruit ion channels, transporters, and cell adhesion molecules to subcellular domains. They are thought to constitute microdomains for intercellular contact and signaling (Bennett and Baines, 2001; Bennett and Chen, 2001). On the other hand, Neurofascin is known as a cell-surface glycoprotein that is shown to mediate axon-axon interactions in vitro (Rathjen et al., 1987). NF186 exhibits subcellular concentration gradient in PCs from the AIS toward the soma and dendrites, being highest at the AIS and very low at the top of the soma and in dendrites (Fig. 4.7A) (Ango et al., 2004). This gradient is already formed at the end of the first postnatal week when basket cell axons first contact the somata of PCs Fig. 4.7A). Evidence for the requirement of ankyrin-G and NF186 in targeting of basket cell axon and formation of the pinceau at the AIS of PC has been presented by the analysis of ankyrin-G knockout mice in which the NF186 gradient in PCs is abolished (Fig. 4.7B) (Ango et al., 2004). In these mice, basket cell axons are not restricted to the AIS but instead are present on the soma and slightly more distal portion of PC axons (Fig. 4.7B). Although some basket cell axon bundles successfully reach the AIS, they are very thin, extend along PC axons abnormally, and follow the ectopic localization of NF186. The synaptic contacts visualized by a GABA synthesizing enzyme, GAD65, are greatly reduced at the pinceau. These results strongly suggest that NF186 is a substrate for the growth of basket cell axons and that NF186 bound to ankyrin-G at the AIS is required for the stabilization of the pinceau. Indeed, expression of the dominant-negative form of NF186 into PCs during postnatal development caused abnormal organization of the pinceau (Ango et al., 2004). On the other hand, conditional deletion of NF186 from PCs of developing mice prevents maturation of the AIS and causes

(A)

(B)

Wild-type

Molecular layer BG

PC

BC

SC

BC BG

BG

CHL1 −/−

SC

SC

BC

(C)

Ankyrin-G −/−

PC

BG

BG

PC

BG

Purkinje layer

Neurofascin gradient

Granule layer

FIGURE 4.7 Subcellular specificity of basket cell and stellate cell inhibitory connections to PCs. Schematics showing the inhibitory synaptic connectivity patterns of the wild type (A), ankyrin-G deficient (B), and CHL1 deficient (C) mice. In wild type mice, a sharp gradient of neurofascin is present from AIS toward the soma. In ankyrin-G deficient mice, this gradient is no longer restricted to the AIS, which causes mistargeting of basket cell axons and reduced synapse formation. In CHL1 deficient mice, stellate cell axons are not properly guided by Bergman glia fibers and synapse formation is decreased. SC, stellate cell; BC, basket cell; PC, Purkinje cell; BG, Bergman glia. Reproduced from Williams ME, de Wit J, and Ghosh A (2010) Molecular mechanisms of synaptic specificity in developing neural circuits. Neuron 68: 9e18, with permission.

94

PART | I Circuit development

disorganization of the pinceau (Buttermore et al., 2012). It is also reported that conditional deletion of NF186 from both PCs and basket cells causes abnormal basket cell axon collateral branching and targeting to the PC soma/AIS, and severe disorganization of the pinceau (Buttermore et al., 2012). These results indicate that NF186 in PCs is required for AIS maturation and for organization of the pinceau, whereas NF186 in basket cells in combination with that in PCs is required for proper basket cell axon collateral outgrowth and targeting to the PC soma/AIS (Buttermore et al., 2012). Furthermore, inducible conditional deletion of NF186 from adult mouse PCs causes disintegration of the AIS and disorganization of the pinceau (Zonta et al., 2011), indicating that NF186 in PCs is required for maintaining basket cell-PC wiring and integration of the AIS in the adult cerebellum. Later studies show that besides NF186, Sema3A and its cognate receptor Neuropilin-1 (NRP1) are essential for establishment of basket cell-PC wiring during postnatal development (Cioni et al., 2013; Telley et al., 2016). Sema3A is presumably secreted from PC and concentrated around the PC layer. In Sema3A knockout mice and mice with specific deletion of NRP1 from basket cells, the axon shaft and collaterals of basket cells are severely disorganized, formation of the pinceau is impaired (Telley et al., 2016), and branching of basket cell axons around the PC soma/AIS is reduced (Cioni et al., 2013). NRP1 in basket cell axons transsynaptically binds to NF186 in the PC AIS during construction of the pinceau (Telley et al., 2016). These results suggest that NRP1 in basket cells contributes to axon collateral guidance to the PC soma (NRP1 in basket cell axons is attracted by secreted Sema3A) (Telley et al., 2016), collateral contact and recognition of the PC soma/AIS (Sema3A stabilizes NRP1 on basket cell axon collaterals and facilitates target recognition through transsynaptic interaction with NF186) (Telley et al., 2016), and terminal branching of basket cell axons at the PC AIS (through interaction with extracellular Sema3A and the Src kinase Fyn in basket cell axons) (Cioni et al., 2013).

4.4.2 Formation of stellate Cell-PC synapses Stellate cells are GABAergic inhibitory interneurons whose somata are located in the molecular layer. In the mature cerebellum, their axons innervate dendrites of PCs with ascending and descending collaterals, and with a plexus of finer branches and terminals. Similarly to basket cells, stellate cells are derived from dividing progenitors in the white matter of postnatal cerebellum (Zhang and Goldman, 1996). Stellate cell precursors migrate into the molecular layer a few days later than basket cell precursors, with a peak between P8 and P11 but continuing till P14 (Yamanaka et al., 2004). By using a green fluorescent protein-bacterial artificial chromosome transgenic reporter mouse line in which GFP was mainly expressed in stellate cells and basket cells, Huang and colleagues demonstrated how stellate cell axons establish their innervations of PC dendrites after they reach the molecular layer (Ango et al., 2008). Between P12 and P16, stellate cells become bipolar and extend neurites in horizontal orientation. Then, at P16eP18, stellate cell axons send ascending and descending collaterals, which are further elaborated with appearance of plexus of finer branches up to P40. Importantly, both ascending and descending collaterals of stellate cell axons are strictly associated with the fibers of Bergmann glia that are visualized by the staining of the glia-specific cytoskeleton protein glial fibrillary acidic protein (GFAP). During the third and fourth postnatal weeks, Bergmann fibers are known to extend lateral varicoses and fine processes and form an extensive reticular meshwork. Radial fibers from neighboring Bergmann glia are aligned to form palisades in the plane perpendicular to PC dendrites (Altman and Bayer, 1997). In contrast to close association to Bergmann fibers, stellate cell axons do not follow PC dendrites. Bergmann fibers enwrap segments of PC dendrites in a patchy, en passant pattern, which is in contrast to close association to stellate cell axons. Triple immunolabeling of stellate cell axon terminals, Bergmann fibers and PC dendrites indicates that stellate cell boutons are formed at the intersection between Bergmann fibers and PC dendrites. Thus, Bergmann fibers may function as an intermediate scaffold to guide stellate cell axons along the characteristic trajectories toward multiple PC dendrites and to form synaptic contacts (Fig. 4.7A) (Ango et al., 2008). Since neurofascin, a member of the L1 cell adhesion molecule (L1CAM) subfamily, is crucial for targeting of basket cell axons to the AIS of PC, it is possible that other members of L1CAM might be important for the targeting of stellate cell axons to PC dendrites. A systematic survey of the expression patterns of L1CAMs during postnatal cerebellar development revealed that close homolog of L1 (CHL1) is distributed in a radial stripe pattern that exactly matches the expression of the Bergmann glia marker GFAP, but not the PC marker calbindin (Ango et al., 2008). CHL1 expression in Bergmann fibers is prominent as early as P8, reaching higher level around P18, and then declining in adulthood. CHL1 is also expressed in stellate cells, but the expression is delayed, being undetectable at P8, becoming obvious around P14, and remaining detectable in adulthood (Ango et al., 2008). These expression patterns suggest the involvement of CHL1 in the stellate cell axon targeting to PC dendrites. A series of experiments using CHL1-deficient mice have clarified the importance of Bergmann fibers and CHL1 in organizing stellate cell innervations of PC dendrites (Fig. 4.7C) (Ango et al., 2008). In CHL1 knockout mice, stellate cell axons exhibit abnormal trajectories and orientations, and aberrant innervations of PC dendrites (Fig. 4.7C). In addition,

Cerebellar circuits Chapter | 4

95

there is a clear reduction in the staining of GAD65, a marker for GABAergic inhibitory synaptic terminal, and the density of stellate cell synapses along PC dendrites. Such aberrant stellate cell axons can form morphologically normal synapses onto PC dendrites albeit with reduced efficacy and density, but many of them are not maintained and stellate cell axon terminals become atrophic with age. In contrast, there is no change in the GAD65 staining in the AIS of PC (Fig. 4.7C). Furthermore, there is no change in excitatory synapses from PFs and CFs in CHL1-deficient mice. Importantly, the selective defect in stellate cell innervation and synapse formation on PC dendrites is observed in mice with conditional deletion of CHL1 in Bergmann glia. These results demonstrate that Bergmann fibers function as guiding scaffolds, and CHL1 is a molecular signal for organization of stellate cell axon arbors and directing their innervations of PC dendrites (Huang et al., 2007; Williams et al., 2010).

4.4.3 Activity-dependent remodeling of inhibitory synapses Several lines of evidence indicate that GABAA receptor-mediated signaling coordinates pre- and postsynaptic maturation during activity-dependent development of inhibitory synapses (Huang, 2009; Huang et al., 2007; Huang and Scheiffele, 2008). For example, altering GABA synthesis by manipulating the expression of GAD67 greatly influences inhibitory synaptic innervation in the visual cortex (Chattopadhyaya et al., 2007). Acute suppression of the g2 subunit of GABAA receptor not only disrupts GABAA receptor clustering but also reduces innervations of the g2-deficient neurons by GABAergic terminals (Li et al., 2005). In cerebellar PCs, genetic deletion of the a1 subunit of GABAA receptor in mice causes complete loss of functional GABAA receptors and synaptic inhibition in PCs by P18 (Fritschy et al., 2006). Morphologically, GABAergic synaptic terminals from stellate cells are reduced by 75%, whereas basket cell synapses on the PC soma are not affected. During postnatal development, GABAergic terminals from stellate cells are initially formed normally onto PC dendritic shafts. PCs of the a1 knockout mice transiently express the a3 subunit of GABAA receptor and have functional GABAA receptors during early postnatal development (Patrizi et al., 2008). However, subsequent downregulation of the a3 subunit results in complete loss of GABAergic currents and a decreased rate of GABAergic synaptogenesis (Patrizi et al., 2008). Simultaneously, ectopic mismatched synapses begin to be formed between GABAergic terminals and PC dendritic spines (Fritschy et al., 2006; Patrizi et al., 2008) on which normally glutamatergic excitatory terminals make synaptic contacts. The postsynaptic adhesion molecule neuroligin-2 is correctly targeted to inhibitory synapses lacking GABAA receptors, whereas neuroligin-2 is absent from the mismatched synapses albeit the presence of GABAergic terminals (Patrizi et al., 2008). The same group generated another mouse model in which downregulation of the g2 subunit of GABAA receptor results in mosaic expression of GABAA receptors in neighboring PCs (Frola et al., 2013). In PCs lacking g2 subunit, the number of GABAergic synapses on dendrites is reduced. In contrast, in neighboring PCs expressing the g2 subunit, GABAergic innervation on dendrites is significantly increased when compared to PCs in wildtype mice (Frola et al., 2013). Notably, ectopic GABAergic terminals on dendritic spines, which are frequently seen in the a1 subunit knockout mice (Fritschy et al., 2006), are not observed, indicating that the presence of synaptic GABAA receptors in subsets of PCs prevents the formation of aberrant GABAergic contacts (Frola et al., 2013). Taken together, these results suggest that GABAA receptors are not required for the formation of synapses, but they appear to be crucial for activity-dependent regulation of synaptic density, presumably through promoting the stabilization of transient axodendritic contact into mature inhibitory synapses.

4.5 Summary and conclusions The cerebellum provides a good system to study how microcircuits are formed during peri- and postnatal development. The cerebellar cortex consists of only seven types of neurons, that is, PC, GC, basket cell, stellate cell, Golgi cell, Lugaro cell, and unipolar brush cell, and there are two glutamatergic excitatory afferents. The PC is the sole output neuron of the cerebellar cortex and inhibits neurons in the DCN. Bergmann glia, the characteristic astrocyte in the cerebellar cortex, plays multiple roles in neural circuit formation and synaptic transmission such as migration of GCs and guidance of stellate cell axons. The neural circuits made by these cell types and afferents are basically the same throughout the cerebellum. In the first section of this chapter, we briefly described the cell types and the synaptic organization of the cerebellum and how these cells are generated and migrate to their final positions. We also mentioned the mediolateral compartmentalization based on olivocerebellar projection and some molecular background of the compartmentalization. In the second section, we made an overview of postnatal development of CFePC synapses, which is one of the beststudied examples of synapse elimination in the brain. Shortly after birth, each PC is innervated by multiple CFs with similar synaptic strengths on the soma. Subsequently, a single CF is selectively strengthened during the first postnatal week. Then, at around P9, only the strongest CF (“winner” CF) starts to extend its innervation to PC dendrites. In contrast,

96

PART | I Circuit development

synapses of the weaker CFs (“loser” CFs) remain on the soma and the most proximal portion of the dendrite, and they are eliminated progressively during the second and third postnatal weeks. From P7 to P11, the elimination proceeds independently of PFePC synapse formation. From P12 and thereafter, the elimination of weaker CFs requires normal PFePC synapse formation and is dependent on the PF synaptic inputs that activate mGluR1 and its downstream signaling in PCs. P/Q-VDCCs in PCs play crucial roles in all of the four phases of CF synapse elimination, whereas GABAergic inhibition regulates mainly the late phase. We also described other key molecules such as Arc and Sema7A that are involved in respective phases of CF synapse elimination. In the third section, we described how PF synapses are formed and maintained on dendritic spines of PCs. We introduced how the GluD2eCbln1eneurexin system stabilizes and maintains PFePC synapses. We also showed that innervation territories of PFs and CFs on PC dendrites stand on the equilibrium caused by heterosynaptic competition between PFs and CFs and homosynaptic competition between multiple CFs. Furthermore, we described PF synapse elimination from PC proximal dendrites from around P15 to P30 after the establishment of CF mono-innervation. In the fourth section, we summarized how GABAergic inhibitory synapses from basket and stellate cells are targeted to the PC soma/AIS and dendrites, respectively. Basket cell axons seem to be guided to the PC soma/AIS following the gradient of neurofascin, a member of L1CAM. This gradient is caused by cross-linking of neurofascin to ankyrin-G that is localized to the AIS. We also described the roles of Sema3A and NRP1 in basket cell axon collateral guidance to the PC soma, collateral contact and recognition of the PC soma/AIS, and terminal branching of basket cell axons at the PC AIS. On the other hand, stellate cells direct their axons along Bergmann glia fibers to PC dendrites. The association of stellate cell axons to Bergmann fibers is mediated by CHL1, another member of L1CAM. We also mentioned that GABAA receptors appear to be crucial for activity-dependent regulation of the density of inhibitory synapses on PCs. The cerebellum has been attracting many neuroscientists who pursue the mechanisms of synapse formation, synapse elimination, and synapse remodeling. The small number of cell types and little regional variation in the layer structure and synaptic organization in the cerebellum enable us to perform quantitative and detailed morphological and electrophysiological analyses, when compared to other brain areas. An example is the study of developmental synapse elimination. While this process is intensively studied in peripheral synapses such as neuromuscular junction and autonomic ganglia, it is generally very difficult to do detailed analyses of synapse elimination in the CNS because of small synapse size, heterogeneity and abundance of synaptic inputs to each neuron, and the complexity of synaptic organization. The CF to PC synapse is one of a few examples in the CNS in which developmental synapse elimination can be studied quantitatively by electrophysiological and morphological techniques. Synapses from retinal ganglion cells to the lateral geniculate nucleus and those from the lateral lemniscus to the ventral basal thalamus are also known to undergo massive elimination during postnatal development and have been studied intensively. Formation of inhibitory topographic map in the auditory brainstem is another example in which developmental refinement occurs through synapse elimination. As for the molecular and cellular mechanisms, developmental synapse elimination at CF to PC connection is best characterized among the four types of synapses, as detailed in this review. Thus, CF synapse elimination is an excellent model system for the study of developmental synapse refinement, which is comparable to that of the visual cortex. Continuing researches on cerebellar microcircuits will elucidate fundamental mechanisms of the formation, elimination, maturation, and maintenance of neural circuits in developing CNS.

Acknowledgments We thank Takaki Watanabe for updating Fig. 4.4. This work has been supported in part by Grants-in-Aid for Scientific Research 25000015 (M.K.) and 24220007 (M.W.) from JSPS, Japan, by Brain/MINDS from AMED, Japan, and by SRPBS from AMED, Japan.

References Altman, J., Bayer, S.A., 1997. Development of the Cerebellar System : In Relation to its Evolution, Structure, and Functions. CRC Press, Boca Raton, FL. Andjus, P.R., Zhu, L., Cesa, R., Carulli, D., Strata, P., 2003. A change in the pattern of activity affects the developmental regression of the Purkinje cell polyinnervation by climbing fibers in the rat cerebellum. Neuroscience 121, 563e572. Ango, F., di Cristo, G., Higashiyama, H., Bennett, V., Wu, P., Huang, Z.J., 2004. Ankyrin-based subcellular gradient of neurofascin, an immunoglobulin family protein, directs GABAergic innervation at purkinje axon initial segment. Cell 119, 257e272. Ango, F., Wu, C., Van der Want, J.J., Wu, P., Schachner, M., Huang, Z.J., 2008. Bergmann glia and the recognition molecule CHL1 organize GABAergic axons and direct innervation of Purkinje cell dendrites. PLoS Biol. 6, e103. Apps, R., Hawkes, R., 2009. Cerebellar cortical organization: a one-map hypothesis. Nat. Rev. Neurosci. 10, 670e681. Araki, K., Meguro, H., Kushiya, E., Takayama, C., Inoue, Y., Mishina, M., 1993. Selective expression of the glutamate receptor channel d2 subunit in cerebellar Purkinje cells. Biochem. Biophys. Res. Commun. 197, 1267e1276.

Cerebellar circuits Chapter | 4

97

Armengol, J.A., Sotelo, C., 1991. Early dendritic development of Purkinje cells in the rat cerebellum. A light and electron microscopic study using axonal tracing in ’in vitro’ slices. Brain Res. Dev. Brain Res. 64, 95e114. Bailly, Y., Rabacchi, S., Sherrard, R.M., Rodeau, J.L., Demais, V., Lohof, A.M., Mariani, J., 2018. Elimination of all redundant climbing fiber synapses requires granule cells in the postnatal cerebellum. Sci. Rep. 8, 10017. Bao, D., Pang, Z., Morgan, J.I., 2005. The structure and proteolytic processing of Cbln1 complexes. J. Neurochem. 95, 618e629. Bao, J., Reim, K., Sakaba, T., 2010. Target-dependent feedforward inhibition mediated by short-term synaptic plasticity in the cerebellum. J. Neurosci. 30, 8171e8179. Barbour, B., 1993. Synaptic currents evoked in Purkinje cells by stimulating individual granule cells. Neuron 11, 759e769. Batchelor, A.M., Madge, D.J., Garthwaite, J., 1994. Synaptic activation of metabotropic glutamate receptors in the parallel fibre-Purkinje cell pathway in rat cerebellar slices. Neuroscience 63, 911e915. Bennett, V., Baines, A.J., 2001. Spectrin and ankyrin-based pathways: metazoan inventions for integrating cells into tissues. Physiol. Rev. 81, 1353e1392. Bennett, V., Chen, L., 2001. Ankyrins and cellular targeting of diverse membrane proteins to physiological sites. Curr. Opin. Cell. Biol. 13, 61e67. Bishop, D.L., Misgeld, T., Walsh, M.K., Gan, W.B., Lichtman, J.W., 2004. Axon branch removal at developing synapses by axosome shedding. Neuron 44, 651e661. Blot, A., Barbour, B., 2014. Ultra-rapid axon-axon ephaptic inhibition of cerebellar Purkinje cells by the pinceau. Nat. Neurosci. 17, 289e295. Bosman, L.W., Hartmann, J., Barski, J.J., Lepier, A., Noll-Hussong, M., Reichardt, L.F., Konnerth, A., 2006. Requirement of TrkB for synapse elimination in developing cerebellar Purkinje cells. Brain Cell Biol. 35, 87e101. Bosman, L.W., Takechi, H., Hartmann, J., Eilers, J., Konnerth, A., 2008. Homosynaptic long-term synaptic potentiation of the "winner" climbing fiber synapse in developing Purkinje cells. J. Neurosci. 28, 798e807. Bravin, M., Rossi, F., Strata, P., 1995. Different climbing fibres innervate separate dendritic regions of the same Purkinje cell in hypogranular cerebellum. J. Comp. Neurol. 357, 395e407. Brummendorf, T., Kenwrick, S., Rathjen, F.G., 1998. Neural cell recognition molecule L1: from cell biology to human hereditary brain malformations. Curr. Opin. Neurobiol. 8, 87e97. Buttermore, E.D., Piochon, C., Wallace, M.L., Philpot, B.D., Hansel, C., Bhat, M.A., 2012. Pinceau organization in the cerebellum requires distinct functions of neurofascin in Purkinje and basket neurons during postnatal development. J. Neurosci. 32, 4724e4742. Cajal, S.R., 1911. Histologie du systeme nerveux de l’homme et des vertebres, vol. II (Paris: Maloine). Carrillo, J., Nishiyama, N., Nishiyama, H., 2013. Dendritic translocation establishes the winner in cerebellar climbing fiber synapse elimination. J. Neurosci. 33, 7641e7653. Cesa, R., Scelfo, B., Strata, P., 2007. Activity-dependent presynaptic and postsynaptic structural plasticity in the mature cerebellum. J. Neurosci. 27, 4603e4611. Chattopadhyaya, B., Di Cristo, G., Wu, C.Z., Knott, G., Kuhlman, S., Fu, Y., Palmiter, R.D., Huang, Z.J., 2007. GAD67-mediated GABA synthesis and signaling regulate inhibitory synaptic innervation in the visual cortex. Neuron 54, 889e903. Chedotal, A., Sotelo, C., 1993. The ’creeper stage’ in cerebellar climbing fiber synaptogenesis precedes the ’pericellular nest’eultrastructural evidence with parvalbumin immunocytochemistry. Brain Res. Dev. Brain Res. 76, 207e220. Chen, S., Hillman, D.E., 1982. Plasticity of the parallel fiber-Purkinje cell synapse by spine takeover and new synapse formation in the adult rat. Brain Res. 240, 205e220. Choo, M., Miyazaki, T., Yamazaki, M., Kawamura, M., Nakazawa, T., Zhang, J., Tanimura, A., Uesaka, N., Watanabe, M., Sakimura, K., Kano, M., 2017. Retrograde BDNF to TrkB signaling promotes synapse elimination in the developing cerebellum. Nat. Commun. 8, 195. Cioni, J.M., Telley, L., Saywell, V., Cadilhac, C., Jourdan, C., Huber, A.B., Huang, J.Z., Jahannault-Talignani, C., Ango, F., 2013. SEMA3A signaling controls layer-specific interneuron branching in the cerebellum. Curr. Biol. 23, 850e861. Crepel, F., 1971. Maturation of climbing fiber responses in the rat. Brain Res. 35, 272e276. Crepel, F., Delhaye-Bouchaud, N., Dupont, J.L., 1981. Fate of the multiple innervation of cerebellar Purkinje cells by climbing fibers in immature control, X-irradiated and hypothyroid rats. Brain Res. 227, 59e71. Crepel, F., Mariani, J., Delhaye-Bouchaud, N., 1976. Evidence for a multiple innervation of Purkinje cells by climbing fibers in the immature rat cerebellum. J. Neurobiol. 7, 567e578. Davis, J.Q., Lambert, S., Bennett, V., 1996. Molecular composition of the node of Ranvier: identification of ankyrin-binding cell adhesion molecules neurofascin (mucinþ/third FNIII domain-) and NrCAM at nodal axon segments. J. Cell. Biol. 135, 1355e1367. De Zeeuw, C.I., Hansel, C., Bian, F., Koekkoek, S.K., van Alphen, A.M., Linden, D.J., Oberdick, J., 1998. Expression of a protein kinase C inhibitor in Purkinje cells blocks cerebellar LTD and adaptation of the vestibulo-ocular reflex. Neuron 20, 495e508. De Zeeuw, C.I., Holstege, J.C., Calkoen, F., Ruigrok, T.J., Voogd, J., 1988. A new combination of WGA-HRP anterograde tracing and GABA immunocytochemistry applied to afferents of the cat inferior olive at the ultrastructural level. Brain Res. 447, 369e375. Eccles, J.C., Llinas, R., Sasaki, K., 1966. The excitatory synaptic action of climbing fibres on the Purkinje cells of the cerebellum. J. Physiol. 182, 268e296. Englund, C., Kowalczyk, T., Daza, R.A., Dagan, A., Lau, C., Rose, M.F., Hevner, R.F., 2006. Unipolar brush cells of the cerebellum are produced in the rhombic lip and migrate through developing white matter. J. Neurosci. 26, 9184e9195. Ferraguti, F., Crepaldi, L., Nicoletti, F., 2008. Metabotropic glutamate 1 receptor: current concepts and perspectives. Pharmacol. Rev. 60, 536e581. Finch, E.A., Augustine, G.J., 1998. Local calcium signalling by inositol-1,4,5-trisphosphate in Purkinje cell dendrites. Nature 396, 753e756.

98

PART | I Circuit development

Fink, A.J., Englund, C., Daza, R.A., Pham, D., Lau, C., Nivison, M., Kowalczyk, T., Hevner, R.F., 2006. Development of the deep cerebellar nuclei: transcription factors and cell migration from the rhombic lip. J. Neurosci. 26, 3066e3076. Fritschy, J.M., Panzanelli, P., Kralic, J.E., Vogt, K.E., Sassoe-Pognetto, M., 2006. Differential dependence of axo-dendritic and axo-somatic GABAergic synapses on GABAA receptors containing the a1 subunit in Purkinje cells. J. Neurosci. 26, 3245e3255. Frola, E., Patrizi, A., Goetz, T., Medrihan, L., Petrini, E.M., Barberis, A., Wulff, P., Wisden, W., Sassoe-Pognetto, M., 2013. Synaptic competition sculpts the development of GABAergic axo-dendritic but not perisomatic synapses. PLoS One 8, e56311. Furutani, K., Okubo, Y., Kakizawa, S., Iino, M., 2006. Postsynaptic inositol 1,4,5-trisphosphate signaling maintains presynaptic function of parallel fiberPurkinje cell synapses via BDNF. Proc. Natl. Acad. Sci. USA 103, 8528e8533. Galliano, E., Baratella, M., Sgritta, M., Ruigrok, T.J., Haasdijk, E.D., Hoebeek, F.E., D’Angelo, E., Jaarsma, D., De Zeeuw, C.I., 2013. Anatomical investigation of potential contacts between climbing fibers and cerebellar Golgi cells in the mouse. Front. Neural Circuits 7, 59. Good, J.M., Mahoney, M., Miyazaki, T., Tanaka, K.F., Sakimura, K., Watanabe, M., Kitamura, K., Kano, M., 2017. Maturation of cerebellar Purkinje cell population activity during postnatal refinement of climbing fiber network. Cell Rep. 21, 2066e2073. Guastavino, J.M., Sotelo, C., Damez-Kinselle, I., 1990. Hot-foot murine mutation: behavioral effects and neuroanatomical alterations. Brain Res. 523, 199e210. Hansel, C., de Jeu, M., Belmeguenai, A., Houtman, S.H., Buitendijk, G.H., Andreev, D., De Zeeuw, C.I., Elgersma, Y., 2006. aCaMKII is essential for cerebellar LTD and motor learning. Neuron 51, 835e843. Hashimoto, K., Ichikawa, R., Kitamura, K., Watanabe, M., Kano, M., 2009. Translocation of a "winner" climbing fiber to the Purkinje cell dendrite and subsequent elimination of "losers" from the soma in developing cerebellum. Neuron 63, 106e118. Hashimoto, K., Ichikawa, R., Takechi, H., Inoue, Y., Aiba, A., Sakimura, K., Mishina, M., Hashikawa, T., Konnerth, A., Watanabe, M., Kano, M., 2001. Roles of glutamate receptor d2 subunit (GluRd2) and metabotropic glutamate receptor subtype 1 (mGluR1) in climbing fiber synapse elimination during postnatal cerebellar development. J. Neurosci. 21, 9701e9712. Hashimoto, K., Kano, M., 2003. Functional differentiation of multiple climbing fiber inputs during synapse elimination in the developing cerebellum. Neuron 38, 785e796. Hashimoto, K., Kano, M., 2005. Postnatal development and synapse elimination of climbing fiber to Purkinje cell projection in the cerebellum. Neurosci. Res. 53, 221e228. Hashimoto, K., Tsujita, M., Miyazaki, T., Kitamura, K., Yamazaki, M., Shin, H.S., Watanabe, M., Sakimura, K., Kano, M., 2011. Postsynaptic P/Q-type Ca2þ channel in Purkinje cell mediates synaptic competition and elimination in developing cerebellum. Proc. Natl. Acad. Sci. USA 108, 9987e9992. Hashimoto, K., Watanabe, M., Kurihara, H., Offermanns, S., Jiang, H., Wu, Y., Jun, K., Shin, H.S., Inoue, Y., Wu, D., et al., 2000. Climbing fiber synapse elimination during postnatal cerebellar development requires signal transduction involving Gaq and phospholipase C b4. Prog. Brain Res. 124, 31e48. Hashimoto, K., Yoshida, T., Sakimura, K., Mishina, M., Watanabe, M., Kano, M., 2009. Influence of parallel fiber-Purkinje cell synapse formation on postnatal development of climbing fiber-Purkinje cell synapses in the cerebellum. Neuroscience 162, 601e611. Hashizume, M., Miyazaki, T., Sakimura, K., Watanabe, M., Kitamura, K., Kano, M., 2013. Disruption of cerebellar microzonal organization in GluD2 (GluRd2) knockout mouse. Front. Neural Circuits 7, 130. Hawkes, R., Leclerc, N., 1987. Antigenic map of the rat cerebellar cortex: the distribution of parasagittal bands as revealed by monoclonal anti-Purkinje cell antibody mabQ113. J. Comp. Neurol. 256, 29e41. Hirai, H., Pang, Z., Bao, D., Miyazaki, T., Li, L., Miura, E., Parris, J., Rong, Y., Watanabe, M., Yuzaki, M., Morgan, J.I., 2005. Cbln1 is essential for synaptic integrity and plasticity in the cerebellum. Nat. Neurosci. 8, 1534e1541. Hirasawa, M., Xu, X., Trask, R.B., Maddatu, T.P., Johnson, B.A., Naggert, J.K., Nishina, P.M., Ikeda, A., 2007. Carbonic anhydrase related protein 8 mutation results in aberrant synaptic morphology and excitatory synaptic function in the cerebellum. Mol. Cell Neurosci. 35, 161e170. Hirota, J., Ando, H., Hamada, K., Mikoshiba, K., 2003. Carbonic anhydrase-related protein is a novel binding protein for inositol 1,4,5-trisphosphate receptor type 1. Biochem. J. 372, 435e441. Hoshino, M., Nakamura, S., Mori, K., Kawauchi, T., Terao, M., Nishimura, Y.V., Fukuda, A., Fuse, T., Matsuo, N., Sone, M., et al., 2005. Ptf1a, a bHLH transcriptional gene, defines GABAergic neuronal fates in cerebellum. Neuron 47, 201e213. Huang, Z.J., 2009. Activity-dependent development of inhibitory synapses and innervation pattern: role of GABA signalling and beyond. J. Physiol. 587, 1881e1888. Huang, Z.J., Di Cristo, G., Ango, F., 2007. Development of GABA innervation in the cerebral and cerebellar cortices. Nat. Rev. Neurosci. 8, 673e686. Huang, Z.J., Scheiffele, P., 2008. GABA and neuroligin signaling: linking synaptic activity and adhesion in inhibitory synapse development. Curr. Opin. Neurobiol. 18, 77e83. Ichikawa, R., Hashimoto, K., Miyazaki, T., Uchigashima, M., Yamasaki, M., Aiba, A., Kano, M., Watanabe, M., 2016. Territories of heterologous inputs onto Purkinje cell dendrites are segregated by mGluR1-dependent parallel fiber synapse elimination. Proc. Natl. Acad. Sci. USA 113, 2282e2287. Ichikawa, R., Miyazaki, T., Kano, M., Hashikawa, T., Tatsumi, H., Sakimura, K., Mishina, M., Inoue, Y., Watanabe, M., 2002. Distal extension of climbing fiber territory and multiple innervation caused by aberrant wiring to adjacent spiny branchlets in cerebellar Purkinje cells lacking glutamate receptor d2. J. Neurosci. 22, 8487e8503. Ichikawa, R., Sakimura, K., Watanabe, M., 2016. GluD2 endows parallel fiber-Purkinje cell synapses with a high regenerative capacity. J. Neurosci. 36, 4846e4858. Ichise, T., Kano, M., Hashimoto, K., Yanagihara, D., Nakao, K., Shigemoto, R., Katsuki, M., Aiba, A., 2000. mGluR1 in cerebellar Purkinje cells essential for long-term depression, synapse elimination, and motor coordination. Science 288, 1832e1835.

Cerebellar circuits Chapter | 4

99

Iijima, T., Miura, E., Matsuda, K., Kamekawa, Y., Watanabe, M., Yuzaki, M., 2007. Characterization of a transneuronal cytokine family Cblneregulation of secretion by heteromeric assembly. Eur. J. Neurosci. 25, 1049e1057. Ito-Ishida, A., Miura, E., Emi, K., Matsuda, K., Iijima, T., Kondo, T., Kohda, K., Watanabe, M., Yuzaki, M., 2008. Cbln1 regulates rapid formation and maintenance of excitatory synapses in mature cerebellar Purkinje cells in vitro and in vivo. J. Neurosci. 28, 5920e5930. Ito, M., 1984. The cerebellum and neural control. Raven Press, New York. Iwakura, A., Uchigashima, M., Miyazaki, T., Yamasaki, M., Watanabe, M., 2012. Lack of molecular-anatomical evidence for GABAergic influence on axon initial segment of cerebellar Purkinje cells by the pinceau formation. J. Neurosci. 32, 9438e9448. Johnson, E.M., Craig, E.T., Yeh, H.H., 2007. TrkB is necessary for pruning at the climbing fibre-Purkinje cell synapse in the developing murine cerebellum. J. Physiol. 582, 629e646. Juttner, R., Montag, D., Craveiro, R.B., Babich, A., Vetter, P., Rathjen, F.G., 2013. Impaired presynaptic function and elimination of synapses at premature stages during postnatal development of the cerebellum in the absence of CALEB (CSPG5/neuroglycan C). Eur. J. Neurosci. 38, 3270e3280. Kakegawa, W., Kohda, K., Yuzaki, M., 2007. The d2 ’ionotropic’ glutamate receptor functions as a non-ionotropic receptor to control cerebellar synaptic plasticity. J. Physiol. 584, 89e96. Kakegawa, W., Mitakidis, N., Miura, E., Abe, M., Matsuda, K., Takeo, Y.H., Kohda, K., Motohashi, J., Takahashi, A., Nagao, S., et al., 2015. Anterograde C1ql1 signaling is required in order to determine and maintain a single-winner climbing fiber in the mouse cerebellum. Neuron 85, 316e329. Kakegawa, W., Miyazaki, T., Emi, K., Matsuda, K., Kohda, K., Motohashi, J., Mishina, M., Kawahara, S., Watanabe, M., Yuzaki, M., 2008. Differential regulation of synaptic plasticity and cerebellar motor learning by the C-terminal PDZ-binding motif of GluRd2. J. Neurosci. 28, 1460e1468. Kakegawa, W., Miyazaki, T., Hirai, H., Motohashi, J., Mishina, M., Watanabe, M., Yuzaki, M., 2007. Ca2þ permeability of the channel pore is not essential for the d2 glutamate receptor to regulate synaptic plasticity and motor coordination. J. Physiol. 579, 729e735. Kakegawa, W., Miyazaki, T., Kohda, K., Matsuda, K., Emi, K., Motohashi, J., Watanabe, M., Yuzaki, M., 2009. The N-terminal domain of GluD2 (GluRd2) recruits presynaptic terminals and regulates synaptogenesis in the cerebellum in vivo. J. Neurosci. 29, 5738e5748. Kakizawa, S., Miyazaki, T., Yanagihara, D., Iino, M., Watanabe, M., Kano, M., 2005. Maintenance of presynaptic function by AMPA receptor-mediated excitatory postsynaptic activity in adult brain. Proc. Natl. Acad. Sci. USA 102, 19180e19185. Kakizawa, S., Yamada, K., Iino, M., Watanabe, M., Kano, M., 2003. Effects of insulin-like growth factor I on climbing fibre synapse elimination during cerebellar development. Eur. J. Neurosci. 17, 545e554. Kakizawa, S., Yamasaki, M., Watanabe, M., Kano, M., 2000. Critical period for activity-dependent synapse elimination in developing cerebellum. J. Neurosci. 20, 4954e4961. Kano, M., Hashimoto, K., Chen, C., Abeliovich, A., Aiba, A., Kurihara, H., Watanabe, M., Inoue, Y., Tonegawa, S., 1995. Impaired synapse elimination during cerebellar development in PKCg mutant mice. Cell 83, 1223e1231. Kano, M., Hashimoto, K., Kurihara, H., Watanabe, M., Inoue, Y., Aiba, A., Tonegawa, S., 1997. Persistent multiple climbing fiber innervation of cerebellar Purkinje cells in mice lacking mGluR1. Neuron 18, 71e79. Kano, M., Hashimoto, K., Watanabe, M., Kurihara, H., Offermanns, S., Jiang, H., Wu, Y., Jun, K., Shin, H.S., Inoue, Y., et al., 1998. Phospholipase Cb4 is specifically involved in climbing fiber synapse elimination in the developing cerebellum. Proc. Natl. Acad. Sci. USA 95, 15724e15729. Kashiwabuchi, N., Ikeda, K., Araki, K., Hirano, T., Shibuki, K., Takayama, C., Inoue, Y., Kutsuwada, T., Yagi, T., Kang, Y., et al., 1995. Impairment of motor coordination, Purkinje cell synapse formation, and cerebellar long-term depression in GluRd2 mutant mice. Cell 81, 245e252. Kawamura, Y., Nakayama, H., Hashimoto, K., Sakimura, K., Kitamura, K., Kano, M., 2013. Spike timing-dependent selective strengthening of single climbing fibre inputs to Purkinje cells during cerebellar development. Nat. Commun. 4, 2732. Kawata, S., Miyazaki, T., Yamazaki, M., Mikuni, T., Yamasaki, M., Hashimoto, K., Watanabe, M., Sakimura, K., Kano, M., 2014. Global scaling down of excitatory postsynaptic responses in cerebellar Purkinje cells impairs developmental synapse elimination. Cell Rep. 8, 1119e1129. Kishimoto, Y., Kawahara, S., Fujimichi, R., Mori, H., Mishina, M., Kirino, Y., 2001. Impairment of eyeblink conditioning in GluRd2-mutant mice depends on the temporal overlap between conditioned and unconditioned stimuli. Eur. J. Neurosci. 14, 1515e1521. Kulik, A., Nakadate, K., Hagiwara, A., Fukazawa, Y., Lujan, R., Saito, H., Suzuki, N., Futatsugi, A., Mikoshiba, K., Frotscher, M., Shigemoto, R., 2004. Immunocytochemical localization of the a1A subunit of the P/Q-type calcium channel in the rat cerebellum. Eur. J. Neurosci. 19, 2169e2178. Kurihara, H., Hashimoto, K., Kano, M., Takayama, C., Sakimura, K., Mishina, M., Inoue, Y., Watanabe, M., 1997. Impaired parallel fiber–>Purkinje cell synapse stabilization during cerebellar development of mutant mice lacking the glutamate receptor d2 subunit. J. Neurosci. 17, 9613e9623. Laine, J., Axelrad, H., 2002. Extending the cerebellar Lugaro cell class. Neuroscience 115, 363e374. Landsend, A.S., Amiry-Moghaddam, M., Matsubara, A., Bergersen, L., Usami, S., Wenthold, R.J., Ottersen, O.P., 1997. Differential localization of d glutamate receptors in the rat cerebellum: coexpression with AMPA receptors in parallel fiber-spine synapses and absence from climbing fiber-spine synapses. J. Neurosci. 17, 834e842. Laube, G., Roper, J., Pitt, J.C., Sewing, S., Kistner, U., Garner, C.C., Pongs, O., Veh, R.W., 1996. Ultrastructural localization of shaker-related potassium channel subunits and synapse-associated protein 90 to septate-like junctions in rat cerebellar pinceaux. Brain Res. Mol. Brain Res. 42, 51e61. Letellier, M., Wehrle, R., Mariani, J., Lohof, A.M., 2009. Synapse elimination in olivo-cerebellar explants occurs during a critical period and leaves an indelible trace in Purkinje cells. Proc. Natl. Acad. Sci. USA 106, 14102e14107. Leto, K., Bartolini, A., Yanagawa, Y., Obata, K., Magrassi, L., Schilling, K., Rossi, F., 2009. Laminar fate and phenotype specification of cerebellar GABAergic interneurons. J. Neurosci. 29, 7079e7091. Levenes, C., Daniel, H., Jaillard, D., Conquet, F., Crepel, F., 1997. Incomplete regression of multiple climbing fibre innervation of cerebellar Purkinje cells in mGluR1 mutant mice. Neuroreport 8, 571e574.

100 PART | I Circuit development

Li, R.W., Yu, W., Christie, S., Miralles, C.P., Bai, J., Loturco, J.J., De Blas, A.L., 2005. Disruption of postsynaptic GABA receptor clusters leads to decreased GABAergic innervation of pyramidal neurons. J. Neurochem. 95, 756e770. Llinas, R., Baker, R., Sotelo, C., 1974. Electrotonic coupling between neurons in cat inferior olive. J. Neurophysiol. 37, 560e571. Llinas, R., Sasaki, K., 1989. The functional organization of the olivo-cerebellar system as examined by multiple Purkinje cell recordings. Eur. J. Neurosci. 1, 587e602. Lomeli, H., Sprengel, R., Laurie, D.J., Kohr, G., Herb, A., Seeburg, P.H., Wisden, W., 1993. The rat d-1 and d-2 subunits extend the excitatory amino acid receptor family. FEBS Lett. 315, 318e322. Lorenzetto, E., Caselli, L., Feng, G., Yuan, W., Nerbonne, J.M., Sanes, J.R., Buffelli, M., 2009. Genetic perturbation of postsynaptic activity regulates synapse elimination in developing cerebellum. Proc. Natl. Acad. Sci. USA 106, 16475e16480. Mariani, J., Changeux, J.P., 1981. Ontogenesis of olivocerebellar relationships. I. Studies by intracellular recordings of the multiple innervation of Purkinje cells by climbing fibers in the developing rat cerebellum. J. Neurosci. 1, 696e702. Marzban, H., Chung, S., Watanabe, M., Hawkes, R., 2007. Phospholipase Cb4 expression reveals the continuity of cerebellar topography through development. J. Comp. Neurol. 502, 857e871. Matsuda, K., Miura, E., Miyazaki, T., Kakegawa, W., Emi, K., Narumi, S., Fukazawa, Y., Ito-Ishida, A., Kondo, T., Shigemoto, R., et al., 2010. Cbln1 is a ligand for an orphan glutamate receptor d2, a bidirectional synapse organizer. Science 328, 363e368. Miale, I.L., Sidman, R.L., 1961. An autoradiographic analysis of histogenesis in the mouse cerebellum. Exp. Neurol. 4, 277e296. Mikuni, T., Uesaka, N., Okuno, H., Hirai, H., Deisseroth, K., Bito, H., Kano, M., 2013. Arc/Arg3.1 is a postsynaptic mediator of activity-dependent synapse elimination in the developing cerebellum. Neuron 78, 1024e1035. Mintz, I.M., Adams, M.E., Bean, B.P., 1992. P-type calcium channels in rat central and peripheral neurons. Neuron 9, 85e95. Miura, E., Iijima, T., Yuzaki, M., Watanabe, M., 2006. Distinct expression of Cbln family mRNAs in developing and adult mouse brains. Eur. J. Neurosci. 24, 750e760. Miura, E., Matsuda, K., Morgan, J.I., Yuzaki, M., Watanabe, M., 2009. Cbln1 accumulates and colocalizes with Cbln3 and GluRd2 at parallel fiberPurkinje cell synapses in the mouse cerebellum. Eur. J. Neurosci. 29, 693e706. Miyakawa, H., Lev-Ram, V., Lasser-Ross, N., Ross, W.N., 1992. Calcium transients evoked by climbing fiber and parallel fiber synaptic inputs in Guinea pig cerebellar Purkinje neurons. J. Neurophysiol. 68, 1178e1189. Miyazaki, T., Hashimoto, K., Shin, H.S., Kano, M., Watanabe, M., 2004. P/Q-type Ca2þ channel a1A regulates synaptic competition on developing cerebellar Purkinje cells. J. Neurosci. 24, 1734e1743. Miyazaki, T., Hashimoto, K., Uda, A., Sakagami, H., Nakamura, Y., Saito, S.Y., Nishi, M., Kume, H., Tohgo, A., Kaneko, I., et al., 2006. Disturbance of cerebellar synaptic maturation in mutant mice lacking BSRPs, a novel brain-specific receptor-like protein family. FEBS Lett. 580, 4057e4064. Miyazaki, T., Yamasaki, M., Hashimoto, K., Kohda, K., Yuzaki, M., Shimamoto, K., Tanaka, K., Kano, M., Watanabe, M., 2017. Glutamate transporter GLAST controls synaptic wrapping by Bergmann glia and ensures proper wiring of Purkinje cells. Proc. Natl. Acad. Sci. USA 114, 7438e7443. Miyazaki, T., Yamasaki, M., Hashimoto, K., Yamazaki, M., Abe, M., Usui, H., Kano, M., Sakimura, K., Watanabe, M., 2012. Cav2.1 in cerebellar Purkinje cells regulates competitive excitatory synaptic wiring, cell survival, and cerebellar biochemical compartmentalization. J. Neurosci. 32, 1311e1328. Miyazaki, T., Yamasaki, M., Takeuchi, T., Sakimura, K., Mishina, M., Watanabe, M., 2010. Ablation of glutamate receptor GluRd2 in adult Purkinje cells causes multiple innervation of climbing fibers by inducing aberrant invasion to parallel fiber innervation territory. J. Neurosci. 30, 15196e15209. Morando, L., Cesa, R., Rasetti, R., Harvey, R., Strata, P., 2001. Role of glutamate d-2 receptors in activity-dependent competition between heterologous afferent fibers. Proc. Natl. Acad. Sci. USA 98, 9954e9959. Mugnaini, E., Floris, A., 1994. The unipolar brush cell: a neglected neuron of the mammalian cerebellar cortex. J. Comp. Neurol. 339, 174e180. Mukamel, E.A., Nimmerjahn, A., Schnitzer, M.J., 2009. Automated analysis of cellular signals from large-scale calcium imaging data. Neuron 63, 747e760. Nakamura, M., Sato, K., Fukaya, M., Araishi, K., Aiba, A., Kano, M., Watanabe, M., 2004. Signaling complex formation of phospholipase Cb4 with metabotropic glutamate receptor type 1a and 1,4,5-trisphosphate receptor at the perisynapse and endoplasmic reticulum in the mouse brain. Eur. J. Neurosci. 20, 2929e2944. Nakayama, H., Abe, M., Morimoto, C., Iida, T., Okabe, S., Sakimura, K., Hashimoto, K., 2018. Microglia permit climbing fiber elimination by promoting GABAergic inhibition in the developing cerebellum. Nat. Commun. 9, 2830. Nakayama, H., Miyazaki, T., Kitamura, K., Hashimoto, K., Yanagawa, Y., Obata, K., Sakimura, K., Watanabe, M., Kano, M., 2012. GABAergic inhibition regulates developmental synapse elimination in the cerebellum. Neuron 74, 384e396. Napper, R.M., Harvey, R.J., 1988. Number of parallel fiber synapses on an individual Purkinje cell in the cerebellum of the rat. J. Comp. Neurol. 274, 168e177. Nishiyama, H., Fukaya, M., Watanabe, M., Linden, D.J., 2007. Axonal motility and its modulation by activity are branch-type specific in the intact adult cerebellum. Neuron 56, 472e487. Ohtani, Y., Miyata, M., Hashimoto, K., Tabata, T., Kishimoto, Y., Fukaya, M., Kase, D., Kassai, H., Nakao, K., Hirata, T., et al., 2014. The synaptic targeting of mGluR1 by its carboxyl-terminal domain is crucial for cerebellar function. J. Neurosci. 34, 2702e2712. Oostland, M., Buijink, M.R., van Hooft, J.A., 2013. Serotonergic control of Purkinje cell maturation and climbing fibre elimination by 5-HT3 receptors in the juvenile mouse cerebellum. J. Physiol. 591, 1793e1807. Ottersen, O.P., Storm-Mathisen, J., Somogyi, P., 1988. Colocalization of glycine-like and GABA-like immunoreactivities in Golgi cell terminals in the rat cerebellum: a postembedding light and electron microscopic study. Brain Res. 450, 342e353.

Cerebellar circuits Chapter | 4

101

Palay, S.L., Chan-Palay, V., 1974. Cerebellar Cortex. Springer-Verlag, New York. Pang, Z., Zuo, J., Morgan, J.I., 2000. Cbln3, a novel member of the precerebellin family that binds specifically to Cbln1. J. Neurosci. 20, 6333e6339. Patrizi, A., Scelfo, B., Viltono, L., Briatore, F., Fukaya, M., Watanabe, M., Strata, P., Varoqueaux, F., Brose, N., Fritschy, J.M., Sassoe-Pognetto, M., 2008. Synapse formation and clustering of neuroligin-2 in the absence of GABAA receptors. Proc. Natl. Acad. Sci. USA 105, 13151e13156. Pijpers, A., Voogd, J., Ruigrok, T.J., 2005. Topography of olivo-cortico-nuclear modules in the intermediate cerebellum of the rat. J. Comp. Neurol. 492, 193e213. Rabacchi, S., Bailly, Y., Delhaye-Bouchaud, N., Mariani, J., 1992. Involvement of the N-methyl D-aspartate (NMDA) receptor in synapse elimination during cerebellar development. Science 256, 1823e1825. Rakic, P., 1971. Neuron-glia relationship during granule cell migration in developing cerebellar cortex. A Golgi and electronmicroscopic study in Macacus Rhesus. J. Comp. Neurol. 141, 283e312. Rathjen, F.G., Wolff, J.M., Chang, S., Bonhoeffer, F., Raper, J.A., 1987. Neurofascin: a novel chick cell-surface glycoprotein involved in neurite-neurite interactions. Cell 51, 841e849. Ribar, T.J., Rodriguiz, R.M., Khiroug, L., Wetsel, W.C., Augustine, G.J., Means, A.R., 2000. Cerebellar defects in Ca2þ/calmodulin kinase IV-deficient mice. J. Neurosci. 20, RC107. Roth, A., Hausser, M., 2001. Compartmental models of rat cerebellar Purkinje cells based on simultaneous somatic and dendritic patch-clamp recordings. J. Physiol. 535, 445e472. Schafer, D.P., Lehrman, E.K., Kautzman, A.G., Koyama, R., Mardinly, A.R., Yamasaki, R., Ransohoff, R.M., Greenberg, M.E., Barres, B.A., Stevens, B., 2012. Microglia sculpt postnatal neural circuits in an activity and complement-dependent manner. Neuron 74, 691e705. Schilling, K., Oberdick, J., 2009. The treasury of the commons: making use of public gene expression resources to better characterize the molecular diversity of inhibitory interneurons in the cerebellar cortex. Cerebellum 8, 477e489. Schultz, S.R., Kitamura, K., Post-Uiterweer, A., Krupic, J., Hausser, M., 2009. Spatial pattern coding of sensory information by climbing fiber-evoked calcium signals in networks of neighboring cerebellar Purkinje cells. J. Neurosci. 29, 8005e8015. Seto, Y., Nakatani, T., Masuyama, N., Taya, S., Kumai, M., Minaki, Y., Hamaguchi, A., Inoue, Y.U., Inoue, T., Miyashita, S., et al., 2014. Temporal identity transition from Purkinje cell progenitors to GABAergic interneuron progenitors in the cerebellum. Nat. Commun. 5, 3337. Sherrard, R.M., Dixon, K.J., Bakouche, J., Rodger, J., Lemaigre-Dubreuil, Y., Mariani, J., 2009. Differential expression of TrkB isoforms switches climbing fiber-Purkinje cell synaptogenesis to selective synapse elimination. Dev. Neurobiol. 69, 647e662. Sigoillot, S.M., Iyer, K., Binda, F., Gonzaález-Calvo, I., Talleur, M., Vodjdani, G., Isope, P., Selimi, F., 2015. The secreted protein C1QL1 and its receptor BAI3 control the synaptic connectivity of excitatory inputs converging on cerebellar Purkinje cells. Cell Rep 10, 820e832. Simat, M., Parpan, F., Fritschy, J.M., 2007. Heterogeneity of glycinergic and GABAergic interneurons in the granule cell layer of mouse cerebellum. J. Comp. Neurol. 500, 71e83. Slemmon, J.R., Blacher, R., Danho, W., Hempstead, J.L., Morgan, J.I., 1984. Isolation and sequencing of two cerebellum-specific peptides. Proc. Natl. Acad. Sci. USA 81, 6866e6870. Song, J.W., Misgeld, T., Kang, H., Knecht, S., Lu, J., Cao, Y., Cotman, S.L., Bishop, D.L., Lichtman, J.W., 2008. Lysosomal activity associated with developmental axon pruning. J. Neurosci. 28, 8993e9001. Sotelo, C., 2004. Cellular and genetic regulation of the development of the cerebellar system. Prog. Neurobiol. 72, 295e339. Sotelo, C., Llinas, R., Baker, R., 1974. Structural study of inferior olivary nucleus of the cat: morphological correlates of electrotonic coupling. J. Neurophysiol. 37, 541e559. Spacek, J., 1985. Three-dimensional analysis of dendritic spines. III. Glial sheath. Anat. Embryol. (Berl) 171, 245e252. Stea, A., Tomlinson, W.J., Soong, T.W., Bourinet, E., Dubel, S.J., Vincent, S.R., Snutch, T.P., 1994. Localization and functional properties of a rat brain a1A calcium channel reflect similarities to neuronal Q- and P-type channels. Proc. Natl. Acad. Sci. USA 91, 10576e10580. Sugihara, I., 2005. Microzonal projection and climbing fiber remodeling in single olivocerebellar axons of newborn rats at postnatal days 4-7. J. Comp. Neurol. 487, 93e106. Sugihara, I., Lang, E.J., Llinas, R., 1993. Uniform olivocerebellar conduction time underlies Purkinje cell complex spike synchronicity in the rat cerebellum. J. Physiol. 470, 243e271. Sugihara, I., Shinoda, Y., 2004. Molecular, topographic, and functional organization of the cerebellar cortex: a study with combined aldolase C and olivocerebellar labeling. J. Neurosci. 24, 8771e8785. Sugihara, I., Shinoda, Y., 2007. Molecular, topographic, and functional organization of the cerebellar nuclei: analysis by three-dimensional mapping of the olivonuclear projection and aldolase C labeling. J. Neurosci. 27, 9696e9710. Sugihara, I., Wu, H.S., Shinoda, Y., 2001. The entire trajectories of single olivocerebellar axons in the cerebellar cortex and their contribution to cerebellar compartmentalization. J. Neurosci. 21, 7715e7723. Sultan, F., Bower, J.M., 1998. Quantitative Golgi study of the rat cerebellar molecular layer interneurons using principal component analysis. J. Comp. Neurol. 393, 353e373. Takacs, J., Hamori, J., 1994. Developmental dynamics of Purkinje cells and dendritic spines in rat cerebellar cortex. J. Neurosci. Res. 38, 515e530. Takagishi, Y., Hashimoto, K., Kayahara, T., Watanabe, M., Otsuka, H., Mizoguchi, A., Kano, M., Murata, Y., 2007. Diminished climbing fiber innervation of Purkinje cells in the cerebellum of myosin Va mutant mice and rats. Dev. Neurobiol. 67, 909e923. Takayama, C., Nakagawa, S., Watanabe, M., Mishina, M., Inoue, Y., 1995. Light- and electron-microscopic localization of the glutamate receptor channel d2 subunit in the mouse Purkinje cell. Neurosci. Lett. 188, 89e92. Takechi, H., Eilers, J., Konnerth, A., 1998. A new class of synaptic response involving calcium release in dendritic spines. Nature 396, 757e760.

102 PART | I Circuit development

Takeuchi, T., Miyazaki, T., Watanabe, M., Mori, H., Sakimura, K., Mishina, M., 2005. Control of synaptic connection by glutamate receptor d2 in the adult cerebellum. J. Neurosci. 25, 2146e2156. Telley, L., Cadilhac, C., Cioni, J.M., Saywell, V., Jahannault-Talignani, C., Huettl, R.E., Sarrailh-Faivre, C., Dayer, A., Huber, A.B., Ango, F., 2016. Dual function of NRP1 in axon guidance and subcellular target recognition in cerebellum. Neuron 91, 1276e1291. Tsutsumi, S., Yamazaki, M., Miyazaki, T., Watanabe, M., Sakimura, K., Kano, M., Kitamura, K., 2015. Structure-function relationships between aldolase C/zebrin II expression and complex spike synchrony in the cerebellum. J. Neurosci. 35, 843e852. Uemura, T., Kakizawa, S., Yamasaki, M., Sakimura, K., Watanabe, M., Iino, M., Mishina, M., 2007. Regulation of long-term depression and climbing fiber territory by glutamate receptor d2 at parallel fiber synapses through its C-terminal domain in cerebellar Purkinje cells. J. Neurosci. 27, 12096e12108. Uemura, T., Lee, S.J., Yasumura, M., Takeuchi, T., Yoshida, T., Ra, M., Taguchi, R., Sakimura, K., Mishina, M., 2010. Trans-synaptic interaction of GluRd2 and neurexin through Cbln1 mediates synapse formation in the cerebellum. Cell 141, 1068e1079. Uemura, T., Mishina, M., 2008. The amino-terminal domain of glutamate receptor d2 triggers presynaptic differentiation. Biochem. Biophys. Res. Commun. 377, 1315e1319. Uemura, T., Mori, H., Mishina, M., 2004. Direct interaction of GluRd2 with Shank scaffold proteins in cerebellar Purkinje cells. Mol. Cell Neurosci. 26, 330e341. Uesaka, N., Abe, M., Konno, K., Yamazaki, M., Sakoori, K., Watanabe, T., Kao, T.H., Mikuni, T., Watanabe, M., Sakimura, K., Kano, M., 2018. Retrograde signaling from progranulin to Sort1 counteracts synapse elimination in the developing cerebellum. Neuron 97, 796e805 e795. Uesaka, N., Kano, M., 2018. Presynaptic mechanisms mediating retrograde semaphorin signals for climbing fiber synapse elimination during postnatal cerebellar development. Cerebellum 17, 17e22. Uesaka, N., Uchigashima, M., Mikuni, T., Nakazawa, T., Nakao, H., Hirai, H., Aiba, A., Watanabe, M., Kano, M., 2014. Retrograde semaphorin signaling regulates synapse elimination in the developing mouse brain. Science 344, 1020e1023. Urade, Y., Oberdick, J., Molinar-Rode, R., Morgan, J.I., 1991. Precerebellin is a cerebellum-specific protein with similarity to the globular domain of complement C1q B chain. Proc. Natl. Acad. Sci. USA 88, 1069e1073. Voogd, J., Ruigrok, T.J., 2004. The organization of the corticonuclear and olivocerebellar climbing fiber projections to the rat cerebellar vermis: the congruence of projection zones and the zebrin pattern. Neurocytol. 33, 5e21. Wassef, M., Chedotal, A., Cholley, B., Thomasset, M., Heizmann, C.W., Sotelo, C., 1992. Development of the olivocerebellar projection in the rat: I. Transient biochemical compartmentation of the inferior olive. J. Comp. Neurol. 323, 519e536. Wassef, M., Zanetta, J.P., Brehier, A., Sotelo, C., 1985. Transient biochemical compartmentalization of Purkinje cells during early cerebellar development. Dev. Biol. 111, 129e137. Watanabe, M., Nakamura, M., Sato, K., Kano, M., Simon, M.I., Inoue, Y., 1998. Patterns of expression for the mRNA corresponding to the four isoforms of phospholipase Cb in mouse brain. Eur. J. Neurosci. 10, 2016e2025. Watase, K., Hashimoto, K., Kano, M., Yamada, K., Watanabe, M., Inoue, Y., Okuyama, S., Sakagawa, T., Ogawa, S., Kawashima, N., et al., 1998. Motor discoordination and increased susceptibility to cerebellar injury in GLAST mutant mice. Eur. J. Neurosci. 10, 976e988. Williams, M.E., de Wit, J., Ghosh, A., 2010. Molecular mechanisms of synaptic specificity in developing neural circuits. Neuron 68, 9e18. Yamada, K., Fukaya, M., Shibata, T., Kurihara, H., Tanaka, K., Inoue, Y., Watanabe, M., 2000. Dynamic transformation of Bergmann glial fibers proceeds in correlation with dendritic outgrowth and synapse formation of cerebellar Purkinje cells. J. Comp. Neurol. 418, 106e120. Yamada, M., Seto, Y., Taya, S., Owa, T., Inoue, Y.U., Inoue, T., Kawaguchi, Y., Nabeshima, Y., Hoshino, M., 2014. Specification of spatial identities of cerebellar neuron progenitors by ptf1a and atoh1 for proper production of GABAergic and glutamatergic neurons. J. Neurosci. 34, 4786e4800. Yamanaka, H., Yanagawa, Y., Obata, K., 2004. Development of stellate and basket cells and their apoptosis in mouse cerebellar cortex. Neurosci. Res. 50, 13e22. Yuzaki, M., 2004. The d2 glutamate receptor: a key molecule controlling synaptic plasticity and structure in Purkinje cells. Cerebellum 3, 89e93. Zhang, L., Goldman, J.E., 1996. Generation of cerebellar interneurons from dividing progenitors in white matter. Neuron 16, 47e54. Zonta, B., Desmazieres, A., Rinaldi, A., Tait, S., Sherman, D.L., Nolan, M.F., Brophy, P.J., 2011. A critical role for neurofascin in regulating action potential initiation through maintenance of the axon initial segment. Neuron 69, 945e956.

Chapter 5

Cortical columns Zolta´n Molna´r1 and Kathleen S. Rockland2 1

Department of Physiology, Anatomy and Genetics, University of Oxford, Oxford, United Kingdom; 2Department of Anatomy and Neurobiology,

Boston University School of Medicine, Boston, MA, United States

Chapter outline 5.1. The loose and uncritical use of the term in ways that are so generalized as to be unhelpful and even confusing 107 5.2. A lack of a universal presence of certain columns within cortical areas, brains, and species is undermining the idea that similar building blocks comprise all cortical circuits 108 5.2.1. The use of physiological methods to reveal columns 108 5.2.2. Columnar organization of some afferent and efferent projections 108 5.2.2.1. Modules of visual cortex 110 5.2.2.2. Ocular dominance columns/stripes 110 5.2.2.3. Orientation columns 110 5.2.3. Gene expression in the cortex in “columnar” fashion110 5.2.3.1. A system of interleaving modules in rodent layer II 110 5.2.3.2. Overlap between columnar entities within the same structures; combining physiological and anatomical definitions 111 5.3. General concept that the cortical column (even just an arbitrary unit column that includes the full depth of the cortex) has a universal constant number of neurons associated with it 112 5.3.1. Number of neurons in a cortical column 112 5.4. Lack of correlation between the absence or presence of particular columns and a specific sensory or cognitive processing network (comparisons across the same brain and across close and more distant species) 112

5.4.1. Microscopic and macroscopic cell patterning defining cortical modules 112 5.4.2. Are barrels cortical columns? 113 5.4.2.1. A system of interleaving modules in rodent layer VI 114 5.4.2.2. Function of barrels 115 5.4.2.3. Microcolumns and apical dendritic bundles 115 5.4.3. Complex relationship relations between minicolumns and dendritic bundles 117 5.4.3.1. Columns outside the mammalian isocortex 117 5.4.3.2. Columns in nonmammals 118 5.4.4. What is the function of a cortical column? 118 5.4.5. Columns in neuropathology 118 5.5. What is the correlation between the columnar development of the brain and future columns 119 5.5.1. Cortical columns during development 119 5.5.1.1. Ontogenic units/columnsdthe fundamental building blocks in the developing neocortex119 5.5.1.2. Sibling neuron circuits in the developing columns 120 5.5.2. Transient columnar domains during development 121 5.5.3. The way forward 122 Acknowledgments 123 References 123

Columns are ubiquitous in the brain but in no way obligatory, and comprehensive descriptions of the various specific forms of “columns” in the brain are still in progress. Up to the present, comparative studies have failed to identify a specific sensory, motor, or cognitive function that is specifically associated with the presence or absence of a particular form of cortical column. The developing cortex contains numerous radial determinants. Pyramidal neurons are generated in “ontogenic units” and subsequently disperse radially. It has been shown that sibling cells have a stronger tendency to establish synaptic connections with each other in the cortical plate. However, the link between the embryonic and adult columnar constellations is currently not known. There are, therefore, several problems about the term and concept of cortical column. (1) The first is the often loose, very general, and uncritical use of the term that can be confusing. (2) The second is related to the lack of a universal

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00005-6 Copyright © 2020 Elsevier Inc. All rights reserved.

103

104 PART | I Circuit development

presence of certain types of column within some cortical areas, brains or species that undermines the idea that similar building blocks comprise all cortical circuits. (3) The concept that the cortical column (or even just an arbitrary columnar unit along the depth of the cortex) has a universally constant number of neurons associated with it with only the primate visual cortex showing a difference. (4) The fourth is related to the lack of correlation between the absence or presence of particular types of column and specific mode of sensory or cognitive processing capacities (across the same brain or across close or more distant species). (5) Although there is evidence for the overall radial disposition of the pyramidal neuron clones and a higher probability of synapse formation between sibling cells, there is a lack of correlation between the columnar development of the brain and columns in the adult. Knowledge of the laminar and columnar organization of the cerebral cortex is continuously advancing, and with this, the conceptual details of the columnar organization are also changing. The time may arrive when both the concept and the nomenclature will also have to adapt to these changes. The hypothesis of the column as the fundamental processing unit of the cerebral cortex was formulated by Mountcastle (1957) from studies of cells responding to a single modality of tactile stimuli (cutaneous or deep joint receptors) in the somatosensory cortex of the cat. From Mountcastle’s work, the concept emerged and further developed over five decades claiming that the cerebral cortex can be further subdivided into “complex processing and distributing units that link a number of inputs to a number of outputs via overlapping internal processing chains” (Mountcastle, 1957). By exploring the physiological, anatomical, genetic, and developmental properties of the cerebral cortex, more details of its organization have been revealed, and many of these new entities were referred to as “column.” The emerging concept in cerebellar circuits by Eccles et al. (1967) fueled the quest for a fundamental cortical processing unit, for an archetypical cortical column, intensified in the hope of identifying modules that are general for all cortical areas. There are references to functional columns, minicolumns, hypercolumns, ontogenetic or embryonic columns, ocular dominance columns, orientation columns, and barrel columns. The only common theme linking these terms is that they refer to a structural, physiological, or developmental organization that transcends the laminar pattern and is perpendicular to it. None of these several types of column are general to all cortical areas, and several are restricted to primary sensory areas. Table 5.1 gives three definitions, and Table 5.2 gives a list of some of the terms that refer to cortical columnar structures. There are so

TABLE 5.1 Examples for definitions of cortical columns. 1. Mountcastle (1997) “The modular organization of nervous systems is a widely documented principle of design for both vertebrate and invertebrate brains of which the columnar organization of the neocortex is an example. The classical cytoarchitectural areas of the neocortex are composed of smaller units, local neural circuits repeated iteratively within each area. Modules may vary in cell type and number, in internal and external connectivity, and in mode of neuronal processing between different large entities; within any single large entity they have a basic similarity of internal design and operation. Modules are most commonly grouped into entities by sets of dominating external connections. This unifying factor is most obvious for the heterotypical sensory and motor areas of the neocortex. Columnar defining factors in homotypical areas are generated, in part, within the cortex itself. The set of all modules composing such an entity may be fractionated into different modular subsets by different extrinsic connections. Linkages between them and subsets in other large entities form distributed systems. The neighborhood relations between connected subsets of modules in different entities result in nested distributed systems that serve distributed functions. A cortical area defined in classical cytoarchitectural terms may belong to more than one and sometimes to several distributed systems. Columns in cytoarchitectural areas located at some distance from one another, but with some common properties, may be linked by long-range, intracortical connections.” 2. http://en.wikipedia.org/wiki/Cortical_minicolumn “A cortical column, also called hypercolumn or sometimes cortical module,a is a group of neurons in the braincortex which can be successively penetrated by a probe inserted perpendicularly to the cortical surface, and which have nearly identical receptive fields. Neurons within a minicolumn encode similar features, whereas a hypercolumn “denotes a unit containing a full set of values for any given set of receptive field parameters.”b A cortical module is defined as either synonymous with a hypercolumn (Mountcastle) or as a tissue block of multiple overlapping hypercolumns (Hubel&Wiesel). Various estimates suggest there are 50e100 cortical minicolumns in a hypercolumn, each comprising around 80 neurons. An important distinction is that the columnar organization is functional by definition, and reflects the local connectivity of the cerebral cortex. Connections “up” and “down” within the thickness of the cortex are much denser than connections that spread from side to side.” 3. Boucsein et al.c http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3072165/ “This slightly ambiguous term loosely describes the concept of vertically arranged groups of cells that share certain functional and/or anatomical properties and could represent a “basic functional unit” in cortical processing. They are ubiquitous in the brain but in no way obligatory, and a comprehensive description of the various forms of “columns” in the brain is still lacking.” a

Johansson, C., Lansner, A., 2007. Towards cortex sized artificial neural systems. Neural Network. 20 (1), 48e61, Elsevier. Mountcastle, V.B., 1997. The columnar organization of the neocortex. Brain 20 (4), 701e722, Oxford University Press. Boucsein, C., Nawrot, M.P., Schnepel, P., Aertsen, A., 2011. Beyond the cortical column: abundance and physiology of horizontal connections imply a strong role for inputs from the surround. Front Neurosci. (5), 32. b c

TABLE 5.2 Examples for terms that refer to columnar structures in the cortex. Cortical area

Definition

Dimension

References

Cortical column/ module

S1

Penetrations parallel to the pial surface and crossing the vertical axis of the cortex pass through 300- to 500-mm-sized blocks of tissue in each of which neurons with identical properties are encountered. Sharp transitions are observed from a block with one set of properties to the adjacent block with different properties. The defining property for place is the peripheral receptive field, the zone on the body surface within which an adequate stimulus evokes a response of cortical cells.

300e500 mm

Mountcastle (1957), Powell and Mountcastle (1959)

Ocular dominance column

V1

Ocular dominance (OD) columns or ocular dominance stripes are regions of neurons in the visual cortex that respond to the stimulation from either the left or right eye, and they can be defined both anatomically and physiologically.

Variable

Hubel and Wiesel (1969)

Orientation columns

V1

Form orientation slabs that measure 0.5 e1.0 mm in the iso-orientation direction, and in which a full 180
rotation of orientation preference is repeated.

560 mm

Hubel and Wiesel (1968)

Blobs

V1

Metabolic activity 2DG or cytochrome oxidase staining. Blob cells respond differentially at low spatial frequencies (1.1 cycles per degree), interblob cells at higher frequencies (3.8 cycles per degree).

150 mm diameter, most prominent in layers II and III. Repeat interval of 500e550um, the parallel rows are 350 mm apart

Livingston and Hubel (1984)

Isofrequency bands

A1

Neurons of similar frequency preference are arranged in isofrequency bands (IFBs) across the primary auditory cortex (AI) of many mammals.

No wider than 200 mm, and 5e7 mm in length extending across the gyrus

Binaural summation columns

A1

Most neurons arrayed in a column perpendicular to the cortical surface display the same aural dominance and binaural interaction. Summation columns occupy about two-thirds of the area sampled; suppression columns, about one-third. Within most suppression columns, the contralateral ear was dominant. Within summation columns, aural dominance varied. Summation columns appear to be composed of smaller columns differing in aural dominance.

The sizes of binaural interaction columns vary considerably; some occupy several square millimeters of cortical surface. At least some binaural interaction columns occupy stripes of cortex-oriented orthogonal to isofrequency contours

Cortical columns Chapter | 5

Module

105

Continued

Module

Cortical area

Definition

Dimension

References

Motor columnar aggregates

Motor cortex

Pyramidal and nonpyramidal cells are clustered into columnar aggregates.

300 mm wide, separated by 100 mm cell-sparse zones

Meyer et al. (2010)

Motion columns

MT

Neurons in monkey MT with similar axes of motion preference are arranged in vertical columns, and that these columns are themselves arranged in slabs in which a full rotation of 180
of axis of motion is represented.

400e500 mm

Shape and face recognition columns

Homotypical inferotemporal cortex

Require moderately complex features (shapes and faces) for their activation.

Microcolumns

All cortical areas

The dendrites of 3e20 large pyramidal cells of layer V form clusters that ascend together through layer IV.

These modules are 30 mm in diameter, and cooccur with center-to-center spacing that varies from 20 to 80 mm; the wider spacing occurs in the larger brains of the macaque monkey and man

Fleischhauer et al. (1972), Peters and Walsh (1972)

Barrels

Some rodent S1

Cytoarchitectonic patterning of the layer IV neurons forms a ringlike structure on tangential sections.

Variable 300 mm

Woolsey and Van der Loos, 1970

Synaptic ZN

Monkey primary visual cortex

Synaptic Zn patches correspond to a subset of corticocortical terminations.

VGLUT-2 columns

Rat and mouse barrel field

VGLUT-2 marker for thalamocortical termination.

Variable 300 mm

Ontogenic units/ columns

Monkey

The progenitor cells that generate the minicolumn.

Each proliferative unit in the ventricular zone of the monkey consists of 3e5 stem cells, a number that gradually increases to 10e12 stem cells during development; the units are separated by glial septa (Rakic, 1988)

Domains

Early postnatal rat cortex

Domains of spontaneously coactive neurons using optical recordings of brain slices labeled with the fluorescent calcium indicator fura-2 in early postnatal rat cortex.

The functional domains were 50e120 mm in diameter on tangential slices; they spanned several cortical layers and resembled columns found in the adult cortex in coronal slices

Rakic (1988)

106 PART | I Circuit development

TABLE 5.2 Examples for terms that refer to columnar structures in the cortex.dcont’d

Cortical columns Chapter | 5

107

many varieties of “cortical columns” defined by different criteria and by different authors that it is difficult to define, relate, or compare these columns. We are not any closer to defining an archetypical cortical column than when the concept was first introduced. In this chapter, (1) we discuss the problems associated with our current nomenclature; (2) discuss the evidence for and against the idea that columns are the common building blocks of the cortex; (3) examine the question of how constant the cell numbers are within a column and how homogeneous is the structure of the various columns; (4) examine the possible functions of the columns; and (5) examine our current knowledge of the columnar development in the cortex. (i) The cortical column nomenclature reflects the history of changing concepts on cortical anatomy, functional representation, and development. The cortex is organized horizontally into six laminae and vertically into groups of cells linked synaptically across and along the horizontal layers. The definitions of laminae and radial modules are both a historic convention, rather than a biologically or functionally related demonstrable reality. All mammalian cerebral neocortices have a laminar structure that has been historically and arbitrarily divided into six layers (Brodmann, 1909; von Economo and Koskinas, 2008; Lorente de Nó, 1922, 1949). This basic master plan is modified according to variations. These variations expose the problem of fitting the hexalaminar universal model to all mammalian cortices. Subdivisions of layers III, IV, V, and VI in some species, missing layers (IV) in some cortical areas, or fused layers (II/III) in other areas indicate that there is uncertainty and a random element to current laminar nomenclature (Molnár, 2011). The radially oriented apical dendrites and processes and the general radial orientation within the cortex have been widely known since the time the cerebral cortex was first examined microscopically (see, e.g., Cajal, 1909). Also the vertical connectivity linking neurons across cortical layers had been described by Lorente de Nó (1949) in the primary somatosensory cortex of the mouse. The concept of a column emerged from the functional properties of the cortex and received attention after Vernon Mountcastle discovered that the neurons are arranged vertically (or radially in the convoluted cerebrum) in the form of columns spanning the width of the primate somatosensory cortex with cells in each column responding with distinct receptive field properties (superficial as compared with deep skin receptors) to a stimulus within a single receptive field at the periphery (Mountcastle et al., 1957). Although now we know that these functionally distinct “columns” were separate, distinct cortical areas, and not functional units within a cortical area (Kaas et al., 2011), these observations drove further discoveries of an array of iterative neuronal groups (also called modules) that extend radially across cellular layers VI to II with layer I at the top. Subsequently, Hubel and Wiesel (1968) revealed the orientation and ocular dominance columns in the primary visual cortex, and this was followed by the observations of Abeles and Goldstein (1970) in the primary auditory cortex (Table 5.2). These physiological observations led to the concept that “neurons within a given column are stereotypically interconnected in the vertical dimension, share extrinsic connectivity, and hence act as basic functional units subserving a set of common static and dynamic cortical operations that include not only sensory and motor areas but also association areas subserving the highest cognitive functions” (Jones and Rakic, 2010). The inclusion of the highest cognitive functions was, of course, an extrapolation that lacked evidence. The concept that the cortex comprises similarly structured units is an attractive one, but it seems that there are far too many variations and individual units that can be highly specialized and vary within certain cortical areas or sectors within areas.

5.1 The loose and uncritical use of the term in ways that are so generalized as to be unhelpful and even confusing Although the anatomical and functional columnarity of the neocortex has never been in doubt, over time and with more discoveries of radial arrangements in the cortex, the term “cortical column” became looser as columns were defined by cell constellation, dendritic bundling, pattern of connectivity, myelin content, staining property, magnitude of gene expression, or functional properties (Rockland, 2010) (Table 5.2). Although the term column is only used by some to refer to “interconnected neurons, with common input, common output, and common response properties extending through the thickness of the cortex,” others do not use these criteria and the term “column” evolved into a loose and somewhat ambiguous term referring to some aspect of vertical organization of the cortex (Table 5.1, third definition). Montcastle’s definition of column from 1997 is different from the one formulated in 1957, and it includes references to physiological, anatomical, and embryological aspects: “The basic unit of the mature neocortex is the minicolumn, a narrow chain of neurons extending vertically across the cellular layers IIeVI, perpendicular to the pial surface.

108 PART | I Circuit development

Each minicolumn in primates contains 80e100 neurons, except for the striate cortex where the number is 2.5 times larger. Minicolumns contain all the major cortical neural cell phenotypes, interconnected in the vertical dimension. The minicolumn is produced by the interactive division of a small cluster of progenitor cells, a polyclone, in the neuroepithelium, via the intervening ontogenetic unit in the cortical plate of the developing neocortex” (Mountcastle, 1997). This excerpt shows how, over the decades, the increasingly protean imagery evoked by the term “column” now obliges investigators to acknowledge its conceptual and linguistic shortcomings (Rockland, 2010). Structural, functional, and embryological definitions are used loosely, without proper and unequivocal definitions. Therefore, it is difficult to define what constitutes a particular “cortical column.” Moreover the use of the terminology is not stringent. Most columns have no definable “solid” borders; some of the structures referred to as a column do not extend across the entire thickness of the cortex from the pial surface to the white matter (e.g., barrels, microcolumns). The term “column” has become too general. To convey adequately the complex aspects of cortical organization, additional adjectives are required to specify a particular entity. Mountcastle used the terms column and modules interchangeably, but nowadays modules are used more loosely. Other terms, “patch” or “domain,” to varying degrees suffer from the same problem (Rockland, 2010).

5.2 A lack of a universal presence of certain columns within cortical areas, brains, and species is undermining the idea that similar building blocks comprise all cortical circuits 5.2.1 The use of physiological methods to reveal columns Evidence for neocortical columnar organization was initially obtained in electrophysiological studies of single neurons in the somatic sensory cortex in anesthetized cats and monkeys (Mountcastle, 1957; Powell and Mountcastle, 1959). Microelectrode penetrations made normal to the pial surface encounter neurons in each cellular layer with similar properties of place and modality. Penetrations parallel to the pial surface and crossing the vertical axis of the cortex pass through 300e500 mm sized blocks of tissue in each of which neurons with identical properties are encountered. Sharp transitions are observed from a block with one set of properties to the adjacent block with different properties. The defining property for place is the peripheral receptive field, the zone on the body surface within which an adequate stimulus evokes a response of cortical cells. To reveal functional modularity in the cortex 2-deoxy-glucose, optical recording of intrinsic signal, voltage- and calcium-sensitive dyes, and expression of immediate early genes have also been used in both somatosensory and visual cortical areas (Horton and Adams, 2005).

5.2.2 Columnar organization of some afferent and efferent projections Both intrinsic and extrinsic cortical connections are often patchy and appear columnar in cross sections of the cerebral cortex. Various anterograde tracers injected in vivo are often dramatically patterned in cross section, especially in layer 4 and adjacent layers. The problem here is that very commonly a patchy distribution of label that may involve only one or two laminae at most is interpreted as columnar, whereas the label can be a stripe, area, or spot and the “column” is projected to the structure because of the investigator’s interpretation (Fig. 5.1). By contrast, cortical or thalamic terminations in layer 1 are in fact transcolumnar, typically extending over several millimeters. “Bundles of axons from cells of thalamic modules project to columnar zones of termination in layers IV and IIb of the postcentral cortex, forming clusters separated by zones in which terminals are much less dense. Clustering obtains also for the ipsilateral cortico-cortical and transcallosal systems” (Mountcastle, 1997). It is widely considered that the effective unit of operation in such a distributed system is not the single neuron and its axon, but groups of cells with similar functional properties and anatomical connections (Jones, 1999). This modular arrangement might allow a large number of neurons to be connected without a significant increase in cortical volume. Mitchison (1992) estimated that fusing 100 cortical columns would lead to a 10-fold increase in cortical volume. The explanation for this surprising estimate is that within a column only restricted subsets of neurons are involved in longdistance connections, whereas the majority is only connected locally within the columns. Consequently, the length of axons that interconnect neurons is shortened, reducing also the cortical volume. The hypothesis requires that nerve cells in the middle layers of the cortex, in which thalamic afferents terminate, should be joined by narrow vertical connections to cells in layers lying superficial and deep to them, so that all cells in the column are excited by incoming stimuli with only small latency differences. Experiments in monkey, however, did not show such homogeneous arrangement and revealed that terminal arbors of individual thalamocortical axons are often smaller than the cross-sectional width of the region that

Cortical columns Chapter | 5

109

(A)

(B)

(C)

(D)

(E)

(F)

250 μm FIGURE 5.1 Examples of modular circuitry in the mammalian brain. (A) Ocular dominance columns in layer IV in primary visual cortex (V1) of a rhesus monkey (autoradiograph after injection radioactive proline into one eye). (B) Blobs in layers IIeIII in V1 of a squirrel monkey (cytochrome oxidase histochemistry). (C) Stripes in layers IIeIII in V2 of a squirrel monkey (cytochrome oxidase histochemistry). (D) Barrels in layer IV in primary sensory cortex (S1) of a rat (succinic dehydrogenase histochemistry). (E) Barreloids in the ventrobasal nucleus of the thalamus in a rat (succinic dehydrogenase histochemistry). (F) Glomeruli in the olfactory bulb of a mouse (Sudan Black staining). From Purves, D., Riddle, D.R., LaMantia, A.S., 1992. Iterated patterns of brain circuitry (or how the cortex gets its spots). Trends Neurosci. 15 (10), 362e368.

showed response revealed by optical recording in the so-called “activity columns” (Blasdel and Lund, 1983; Freund et al., 1989). Thus, “activity columns” are assembled from the convergence of smaller units defined by arbors in a 300e500 mm wide space and not only defined by activity related or molecular factors (Inan and Crair, 2007). Moreover, analysis of a large data set of recordings has revealed that, within a cortical column, connectivity is highly nonuniform (Song et al., 2005).

110 PART | I Circuit development

5.2.2.1 Modules of visual cortex The visual cortex processes information concerning form, pattern, and motion within functional maps that reflect the layout of neuronal circuits. The columnar organization in the primate V-1 is defined by the neuronal properties of ocularity and place imposed by geniculate input and by orientation specificity generated by intracortical processing. The neurons studied in tangential penetrations vary systematically in ocularity (Fig. 5.1A) and orientation selectivity (Hubel and Wiesel, 1969). Since the primary visual cortex has been studied in most detail, this is the area where the cortical columns have been identified most methodically: ocular dominance columns, orientation columns, hypercolumns, alternating callosal, and ipsilateral columns (see Table 5.2).

5.2.2.2 Ocular dominance columns/stripes Ocular dominance (OD) columns or ocular dominance stripes are regions of neurons in the visual cortex that respond to the stimulation from either the left or right eye, and they can be defined both anatomically and physiologically (Hubel and Wiesel, 1969). Thalamocortical projections carry signals either from one eye or from the other synapse mostly within layer 4. In a normally developed visual system, the area of dominance columns for each eye is the same, and each cortical cell responds to visual input predominantly according to its column. OD columns were revealed by single unit recording and transneuronal transport across the lateral geniculate synapse of radioactive amino acids (Hubel and Wiesel, 1969). OD columns are slablike domains; columnar width is variable as a function of the visual field; that is, they are larger in the foveal representation (Hubel and Wiesel, 1977). In the peripheral visual field representation, the slablike confirmation breaks up into patches (Adams et al., 2007). Monocular deprivation during early life prevents this balance from developing, and the nondeprived eye assumes control of nearly all cortical cells. These effects were largely identified by Wiesel and Hubel, through studies on cats and monkeys (Hubel and Wiesel, 1969). Similarly, for ocular dominance columns of primate visual cortex, classical anatomical and physiological studies identified core and edge regions, functionally distinguished by different degrees of monocular bias (LeVay et al., 1975). More recently, different conditions of visual deprivation have revealed functional subcompartments within ocular dominance columns, visualized either by changes in cytochrome oxidase activity (Horton and Hocking, 1998) or by differential expression of immediate-early genes (Takahata et al., 2009). Species variation in columnar structure is hard to reconcile with ideas about the fundamental importance of columns. Some members of single species, e.g., in squirrel monkeys show enormous variability in the expression of ocular dominance columns (Adams and Horton, 2003). Some individuals have normal ocular dominance columns throughout the visual cortex, others only in part, or nearly absent (Fig. 5.2).

5.2.2.3 Orientation columns Within an orientation column, neurons throughout the vertical thickness of the cortex respond to stimuli oriented at the same angle (Hubel and Wiesel, 1968; Hubel et al., 1977, 1978). A neighboring column will then have neurons responding to a slightly different orientation from the one next to it. The functional maps of orientation preference in the ferret, tree shrew, and galagodthree species separated since the basal radiation of placental mammals more than 65 million years agodshare this common organizing principle (Kaschube et al., 2010). Maps of orientation tuning as viewed from the cortical surface (not in sections) contain singularities where orientation columns converge (Blasdel and Salama, 1986), also called pinwheels (Bonhoeffer and Grinvald, 1991).

5.2.3 Gene expression in the cortex in “columnar” fashion Zones of heightened cytochrome oxidase levels can reveal metabolic zones arranged in a modular fashion across the cortex (Livingstone and Hubel, 1984; Fig. 5.1B), but these do not extend through all cortical layers. In monkeys, primary visual cortex cytochrome oxidase (CO) staining reveals metabolic activity in a nonuniform fashion. The patches of CO activity correspond to thalamocortical terminations (Livingstone and Hubel, 1982). Adjacent sections reacted for synaptic zinc show patches that correspond to another subset of corticocortical terminations (e.g., Ichinohe and Rockland, 2004). There are numerous other differences that distinguish cortical modules; some of these are associated to thalamic inputs (e.g., Vglut2; 5HTT).

5.2.3.1 A system of interleaving modules in rodent layer II Primary visual cortex of both mice and rats exhibits a striking anatomical modularity in layer II. Here, the primary components are input-specific patches, where thalamic input (visualized by VGluT2) interdigitates with zinc-expressing

Cortical columns Chapter | 5

(A)

(B)

(C)

111

(D)

MC



MC



MC





MC

5 mm

FIGURE 5.2 Variable appearance of ocular dominance columns in normal squirrel monkeys. Shown are photographic montages of layer 4C from the left striate cortex of four squirrel monkeys with normal vision 10 days after left eye enucleation. Flatmounts were reacted for CO, and examples for the individual variations are presented: (A) large, crisply segregated columns, (B) intermediate columns, (C) fine, indistinct columns, (D) rudimentary columns. The columns are essentially absent, although hints were visible in a few peripheral regions of cortex. Note thin profiles (arrows) radiating from the blind spot in (C and D), which represent shadows (angioscotomas) from retinal blood vessels. Columns were virtually absent in one-third of the animals (D). *, blind spot; MC, monocular crescent. From Adams, D.L., Horton, J.C., 2003. Capricious expression of cortical columns in the primate brain. Nat. Neurosci. 6, 113e114.

putative cortical connections and parvalbumin-expressing interneurons. M2 muscarinic acetylcholine receptors have a modular pattern, which matches thalamic terminations from the lateral geniculate nucleus. Cellular clusters distinguished in the larger fields produced by sections cut tangentially across layer II, align preferentially with thalamic terminations, and interdigitate with dendritic bundles of underlying neurons in layers III and V. By size, the modules are about 120 mm apart. They show distinct functional specializations, not for ocularity, but rather for spatial or temporal acuity (respectively, the M2-rich and the M2-poor patches). Neuron clusters are prominent in layer II of entorhinal cortex. In a quantitative comparative study, provide evidence for conserved size and periodicity of pyramidal cell patches in layer II of five species and further suggest an evolutionary continuity with neocortical modules in layer II. Cellular modularity has also been reported in insula cortex of dolphins.

5.2.3.2 Overlap between columnar entities within the same structures; combining physiological and anatomical definitions Individual columns are embedded within distributed networks, and cortical modules are composed of groups of minicolumns. The same column can be part of different networks (e.g., ocular dominance and orientation columns or hypercolumns). Despite decades of work, the organization of these modules and their connections, singly or in relation to each other, is only poorly understood. Anatomical observations are often linked to modular patterns of increased metabolic activity; blood flow studies, 2DG uptake, or expression of immediate-early genes depend on the stimulus and may be limited to a single layer, a few layers, or a whole cortical column. However, some of these studies revealed no discrete anatomical arrangements that would explain modularity of function. Or perceived anatomical arrangements are not in line with detected physiological changes raising the question: How can an “imperfect” anatomical arrangement generate functionally distinct modules? At the cellular level, there is growing evidence that cortical columns contain multiple, highly specific, fine-scale subcircuits (Yoshimura et al., 2005; Otsuka and Kawaguchi, 2008) and that within columns there are locally heterogeneous response properties (Sato et al., 2007).

112 PART | I Circuit development

5.3 General concept that the cortical column (even just an arbitrary unit column that includes the full depth of the cortex) has a universal constant number of neurons associated with it The various cortical columns that have been described by different anatomical or physiological methods have very different sizes, shapes, and diameters (examples shown in Fig. 5.1). As we discussed, the term cortical “column” is ambiguousdit can refer to small-scale minicolumns (diameter 50 mm), to larger-scale macrocolumns (diametere300e500 mm), or to multiple different structures within both categories (Jones, 2000; Rakic, 2007; Rockland, 2010).

5.3.1 Number of neurons in a cortical column Although there is very little quantitative work on the number of neurons in anatomically or physiologically identified cortical columns, it is expected that they are different. It is also generally accepted that the cortical surface areas vary much more than the radial thickness of the cortex. Powell after returning to Oxford from Mountcastle’s laboratory was influenced by the concept of the column and set out to quantify parameters within a cortical segment that had roughly the same dimensions as the physiological columns that were estimated (Jones, 1999). After quantification in selected species, it has been proposed that regardless of the thickness of the cortex within an arbitrary (30-mm-wide, 25-mm-deep) vertical “column” between the pial surface and the white matter of the cortex, the number of neurons is 110 in all cytoarchitectonic areas (Rockel et al., 1974, 1980). Under such conditions, the neuronal number was claimed constant in all mammalian species (mouse, rat, cat, Old World monkey, human) and for all cortical thicknesses, the numbers of cells in these arbitrary columns in prefrontal, primary motor, somatosensory (posterior) parietal, and temporal neocortex (except the primary visual cortex in primates) all being the same. There was only one area in the cortex that showed a difference from this constant number; in all primates studied (Galago, marmoset, squirrel monkey, macaque monkey, baboon, and human), the number per 30-mm-wide, 25-mm-deep column of visual cortex was increased to about 260e270. This increase is a reflection of the much higher packing density of cells in the true striate cortex. In a later study with Anita Hendrickson, Powell found that the neuronal number remained constant across both the monocular and binocular segments of the macaque visual cortex (Powell and Hendrickson, 1981). The changes in packing density of neurons in the arbitrary unit columns were inversely related to the volume of neuropil. Using similar methods in marsupials, it has been established that the neuronal numbers are half the ones observed in mouse in a similar arbitrary unit column (Cheung et al., 2007, 2010; Fig. 5.3). Using a more recent “unbiased” stereology method, Herculano-Houzel and her colleagues showed that the density of neurons in the neocortex varies as much as three times even among the highly related primate species (Herculano-Houzel et al., 2008; Collins et al., 2010). In spite of these observations. it is still true that cortical expansion in evolution is achieved by expanding cortical surface area, with relatively little change in thickness. The ratio for the cerebral cortical surface areas in mouse, macaque, and human is 1:100:1000, whereas for the cortical thickness the ratio is more like 1:1:1 since it is in the range of 2e4 mms in all three species (Rakic, 2009). However, the idea that all mammalian cortices in most areas have a very similar numerical constancy has to be abandoned. In fact, the differences might hold a key to understanding cortical specializations for specific functions.

5.4 Lack of correlation between the absence or presence of particular columns and a specific sensory or cognitive processing network (comparisons across the same brain and across close and more distant species) 5.4.1 Microscopic and macroscopic cell patterning defining cortical modules Most cortical columns can be related to some forms of cellular patterning in the cortex. These can be from subtle microscopic patterns to macroscopically identifiable features. Some distinctive body parts with homologously shaped neuronal maps are even recognizable in the somatosensory cortex. Examples include the barrel cortex of the mouse; representation of the nasal proboscis of the star-faced mole (Catania and Kaas, 1995); the representation of the raccoon hand (Welker et al., 1964), or the primate hand (Jain et al., 1998). These macroscopically visible columns received special status in neurobiology because they helped investigators to understand questions related to synaptic plasticity and map formation. The barrel field is one of the best-studied model systems in the mouse cortex; yet we still do not comprehend its functional significance. Are they structures without any particular function (Horton and Adams, 2005)?

Cortical columns Chapter | 5

Mouse

Opossum

Wallaby

(D) 743±9

Total

(B)

Glia 117±12

Wallaby

Neurons

Glia

Neurons

Glia

418±29

89±7

544±13

130±8

L1

13±3

5±3

14±4

7±2

15±3

5±3

L2/3

183±5

22±5

122±17

14±2

72±23

22±5

L4

186±5

15±5

105±9

14±2

204±11

15±5

L5

130±3

31±3

96±13

27±3

102±8

31±3

L6

233±7

44±7

81±7

27±5

82±13

44±7

(C) (E)

L1

Opossum

Mouse Neurons

(A)

113

(E’) 100%

800

2

3

13

3 13

25

L1

29

80%

183

600

L2/3

L2 L1 15

186

L2

L3

14

400

72

122

L4

130

L3

204

L1 L2/3 L4 L5 L6

60%

25

40%

18 23

105

200

37 25

L1 L2/3 L4 L5 L6

19

20% 233

L4 L4

L5

96

102

81

82

0

31 19

15

Opossum

Wallaby

0% Mouse

Opossum

Wallaby

Mouse

L5

L5 L6 L6 L6

(A’)

100 μm

(B’)

100 μm

(C’)

100 μm

FIGURE 5.3 Quantification of the number of neurons in mouse, opossum, and wallaby. (AeC) Cresyl violetestained sections of adult (A) mouse, (B) opossum, and (C) tammar wallaby. An arbitrary “unit column” (a 100-mm wide) spanning from layer 1 to 6) was marked in the primary somatosensory/visual area (boxed areas in AeC, higher magnification in A0 eC0 ). The number of neurons and glia was quantified in each layer and expressed as mean  SEM in (D). (E) The mean number of neurons present in each cortical layer, showing that the number of neuron in a unit column is not constant between different infraclass within mammals. (E0 ) The proportion of neurons in each cortical layer. Scale bar: AeB ¼ 500 mm, C ¼ 1 mm. Reproduced with permission from Cheung, A.F., Kondo, S., Abdel-Mannan, O., Chodroff, R.A., Sirey, T.M., Bluy, L.E., Webber, N., DeProto, J., Karlen, S.J., Krubitzer, L., Stolp, H.B., Saunders, N.R., Molnár, Z., 2010. The subventricular zone is the developmental milestone of a 6-layered neocortex: comparisons in metatherian and eutherian mammals. Cerebr. Cortex. 20 (5), 1071e1081.

5.4.2 Are barrels cortical columns? It is puzzling that barrel fields are present in rats, mice, squirrels, rabbits, possums, and porcupines, but not in raccoons, beavers, or cats (Woolsey et al., 1975). The presence or absence of barrels is not related to the presence of actively mobile whiskers (whisking behavior), since guinea pigs do not whisk but nonetheless have a barrel field. The peripheral somatic sensory input is relayed through the brain stem and the ventrobasal complex (VB) of the thalamus before it is transmitted to layer IV, the gateway of the sensory cortical circuitry. Thalamic axons form arbors and establish synapses in a peripheryrelated pattern in layer IV. This pattern formed by thalamocortical axons can be present in the absence of the cytoarchitectonic pattern that was originally termed barrels (see López-Bendito and Molnár, 2003). The individual thalamocortical axon clusters that form periphery related pattern first are surrounded by densely packed layer IV cells that form the walls of the actual cytoarchitectural “barrels.” In the middle of each barrel is a plexus of thalamic fibers carrying signals from one corresponding whisker (Fig. 5.4). In the barrel field of the rodent somatosensory cortex, dendritic bundles are mostly located in the barrel walls and septa, avoiding hollows (mouse, Escobar et al., 1986). In rodent barrel cortex, dendrites of neurons in layer 4 conform to barrel limits (Harris and Woolsey, 1979), but this seems to be an exceptional case. The best documented example of two thalamic systems in the same layer is from rodent barrel cortex (Alloway, 2008). In layer 4, thalamocortical projections from the ventral posterior medial and the posterior

114 PART | I Circuit development

P0

P7

FIGURE 5.4 Schematic overview of the development of the periphery-related thalamocortical patterning and the cytoarchitectonic barrel formation in the mouse. Thalamocortical fiber clusters arise from an initially uniform distribution of thalamocortical arbors (red), and they impose the characteristic patterning of layer IV neurons (blue). Left column represents coronal, right column tangential sections. From Molnár, Z., Molnár, E., 2006. Calcium and NeuroD2 control the development of thalamocortical communication. Neuron 49 (5), 639e642.

nuclei, respectively, target the barrels (lemniscal pathway) and their intervening septa (paralemniscal pathway). A similar segregation occurs more generally but with a segregation in different layers. The functional significance of the thalamocorticalecorticocortical patterning is unknown but could be related to differential processing by distinct postsynaptic populations. Barrels show variation in size and shape across S1. There have been few quantitative studies of the differences and their relevance. Sakmann’s group has presented a set of papers in which they define a “standard” column in the rat somatosensory cortex as based on the topographically specific input from the large bundle of thalamocortical axons emanating from a single “barreloid” in the ventral posterior medial (VPM) nucleus of the thalamus and terminating in one of the “barrels,” in the layer IV aggregations of neurons (Helmstaedter et al., 2007; Meyer et al., 2010; Wimmer et al., 2010). In this case, then, their column is of the kind defined originally by Mountcastle and not a minicolumn, although it may contain minicolumns as defined earlier. Based on measurements of concentrations of thalamocortical axon terminals labeled by green fluorescent protein expressed in their parent cells and extending the width of the periodic densities of terminations, which in layer IV are approximately 300-mm wide, across the depth of the cortex, this column has a cross-sectional area of about 121 000 square microns and a depth from pia to white matter of approximately 1840 mm. A second kind of column defined by the authors has its basis in the terminations of axons arriving from the posterior medial (Pom) nucleus of the thalamus and ending deep and superficial to the barrels and especially in the zones of reduced cell density or “septa” lying between them. This column, as measured from septum to septum and across the intervening barrel, is thus a little wider than the column defined by inputs to the barrels; it has a cross-sectional area of approximately 124 000 square microns but when projected across the depth of the cortex has the same length as the VPM-based column. The measurement of the Pom-based barrel might be rather arbitrary since the authors describe the axons of Pom neurons as spreading horizontally for seemingly wider extents than those from VPM (Jones and Rakic, 2010). The barrels in rodent somatosensory cortex are not stereotyped. Hollow barrels, with cell sparse cores, are typical of mice, young rats, and the anterolateral subfield of mature rats, but solid columns, with cell-dense cores, are typical of the main posteromedial field in rats (Rice, 1995). Variability is not reported for other columnar systems of connections, but this is likely because many of the systems are harder to visualize globally or require specialized tissue processing.

5.4.2.1 A system of interleaving modules in rodent layer VI Layer VI of mouse somatosensory cortex is reported to contain readily identifiable “infrabarrels” that align with the layer IV barrels. These consist of corticothalamic (CT) projecting neurons, which interdigitate with corticocortically (CC) projecting neurons. However, it is the CC neurons, not the CT neurons, that receive dense thalamic input, and these are also distinguishable by greater excitability.

Cortical columns Chapter | 5

115

Layer VI in many cortical areas is recipient of at least collaterals of thalamic inputs and contains populations of both CC and CT projecting neurons. So far, overt modularity has only been reported in mouse barrel cortex. In nonlaboratory species, however, pronounced cellular modularity, usually in the form of narrow cellular strings or “minicolumns,” has frequently been documented; for example, dorsal parietal cortex of the giraffe and the bill representation in the primary somatosensory cortex of the platypus (Fig. 5.1). Short vertical clusters of cells are commonly observed in the deep layers of marine mammals. These occur in isolation from layer IV, since the neocortex in cetaceans and hippopotamids is completely agranular.

5.4.2.2 Function of barrels The function of the barrels in the rodent primary somatosensory cortex is not known. The cortical architecture can be missing or significantly disrupted and yet apparently remains functionally intact. For example, the disrupted barrel cortex in the reeler and in other mutant or transgenic mice is not associated with marked somatosensory deficits (Rakic and Caviness, 1995; López-Bendito and Molnár, 2003). The degree to which cortex is modifiable and, by what mechanisms, has been extensively investigated under various environmental manipulations. Although we do not know what the functional relevance (if any) of the barrel arrangements may be, this system helped the understanding of various aspects of cortical circuit formation and plasticity. Study of the barrel field in various mouse mutants proved to be instrumental in the understanding of the molecular mechanisms of these interactions (Erzurumlu and Kind, 2001). The development of the periphery-related patterning of the thalamocortical projections and the induction of the cytoarchitectonic barrels require both pre- and postsynaptic interactions. During the first days of postnatal development, thalamic projections assume a periphery-related pattern within layer IV precisely mirroring the arrangements of the whiskers. Thalamocortical axon segregation is soon followed by the relocation of layer IV cells from an initially homogeneous distribution to the walls of the barrels surrounding the clustered thalamic projections (Fig. 5.4). Van der Loos and Woolsey (1973) provided evidence for the environmental influence on cortical cytoarchitectonic differentiation by demonstrating that changing or blocking the flow of sensory input from specific whiskers during the early stages of development results in a cascade of events that will change the arrangements and somatodendritic morphology of layer IV cells. With the development of finer techniques of clonal analysis and neuronal cell-type specification, one can anticipate a new generation of genetic and molecular manipulations that will help us elucidate the underlying mechanisms of barrel formation. Overexpression of NT3 is reported to result in an enhanced expression of dendritic bundles (“minicolumns”) in rat barrel cortex (Miyashita et al., 2010). However, it is not clear to what extent the barrels represent a general and valid model for cortical columns.

5.4.2.3 Microcolumns and apical dendritic bundles There are a number of examples of repeating microarrays of intracortical elements that are interpreted as conforming to a microcolumnar pattern of vertical connections. The observations on patterning of apical dendrites of pyramidal cells with somata located in layers II, III, and V have led to the introduction of the term minicolumns or microcolumns (Fleischhauer et al., 1972; Peters and Walsh, 1972). Innocenti and Vercelli (2010) distinguished minicolumns and bundles, whereas some investigators have used these terms interchangeably. Minicolumns of radially aligned cell bodies can be demonstrated by standard Nissl preparations or other histological methods that reveal cell bodies. Bundles comprise the apical dendrites of pyramidal neurons whose cell bodies are in different layers and can be seen in material prepared by the Golgi technique, stained with osmium for electron microscopical analysis or with markers of somatodendritic morphology (e.g., microtubule-associated protein 2 [MAP2] or SMI32) (Peters and Walsh, 1972; Fig. 5.5). Innocenti and Vercelli (2010) demonstrated dendritic bundles using retrograde transport of lipophilic tracers or intracellular injection of neurons in slice preparations. Myelinated axons are also organized in bundles; these bundles course close to those of the dendrites, and at least some of them originate from neurons whose apical dendrites are in a bundle (monkey primary visual cortex: Peters and Sethares, 1996). Depending also on tangential location and depth, the minicolumns and bundles can be more or less distinct. An average bundle comprises the dendrites of 3e20 large pyramidal cells of layer V that form clusters that ascend together through layer IV. They are joined in the supragranular layers by the successive addition of the apical dendrites of pyramidal cells of layers II and III, and all ascend further, many sending their distal tufts to layer I (Fig. 5.5). Reconstructions have revealed that individual dendrites change their neighborhood relations along a bundle; superficial dendrites can be added between the dendrites from deeper layers; and individual dendrites can bifurcate to sending

116 PART | I Circuit development

AB No A- n er gi c G AB Ae rg ic

No. of neurons

23 μm

23 μm

G

23 μm

0.4

I

II/III

18

5

7

2

6

2

7

1

10

2

V

6

2

VI

8

1

VIb

2

1

64

16

IVB

1600 μm

IVA

IVCα

IVCβ

Total:

FIGURE 5.5 Bundles and microcolumns in the primary visual cortex of the macaque monkey. Left panel: Schematic representation of the arrangements of the apical dendrites of pyramidal cells. Layer II/III, IVA and V pyramidal cells are shown in red, layer VI in green. Gray represents neurons of IVB and IVC without dendrites; azure represents GABAergic neurons. Right panel: Pyramidal minicolumns are represented adjacent to the dendritic bindles. Axons of pyramidal cells are depicted in blue. From Peters, A., Sethares, C., 1996. Myelinated axons and the pyramidal cell modules in monkey primary visual cortex. J. Comp. Neurol. 365, 232e255.

branches to neighboring bundles (Massing and Fleischhauer, 1973; Fig. 5.3 in Rockland and Ichinohe, 2004). In the monkey visual cortex, the microcolumns are estimated to consist of the dendrites of z142 pyramidal neurons. These modules are 30 mm in diameter and occur with center-to-center spacing that varies from 20 to 80 mm, the wider spacing occurring in the larger brains of the macaque monkey and man. Their estimated density is z1270 per mm2 in the monkey visual cortex. In the visual cortex, the mean spacing between modules was found to be 60 mm in the rat, 56 mm in the cat, and 23 mm in the rhesus monkey (Peters, 1997). Not all apical dendrites from layer V enter into the composition of dendritic bundles (Rockland and Ichinohe, 2004). The presence of layer V is not considered the prerequisite for bundle formations (Innocenti and Vercelli, 2010); however, this issue has not been investigated in mutant mouse cortex that lacks a particular subtype of layer V. Apical dendrites of layer VI neurons have been described as clustering together, but forming an independent system of “swathes” concentrated deep to the border of layers III and IV (Escobar et al., 1986; Lev and White, 1997). Maruoka and colleagues used three-dimensional imaging to study the lattice structure that distinct types of excitatory and inhibitory neurons form. They described in many brain areas that major cell types in neocortical layer V form cell

Cortical columns Chapter | 5

117

typeespecific radial clusters, microcolumns (Maruoka et al., 2017). These microcolumns are patterned into a hexagonal mosaic, and microcolumn neurons demonstrate synchronized in vivo activity and visual responses with similar orientation preference and ocular dominance. In early postnatal development, microcolumns are coupled by cell typeespecific gap junctions and attract convergent synaptic inputs to organize into a brain-wide modular system that will be the template for cortical processing (Maruoka et al., 2017).

5.4.3 Complex relationship relations between minicolumns and dendritic bundles Cell bodies of neurons in a minicolumn can be seen to orient obliquely to engage their apical dendrite into the neighboring dendritic bundles already in layer V and more so in layer III (Peters and Walsh, 1972; Peters and Kara, 1987; Gabbott, 2003; Fig. 5.5). The progressive addition of dendrites to the bundle from depth to surface in cortex (“like onions held by their stem”; Peters and Kara, 1987) also indicates that the bundles collect dendrites from more than one minicolumn of cell bodies (Massing and Fleischhauer, 1973; http://www.frontiersin.org/neuroanatomy/10.3389/neuro.05.011.2010/full; Peters and Kara, 1987; Vercelli et al., 2004). Peters and Sethares (1996) observed that neurons with apical dendrites in the same bundle also have cobundled myelinated axons, indicating that neurons in a dendritic bundle might send their axons to the same target. Subsequently, Lev and White (1997) showed that, in the mouse area MsI, following injection of horseradish peroxidase in the contralateral hemisphere, all dendrites in a labeled bundle belonged to callosally projecting neurons, thus suggesting that dendritic bundles are target specific. This issue has been further investigated by Vercelli et al. (2004) in the visual cortex of the rat for different targets (ipsilateral cortex, superior colliculus, pons, lateral geniculate nucleus, striatum). This study strongly supports the concept that dendritic bundles are related to target specificity. Moreover, the composition of dendritic bundles does not seem to depend on the age of the animal and is already established at P3. It is an interesting possibility that some transient elements join the bundles at early developmental stages from the early-generated subplate neurons. Thus, it has been recently demonstrated that subplate apical dendrites have a close association with layer V apical dendrites having the same targets in the early postnatal barrel cortex (Hoerder-Suabedissen and Molnár, 2011). It is not known whether dendrodendritic synapses occur in particular dendritic bundles or not. Unlike what might have been expected, neurons of the same bundle are not more interconnected than neurons of different bundles (Krieger et al., 2007). There are preferential connections between certain output neurons, interestingly between corticocortical and corticotectal neurons, whose apical dendrites lie in separate dendritic bundles in mouse (Brown and Hestrin, 2009). Proposed that neurons in the different layers of one minicolumn, projecting to different targets, send their apical dendrites to separate dendritic bundles where they join apical dendrites of neurons from neighboring minicolumns, projecting to the same target or combination of targets. They propose that in a given cortical locale (area or part of an area) an assembly of apical dendritic bundles constitutes a “cortical output unit.” Such a unit would include outputs to distant cortical or subcortical structures, the parent somata of origin, and basal dendrites and the portion of the neuropil that pertains to them. Dendritic bundles and microcolumns can be identified in all cortical areas in the cerebral cortex of different mammalian species, such as rodents, carnivores, and primates including human. The dendritic bundling seems to offer two important advantages. It might minimize the length of the axonal arbors that contact specific neuronal classes and, in development, it might simplify the axonal search and recognition of targets. The link, if any, between the minicolumn and apical dendritic bundles and functional cortical units is not yet established. The link between the anatomical organization and the physiological units is not clear.

5.4.3.1 Columns outside the mammalian isocortex Columnar arrangements are present in numerous structures in mammals. Iterated circuitry is present in the olfactory bulb glomeruli (Fig. 5.1F), in the barreloids (Fig. 5.1E) and barrelettes in the ventrobasal thalamic nucleus and brain stem, respectively. Columnar structures are also present in the laminated architecture of the superior colliculus (SC) (Illing and Graybiel, 1986; Harting et al., 1992). These can be revealed by acetylcholinesterase reactivity as 200- to 600-mm-wide patches (Harting et al., 1992; Mana and Chevalier, 2001). Gives an overview of these extracortical patterns: “The periaqueductal gray contains longitudinal columns of afferent inputs, output neurons and intrinsic interneurons thought to co-ordinate different strategies for coping with different types of aversive stimuli (Bandler and Shipley, 1994; Keay and Bandler, 2001). The lateral septal nucleus is reported to have a complex system of chemically and connectionally distinct zones of transverse sheets (Risold and Swanson, 1998).

118 PART | I Circuit development

Some thalamic nuclei have distinct domains, which are neurochemically and connectionally distinguishable (Rausell and Jones, 1991). The basal ganglia are organized into neurochemically and connectionally distinct striosomes and matrix (Graybiel and Ragsdale, 1978).” In the cerebellar cortex, an elaborate array of modular subdivisions is revealed by histochemical markers, the topography of afferent projections and some efferent projections, and by gene expression in subpopulations of Purkinje cells (PCs) (Voogd and Glickstein, 1998; Sillitoe and Joyner, 2007). Zebrin II expression in Purkinje cells reveals a parasagittal stripy pattern. Each stripe consisting of a few hundreds to thousands of Purkinje cells, which are highly reproducible, activity independent, and conserved across species. Other molecular and connectivity markers have an orderly relation to zebrinþ or zebrin stripes (Larouche and Hawkes, 2006). The functional importance of this striking organization remains to be elucidated, but, comparable with the mosaicism of the superior colliculus, it has been suggested to subserve a massively parallel architecture with a high number of processing channels (Larouche and Hawkes, 2006). Minicolumn-like dendritic bundles can also be found in numerous noncortical structures (e.g., Roney et al., 1979).

5.4.3.2 Columns in nonmammals The columnar organization of the cerebral cortex is a broadly documented principle of design preserved throughout mammalian evolution (Mountcastle, 1997). Although it is often considered unique to mammals, Karten has questioned the assumption of this uniqueness and has argued for a similar laminar and columnar organization in the avian brain (Wang et al., 2010). Using contemporary methods, Karten and colleagues demonstrated the existence of comparable columnar functional modules in the laminated auditory telencephalon of an avian species (Gallus gallus). Tracer placed into individual layers of the telencephalon within the cortical region that is considered similar to mammalian auditory cortex by Karten and colleagues revealed extensive interconnections across layers and between neurons within narrow radial columns perpendicular to the laminae (Wang et al., 2010). This columnar organization was further confirmed by visualization of radially oriented axonal collaterals of individual intracellularly filled neurons. These findings indicate that laminar and columnar properties of the neocortex are not unique to mammals and may have evolved from cells and circuits found in more ancient vertebrates (Shepherd and Grillner, 2010).

5.4.4 What is the function of a cortical column? The functional rationale put forth for columnar organization of the cerebral cortex includes arguments for “augmenting cortical surface area during speciation; modular segregation of inputs and facilitate computation by enhancing information processing” (Purves et al., 1992). Modular clustering is believed to be important to allow a large number of neurons to be connected without a significant increase in cortical volume (Mitchison, 1992). However, if modules are essential for information processing, why is it that they are present in some species but not in others without any noticeable perceptual differences (Purves et al., 1992; Horton and Adams, 2005)? Or, if they are essential in enhanced computation why are they not present in higher motor and association areas (Purves et al., 1992)? The criteria for the identification for modules/columns are so diverse that it is possible that some variables that might define patchy or modular arrangements might yet be identified in the future. The experimental paradigms provided by the barrel and the ocular dominance columnarity tend to influence the way we look at the cerebral cortex as a whole, but neither is clearly built up from microcolumnar units of cells or connections. None of these cortical arrangements are associated to a particular cognitive or perceptual ability in species where they are present or absent. Horton and Adams highlight the seeming lack of anatomical and functional correlations across species; that is, for visual function, species with or without ocular dominance columns, and for the whisker system, those with and without a cortical barrel. On this basis, they strongly argue for the “lack of particular function of these striking, but inconstantly expressed anatomical features” (Horton and Adams, 2005). Dendritic bundles have been found throughout the mammalian brain and are believed to serve fundamental roles in the brain’s functioning. However, no physiological experiments have yet succeeded in determining the function of these wellestablished anatomical units. The function of the microcolumns as fundamental anatomical units of organization is also not clear. We need much more comprehensive analysis of the intricacies of intracortical connectivity and of the anatomy and physiology of microcolumns in all cortical areas of several species. Combining these approaches could clarify several issues (see Bock et al., 2011); for example, microcolumns represent a fine-grain functional modularity of cortex.

5.4.5 Columns in neuropathology Cortical modules have received attention from neuropathologists not because we know their function, but because they can be visualized, quantified, and compared between subjects. They might be epiphenomena, but they are detectable entities

Cortical columns Chapter | 5

119

and therefore can be used as diagnostic signs for abnormal cortical organization. The periodicities of microcolumnar structures (that contain about 11 neurons and have a periodicity of about 80 microns) were disrupted in two examples of neurodegenerative disease in human (Jones, 2000). Some alterations of the microcolumnar structures have been described in the brain of the elderly (Peters, 1997). Alterations in the clonal dispersion of neurons have been linked to neuropsychiatric disorders associated with abnormal columnar organization (Torii et al., 2009). Currently, it is not known what degree of radial allocation and lateral neuronal dispersion seems to be essential for the proper radial delivery and intermixing of neuronal types in cortical columns (Molnár et al., 2019).

5.5 What is the correlation between the columnar development of the brain and future columns 5.5.1 Cortical columns during development Mountcastle (1997) emphasized that the mode of generation of the cortex already reflects its basic columnar organization. From Golgi preparations and from Nissl-stained material, the radial orientation of neurons within the developing cerebral cortex is apparent from very early stages (Cajal, 1909). Neurons assume a radial orientation and dendritic polarity shortly after their generation. These observations triggered theories that much of the anatomical substrate for a columnar organization would already be specified at early developmental stages, before activity-dependent mechanisms could take place (Rakic, 1988).

5.5.1.1 Ontogenic units/columnsdthe fundamental building blocks in the developing neocortex There is strong evidence for the overall radial migration of the pyramidal neurons in all mammalian cortices. Clonally related postmitotic pyramidal neurons are initially deployed in a geometrically columnar pattern in the embryonic primate cerebrum. Rakic (1988) proposed that the location of the cohorts of cortical neurons from a single neuronal progenitor is not random but is largely predictable. Neurons in mature cortical “minicolumns” are each derived from one of a small group of progenitors forming a polyclonal group already in the ventricular zone (Kornack and Rakic, 1995). The progenitor cells that generate the minicolumn were termed ontogenic columns (Rakic, 1988). Rakic estimated that “each proliferative unit in the ventricular zone of the monkey consists of 3e5 stem cells, a number that gradually increases to 10e12 stem cells during development; the units are separated by glial septa” (Rakic, 1988). According to this theory, the surface area, thus the size, of any neocortex is determined by the number of ontogenetic units, set by the number of symmetric divisions of progenitor cells in the neural epithelium before migration begins (Rakic, 1988, 2009). It has been suggested that one important phenomenon for the increased cerebral complexity during evolution may be the multiplication of neuronal columns throughout the cerebral cortex (Rakic and Caviness, 1995). According to this theory, functional columns in the adult cerebral cortex must consist of several ontogenic columns (polyclones). The actual visualization of these ontogenetic units has not been achieved, and it is still unclear to what extent and how gene expression in the ventricular zone (VZ) could play a role in the development of discrete functional units, such as minicolumns or columns. Cell lineage experiments using replication-incompetent retroviral vectors have shown that the pyramidal neuron progeny of a single neuroepithelial/radial glial cell in the dorsal telencephalon is organized into discrete radial clusters of sibling excitatory neurons (Kornack and Rakic, 1995; Noctor et al., 2001). Costa et al. (2009) noted that most neuronal clones derived from E13 progenitors span 150e250 mm in the horizontal axis and contribute to all cortical layers generated after that embryonic stage. The same authors performed mathematical extrapolations for injections performed at the onset of neurogenesis in the cerebral cortex (E10e11) and suggested that neuronal siblings would not disperse by more than 400e500 mm. Thus, both the radial and horizontal dispersion of excitatory neuronal clones fit well with the possibility that they could help to create a structural basis for the future specification of columns. How these developmental neural clusters relate to adult anatomical and physiological columns has not been addressed. Neurons from different clones intermix with the adjacent columns, as they migrate across the intermediate zone. In addition to the radial allocation of clonally related neurons, short lateral shifts and transfers from their parental to the neighboring radial glial fibers have also been described (Tan and Breen, 1993; Kornack and Rakic, 1995; Noctor et al., 2001). These dispersed neurons intermix with neurons originating from neighboring proliferative units. The molecular mechanisms, their role, and the significance of this lateral dispersion for cortical development are not understood. A recent study revealed that the lateral dispersion depends on the expression levels of EphAs and ephrin-As during neuronal migration.

Control EP

120 PART | I Circuit development

E12.5 EP > E15.5

E12.5 EP > P4

EphA7 EP

E12.5 EP > E18.5

FIGURE 5.6 Eph receptor A (EphA) and ephrin A (Efna) signaling-dependent shift in the allocation of clonally related neurons changes the level of lateral dispersion in embryonic brain. Torii and colleagues demonstrated that lateral dispersion depends on the expression levels of EphAs and ephrin-As. Increased EphAs and ephrins during neuronal migration lead to increased tangential sorting of cortical neurons. The figure represents the distribution of EYFP-labeled neurons (green or black) during development in the cortex in which control or EphA7 expression plasmid was delivered using electroporation (EP) with EYFP plasmid. Scale bars, 600 mm (for E18.8), 50 mm (for E15.5), and 200 mm (for P4). From Torii, M., Hashimoto-Torii, K., Levitt, P., Rakic, P., 2009. Integration of neuronal clones in the radial cortical columns by EphA and ephrin-A signalling. Nature 461 (7263), 524e528.

Torii et al. (2009) demonstrated that Eph receptor A (EphA) and ephrin A (Efna) signaling-dependent shift in the allocation of clonally related neurons is essential for the proper assembly of cortical columns in the mouse cerebral cortex (Fig. 5.6). Currently, it is not known what degree of radial allocation and lateral neuronal dispersion seems to be essential or optimal for the proper radial delivery and intermixing of neuronal types in the cortical columns. The degrees of mixing of derivatives of different progenitors have not been estimated for different species.

5.5.1.2 Sibling neuron circuits in the developing columns The developing cortex and the adult cortical columns both have overwhelmingly radial arrangements. In the developing brain, the clonally related neurons have higher chance of being situated within the same radial volume of cortical tissue (Tan and Breen, 1993). It has been proposed that the initial columnar organization may act as a seed to establish the primary information-processing unit in the cortex. This raises the question as to whether neurons from the same clone are preferentially interconnected. Are they more likely to develop chemical synapses with each other rather than with neighboring nonsiblings? Cell lineage or clonal analysis studies have been combined with recording experiments to study the possibility that the cell lineage of single neuroepithelial/radial glia cells could form radial columns of sibling, interconnected neurons. Yu et al. (2009) identified individual clones of cortical pyramidal neurons by injecting enhanced green fluorescent protein (EGFP)eexpressing retroviruses into the lateral ventricle of mouse embryos at early stages of neurogenesis. They made simultaneous wholecell recordings of two EGFP-expressing sister neurons and observed that these cells displayed unidirectional synaptic connections in 35% of pairs. In contrast, less than 7% of radially situated nonsister excitatory neurons were connected (Yu et al., 2009). This experiment provides strong support for the idea that excitatory neurons generated from the same progenitor keep spatial relationships and display (mutual) connectional preferences, but it stops short of relating the clones to adult cortical columns. There is a distinction between the idea that sibling neurons have predictable arrangements and connectivity and the idea that adult columns are “preformed” and prespecified in ontogenic units in the ventricular zone. More work is needed to clarify these issues. The potential molecular mechanisms involved in the establishment of sibling neuron circuits are not known. It has been hypothesized that neurons derived from the same progenitor are more likely to display similar chemical and physical properties due to their genetic inheritance (Costa and Hedin-Pereira, 2010). Sister neurons might share more of the combinatorial transcription code that has been present in the common cortical progenitors, and therefore the sister neurons might share a similar set of surface molecules that are important to cellecell recognition or to molecular guidance cues. The molecular determinants of the cell-intrinsic properties for cellecell recognition between sibling neurons in the cortex are currently a largely uncharted territory.

Cortical columns Chapter | 5

121

FIGURE 5.7 Mediator of cell motility 1 (Memo1) is a critical determinant of radial glial tiling during neocortical development. Memo1deletion or knockdown leads to hyperbranching of RGC basal processes and disrupted RGC tiling, resulting in aberrant radial unit assembly and neuronal layering. (From Nakagawa, N., Plestant, C., Yabuno-Nakagawa, K., Li J, Lee, J., Huang, C.W., Lee, A., Krupa, O., Adhikari, A., Thompson, S., Rhynes, T., Arevalo, V., Stein, J.L., Molnár, Z., Badache, A., Anton, E.S. (2019). Memo1-Mediated Tiling of Radial Glial Cells Facilitates Cerebral Cortical Development. Neuron 103 (5),:836e852.e5.)

Polarized, nonoverlapping, regularly spaced, tiled organization of radial glial cells (RGCs) is important for normal cortical development. This arrangement serves as a framework to generate and organize cortical neuronal columns, layers, and circuitry. Recent findings identify mediator of cell motility 1 (Memo1) as a mediator of radial glial progenitor cell scaffold tiling, necessary to generate and organize neurons into functional ensembles in the developing cerebral cortex (Nakagawa et al., 2019). This study demonstrated that Memo1 deletion or knockdown leads to hyperbranching of RGC basal processes and disrupted RGC tiling, resulting in aberrant radial unit assembly and neuronal layering (Fig. 5.7).

5.5.2 Transient columnar domains during development The coordinated calcium fluctuation patterns underlying gap junctionalemediated communication were suggested as a possible basis for the formation of initial functional cell assemblies in postnatal cerebral cortex. Yuste et al. (1992) observed distinct domains of spontaneously coactive neurons using optical recordings of brain slices labeled with the fluorescent calcium indicator fura-2 in early postnatal rat cortex. Their observations emphasized the discrete multicellular patterns that are mediated through communication via gap junction. The functional domains were 50e120 mm in diameter on tangential slices; they spanned several cortical layers and resembled columns found in the adult cortex in coronal slices. In developing somatosensory cortex, domains were smaller than, and distinct from, the barrels. Gap junctions coupled the neurons within each domain. Gap junction domains persisted after blockade of sodium- and calcium-dependent action potentials, suggesting that they may promote metabolic- rather than activity-related assemblies (Kandler and Katz, 1998). There are modules and columns within the developing cortex that are present only transiently during development but not in the adult. Numerous stains are transient during barrel development (Erzurumlu et al., 1990; Mitrovic and Schachner,

122 PART | I Circuit development

1996). Ocular dominance columns are more apparent during development or upon visual deprivation than in normal adult in the marmoset, and the presence or absence of ocular dominance columns has been debated in the adult marmoset (Spatz, 1989; Chappert-Piquemal et al., 2001). There are transient circuits that show a pattern that is transiently related to the (sensory) periphery of the vibrissae in the barrel cortex in the mouse. This involves the neurites of the early-generated and largely transient subplate neurons (Piñon et al., 2009). These changes in cortical patterning may reflect the development of synaptic integration that will provide coherent activity among groups of target cells, but it has been questioned whether the observed patterns themselves have any functional relevance (Purves et al., 1992). Induction of visible and distinguishable barrel patterns or of ocular dominance columns has not been linked to a particular sensory or motor capacity (Purves et al., 1992; Horton and Adams, 2005). Prior to birth, monocular transduction pathways are already established through a process known as Hebbian learning. Spontaneous retinal activity in one eye of the developing fetus leads to neuronal depolarization (Galli and Maffei, 1988) that can propagate through the thalamus (Mooney et al., 1996). Synapses that receive multiple inputs are more likely to propagate the signal, whereas errant connections will not be sufficient to trigger another action potential. If glutamate has been released by the presynaptic axon terminal, postsynaptic neurons that depolarize become permeable to calcium ions. Calcium entry leads to a chemical process that strengthens the synapse, making it more likely to survive than other connections. Although orientation columns can develop without any externally elicited sensory visual input (before birth), their maintenance relies on postnatal sensory driven visual activity (Crair et al., 1998).

5.5.3 The way forward (i) There are several problems associated to our current nomenclature of columns. The concept of a “universal cortical column” is very captivating in anatomical, physiological, and developmental models of the cerebral cortex, and this is reflected in the current terminology in that this aims to gloss over differences rather than expose them. There is overwhelming evidence for various forms of radial organization. However, there are various types of cortical columns, and we shall need to specify these as we get better understanding of their properties. The term “column” might be modified (e.g., “cellular column,” “afferent column”) or all together abandoned when we have more information about types of cortical circuits and the range of their operations. We stick to the term column, because we want to stick to the captivating concept, and for the time being, there is no easy alternative. We have to establish more specific terminology that will allow specific reference for particular entities. As the cell types of the cerebral cortex become better characterized morphologically, chemically, and physiologically, the details of the types of connections and circuits that they establish with one another within the cortex will be better understood. (ii) We now know that columns/modules are characteristic of neocortex, but there is no single structure or function that is the common building block of all cortical areas in all mammals. The observations made in the barrels in S1 and the ocular dominance columns in V1 had an enormous influence on the way we look at the cerebral cortex as a whole; however, it is increasingly apparent that we cannot generalize to all regions. Horton and Adams write: “At some point, one must abandon the idea that columns are the basic functional entity of the cortex. It now seems doubtful that any single, transcendent principle endows the cerebral cortex with a modular structure. Each individual area is constructed differently, and each will need to be taken apart cell by cell, layer by layer, circuit by circuit and projection by projection to describe fully the architecture of the cortex” (Horton and Adams, 2005). Under the influence of new data, our concept will also gradually change, and we may finally let go of the elusive idea of a “universal cortical unit” that is extending radially as a homogeneous building block in all areas in all mammals. (iii) The previous finding of constant cell numbers within an arbitrary unit column and the homogeneous structure of the column is not supported by recent observations. On the contrary, it is apparent that cortical areas exhibit huge differences in cell composition, cell numbers, and connectivity. (iv) The possible functions specifically associated with the presence or absence of a particular column (e.g., ocular dominance columns, barrels) is not clear. Comparative studies have yet to identify a specific sensory, motor, or cognitive function that is specifically associated with a particular form of cortical column. Purves et al. (1992) postulated that the columnar patterns arise because they are “incidental consequence of the rules of synapse formation”. (v) There is overwhelming evidence for early columnar allocation of the developing pyramidal neurons. It is also evident that at early developmental stages, an early organization has already been specified, before activity-dependent mechanisms could take place. The link between the clonally related neuronal assemblies and future modules of the cortex is not yet clear, but with the current repertoire of methodologies, these issues can now be addressed.

Cortical columns Chapter | 5

123

Acknowledgments We are grateful to Ray Guillery, Tony Rockel, Anna Hoerder-Suabedissen, and Andrá Marques-Smith for discussions and comments. This chapter is the updated version of Molnár (2012).

References Abeles, M., Goldstein Jr., M.H., 1970. Functional architecture in cat primary auditory cortex: columnar organization and organization according to depth. J. Neurophysiol. 33 (1), 172e187. Adams, D.L., Horton, J.C., 2003. Capricious expression of cortical columns in the primate brain. Nat. Neurosci. 6, 113e114. Adams, D.L., Sincich, L.C., Horton, J.C., 2007. Complete pattern of ocular dominance columns in human primary visual cortex. J. Neurosci. 27 (39), 10391e10403. Alloway, K.D., 2008. Information processing streams in rodent barrel cortex: the differential functions of barrel and septal circuits. Cerebr. Cortex 18, 979e989. Bandler, R., Shipley, M.T., 1994. Columnar organization in the midbrain periaqueductal gray: modules for emotional expression? Trends Neurosci. 17, 379e389. Blasdel, G.G., Lund, J.S., 1983. Termination of afferent axons in macaque striate cortex. J. Neurosci. 3, 1389e1413. Blasdel, G.G., Salama, G., 1986. Voltage-sensitive dyes reveal a modular organization in monkey striate cortex. Nature 321 (6070), 579e585. Bock, D.D., Lee, W.C., Kerlin, A.M., Andermann, M.L., Hood, G., Wetzel, A.W., Yurgenson, S., Soucy, E.R., Kim, H.S., Reid, R.C., 2011. Network anatomy and in vivo physiology of visual cortical neurons. Nature 471 (7337), 177e182. Bonhoeffer, T., Grinvald, A., 1991. Iso-orientation domains in cat visual cortex are arranged in pinwheel-like patterns. Nature 353 (6343), 429e431. Brodmann, K., 1909. Vergleichende Lokalizationslehre der Grosshirnrinde in ihren Prinzipien dargestellt auf Grund des Zellenbaues. Barth, Leipzig. Brown, S.P., Hestrin, S., 2009. Intracortical circuits of pyramidal neurons reflect their long-range axonal targets. Nature 457, 1133e1137. Cajal, S.R.y., 1909. 1909e1911. Histologie du système nerveux de l’homme et des vertébrés, vol. 2, Translated by Azoulay, L., Maloine, A., Paris. English translation by Swanson, N., Swanson, L.W., 1995. Histology of the Nervous System of Man and Vertebrates, vol. 2. Oxford Univ. Press, New York. Catania, K.C., Kaas, J.H., 1995. Organization of the somatosensory cortex of the star-nosed mole. J. Comp. Neurol. 351 (4), 549e567. Chappert-Piquemal, C., Fonta, C., Malecaze, F., Imbert, M., 2001. Ocular dominance columns in the adult new World monkey Callithrix jacchus. Vis. Neurosci. 18 (3), 407e412. Cheung, A.F., Pollen, A.A., Tavare, A., DeProto, J., Molnár, Z., 2007. Comparative aspects of cortical neurogenesis in vertebrates. J. Anat. 211 (2), 164e176. Cheung, A.F., Kondo, S., Abdel-Mannan, O., Chodroff, R.A., Sirey, T.M., Bluy, L.E., Webber, N., DeProto, J., Karlen, S.J., Krubitzer, L., Stolp, H.B., Saunders, N.R., Molnár, Z., 2010. The subventricular zone is the developmental milestone of a 6-layered neocortex: comparisons in metatherian and eutherian mammals. Cerebr. Cortex 20 (5), 1071e1081. Collins, C.E., Airey, D.C., Young, N.A., Leitch, D.B., Kaas, J.H., 2010. Neuron densities vary across and within cortical areas in primates. Proc. Natl. Acad. Sci. U.S.A. 107 (36), 15927e15932. Costa, M.R., Hedin-Pereira, C., 2010. Does cell lineage in the developing cerebral cortex contribute to its columnar organization? Front. Neuroanat. 4, 26. Costa, M.R., Bucholz, O., Schroeder, T., Götz, M., 2009. Late origin of glia-restricted progenitors in the developing mouse cerebral cortex. Cerebr. Cortex (Suppl. 1), i135ei143. Crair, M.C., Gillespie, D.C., Stryker, M.P., 1998. The role of visual experience in the development of columns in cat visual cortex. Science 279 (5350), 566e570. Eccles, J.C., Ito, M., Szentágothai, J., 1968. Berlin, Heidelberg, New York: Springer-Verlag. 1967. Pp. 335. 62s. Exp. Physiol. 53 (3), 337e338. Erzurumlu, R.S., Kind, P.C., 2001. Neural activity: sculptor of ’barrels’ in the neocortex. Trends Neurosci. 24 (10), 589e595. Erzurumlu, R.S., Jhaveri, S., Benowitz, L.I., 1990. Transient patterns of GAP-43 expression during the formation of barrels in the rat somatosensory cortex. J. Comp. Neurol. 292 (3), 443e456. Escobar, M.I., Pimienta, H., Caviness Jr., V.S., Jacobson, M., Crandall, J.E., Kosik, K.S., 1986. Architecture of apical dendrites in the murine neocortex: dual apical dendritic systems. Neuroscience 17, 975e989. Fleischhauer, K., Petsche, H., Wittkowski, W., 1972. Vertical bundles of dendrites in the neocortex. Z. Anat. Entwicklungsgesch. 136, 213e223. Freund, T.F., Martin, K.A., Soltesz, I., Somogyi, P., Whitteridge, D., 1989. Arborisation pattern and postsynaptic targets of physiologically identified thalamocortical afferents in striate cortex of the macaque monkey. J. Comp. Neurol. 289, 315e336. Gabbott, P.L., 2003. Radial organisation of neurons and dendrites in human cortical areas 25, 32, and 320 . Brain Res. 992, 298e304. Galli, L., Maffei, L., 1988. Spontaneous impulse activity of rat retinal ganglion cells in prenatal life. Science 242 (4875), 90e91. Graybiel, A.M., Ragsdale, C.W., 1978. Histochemically distinct compartments in the striatum of humans, monkeys, and cat demonstrated by acetylthiocholinesterase staining. Proc. Natl. Acad. Sci. U.S.A. 75, 5723e5726. Harris, R.M., Woolsey, T.A., 1979. Morphology of Golgi-impregnated neurons in mouse cortical barrels following vibrissae damage at different post-natal ages. Brain Res. 161, 143e149. Harting, J.K., Updyke, B.V., Lieshout, D.P.V., 1992. Corticotectal projections in the cat: anterograde transport studies of twenty-five cortical areas. J. Comp. Neurol. 324, 379e414.

124 PART | I Circuit development

Helmstaedter, M., de Kock, C.P.J., Feldmeyer, D., Bruno, R.M., Sakmann, B., 2007. Reconstruction of an average cortical column in silico. Brain Res. Rev. 55, 193e203. Herculano-Houzel, S., Collins, C.E., Wong, P., Kaas, J.H., Lent, R., 2008. The basic nonuniformity of the cerebral cortex. Proc. Natl. Acad. Sci. U.S.A. 105 (34), 12593e12598. Hoerder-Suabedissen, A., Molnár, Z., July 28, 2011. The morphology of mouse subplate cells with identified projection targets changes with age. J. Comp. Neurol. https://doi.org/10.1002/cne.22725. Horton, J.C., Adams, D.L., 2005. The cortical column: a structure without a function. Philos. Trans. R. Soc. Lond. B Biol. Sci. 360 (1456), 837e862. Horton, J.C., Hocking, D.R., 1998. Monocular core zones and binocular border stripes in primate striate cortex revealed by contrasting effects of enucleation, eyelid suture, and retinal laser lesions on cytochrome oxidase activity. J. Neurosci. 15, 5433e5455. Hubel, D.H., Wiesel, T.N., 1968. Receptive fields and functional architecture of monkey striate cortex. J. Physiol. 195 (1), 215e243. Hubel, D.H., Wiesel, T.N., 1969. Anatomical demonstration of columns in the monkey striate cortex. Nature 221, 747e750. Hubel, D.H., Wiesel, T.N., 1977. Functional architecture of macaque monkey visual cortex. Proc. R. Soc. Lond. Ser. B 198, 1e59. Hubel, D.H., Wiesel, T.N., Stryker, M.P., 1977. Orientation columns in macaque monkey visual cortex demonstrated by the 2-deoxyglucose autoradiographic technique. Nature 269 (5626), 328e330. Hubel, D.H., Wiesel, T.N., Stryker, M.P., 1978. Anatomical demonstration of orientation columns in macaque monkey. J. Comp. Neurol. 177 (3), 361e380. Ichinohe, N., Rockland, K.S., 2004. Region specific micromodularity in the uppermost layers in primate cerebral cortex. Cerebr. Cortex 14, 1173e1184. Illing, R.-B., Graybiel, A.M., 1986. Complementary and non-matching afferent compartments in the cat’s superior colliculus: innervation of the acetycholinesterase-poor domain of the intermediate zone. Neuroscience 18, 373e394. Inan, M., Crair, M.C., 2007. Development of cortical maps: perspectives from the barrel cortex. Neuroscientist 13, 49e60. Innocenti, G.M., Vercelli, A., 2010. Dendritic bundles, minicolumns, columns, and cortical output units. Front. Neuroanat. 12 (4), 11. Jain, N., Catania, K.C., Kaas, J.H., 1998. A histologically visible representation of the fingers and palm in primate area 3b and its immutability following long-term deafferentations. Cerebr. Cortex 8 (3), 227e236. Jones, E.G., 1999. Making brain connections: neuroanatomy and the work of TPS Powell, 1923-1996. Annu. Rev. Neurosci. 22, 49e103. Jones, E.G., 2000. Microcolumns in the cerebral cortex. Proc. Natl. Acad. Sci. U.S.A. 97, 5019e5021. Jones, E.G., Rakic, P., 2010. Radial columns in cortical architecture: it is the composition that counts. Cerebr. Cortex 20 (10), 2261e2264. Epub 2010 Jul 28. Kaas, J.H., Gharbawie, O.A., Stepniewska, I., 2011. The organization and evolution of dorsal stream multisensory motor pathways in primates. Front. Neuroanat. 5, 34. Kandler, K., Katz, L.C., 1998. Coordination of neuronal activity in developing visual cortex by gap junction-mediated biochemical communication. J. Neurosci. 18 (4), 1419e1427. Kaschube, M., Schnabel, M., Löwel, S., Coppola, D.M., White, L.E., Wolf, F., 2010. Universality in the evolution of orientation columns in the visual cortex. Science 330 (6007), 1113e1116. Keay, K.A., Bandler, R., 2001. Parallel circuits mediating distinct emotional coping reactions to different kinds of stress. Neurosci. Biobehav. Rev. 25, 669e678. Kornack, D.R., Rakic, P., 1995. Radial and horizontal deployment of clonally related cells in the primate neocortex: relationship to distinct mitotic lineages. Neuron 15 (2), 311e321. Krieger, P., Kuner, T., Sakmann, B., 2007. Synaptic connections between layer 5B pyramidal neurons in mouse somatosensory cortex are independent of apical dendrite bundling. J. Neurosci. 27, 11473e11482. Larouche, M., Hawkes, R., 2006. From clusters to stripes: the developmental origins of adult cerebellar compartmentation. Cerebellum 5, 77e88. Lev, D.L., White, E.L., 1997. Organization of pyramidal cell apical dendrites and composition of dendritic clusters in the mouse: emphasis on primary motor cortex. Eur. J. Neurosci. 9, 280e290. LeVay, S., Hubel, D.H., Wiesel, T.N., 1975. The pattern of ocular dominance columns in macaque visual cortex revealed by a reduced silver stain. J. Comp. Neurol. 159 (4), 559e576. Livingstone, M.S., Hubel, D.H., 1982. Thalamic inputs to cytochrome oxidase-rich regions in monkey visual cortex. Proc. Natl. Acad. Sci. U.S.A. 79 (19), 6098e6101. Livingstone, M.S., Hubel, D.H., 1984. Specificity of intrinsic connections in primate primary visual cortex. J. Neurosci. 4 (11), 2830e2835. López-Bendito, G., Molnár, Z., 2003. Thalamocortical development: how are we going to get there? Nat. Rev. Neurosci. 4 (4), 276e289. Lorente de Nó, R., 1922. Corteza cerebral del ratón. I. Lacorteza acústica. In: Trabajos del Laboratorio de Investigaciones Biológicas, vol. 20, pp. 41e78. Lorente de Nó, R., 1949. Cerebral cortex. Architecture, intracortical connections, motor projection. In: Fulton, J.F. (Ed.), Physiology of the Nervous System. Oxford University Press, Oxford, pp. 274e313. Mana, S., Chevalier, G., 2001. Honeycomb-like structure of the intermediate layers of the rat superior colliculus: afferent and efferent connections. Neuroscience 103, 673e693. Maruoka, H., Nakagawa, N., Tsuruno, S., Sakai, S., Yoneda, T., Hosoya, T., 2017. Lattice system of functionally distinct cell types in the neocortex. Science 358 (6363), 610e615. Massing, W., Fleischhauer, K., 1973. Further observations on vertical bundles of dendrites in the cerebral cortex of the rabbit. Z. Anat. Entwicklungsgesch. 141, 115e123.

Cortical columns Chapter | 5

125

Meyer, H.S., Wimmer, V.C., Oberlaender, M., de Kock, C.P.J., Sakmann, B., Helmstaedter, M., 2010. Number and laminar distribution of neurons in a thalamocortical projection column of rat vibrissal cortex. Cerebr. Cortex 20 (10), 2277e2286. https://doi.org/10.1093/cercor/bhq067. Epub 2010 Jun 9. Mitchison, G., 1992. Axonal trees and cortical architecture. Trends Neurosci. 15 (4), 122e126. Mitrovic, N., Schachner, M., 1996. Transient expression of NADPH diaphorase activity in the mouse whisker to barrel field pathway. J. Neurocytol. 25 (8), 429e437. Miyashita, T., Wintzer, M., Kurotani, T., Konishi, T., Ichinohe, N., Rockland, K.S., 2010. Neurotrophin-3 is involved in the formation of apical dendritic bundles in cortical layer 2 of the rat. Cerebr. Cortex 20, 229e240. Molnár, Z., 2011. The need for research on human brain development. Brain 34 (7), 2177e2185. Molnár, Z., 2012. Cortical columns (Chapter 137). In: Rakic, P., Rubenstein, J. (Eds.), Neural Circuit Development and Function in the Healthy and Diseased Brain: Comprehensive Developmental Neuroscience, vol. 3. Molnár, Z., Molnár, E., 2006. Calcium and NeuroD2 control the development of thalamocortical communication. Neuron 49 (5), 639e642.  Molnár, Z., Clowry, G.J., Sestan, N., Alzu’bi, A., Bakken, T., Hevner, R.F., Hüppi, P.S., Kostovic, I., Rakic, P., Anton, E.S., Edwards, D., Garcez, P., Hoerder-Suabedissen, A., Kriegstein, A., 2019. New insights into the development of the human cerebral cortex. J. Anat. https://doi.org/10.1111/ joa.13055. (Epub ahead of print). Mooney, R., Penn, A.A., Gallego, R., Shatz, C.J., 1996. Thalamic relay of spontaneous retinal activity prior to vision. Neuron 17 (5), 863e874. Mountcastle, V.B., 1957. Modality and topographic properties of single neurons of cat’s somatic sensory cortex. J. Neurophysiol. 20 (4), 408e434. Mountcastle, V.B., 1997. The columnar organization of the neocortex. Brain 120 (Pt 4), 701e722. Mountcastle, V.B., Davies, P.W., Berman, A.L., 1957. Response properties of neurons of cat’s somatic sensory cortex to peripheral stimuli. J. Neurophysiol. 20 (4), 374e407. Nakagawa, N., Plestant, C., Yabuno-Nakagawa, K., Li, J., Lee, J., Huang, C.W., Lee, A., Krupa, O., Adhikari, A., Thompson, S., Rhynes, T., Arevalo, V., Stein, J.L., Molnár, Z., Badache, A., Anton, E.S., 2019. Memo1-Mediated tiling of radial glial cells facilitates cerebral cortical development. Neuron. https://doi.org/10.1016/j.neuron.2019.05.049 pii: S0896-6273(19)30525-2. (Epub ahead of print). Noctor, S.C., et al., 2001. Neurons derived from radial glial cells establish radial units in neocortex. Nature 409, 714e720. Otsuka, T., Kawaguchi, Y., 2008. Firing-pattern-dependent specificity of cortical excitatory feed-forward subnetworks. J. Neurosci. 28 (44), 11186e11195. Peters, A., 1997. Pyramidal cell modules in cerebral cortex. In: Motta, P.M. (Ed.), Recent Advances in Microscopy of Cells, Tissues and Organs. A. Delfino, Rome, pp. 187e192. Peters, A., Kara, D.A., 1987. The neuronal composition of area 17 of rat visual cortex. IV. The organization of pyramidal cells. J. Comp. Neurol. 260, 573e590. Peters, A., Sethares, C., 1996. Myelinated axons and the pyramidal cell modules in monkey primary visual cortex. J. Comp. Neurol. 365, 232e255. Peters, A., Walsh, M.A., 1972. Study of the organization of apical dendrites in the somatic sensory cortex of the rat. J. Comp. Neurol. 144, 253e268. Piñon, M.C., Jethwa, A., Jacobs, E., Campagnoni, A., Molnár, Z., 2009. Dynamic integration of subplate neurons into the cortical barrel field circuitry during postnatal development in the Golli-tau-eGFP (GTE) mouse. J. Physiol. 587 (Pt 9), 1903e1915. Powell, T.P., Hendrickson, A.E., 1981. Similarity in number of neurons through the depth of the cortex in the binocular and monocular parts of area 17 of the monkey. Brain Res. 216 (2), 409e413. Powell, T.P., Mountcastle, V.B., 1959. Some aspects of the functional organization of the cortex of the postcentral gyrus of the monkey: a correlation of findings obtained in a single unit analysis with cytoarchitecture. Bull. Johns Hopkins Hosp. 105, 133e162. Purves, D., Riddle, D.R., LaMantia, A.S., 1992. Iterated patterns of brain circuitry (or how the cortex gets its spots). Trends Neurosci. 15 (10), 362e368. Rakic, P., 1988. Specification of cerebral cortical areas. Science 241, 170e176. Rakic, P., 2007. The radial edifice of cortical architecture: from neuronal silhouettes to genetic engineering. Brain Res. Rev. 55, 204e219. Rakic, P., 2009. Evolution of the neocortex: a perspective from developmental biology. Nat. Rev. Neurosci. 10 (10), 724e735. Rakic, P., Caviness Jr., V.S., 1995. Cortical development: view from neurological mutants two decades later. Neuron 14 (6), 1101e1104. Rausell, E., Jones, E.G., 1991. Chemically distinct compartments of the thalamic VPM nucleus in monkeys relay principal and spinal trigeminal pathways to different layers of the somatosensory cortex. J. Neurosci. 11, 226e237. Rice, F.L., 1995. Comparative aspects of barrel structure and development. In: Peters, A., Jones, E.G., Diamond, I.T. (Eds.), Cerebral Cortex: The Barrel Cortex of Rodents, vol. 11. Plenum Press, New York, pp. 1e76. Risold, P.Y., Swanson, L.W., 1998. Connections of the rat lateral septal complex. Brain Res. Rev. 24, 115e195. Rockel, A.J., Hiorns, R.W., Powell, T.P., 1974. Numbers of neurons through full depth of neocortex. J. Anat. 118 (Pt 2), 371. Rockel, A.J., Hiorns, R.W., Powell, T.P., 1980. The basic uniformity in structure of the neocortex. Brain 103 (2), 221e244. Rockland, K.S., 2010. Five points on columns. Front. Neuroanat. 4, 22. Rockland, K.S., Ichinohe, N., 2004. Some thoughts on cortical minicolumns. Exp. Brain Res. 158, 265e277. Roney, K.J., Scheibel, A.B., Shaw, G.L., 1979. Dendritic bundles: survey of anatomical experiments and physiological theories. Brain Res. 180 (2), 225e271. Sato, T.R., Gray, N.W., Mainen, Z.F., Svoboda, K., 2007. The functional microarchitecture of the mouse barrel cortex. PLoS Biol. 5, e189. https://doi.org/ 10.1371/journal.pbio.0050189. Shepherd, G., Grillner, S., 2010. Handbook of Brain Microcircuits. Oxford University Press, New York, p. 544.

126 PART | I Circuit development

Sillitoe, R.V., Joyner, A.L., 2007. Morphology, molecular codes, and circuitry produce the three-dimensional complexity of the cerebellum. Annu. Rev. Cell Dev. Biol. 23, 549e577. Song, S., Sjöström, P.J., Reigl, M., Nelson, S., Chklovskii, D.B., 2005. Highly nonrandom features of synaptic connectivity in local cortical circuits. PLoS Biol. 3 (3), e68. Epub 2005 Mar 1. Spatz, W.B., 1989. Loss of ocular dominance columns with maturity in the monkey, Callithrix jacchus. Brain Res. 488 (1e2), 376e380. Takahata, T., Higo, N., Kaas, J.H., Yamamori, T., 2009. Expression of immediate-early genes reveals functional compartments within ocular dominance columns after brief monocular inactivation. Proc. Natl. Acad. Sci. U.S.A. 106, 12151e12155. Tan, S.S., Breen, S., 1993. Radial mosaicism and tangential cell dispersion both contribute to mouse neocortical development. Nature 362, 638e640. Torii, M., Hashimoto-Torii, K., Levitt, P., Rakic, P., 2009. Integration of neuronal clones in the radial cortical columns by EphA and ephrin-A signalling. Nature 461 (7263), 524e528. Van der Loos, H., Woolsey, T.A., 1973. Somatosensory cortex: structural alterations following early injury to sense organs. Science 179 (71), 395e398. Vercelli, A., Garbossa, D., Curtetti, R., Innocenti, G.M., 2004. Somatodendritic minicolumns of output neurons in the rat visual cortex. Eur. J. Neurosci. 20, 495e502. von Economo, C., Koskinas, G.N., 2008. In: Triarhou, L.C. (Ed.), Atlas of the Cytoarchitectonics of the Adult Human Cerebral Cortex. Karger, Basel. Voogd, J., Glickstein, M., 1998. The anatomy of the cerebellum. Trends Neurosci. 21, 370e375. Wang, Y., Brzozowska-Prechtl, A., Karten, H.J., 2010. Laminar and columnar auditory cortex in avian brain. Proc. Natl. Acad. Sci. U.S.A. 107 (28), 12676e12681. Welker, W.I., Johnson Jr., J.I., Pubols Jr., B.H., 1964. Some morphological and physiological characteristics of the somatic sensory system in raccoons. Am. Zool. 4, 75e94. Wimmer, V.C., Bruno, R.M., de Kock, C.P.J., Kuner, T., Sakmann, B., 2010. Dimensions of a projection column and architecture of VPM and POm axons in rat vibrissal cortex. Cerebr. Cortex 20 (10), 2265e2276. Woolsey, T.A., Van der Loos, H., 1970. The structural organization of layer IV in the somatosensory region (SI) of mouse cerebral cortex. The description of a cortical field composed of discrete cytoarchitectonic units. Brain Res. 17 (2), 205e242. Woolsey, T.A., Welker, C., Schwartz, R.H., 1975. Comparative anatomical studies of the SmL face cortex with special reference to the occurrence of “barrels” in layer IV. J. Comp. Neurol. 164 (1), 79e94. Yoshimura, Y., Dantzker, J.L., Callaway, E.M., 2005. Excitatory cortical neurons form fine-scale functional networks. Nature 433 (7028), 868e873. Yu, Y.C., Bultje, R.S., Wang, X., Shi, S.H., 2009. Specific synapses develop preferentially among sister excitatory neurons in the neocortex. Nature 458, 501e504. Yuste, R., Peinado, A., Katz, L.C., 1992. Neuronal domains in developing neocortex. Science 257 (5070), 665e669.

Chapter 6

Spike timingedependent plasticity Daniel E. Feldman Department of Molecular & Cell Biology, Helen Wills Neuroscience Institute, UC Berkeley, Berkeley, CA, USA

Chapter outline 6.1. Introduction: synaptic plasticity and synaptic learning rules 6.2. Discovery of STDP 6.3. Definition and forms of STDP 6.3.1. Hebbian STDP 6.3.2. Anti-Hebbian STDP 6.4. STDP is part of a broader, multifactor plasticity rule 6.5. Functional properties of STDP 6.5.1. Properties of Hebbian STDP 6.5.2. Properties of anti-Hebbian STDP 6.5.3. STDP and circuit homeostasis

127 128 128 128 129 130 131 131 131 132

6.5.4. Neuromodulation of STDP 6.6. Cellular mechanisms for STDP 6.6.1. Mechanisms for LTP and LTD components of STDP 6.6.2. Dendritic excitability and STDP 6.7. Is STDP a realistic learning rule in vivo? 6.8. How does STDP contribute to development of neural circuits? 6.9. How does STDP contribute to adult plasticity and learning? 6.10. Summary: role of spike timing in synaptic plasticity References

132 132 132 134 135 135 136 137 138

6.1 Introduction: synaptic plasticity and synaptic learning rules Synaptic plasticity is the principal mechanism for information storage and learning in the brain. It also plays a critical role in use-dependent refinement of neural circuits during development. Together, these processes build circuits that are shaped by an individual’s experience, and store sensory, motor, declarative, and emotional memories. Understanding synaptic plasticity has long been a focus of neuroscience research. Long-term associative synaptic plasticity was proposed as a basis for learning and memory in the 19th century (James, 1890). The modern formulation of this hypothesis emerged in 1949, when Hebb famously proposed that when a cell (A) contributes reliably to spiking of a postsynaptic cell (B), the strength of the A / B synapse is increased (Hebb, 1949). This synaptic change was proposed to store associative memories in neural networks. In the 1970s, this idea was extended to include weakening of ineffective synapses (von der Malsburg, 1973; Stent, 1973). An intensive search ensued for physiological evidence of such plasticity at synapses. This led to the discovery in the 1970se90s of long-term potentiation (LTP) and depression (LTD), which implement these rules at many synapses. Substantial evidence now implicates LTP and LTD as critical neurobiological mechanisms for memory and for activitydependent development. How does neural activity drive induction of LTP and LTD at appropriate synapses? What neural activity patterns are critical? The mathematical relationship is encapsulated in synaptic learning rules, which describe the induction rules for plasticity as a relationship between neural activity patterns and the resulting change in synapse strength. Synaptic learning rules are determined by the underlying cellular mechanisms for plasticity. They are useful because they predict the functional consequences of plasticity, including its potential role in learning, and can be incorporated into predictive models. Brain-inspired learning rules are widely used outside of the neuroscience research domain in neural networks for artificial intelligence, image analysis, and other applications. The first learning rules to describe LTP and LTD induction were based on presynaptic firing frequency. Early experiments showed that high- and low-frequency presynaptic stimulation drive LTP and LTD, respectively (Bliss and Lomo, 1973; Dudek and Bear, 1992; Mulkey and Malenka, 1992). This is quantified within the Bienenstock-CooperMunro (BCM) learning rule (Bienenstock et al., 1982). Subsequent experiments showed that the critical parameter was the

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00006-8 Copyright © 2020 Elsevier Inc. All rights reserved.

127

128 PART | I Circuit development

temporal correlation between presynaptic spiking and postsynaptic depolarization, with strong depolarization driving LTP, and weaker depolarization driving LTD (Artola and Singer, 1993). This is a direct consequence of the joint glutamate- and voltage-dependence of postsynaptic NMDA receptors, which provide calcium to trigger LTP and LTD. Most early studies suggested that pre- and postsynaptic activity had to co-occur within w
100 ms to drive plasticity (Baranyi and Feher, 1981; Gustafsson et al., 1987). This idea that LTP and LTD depend on correlated pre- and postsynaptic activity in a relatively broad window, without regard for precise spike timing within that window, is referred to as correlationdependent plasticity (CDP). Some early experiments noticed a relationship between LTP or LTD and the precise order of pre- and postsynaptic spikes within this broad window. In particular, LTP occurred when presynaptic inputs led or were synchronous with postsynaptic spikes, and LTD occurred when presynaptic input followed postsynaptic spikes (Levy and Steward, 1983; Debanne et al., 1994, 1997). Over the next few years, this finding was replicated at other synapses. The induction of LTP and LTD was often found to be critically dependent not only on the order, but also the relative timing of single spikes, down to the millisecond time scale (Debanne et al., 1994, 1997; Bell et al., 1997; Markram et al., 1997; Bi and Poo, 1998). This precise time- and order-dependence is fundamentally distinct from CDP, in which the temporal requirement for preand postsynaptic coactivation was coarse. This precise, spike-timing dependent form of LTP and LTD became known as spike timingedependent plasticity, or STDP. This chapter reviews STDP, its properties, its mechanisms, and its role in brain development and adult learning in vivo. There are several forms of STDP, and evidence suggests that these constitute an important class of synaptic plasticity rules in the brain. At the cellular mechanistic level, STDP uses overlapping mechanisms with CDP. And at many synapses, a mixture of both learning rules holds, suggesting that STDP may be best described as the spike timingedependent component of a broader “multi-factor” plasticity rule that includes firing rate and spike timing, as well as other factors such as postsynaptic depolarization and neuromodulation. Several recent reviews discuss STDP and its relationship to CDP (Caporale and Dan, 2008; Sjöström et al., 2008; Froemke et al., 2010; Feldman, 2012).

6.2 Discovery of STDP As introduced above, most early studies indicated a correlation requirement of about
100 ms for plasticity (Baranyi and Feher, 1981; Gustafsson et al., 1987). Levy and Steward (1983) first noted that presynaptic input occurring just prior to postsynaptic spikes induced LTP, while the opposite order evoked LTD. Debanne and colleagues showed clearly that postleading-pre spike order generated LTD (Debanne et al., 1994, 1997). In 1997, Markram et al. imposed precise pre- and postsynaptic spike timing using dual whole-cell recording, and discovered that both the sign and magnitude of LTP and LTD depended on the order and timing of pre- and postsynaptic spikes on the 10 ms time scale (Markram et al., 1997). Several groups made similar discoveries at the same time (Bell et al., 1997; Debanne et al., 1997; Bi and Poo, 1998). Bi and Poo (1998) first characterized the plasticity rule as a function of pre-post time delay with millisecond-scale resolution, which demonstrated its precise time dependence. Confirmation at many other synapses followed. The phenomenon was named “spike timingedependent plasticity” (STDP) by Song et al. (2000). For a comprehensive history of STDP, see Markram et al. (2011).

6.3 Definition and forms of STDP STDP occurs in several different forms. In canonical STDP, first defined at excitatory synapses onto neocortical and hippocampal pyramidal cells, pre-leading-post spiking by 0 to 10e20 ms induces LTP. LTD is induced by post-leadingpre spiking within a short interval (0e20 ms) or a long interval (0e100 ms), depending on the synapse (e.g., Markram et al., 1997; Bi and Poo, 1998; Feldman, 2000; Froemke and Dan, 2002). Thus, this canonical STDP is bidirectional and order-dependent. It is also called Hebbian STDP, which is defined below. These basic properties are illustrated in Fig. 6.1. It is also generally thought to be homosynaptic (that is, specific to the synapse(s) that experienced pre-post spike pairing), although this may not be strictly true. The term “STDP” is also used more broadly to describe additional forms of LTD and LTD that depend on precise spike timing, but are not bidirectional or order-dependentdfor example, that may contain only LTD. The most common forms are illustrated in Fig. 6.2. This review only considers STDP at excitatory synapses, because inhibitory synapses have been less studied.

6.3.1 Hebbian STDP Hebbian STDP is the canonical form of STDP in which LTP occurs when presynaptic spikes precede postsynaptic spikes by w0 to 20 ms, while LTD is induced when post leads pre by w0 to 20e100 ms (Fig. 6.2). This is distinct from classical

Spike timingedependent plasticity Chapter | 6

129

FIGURE 6.1 Basics of SDTP. (A) STDP is induced when a presynaptic spike, and the EPSP it induces in the postsynaptic cell, interact with a postsynaptic spike backpropagating through the dendrites (bAP). This results in a long-term change in the strength of the specific synapse(s) that experienced pre- and postsynaptic spiking (gray). (B) In canonical STDP, pairing of a presynaptic spike 10 ms prior to a postsynaptic spike causes LTP. The reverse order causes LTD. In reality, the pairing must be repeated w10e100 times to induce measurable plasticity. (C) Learning rule for STDP. Each circle is the result obtained in one cell, tested at one pre-post spike delay (Dt). (C) Replotted from Bi, G.Q., Poo, M.M., 1998. Synaptic modifications in cultured hippocampal neurons: dependence on spike timing, synaptic strength, and postsynaptic cell type. J. Neurosci. 18, 10464e10472.

(A)

(B)

Hebbian STDP

(C)

high firing rate No Δ

low firing rate -100

-50

0

50

Long LTD window 100

-100 -50 post leads pre

Anti-Hebbian STDP Bimodal STDP

Short LTD window

LTD

Change in strength

LTP

Correlation-dependent plasticity

0

50 100 pre leads post

LTD only -100

-50

0

50

100

Pre-post spike timing (ms) FIGURE 6.2 Different forms of STDP. (A) In classical correlation-dependent plasticity, high presynaptic firing rate causes strong, correlated activity in both pre- and postsynaptic neurons, inducing LTP. Low presynaptic firing rate causes less postsynaptic depolarization, inducing LTD. Plasticity is determined by firing correlation on a coarse w100 ms time scale. (B) In Hebbian STDP, pre-leading-post spiking within 0e10 ms Dt evokes LTP, while post-leading-pre spiking within either a similar, short window (0e10 ms) causes LTD. Some synapses exhibit a longer temporal window for LTD induction (0 to 50e100 ms). (C) In anti-Hebbian STDP, the classical STDP rule is reversed. Some synapses exhibit bimodal anti-Hebbian STDP with both LTP and LTD components; but other synapses exhibit only the LTD component, with varying degrees of temporal precision. In (B) and (C), different curves show different subtypes of STDP.

CDP, which lacks the precise time- and order-dependence. Hebbian STDP occurs at many excitatory synapses onto neocortical and hippocampal pyramidal neurons (e.g., Markram et al., 1997; Bi and Poo, 1998; Feldman, 2000; Sjöström et al., 2001; Nevian and Sakmann, 2006), and also at many excitatory synapses in other areas including auditory brainstem, striatum, and cerebellum (e.g., Fino et al., 2008; Pawlak and Kerr, 2008; Sgritta et al., 2017 ). Within the family of Hebbian STDP, some synapses exhibit short LTD windows that have the same duration as the LTP window (e.g., Bi and Poo, 1998), while others exhibit long LTD windows that produce a net bias toward LTD (e.g., Debanne et al., 1998; Feldman, 2000; Sjöström et al., 2001; Froemke et al., 2005). The form of STDP that occurs at one synapse is not fixed, but can change as synapse mature (e.g., Itami and Kimura, 2012; Valtcheva et al., 2017). Hebbian STDP implements Hebb’s postulate by strengthening synapses whose activity is causal for (i.e., contributes to) postsynaptic spiking, and weakening noncausal synapses, defined as those that do not evoke spikes but are located on otherwise active postsynaptic cells (Abbott and Nelson, 2000; Paulsen and Sejnowski, 2000). In contrast, CDP only approximates Hebb’s rule by assuming that synapses whose activity is loosely correlated with postsynaptic activation (and may follow that activation, rather than help cause it) are potentiated. Hebbian STDP also has a variety of other computationally useful properties, which are discussed in Section 6.5.

6.3.2 Anti-Hebbian STDP Anti-Hebbian STDP has the opposite order dependence from Hebbian STDP, with pre-leading-post spike order driving LTD, and post-leading-pre spiking driving LTP (Fig. 6.2). This bimodal, anti-Hebbian STDP has been observed at a few

130 PART | I Circuit development

synapses including excitatory synapses onto medium spiny neurons and cholinergic interneurons in striatum, and onto L5 pyramids in somatosensory cortex in response to spike bursts. For citations to specific studies reporting STDP at different synapses, see Feldman (2012). At most synapses, however, anti-Hebbian STDP contains only the LTD component, and is referred to simply as anti-Hebbian LTD (Han et al., 2000; Requarth and Sawtell, 2011; Zhao and Tzounopoulos, 2011). This is often temporally asymmetric, with stronger LTD for pre-leading-post spike order. Anti-Hebbian LTD occurs most commonly at excitatory inputs onto GABAergic neurons, including in neocortex, cerebellum, and dorsal cochlear nucleus. It is expressed strongly at parallel fiber synapses onto Purkinje-like neurons in the electrosensory lobe of the electric fish, where it cooccurs with timing-independent LTP (Bell et al., 1997; Han et al., 2000). Classical parallel fiber-Purkinje cell LTD in cerebellum is anti-Hebbian, with maximal LTD when parallel fiber stimulation precedes postsynaptic spiking by 80e150 ms (Wang et al., 2000; Safo and Regehr, 2008). While these examples are all of excitatory synapses onto inhibitory neurons, anti-Hebbian LTD can also occur at excitatory synapses onto cortical pyramidal cells under some conditions.

6.4 STDP is part of a broader, multifactor plasticity rule

Bimodal STDP

LTD only

LTD

No Δ

LTP

Spike timing is not the sole factor governing plasticity induction in STDP rules. Rather, the magnitude of LTP and LTD depends jointly on spike timing, presynaptic firing rate, postsynaptic voltage, and synaptic cooperativity (Sjöström et al., 2001). This is clearly shown at unitary connections between cortical pyramidal cells, where Hebbian STDP is evident only when pre-post spike pairing takes place at moderate spike rates (10e20 Hz). Lower firing rates (30 Hz) induce LTP independent of spike timing (Markram et al., 1997; Sjöström et al., 2001; Wittenberg and Wang, 2006; Zilberter et al., 2009). Thus, Hebbian STDP operates in a permissive middle range of firing frequency, superimposed on the BCM plasticity function in which high firing rates drive LTP, and low firing rates drive LTD (Fig. 6.3). The major reason for this behavior is that the LTP component of STDP requires additional postsynaptic depolarization beyond that supplied by a single presynaptic spike and its associated excitatory postsynaptic potential (EPSP). Because unitary connections involve very few synapses, this extra depolarization must be supplied by temporal summation across EPSPs, which occurs at moderate firing rates. Alternatively, this depolarization can also be provided by cooperative activation of multiple nearby synapses, which allows

post leads -50

pre leads 0

50

post leads -50

0

50

pre leads

-50

0

50

-50

0

50

LTP only

No Δ

Change in strength

LTP

Pre-post spike timing (ms)

LTD

BCM learning rule post leads

low

high

Firing rate (Hz)

-50

pre leads 0

50

Pre-post spike timing (ms)

FIGURE 6.3 STDP is part of a multifactor plasticity rule. Lower left panel, the black curve shows the classical BCM learning rule for LTP or LTD induction as a function of presynaptic firing rate. High firing rates induce LTP (because they evoke strong postsynaptic depolarization and calcium signals), low-to-moderate firing rates induce LTD (because they evoke moderate depolarization and calcium signals), and very low firing rates do not induce plasticity. Top and right panels show how plasticity depends on pre-post spike timing for different ranges of firing rate, illustrated by the colored boxes and arrows. In a middle range of firing frequencies, standard bimodal (Hebbian) STDP occurs. At low firing rates, only LTD can be induced, and at high firing rates, LTP is induced irrespective of spike timing. Thus, plasticity reflects an interaction between pre-post spike timing and firing rate (and depolarization, not shown here).

Spike timingedependent plasticity Chapter | 6

131

Hebbian STDP to be induced at lower frequency (Sjöström et al., 2001; Stuart and Hausser, 2001; Sjöström and Hausser, 2006). Thus, STDP also depends on postsynaptic depolarization, firing rate, and synaptic cooperativity. The requirement for depolarization also means that dendritic inhibition powerfully regulates STDP by regulating local dendritic depolarization (Paille et al., 2013). In addition, STDP also depends on other factors including neuromodulators, which can gate and shape STDP when applied during pre-post spike pairing, or even long afterward (Seol et al., 2007; Pawlak and Kerr, 2008; Shen et al., 2008; Cassenaer and Laurent, 2012). Together, these findings demonstrate that STDP is a highly overlapping process with CDP. What we call STDP may best be described as the spike timingedependent component of a broader, multifactor plasticity mechanism that also mediates rate- and depolarization-dependent LTP and LTD (that is, CDP). At some synapses, the timing-dependent component is strong; at other synapses, or under different activity regimes, it is modest; and at some synapses, postsynaptic somatic spikes have no role in plasticity at all. Thus STDP is not a universal plasticity rule, but is an important functional component of plasticity at some synapses.

6.5 Functional properties of STDP STDP requires postsynaptic somatic spikes, and depends on pre- versus postsynaptic spike order and precise (10 ms scale) spike timing. In contrast, CDP is driven by local postsynaptic depolarization from any source, and it is the magnitude of this depolarization, not precise spike timing and order, that determines the sign of plasticity. CDP does not explicitly require postsynaptic somatic spikes, and the local depolarization can be generated either by very local integration of nearby synaptic potentials, branch-level dendritic spikes elicited by cooperative activity in nearby synapses, and/or somatic spikes that backpropagate into the region. In contrast, STDP explicitly depends on postsynaptic spikes that initiate in the soma and backpropagate through the dendritic tree to synapses, providing the associative signal for plasticity (Magee and Johnston, 1997) (schematized in Fig. 6.1A). As a result, STDP can store associations between widespread inputs that contribute to somatic spiking, while CDP will store associations between local synapses that can interact without causing somatic spikes. In addition to this general difference between STDP and CDP, the two major forms of STDP (Hebbian and antiHebbian) have different functional properties, which are discussed in the next two sections.

6.5.1 Properties of Hebbian STDP Hebbian STDP has powerful functional properties that have been defined theoretically and computationally. These are discussed in several reviews (Abbott and Nelson, 2000; Morrison et al., 2008; Clopath et al., 2010), and only basic properties are summarized here. As mentioned above, Hebbian STDP implements the exact causal nature of Hebb’s postulate by strengthening synapses whose activity leads postsynaptic spikes, and weakening synapses whose activity lags postsynaptic spikes, which represent ineffective synapses onto otherwise active neurons. The form of Hebbian STDP with a long LTD window exhibits an overall bias toward LTD that powerfully depresses inputs that are uncorrelated with postsynaptic spiking (Feldman, 2000). These properties enable Hebbian STDP to create neural ensembles to efficiently store associations in recurrent networks, and to build topographic maps and receptive fields based on temporal correlations in input activity. In addition, Hebbian STDP implements competition between noncorrelated convergent inputs, and supports learning of temporal sequences including the ability to predict future events from past stimuli (e.g., Mehta et al., 2000; Fiete et al., 2010). It also enforces synchronous spiking during feedforward network activation, which is a common feature in vivo. These functional properties are generally not predicted for CDP, but are robust for Hebbian STDP. Hebbian STDP can also mediate temporal difference learning and reinforcement learning, and can tune neurons for temporal features of input. As discussed above, Hebbian STDP is actually part of a broader multifactor plasticity rule. Theory has also shed light on the functional properties of this multifactor STDP rule. A phenomenological model of multifactor STDP (Clopath et al., 2010) (built on earlier work by Pfister and Gerstner, 2006) captures the joint timing, rate, and voltage-dependence of plasticity at individual synapses. On the network level, this learning rule stores information about both slow input correlations and rapid spatiotemporal sequences, depending on the structure of spike train input, thus capturing functional aspects of firing rateedependent plasticity and STDP (Clopath et al., 2010).

6.5.2 Properties of anti-Hebbian STDP Anti-Hebbian STDP detects functional sets of inputs that are anticorrelated with postsynaptic spiking, and can store precise information about these signals and their timing relative to a dominant, spike-eliciting input to the postsynaptic cell. In the

132 PART | I Circuit development

cerebellum-like electrosensory lobe of electric fish, the LTD component of this plasticity (anti-Hebbian LTD) stores negative images of predicted sensory input, so that novel (unexpected) sensory inputs can be better represented (Roberts and Bell, 2000; Requarth and Sawtell, 2011). Anti-Hebbian LTD at parallel fiber-Purkinje cell synapses in mammalian cerebellum may perform a similar computation. Anti-Hebbian LTD is also prominent in distal dendrites of cortical pyramidal cells, but its computational role is less understood.

6.5.3 STDP and circuit homeostasis CDP has a well-known drawback of positive feedback instability, in which initial LTP of some synapses raises firing rates, thus promoting further LTP and, eventually, runaway network activity. Computational models avoid this by adding metaplasticity or additional, homeostatic mechanisms. Hebbian STDP promotes a more stable distribution of synaptic weights (van Rossum et al., 2000), but other instabilities can persist, including firing rate instability (Zenke et al., 2013). Other, novel features of STDP may further stabilize network function. For example, STDP-LTP of synapses onto cortical pyramidal cells may drive heterosynaptic weakening of other nearby synapses, thus preserving total synaptic strength (El-Boustani et al., 2018). Alternatively, Hebbian STDP may shift toward an LTD-biased STDP rule during network “up” states prominent in slow-wave sleep. This could weaken all but the most effective synapses during sleep, preventing buildup of synaptic strength (Gonzalez-Rueda et al., 2018). Network activity may also be stabilized by STDP at inhibitory synapses. Cortical networks are often thought to operate in a “precisely balanced” state in which synaptic excitation and inhibition are co-tuned and temporally correlated, which enables rapid responses to input while maintaining global firing rate stability (van Vreeswijk and Sompolinsky, 1996; Brunel, 2000). How this precise balance is maintained is not known. Inhibitory synapses onto cortical and hippocampal pyramidal cells exhibit STDP in which near-synchronous pre- and postsynaptic spiking (irrespective of order) drive inhibitory LTP (Woodin et al., 2003; D’amour and Froemke, 2015). Modeling shows that this form of inhibitory STDP adjusts inhibition to precisely balance excitation and inhibition onto individual pyramidal cells, stabilize overall network activity, and enable realistic longterm memory storage (Vogels et al., 2011). This may explain how inhibition dynamically reorganizes to match excitation during cortical plasticity (Froemke et al., 2007). Other homeostatic plasticity mechanisms, including synaptic scaling or other inhibitory circuit plasticity, could also stabilize networks with excitatory STDP (Zenke et al., 2013).

6.5.4 Neuromodulation of STDP Hebbian synaptic plasticity, including Hebbian STDP, is a mechanism for unsupervised learning, in which a system learns information about the statistical properties of its input, but does not learn anything about which inputs are useful to guide behavior. The brain uses unsupervised learning widely, e.g., to learn sensory associations, or to predict sensory consequences of motor commands. But in many other cases, the brain uses reward and punishment signals to reinforce neural activity patterns that lead to optimal behavioral outcomes, or novelty signals to instruct the brain when to enable learning. These are supervised forms of learning, gated or driven by neuromodulatory signals, and cannot be explained by Hebbian plasticity alone. Neuromodulators are expected to interact with Hebbian and other local plasticity rules to mediate these processes. Neuromodulators including dopamine, norepinephrine, and acetylcholine, strongly affect STDP. In some cases, a neuromodulator is required to permissively enable STDP (e.g., Pawlak and Kerr, 2008; Fisher et al., 2017). In other cases, neuromodulators change the shape of STDP rules by selectively promoting LTP or LTD, or by altering timing dependence (e.g., Seol et al., 2007; Salgado et al., 2012). This may occur by direct modulation of plasticity machinery at the synapse, or indirectly by regulation of dendritic excitability. Reinforcement signals are often delayed by several seconds relative to the original synaptic activation, creating challenge in reinforcing the correct synapses. This is called the credit assignment problem. Remarkably, STDP in the striatum in vivo is enabled by dopamine and adenosine presented 1e2 s after pre-post spike pairing. This matches the expected delay to natural reinforcement and may solve the credit assignment problem (Brzosko et al., 2015; Fisher et al., 2017). For detailed review, see Fremaux and Gerstner (2015). Thus, neuromodulators are an additional factor within the broad multifactor plasticity rule, and can operate long after pre-post spiking to shape STDP.

6.6 Cellular mechanisms for STDP 6.6.1 Mechanisms for LTP and LTD components of STDP Within Hebbian STDP, LTP and LTD are mediated by three major signaling pathways, which are the same as for most classical, firing rate-, and correlation-dependent LTP and LTD. These are: (1) NMDA receptor (NMDAR)-dependent LTP,

Spike timingedependent plasticity Chapter | 6

NMDA-LTP

NMDA-LTD

pre leads post

post leads pre

coincidence pre detector

N

GLU

GLU

A Ca LTP

bAP (boosted when pre leads post)

CB1 eCB

GLU

A

depol.

VSCC

(C) NMDALTP

Pr

post

(B)

mGluR-CB1-LTD

N mGluR

Ca

Ca

VSCC

LTD

pre

IP3R

PLC

coincidence detector?

post

Ca

Change in strength

(A)

133

NMDALTD -50

0

50

NMDALTP

mGluRCB1-LTD

No STDP (local assoc. plasticity) Mostly LTD; STDP with associave boosng

Hebbian STDP

no effective bAP weak bAP

strong bAP

VSCC

bAP

-50

bAP

0

50

Pre-post spike ming (ms)

FIGURE 6.4 Cellular mechanisms for Hebbian STDP. (A) Major signaling pathways for plasticity. Left, NMDA-LTP and NMDA-LTD. NMDA receptors (N), AMPA receptors (A), voltage-sensitive calcium channels (VSCC), and glutamate (GLU) are shown. In STDP, the NMDA receptor is one major coincidence detector, generating strong calcium signals only when coactivated by presynaptic glutamate and strong postsynaptic depolarization. Dendritic amplification of the bAP also contributes to coincidence detection by boosting the bAP during pre-leads-post spike order (not shown). For postleading-pre spike order, this amplification is lost, and other processes can further suppress NMDA-mediated calcium signals, leading to LTD. Right, mGluR-CB1-LTD involves postsynaptic signaling and generation of eCB, which diffuses to activate presynaptic CB1 receptors, driving a long-term decrease in release probability. The coincidence detection mechanism may involve PLC, which requires mGluR activation and calcium to maximally generate eCB. (B) Contribution of different molecular forms of LTP and LTD to Hebbian STDP with short or long LTD windows. (C) Hypothesis for a spatial zone of STDP in pyramidal neurons. In a proximal zone, bAP propagation is strong and STDP is effective. At intermediate locations, bAPs are weak and can only drive LTD, unless local synaptic interactions generate additional depolarization to enable LTP. Most distally, bAPs are so small that they do not effectively drive plasticity, and LTP and LTD are induced instead by associative interactions between local synapses.

(2) NMDAR-dependent LTD, and (3) metabotropic glutamate receptor (mGluR)-dependent and/or cannabinoid type 1 receptor (CB1R)-dependent LTD (Fig. 6.4A). In classical NMDAR-LTP and NMDAR-LTD, the combination of presynaptic glutamate release and postsynaptic depolarization causes calcium influx through postsynaptic NMDARs. Voltage-sensitive calcium channels (VSCCs) also contribute calcium. Brief high calcium generates LTP, sustained moderate calcium generates LTD, and low calcium induces no plasticity (Lisman, 1989). Expression is primarily through addition or removal of postsynaptic AMPA receptors (AMPARs) and changes in AMPAR single-channel conductance, accompanied by structural changes in dendritic spines. Changes in presynaptic release probability may also occur. In mGluR- and/or CB1R-dependent LTD, postsynaptic NMDARs are not involved, and LTD is expressed via a decrease in presynaptic transmitter release. Of multiple forms, the most relevant is CB1R-dependent LTD, in which postsynaptic calcium and mGluR activation drive postsynaptic synthesis of an endocannabinoid (eCB). This diffuses retrogradely to activate CB1Rs on the glutamatergic presynaptic terminal and drive a long-lasting decrease in release probability (Chevaleyre et al., 2006). Other forms of mGluR-LTD are CB1R-independent and postsynaptically expressed, but are less linked to STDP. Hebbian STDP is mediated by these three mechanisms, with postsynaptic spikes providing a critical component of postsynaptic depolarization for plasticity. There are two major, biochemically distinct forms of Hebbian STDP. One is composed of NMDAR-dependent LTP and NMDAR-dependent LTD (Fig. 6.4B, top). Here, the magnitude of the NMDAR calcium signal (augmented by calcium from VSCCs) determines the sign of plasticity. With pre-leading-post spike order, the EPSP coincides with, and boosts, the bAP to produce a strong supralinear NMDAR calcium signal, while a post-leading-pre spike order triggers a weaker, sublinear calcium signal (Magee and Johnston, 1997; Koester and Sakmann, 1998; Nevian and Sakmann, 2006). The primary coincidence detector for pre-post spike time is the NMDA receptor, with additional contribution from nonlinear dendritic boosting of the bAP (see below). This form occurs at CA3-CA1 hippocampal synapses, and at some synapses onto neocortical pyramidal cells. A second form of Hebbian STDP is composed of NMDAR-dependent LTP and mGluR- and/or CB1R-dependent LTD (Fig. 6.4B, bottom). Here, postsynaptic NMDARs are required for spike timingedependent LTP, but not LTD. LTD instead requires postsynaptic group I mGluRs, low-threshold VSCCs, and calcium release from IP3 receptor-gated internal stores. Coincident activation of mGluRs and VSCCs synergistically activates phospholipase C (PLC) (Hashimotodani et al., 2005), leading to generation of endocannabinoid (eCB). Retrograde eCB signaling activates presynaptic CB1Rs, and LTD expression occurs by a decrease in presynaptic transmitter release probability. Thus, this form of STDP involves two separate coincidence detectors: NMDARs detect pre-leading-post spike intervals and exclusively trigger LTP, whereas a separate mechanism within the mGluR-VSCC-PLC-CB1 pathway detects post-leading-pre spike intervals and exclusively

134 PART | I Circuit development

triggers LTD (Bender et al., 2006; Nevian and Sakmann, 2006; Fino et al., 2010). How spike timing dependence arises for mGluR-CB1-LTD remains unclear, but may involve coincidence detection by PLC. This form of STDP occurs at several synapses onto L2/3 and L5 cortical pyramidal neurons, at cortical synapses onto striatal medium spiny neurons, and at juvenile CA3-CA1 synapses (Andrade-Talavera et al., 2016). There is a controversy about whether the LTD component also requires presynaptic NMDARs, or a nonionic signal originating from postsynaptic NMDARs, but this remains to be resolved (Rodriguez-Moreno and Paulsen, 2008; Carter and Jahr, 2016). Anti-Hebbian LTD can involve several different CB1R-dependent and mGluR-dependent mechanisms, and appears to have multiple forms. For example, anti-Hebbian LTD at excitatory synapses onto inhibitory cartwheel cells in the dorsal cochlear nucleus is presynaptic and CB1R-dependent (Tzounopoulos et al., 2007). Anti-Hebbian LTD at cerebellar parallel fiber-Purkinje cell synapses involves postsynaptic mGluRs, VSCCs, IP3Rs, and presynaptic CB1R activation, but is expressed postsynaptically by AMPAR internalization (Steinberg et al., 2006; Safo and Regehr, 2008). Strong evidence suggests that the order-dependent coincidence detector is the IP3 receptor, which is coactivated by PLC-produced IP3 and VSCC-derived cytosolic calcium (Nakamura et al., 1999; Wang et al., 2000; Sarkisov and Wang, 2008). Thus, the timing dependence of STDP reflects, in part, well-known molecular coincidence detectors within classical LTP and LTD signaling pathways, including NMDARs, PLC, and IP3Rs. These pathways also mediate firing rate- and correlation-dependent LTP and LTD at many synapses, explaining the convergence of STDP, firing rate, and depolarization within one multifactor learning rule.

6.6.2 Dendritic excitability and STDP The precise time dependence for STDP is also generated, in part, by the dynamics of electrical signaling in dendrites, caused by interactions between AMPAR-mediated EPSPs, NMDARs, and bAPs. The key associative signal for STDP is the backpropagation of somatic spikes from the axonal initiation site to the relevant synapses on the dendrites (Magee and Johnston, 1997). Depolarization from the bAP must summate with the local AMPA-EPSP to provide sufficient depolarization to activate NMDARs for STDP-LTP (Holbro et al., 2010). However, the backpropagation process is incomplete and context-dependent, which provides important constraints on STDP. bAPs propagate decrementally, losing half their amplitude within several hundred microns of the soma, and failing completely in distal branches (Spruston, 2008). This results in postsynaptic depolarization that is sufficient only for LTD, but not LTP, at distal synapses. Full STDP requires enhancement of bAP propagation or additional sources of depolarization. For example, in L5 pyramidal cell distal dendrites, EPSPs occurring

LTD d

before training

LTP before

change in tuning

before after

after

LTD

FIGURE 6.5 Examples of STDP in neural development and plasticity in vivo. (A) STDP can drive development of motion directioneselective responses, as studied in the Xenopus retinotectal system. Four neurons (aed) are hypothesized to be driven sequentially by a visual stimulus moving rightward. Right, Vm traces from the postsynaptic cell before and after visual motion training. In Epoch 1, the synapses are initially weak, but EPSPs summate to evoke a spike at the end of the motion stimulus. After repeated training with this motion stimulus, the synapses whose spikes precede the postsynaptic spike by in CA1, and differences in degree of axonal collateralization (CA3 > CA1), in addition to mix of afferents (Chapter 3 in Andersen et al., 2007). The entorhinal cortex (EC) is not technically part of the HF but is a major source of extrinsic input and often considered the “start point” of hippocampal circuitry, somewhat in the place of what would be thalamic input in neocortical areas (Fig. 9.1). The EC-HF circuit is often schematized as a relay from the EC upper layers to the HF fields (layer 2 projecting to DG, CA3, and CA2; layer 3 to CA1 and subiculum), with multiple stations of intrinsic processing, and return projections from CA1 and the subiculum to the EC, in the deeper layers. In some functional circumstances, the implied sequential activation may in fact dominate. However, the across-subfield progression has many alternate routings, which are important to keep in mind, and for this reason alone, the idea of the EC as a “stand-in” for thalamic input to other areas may not be helpful beyond a point. 1. Processing streams. EC consists of multiple subfields. In rodents, the main distinction is between the lateral (LEC) and medial (MEC) subdivisions. These have differentiated connections and are associated preferentially with spatial processing (MEC) versus object recognition, including olfactory (LEC). These subdivisions overlap within the DG but have complicated and preferential proximal-distal patterns in CA3 and CA1. In primates, at least seven interconnected specialized subfields have been identified within EC (Mohedano-Moriano et al., 2008). 2. Direct and indirect pathways. Neurons in layer 2 of both MEC and LEC project to DG, CA3, and CA2, but projections to CA1 and the subiculum originate from neurons in layer 3. Neurons in layer 2 are further distinguished as reelin þ large stellate cells or reelin/calbindinþ pyramidal neurons. Calbindinþ pyramidal neurons innervate the stratum lacunosum of CA1 where they establish synaptic connections with local interneurons (Kitamura et al., 2014). Some neurons in the deeper layers also contribute to the hippocampal projections (van Strien et al., 2009; Witter et al., 2014), but these have been less investigated. Coinnervation can occur by collateralization; for example, EC layer 3 neurons are reported to have collaterals to both CA1 and the subiculum, possibly as another means of parallel processing. 3. Non-EC inputs (i.e., other circuits). Other major inputs to the hippocampal subfields are from the septum (to all fields), supramammillary bodies (to DG and CA2), and for CA1 and subiculum, from reuniens and several cortical fields, especially in primate. The septal connections, important for hippocampal rhythmic activity, are established relatively early (see below), but developmental data are not available for reuniens or non-EC cortical inputs. 4. Non-EC outputs (i.e., other circuits). DG and CA3 are primarily input fields. However, CA1 and the subiculum both project to a wide array of cortical, thalamic, and subcortical targets; for example, CA1 projects to reuniens, to part of frontal cortex, to non-EC temporal areas, lateral septum and, from ventral CA1, to amygdala, among other subcortical structures. Inhibitory neurons in CA1 have direct projections to layer 1 of retrosplenial cortex (Miyashita and Rockland, 2007). For the subiculum, the number of outputs is even larger: to retrosplenial cortex (RC) and, through RC to parietal areas likely to be associated with spatial aspects. The pre- and parasubiculum form an intricate complex with the subiculum and RC, which is in turn interconnected with multiple cortical areas, including but not limited to EC.

9.1.1.1 Subfield features and numbers A detailed mapping of adult microcircuitry and its development requires a catalog of cell types, and spatial quantitative data on the divergence and convergence of connections. In the visual system, the understanding of the parvocellular versus magnocellular organization was facilitated by visualization of how the thalamic arbors terminated in primary visual cortex. In the hippocampus, relatively limited data are available. DG granule cells, for example, are known to have a small number of large boutons (w12), which contact the same number of postsynaptic CA3 pyramidal cells by a single dendritic site, but one which has an unusually large number of active sites (w50). The same axon will have w10X more, smaller filopodial terminations, which preferentially contact GABAergic neurons in CA3 (Acsady et al., 1998). The total number and distribution of inhibitory (1700) and excitatory (30,000) synapses have been mapped for a small number of CA1 pyramids (rat: Megias et al., 2001). An extensive survey of quantitative data has been calculated, but only as estimate, for CA1 local circuits in the rat (Bezaire and Soltesz, 2013). Continued progress in three-dimensional electron microscopy will be an important source of new data; for example, datasets for excitatory axons in MEC of P25 rats reveal path-length-dependent mapping of axonal synapses (over reconstructed lengths of 555 mm). A high-precision circuit sequentially targets inhibitory or excitatory postsynaptic targets over regular intervals (Schmidt et al., 2017).

204 PART | I Circuit development

Dentate gyrus. The DG receives extrinsic inputs from the medial septum (cholinergic), supramammillary nucleus, and EC, arrayed in a proximal-distal order along the dendrite. EC inputs originate from reelinþ stellate cells in layer 2 (not layer 2 pyramidal cells). Main features of dentate granule cells are their large number and their axonal specialization, the mossy fibers (MF) (Fig. 9.2 and see Chapter 3 in Andersen et al., 2007, and Chapter 1 in Scharfman, 2007, as two of many general references). There are additional contacts, as recently reported, with neurons in CA2 (Kohara et al., 2014); and additional local collaterals within the hilar region (i.e., polymorphic layer) (Scharfman, 2007). CA3. Pyramidal cells in CA3 are postsynaptic to extrinsic input from the medial septum and stellate cells in layer 2 of EC, to intrinsic connections from MF in DG, and to an abundant autoassociational system of CA3-CA3 collaterals. The several systems observe a dendritic compartmentalization, from distal to proximal along the apical dendrite (in order: EC, MF, CA3 collaterals). The CA3 axonal tree (CA3 collaterals and extrinsic Schaffer collaterals to CA1) is estimated to bear 30,000e60,000 synapses (Chapter 3 in Andersen et al., 2007), which suggests a postsynaptic network of 6000e12,000 neurons at least (allowing for up to five contacts on one neuron). Neither the spatial distribution of the full set of postsynaptic neurons nor their possible heterogeneity (see below: “sister neurons”) is known. One conspicuously distinct subpopulation (10%e20%) of pyramidal cells has been recently identified, which lacks the classical thorny excrescence, lacks MF input, is differentially modulated by acetylcholine, and has high initial firing frequencies (“bursting”; Hunt et al., 2018). The spatial pattern of CA3 axon arborizationdand therefore the functional microcircuitrydappears to have subregionspecific specializations. That is, individual neurons have differing ratios of CA3-CA3 and CA3-CA1 axon lengths, as well as variable preference for terminating in upper and/or deeper layers of CA1 (Ropireddy et al., 2011). CA2. CA2 was recognized in classic studies as a distinct subfield by several anatomical criteria: cell size (larger than pyramids in CA3, but smaller than those in CA1), and subcortical connections with both the septum and supramammillary nucleus (SUM-medial) (Cui et al., 2013). The input from SUM is denser than that from the septum and is associated with theta oscillations (see Section 9.2.7). As a subfield, CA2 has been specifically implicated in social memory processing, with a crucial role in the consolidation of socially relevant information into long-term memory, and in temporal encoding (Dudek et al., 2016). Genetically targeted inactivation of CA2 pyramidal neurons caused a pronounced loss of social memory (Hitti and Siegelbaum, 2014); and consistent with a role in social processes, CA2 neurons receive vasopressin expressing connections from the paraventricular nucleus of the hypothalamus (Cui et al., 2013). The pattern of input connections to CA2 has been recently reexamined and suggested to form a major circuit separable from and in parallel with EC-HF. CA2 receives inputs on distal apical dendrites from EC layer 2, from MF (but from synapses that are smaller than those to CA3), and, on more proximal dendritic portions, from CA3 (Kohara et al., 2014). Inputs from EC are more efficacious, but in a neuron-specific manner, only for those CA2 pyramids which extend their apical dendrites into SLM (Helton et al., 2019). CA1. With the subiculum, CA1 is a major source of widely distributed hippocampal outputs, and is implicated in novelty detection, input comparison, and “enrichment” of hippocampal output (Soltesz and Losonczy, 2018). Distinctive features, compared with CA3 and CA2, are input from EC layer 3 (not layer 2), input from thalamic and some neocortical regions, and a bilaminar organization. Input from the EC, as in CA3 and CA2, targets distal apical dendrites, but in CA1, this converges with input from reuniens thalamus (Dolleman-van der Weel et al., 2017). Input from Schaffer collaterals, as in CA2, terminates more proximally. Subiculum. The subiculum receives projections from EC and CA1, and is also part of multiple wider networks, with abundant bidirectional thalamic and cortical connectivity. Like CA1, it projects back to the EC, in the deeper layers; but it also participates in a second spatially relevant network composed of the presubiculum, parasubiculum, and RC. The subiculum (but not CA1) sends excitatory projections from segregated pyramidal layers to the anterior thalamus, mammillary bodies, and RC, a network associated with episodic memory in humans and spatial memory in rodents (e.g., Aggleton and Christensen, 2015). Recent studies using single cell RNAseq further refine subicular pyramidal cells into eight separable classes, producing an ordered but complex spatial heterogeneity (Cembrowski et al., 2018).

9.1.1.2 Zonal specializations. Example: upper and deep CA1 Other structures in the brain have a complex zonal organization, perhaps most notably the cerebellum. The hippocampal circuitry differs with regard to location in the dorsal versus ventral sectors, with major differences in cellular arrangement, transcriptomic profiles, and connections, and in the transverse axis (proximal, near DGdor distal, near the subiculum). Whether this is best viewed as an organization by gradients remains under active investigation (e.g., Strange et al., 2014; Cembrowski et al., 2018).

FIGURE 9.2 Schematic summary of EC-DG-CA3 projections. (A) EC layer 2 (stellate cell) projects divergently to granule neurons in the DG (Tamamaki and Nojo, 1993). Cell bodies in black, and stellate cell axon in red. (B) Two granule cells (gc) are shown sending their MF projections to pyramidal cells (pc) in CA3. MF (in red) arborize across CA3, and have local collaterals in the polymorphic layer (PL). gcl, granule cell layer; ML, molecular layer; pcl, pyramidal cell layer; pl PL ¼ polymorphic layer. Note the small number but divergent spatial extent of large MF boutons in CA3. With permission from Amaral, D.G. et al. (2007). The dentate gyrus: fundamental neuroanatomical organization (dentate gyrus for dummies). Progress Brain Research (ed. Helen Scharfman) 163: 3e22. (C) Higher magnification of three MF axons, with typical distribution of large boutons (arrowheads) and smaller filopodia (arrows). (D) (left) The same three axons rotated to better show the spatial extent. A fourth DG axon is included, and two of the originating GCs. D) (right) The same axons, rotated to a dorsal view. The insets at left and right are for orientation, with respect to the transverse and dorsal planes. With permission from Acsady, L., Kamondi, A., Sik, A., Freund, T., Buzsaki, G., 1998. GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. Journal of Neurosci. 18, 3386e3403. Scale bars: 50 mm in A, 400 mm in B. str. Luc ¼ stratum lucidum; str pyr ¼ stratum pyramidale.

206 PART | I Circuit development

An example of depthwise organization is the two broad layers in CA1 of rodents, which have been characterized by multiple criteria (Lee et al., 2014; Valero and Menendez de la Prida, 2018). Two sublayers are recognized as superficial (nearer the hippocampal fissure) or deep (nearer stratum oriens white matter). Pyramidal neurons in the deeper layers preferentially receive inputs from CA2 (Kohara et al., 2014) and from MEC, are more active, and are more likely to form place fields. Deeper neurons have been proposed to represent dynamic features while the superficial population, which receives stronger inputs from LEC, conveys a more stable map of the environment (reviewed in Soltesz and Losonczy, 2018). EC inputs, however, exhibit a proximo-distal gradient, in which MEC input, strongest in proximal CA1, preferentially targets pyramidal neurons in the deeper layer, but LEC input is strongest distally and preferentially targets the superficial subpopulation (Soltesz and Losonczy, 2018). Neurons in the upper layers, but not the deeper, are positive for calbindin and use synaptic zinc, an activity related marker (Slomianka et al., 2011). In a further dissociation, most distinct in NHP, neurons in the upper layers project to perirhinal cortex, while those in the deeper layers project preferentially to parts of frontal cortex (Ichinohe and Rockland, 2005). In ventral hippocampus, neurons in the deeper layers project to nucleus accumbens and the amygdala. Neurons in the lower layers, due to the inside-out pattern of neurogenesis in CA1, are born earlier than those in the upper layers (Chapter 4 in Andersen et al., 2007). The ontological sequences beyond the differential birthdates are currently unclear. Summary. The hippocampus has several distinctive architectural features in comparison with neocortex: (1) the large bouton, multisynaptic MFs, each having a small number of boutons; (2) the very extensive CA3-CA3, CA3-CA1 collateralization; (3) the lack of major thalamic input at early stages (e.g., DG, CA3, CA2); the robust dendritic compartmentalization of input systems; and the gradientwise (transverse and proximodistal) topographic organization of input and output connections. Common features between six-layered neocortex and the three-layered HF, however, have been discussed (e.g., Mercer and Thomson, 2017).

9.1.2 Functional backdrop 9.1.2.1 The spatial navigation system in rodents In rodents, a basic assay of mature hippocampal function is the emergence of spatial abilities and the emergence of characteristic rhythmic activity. The four main classes of spatially tuned neurons emerge postnatally: head direction cells (P12), place and boundary cells, and lastly grid cells (P21) (in rat; Bjerknes et al., 2014; 2018; Witter et al., 2014; Tan et al., 2017). The early emergence of head direction responses, even before eye opening (at P15), can be taken as indicative that an underlying network of sensory connectivity is already well established. This has been attributed to olfactory or tactile modalities (both adultlike by P11eP12), in combination with a hard-wired, genetically specified component (Fig. 9.3).

9.1.2.2 Developmental milestones in humans In humans and NHP, the maturational progression of hippocampal subfields has been studied extensively by functional imaging. One anatomical study (Lavenex and Lavenex, 2013) has related hippocampal subfield development to three behavioral stages: the earliest epoch of infant amnesia (under 2 years old), sporadic memory capacity in childhood amnesia (3e5 years old), and adultlike episodic memory (at about 7 years). Anatomical stages were defined largely by volumetric laminar changes in the respective subfields and are provisionally correlated with the emergence of specific hippocampal dependent aspects of memory; namely, in order of early to later, path integration (group 1: subiculum, pre- and parasubiculum, and CA2), basic allocentric spatial memory (group 2: CA1), pattern separation and high-resolution spatial memory (group 3: CA3). The emergence of episodic memory requires the maturation of all hippocampal circuits. The earliest maturing hippocampal subfields (subiculum, presubiculum, parasubiculum, and CA2) are fields most highly interconnected with subcortical regions. An intermediate group is identified as CA1, for which the direct connections with EC layer 3 are proposed as a major influential factor. Lastly, CA3 is identified as relatively late maturing, owing to its close association with the DG, which is also late maturating (Lavenex and Lavenex, 2013). In macaque monkeys and humans, pyramidal cell neurogenesis occurs early, during the first half of gestation (Berger et al., 1993). Entorhinal inputs, both from layers 2 and 3, are in place in the neonatal monkey, and appear generally similar in laminar and topographic distribution as in adults (Amaral et al., 2014).

37

Orienting towards maternal odor 30 Active sniffing emerges Olfactory

Vestibular

Adult-like olfactory discrimation of complex odor combinations

50,51

38,40

Basic associative olfactory learning 49 First central vestibular Adult-like central vestibular response response48 35

Rudimentary righting reflex 41 First cortical barrel response from passive whisker stimulation Tactile 52,53

Active whisking 54 55 First auditory hair cell potential Ear canals unfuse

Adult-like auditory nerve responses

54

Auditory 57

56 Tone discrimination Orienting towards sound source 47 58-60 V1 response to flash stimulation Eyelids unfuse

Adult-like V1 responses

45

Visual Limited crawling

42

Pivoting on abdomen

First V1 response to drifting gratings 35 Quadrapedal walking

35

45,46

Pivot towards nest Non-hippocampus dependent spatial activity

Nipple seeking

Hippocampus dependent spatial activity

42,64

Spontaneous exploration when placed in open field

36

66,69

67,68

Crawl to nest after displacement Spontaneous exit of nest in lab conditions Spontaneous exit of nest in the wild 70-72 76-78 Morris Water Maze (cued platform) Radial 8-arm maze / T-maze spontaneous alternation

Immature SPWs 140 in behaving rats

Hippocampal network activity

36

Immature gamma bursts

Immature theta oscillations/ 139 Start of increasing gamma oscillations

63

70,71,73,74

Morris Water Maze (hidden platform) 85 Adult-like proportion of CA1/mEC cells modulated by theta oscillations 140,141

139

Emergence of fast-field ripples 87 Visual cue control

Head direction cells First immature HD cell

87,88

Adult-like HD signal

85,86

Place cells First instance of adult-like place cell Boundary responsive cells

112

Adult-like boundary signal

Adult-like place signal

70,113,114

121,122

Grid cells Adult-like grid signal 0

5

10

15

20

85

25

30

35

40

45

50

Postnatal age (days)

207

FIGURE 9.3 Developmental timeline of sensorimotor capacities (top), spatial behavior (middle), and spatial cell activity (at bottom) in the rat, at P0eP50, as indicated in the x-axis. Colored bars indicate the time course of emergence and maturation (shading with slants), corresponding to modality (top), behavior (middle), or cell type (bottom). Red dots þ emergence of a particular phenomenon. With permission from Tan, H.M., Wills, T.J., Cacucci, F., 2017. The development of spatial and memory circuits in the rat. Wiley Interdiscip. Rev. Cogn. Sci. 8, 1e16.

Organization and development of hippocampal circuits Chapter | 9

Motor

208 PART | I Circuit development

9.2 Circuit development 9.2.1 Early stages The early stages of circuit development occur in the hippocampal neuroepithelium and, for eventual GABAergic neurons, in the lateral and medial ganglionic eminences (LGE and MGE). These stages are similar to neocortical development, except that the migratory routes are more complex owing to the complicated geometry of the adult hippocampus. It takes four days for CA1 pyramidal neurons to reach the hippocampal plate and longer for those in CA3 (e.g., Danglot et al., 2006). The hippocampus arises from the caudomedial edge of the dorsal telencephalic neuroepithelium. The hippocampal primordium is initially populated by Cajal-Retzius cells and radial glia cells, which proliferate in the dentate neuroepithelium. Radial glia cells expand into the tissue producing embryonic neural stem cells that will later migrate to the subgranular zone to give rise to DG neurons (in the proximal location). Cajal-Retzius cells are important in the assembly of the developing hippocampal circuit (Ceranik et al., 2000). They constitute a class of calretinin positive excitatory cells, releasing glutamate (Quattrocolo and Maccaferri, 2014). In the DG, they are located in the outer molecular layer, the target layer of the EC stellate cell afferents. Cajal-Retzius cells project to the rat EC by E17, before the arrival of EC afferents in the HF, and presumably act as guideposts for EC fibers. Their presence decreases by 85% from P8 to P60 (Anstotz et al., 2016).

9.2.2 Neurogenesis The timeline of hippocampal neurogenesis has an overall early progression, and significantly precedes the emergence of spatial abilities and functional cellular characteristics (Section 9.1.2 Figs. 9.3 and 9.4). Pyramidal cells in CA3-CA2-CA1 are generated at E10eE18 in mouse (of 19 days gestation) and E16eE21 in rats (of 21 days). The peak for CA3 is earlier than that for CA1 (respectively, in rat: E17 and E19; in mouse: E14e15 and E16e17). While the pyramidal cell layers are recognizable at birth, some late-born pyramidal neurons are still migrating. This timeline is generally compatible with a one-directional serial processing across CA3-CA2-CA1-subiculum, with the notable exception of the DG. Only about 15% of DG granule cells are generated before birth (starting at E20 in the rat, and peaking during the first postnatal week (reviewed in Danglot et al., 2006, Chapter 4 in Andersen et al., 2007). Thus, much of the microcircuitry development ongoing in the first two postnatal weeks of the rodent hippocampus is independent of DG input, although there might be a significant “guidepost” role from the small early born component. Importantly, GABAergic interneurons are generated early: between E13 and E18 in rat, and E11 and E17 in mouse. Interneurons in CA3 and CA1 as a group are generated at E12eE13. Those in DG lag but only slightly (E13eE14). Thus,

FIGURE 9.4 (A) Increase in rat brain size from (left to right) E18, P0, P4, P8, P12, P16, P20 to adult. Gross brains overlay a 1  1 cm grid. (B) Development of mean laminar width in the MEC, from P0 to adult, as shown. From Ray, S., Brecht, M., 2016. Structural development and dorsoventral maturation of the medial entorhinal cortex. Elife 5.

Organization and development of hippocampal circuits Chapter | 9

209

early development of microcircuitry is disproportionately influenced by interneurons. These, however, exert a developmentally specific depolarizing effect up to about P10 (Danglot et al., 2006). Interneurons are characterized by great heterogeneity, but nevertheless it is still possible to identify some general developmental patterns. Hippocampal interneurons are generated in two neurogenic waves between E9 and E12, and E12 and E16 from the medial and caudal ganglionic eminence (MGE and CGE), infiltrating the hippocampus by E14. MGE generates parvalbumin (PV), somatostatin (SOM), and nitric oxide synthase interneurons including fast-spiking basket, axo-axonic, bistratified, ivy, neurogliaform, and OLM cells. On the other hand, CGE cells give rise to cholecystokinin (CCK), calretinin (CR), vasoactive intestinal peptide (VIP), and reelin cells including mossy fibereassociated cells, Schaffer collateraleassociated cells, trilaminar cells, non-fast-spiking basket cells, and other neurogliaform cells (Tricoire et al., 2011). Differential distributions have been reported across the subfields (see Booker and Vida, 2018).

9.2.3 Connections: EC EC afferents invade the DG at E19 (mouse: Super and Soriano, 1994), and are present at birth in CA3 and CA1. This, however, precedes the appropriate positioning of postsynaptic pyramidal cell dendrites, which have not yet reached the molecular layer (but see above, about interneurons). Moreover, even though the axons are in place, maturational processes are ongoing postnatally in EC. For example, doublecortin (DCX) expression, a marker for immature neurons, is detected through the first two to three postnatal weeks (in mouse: Donato et al., 2017): P14 (MEC layer 2 stellate cells), P20 (MEC layer 2 pyramidal cells), P26 (MEC and LEC layer 5), and P30 (LEC layer 2). In the rat Ray and Brecht (2016) report DCX expression in layer 2 pyramidal cells as no longer detectable at P16 in the earlier maturing dorsal MEC (but persisting in the ventral MEC at that stage).

9.2.4 Cell autonomous organization The early stages of circuit organization are heavily influenced by cell autonomous factors, which appear to rely on temporally matched molecular signals. Thus, GABAergic synapses appear first on older, more mature dentate granular cells, and are then time matched within the postsynaptic populations, occurring earlier on CA3, and then on CA1 neurons (Danglot et al., 2006). Clonally related sister pyramidal neurons (four to five neurons), labeled from injections at E12, have been identified and tracked postnatally (Xu et al., 2014). In CA1, excitatory neurons of the same sister lineage share synchronous activity and inhibitory input from fast-spiking (FS), but not non-fast-spiking interneurons. Putative functional units are thus established by common inhibitory, but not direct connectivity. Sister neurons are horizontally dispersed, unlike the radial dispersion more common in neocortex. A still open question is whether these early specified clonal units encode similar place fields and how these might relate to anatomically organized functional clusters (Dombeck et al., 2010). Several other studies provide evidence that parallel hippocampal circuits are assembled from distinct subpopulations through matched developmental time windows. Deguchi et al., (2011) use sparse Thy1 reporter lines (Lsi1 and LSi2) with sparse expression of GFP to identify two early specified pyramidal cell subpopulations (Lsi1 and Lsi2) with distinct temporal signatures of neuronal specification and synaptogenesis. These are traced to separable populations of radial glia progenitors, with Lsi1 neurons deriving from radial glia at E10.5, an earlier cohort than for Lsi2 neurons. These two temporally distinct populations form preferential connections with matched sister lineages in DG-CA3 and CA3-CA1. Donato et al., (2015) tracked the early developmental responses in CA1 of two PV basket cell populations, distinguishable as early born (EB) or late born (LB) (respectively, E9.5 vs. 13.5 in mouse). These had distinctive levels of PV and GAD67 (high for EB or low for LB), distinctive ratios of excitatory to inhibitory synapses (respectively, 4:1 and 1:2), and distinguishable features relatable to plasticity responses. That is, the EB population was associated with what appeared to be “adherence to rules,” whereas the LB population was more responsive in paradigms drawing on flexible task-relevant information. The study suggests that EB PV cells predominantly target deep pyramidal cells, which are born earlier (E16/ 17) than the pyramidal cells in the superficial layers, but later than either of the two PV basket cell groups.

9.2.5 Neural activity Neural activity is a crucial regulator of neural network development. How intrinsic excitability affects integration of a neuron into developing circuits has been probed by assessing changes in spine density after differential silencing of CA1, CA3, or DG granule cells (Johnson-Venkatesh et al., 2015). Unique excitability-dependent mechanisms are reported to govern the development of hippocampal neurons in different ways. Thus, intrinsic excitability is important for spine

210 PART | I Circuit development

development of CA1 pyramidal neurons, regardless of input origin (EC or Schaffer collaterals), but only at a later stage (P11eP15, but not at P6eP11). Silenced CA3 neurons appear to develop normally. Silenced DG granule cells show perforant pathway-specific decreases in spine density. The authors suggest that these outcomes reflect unique activitydependent coupling to transcription networks. Neural activity acts in conjunction with cell autonomous mechanisms in the setting-up of mature circuits. The EC is a major extrinsic source of neural activation, and silencing MEC layer 2 stellate cells is reported to result in major changes as assessed by three markers of maturational state (lack of expression of DCX, upregulation of PV, and synaptic density). Note that finer subdivisions (as, proximal-distal differences, and upper and lower strata in CA1) are not accounted for, and multiple factors, such as indirect effects on interneurons and on other extrinsic and intrinsic afferent systems, need to be investigated (Donato et al., 2017).

9.2.6 Maturational events All hippocampal fields undergo robust postnatal maturation in dendritic length, spine density, and synaptogenesis. Sequences of structural maturation have been scored by three different criteria (Donato et al., 2017 mouse), and are staggered across subfields. With the notable exception of the late maturing DG, the sequences approximately correlate with the traditional subdivisions. Cellular maturity, as assayed by lack of expression of DCX, proceeds from layer 2 stellate cells of MEC (P14), layer 2 pyramidal cells of MEC and CA3 (P20), pyramidal cells of CA1 (P23); the cluster of DG, SUB, layer 5 of both MEC and LEC (P26); and lastly, layer 2 of LEC (after P30). A similar maturational progression was identified, at slightly different time points, by upregulation of PV: layer 2 MEC (P17), followed by CA3 (P20); followed by the cluster of CA1, MEC layer 5, SUB, and DG (all P26); and lastly, layer 5 LEC (P30) and layer 2 LEC (after P30). For synaptogenesis, assayed by increased density of presynaptic zones (visualized by Bassoon; see Fig. 9.1 in Donato et al., 2017), the timeline was layer 2 MEC (P17), CA3 (P20), CA1 (P23); followed by DG, SUB, layer 5 MEC, and layer 5 LEC (P26); and lastly, layer 2 LEC (beyond P30). The first adultlike granule cells of the DG have been identified at P7, but with neurons showing active dendritic remodeling as late as P60. The MF of the oldest granule cells arrive at CA3 by P7, the time at which electrical stimulation of the DG induces LTP in CA3 (rats: Chapter 10 in Scharfman, 2007). Zinc levels, an indication of mature MF, which use synaptic zinc, are detectable in CA3 from P3 and reach adult levels by P15 (Slomianka and Geneser, 1997). In CA3, the postsynaptic thorny excrescences emerge proximally at P9, slightly later than the first MFs, but in a simplified, immature form (Amaral and Dent, 1981 and see Seress and Ribak, 1995 for primates). Dendrogenesis is an important factor, since afferent connections are only efficacious if the postsynaptic targets are in place. This has been discussed in detail for granule cells of the DG (e.g., Chapters 1 and 10 in Scharfman, 2007). Less is known about the progression of axons. Reelin is a necessary factor for normal development of entorhinal axons, in terms of collateralization and synaptic density, although the EC-HF pathway otherwise develops normally in the absence of reelin (Chapter 3 in Andersen et al., 2007). Overall, there does not seem to be an exuberant overgrowth (Chapter 2 in Andersen et al., 2007), but CA3 is reported to exhibit a transient hyperexcitability in the second postnatal week (rats: Gomez-di Cesare et al., 1997). With maturation, short-ranging profusely branched axons are replaced by those with longerranging arbors. The relative outgrowth of the axonal arborization in CA3 and CA1 has not been quantitatively compared. Large-scale developmental gradients have been identified. Maturation proceeds earlier for dorsal MEC than for ventral MEC. Superficial layers mature earlier, where small spatial scales are represented; and maturation of ventral MEC microcircuits, where larger spatial scales are represented, is later, coincident with the onset of exploratory behavior (Ray and Brecht, 2016). The differentiation of the dorsal and ventral hippocampus is already in place at birth, as regards gene expression and topography of EC inputs, but global length and curvature continue to increase through the first two weeks postnatal. Volumetric changes are attributed to cell division, especially of the DG granule cells, as well as dendritic arborization and myelination (O’Reilly et al., 2015). Two anatomical markers are available to assay the maturation of CA2. One is the field-specific expression of PCP4 (mouse: San Antonio et al., 2014). This is described as first detected at P4-5 but reaching adult levels only by P21. Second is the expression on perineuronal nets (PNN), which is associated with suppression of plasticity and, in the case of CA2, might contribute to the plasticity-resistant features of CA2 synapses (i.e., lack of LTP of excitatory synapses of the CA3 synapses; Carstens et al., 2016; Lensjo et al., 2017). PNN are identifiable at P14, but only fully developed at P45. PV and grid cells (mouse) in EC. Grid cells in the EC (equivalent to about 25% of stellate cells in layer 2, Rowland et al., 2018) start to establish a stable activity pattern at P16, and an adult number of mature grid cells is reached at about P22 (Berggaard et al., 2018). Multiple anatomical substrates underlie grid cell functional organization. The role of inhibition by PVþ interneurons has been specifically investigated by pharmacological silencing, a procedure which impaired

Organization and development of hippocampal circuits Chapter | 9

211

the hexagonal spatial selectivity of excitatory neurons in layer 2 of MEC. Relevant for circuit definition, this was in contrast with the result of silencing SOMþ interneurons, a procedure which resulted in decreased spatial selectivity of cells with aperiodic firing fields (mice: Miao et al., 2017). This result indicates at least two differentiable networks, in this case, PVþ versus SOMþ. PVþ maturational progression in MEC is approximately in parallel with grid cell features (Berggard et al., 2018). That is, soma size of PVþ interneurons and the number and density of PVþ terminations onto principal neurons increase between P10 and P15 (before eye opening and before the emergence of grid cell properties). During the later stabilization of grid activity, PVþ terminations increase in size with an increase in the percentage of PVþ terminations containing mitochondria (Berggaard et al., 2018).

9.2.7 Coordinated network activity In the adult hippocampus, distinctive oscillations are associated with different behavioral conditions and different facets of memory processing. Theta-band oscillations (7e12 Hz), informed by afferent activity from neocortex and EC, are associated with exploratory and active behaviors and a labile form of memory trace. Spontaneous sharp-wave (SPW) bursts are initiated by cortical activity during slow-wave state (Hahn et al., 2006) and associated with sleep and other consummatory behaviors (reviewed in Buzsaki, 2015). The complex underlying mechanisms are likely to involve multiple sources, including glutamatergic population bursts mediated by Schaffer collaterals of pyramidal neurons in CA3, large synaptic events in CA1, extrinsic inputs from EC, medial septum, and supramammillary nucleus; and temporal and spatial coordination by the large diversity of GABAergic interneurons (Unal et al., 2018) (Fig. 9.5). Throughout the brain, spontaneous patterns of correlated neural activity figure in the formation of neural circuits. In the immature hippocampus, early SPW activity emerges in the first postnatal week, associated with periods of immobility, sleep, and feeding. Myoclonic twitches can trigger hippocampal early SPW activity, probably secondary to motor cortical influence on MEC (Buzsaki, 2015; Veleeva et al., 2019).

FIGURE 9.5 Schematic to illustrate network oscillations in the adult rat. (A) Basic anatomical circuit, consisting of (1) excitatory pyramidal neurons and four populations of inhibitory interneurons, (2) inputs from CA3 and the septum, (3) the extended network (at left). (B) Phase relations between theta or ripple oscillations and firing probability of pyramidal cells and interneurons. With permission from Klausberger, T., Somogyi, P., 2008. Neuronal diversity and temporal dynamics: the unity of hippocampal circuit operations. Science 321, 53e57.

212 PART | I Circuit development

Adultlike ripple oscillations emerge at the end of the second postnatal week and are recognizable as adult by P20 (Buhl and Buzsaki, 2005; Buzsaki, 2015). This time course parallels the switch in polarity in GABAA receptor mediated effect from depolarization to hyperpolarization. Extrinsic entorhinal and septo-hippocampal projections are present before birth, preceding oscillatory patterns, but their maturation continues through the second postnatal week. Another factor in coordinated network activity is the occurrence of widespread inhibitory projections, originating from a subpopulation of early generated (from E10) GABAergic “hub” cells. These are morphologically heterogeneous, occurring mainly in CA3 but also in CA1 and DG, and, given their early and persisting widespread axonal projections within the septal network, are interesting candidates to coordinate the timing of neural activity across structures (Picardo et al., 2011; Villette et al., 2016). Intrinsic interneurons are a source of widespread inhibition; for example, a VIP neuron has long-range projections arborizing both within CA1 (presynaptic to interneurons) and the subiculum (presynaptic to both interneurons and pyramidal cells). This functional specialization operates to control cell ensembles. Developmental trajectories have not yet been investigated (Francavilla et al., 2018).

9.3 Postnatal development of electrophysiological patterns In this section we present results concerning (i) the postnatal development of single cell electrophysiological properties, (ii) the early in vivo developmental patterns of activity, and (iii) the development of major hippocampal rhythms in relation to place cells, head direction cells, and grid cells. Overall the electric properties of hippocampal cells produce a very highly excitable state immediately after birth that progressively decreases until the third week. These changes in electric properties are paralleled by the maturation of the dendritic arborization (see Section 9.2.6).

9.3.1 Postnatal development of single cell electrophysiological properties 9.3.1.1 Entorhinal cortex layer II stellate cells The stellate cells of the MEC are the first class of cell to mature and provide a major excitatory input to instruct circuit formation during postnatal development (Donato et al., 2017). Their activity is necessary to generate a stereotyped wave of maturation, which at least approximately conforms to a sequential recruitment: CA3, CA1, DG, layer 5 of medial and lateral entorhinal cortices (MEC and LEC), and finally layer 2 of the LEC. This maturation occurs between P5 and P30 and pharmacogenetic silencing of MEC stellate cells at P14eP17, but not after P20, can impair the entire circuit formation. Layer 2 of the MEC reaches a stable width around P12 (Ray and Brecht, 2016). Mature stellate cells exhibit subthreshold membrane potential oscillations (4e12 Hz) in vitro in response to constant depolarizing current. These subthreshold oscillations appear between P14 and P18 becoming stable at P28. These changes in intrinsic electrical properties depend on the progressive expression of Ih and INaP currents, which are necessary, respectively, for the slow rectification required for resonance, and for amplification of the resonance. Synaptic inputs, assessed by spine density, follow a similar trend. These findings predict that theta rhythmicity can be immature until P18 (Burton et al., 2008).

9.3.1.2 Postnatal development of the dentate gyrus The DG presents an extended developmental period beyond P15 (Yu et al., 2014). Interestingly, DG granule cells coexpress both GABA and glutamate (Walker et al., 2001), and during the first postnatal days GABA displays a depolarizing effect (Minlebaev et al., 2013), which is then overridden by AMPA/kainate receptor mediated excitatory currents (Marchal and Mulle, 2004). In the same period, the DG cells display hyperexcitable properties characterized by high resting membrane potential (40 mV), high input resistance (>1000 MU), low threshold calcium spikes, and synchronous network activity reminiscent of Giant Depolarizing Potentials (GDP) (Pedroni et al., 2014, and see below). Both resting membrane potential and input resistance are inversely correlated with the extension of the dendritic arborization, and by P14 the high excitability properties progressively shift to lower values (Liu et al., 2000). Furthermore, the dendritic arborization reaches maximum extension at P14 and regresses to a mature size by P20 (Rihn and Claiborne, 1990).

9.3.1.3 Postnatal development of CA3 pyramidal cells The CA3 region receives direct inputs from EC stellate cells through the perforant path and DG inputs from MFs, and conveys its output to CA1 through the Schaffer collaterals. During the postnatal period these cells display a remarkable change in their electrical properties. Immediately after birth, at P0, the resting membrane potential is highly depolarized

Organization and development of hippocampal circuits Chapter | 9

213

(40 mV) and the input resistance is extremely high (>1000 MU) generating a very high excitable state. This excitable state linearly decreases to adult values by P20eP25 (Tyzio et al., 2003). This change in cellular excitability is partially due to a progressive expression of Ih currents from P0 to P20 (Vasilyev and Barish, 2002). In the early postnatal days, until P10, GABA application has a depolarizing effect (Minlabaev et al., 2013). Morphologically, MFs are present in the hilus in the first postnatal day, and by P3, some fasciculi are already detectable in the stratum lucidum of CA3. MF giant boutons appear around P7, reaching adult size by P14.

9.3.1.4 Postnatal development of CA1 pyramidal cells The main electric signature of CA1 pyramidal cell, the sag in response to a current pulse, depends on Ih currents. These currents control not only input resistance and dendritic integration but are also responsible for the resonance properties of the cells. This physiological characteristic is essential to follow rhythmic network activity such as the one provided by theta oscillations. Ih currents are conducted by HCN channels and during postnatal development these channels display a progressive increase in expression. The most relevant isoforms of the HCN family, the HCN1-2, increases from P6 to P26 (Surges et al., 2006). The result is an enhancement of the Ih current (Vasilyev and Barish, 2002). Furthermore, at around P20, there is a shift in the composition of AMPA receptors, characterized by a decrease in GluA1 and an increase in GluA3 and AMPA regulatory protein (TARP). This increases AMPA receptor response duration and postsynaptic excitability (Blair et al., 2013).

9.3.1.5 Postnatal development of interneurons Interneurons are responsible for inhibition and their activity provides a cardinal regulation of the dynamic of the excitatory cells. Importantly, their activity is crucial to the expression of major hippocampal rhythms such as theta oscillations and sharp-wave ripples. Although hippocampal interneurons are generated prenatally, their maturation extends postnatally (Danglot et al., 2006). Usually in the early postnatal period, interneurons with no spontaneous or evoked postsynaptic activity are characterized by a poor dendritic arborization, with the exception of interneurons with only GABAergic postsynaptic potentials. Interneurons with GABAergic and glutamatergic postsynaptic potentials show more developed dendrites (Hennou et al., 2002). The maturation of dendrites increases between P0 and P5 with a further increase between P10 and P20 (Lang and Frotscher, 1990). Electron microscopy analysis of axosomatic synapses in the CA3 region shows a strong increase in synaptic connections between P7 and P21. In addition, the density of the GABAergic terminals around pyramidal cells also increases by P20 (Serress and Ribak, 1988). The maturation of the dendritic and axonal arborization is strictly dependent on the cell type. For example, in the rat, parvalbumin (PV) basket cell dendrites progressively mature between P2 and P6 in the DG (Serress and Ribak, 1990) and their axonal arborization is concentrated in the granule cell layer by P16. Numerous interneuron markers reach mature levels by P21. More specifically PV immunoreactivity appears at P7eP8 first in CA3 and then in CA1. Calbindin (CB) is expressed at P2 in several interneurons of the stratum oriens and radiatum as well as in DG granule cells. Calretinin (CR) is expressed in interneurons at P10. Somatostatin (SOM) is already detected in the stratum oriens at P0 reaching a peak at P10eP15 and then decreasing to adult levels. Similarly, Cholecystokininpositive (CCK) interneurons are already detected at P0 reaching the peak at P10. Vasoactive intestinal peptide (VIP) interneurons appear at P5, whereas NPY immunoreactivity is first detected at P0 reaching adult levels at P21 (Danglot et al., 2006). Importantly, PV interneurons are coupled by gap junctions. These electric connections are mediated by Connexin 36. The expression of Connexin 36 increases until P7eP16 and then decreases between P14 and P28 (Belluardo et al., 2000). The developmental changes in electric properties have been well characterized in PV fast-spiking basket cells, and the fast-spiking property is known to depend on the expression of voltage gated potassium channels Kv3.1 and 3.2 (Martina et al., J. 1998). The Kv3.1b subunit is first detected at P8, reaching adult levels at P40 (Du et al., 1996). Similarly, Kv3.2 is detected at P7 reaching full expression at P21 (Tansey et al., 2002). It is crucial to remark that GABA neurotransmitter has a depolarizing effect until P10, providing an important excitatory drive to the developing hippocampus (Ben Ari et al., 1989). The depolarizing action of GABA is due to active uptake of Cl by NHCC1, a Naþ/Kþ/2Cl- cotransporter (Marty et al., 2002). The switch from depolarizing to hyperpolarizing effect of the GABAa current is hypothesized to depend on a shift in the equilibrium potential of the Cl (ECl), from values more positive of the transmembrane potential to more negative ones (Owens et al., 1996), due to the expression of Cl extruding Kþ/Cl-cotransporter KCC2 (Rivera et al., 1999). These changes in electrical properties are tightly correlated to very specific electrophysiological brain patterns.

214 PART | I Circuit development

9.3.2 Early developmental patterns of brain activity The developing nervous system generates stereotypical patterns of electrical activity (Ben Ari, 2001). This activity is believed to be necessary for circuit maturation. In the early postnatal period, during synaptogenesis, the electrical activity of the hippocampus is characterized by the occurrence of synchronous GDPs. In the hippocampus, GDPs appear immediately after birth lasting until P10 (Wester and McBain, 2016). Recordings in vivo have shown recurrent activity (0.33e0.1 Hz), which is consistent with GDP recorded in vitro (Leinekugel et al., 2002). In addition, this recurrent activity shows a high degree of correlation between EC and hippocampus (Valeeva et al., 2019). GDPs are the result of bursts of polysynaptic activity lasting several hundred milliseconds and traveling as a wave. At early stages, the lack of robust synaptic connectivity necessary for network activity is counterbalanced by the high intrinsic neuronal excitability. A crucial role in GDP generation is mediated by the excitatory effect of GABA. GABA has a depolarizing effect until P10 and is supported by several interneuronal populations. More specifically, due to their high degree of connectivity, MGE-derived interneurons drive the generation of GDP (Wester and McBain, 2016). Among the MGE-derived cells, long-range projecting SOM cells are thought to play the major role of “hub cells” (Picardo et al., 2011), innervating large numbers of neighboring cells with a diffuse and dense axonal arborization. These cells also display intrinsic electrical properties typical of a high excitability state, such as a depolarized resting membrane potential (47 mV) and a high input resistance (600 MU). It has been proposed that the recurrent coactivation of cellular assemblies may promote growth, wiring, and more generally, plasticity (Leinekugel, 2003).

9.3.3 Development of major hippocampal rhythms The activity of the mature hippocampus can be differentiated in two major states, both occurring during sleep and wakefulness. These circuit states are defined as theta oscillations and sharp-wave ripples (SWR). Theta oscillations consist of periodic fluctuations of the hippocampal cell membrane potential in the 7e12 Hz range. These regular fluctuations are associated with movement in the awake animal and with REM sleep in the sleeping animal (Pignatelli et al., 2011). The rhythmic GABA release, from terminals arising from the medial septum and horizontal and vertical limb of the diagonal band of Broca, modulates hippocampal PV basket cell activity, which in turn generates theta oscillations. Thus, these septal GABAergic cells control the theta synchronization of the entire hippocampus (Freund and Antal, 1988). SWRs consist of periodic events (1 Hz) which are generated by neocortical slow-wave activity (Hahn et al., 2006) and include phasic high-frequency (150 Hz) wavelets named ripples, which are recorded exclusively in the CA1-CA3 pyramidal layer (Buzsaki et al., 1992). SWRs may occur during sleep or wakefulness, associated with immobility and consummatory behavior. During ripples, CA1 pyramidal cells may present the same firing pattern that occurred during the preceding exploratory period, a phenomenon known as “replay” (Wilson and McNaughton, 1994). It has recently been shown that the firing pattern of the neural assembly recorded during the exploratory period can also occur before the same exploratory period, a phenomenon known as “preplay” (Dragoi and Tonegawa, 2011). During movement, hippocampal theta oscillations are accompanied by the activities of place cells, head direction cells, and grid cells. Place cells are recorded in CA3 and CA1 and increase their firing rates in response to the specific spatial location of the animal. Head direction cells are recorded in several regions of the limbic system including the dorsal presubiculum, RC, EC, lateral mammillary nucleus, striatum, anterodorsal and laterodorsal thalamic nuclei, and dorsal tegmental nucleus, and increase their firing rates only when the animal’s head points to a specific direction. Grid cells are recorded in the medial EC, presubiculum, and parasubiculum, and have multiple firing fields that span the entire available space in a periodic hexagonal pattern (Buzsaki and Moser, 2013). It has been hypothesized that the activity of place cells, head direction cells, and grid cells is necessary to construct a cognitive representation of the surrounding environment. A rudimentary map of space is already present at P18 when the pup explores an open environment outside the nest for the first time (Langston et al., 2010; Wills et al., 2010). Head direction cells display adultlike properties from the beginning at P11, before eye opening (Bjerknes et al., 2014; Tan et al., 2017). Place cells are also present but develop more gradually, with grid cells showing the slowest development. By P28 all functional kinds of cells display full maturity.

9.3.3.1 Development of theta oscillations Connections between the HF and septum follow an early time course: septo-hippocampal fibers from the medial septum are detected as early as E15, with the first septo-hippocampal fibers arriving 2 days later, at E17 (Chapter 4 in

Organization and development of hippocampal circuits Chapter | 9

215

Andersen et al., 2007). This precedes the arrival of EC connections, and can be seen as in accord with the early maturation of GABAergic projections and a sign that oscillatory activity, through the septal nuclei, is established before EC-driven events. The time course for other extrinsic afferentsdfor example, from the amygdala, reuniens thalamus, and cortical areas to CA1dis not known, and how this compares with maturational properties of EC afferents is not clear. The hippocampus is extensively interconnected with multiple wider networks, associated with emotional or spatial aspects (e.g., amygdala, and the intertwined loops of the retrosplenial and parietal cortex, along with the subiculum, presubiculum, parasubiculum (Chapter 3 in Andersen et al., 2007)). In the context of coactivations, the supramammillary nucleus (SUM) is an important component of the septo-hippocampal system and participates in the genesis and control of hippocampal theta activity. SUM afferents, identified by immunocytochemistry, reach their targets in DG and CA2 at E109 in monkeys (of 165 days) and slightly earlier in humans (mid-gestation: Berger et al., 2001). Data are not available for rodents. The power and frequency of theta oscillations are low at P16, reaching adult levels by P26. Also, the speed-frequency relation increases as a function of age (Wills et al., 2010). Pharmacological inactivation of the cholinergic cells of the medial septum increases the amplitude of postnatal developmental theta bursts in the hippocampus (Janiesch et al., 2011). Adult prefrontal and hippocampal networks exhibit oscillatory coactivation. At neonatal ages, hippocampal theta bursts, from ventral CA1, are reported to drive the generation of theta-gamma oscillations in frontal cortex (FC). Communication from FC, in the absence of direct FC to hippocampal projections, is via direct projections from LEC to HF and reuniens to HF (Hartung et al., 2016). In the adult, enhanced coordination to CA1 theta rhythm has been demonstrated as in parallel with a similar enhanced coordination between CA1 and neurons in the SUM, but developmental trajectories are unknown (Ito et al., 2018; and, among others, Leranth et al., 1999 on entorhinal-septo-SUM network).

9.3.3.2 Development of SWR The incidence, duration, and power of ripples are also low at P12, reaching adult levels by P24 (Buhl and Buzsaki, 2005; Farooq and Dragoi, 2019). Interestingly, “preplay” firing sequences of CA1 pyramidal cell activity appear at P21, whereas “replay” firing sequences of CA1 pyramidal cell activity appear later at P28 (Farooq and Dragoi, 2019). 9.3.3.2.1 Summary Immediately after birth, the electrical properties of hippocampal cells generate a very high excitable state that progressively decreases until the third week. These changes in electrical properties are paralleled by the maturation of the dendritic arborization and interneurons. A crucial role in changing this high excitability state is played by the switch around P10 of the GABAa current, from excitatory to inhibitory. Before P10, the hippocampal circuit activity is characterized by the presence of recurrent GDP, which are supported by long-range projecting SOM interneurons, the “hub cells.” After P10, major hippocampal rhythms, such as theta oscillations and SWR, emerge progressively. The development of these rhythms is associated with the appearance of place cells, head direction cells, and later, grid cells (Fig. 9.6).

FIGURE 9.6 Graphs depicting progressive changes in physiological activity across postnatal days P0eP30. (A) Giant Depolarizing Potential (GDP) frequency decreases significantly at P10, caused by the switch in the polarity of the GABAergic effect. (B) Progressive decrease in input resistance (Ri), a measurement of cellular excitability, in DG granule cells (green), CA3 pyramidal cells (blue), and CA1 pyramidal cells (black). (C) Emergence of mature electrophysiological patterns (theta oscillations in red, and ripples in black).

216 PART | I Circuit development

9.4 Conclusion New aspects of hippocampal organization are rapidly coming into better focus. As we have briefly surveyed here, cell populations are being further defined in terms of molecular, anatomic, and physiological features; and progress is being made on how these are grouped in functional assemblies in adult. A key issue is how these microcircuits vary within and across hippocampal subdivisions and how these interact under task- and stage-specific conditions. Developmental timelines are particularly important as informing behavioral stages and network dynamics across postnatal development. Moreover, new data may be expected from probing how developmental fates and transcriptomes influence activity driven features of the neural organization. Fundamentally, the hippocampal subdivisions present as a tapestry of multiple, partially overlapping circuits, and we can look forward with confidence to an abundance of fascinating new results.

References Acsady, L., Kamondi, A., Sik, A., Freund, T., Buzsaki, G., 1998. GABAergic cells are the major postsynaptic targets of mossy fibers in the rat hippocampus. J. Neurosci. 18, 3386e3403. Aggleton, J.P., Christensen, K., 2015. The subiculum: the heart of the extended hippocampal System. Prog. Brain Res. 219, 65e82. Amaral, D.G., Dent, J.A., 1981. Development of the mossy fibers of the dentate gyrus: I. A light and electron microscopic study of the mossy fibers and their expansions. J. Comp. Neurol. 195, 51e86. Amaral, D.G., Kondo, K., Lavenex, P., 2014. An analysis of entorhinal cortex projections to the dentate gyrus, Hippocampus, and subiculum of the neonatal macaque monkey. J. Comp. Neurol. 522, 1485e1505. Andersen, P., Morris, R., Amaral, D., Bliss, T., O’Keefe, J.O. (Eds.), 2007. The Hippocampus Book. Oxford University Press, New York, pp. 1e832. Anstotz, M., Huang, H., Marchionni, I., Haumann, I., Maccaferri, G., Lubke, J., 2016. Development of the mossy fibers of the dentate gyrus: I. A light and electron microscopic study of the mossy fibers and their expansions. Cereb. Cereb. 26, 855e872. Belluardo, N., Mudò, G., Blum, M., Fuxe, K., 2000. Central nicotinic receptors, neurotrophic factors and neuroprotection. Behav. Brain Res. 113, 21e34. Ben-Ari, Y., Cherubini, E., Corradetti, R., Gaiarsa, J.L., 1989. Giant synaptic potentials in immature rat CA3 hippocampal neurones. J. Physiol. 416, 303e325. Ben-Ari, Y., 2001. Developing networks play a similar melody. Trends Neurosci. 24 (6), 353e360. Berger, B., Alvarez, C., Goldman-Rakic, P., 1993. Neurochemical development of the hippocampal region in the fetal rhesus monkey. I. Calcium-binding proteins, of the hippocampal region in early appearance of peptides, DARPP-32, and monoamine innervation in the entorhinal cortex during the first half of gestation (E47 to E90). Hippocampus 3, 279e305. Berger, B., Esclapez, M., Alvarez, C., Meyer, G., Catala, M., 2001. Human and monkey fetal brain development of the supramammillary-hippocampal projections: a system involved in the regulation of theta activity. J. Comp. Neurol. 429, 515e529. Berggaard, N., Bjerke, I.E., Paulsen, A.E.B., Hoang, L., Skogaker, N.E.T., Witter, M.P., van der Want, J.J.L., 2018. Development of parvalbuminexpressing basket terminals in layer II of the rat medial entorhinal cortex. eNeuro 5, 1e12. Bezaire, M.J., Soltesz, I., 2013. Quantitative assessment of CA1 local circuits: knowledge base for interneuron-pyramidal cell connectivity. Hippocampus 23, 751e785. Bjerknes, T.L., Dagslott, N.C., Moser, E.I., Moser, M.B., 2018. Path integration in place cells of developing rats. PNAS 115, E1637eE1646. Bjerknes, T.L., Moser, E.I., Moser, M.B., 2014. Representation of geometric borders in the developing rat. Neuron 82, 71e78. Blair, M.G., Nguyen, N.N., Albani, S.H., L’Etoile, M.M., Andrawis, M.M., Owen, L.M., Oliveira, R.F., Johnson, M.W., Purvis, D.L., Sanders, E.M., Stoneham, E.T., Xu, H., Dumas, T.C., 2013. Developmental changes in structural and functional properties of hippocampal AMPARs parallels the emergence of deliberative spatial navigation in juvenile rats. J. Neurosci. 33 (30), 12218e12228. Booker, S.A., Vida, I., 2018. Morphological diversity and connectivity of hippocampal interneurons. Cell Tissue Res. 373, 619e641. Buhl, D., Buzsaki, G., 2005. Developmental emergence of hippocampal fast-field “ripple” oscillations in the behaving rat pups. Neuroscience 134, 1423e1430. Burton, B.G., Economo, M.N., Lee, G.J., White, J.A., 2008. Development of theta rhythmicity in entorhinal stellate cells of the juvenile rat. J. Neurophysiol. 100, 3144e3157. Buzsaki, G., 2015. Hippocampal sharp wave-ripple: a cognitive biomarker for episodic memory and planning. Hippocampus 25, 1073e1188. Buzsáki, G., Horváth, Z., Urioste, R., Hetke, J., Wise, K., 1992. High-frequency network oscillation in the hippocampus. Science 256, 1025e1027. Buzsáki, G., Moser, E.I., 2013. Memory, navigation and theta rhythm in the hippocampal-entorhinal system. Nat Neurosci 16 (2), 130e138. Carsten, K.E., Phillips, M.L., Pozzo-Miller, L., Weinberg, R.J., Dudek, S.M., 2016. Perineuronal nets suppress plasticity of excitatory synapses pn CA2 pyramidal neurons. J. Neurosci. 36, 6312e6320. Cembrowski, M.S., Bachman, J.L., Wang, L., Sugino, K., Shields, B.C., Spruston, N., 2016. Spatial gene-expression gradients underlie prominent heterogeneity of CA1 pyramidal neurons. Neuron 89, 351e368. Cembrowski, M.S., Wang, L., Lemire, A.L., Copeland, M., DiLisio, S.F., Clements, J., Spruston, N., 2018. The subiculum is a patchwork of discrete subzones. Elife 7, 1e21. Ceranik, K., Deng, J., Heimrich, B., Lübke, J., Zhao, S., Förster, E., Frotscher, M., 2000. Development of the entorhino-hippocampal projection: guidance by Cajal-Retzius cell axons. Ann N Y Acad Sci 911, 43e54.

Organization and development of hippocampal circuits Chapter | 9

217

Cui, Z., Gerfen, C.R., Young, W.S., 2013. Hypothalamic and other connections with the dorsal CA2 area of the mouse hippocampus. J. Comp. Neurol. 521, 1844e1866. Danglot, L., Triller, A., Marty, S., 2006. The development of hippocampal interneurons in rodents. Hippocampus 16, 1032e1060. Deguchi, Y., Donato, F., Galimberti, I., Cabuy, E., CaroniP, 2011. Temporally matched subpopulations of selectively interconnected principal neurons in the hippocampus. Nat. Neurosci. 14, 495e504. Dolleman-van der Weel, M.J., Lopes da Silva, F.H., Witter, M.P., 2017. Interaction of nucleus reuniens and entorhinal cortex projections in the hippocampal field CA1 of the rat. Brain Struct. Funct. 222, 2421e2438. Dombeck, D.A., Harvey, C.D., Tian, L., Looger, L.L., Tank, D.W., 2010. Functional imaging of hippocampal place cells at cellular resolution during virtual navigation. Nat. Neurosci. 13, 1433e1440. Donato, F., Chowdhury, A., Lahr, M., Caroni, P., 2015. Eary- and late-born parvalbumin basket cell subpopulations exhibiting distinct regulation and roles in learning. Neuron 85, 770e786. Donato, F., Jacobsen, R.I., Moser, M.-B., Moser, E.I., 2017. Stellate cells drive the maturation of the entorhinal-hippocampal circuit. Science 355, 1e10. Dragoi, G., Tonegawa, S., 2011. Preplay of future place cell sequences by hippocampal cellular assemblies. Nature 469, 397e401. Du, J., Zhang, L., Weiser, M., Rudy, B., McBain, C.J., 1996. Developmental expression and functional characterization of the potassium-channel subunit Kv3.1b in parvalbumin-containing interneurons of the rat hippocampus. J. Neurosci 16, 506e518. Dudek, S.M., Alexander, G.M., Farris, S., 2016. Rediscovering area CA2: unique properties and functions. Nat. Neurosci. Rev. 17, 89e102. Farooq, U., Dragoi, G., 2019. Cholinergic control in developing prefrontal-hippocampal networks. Science 363, 168e173. Forster, E., Zhao, S., Frotscher, E., 2006. Laminating the hippocampus. Nat. Rev. Neurosci. 7, 259e267. Francavilla, R., Villette, V., Luo, X., Chamberland, S., Munoz-Pino, E., Camire, O., Wagner, K., Kis, V., Somogyi, P., Topolnik, L., 2018. Connectivity and network state-dependent recruitment of long-range VIP-GABAergic neurons in the mouse hippocampus. Nat. Commun. 9, 1e17. Freund, T.F., Antal, M., 1988. GABA-containing neurons in the septum control inhibitory interneurons in the hippocampus. Nature 336, 17e173. Freund, T.F., Buzsáki, G., 1996. Interneurons of the hippocampus. Hippocampus 6, 347e470. Gomez-di Cesare, C., Smith, K.L., Rice, F.L., Swann, J.W., 1997. Axonal remodeling during postnatal maturation of CA3 hippocampal pyramidal neurons. J. Comp. Neurol. 384, 165e180. Hahn, T.T., Sakmann, B., Mehta, M.R., 2006. Phase-locking of hippocampal interneurons’ membrane potential to neocortical up-down states. Nat. Neurosci. 9 (11), 1359e1361. Hartung, H., Brokmann, M.D., Poschel, B., De Feo, V., Hanganu-Opatz, I.L., 2016. Thalamic and entorhinal network activity differently modulates the functional development of prefrontalehippocampal interactions. J. Neurosci. 36, 3676e3690. Helton, T.D., Zhao, M., Farris, S., Dudek, S.M., 2019. Diversity of dendritic morphology and entorhinal cortex synaptic effectiveness in mouse CA2 pyramidal neurons. Hippocampus 29, 78e92. Hennou, S., Khalilov, I., Diabira, D., Ben-Ari, Y., Gozlan, H., July 2002. Early sequential formation of functional GABA(A) and glutamatergic synapses on CA1 interneurons of the rat foetal hippocampus. Eur. J. Neurosci. 16 (2), 197e208. Hitti, F.L., Siegelbaum, S.A., 2014. The hippocampal CA2 region is essential for social memory. Nature 508, 88e92. Hunt, D.L., Linaro, D., Si, B., Romani, S., Spruston, N., 2018. A novel pyramidal cell type promotes sharp-wave synchronization in the hippocampus. Nat. Neurosci. 21, 985e995. Ichinohe, N., Rockland, K.S., 2005. Zinc-enriched amygdalo- and hippocampo-cortical connections to the inferotemporal cortices in macaque monkey. Neurosci. Res. 53, 57e68. Ito, H.T., Moser, E.I., Moser, M.-B., 2018. Supramammillary nucleus modulates spike-time coordination in the prefrontal-thalamo-hippocampal circuit during navigation. Neuron 99, 576e587. Janiesch, P.C., Krüger, H.S., Pöschel, B., Hanganu-Opatz, I.L., 2011. Cholinergic control in developing prefrontal-hippocampal networks. J. Neurosci. 31 (49), 17955e17970. Johnson-Venkatesh, E.M., Khan, M.N., Murphy, G.G., Sutton, M.A., Umemori, H., 2015. Excitability governs neural development in a hippocampal region-specific manner. Development 142, 3879e3891. Kitamura, T., Pignatelli, M., Suh, J., Kohara, K., Yoshiki, A., Abe, K., Tonegawa, S., 2014. Island cells control temporal association memory. Science 343, 896e901. Klausberger, T., Somogyi, P., 2008. Neuronal diversity and temporal dynamics: the unity of hippocampal circuit operations. Science 321, 53e57. Kohara, K., Pignatelli, M., Rivest, A.J., Jung, H.Y., Kitamura, T., Suh, J., Frank, D., Kajikawa, K., Mise, N., Obata, Y., Wickersham, I.R., Tonegawa, S., 2014. Cell type-specific genetic and optogenetic tools reveal hippocampal CA2 circuits. Nat. Neurosci. 17, 269e279. Lang, U., Frotscher, M., 1990. Postnatal development of nonpyramidal neurons in the rat hippocampus (areas CA1 and CA3): a combined Golgi/electron microscope study. Anat. Embryol. (Berl) 181 (6), 533e545. Langston, R.F., Ainge, J.A., Couey, J.J., Canto, C.B., Bjerknes, T.L., Witter, M.P., Moser, E.I., Moser, M.B., 2010. Development of the spatial representation system in the rat. Science 328 (5985), 1576e1580. Lavenex, P., Lavenex, P.B., 2013. Building hippocampal circuits to learn and remember: insights into the development of human memory. Behav. Brain Res. 254, 8e21. Lee, S.-H., Marchionni, I., Bezaire, M., Varga, C., Danielson, N., Lovett-Barron, M., Losonczy, A., Soltész, I., 2014. Parvalbumin-positive basket cells differentiate among hippocampal pyramidal cells. Neuron 82, 1129e1144. Leinekugel, X., Khazipov, R., Cannon, R., Hirase, H., Ben-Ari, Y., Buzsáki, G., 2002. Correlated bursts of activity in the neonatal hippocampus in vivo. Science 296 (5575), 2049e2052.

218 PART | I Circuit development

Leinekugel, X., 2003. Developmental patterns and plasticities: the hippocampal model. J. Physiol. Paris 97 (1), 27e37. Lensjo, K.K., Christensen, A.C., Tennoe, S., Fyhn, M., Hafting, T., 2017. Differential expression and cell-type specificity of perineuronal nets in Hippocampus, medial entorhinal cortex, and visual cortex examined in the rat and mouse. eNeuro 4, 1e18. Leranth, C., Carpi, D., Buzsaki, G., Kiss, J., 1999. The entorhino-septo-supramammillary nucleus connection in the rat: morphological basis of a feedback mechanism regulating hippocampal theta rhythm. Neuroscience 88, 701e718. Liu, X., Tilwalli, S., Ye, G., Lio, P.A., Pasternak, J.F., Trommer, B.L., 2000. Morphologic and electrophysiologic maturation in developing dentate gyrus granule cells. Brain Res. 856 (1e2), 202e212. Marchal, C., Mulle, C., 2004. Postnatal maturation of mossy fibre excitatory transmission in mouse CA3 pyramidal cells: a potential role for kainate receptors. J. Physiol. 15, 561. Martina, M., Schultz, J.H., Ehmke, H., Monyer, H., Jonas, P., 1998. Functional and molecular differences between voltage-gated Kþ channels of fastspiking interneurons and pyramidal neurons of rat hippocampus. J. Neurosci. 8 (20), 8111e8125. Marty, S., Wehrlé, R., Alvarez-Leefmans, F.J., Gasnier, B., Sotelo, C., 2002. Postnatal maturation of Naþ, Kþ, 2Cl- cotransporter expression and inhibitory synaptogenesis in the rat hippocampus: an immunocytochemical analysis. Eur. J. Neurosci. 15 (2), 233e245. Megıas, M., Emri, Z., Freund, T., Gulyas, A., 2001. Total number and distribution of inhibitory and excitatory synapses on hippocampal CA1 pyramidal cells. Neuroscience 102, 527e540. Mercer, A., Thomson, A.M., 2017. Cornu Ammonis regions e antecedents of cortical layers? Front. Neuroanat. 11 article 83. Minlebaev, M., Valeeva, G., Tcheremiskine, V., Coustillier, G., Khazipov, R., 2013. Cell-attached recordings of responses evoked by photorelease of GABA in the immature cortical neurons. Front. Cell Neurosci. 7, 83. Miao, C., Cao, Q., Moser, M.B., Moser, E.I., 2017. Parvalbumin and somatostatin interneurons control different Space- Coding networks in the medial entorhinal cortex. Cell 171, 507e521. Miyashta, T., Rockland, K.S., 2007. GABAergic projections from the hippocampus to the retrosplenial cortex in the rat. Europ. J. Neurosci. 26, 1193e1204. Mohedano-Moriano, A., Martinez-Marcos, A., Pro-Sistiaga, P., Blaizot, X., Arroyo-Jimenez, M.M., Marcos, P., Artacho-Pérula, E., Insausti, R., 2008. Convergence of unimodal and polymodal sensory input to the entorhinal cortex in the fascicularis monkey. Neuroscience 151, 255e271. O’Reilly, K.C., Flatberg, A., Islam, S., Olsen, L.C., Kruge, I.U., Witter, M.P., 2015. Identification of dorsaleventral hippocampal differentiation in neonatal rats. Brain Strct. Funct. 220, 2873e2893. Owens, D.F., Boyce, L.H., Davis, M.B., Kriegstein, A.R., 1996. Excitatory GABA responses in embryonic and neonatal cortical slices demonstrated by gramicidin perforated-patch recordings and calcium imaging. J. Neurosci. 16 (20), 6414e6423. Pedroni, A., Minh do, D., Mallamaci, A., Cherubini, E., 2014. Electrophysiological characterization of granule cells in the dentate gyrus immediately after birth. Front. Cell Neurosci. 14 (8), 44. Pelkey, K.A., Chittajallu, R., Craig, M.T., Tricoire, L., Wester, J.C., McBain, C.J., 2017. Hippocampal GABAergic inhibitory interneurons. Physiol. Rev. 97, 1619e1747. Picardo, M.A., Gigue, P., Nonifazi, P., Batista-Brito, R., Allene, C., Ribas, A., Fishell, G., Baude, A., Cossart, R., 2011. Pioneer GABA cells comprise a subpopulation of hub neurons in the developing hippocampus. Neuron 71, 695e709. Pignatelli, M., Beyeler, A., Leinekugel, X., 2011. Neural circuits underlying the generation of theta oscillations. J. Phsiol. Paris 106, 81e92. Quattrocolo, G., Maccaferri, G., 2014. Optogenetic activation of cajal-retzius cells reveals their glutamatergic output and a novel feedforward circuit in the developing mouse hippocampus. J. Neurosci. 34 (39), 13018e13032. Ray, S., Brecht, M., 2016. Structural development and dorsoventral maturation of the medial entorhinal cortex. Elife 5. Rihn, L.L., Claiborne, B.J., 1990. Dendritic growth and regression in rat dentate granule cells during late postnatal development. Brain Res. Dev. Brain Res. 54, 115e124. Rivera, C., Voipio, J., Payne, J.A., Ruusuvuori, E., Lahtinen, H., Lamsa, K., Pirvola, U., Saarma, M., Kaila, K., 1999. The Kþ/Cl- co-transporter KCC2 renders GABA hyperpolarizing during neuronal maturation. Nature 397, 251e255. Ropireddy, D., Scorcioni, R., Lasher, B., Buzsaki, G., Ascoli, G.A., 2011. Axonal morphometry of hippocampal pyramidal neurons semi-automatically reconstructed after in vivo labeling in different CA3 locations. Brain Struc. Funct. 216, 1e15. Rowland, D.C., Obenhaus, H.A., Skytøen, E.R., Zhang, O., Kentros, C.G., Moser, E.I., Moser, M.-B., 2018. Functional properties of stellate cells in medial entorhinal cortex layer II. eLIFE 7, 1e17. San Antonio, A., Liban, K., Ikar, T., Tsyganovskiy, E., Xu, W., 2014. Distinct physiological and developmental properties of hippocampal CA2 subfield revealed by using anti-purkinje cell protein 4 (PCP4) immunostaining. J. Comp. Neurol. 522, 1333e1354. The dentate gyrus. In: Scharfman, H.E. (Ed.), Prog. Brain Res. 163, 1e823. Schmidt, H., Gour, A., Straehle, J., Boergens, K.M., Brecht, M., Helmstaedter, M., 2017. Axonal synapse sorting in medial entorhinal cortex. Nature 549, 469e475. Serress, L., Ribak, C.E., 1988. The development of GABAergic neurons in the rat hippocampal formation. An immunocytochemical study. Brain Res. Dev. Brain Res. 44, 197e209. Serress, L., Ribak, C.E., 1990. Postnatal development of the light and electron microscopic features of basket cells in the hippocampal dentate gyrus of the rat. Anat. Embrol. (Berlin) 181, 547e565. Seress, L., Ribak, C.E., 1995. Postnatal development of CA3 pyramidal neurons and their afferents in the ammon’s horn of rhesus monkeys. Hippocampus 5, 217e231.

Organization and development of hippocampal circuits Chapter | 9

219

Slomianka, L., Geneser, F.A., 1997. Postnatal development of zinc-containing cells and neuropil in the hippocampal region of the mouse. Hippocampus 7, 321e340. Slomianka, L., Amrein, I., Knuesel, I., Sorensen, J.C., Wolfer, D.P., 2011. Hippocampal pyramidal cells: the reemergence of cortical lamination. Brain Struc. Funct. 216, 301e317. Soltesz, I., Losonczy, A., 2018. CA1 pyramidal cell diversity enabling parallel information processing in the hippocampus. Nat. Neursci. 21, 484e493. Strange, B.A., Witter, M.P., Lein, E.S., Moser, E.I., 2014. Functional organization of the hippocampal longitudinal axis. Nat. Rev. Neurosci. 15, 655e669. Super, H., Soriano, E., 1994. The organization of the embryonic and early postnatal murine Hippocampus. 11. Development of entorhinal, commissural, and Septa1 connections studied with the lipophilic tracer DiI. J. Comp. Neurol. 344, 101e120. Surges, R., Brewster, A.L., Bender, R.A., Beck, H., Feuerstein, T.J., Baram, T.Z., 2006. Regulated expression of HCN channels and cAMP levels shape the properties of the h current in developing rat hippocampus. Eur. J. Neurosci. 24 (1), 94e104. Tamamaki, N., Nojyo, Y., 1993. Projection of the entorhinal layer II neurons in the rat as revealed by intracellular pressure-injection of neurobiotin. Hippocampus 3 (4), 471e480. Tan, H.M., Wills, T.J., Cacucci, F., 2017. The development of spatial and memory circuits in the rat. Wiley Interdiscip. Rev. Cogn. Sci. 8, 1e16. Tansey, E.P., Chow, A., Rudy, B., McBain, C.J., 2002. Developmental expression of potassium-channel subunit Kv3.2 within subpopulations of mouse hippocampal inhibitory interneurons. Hippocampus 12, 137e148. Tricoire, L., Pelkey, K.A., Erkkila, B.E., Jeffries, B.W., Yuan, X., McBain, C.J., 2011. A blueprint for the spatiotemporal origins of mouse hippocampal interneuron diversity. J. Neurosci. 31, 10948e10970. Tyzio, R., Ivanov, A., Bernard, C., Holmes, G.L., Ben-Ari, Y., Khazipov, R., 2003. Membrane potential of CA3 hippocampal pyramidal cells during postnatal development. J. Neurophysiol. 90 (5), 2964e2972. Unal, G., Crump, M.G., Viney, T.J., Eltes, T., Katona, L., Klausberger, T., Somogyi, P., 2018. Spatio-temporal specialization of GABAergic septohippocampal neurons for rhythmic network activity. Brain Struc. Funct. 223, 2409e2432. Veleeva, G., Janackova, S., Nasretdinov, A., Rychkova, V., Makarov, R., Holmes, G.L., Khazipov, R., Lenck-Santini, P.P., 2019. Emergence of coordinated activity in the developing entorhinal-hippocampal network. Cereb. Cortex 29, 906e920. Valero, M., Menendex de la Prida, L., 2018. The hippocampus in depth: a sublayer-specific perspective of entorhinal-hippocampal function. Curr. Opin. Neurobiol. 52, 107e114. van Strien, N.M., Cappaert, N.L., Witter, M.P., 2009. The anatomy of memory: an interactive overview of the parahippocampal hippocampal network. Nat. Rev. Neurosci. 10, 272e282. Vasilyev, D.V., Barish, M.E., 2002. Postnatal development of the hyperpolarization-activated excitatory current Ih in mouse hippocampal pyramidal neurons. J. Neurosci. 22 (20), 8992e9004. Villette, V., Guigue, P., Picardo, M.A., Sousa, V.H., Leprince, E., Lachamp, P., Malveche, A., Tressard, T., Cossart, R., Baude, A., 2016. Development of early-born c-aminobutyric acid hub neurons in mouse Hippocampus from embryogenesis to adulthood. J. Comp. Neurol. 524, 2440e2461. Walker, M.C., Ruiz, A., Kullman, D.M., 2001. Monosynaptic GABAergic signaling from dentate to CA3 with a pharmacological and physiological profile typical of mossy fiber synapses. Neuron 29, 703e715. Wester, J.C., McBain, C.J., 2016. Interneurons differentially contribute to spontaneous network activity in the developing Hippocampus dependent on their embryonic lineage. J. Neurosci. 36 (9), 2646e2662. Wills, T.J., Cacucci, F., Burgess, N., O’Keefe, J., 2010. Development of the hippocampal cognitive map in preweanling rats. Sci. 328 (5985), 1573e1576. Wilson, M.A., McNaughton, B.L., 1994. Science. Reactivation of hippocampal ensemble memories during sleep, 265 (5172), 676e679. Witter, M.P., Kleven, H., Flatmoen, A.K., 2017. Comparative comtemplations on the hippocampus. Brain, Behav. Evol. 90, 15e24. Witter, M.P., Canto, C.B., Couey, J.J., Koganezawa, N., O’Reilly, K.C., 2014. Architecture of spatial circuits in the hippocampal region. Philos. Trans. R. Soc. B 369. Xu, H.T., Han, Z., Gao, P., He, S., Li, Z., Shi, W., Kodish, O., Shao, W., Brown, K.N., Huang, K., Shi, S.-H., 2014. Distinct lineage-dependent structural and functional organization of the hippocampus. Cell 157, 1552e1564. Yu, D.X., Marchetto, M.C., Gage, F.H., 2014. How to make a hippocampal dentate gyrus granule neuron. Development 141 (12), 2366e2375.

Chapter 10

Basal ganglia circuits Aryn H. Gittis1, Bryan M. Hooks2 and Charles R. Gerfen3 1

Department of Biological Sciences and Center for the Neural Basis of Cognition, Carnegie Mellon University, Pittsburgh, PA, United States;

2

Department of Neurobiology, University of Pittsburgh School of Medicine, Pittsburgh, PA, United States; 3Intramural Research Program, NIMH, Bethesda, MD, United States

Chapter outline 10.1. 10.2. 10.3. 10.4.

Introduction General organization of the basal ganglia Organization of corticostriatal projections The striatum 10.4.1. Physiology 10.4.2. Striatal interneurons 10.5. The effect of direct and indirect striatal output pathways on behavior

221 221 222 228 230 230

10.6. Direct and indirect striatal output pathways 10.7. External segment of the globus pallidus 10.8. The subthalamic nucleus 10.9. Development of the basal ganglia 10.10. Summary References

232 233 234 235 236 237

232

10.1 Introduction The basal ganglia connect the cerebral cortex to neuronal systems that affect a wide range of behaviors. Classically the basal ganglia are associated with affecting motor behavior since diseases involving them result in bradykinetic and dyskinetic movement disorders, such as Parkinson’s disease and Huntington’s chorea. Normal behavioral functions attributed to the basal ganglia are more wide ranging. This is due to the fact that as most cortical areas provide inputs to the basal ganglia, many behaviors involving the cerebral cortex are affected by the basal ganglia. Inputs from the cerebral cortex targeting the input nucleus of the basal ganglia, the striatum, are organized in parallel loops through the basal ganglia based on the cortical areas from which those inputs arise, which affect distinct behaviors (Alexander et al., 1986). The distinct behaviors these loops govern include motor learning, habit formation, and the selection of actions based on desirable outcomes (Balleine et al., 2007; Cisek and Kalaska, 2010; Graybiel et al., 1994; Hikosaka et al., 2000; Mink, 1996, 2003; Redgrave et al., 1999; Wichmann and DeLong, 2003; Yin and Knowlton, 2006). Prefrontal and cingulate cortical areas project to the dorsal medial striatum, which is implicated in motor learning coding actioneoutcome associations. Sensorimotor cortical areas project to the dorsolateral striatum, which is implicated in habit formation that code stimuluseresponse associations (Yin et al., 2009; Redgrave et al., 2010). The basal ganglia operate in parallel with other corticofugal outputs. These may have a more primary role in the actual generation of behavior. For example, the frontal cortical areas are involved in the planning and execution of movement and include direct corticospinal projections as well as input to basal ganglia. Understanding the role of the basal ganglia in behavior depends on determining the computations that are performed on cortical information through its neuroanatomical circuits.

10.2 General organization of the basal ganglia The general organization of the basal ganglia is diagrammed in Fig. 10.1. The major input nucleus is the striatum, which in the mouse consists of the dorsal striatum (also referred to as the caudateeputamen) and the ventral striatum, which includes the nucleus accumbens. The major glutamatergic excitatory input to the striatum arises from layer 5 and to a lesser extent from layer 2/3 of virtually all neocortical areas, from allocortical areas, and from the basolateral amygdala.

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00010-X Copyright © 2020 Elsevier Inc. All rights reserved.

221

222 PART | I Circuit development

FIGURE 10.1 Major circuits of the basal ganglia. Corticostriatal inputs arise from two major subtypes: intratelencephalic (IT) corticostriatal neurons, which provide bilateral inputs to the striatum, and pyramidal tract (PT) corticostriatal neurons, which project an axon ipsilaterally with collaterals to the striatum, thalamus, subthalamic nucleus (STN), superior colliculus (SC), pons, and spinal cord. Two main subtypes of projection neurons in the striatum give rise to direct and indirect pathways. The direct pathway provides direct projections to the output nuclei of the basal ganglia, the internal segment of the globus pallidus (GPi), and the substantia nigra pars reticulata (SNr). The indirect pathway projects to the external segment of the globus pallidus (GPe), which connects indirectly through the STN to the GPi and SNr. The major output of the basal ganglia originates from GABAergic neurons in the GPi and SNr, which provide inhibitory input to the thalamus, SC, and pedunculopontine nucleus (PPN). Adapted from Gerfen, C.R., Paletzki, R., Heintz, N., 2013. GENSAT BAC Cre-recombinase driver lines to study the functional organization of cerebral cortical and basal ganglia circuits. Neuron 80, 1368e1383.

Neocortical areas project mainly to the dorsal striatum, whereas allocortical areas project to the ventral striatum. Glutamatergic excitatory inputs to the striatum also arise from the intralaminar thalamic nuclear complex, including the parafascicular nucleus. Outputs from the basal ganglia arise from GABAergic neurons in the internal segment of the globus pallidus (GPi, also referred to in mice as the entopeduncular nucleus, EP) and the substantia nigra pars reticulata (SNr). These nuclei provide tonic inhibitory input to thalamic nuclei including the ventromedial and paralamellar mediodorsal nuclei, which project to frontal motor related areas, to the intermediate layers of the superior colliculus, and to the pedunculopontine nucleus. The main input nucleus of the basal ganglia, the striatum, comprises GABAergic spiny projection neurons (SPNs), which constitute some 90% of striatal neurons, and various interneuron subtypes, including cholinergic, parvalbumin, somatostatin, vasoactive intestinal peptide, and other subtypes. The SPNs provide the only output pathways from the striatum and are divided into two main subtypes based on their axonal projections, which are intermingled throughout the striatum. One subtype gives rise to projections directly to the output nuclei of the basal ganglia, the GPi and SNr, which are referred to as the direct pathway SPNs (dSPNs). These neurons also have a collateral to the external segment of the globus pallidus (GPe). The other subtype gives rise to the indirect pathway (iSPNs). The axonal projections of these neurons target only the GPe. Neurons in the GPe, which are GABAergic project to the subthalamic nucleus (STN), whose excitatory neurons project to the GPi and SNr, such that projections from iSPNs are indirectly connected to the output nuclei of the basal ganglia. While the glutamatergic and GABAergic neurotransmitters in these circuits may be classified as fast acting, other inputs to the striatum are considered to have a neuromodulatory function. The most important of these is the dense dopaminergic projection from the midbrain dopamine neurons to the striatum, along with a sparser serotonergic input from the dorsal raphe and noradrenergic input from the locus coeruleus. In summary, inputs to the basal ganglia arise from glutamatergic excitatory neurons in the cerebral cortex and thalamus that drive the activity of SPNs in the striatum. SPN projections through the direct and indirect striatal output pathways regulate the GABAergic inhibitory output of the basal ganglia directed to thalamic nuclei projecting to frontal motor cortical areas and to midbrain motor effector circuits originating in the superior colliculus and pedunculopontine midbrain nuclei.

10.3 Organization of corticostriatal projections The striatum is the main input structure of the basal ganglia. The vast majority of its neurons are SPNs, whose activity is determined by excitatory inputs from the cerebral cortex and thalamus. Consequently, the information that striatal

Basal ganglia circuits Chapter | 10

223

projection neurons process within the circuits of the basal ganglia is largely determined by the organization of activity of corticostriatal (and thalamostriatal) inputs. Corticostriatal projections arise mainly from layer 5 and to a smaller degree from layer 2/3 pyramidal neurons. There are two main subtypes of corticostriatal neurons (Cowan and Wilson, 1994; Wilson, 1987; Wise and Jones, 1977; Zheng and Wilson, 2002; Shepherd, 2013). One is the intratelencephalic (IT) subtype with axon collateral projections distributed bilaterally and confined to the cerebral cortex and striatum. The somata of these cells are located in layer 2/3 and for the most part in upper layer 5. The other is the pyramidal tract (PT) subtype with axon collaterals projecting ipsilaterally within the cerebral cortex and striatum and subcortically to the thalamus, STN, superior colliculus, and other midbrain structures, to brain stem nuclei and some to the spinal cord. The somata of these cells are located primarily in lower parts of layer 5 (Kita and Kita, 2012). Corticostriatal projections are organized in a general topographic manner such that the spatial relationships of the areas of the cerebral cortex are maintained in the projection of neurons in these areas to the striatum (Carman et al., 1963; Kemp and Powell, 1970; Webster, 1961). However, these projections do not map precisely in that the projections from individual cortical areas are distributed broadly with considerable overlap of projections from different cortical areas. The patterns of overlap of corticostriatal projections are critical in determining how information from the cortex is integrated through basal ganglia circuits. A long-standing model is that cortical areas that are interconnected with each other provide converging inputs to the striatum (Alexander et al., 1986). In their model based on data from primates, corticobasal ganglia circuits include parallel motor, oculomotor, and prefrontal pathways. Similar parallel circuits have been identified in the mouse that include distinct parallel sensorimotor and association pathways that target different topographically related regions of the striatum (Hintiryan et al., 2016). The organization of corticostriatal projections into parallel functional circuits related to the mapping of interconnected cortical areas into the striatum underlies the functional roles of different areas of the striatum. For example, the dorsal medial striatum, which receives inputs from prefrontal and cingulate cortical areas, is implicated in motor learning coding of actioneoutcome associations. The dorsolateral striatum, which receives input from sensorimotor cortical areas, is implicated in habit formation that code stimuluseresponse associations (Yin et al., 2009; Redgrave et al., 2010). While the general principles governing the organization of cortical input to striatal regions are related to the topographic mapping of the organization of corticocortical connections of functionally related areas, the organization of these inputs is more complex due to differences in the specific projections from different cortical areas and of subtypes of corticostriatal neurons within these areas. The precision of the topographic projections from the cortex to the striatum varies dependent on the cortical origin (Hooks et al., 2018). In general, the topographic relationship between different cortical areas is maintained in their projections to the striatum such that lateral cortical areas project into the lateral striatum compared with medial areas. The projections from specific cortical areas, such as the somatosensory and motor areas, are each organized topographically. For example, the somatosensory cortex is organized somatotopically with distinct mouth, barrel field, and upper and lower limb regions, and the projections of these areas to the striatum maintain their somatotopic organization. Similarly, projections from primary and secondary motor areas each also project in a topographic manner. These patterns of somatotopic corticostriatal projections are apparent by comparing the projections from discrete areas within each of these areas and applying a filter that ascribes the area within the striatum that receives the most dominant input (Fig. 10.2). The topographic organization of cortical projections is also maintained in projections to the thalamus, superior colliculus, and pontine nuclei. This type of mapping reveals the topographic organization of the dominant projections from discrete cortical areas. However, the projections from each discrete cortical area distribute over a larger domain, overlapping with projections from other cortical areas. Analysis of the overlap from different cortical injections within functional cortical areas dependent on the distance between the injection sites provides a measure of the precision of their somatotopic organization within the striatum. Such analysis demonstrates a progression of decreasing somatotopic precision from primary somatosensory cortex to primary and then secondary motor cortical areas (Hooks et al., 2018). The precision of projections to the striatum from these different cortical areas reflects the information they contain. Somatosensory cortical areas provide information related to specific areas of the animal’s body surface, whereas motor cortical areas encode information related to movements each of which involve multiple muscle groups. Secondary motor cortical areas encode more complex information related to actions-associated specific stimuli and preparatory activity. Thus, the information encoded from sensory to motor and secondary motor areas is progressively more complex, and there is a corresponding progressive distribution of this information into the striatum. Another principle of the organization of corticostriatal projections is that cortical areas connected with each other provide convergent inputs to the striatum (Alexander et al., 1986). This organization was studied in detail in a study in which multiple AAV-Cre-dependent constructs with different fluorophore labels were injected into the vibrissae somatosensory area (vS1) and into either the vibrissae motor area (vM1) or the forelimb motor area (fM1) using transgenic approaches to target only cortical IT neurons. IT-type neuronal projections were analyzed for the amount of overlap

224 PART | I Circuit development

FIGURE 10.2 Topographic projections in corticostriatal projections. Injection cases into five sites in the primary somatosensory cortex from the Allen Institute Mouse Connectivity Database and their axonal projections are mapped in the Common Coordinate Framework mouse reference atlas (Oh et al., 2014). Injections of five sites in the primary somatosensory cortex (SSp injections) are color coded with sites in the mouth region (blue), upper limb (red), and lower limb (green) and two sites in the barrel field (yellow and purple). In the sagittal plane, the projections from the cortex are shown to project through the striatum to the thalamus and superior colliculus. Coronal sections at four levels through the striatum (A, B), thalamus (C), and midbrain at the level of the superior colliculus and pontine nucleus (D) show that the topographic relationship of the cortical areas in the SSp, MOp, and MOs is maintained in the axonal projections to the striatum (A,B), thalamus (C), and the superior colliculus and pontine nucleus (D). MOp, primary motor cortex; MOs, secondary motor cortex.

(Hooks et al., 2018). Results demonstrated that striatal projections from cortical area vS1, which project to vM1, and striatal projections from vM1 are highly correlated in their projections to the striatum (Fig. 10.3). On the other hand, striatal projections from cortical area vS1 and those from cortical area fM1, which does not receive inputs from vS1, are not correlated in their projections to the striatum (Fig. 10.3). These data support the concept that information from interconnected cortical areas is integrated by converging to the same area of the striatum. Studies of the organization of corticostriatal projections have mostly analyzed projections from populations of cortical neurons. Recently, microscopes capable of tracing axons of single neurons through the whole brain, a substantial technological advance, have provided data on the projections of over 100 single corticostriatal projecting neurons (Economo et al., 2016). Analysis of the axonal projections of individual layer 5 IT corticostriatal neurons from the same cortical area has very diverse axonal projections (Fig. 10.4). The projections of multiple single cortical neurons from the same area project to the striatum in general topographic manner. However, individual cortical neurons in the same area provide different patterns of axonal projections both to other cortical areas and to the striatum. When analyzed as a population, primary and secondary motor cortex (MOp and MOs) layer 5 IT cortical neurons provide bilateral projections to multiple cortical areas, including projections to other parts of motor cortex, to the primary and secondary somatosensory cortical areas, to the cingulate cortex, and to the ectorhinal cortex. However, individual neurons in MOp and MOs project to different subsets of these areas (Fig. 10.4). For example, for two adjacent MOp neurons, one projects bilaterally to rostral cortical pole and to the ectorhinal cortex, whereas the other projects bilaterally to the primary somatosensory cortex and

Basal ganglia circuits Chapter | 10

225

FIGURE 10.3 Corticostriatal projections share the topography of corticocortical connectivity. (A and D) Fluorescent tracing injection cases contrast nonsomatotopically aligned injection sites (A, green forelimb M1 and magenta whisker S1) with a pair of strongly connected cortical areas (D, magenta whisker S1 and green whisker M1 injections). The overlap in cortex and striatum is scored for correlation. (B) Cartoon illustrates injection location in coronal sections of mouse sensory and motor cortex as well as striatal targets. (C) Scatterplot of correlation scores for corticocortical connectivity (using injection site overlap) and corticostriatal connectivity for axonal projections. Example pair, drawn from intratelencephalic-type injection cases, is illustrated, with points of low (green) and high (teal) cocorrelation for the examples indicated on the scatterplot. Scatterplot represents data from pairwise comparison of w92 cases (N ¼ w4186 comparisons). Each individual point represents the corticostriatal correlations (x-axis) and cortical injection site correlation (y-axis) for a single pair of injection cases. Adapted from Hooks, M., Papale, A.E., Paletzki, M., Eastwood, B.S., Couey, J.J, Winnubst, J., Chandrashekar, J., Gerfen, C.R., 2018. Topographic precision in sensory and motor corticostriatal projections varies across cell type and cortical area Nat. Commun. 9,3549.

rostral cortical pole but not to the ectorhinal cortex. In addition, for both MOp and MOs neurons from the same area, there is diversity in the laminar distribution and the bilateral symmetry of their intracortical projections. There is a similar diversity of individual layer 5 IT neurons in their striatal projections. Corticostriatal projections from individual MOp and MOs cortical neurons differ in whether they distribute projections broadly to the striatum or to discrete areas and whether the projections are bilaterally symmetric. Some cases are bilaterally symmetric, and other cases project predominantly either ipsi- or contralaterally. Moreover, there is no apparent correlation between the pattern of corticocortical and corticostriatal axonal projections of individual cortical neurons. These data suggest that information encoded in individual cortical neurons in the same cortical area is diverse and that this diverse information is distributed to different cortical and striatal areas. Subtypes of cortical pyramidal neurons with distinct axonal projection patterns to which specific functions are ascribed have been described in layer 5 PT neurons in the rostral part of MOs, a region called anterior lateral motor (ALM) cortex (Economo et al., 2018). As a population, PT neurons are distributed in the deeper parts of layer 5 and provide projections to the striatum, thalamus, STN, superior colliculus, brain stem motor nuclei, and the spinal cord (Kita and Kita, 2012; Shepherd, 2013). Retrograde labels suggested that PT neurons projecting to the thalamus are distinct from those projecting to the brain stem and are distributed in upper and lower parts of the deep parts of layer 5B in the ALM (Economo et al., 2018). Sensory cortex similarly contains projection-specific subsets of PT neurons (Rojas-Piloni et al., 2017). Tracing of axonal projections of individual PT neurons in ALM confirmed that these two subtypes in fact projected either to the thalamus and to the brain stem. Both populations did provide axon collaterals to the superior colliculus. Moreover, thalamic and brain stem projecting PT neurons express different genetic markers. Significantly, while these PT subtypes in ALM are distributed in distinct layer 5 sublayers, they are intermingled in other cortical areas. Prior studies show that in a

226 PART | I Circuit development

FIGURE 10.4 Axonal projections of single intratelencephalic (IT) corticostriatal neurons. Tracings of axonal projections of single IT corticostriatal neurons demonstrate both topographic organization and diversity of subtypes (from Janelia MouseLight Project, Economo et al., 2016). Topographic organization is demonstrated by coloring axons of clusters of four MOp neurons red and four MOs neurons green and one MOs neuron in between these blue (AeD). (A) Top-down view shows the four MOp axons (red) are distributed within the cortex lateral to those of the four MOs axons (green), with the MOs neuron in between these clusters distributing axons in between. Coronal sections through rostral (B), middle (C), and caudal (D) levels through the striatum demonstrate a similar topographic organization in projections to the striatum. The axons of the cluster of four MOp neurons (red) distribute laterally within the striatum relative to those of the four MOs neurons (green). Top-down views of the axonal projections of the individual MOp and MOs neurons demonstrate the diversity of axonal projections to the cortex and striatum. Together, the cortical projections of four MOp neurons (E, F, G, H) project to the ectorhinal (ec), somatosensory (SS), and frontal pole (fp) cortical areas. But each projects to a different subset of these areas. Diversity of cortical projections of four MOs neurons to the striatum (I, J, K, L) demonstrates that some project in a bilaterally symmetric pattern but to rostral and caudal regions (I, J), whereas another projects principally to the contralateral striatum (K) and another projects to the ipsilateral striatum (L). MOp, primary motor cortex; MOs, secondary motor cortex.

two-alternative choice task in which a specific response is determined by different sensory cues, different neurons in ALM are active during different phases of the task. The phases include a sensory phase when the cue is presented, the preparatory phase during a delay during which the animal is selecting the appropriate action, and the initiation of the specific action (Li et al., 2015). The preparatory activity is generated by reciprocal connections between ALM and the thalamus (Guo et al., 2017). Activity in ALM PT neurons projecting to the thalamus is correlated with the preparatory phase, and that activity of neurons projecting to the brain stem is correlated with the initiation of the movement. These results demonstrate distinct functional roles for subtypes of PT neurons based on their axonal projection targets to the thalamus and brain stem (Fig. 10.5). There also appear to be distinct subtypes of corticostriatal neurons that selectively target specific components within the striatum. An example of such specific targeting relates to the striatal patch and matrix compartments. The distribution of SPNs in the striatum is rather homogenous lacking distinct cytoarchitectural structure that is evident in neocortical laminar organization. However, overlain on this homogeneity, there are two distinct macroscopic compartments, termed the striatal

Basal ganglia circuits Chapter | 10

227

FIGURE 10.5 Pyramidal tract (PT) corticostriatal subtypes. Axonal projections of subtypes of single PT neurons are compared with the projection of an intratelencephalic (IT) neuron from the secondary motor cortex (MOs) from the Janelia MouseLight Project (Economo et al., 2016; A-F: Economo et al., 2018; G-H: Hooks et al., 2018). (A,B) Top-down view of axonal tracings of MOs PT neurons that project to the striatum, thalamus, and superior colliculus (SC) but not to the medulla. (C, D) Top-down view of axonal tracings of MOs PT neurons that project to the striatum, SC, and medulla but not to the thalamus. Top-down (E) and sagittal (F) views combining tracings of thalamic (green and light green) and medulla (red and orange) projecting PT MOs neurons with tracing of MOs IT neuron (blue). Both thalamic and medulla projecting PT neurons project to the SC and pons. (G) Coronal view of location of somata in MOs of IT (blue), PT thalamic projecting (green and light green) and PT medulla projecting (red and orange) neurons. Somata of IT and PT thalamic projecting neurons are more superficial than PT medulla projecting neurons. (H) Coronal view of axonal tracings of IT (blue) and PT (green, light green, red, orange) in the striatum. IT axons distribute bilaterally in the cortex and striatum. PT axons distribute within the same area of the striatum that close by IT neurons do but with many fewer collaterals and branches.

patch and matrix, which are distinguished by patterns of neurochemical markers and by the organization of their inputs and outputs (Gerfen, 1984, 1992). These compartments were first identified based on the early developmental distribution of dopamine inputs to islands in the striatum and later by higher levels of acetylcholinesterase staining in the matrix (Graybiel and Ragsdale, 1978), by the distribution of m-opioid receptor binding in the patches (Herkenham and Pert, 1981), and by the calcium-binding protein calbindin being selectively expressed in SPNs in the matrix (Gerfen, 1985). The SPNs in the patch and matrix compartments target different parts of the substantia nigra, with matrix neurons principally targeting the

228 PART | I Circuit development

GABAergic neurons in the substantia nigra pars reticulata (SNr) and patch neurons targeting the ventral tier of dopamine neurons in the substantia nigra pars compacta (SNc) (Gerfen, 1985). Subsequent studies using rabies virus constructs to label input to SNc dopamine neurons have for the most part confirmed their selective innervation from the patch compartment (Watabe-Uchida et al., 2012) although there appears to be a significant number of matrix neurons in the ventral striatum that also innervate the SNc dopamine neurons (Smith et al., 2016). There also appear to be selective inputs from different subtypes of SNc dopamine neurons to the striatal patch and matrix compartments (Gerfen et al., 1987). The dorsal tier SNc neurons, which express calbindin, project selectively to the striatal matrix, whereas the ventral tier dopamine neurons provide input to the patch compartment. Cortical inputs also display specific projections to the patch and matrix compartments. Initial studies reported that inputs to the patch compartment arose mainly from allocortical areas such as the prelimbic cortex, whereas neocortical areas projected primarily to the matrix compartment (Gerfen, 1984; Donoghue and Herkenham, 1986). Subsequent studies showed that although there is a more pronounced input to the patch compartment from allocortical areas such as the prelimbic and anterior cingulate cortex, neocortical areas including the somatosensory and motor cortical areas also provide input to the matrix compartment (Gerfen, 1989). Conversely, input from these neocortical areas provides more input to the matrix, and periallocortical areas provide some inputs to the matrix compartment (Gerfen, 1989; Smith et al., 2016). There also appeared to be a relationship between the laminar origin of corticostriatal inputs to the compartments with more superficial neurons providing inputs to the matrix and deeper layer neurons providing inputs to the patch compartment (Gerfen, 1989). But using Cre driver mouse lines that express selectively in either the patch or matrix compartment and trans-synaptically labeling cortical afferents in starter neurons with rabies virus, the laminar origin of inputs to the two compartments was not observed to be segregated in different cortical layers in some cortical areas (Smith et al., 2016). This difference could be due to the fact that Cre expression in these lines is not exclusively restricted to only one compartment. Alternatively, as is the case with other corticostriatal subtypes with distinct axonal projections, such as the subtypes of PT neurons projecting to either the thalamus or brain stem, the laminar distribution of such projectionspecific subtypes may be segregated in different layers in some cortical areas and intermingled in other cortical areas (Economo et al., 2018). Regardless, the fact that injections of multiple types of anterograde tracers into the cortex that selectively label projections that are segregated to either the patch or matrix compartments demonstrates that there are subtypes of corticostriatal neurons that differentially target these compartments. This suggests that there are specific functional channels originating from different corticostriatal neurons through the striatal patch and matrix compartment to selectively affect either the dopamine neurons in the SNc or the GABAergic output neurons in the SNr (Gerfen, 1992). The organization of corticostriatal projections to the striatal patch and matrix compartments identifies functional channels through the striatum at a macroscopic level. Another question is whether distinct corticostriatal neuron subtypes specifically target direct and indirect SPNs, which are intermingled. In an initial study, it was reported that IT and PT corticostriatal neurons preferentially innervate dSPNs and iSPNs, respectively (Lei et al., 2004). Follow-up studies demonstrated that both IT and PT corticostriatal neurons innervate both dSPNs and iSPNs but that the relative number of axospinous inputs differs such that there is a greater number of IT corticostriatal synapses on dSPNs and a greater number of PT corticostriatal synapses on iSPNs (Deng et al., 2015). This is consistent with data from physiologic studies that demonstrated that both dSPNs and iSPNs receive convergent input from both IT and PT corticostriatal neurons (Kress et al., 2013). Another study has suggested that the relative innervation of dSPNs and iSPNs from corticostriatal neurons varies dependent on the cortical area of origin with limbic and sensory cortical areas preferentially targeting dSPNs and motor cortical areas targeting iSPNs (Wall et al., 2013). Taken together, these data suggest that there appears to be some specificity in the innervation of dSPNs and iSPNs by IT and PT corticostriatal neurons and that there are also differences in the cortical area of origin of these inputs. Given the diversity of the projections of individual corticostriatal neurons, several questions remain. One is whether individual corticostriatal neurons project selectively to dSPNs or iSPNs or whether there are individual neurons that project to both. Additionally, differences in the projection patterns of individual corticostriatal neurons from sensory and limbic areas compared with motor areas to dSPNs and iSPNs will be important to determine the information provided from the cortex that regulates the activity of the striatal output pathways.

10.4 The striatum The main neuron subtype in the striatum is the spiny projection neuron (SPN), which accounts for some 90%e95% of striatal neurons. These neurons are homogenously distributed such that the striatum lacks a distinct cytoarchitectural organization, contrasted with the laminar organization of the cerebral cortex. In the primate, the striatum comprises the caudate and putamen nuclei, which are separated by the corticofugal fiber bundles that coalesce as the internal capsule. In the rodent, corticofugal fiber bundles do not fasciculate such that there is no clear distinction of the caudate and putamen

Basal ganglia circuits Chapter | 10

229

nuclei in the dorsal striatum. The dorsomedial and dorsolateral parts of the striatum are generally comparable with the primate caudate and putamen nuclei, respectively, based on the organization of the corticostriatal inputs to these regions from neocortical areas. Allocortical and limbic cortical areas project primarily to the ventral parts of the striatum, including the nucleus accumbens, although there is no clear boundary that definitively separates the dorsal and ventral striatum. Spiny projection neurons (Fig. 10.6) have a cell body that is approximately 12e20 mm in diameter from which radiate 7e10 moderately branched dendrites that are densely covered with spines (Bishop et al., 1982; Chang et al., 1982; DiFiglia et al., 1976; Wilson and Groves, 1980). These neurons have a local axon collateral that is distributed in an area approximately 300 mm around the cell body that covers an area roughly equivalent to that of their dendrites, though not necessarily overlapping (Bishop et al., 1982; Kawaguchi et al., 1990). A small number of SPNs have a local axon collateral that extends as much as 1 mm from the cell body (Kawaguchi et al., 1990). SPNs are the only striatal neuron that extend axons that project out of the striatum. There are two main subtypes based on the targets of these projections. One subtype has an axon that projects exclusively to the GPe. The other SPN subtype has an axon collateral that courses through the GPe, extending a small collateral branch into the nucleus, and another axon collateral that projects to the GPi and SNr/SNc. As these latter SPN neurons project to the output nuclei of the basal ganglia, the GPi and SNr, they give rise to the socalled direct striatal projection pathway. The SPN neurons that project exclusively to the GPe give rise to the indirect striatal projection pathway, as they affect the basal ganglia output through multisynaptic circuits. Neurons giving rise to the direct striatal pathway (dSPNs) and indirect striatal pathway (iSPNs) are both GABAergic providing inhibitory inputs to their targets. However, dSPNs and iSPNs are distinguished from each other by the expression of other proteins and peptides. Significantly, dSPNs express the D1 dopamine receptor (Drd1a), whereas iSPNs

FIGURE 10.6 The striatal spiny projection neuron (SPN). (A) Photomicrograph of a single SPN intracellularly filled with biocytin. (A’) High magnification of the filled SPN in (A). (B) Tracings of an indirect and a direct pathway SPN drawn in place on a sagittal brain diagram (from Kawaguchi et al., 1990). The indirect pathway SPN has a projection axon that extends into the GPe, where it arborizes extensively, but does not extend beyond this nucleus. Direct pathway SPNs have projection axons that extend some collaterals into the GPe and project to the GPi and the SNr. Higher magnification (upper right) of the indirect and direct pathway neurons shows their dendrites (red and green) and local axon collaterals within the striatum (orange and blue). GPe, external segment of the globus pallidus; GPi, internal segment of the globus pallidus; RR, retrorubral area; SNr, substantia nigra pars reticulata; VTA, ventral tegmental area. (C) Diagram summarizing the main features of the direct and indirect pathway SPNs. Both neurons are GABAergic and receive glutamatergic corticostriatal inputs. Direct pathway SPNs express the D1 receptor subtype, the Gs and Golf stimulatory Gproteins, as well as the neuropeptides substance P (SP) and dynorphin (DYN). These neurons project to the GPe, GPi, and SNr. Indirect pathway medium spiny projection neurons (MSNs) express the D2 receptor, the A2A adenosine receptor, and the neuropeptide enkephalin (ENK). The D2 receptor is coupled to the inhibitory Gi G-atein, whereas the A2A receptor is coupled to the stimulatory Golf G-protein. These neurons project only to the GPe. (D) Functional dissociation of direct and indirect pathway SPNs. In the dopamine-depleted striatum, D1 dopamine receptor stimulation results in phosphorylation of ERK1/2/MAP kinase (green immunoreactive neurons) selectively in direct pathway MSNs. Indirect pathway SPNs are labeled by localization of mRNA encoding ENK (red neurons). This functional dissociation reflects the differential expression of D1 and D2 receptor subtypes by direct pathway SPNs and indirect pathway MSNs, respectively (Gerfen et al., 2002).

230 PART | I Circuit development

express the D2 dopamine receptor (Drd2) (Gerfen et al., 1990). While initially there was some controversy as to whether Drd1a and Drd2 receptors are colocalized in SPNs (Surmeier et al., 1992), there is a general consensus that these receptors are segregated in all but a very small number of SPNs, with as few as 5% displaying coexpression mostly in the nucleus accumbens. The neuropeptides dynorphin and substance P coexpress in dSPNs, whereas enkephalin coexpresses in iSPNs (Beckstead and Kersey, 1985; Gerfen and Young, 1988; Haber and Watson, 1983; Young et al., 1986). The segregation of these peptides in dSPNs and iSPNs was important in early studies of the functional role of the direct and indirect pathways (Young et al., 1986; Gerfen et al., 1990). Recent advances in the ability to perform sequence analysis on expressed mRNAs on large populations of single neurons provide a detailed list of gene expression that reveals similarities and differences between dSPNs and iSPNs and identifies further subtypes of each (Saunders et al., 2018).

10.4.1 Physiology Spiny projection neurons constitute approximately 90%e95% of the neurons in the striatum and are the only neurons providing output projections. Consequently, activity of SPNs determines the effect the basal ganglia have on behavior. In their resting state, SPNs are dominated by an inwardly rectifying Kir2 Kþ channels that hold their membrane potential at about 90 mV, in the so-called down state during which they do not exhibit spiking activity (Shen et al., 2007; Wilson, 1993). In response to correlated glutamatergic synaptic input from the cortex, SPNs depolarize, bringing their membrane potential into an “up-state,” during which they may display spiking activity. Hyperpolarized, SPNs display few action potentials (Wilson and Kawaguchi, 1996; Day et al., 2008; Wilson, 1993). Single cortical neurons are estimated to make only a single synapse onto an individual SPN (Cowan and Wilson, 1994). While transition to the up-state is dependent on coherent corticostriatal inputs, it is affected by dopamine (Surmeier et al., 2007). Thus, the organization of corticostriatal inputs determines what information is processed through the basal ganglia. These spikes are typically not correlated with the transition to the up-state, which suggests that they are driven by an independent synaptic input (Stern et al., 1998). DA modulates the glutamatergic synapses responsible for the transition to the up-state and the ion channels controlling spiking. The qualitative features of the modulation depend on which DA receptor is being stimulated. As originally hypothesized on the basis of changes in gene expression induced by DA depletion (Gerfen et al., 1990), D2 receptor signaling impedes the up-state transition and diminishes up-state spiking in indirect pathway SPNs, whereas D1 receptor signaling does precisely the opposite in direct pathway SPNs (Surmeier et al., 2007). The role of dopamine in affecting activity of striatal SPNs has been studied extensively with demonstrations of diverse roles including synaptic plasticity (for review, see Gerfen and Surmeier, 2011) (Fig. 10.7).

10.4.2 Striatal interneurons Striatal interneurons, which give rise to axons that remain within the striatum, make up some 5%e10% of the striatal neuron population (Bishop et al., 1982; Chang et al., 1982; DiFiglia et al., 1976; Kemp and Powell, 1971). Two major subtypes are recognized: the large aspiny neuron that utilizes acetylcholine as a transmitter (Bolam et al., 1984; Kawaguchi, 1992, 1993; Wilson et al., 1990) and medium-sized aspiny GABAergic interneurons, of which there are numerous subtypes (Kawaguchi et al., 1995; Kita, 1993; Tepper and Koos, 2017; Tepper et al., 2010; Munoz-Manchado et al., 2018). The large aspiny cholinergic striatal interneurons have a very large cell body, up to 40 mm in diameter from which extend aspiny dendrites that may split into secondary and tertiary branches, which cover an area of over 1 mm. Axons from these neurons are extremely fine and extend over an area of as much as 2 mm. Cholinergic striatal interneurons fire tonically but exhibit burstepause firing patterns in response to behaviorally salient stimuli, which are thought to contribute to behavioral flexibility and the ability to learn associations between cues and action outcomes (Aosaki et al., 1994, 2010; Matsumoto et al., 2001; Schulz and Reynolds, 2013). Although cholinergic interneurons receive input from both cortical and thalamic sources (Doig et al., 2014; Guo et al., 2017; Klug et al., 2018), thalamic inputs, particularly from the parafascicular nucleus, are primarily responsible for shaping burstepause firing patterns (Assous et al., 2017; Ding et al., 2010; Lapper and Bolam, 1992; Meredith and Kang, 2006; Wilson et al., 1990). Cholinergic interneurons influence SPN activity through a variety of mechanisms including direct excitation through cholinergic signaling (Akins et al., 1990; Shen et al., 2005; Zucca et al., 2018), corelease of glutamate (Higley et al., 2011; Nelson et al., 2014a), and modulation of local inhibitory networks (English et al., 2011; Faust et al., 2016; Koos and Tepper, 2002; Lv et al., 2017; Nelson et al., 2014b). GABAergic interneurons in the striatum are composed of an increasingly diverse population of neurons that form a complex local inhibitory network within the striatum (Kawaguchi, 1993, 1995; Tepper and Koos, 2017). The best characterized class of GABAergic interneurons in the striatum are parvalbumin-expressing fast-spiking interneurons (PV or FSIs; Gerfen, 1985; Kubota et al., 1993). FSIs receive excitatory inputs predominantly from the cerebral cortex and make

Basal ganglia circuits Chapter | 10

231

FIGURE 10.7 Canonical basal ganglia microcircuits. (A) Corticostriatal and direct/indirect pathway canonical circuits. Layer 5 cortical pyramidal neurons provide excitatory glutamatergic inputs to the spines of striatal D1- and D2-receptor-expressing medium spiny projection neurons (D1-SPNs and D2-SPNs). D1-SPNs give rise to direct pathway projections to the output nuclei of the basal ganglia (GPi/SNr), whereas D2-MSNs give rise to indirect pathway projections to basal ganglia output nuclei. Dopamine input through the nigrostriatal pathway is directed to the spine necks of D1- and D2-SPNs to modulate corticostriatal inputs. (B). Feedforward, feedback, and intrinsic striatal circuits. One feedforward circuit involves FS, parvalbumin (PV)/GABAergic interneurons that provide perisomatic synapses on both D1- and D2-SPNs. These PV neurons receive excitatory inputs from layer 5 corticostriatal neurons and are inhibited by the GPe. Intralaminar thalamic nuclei provide inputs to D1- and D2-SPNs and contribute to a feedforward circuit involving thalamostriatal inputs to cholinergic (ChAT) interneurons that provide input to both D1- and D2-SPNs. Cholinergic neuron activity is also affected by dopamine inputs. Feedback striatal microcircuits involve interconnections between local axonal collaterals of D1- and D2-SPNs that make synaptic contact with other SPNs. GPi, internal segment of the globus pallidus; MSNs, medium spiny projection neurons; SNr, substantia nigra pars reticulata. Adapted from Gerfen, C.R., Surmeier, D.J., 2011. Modulation of striatal projection systems by dopamine. Annu. Rev. Neurosci. 34, 441e466.

dense, basket-like synapses that powerfully inhibit spiking in SPNs (Koos and Tepper, 1999; Gittis et al., 2010; Kita and Kitai, 1987; Planert et al., 2010; Bennett and Bolam, 1994; Choi et al., 2018). Reduced activity of striatal FSIs produces a spectrum of behavioral deficits, including motor abnormalities (Kalanithi et al., 2005; Gittis et al., 2011a; Rapanelli et al., 2017; Xu et al., 2016) and impaired learning and decision making (Owen et al., 2018; Lee et al., 2017, 2018; Friedman et al., 2017). Although FSIs are thought to mediate feedforward inhibition and regulate neuronal ensembles in the striatum (Parthasarathy and Graybiel, 1997; Koos and Tepper, 1999; Owen et al., 2018; Gittis et al., 2010, 2011b; Yu et al., 2017; O’Hare et al., 2017), evidence of these effects in vivo has been mixed (Gage et al., 2010; Lee et al., 2017). Their activity might be more heavily influenced by inhibition either from other interneurons (Tepper and Koos, 2017) or the GPe (Gittis et al., 2014), which might create more coordinated activity across the population and exert greater influence over SPNs. Several recent studies have provided new insights into the effects of striatal interneurons on SPNs. First, interneurons contributing to feedforward inhibition include both PV and low threshold spiking somatostatin (som) interneurons that affect activity of SPNs in distinct ways (Straub et al., 2016). While both receive inputs from the cerebral cortex and inhibit SPNs, PV neurons direct their input to proximal dendrites, whereas som neuron input is directed to distal dendrites, similar to the arrangement of inputs of these interneurons in the cortex onto cortical pyramidal neurons. Moreover, PV interneurons connect to SPNs that are in close proximity, whereas som interneurons target SPNs over a larger area. Second, in some cases, striatal interneurons provide feedback inhibition of SPN activity indirectly through connections between interneuron subtypes. An example of this are connections of cholinergic interneurons with Htr3a or neuroglioform (NGF) striatal GABAergic interneurons, which provide inhibition of SPNs (Faust et al., 2016). Third, there appears to be an inverse relationship between the source of inputs to striatal interneuron subtypes from the cerebral cortex and thalamus

232 PART | I Circuit development

(Assous et al., 2018). Inputs to PV and som interneurons providing feedforward inhibition arise principally from the cerebral cortex, whereas inputs to cholinergic and NGF interneurons providing feedback inhibition arise principally from the thalamus. Taken together, these findings suggest that the engagement of different striatal interneurons has distinct effects on shaping the activity of the striatal direct and indirect output pathways.

10.5 The effect of direct and indirect striatal output pathways on behavior For the past 30 years, the predominant theory that the effect of the basal ganglia has on movement behavior is dependent on the relative activity between the direct and indirect striatal output pathways. This concept originated from physiologic studies demonstrating that stimulation of the cerebral cortex results in activity of striatal SPNs whose inhibitory projections inhibit the tonic activity of GABAergic SNr neurons, which results in disinhibition of the targets of the SNr (Chevalier et al., 1985; Deniau and Chevalier, 1985). The effect of the disinhibition through the direct pathway of the SNr on behavior is demonstrated by the finding that pauses in the activity of SNr neurons and the corresponding activity in the superior colliculus are directly correlated with the generation of eye movements (Hikosaka and Wurtz, 1983). These studies suggested that activity through the direct striatal output pathway is responsible for the generation of movements. An influential paper by Albin et al. (1989) introduced the idea that clinical disorders resulting in dyskinetic movements are due to increased activity through the direct pathway, whereas bradykinetic disorders such as Parkinson’s disease result from increased activity in the indirect pathway. The demonstration that D1 and D2 dopamine receptors are segregated, respectively, in direct and indirect striatal pathway neurons provided the mechanism to account for how the relative increase in activity in the indirect pathway compared with the direct pathway occurs in Parkinson’s disease (Gerfen et al., 1990). In direct pathway neurons the D1 receptor is linked to the stimulatory G protein Gs, whereas in indirect pathway neurons, the D2 receptor is linked to the inhibitory G protein Gi. With the degeneration of dopamine input to the striatum, dopamine no longer provides a stimulatory effect on direct pathway neurons, and indirect pathway neurons would be disinhibited. These opponent effects on direct and indirect pathway neurons by dopamine are demonstrated in the dopamine-depleted striatum by decreased and increased gene expression, respectively, in direct and indirect pathway neurons, which are reversed by selective D1 and D2 agonist treatments (Gerfen et al., 1990). Additionally, it was demonstrated that lesions of the STN reverse the behavioral bradykinesia in nonhuman primates with dopamine depletion (Bergman et al., 1990). These findings support the idea that bradykinesia results from increased activity in the indirect pathway, leading to clinical treatments involving lesions or deep brain stimulation of the targets of the indirect pathway that effectively treated patients with Parkinson’s disease (DeLong, 1990). With the development of optogenetic techniques, it was demonstrated that select activation of the direct pathway resulted in increased locomotion, whereas conversely activation of the indirect pathway inhibited movement (Kravitz et al., 2010). While there is compelling support for the idea that activity through the direct pathway enhances movement behavior while activity through the indirect pathway suppresses behavior, other studies suggest that the relative activity in direct and indirect pathway neurons during normal movement behavior is more complex. In one such study, it was demonstrated that while nonhuman primates are performing a reaching task that there is increased activity in the GPi (Turner and Anderson, 1997), which would appear to contradict the idea that movements occur as a result of disinhibition of the direct striatal pathway. To explain this discrepancy, Mink (1996, 2003) proposed that normal movement requires selective activation through the direct pathway of circuits involved in a particular movement and inhibition through the indirect pathway of circuits involved in movements that would compete with the action selected. Recent studies employing optogenetic techniques to monitor the activity of direct and indirect pathway neurons demonstrate that during normal movement behavior, both pathways display increased activity (Cui et al., 2013; Barbera et al., 2016). For such results to support the model proposed by Mink (1996, 2003), it will be necessary to determine the specific effect of activity in direct and indirect striatal neurons on specific movements. Techniques to resolve this question have recently been introduced that involve determining the activity in these pathways in a wide range of movements (Markowitz et al., 2018; Cui et al., 2013). These studies demonstrate the necessity of determining the neuroanatomical circuits that are involved in regulating the activity in the direct and indirect pathways.

10.6 Direct and indirect striatal output pathways The output nuclei of the basal ganglia are the GPi and SNr, which comprise GABAergic neurons that are tonically active to provide inhibitory tone to their targets in the thalamus, superior colliculus, and midbrain motor nuclei including the PPN. Neurons in the striatum that project directly to these output nuclei constitute the direct striatal pathway (dSPNs) (Kawaguchi et al., 1990), which are distinguished by their expression of the D1 dopamine receptor and the peptides

Basal ganglia circuits Chapter | 10

233

dynorphin and substance P (Gerfen et al., 1990). Activity of dSPNs, driven by excitatory input from the cerebral cortex, results in inhibition of the tonic activity of the GPi and SNr and results in the disinhibition of the output targets of the basal ganglia (Chevalier et al., 1985; Deniau and Chevalier, 1985). As described earlier, numerous studies have demonstrated that activity through the direct pathway is correlated with increased motor behavior. The indirect pathway originates from striatal SPNs that project to the GPe (iSPNs), which do not extend an axon beyond this nucleus. This pathway is anatomically more complex than the direct pathway and contains two intermediary nuclei: the GPe and the STN. According to the classic model, activity of iSPNs inhibits GABAergic GPe neurons resulting in disinhibition of excitatory STN neurons, which increase the activity of GABAergic neurons in the GPi and SNr. This increase in the inhibitory output of the basal ganglia exerts an inhibitory effect on movement.

10.7 External segment of the globus pallidus The GPe is the direct recipient of inhibitory output from iSPNs in the striatum, and w90% of GPe neurons are inhibited by optogenetic stimulation of iSPNs (Cazorla et al., 2014). Because GPe neurons fire spontaneously, their inhibition has a net disinhibitory effect on basal ganglia output, mediated by the GPe’s projections to the STN (a glutamatergic nucleus that excites GPi/SNr) or GPi/SNr directly. This sign inversion within the indirect pathway predicts that it is the functional opposite of the direct pathway, forming the basis of the “rate model” in which the direct pathway facilitates movement and the indirect pathway suppresses movement (Albin et al., 1989; DeLong, 1990). In support of the rate model, optogenetic stimulation of iSPNs was shown to increase the firing rates of SNr neurons and cause freezing, whereas dSPN stimulation was shown to decrease the firing rates of SNr neurons and increase locomotion (Kravitz et al., 2010). Historically, the GPe has been described as a relay nucleus in the indirect pathway, whose main role is to invert the sign of striatal output before propagating the signal to downstream nuclei. However, this assumption is not well supported by the data. Pharmacological inhibition of the GPe has very little effect on movement (Horak and Anderson, 1984; Kato and Kimura, 1992; Soares et al., 2004), and its pharmacological activation produces a constellation of both motor and nonmotor behaviors (Francois et al., 2004; Grabli et al., 2004; Matsumura et al., 1995). Bicuculline injections into the posterior, ventrolateral “sensorimotor” territory of GPe produce abnormal movements, but injections into other territories of the GPe affect the organization of behavior, rather than movements themselves (Francois et al., 2004; Grabli et al., 2004). Specifically, bicuculline injections into the dorsomedial, “associative” territory of the GPe produced hyperactivity, defined as an increase in movements related to exploratory activity and disorganized task execution. In contrast, bicuculline injections into the anterior, ventral “limbic” territory of the GPe produced stereotypies: abnormal, persistent repetitions of an otherwise normal behavior such as licking or nail biting. The circuits responsible for the GPe’s diverse effects on behavior have not been fully elucidated, but both cfos and fMRI studies indicate that in addition to its effects on basal ganglia activity, the GPe also influences the activity of neurons in frontal cortex and limbic circuits, including the septum, amygdala, and hippocampus (Sawamura et al., 2002; Van Den Berge et al., 2017), suggesting that the GPe is much more than a simple relay nucleus. A key to understanding the GPe’s diverse effects on behavior is its different neuronal subpopulations. The GPe contains a number of different cell types, but only recently have the implications of this neuronal diversity for behavioral outcomes become apparent (Gittis et al., 2014; Hegeman et al., 2016; Saunders et al., 2018). The number and proportions of neuronal subpopulations in the GPe are an active area of research. For the purposes of this review, we will discuss three broad categories of GPe neurons: (1) “prototypical” GPe neurons, whose main projections target the STN, (2) “arkypallidal” GPe neurons, whose main projections target the striatum, and (3) “frontal cortex” GPe neurons, whose main projections target the frontal cortex. Prototypical GPe neurons are the most abundant category, making up approximately w75%e80% of all GPe neurons (Abdi et al., 2015; Dodson et al., 2015). They are characterized in vivo by their high, regular firing rates (30e70 Hz) (Mallet et al., 2012, 2016). Although the firing rates of most prototypical GPe neurons are modulated by movement, nearly half are inhibited by movement, whereas the other half are excited by movement (Dodson et al., 2015). Prototypical GPe neurons have also been found to be molecularly diverse: approximately w50% express high levels of parvalbumin (PV), whereas the other 50% are more enriched in lim homeobox 6 (Lhx6) (Abrahao and Lovinger, 2018; Hernandez et al., 2015; Mastro et al., 2014; Saunders et al., 2018). Although expression of these markers is graded across cells, creating partial overlap between PV and Lhx6 subpopulations at the molecular level, functionally these neuronal populations have been shown to be distinct at the physiological, anatomical, and behavioral level (Abrahao and Lovinger, 2018; Mastro et al., 2014; Mastro et al., 2017; Wallace et al., 2017). Arkypallidal GPe neurons are characterized by their dense, “netlike” innervation of the striatum and lack of canonical projections to downstream basal ganglia nuclei (Fujiyama et al., 2016; Mallet et al., 2012). A number of features distinguish these neurons from other types of GPe neurons. At the molecular level, arkypallidal neurons are molecularly

234 PART | I Circuit development

(A)

(B) FIGURE 10.8 Indirect pathway circuits. (A) Tracings of single GPe (orange/red) and STN (green/light green) neurons and their axonal projections are diagrammed in the sagittal plane. Single GPe neuron (dendrites in orange in sagittal section) and its axon (red) provide collaterals to the striatum, STN, and SNr. Single STN neuron (dendrites in green in sagittal section) and its axon (light green) provide collateral inputs to the GPe, GPi, and SNr. (B) Diagram of indirect pathway circuits. Pyramidal neurons in the cortex provide inputs to striatal D2 and D1 SPNs and to PV and som striatal interneurons. The indirect pathway originates from D2 expressing SPNs, which project to the GPe. The GPe contains two main subtypes. Arkypallidal (ArkyP) GPe neurons project back to D1 and D2 SPNs. Prototypical (ProtoP) GPe neurons provide GABAergic projections back to the striatum, which target primarily PV and som striatal interneurons and downstream projections to the STN and the GPi an SNr. The STN receives inhibitory input from ProtoP GPe neurons and provides excitatory input to dopamine neurons in the SNc and GABAergic neurons in the GPi/SNr. There are two main subtypes of neurons in the STN. Those expressing the a7 nicotinic receptor (nACHr) projects principally to the SNc, whereas those expressing the a4B2 nACHr project to the GPi/SNr. The direct pathway originates from D1 expressing striatal SPNs project to the GPe and to the GPi/SNr. Pyramidal tract (PT) neurons in the cortex provide direct inputs to the STN a4B2 nACHr neurons that comprise the hyperdirect cortical pathway. GPe, external segment of the globus pallidus; GPi, internal segment of the globus pallidus; SNr, substantia nigra pars reticulata; SPN, spiny projection neuron; STN, sunthalamic nucleus. Adapted from Kita, H., Chang, H.T., Kitai, S.T., 1983a. The morphology of intracellularly labeled rat subthalamic neurons: a light microscopic analysis. J. Comp. Neurol. 215, 245e257; Kita, H., Chang, H.T., Kitai, S.T., 1983b. Pallidal inputs to subthalamus: intracellular analysis. Brain Res. 264, 255e265.

well distinguished from other GPe neurons by their expression of FoxP2 and enkephalin (Abdi et al., 2015; Dodson et al., 2015; Hernandez et al., 2015; Mallet et al., 2012), as well as other markers (Saunders et al., 2018). Furthermore, they have been shown to have a distinct developmental origin compared with other classes of GPe neurons (Nobrega-Pereira et al., 2010). In vivo, arkypallidal neurons exhibit low firing, burstlike activity and are uniformly exited by movement (Dodson et al., 2015; Mallet et al., 2016). Paradoxically, however, chemogenetic activation of these neurons (targeted in the Npas1Cre mouse line) was shown to decrease movement (Glajch et al., 2016). A potential explanation for these seemingly contradictory data sets is that arkypallidal neurons might provide a Stop signal to the striatum that is critical for the cancellation of selected actions (Mallet et al., 2016). Finally, a subset of GPe neurons have been shown to project directly to the frontal cortex (FC-GPe) (Saunders et al., 2015). These FC-projecting GPe neurons are clustered along the ventral boundary of the GPe, and although all FC-GPe neurons are GABAergic, a substantial fraction (72%) also coexpress choline acetyltransferase (ChAT). Although some prototypical GPe neurons have also been shown to innervate the frontal cortex, projections from FC-GPe neurons are more dense and exhibit a distinct innervation pattern compared with other subtypes of GPe neurons. Additionally, because FCGPe neurons lack projections to other basal ganglia nuclei, they are thought to have minimal effects on the canonical indirect pathway (Fig. 10.8).

10.8 The subthalamic nucleus The STN is the only glutamatergic nucleus within the basal ganglia. It receives a dense GABAergic innervation from the GPe and projects to all major basal ganglia output nuclei: the GPi, SNr, and PPN. This anatomical organization suggests that the STN plays a key role in propagating information along the indirect pathway. This model is best supported by studies of basal ganglia pathophysiology under dopamine-depleted conditions. Under dopamine-depleted conditions, STN

Basal ganglia circuits Chapter | 10

235

neurons fire at abnormally high rates (due to disinhibition from the GPe), promoting hyperactivity in the GPi and excessive inhibition of movement (DeLong, 1990; Obeso et al., 2000). However, the STN’s role as a relay nucleus in the indirect pathway is complicated by the fact that it also receives dense innervation from the cortex (Coude et al., 2018; Nambu et al., 2000). Information from this “hyperdirect” pathway is propagated by the STN to nuclei both upstream and downstream in the basal ganglia circuit (Smith et al., 1998). Functional evidence of the STN’s dual role in both the indirect and hyperdirect pathways came from a classic study in which the physiological impact of brief cortical stimulations was traced throughout different basal ganglia nuclei (Nambu et al., 2000). At the level of basal ganglia output (GPi), cortical stimulation evoked a short-latency excitation, followed by inhibition, and then a late excitation. Dissection of the circuit origins of this triphasic signal revealed that the short-latency excitation was mediated by the hyperdirect pathway and the later, slower excitatory response was mediated by the indirect pathway. The dual circuitry of the STN gives this nucleus a more nuanced role in the suppression of actions than if it were simply a relay nucleus in the indirect pathway. One model posits that the STN’s role in stopping actions is mediated by a race model between the hyperdirect (Stop) and direct (Go) pathways (Schmidt et al., 2013). Cancellation of “stopped” actions is then mediated by STN recruitment of arkypallidal neurons in the GPe (Mallet et al., 2016). To date, there are few studies of neuronal diversity in the STN that might help to untangle the complex involvement of this nucleus in motor control. However, a recent study found that STN neurons could be subdivided on the basis of their expression of cholinergic receptors (Xiao et al., 2015). Intriguingly, those that expressed a4b2 nicotinic acetylcholine receptors (nAChR) received stronger glutamatergic inputs from the motor cortex but less from the GPe. These neurons in turn project most strongly to the SNr. Conversely, a separate subset of STN neurons expressing a7 nAChR received much denser innervation from the GPe than from cortex and project densely to SNc dopamine neurons, rather than the SNr. Although these molecular subdivisions within the STN have not yet been mapped onto function, they may provide key insights into future studies into the cellular pathways by which the STN regulates behavior.

10.9 Development of the basal ganglia Development of the basal ganglia and differentiation of specific neuron subtypes are determined by the complex interplay of numerous transcription factors during development (for review, see Long et al., 2009). The SPNs of the striatum develop from the germinal zone of the lateral ganglionic eminence (LGE), which migrate radially along radial glia to populate the striatum (Deacon et al., 1994; Olsson et al. 1997). Spiny projection neurons that form the patch and matrix striatal compartments are generated at different stages, with patch neurons being born earliest (in the rat between E13 and E15) and the matrix neurons born later (in the rat from E15 to E20) (van der Kooy and Fishell, 1987). The neurons forming the patches first migrate into the striatum and are homogenously distributed in the striatum between E19 and E21, with the later-born matrix neurons migrating later to restrict the early-born patch neurons to what becomes the adult patch compartment surrounded by the matrix (Fishell and van der Kooy, 1987). As both dSPNs and iSPNs are expressed in equal numbers in both the striatal patch and matrix compartments (Gerfen and Young, 1988), these two striatal subtypes do not appear to be born at different developmental times. The differential genesis of neurons forming the patch and matrix compartments is comparable with the development of cortical layers in which the earliest-born neurons occupy deeper layers with progressively older-born neurons forming more superficial layers (Gilmore and Herrup, 1997). The pattern of early-born cortical neurons forming deeper layers and early-born neurons forming the striatal patch compartment is consistent with the relationship of cortical inputs to the patches originating in deeper cortical layers than inputs to the matrix from more superficial cortical layers (Gerfen, 1989). Dopamine input to the striatum displays a progressive pattern related to the striatal patch and matrix compartments. When the striatal patches become evident around postnatal day 1, dopamine innervation invades the striatum, is first restricted to the striatal patch compartment, and, then during the first postnatal week to P20, innervates the matrix compartment (Moon Edley and Herkenham, 1984). This differential pattern of dopamine innervation is due to the origin of input arising from different subtypes of midbrain dopamine neurons, with dorsal tier dopamine neurons, which express calbindin, innervating the patch compartment, and ventral tier, calbindin-negative, dopamine neurons, innervating the matrix compartment (Gerfen et al., 1987b). The Gsh1/2 homeobox genes expressed in the LGE appear to be responsible for determining the specification of striatal neurons (Corbin et al., 2000; Toresson et al., 2000). Deletion of Gsh2 results in a decrease of the normal LGE genes that specify striatal neuron differentiation and expansion of telencephalic cortical markers. Gsh2 drives the expression of Dlx1/ 2 and Mash1 transcription factors that are necessary for normal striatal development (Toresson et al., 2000). Deletion of Dlx1/2 results in blockade of later-born striatal neurons that form the striatal matrix compartment, whereas the striatal patch

236 PART | I Circuit development

markers are not affected (Anderson et al., 1997). Deletion of Mash1 results in a profound reduction of neuronal progenitors in the ventral part of the LGE and all of the medial ganglionic eminence (MGE), although markers for SPNs are expressed normally (Casarosa et al., 1999). The transcription factor Ctip2, which is expressed in striatal SPN neurons and in cortical PT and corticothalamic neurons, appears to play an important role in the normal differentiation of the striatum (Arlotta et al., 2008). The role of Ctip2 in striatal development is apparent by P0 at which time its deletion is lethal. Deletion of Ctip2 does not affect the generation of either early-born patch or later-born matrix SPNs but does affect both the normal pattern of migration into the striatum, the compartmental organization of the striatum, and subsequent expression of gene markers that normally distinguish striatal patch and matrix SPNs. The GPe develops from neurons generated in both the MGE and LGE (Nobrega-Pereira et al., 2010). Prototypical GPe neurons, which express parvalbumin, are generated in the MGE and express the transcription factor Nkx2-1, whereas arkypallidal GPe neurons are generated in the LGE and express the transcription factor Foxp2 (Flandin et al., 2010; Dodson et al., 2015). Deletion of Nkx2-1 results in the failure of the generation of prototypical parvalbumin GPe neurons while not affecting the generation of arkypallidal NPas1 expressing GPe neurons (Flandin et al., 2010).

10.10 Summary The effect that the basal ganglia has on behavior is dependent on the information that is provided to the striatum, how that information is processed within the striatum and by other components of the basal ganglia circuits, and how that processed information effects activity of the output nuclei of the basal ganglia and the GABAergic neurons of the GPi and SNr. The organization of excitatory glutamatergic input from the cerebral cortex to the striatum is the primary determinant of what information is processed through the basal ganglia. There are several general principles that have emerged as to how this information is organized. First, there is a general topographic organization whereby spatial relationships of cortical areas are maintained in their projections to the striatum. This results in the existence of parallel functional loops through the basal ganglia, with information in each of these loops reflecting the function of the cortical areas associated with them (Alexander et al., 1986). The organization of information from cortical areas to the striatum varies such that corticostriatal input from somatosensory areas is more precisely organized than input from motor and association cortical areas. Second, although generally topographically organized, there is considerable overlap of projections from different cortical areas, which provides for integration of information within the striatum. The pattern of overlap of corticostriatal inputs maps intracortical connections such that interconnected cortical areas provide inputs to the same region of the striatum (Alexander et al., 1986). Third, different subtypes of corticostriatal neurons within each cortical area have distinct projection patterns. The best characterized of these are the IT and PT corticostriatal subtypes. The IT neurons are generally located in upper part of layer 5 and in layers 2/3 and provide axonal projections bilaterally within the cortex and to the striatum. The PT neurons are located in deeper parts of layer 5 and provide inputs ipsilaterally to the striatum, thalamus, STN, superior colliculus, brain stem, and spinal cord. Subtypes of PT neurons have been identified that project to the striatum and to different subset of other subcortical nuclei. For example, two PT subtypes in the ALM cortex project are distinguished by their projections to either the thalamus or brain stem, which are associated with distinct roles in preparatory activity and action initiation (Economo et al., 2018). Perhaps most illuminating regarding the organization of corticostriatal inputs are the diverse patterns of axonal projections of individual IT neurons. Neurons in close proximity to each other in the same cortical area display different patterns of axonal projections within the cortex, projecting to different subsets of distant cortical areas and in their projections to the striatum, both in terms of whether axons are distributed to discrete areas or widespread and in terms of their bilateral symmetry. The diversity of individual IT neurons in the same cortical area raises the possibility that these different neuron subtypes have distinct functional roles in ongoing behavior, reflecting the differences in their axonal projections similar to that, which has been identified for PT subtypes. Another feature of corticostriatal inputs are functional channels originating from subtypes of cortical pyramidal neurons that target specific components within the striatum (Morita et al., 2012). An example of this is the projection from different cortical neurons, either located in different layers of layer 5 or intermingled within the same layer, which project to either the striatal patch or matrix compartments (Gerfen, 1992). These macroscopic compartments themselves target distinct output targets: SNc dopamine neurons in the case of SPNs in the patches and SNr GABAergic neurons in the case of matrix SPNs. There also appear to be selective inputs from distinct cortical pyramidal subtypes to the dSPNs and iSPNs. The specificity of these connections differs for PT and IT corticostriatal neurons as well as for IT neurons, dependent on the cortical area of origin. Questions remain as to whether the diverse subtypes of IT corticostriatal originating from the same cortical area specifically target dSPNs and iSPNs or different striatal interneurons, including PV and som subtypes. The organization of the cortical input to the striatum provides the direct excitatory effects that drive the activity of dSPNs and iSPNs. However, other components of striatal microcircuits including striatal interneurons modify the activity

Basal ganglia circuits Chapter | 10

237

of the striatal output pathways. Two predominant interneuron striatal microcircuits have been well studied. The first include PV striatal interneurons (also corresponding to fast spiking interneurons), which receive inputs from the cerebral cortex and are considered to provide feedforward inhibition to SPNs. These neurons display different patterns of connections with SPNs, with PV neurons connected SPNs locally and targeting proximal dendrites, whereas som neurons connect with more distant SPNs and target more distal dendrites. The second class of striatal interneuron microcircuits include ChAT, NGF, and Htr3a interneurons. These receive inputs from thalamic nuclei and are considered to provide feedback inhibition to SPNs. In addition, dopaminergic inputs to the striatum have opponent effects on dSPNs and iSPNs due to their respective expression of D1 and D2 dopamine receptors. Together the effect of these interneuron feedforward and feedback microcircuits and dopamine shapes the activity of striatal SPNs that is provided by the excitatory inputs of the cortex. The effect that the processing and shaping of cortical input to the striatum has on behavior is dependent on the effect of the activity in the direct and indirect striatal output pathways on the output structures of the basal ganglia and the GABAergic neurons in the GPi and SNr. The effect of the direct pathway is rather straightforward as activity in this pathway inhibits the tonic activity of GPi and SNr GABAergic neurons, which function to disinhibit the targets of these neurons in the thalamus, superior colliculus, and PPN. The effect of the indirect pathway is more complex. According to the classic model, activity in iSPNs inhibits GABAergic neurons in the GPe, resulting in disinhibition of glutamatergic neurons in the STN, which increases the activity of the GABAergic inhibition of the GPi and SNr. However, this classic model does not incorporate the complexity of the indirect pathway circuits. First, there are different subtypes of GPe neurons, which have distinct connections. The prototypical GPe neuron subtypes, including both PV and Lhx neuron subtypes, provide feedback to the striatum, directed primarily to PV striatal interneurons, as well as downstream projections to the STN and SNr. A second GPe neuron subtype, the arkypallidal neuron, provides feedback to the striatum targeting primarily SPNs and does not project to downstream targets. Second, the cerebral cortex provides an important direct excitatory input to the STN through the so-called hyperdirect pathway. Interestingly, recent studies have suggested that the hyperdirect pathway targets a subset of STN neurons, which provide the major input to the GPi and SNr, whereas STN neurons receiving input from the GPe selectively target dopamine neurons in the SNc. Thus, the complexity of the circuits of the indirect pathway suggests that the model of activity of iSPNs increasing the inhibitory output of the basal ganglia is overly simplistic and has led to new insights into its functional role in behavior and neurologic disease (Gittis, 2018).

References Abdi, A., Mallet, N., Mohamed, F.Y., Sharott, A., Dodson, P.D., Nakamura, K.C., Suri, S., Avery, S.V., Larvin, J.T., Garas, F.N., et al., 2015. Prototypic and arkypallidal neurons in the dopamine-intact external globus pallidus. J. Neurosci. 35, 6667e6688. Abrahao, K.P., Lovinger, D.M., 2018. Classification of GABAergic neuron subtypes from the globus pallidus using wild-type and transgenic mice. J. Physiol. 596, 4219e4235. Akins, P.T., Surmeier, D.J., Kitai, S.T., 1990. Muscarinic modulation of a transient Kþ conductance in rat neostriatal neurons. Nature 344, 240e242. Albin, R.L., Young, A.B., Penney, J.B., 1989. The functional anatomy of basal ganglia disorders. Trends Neurosci. 12, 366e375. Alexander, G.E., DeLong, M.R., Strick, P.L., 1986. Parallel organization of functionally segregated circuits linking basal ganglia and cortex. Annu. Rev. Neurosci. 9, 357e381. Anderson, S.A., Qiu, M., Bulfone, A., Eisenstat, D.D., Meneses, J., Pedersen, R., Rubenstein, J.L., 1997. Mutations of the homeobox genes Dlx-1 and Dlx-2 disrupt the striatal subventricular zone and differentiation of late born striatal neurons. Neuron 19, 27e37. Aosaki, T., Tsubokawa, H., Ishida, A., Watanabe, K., Graybiel, A.M., Kimura, M., 1994. Responses of tonically active neurons in the primate’s striatum undergo systematic changes during behavioral sensorimotor conditioning. J. Neurosci. 14, 3969e3984. Aosaki, T., Miura, M., Suzuki, T., Nishimura, K., Masuda, M., 2010. Acetylcholine-dopamine balance hypothesis in the striatum: an update. Geriatr. Gerontol. Int. 10 (Suppl. 1), S148eS157. Arlotta, P., Molyneaux, B.J., Jabaudon, D., Yoshida, Y., Macklis, J.D., 2008. Ctip2 controls the differentiation of medium spiny neurons and the establishment of the cellular architecture of the striatum. J. Neurosci. 28 (3), 622e632. Assous, M., Kaminer, J., Shah, F., Garg, A., Koos, T., Tepper, J.M., 2017. Differential processing of thalamic information via distinct striatal interneuron circuits. Nat. Commun. 8, 15860. Assous, M., Faust, T.W., Assini, R., Shah, F., Sidibe, Y., Tepper, J.M., 2018. Identification and characterization of a novel spontaneously active bursty GABAergic interneuron in the mouse striatum. J. Neurosci. 38 (25), 5688e5699. Balleine, B.W., Delgado, M.R., Hikosaka, O., 2007. The role of the dorsal striatum in reward and decision-making. J. Neurosci. 27, 8161e8165. Barbera, G., Lian, B., Zhang, L., Gerfen, C.R., Culurciello, E., Chen, R., Li, Y., Lin, D.-T., 2016. Spatially compact neural cluster in the dorsal striatum encode locomotion relevant information. Neuron 92, 202e213. Beckstead, R.M., Kersey, K.S., 1985. Immunohistochemical demonstration of differential substance P-, Met-enkephalin-, and glutamic acid decarboxylase-containing cell and axon distributions in the corpus striatum of the cat. J. Comp. Neurol. 232, 481e498.

238 PART | I Circuit development

Bennett, B.D., Bolam, J.P., 1994. Synaptic input and output of parvalbumin-immunoreactive neurons in the neostriatum of the rat. Neuroscience 62, 707e719. Bergman, H., Wichmann, T., DeLong, M.R., 1990. Reversal of experimental parkinsonism by lesions of the subthalamic nucleus. Science 249 (4975), 1436e1438. Bishop, G.A., Chang, H.T., Kitai, S.T., 1982. Morphological and physiological properties of neostriatal neurons: an intracellular horseradish peroxidase study in the rat. Neuroscience 7, 179e191. Bolam, J.P., Wainer, B.H., Smith, A.D., 1984. Characterization of cholinergic neurons in the rat neostriatum. A combination of choline acetyltransferase immunocytochemistry, Golgi-impregnation and electron microscopy. Neuroscience 12, 711e718. Carman, J.B., Cowan, W.M., Powell, T.P.S., 1963. The organization of cortico-striate connexions in the rabbit. Brain 86, 525e562. Casarosa, S., Fode, C., Guillemot, F., 1999. Mash1 regulates neurogenesis in the ventral telencephalon. Development 126, 525e534. Cazorla, M., de Carvalho, F.D., Chohan, M.O., Shegda, M., Chuhma, N., Rayport, S., Ahmari, S.E., Moore, H., Kellendonk, C., 2014. Dopamine D2 receptors regulate the anatomical and functional balance of basal ganglia circuitry. Neuron 81, 153e164. Chang, H.T., Wilson, C.J., Kitai, S.T., 1982. A Golgi study of rat neostriatal neurons: light microscopic analysis. J. Comp. Neurol. 208, 107e126. Chevalier, G., Vacher, S., Deniau, J.M., Desban, M., 1985. Disinhibition as a basic process in the expression of striatal functions. I. The striato-nigral influence on tecto-spinal/tecto-diencephalic neurons. Brain Res. 334, 215e226. Choi, K., Holly, E.N., Davatolhagh, M.F., Beier, K.T., Fuccillo, M.V., 2018. Integrated anatomical and physiological mapping of striatal afferent projections. Eur. J. Neurosci. https://doi.org/10.1111/ejn.13829. Cisek, P., Kalaska, J.F., 2010. Neural mechanisms for interacting with a world full of action choices. Annu. Rev. Neurosci. 33, 269e298. Corbin, J.G., Gaiano, N., Machold, R.P., Langston, A., Fishell, G., 2000. The Gsh2 homeodomain gene controls multiple aspects of telencephalic development. Development 127, 5007e5020. Coude, D., Parent, A., Parent, M., 2018. Single-axon tracing of the corticosubthalamic hyperdirect pathway in primates. Brain Struct. Funct. 223, 3959e3973. Cowan, R.L., Wilson, C.J., 1994. Spontaneous firing patterns and axonal projections of single corticostriatal neurons in the rat medial agranular cortex. J. Neurophysiol. 71, 17e32. Cui, G., Jun, S.B., Jin, X., Pham, M.D., Vogel, S.S., Lovinger, D.M., Costa, R.M., 2013. Concurrent activation of striatal direct and indirect pathways during action initiation. Nature 494, 238e242. Day, M., Wokosin, D., Plotkin, J.L., Tian, X., Surmeier, D.J., 2008. Differential excitability and modulation of striatal medium spiny neuron dendrites. J. Neurosci. 28, 11603e11614. Deacon, T.W., Pakzaban, P., Isacson, O., 1994. The lateral ganglionic eminence is the origin of cells committed to striatal phenotypes: neural transplantation and developmental evidence. Brain Res. 668, 211e219. DeLong, M.R., 1990. Primate models of movement disorders of basal ganglia origin. Trends Neurosci. 13, 281e285. Deng, Y., Lanciego, J., Kerkerian-Le-Goff, L., Coulon, P., Salin, P., Kachidian, P., Lei, W., Del Mar, N., Reiner, A., April 14, 2015. Differential organization of cortical inputs to striatal projection neurons of the matrix compartment in rats. Front. Syst. Neurosci. 9, 51. Deniau, J.M., Chevalier, G., 1985. Disinhibition as a basic process in the expression of striatal functions. II. the striatonigral influence on thalamocortical cells of the ventromedial thalamic nucleus. Brain Res. 334, 227e233. DiFiglia, M., Pasik, P., Pasik, T., 1976. A Golgi study of neuronal types in the neostriatum of monkeys. Brain Res. 114, 245e256. Ding, J.B., Guzman, J.N., Peterson, J.D., Goldberg, J.A., Surmeier, D.J., 2010. Thalamic gating of corticostriatal signaling by cholinergic interneurons. Neuron 67, 294e307. Dodson, P.D., Larvin, J.T., Duffell, J.M., Garas, F.N., Doig, N.M., Kessaris, N., Duguid, I.C., Bogacz, R., Butt, S.J., Magill, P.J., 2015. Distinct developmental origins manifest in the specialized encoding of movement by adult neurons of the external globus pallidus. Neuron 86, 501e513. Doig, N.M., Magill, P.J., Apicella, P., Bolam, J.P., Sharott, A., 2014. Cortical and thalamic excitation mediate the multiphasic responses of striatal cholinergic interneurons to motivationally salient stimuli. J. Neurosci. 34, 3101e3117. Donoghue, J.P., Herkenham, M., 1986. Neostriatal projections from individual cortical fields conform to histochemically distinct striatal compartments in the rat. Brain Res. 365, 397e403. Economo, M.N., Clack, N.G., Lavis, L.D., Gerfen, C.R., Svoboda, K., Myers, E.W., Chandrashekar, J., 2016. A platform for brain-wide imaging and reconstruction of individual neurons. Elife 5, e10566. Economo, M.N., Viswanathan, S., Taxic, B., Baas, E., Winnubst, J., Menon, V., Raybuck, L.T., Nguyen, T.N., Wang, L., Gerfen, C.R., Chandrashekar, J., Zeng, H., Looger, L.L., Svoboda, K., 2018. Distinct descending motor cortex pathways and their roles in movement. Nature 563, 79e84. English, D.F., Ibanez-Sandoval, O., Stark, E., Tecuapetla, F., Buzsaki, G., Deisseroth, K., Tepper, J.M., Koos, T., 2011. GABAergic circuits mediate the reinforcement-related signals of striatal cholinergic interneurons. Nat. Neurosci. 15, 123e130. Faust, T.W., Assous, M., Tepper, J.M., Koos, T., 2016. Neostriatal GABAergic interneurons mediate cholinergic inhibition of spiny projection neurons. J. Neurosci. 36, 9505e9511. Fishell, G., van der Kooy, D., 1987. Pattern formation in the striatum: developmental changes in the distribution of striatonigral neurons. J. Neurosci. 7, 1969e1978. Flandin, P., Kimura, S., Rubenstein, J.L., 2010. The progenitor zone of the ventral medial ganglionic eminence requires Nkx2-1 to generate most of the globus pallidus but few neocortical interneurons. J. Neurosci. 30 (8), 2812e2823. Francois, C., Grabli, D., McCairn, K., Jan, C., Karachi, C., Hirsch, E.-C., Feger, J., Tremblay, L., 2004. Behavioural disorders induced by external globus pallidus dysfunction in primates II. Anatomical study. Brain 127, 2055e2070.

Basal ganglia circuits Chapter | 10

239

Friedman, A., Homma, D., Bloem, B., Gibb, L.G., Amemori, K.-I., Hu, D., Delcasso, S., Truong, T.F., Yang, J., Hood, A.S., et al., 2017. Chronic stress alters striosome-circuit dynamics, leading to aberrant decision-making. Cell 171, 1191e1205.e28. Fujiyama, F., Nakano, T., Matsuda, W., Furuta, T., Udagawa, J., Kaneko, T., 2016. A single-neuron tracing study of arkypallidal and prototypic neurons in healthy rats. Brain Struct. Funct. 221, 4733e4740. Gage, G.J., Stoetzner, C.R., Wiltschko, A.B., Berke, J.D., 2010. Selective activation of striatal fast-spiking interneurons during choice execution. Neuron 67, 466e479. Gerfen, C.R., 1984. The neostriatal mosaic: compartmentalization of corticostriatal input and striatonigral output systems. Nature 311, 461e464. Gerfen, C.R., 1985. The neostriatal mosaic: I. Compartmental organization of projections from the striatum to the substantia nigra in the rat. J. Comp. Neurol. 236, 454_476. Gerfen, C.R., 1989. The neostriatal mosaic: striatal patch-matrix organization is related to cortical lamination. Science 246, 385e388. Gerfen, C.R., 1992. The neostriatal mosaic: multiple levels of compartmental organization. Trends Neurosci. 15, 133e139. Gerfen, C.R., Surmeier, D.J., 2011. Modulation of striatal projection systems by dopamine. Annu. Rev. Neurosci. 34, 441e466. Gerfen, C.R., Young III, W.S., 1988. Distribution of striatonigral and striatopallidal peptidergic neurons in both patch and matrix compartments: an in situ hybridization and fluorescent retrograde tracing study. Brain Res. 460, 161e167. Gerfen, C.R., Herkenham, M., Thibault, J., 1987. The neostriatal mosaic: II. Patch and matrix directed mesostriatal dopaminergic and non-dopaminergic systems. J. Neurosci. 7, 3915e3934. Gerfen, C.R., Baimbridge, K.G., Thibault, J., 1987. The neostriatal mosaic: III. biochemical and developmental dissociation of dual nigrostriatal dopaminergic systems. J. Neurosci. 7, 3935e3944. Gerfen, C.R., Engber, T.M., Mahan, L.C., Susel, Z., Chase, T.N., Monsma, F.J., Sibley, D.R., 1990. D1 and D2 dopamine receptor regulated gene expression of striatonigral and striatopallidal neurons. Science 250, 1429e1432. Gerfen, C.R., Miyachi, S., Paletzki, R., Brown, P., 2002. D1 Dopamine receptor supersensitivity in the dopamine depleted striatum results from a switch in the regulation of ERK1/2/MAP kinase. J. Neurosci. 22, 5042e5054. Gerfen, C.R., Paletzki, R., Heintz, N., 2013. GENSAT BAC Cre-recombinase driver lines to study the functional organization of cerebral cortical and basal ganglia circuits. Neuron 80, 1368e1383. Gilmore, E.C., Herrup, K., 1997. Cortical development: layers of complexity. Curr. Biol. 7 (4), R231. Gittis, A., 2018. Probing new targets for movement disorders. Science 361 (6401), 462. Gittis, A.H., Nelson, A.B., Thwin, M.T., Palop, J.J., Kreitzer, A.C., 2010. Distinct roles of GABAergic interneurons in the regulation of striatal output pathways. J. Neurosci. 30. Gittis, A.H., Leventhal, D.K., Fensterheim, B.A., Pettibone, J.R., Berke, J.D., Kreitzer, A.C., 2011. Selective inhibition of striatal fast-spiking interneurons causes Dyskinesias. J. Neurosci. 31. Gittis, A.H., Hang, G.B., LaDow, E.S., Shoenfeld, L.R., Atallah, B.V., Finkbeiner, S., Kreitzer, A.C., 2011. Rapid target-specific remodeling of fastspiking inhibitory circuits after loss of dopamine. Neuron 71, 858e868. Gittis, A.H., Berke, J.D., Bevan, M.D., Chan, C.S., Mallet, N., Morrow, M.M., Schmidt, R., 2014. New roles for the external globus pallidus in basal ganglia circuits and behavior. J. Neurosci. 34, 15178e15183. Glajch, K.E., Kelver, D.A., Hegeman, D.J., Cui, Q., Xenias, H.S., Augustine, E.C., Hernandez, V.M., Verma, N., Huang, T.Y., Luo, M., et al., 2016. Npas1þ pallidal neurons target striatal projection neurons. J. Neurosci. 36, 5472e5488. Grabli, D., McCairn, K., Hirsch, E.C., Agid, Y., Feger, J., Francois, C., Tremblay, L., 2004. Behavioural disorders induced by external globus pallidus dysfunction in primates: I. Behavioural study. Brain 127, 2039e2054. Graybiel, A.M., Ragsdale Jr., C.W., 1978. Histochemically distinct compartments in the striatum of human, monkeys, and cat demonstrated by acetylthiocholinesterase staining. Proc. Natl. Acad. Sci. U.S.A. 75 (11), 5723e5726. Graybiel, A.M., Aosaki, T., Flaherty, A.W., Kimura, M., 1994. The basal ganglia and adaptive motor control. Science 265, 1826e1831. Guo, Z.V., Inagaki, H.K., Daie, K., Druckmann, S., Gerfen, C.R., Svoboda, K., 2017. Maintenance of persistent activity in a frontal thalamocortical loop. Nature 545, 181e186. Haber, S.N., Watson, S.J., 1983. The comparison between enkephalin-like and dynorphin-like immunoreactivity in both monkey and human globus pallidus and substantia nigra. Life Sci. 33, 33e36. Hegeman, D.J., Hong, E.S., Hernandez, V.M., Chan, C.S., 2016. The external globus pallidus: progress and perspectives. Eur. J. Neurosci. 43, 1239e1265. Herkenham, M., Pert, C.B., 1981. Mosaic distribution of opiate receptors, parafascicular projections and acetylcholinesterase in rat striatum. Nature 291, 415e418. Hernandez, V.M., Hegeman, D.J., Cui, Q., Kelver, D.A., Fiske, M.P., Glajch, K.E., Pitt, J.E., Huang, T.Y., Justice, N.J., Chan, C.S., 2015. Parvalbuminþ neurons and Npas1þ neurons are distinct neuron classes in the mouse external globus pallidus. J. Neurosci. 35, 11830e11847. Higley, M.J., Gittis, A.H., Oldenburg, I.A., Balthasar, N., Seal, R.P., Edwards, R.H., Lowell, B.B., Kreitzer, A.C., Sabatini, B.L., 2011. Cholinergic interneurons mediate fast VGluT3-dependent glutamatergic transmission in the striatum. PLoS One 6. Hikosaka, O., Wurtz, R.H., 1983. Visual and oculomotor functions of monkey substantia nigra pars reticulata. IV. Relation of substantia nigra to superior colliculus. J. Neurophysiol. 49, 1285e1301. Hikosaka, O., Takikawa, Y., Kawagoe, R., 2000. Role of the basal ganglia in the control of purposive saccadic eye movements. Physiol. Rev. 80, 953e978.

240 PART | I Circuit development

Hintiryan, H., Foster, N.N., Bowman, I., Bay, M., Song, M.Y., Gou, L., Yamashita, S., Bienkowski, M.S., Zingg, B., Zhu, M., Yang, X.W., Shih, J.C., Toga, A.W., Dong, H.W., 2016. The mouse cortico-striatal projectome. Nat. Neurosci. 19 (8), 1100e1114. Hooks, M., Papale, A.E., Paletzki, M., Eastwood, B.S., Couey, J.J., Winnubst, J., Chandrashekar, J., Gerfen, C.R., 2018. Topographic precision in sensory and motor corticostriatal projections varies across cell type and cortical area. Nat. Commun. 9, 3549. Horak, F.B., Anderson, M.E., 1984. Influence of globus pallidus on arm movements in monkeys. II. Effects of stimulation. J. Neurophysiol. 52, 305e322. Kalanithi, P.S.A., Zheng, W., Kataoka, Y., DiFiglia, M., Grantz, H., Saper, C.B., Schwartz, M.L., Leckman, J.F., Vaccarino, F.M., 2005. Altered parvalbumin-positive neuron distribution in basal ganglia of individuals with Tourette syndrome. Proc. Natl. Acad. Sci. U.S.A. 102, 13307e13312. Kato, M., Kimura, M., 1992. Effects of reversible blockade of basal ganglia on a voluntary arm movement. J. Neurophysiol. 68, 1516e1534. Kawaguchi, Y., 1993. Physiological, morphological, and histochemical characterization of three classes of interneurons in rat neostriatum. J. Neurosci. 13, 4908e4923. Kawaguchi, Y., Wilson, C.J., Emson, P.C., 1990. Projection subtypes of rat neostriatal matrix cells revealed by intracellular injection of biocytin. J. Neurosci. 10, 3421e3438. Kawaguchi, Y., 1992. Large aspiny cells in the matrix of the rat neostriatum in vitro: Physiological identification, relation to the compartments and excitatory postsynaptic currents. J. Neurophysiol. 67, 1669e1682. Kawaguchi, Y., Wilson, C.J., Augood, S.J., Emson, P.C., 1995. Striatal interneurones: chemical, physiological and morphological characterization. Trends Neurosci. 18, 527e535. Kemp, J.M., Powell, T.P.S., 1970. The cortico-striate projection in the monkey. Brain 93, 525e546. Kemp, J.M., Powell, T.P.S., 1971. The structure of the caudate nucleus of the cat: light and electron microscopy. Philos. Trans. R Soc. Lond. B Biol. Sci. 262, 383e401. Kita, H., 1993. GABAergic circuits of the striatum. Prog. Brain Res. 99, 51e72. Kita, T., Kita, H., 2012. The subthalamic nucleus is one of multiple innervation sites for long-range corticofugal axons: a single-axon tracing study in the rat. J. Neurosci. 32, 5990e5999. Kita, H., Kitai, S.T., 1987. Efferent projections of the subthalamic nucleus in the rat: light and electron microscopic analysis with the PHA-L method. J. Comp. Neurol. 260, 435e452. Kita, H., Chang, H.T., Kitai, S.T., 1983. The morphology of intracellularly labeled rat subthalamic neurons: a light microscopic analysis. J. Comp. Neurol. 215, 245e257. Kita, H., Chang, H.T., Kitai, S.T., 1983. Pallidal inputs to subthalamus: intracellular analysis. Brain Res. 264, 255e265. Klug, J.R., Engelhardt, M.D., Cadman, C.N., Li, H., Smith, J.B., Ayala, S., Williams, E.W., Hoffman, H., Jin, X., 2018. Differential inputs to striatal cholinergic and parvalbumin interneurons imply functional distinctions. Elife 7. Koos, T., Tepper, J.M., 1999. Inhibitory control of neostriatal projection neurons by GABAergic interneurons. Nat. Neurosci. 2, 467e472. Koos, T., Tepper, J.M., 2002. Dual cholinergic control of fast-spiking interneurons in the neostriatum. J. Neurosci. 22, 529e535. Kravitz, A.V., Freeze, B.S., Parker, P.R.L., Kay, K., Thwin, M.T., Deisseroth, K., Kreitzer, A.C., 2010. Regulation of parkinsonian motor behaviours by optogenetic control of basal ganglia circuitry. Nature 466, 622e626. Kress, G.J., Yamawaki, N., Wokosin, D.L., Wickersham, I.R., Shepherd, G.M., Surmeier, D.J., 2013. Convergent cortical innervation of striatal projection neurons. Nat. Neurosci. 16 (6), 665e667. Kubota, Y., Mikawa, S., Kawaguchi, Y., 1993. Neostriatal GABAergic interneurones contain NOS, calretinin or parvalbumin. Neuroreport 5, 205e208. Lapper, S.R., Bolam, J.P., 1992. Input from the frontal cortex and the parafascicular nucleus to cholinergic interneurons in the dorsal striatum of the rat. Neuroscience 51, 533e545. Lee, K., Holley, S.M., Shobe, J.L., Chong, N.C., Cepeda, C., Levine, M.S., Masmanidis, S.C., 2017. Parvalbumin interneurons modulate striatal output and enhance performance during associative learning. Neuron 93. Lee, K., Holley, S.M., Shobe, J.L., Chong, N.C., Cepeda, C., Levine, M.S., Masmanidis, S.C., 2018. Parvalbumin interneurons modulate striatal output and enhance performance during associative learning. Neuron 99, 239. Lei, W., Jiao, Y., Del Mar, N., Reiner, A., 2004. Evidence for differential cortical input to direct pathway versus indirect pathway striatal projection neurons in rats. J. Neurosci. 24 (38), 8289e8299. Li, N., Chen, T.W., Guo, Z., Gerfen, C.R., Svoboda, K., 2015. Flow of information in motor cortex circuit driving voluntary movement. Nature 519, 51e56. Long, J.E., Swan, C., Liang, W.S., Cobos, I., Potter, G.B., Rubenstein, J.L., 2009. Dlx1&2 and Mash1 transcription factors control striatal patterning and differentiation through parallel and overlapping pathways. J. Comp. Neurol. 512, 556e572. Lv, X., Dickerson, J.W., Rook, J.M., Lindsley, C.W., Conn, P.J., Xiang, Z., 2017. M1 muscarinic activation induces long-lasting increase in intrinsic excitability of striatal projection neurons. Neuropharmacology 118, 209e222. Mallet, N., Micklem, B.R., Henny, P., Brown, M.T., Williams, C., Bolam, J.P., Nakamura, K.C., Magill, P.J., 2012. Dichotomous organization of the external globus pallidus. Neuron 74, 1075e1086. Mallet, N., Schmidt, R., Leventhal, D., Chen, F., Amer, N., Boraud, T., Berke, J.D., 2016. Arkypallidal cells send a Stop signal to striatum. Neuron 89, 308e316. Markowitz, J.E., Gillis, W.F., Beron, C.C., Neufeld, S.Q., Robertson, K., Bhagat, N.D., Peterson, R.E., Peterson, E., Hyun, M., Linderman, S.W., Sabatini, B.L., Datta, S.R., 2018. The striatum organizes 3D behavior via moment-to-moment action selection. Cell 174 (1), 44e58. Mastro, K.J., Zitelli, K.T., Willard, A.M., Leblanc, K.H., Kravitz, A.V., Gittis, A.H., 2017. Cell-specific pallidal intervention induces long-lasting motor recovery in dopamine depleted mice. Nat. Neurosci. 20, 815e823. Mastro, K.J.J., Bouchard, R.S.S., Holt, H.A.K.A., Gittis, A.H.H., 2014. Transgenic mouse lines subdivide external segment of the globus pallidus (GPe) neurons and reveal distinct GPe output pathways. J. Neurosci. 34, 2087e2099.

Basal ganglia circuits Chapter | 10

241

Matsumoto, N., Minamimoto, T., Graybiel, A.M., Kimura, M., 2001. Neurons in the thalamic CM-Pf complex supply striatal neurons with information about behaviorally significant sensory events. J. Neurophysiol. 85, 960e976. Matsumura, M., Tremblay, L., Richard, H., Filion, M., 1995. Activity of pallidal neurons in the monkey during dyskinesia induced by injection of bicuculline in the external pallidum. Neuroscience 65, 59e70. Meredith, G.E., Kang, U.J., 2006. Behavioral models of Parkinson’s disease in rodents: a new look at an old problem. Mov. Disord. 21, 1595e1606. Mink, J.W., 1996. The basal ganglia: focused selection and inhibition of competing motor programs. Prog. Neurobiol. 50, 381e425. Mink, J.W., 2003. The basal Ganglia and involuntary movements: impaired inhibition of competing motor patterns. Arch. Neurol. 60, 1365e1368. Moon Edley, S., Herkenham, M., 1984. Comparative development of striatal opiate receptors and dopamine revealed by autoradiography and histofluorescence. Brain Res. 305, 27e42. Morita, K., Morishima, M., Sakai, K., Kawaguchi, Y., 2012. Reinforcement Learning: Computing the Temporal Difference of Values via Distinct Corticostriatal Pathways. Trends Neurosci. 35, 457e467. Munoz-Manchado, A.B., Bengtsson Gonzales, C., Zeisel, A., Munguba, H., Bekkouche, B., Skene, N.G., Lonnerberg, P., Ryge, J., Harris, K.D., Linnarsson, S., et al., 2018. Diversity of interneurons in the dorsal striatum revealed by single-cell RNA sequencing and PatchSeq. Cell Rep. 24, 2179e2190.e7. Nambu, A., Tokuno, H., Hamada, I., Kita, H., Imanishi, M., Akazawa, T., Ikeuchi, Y., Hasegawa, N., 2000. Excitatory cortical inputs to pallidal neurons via the subthalamic nucleus in the monkey. J. Neurophysiol. 84, 289e300. Nelson, A.B., Bussert, T.G., Kreitzer, A.C., Seal, R.P., 2014. Striatal cholinergic neurotransmission requires VGLUT3. J. Neurosci. 34, 8772e8777. Nelson, A.B., Hammack, N., Yang, C.F., Shah, N.M., Seal, R.P., Kreitzer, A.C., 2014. Striatal cholinergic interneurons Drive GABA release from dopamine terminals. Neuron 82, 63e70. Nobrega-Pereira, S., Gelman, D., Bartolini, G., Pla, R., Pierani, A., Marin, O., 2010. Origin and molecular specification of globus pallidus neurons. J. Neurosci. 30, 2824e2834. Obeso, J.A., Rodriguez-Oroz, M.C., Rodriguez, M., Lanciego, J.L., Artieda, J., Gonzalo, N., Olanow, C.W., 2000. Pathophysiology of the basal ganglia in Parkinson’s disease. Trends Neurosci. 23, S8eS19. Oh, S.W., Harris, J.A., Ng, L., Winslow, B., Cain, N., Mihalas, S., Wang, Q., Lau, C., Kuan, L., Henry, A.M., Mortrud, M.T., Ouellette, B., Nguyen, T.N., Sorensen, S.A., Slaughterbeck, C.R., Wakeman, W., Li, Y., Feng, D., Ho, A., Nicholas, E., Hirokawa, K.E., Bohn, P., Joines, K.M., Peng, H., Hawrylycz, M.J., Phillips, J.W., Hohmann, J.G., Wohnoutka, P., Gerfen, C.R., Koch, C., Bernard, A., Dang, C., Jones, A.R., Zeng, H., 2014. A mesoscale connectome of the mouse brain. Nature 508, 207e214. Olsson, M., Campbell, K., Turnbull, D.H., 1997. Specification of mouse telencephalic and mid-hindbrain progenitors following heterotopic ultrasoundguided embryonic transplantation. Neuron 19, 761e772. Owen, S.F., Berke, J.D., Kreitzer, A.C., 2018. Fast-spiking interneurons supply feedforward control of bursting, calcium, and plasticity for efficient learning. Cell 172, 683e695.e15. O’Hare, J.K., Li, H., Kim, N., Gaidis, E., Ade, K., Beck, J., Yin, H., Calakos, N., 2017. Striatal fast-spiking interneurons selectively modulate circuit output and are required for habitual behavior. Elife 6. Parthasarathy, H.B., Graybiel, A.M., 1997. Cortically driven immediate-early gene expression reflects modular influence of sensorimotor cortex on identified striatal neurons in the squirrel monkey. J. Neurosci. 17, 2477e2491. Planert, H., Szydlowski, S.N., Hjorth, J.J., Grillner, S., Silberberg, G., 2010. Dynamics of synaptic transmission between fast-spiking interneurons and striatal projection neurons of the direct and indirect pathways. J. Neurosci. 30, 3499e3507. Rapanelli, M., Frick, L.R., Xu, M., Groman, S.M., Jindachomthong, K., Tamamaki, N., Tanahira, C., Taylor, J.R., Pittenger, C., 2017. Targeted interneuron depletion in the dorsal striatum produces autism-like behavioral abnormalities in male but not female mice. Biol. Psychiatry 82, 194e203. Redgrave, P., Prescott, T.J., Gurney, K., 1999. The basal ganglia: a vertebrate solution to the selection problem? Neuroscience 89, 1009e1023. Redgrave, P., Rodriguez, M., Smith, Y., Rodriguez-Oroz, M.C., Lehericy, S., Bergman, H., Agid, Y., DeLong, M.R., Obeso, J.A., 2010. Goal-directed and habitual control in the basal ganglia: implications for Parkinson’s disease. Nat. Rev. Neurosci. 11 (11), 760e772. Rojas-Piloni, G., Guest, J.M., Egger, R., Johnson, A.S., Sakmann, B., Oberlaender, M., 2017. Relationships between structure, in vivo function and longrange axonal target of cortical pyramidal tract neurons. Nat. Commun. 8 (1), 870. Saunders, A., Oldenburg, I.A., Berezovskii, V.K., Johnson, C.A., Kingery, N.D., Elliott, H.L., Xie, T., Gerfen, C.R., Sabatini, B.L., 2015. A direct GABAergic output from the basal ganglia to frontal cortex. Nature 521, 85e89. Saunders, A., Macosko, E.Z., Wysoker, A., Goldman, M., Krienen, F.M., de Rivera, H., Bien, E., Baum, M., Bortolin, L., Wang, S., Goeva, A., Nemesh, J., Kamitaki, N., Brumbaugh, S., Kulp, D., McCarroll, S.A., 2018. Molecular diversity and specializations among the cells of the adult mouse brain. Cell 174 (4), 1015e1030.e16. Sawamura, A., Hashizume, K., Tanaka, T., 2002. Electrophysiological, behavioral and metabolical features of globus pallidus seizures induced by a microinjection of kainic acid in rats. Brain Res. 935, 1e8. Schmidt, R., Leventhal, D.K., Mallet, N., Chen, F., Berke, J.D., 2013. Canceling actions involves a race between basal ganglia pathways. Nat. Neurosci. 16, 1118e1124. Schulz, J.M., Reynolds, J.N.J., 2013. Pause and rebound: sensory control of cholinergic signaling in the striatum. Trends Neurosci. 36, 41e50. Shen, W., Hamilton, S.E., Nathanson, N.M., Surmeier, D.J., 2005. Cholinergic suppression of KCNQ channel currents enhances excitability of striatal medium spiny neurons. J. Neurosci. 25, 7449e7458. Shen, W., Tian, X., DayM, Ulrich, S., Tkatch, T., et al., 2007. Cholinergic modulation of Kir2 channels selectively elevates dendritic excitability in striatopallidal neurons. Nat. Neurosci. 10, 1458e1466.

242 PART | I Circuit development

Shepherd, G.M., 2013. Corticostriatal connectivity and its role in disease. Nat. Rev. Neurosci. 14 (4), 278e291. Smith, Y., Bevan, M.D., Shink, E., Bolam, J.P., 1998. Microcircuitry of the direct and indirect pathways of the basal ganglia. Neuroscience 86, 353e387. Smith, J.B., Klug, J.R., Ross, D.L., Howard, C.D., Hollon, N.G., Ko, V.I., Hoffman, H., Callaway, E.M., Gerfen, C.R., Jin, X., 2016. Dissecting the monosynaptic inputs and outputs unveils the functional relevance of striatal patch and matrix compartments. Neuron 91, 1069e1084. Soares, J., Kliem, M.A., Betarbet, R., Greenamyre, J.T., Yamamoto, B., Wichmann, T., 2004. Role of external pallidal segment in primate parkinsonism: comparison of the effects of 1-methyl-4-phenyl-1,2,3,6-tetrahydropyridine-induced parkinsonism and lesions of the external pallidal segment. J. Neurosci. 24, 6417e6426. Stern, E.A., Jaeger, D., Wilson, C.J., 1998. Membrane potential synchrony of simultaneously recorded striatal spiny neurons in vivo. Nature 394, 475e478. Straub, C., Saulnier, J.L., Begue, A., Feng, D.D., Huang, K.W., Sabatini, B.L., 2016. Principles of synaptic organization of GABAergic interneurons in the striatum. Neuron 92, 84e92. Surmeier, D.J., Eberwine, J., Wilson, C.J., Cao, Y., Stefani, A., Kitai, S.T., 1992. Dopamine receptor subtypes colocalize in rat striatonigral neurons. Proc. Natl. Acad. Sci. U.S.A. 89, 10178e10182. Surmeier, D.J., Ding, J., Day, M., Wang, Z., ShenW, 2007. D1 and D2 dopamine-receptor modulation of striatal glutamatergic signaling in striatal medium spiny neurons. Trends Neurosci. 30, 228e235. Tepper, J.M., Koos, T., 2017. Striatal GABAergic interneurons. In: Steiner, H., Tseng, K. (Eds.), Handbook of Basal Ganglia Structure and Function, Handbook of Behavioral Neuroscience, second ed., vol. 24. Academic Press, Cambridge, MA, pp. 157e178. Tepper, J.M., Tecuapetla, F., Koós, T., Ibáñez-Sandoval, O., 2010. Heterogeneity and diversity of striatal GABAergic interneurons. Front. Neuroanat. 4, 150. Toresson, H., Potter, S.S., Campbell, K., 2000. Genetic control of dorsal-ventral identity in the telencephalon: opposing roles for Pax6 and Gsh2. Development 127, 4361e4371. Turner, R.S., Anderson, M.E., 1997. Pallidal discharge related to the kinematics of reaching movements in two dimensions. J. Neurophysiol. 77, 1051e1074. Van Den Berge, N., Albaugh, D.L., Salzwedel, A., Vanhove, C., Van Holen, R., Gao, W., Stuber, G.D., Ian Shih, Y.-Y., 2017. Functional circuit mapping of striatal output nuclei using simultaneous deep brain stimulation and fMRI. Neuroimage 146, 1050e1061. van der Kooy, D., Fishell, G., 1987. Neuronal birthdate underlies the development of striatal compartments. Brain Res. 401, 155e161. Wall, N.R., De La Parra, M., Callaway, E.M., Kreitzer, A.C., 2013. Differential innervation of direct- and indirect-pathway striatal projection neurons. Neuron 79 (2), 347e360. Wallace, M.L., Saunders, A., Huang, K.W., Philson, A.C., Goldman, M., Macosko, E.Z., McCarroll, S.A., Sabatini, B.L., 2017. Genetically distinct parallel pathways in the entopeduncular nucleus for limbic and sensorimotor output of the basal ganglia. Neuron 94, 138e152.e5. Watabe-Uchida, M., Zhu, L., Ogawa, S.K., Vamanrao, A., Uchida, N., 2012. Whole-brain mapping of direct inputs to midbrain dopamine neurons. Neuron 74, 858e873. Webster, K.E., 1961. Cortico-striate interrelations in the albino rat. J. Anat. 95, 532e545. Wichmann, T., DeLong, M.R., 2003. Functional neuroanatomy of the basal ganglia in Parkinson’s disease. Adv. Neurol. 91, 9e18. Wilson, C.J., 1987. Morphology and synaptic connections of crossed corticostriatal neurons in the rat. J. Comp. Neurol. 263, 567e580. Wilson, C.J., 1993. The generation of natural firing patterns in neostriatal neurons. Prog. Brain Res. 99, 277e297. Wilson, C.J., Groves, P.M., 1980. Fine structure and synaptic connections of the common spiny neuron of the rat neostriatum: a study employing intracellular injection of horseradish peroxidase. J. Comp. Neurol. 194, 599e615. Wilson, C.J., Kawaguchi, Y., 1996. The origins of two-state spontaneous membrane potential fluctuations of neostriatal spiny neurons. J. Neurosci. 16, 2397e2410. Wilson, C.J., Chang, H.T., Kitai, S.T., 1990. Firing patterns and synaptic potentials of identified giant aspiny interneurons in the rat neostriatum. J. Neurosci. 10, 508e519. Wise, S.P., Jones, E.G., 1977. Cells of origin and terminal distribution of descending projections of the rat somatic sensory cortex. J. Comp. Neurol. 175, 129e158. Xiao, C., Miwa, J.M., Henderson, B.J., Wang, Y., Deshpande, P., McKinney, S.L., Lester, H.A., 2015. Nicotinic receptor subtype-selective circuit patterns in the subthalamic nucleus. J. Neurosci. 35, 3734e3746. Xu, M., Li, L., Pittenger, C., 2016. Ablation of fast-spiking interneurons in the dorsal striatum, recapitulating abnormalities seen post-mortem in Tourette syndrome, produces anxiety and elevated grooming. Neuroscience 324, 321e329. Yin, H.H., Knowlton, B.J., 2006. The role of the basal ganglia in habit formation. Nat. Rev. Neurosci. 7, 464e476. Yin, H.H., Mulcare, S.P., Hilário, M.R., Clouse, E., Holloway, T., Davis, M.I., Hansson, A.C., Lovinger, D.M., Costa, R.M., 2009. Dynamic reorganization of striatal circuits during the acquisition and consolidation of a skill. Nat. Neurosci. 12 (3), 333e341. Young 3rd, W.S., Bonner, T.I., Brann, M.R., 1986. Mesencephalic dopamine neurons regulate the expression of neuropeptide mRNAs in the rat forebrain. Proc. Natl. Acad. Sci. U.S.A. 83, 9827e9831. Yu, J., Yan, Y., Li, K.-L., Wang, Y., Huang, Y.H., Urban, N.N., Nestler, E.J., Schluter, O.M., Dong, Y., 2017. Nucleus accumbens feedforward inhibition circuit promotes cocaine self-administration. Proc. Natl. Acad. Sci. U.S.A. 114, E8750eE8759. Zheng, T., Wilson, C.J., 2002. Corticostriatal combinatorics: the implications of corticostriatal axonal arborizations. J. Neurophysiol. 87, 1007e1017. Zucca, S., Zucca, A., Nakano, T., Aoki, S., Wickens, J., 2018. Pauses in cholinergic interneuron firing exert an inhibitory control on striatal output in vivo. Elife 7.

Chapter 11

Development of the neuronal circuitry of the cerebellar cortex Constantino Sotelo1, 2, *, Fabrice Ango3 and Richard Hawkes4 1

Sorbonne Universités, UPMC Université Paris 06, INSERM, CNRS, Institut de la Vision Paris, France; 2Instituto de Neurociencias de Alicante,

UMH-CSIC, Universidad Miguel Hernández de Elche, Alicante, Spain; 3INM, University of Montpellier, CNRS, INSERM, Montpellier, France; Department of Cell Biology & Anatomy and Hotchkiss Brain Institute, Cumming School of Medicine, University of Calgary, Calgary, Alberta, Canada 4

Chapter outline 11.1. Introduction 11.2. Organization of the adult circuitry of the cerebellar cortex 11.2.1. Cajal and the cerebellar circuit 11.3. The modular organization of the cerebellar cortex 11.4. Development of the heterogeneity of Purkinje cells 11.5. Development and refinement of climbing fiber projections 11.6. Development and refinement of mossy fiber projections 11.7. Cerebellar interneurons in the cerebellar circuit

243 243 243 245 245 247 250 253

11.7.1. Excitatory interneurons 11.7.1.1. Granule cells 11.7.1.2. Unipolar brush cells (UBCs) 11.7.2. Inhibitory interneurons 11.7.2.1. Granule cell layer interneurons 11.7.2.2. Purkinje cell layer interneurons 11.7.2.3. Molecular layer interneurons Acknowledgment References

253 253 255 255 255 256 257 259 259

11.1 Introduction The cerebellum, despite its name of “small brain,” houses almost half of all neurons in the central nervous system (Azevedo et al., 2009). What is the role of such a densely populated nervous center? By its hodological organization the cerebellum is at the crossroads of the motor pathway. First, through its cortico-ponto-cerebellar projections it receives a copy of the motor commands descending from the motor cortex to the spinal cord, while simultaneously it receives feedback/ascending sensory information from the spinal cord and brainstem. As a result the cerebellum is in a privileged position to coordinate planned motor functions.

11.2 Organization of the adult circuitry of the cerebellar cortex 11.2.1 Cajal and the cerebellar circuit Cajal (1911) described that the three-layered cortex comprised a single class of projection neuron, the Purkinje cell (PC), the sole output of the cerebellar cortex. Adult PCs are easily recognized by the large diameter of their cell bodies (about 27 mm in small rodents) arranged in a single row at the interface between the molecular and granular layers and their large dendritic trees, flattened in the parasagittal plane, that occupy almost the complete thickness of the molecular layer. The dendritic tree is highly branched, consisting of a proximal compartment with large and medium-sized branches, and a distal

* Senior author.

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00011-1 Copyright © 2020 Elsevier Inc. All rights reserved.

243

244 PART | I Circuit development

compartment of spiny branchlets. The proximal compartment is contacted by climbing fiber afferents and inhibitory inputs from interneurons, while the distal is primarily the postsynaptic target of the parallel fibers. The PC axons emerge from the basal poles of the somata and extend through the granular layer to the white matter. With the exception of rather extended extracerebellar projections in the brainstem (De Camilli et al., 1984), the great majority of PC axons terminate within the deep cerebellar and lateral vestibular nuclei. While transiting through the granular layer PC axons send recurrent collaterals that contribute to the supraganglionic and subganglionic plexuses, and contact other PCs and molecular layer interneurons (Guo et al., 2016). Regarding the GABAergic interneurons, which are layer-specific, the stellate and basket cells of the molecular layer establish contacts with one another, and with PC somata and dendrites, and also synapse with the dendrites of other molecular layer interneurons, such as dendritic trees of Lugaro and Golgi cells. In the granular layer, Golgi cells contact mainly granule cells. Underlying the granular layer is the white matter, the site of passage for the extracerebellar afferent fibers and the efferent PC axons (Fig. 11.1). Cajal (1911) also described the two main afferent systemsdclimbing and mossy fibers (CF and MF)dand their main cellular targets. CFs have PCs as their main targets. CFs originate solely from the inferior olivary complex, and thus constitute the most convergent afferent system of the whole brain, since each fiber forms synapses with the proximal dendritic segment of only five to seven PCs (in rodents: see Shinoda et al., 2000). Moreover, CFs extend collaterals that synapse on stellate, basket, and Golgi cells (referred to as “Scheibel collaterals”: see references in Palay and Chan-Palay, 1974). In contrast MFs have numerous origins (spinal cord, basilar pontine nuclei, vestibular nerve and nuclei, lateral reticular nucleus, etc.). MF input to the PCs is relayed via the granule cells. This anatomical situation, MF-granule cell-PC, makes MF projections among the most divergent afferent systems in the brain. Each MF terminates in up to 50 large presynaptic varicosities, since each varicosity can contact 20 granule cell dendrites; each glomerulus establishes synaptic contacts with a few hundreds granule cells. In turn the granule cell axons ascend to the molecular layer where they

FIGURE 11.1 Cellular organization of a folium of the cerebellar cortex. Scheme made by Ramón y Cajal in 1892 for a conference at the Medical Sciences Academy of Catalonia, to illustrate the circuit of the cerebellar cortex. All neuronal cells and afferent fibers are drawn in a three-dimensional perspective. In this parasagittal section, dots represent the parallel fibers. (A) molecular layer, (B) granular layer, (C) white matter, (a) Purkinje cell, (b) basket cell, (d) perisomatic baskets and pinceau around the Purkinje cell axons, (e) stellate cell, (f) Golgi cell, (n) climbing fiber, (g) granule cell, (h) mossy fiber, (J) Bergmann glia, (m) astrocyte. Cajal drawing. CSIC Instituto Cajal.

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

245

bifurcate and run the entire width of the folium as parallel fibers that intersect perpendicularly the sagittally flattened dendrites of up to a 100 PCs (see references in Palay and Chan-Palay, 1974). At the same time as the parallel fibers contact PCs they also synapse with all classes of interneuronal cell bodies and dendrites in the molecular layer. Therefore, both MFs and CFs converge on PCs where their inputs are integrated and transmitted via their axons to the neurons of the deep cerebellar nuclei.

11.3 The modular organization of the cerebellar cortex Although the cerebellar cortex may appear to be homogeneous, it is in fact highly compartmentalized, and no feature escapes this underlying architecture. Historically, cerebellar architecture has been constructed backwards, starting from patterns of PC death, proceeding to regional differences in afferent terminal field distributions, and finally in the distributions of PC subsets. Jansen and Brodal (1940) were the first to reveal the columnar organization of the cortical outputs. By using the Marchi technique to map postlesional PC axons, they showed that the corticonuclear projection is organized following “a definite topographical localization” with a clear mediolateral (sagittal) orientation. Later the same was recognized for the cerebellar afferents (e.g., the olivocerebellar projections; see references in Voogd and Ruigrok, 2004). The general conclusion is that all cerebellar mapsdexpression, connectivity, function, pathology, etc.dare congruent (the “one map” hypothesis: Apps and Hawkes, 2009). Cerebellar architecture is best understood starting from the PCs. First, mediolateral boundaries divide the cerebellar cortex into at least five highly conserved transverse zones. In turn, each transverse zone is further subdivided into parasagittal stripes (reviewed in Armstrong and Hawkes, 2014). Zones and stripes are built around PCs of many different phenotypes with a typical cortical module consisting of a few hundred PCs or fewer. The most-studied marker of cerebellar compartmentation is zebrin II (Brochu et al., 1990), an epitope on aldolase C (¼ zebrin II/AldoC: Ahn et al., 1994) but multiple markers reveal further subdivisions within the same PC architecture (e.g., with expression of the small heat-shock protein HSP25dreviewed in Armstrong and Hawkes, 2014). In sum, the mammalian cerebellar cortex is composed of more than 200 parasagittal stripes in five distinct transverse zones. This architecture is highly conserved through evolution: PC subsets are found from fish to humans and a zone-and-stripe architecture is present, albeit adapted to different lifestyles, in all birds and mammals studied (>30 species). Within the transverse zone-and-stripe organization the fine-level architecture is less-well understood and exactly how fine remains a matter of debate (reviewed in Apps et al., 2018). Importantly for this review, cerebellar synapsesdboth afferent and interneuronal terminal fieldsdare restricted at PC stripe boundaries (Fig. 11.2). The development of the functional modular units of the cerebellum results from an interplay between genetic and epigenetic processes. Many of the stages to be followed in the formation of cerebellar connectivitydfate acquisition, migration, axonal elongation, and synaptogenesisdwill not be dealt with here since they are treated in the other chapters. The aim here is to briefly summarize the developmental data related to the circuitry of the cerebellar cortex. We focus on the main players in the specificity of cerebellum circuitrydthe PCs, the main extracerebellar afferents (CFs and MFs), and the cerebellar cortical interneurons. The chapter is divided into three distinct parts: the first briefly reviews the PC heterogeneity, which is the scaffold around which all circuitry is built, next we consider the development of the projection maps of the afferent fibers, and finally we discuss the development of the interneurons.

11.4 Development of the heterogeneity of Purkinje cells The cerebellum arises from the neural tube at the midbrain/hindbrain boundary (reviewed in Leto et al., 2016). At embryonic day 8.5 (E8.5) in the mouse, antagonism between Otx2 and Gbx2 defines the isthmic organizer that in turn defines the cerebellar primordium. A day or so later (E9.5), two germinal neuroepithelia emerge: the rhombic lip (defined by expression of atonal homolog 1 (Atoh1)) located dorsal to the cerebellar plate and the ventricular zone of the fourth ventricle (defined by expression of pancreatic transcription factor 1a (Ptf1a)). Progenitors in the ventricular zone are fated to generate all GABAergic neurons of the cerebellum: the rhombic lip produces all glutamatergic lineages. The PC phenotype is intrinsic and independent of neural activity or afferent innervation (see below). Phenotype specification likely occurs as PCs undergo terminal mitosis (E10eE13 in mice: Miale and Sidman, 1961). Postmitotic PCs migrate dorsally to form the PC plate. It is during this stage that the first evidence of PC subtype specification is seen in Early B-cell factor 2 (Ebf2) expression. Birthdating studies identify two distinct PC populations: an early-born cohort (E10eE11.5) fated to become zebrin II/AldoCþ and a late-born cohort (E11.5eE13) fated to become zebrin II/AldoC(Larouche and Hawkes, 2006). There is a direct correlation between PC birthdate and adult stripe location, so both PC subtype and positional information may be specified at this time. Ebf2 expression is restricted to the late-born PC

246 PART | I Circuit development

FIGURE 11.2 The compartmentation of the mouse cerebellar cortex. (A and B) Whole mount immunoperoxidase staining of the mouse cerebellum with anti-zebrin II/AldoC, a marker of a PC subset. In the vermis, expression reveals transverse zones with up to seven zebrin IIþ/ stripes in each (the anterior zone (AZ): e.g., lobules V and the posterior zone (PZ): e.g., lobule VIII). These alternate with transverse zones in which all PCs are anti-zebrin II immunoreactive (the central zone (CZ): e.g., lobules VIa,b and the nodular zone (NZ): e.g., lobule IX/X). A similar alternation is seen in the hemispheres (e.g., stripes in the ansiform lobules crus I and crus II (l. ans, cI and cII)). The expression pattern is highly reproducible between individuals and strongly conserved across species. (C and D) Transverse serial sections through the posterior cerebellum immunoperoxide-stained for zebrin II/AldoC (C) and HNK1 (D). The patterns are complementary. Panels (A and B) are adapted from Sillitoe, R.V., Hawkes, R., 2002. Whole-mount immunohistochemistry: a high-throughput screen for patterning defects in the mouse cerebellum. J. Histochem. Cytochem. 50, 235e244. Panels (C and D) are adapted from Marzban, H., Sillitoe, R.V., Hoy M., Chung, S.-H., Rafuse, V.F., Hawkes, R., 2004. Abnormal HNK-1 expression in the cerebellum of an N-CAM null mouse. J. Neurocytol. 33, 117e130.

progenitor population fated to become the zebrin II/AldoC-parasagittal stripes (Chung et al., 2008). It seems that Ebf2 acts to repress the zebrin II/AldoCþ phenotype in at least some late-born PCs: consistent with this, Ebf2 null PCs fail to downregulate markers of the zebrin II/AldoCþ phenotype and therefore many zebrin II/AldoC-PCs ectopically express zebrin II/AldoCþ markers. The pathways that lead to the specification of other PC subtypes (e.g., zebrin II/ AldoCþ:HSP25þ/) are unknown (Fig. 11.3). As the postmitotic PCs migrate into the cerebellar primordium the early- and late-born cohorts coalesce so that by E17eE18 a stereotyped array of clusters has emerged (reviewed in Dastjerdi et al., 2012). Whether this migration is the mechanism that specifies cluster architecture or whether the clusters are already specified in the cerebellar plate, is not known. Grafts of dissociated PCs are able to organize into discrete, ectopic zebrin II/AldoCþ/ aggregates (Rouse and Sotelo, 1990), pointing to cell-cell adhesion molecules as possible organizers: cadherins (reviewed in Redies et al., 2011) and integrins (e.g., Graus-Porta et al., 2001) are plausible candidates. It was at the cluster stage that Wassef and Sotelo (1984) first reported a reproducible mosaic of antigen positive and negative PC clusters within the embryonic PC population (Armstrong and Hawkes, 2014). These clusters are the mustering areas in which cerebellar afferents and interneurons become topographically ordered (reviewed in Apps and Hawkes, 2009). Starting at around E18 in mouse cerebellar foliation begins to develop. Because cerebellar folia stretch the cerebellar cortex some 25-fold in length along the rostrocaudal axis the embryonic clusters become progressively extended into the adult stripes. PC cluster dispersal is triggered by Reelin signaling (reviewed in D’Arcangelo, 2014). Reelin secreted by neurons of the external granular layer and cerebellar nuclei binds two PC receptorsdthe apolipoprotein E receptor 2 (Apoer2) and the very low-density lipoprotein receptor (Vldlr: Hiesberger et al., 1999). If Reelin is absent (e.g., the reeler mouse (Relnrl)), PC cluster dispersal is blocked and the adult mouse retains the embryonic cluster (Mariani et al., 1977). The same is the case if both Reelin receptors are deleted (Trommsdorff et al., 1999; if only one receptor is deleted, selective PC ectopias resultdLarouche et al., 2008). Reelin binding causes the Apoer2/Vldlr receptors to cluster (Strasser et al., 2004), which triggers a Fyn and Src protein kinase cascade (Howell et al., 1997) that results in tyrosine phosphorylation of Disabled1 (e.g., Howell et al., 1997). The result seems to be a drop in PC-PC adhesion that permits the embryonic PC clusters to disperse into stripes as lobules form. As the PCs disperse into stripes the CF and MF afferents and interneurons move along with them, thereby generating the adult stripe patterns.

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

247

FIGURE 11.3 Development of PC heterogeneity. (A) Section taken from a rat E19 cerebellum cut in the coronal plane (magnification 34). (B) Camera lucida drawings of serial coronal sections of the cerebellum of an E19 rat embryo immunostained with antiserum against guanosine 3:5phosphate-dependent protein kinase specific for labeling all PCs. The section indicated by the arrowhead corresponds to that of Fig. 11.1A. The shaded areas represent clusters of PC labeling. Note the alternation of labeled clusters and zones devoid of labeling (which also contain PCs as observed in cresyl violet stained preparations). Adapted from Wassef, M., Sotelo, C., 1984. Asynchrony in the expression of guanosine 30 :50 -phosphate-dependent protein kinase by clusters of Purkinje cells during the perinatal development of rat cerebellum. Neuroscience 13, 1217e1241.

11.5 Development and refinement of climbing fiber projections CFs are derived exclusively from the inferior olivary complex. The earliest CF innervation is found at the PC embryonic cluster stage. CF axons from the contralateral inferior olive cross the midline at the base of the brainstem, enter the cerebellar cortex via the inferior cerebellar peduncles (by E14/15 in micedParadies and Eisenman, 1993), and ramify among the PCs (reviewed in Watanabe and Kano, 2011), with CF subsets targeting specific PC clusters, thereby establishing the olivocerebellar topography (see references in Sotelo, 2004). Within the cerebellum, after sending collaterals that terminate in the deep cerebellar nuclei, they enter the cortex and synapse first on a PC transient apical dendrite (the “creeper” stage: Chédotal and Sotelo, 1993), then postnatally they occupy the apical portion of the PC soma and nascent dendrite. Finally, they “climb” up the proximal dendritic tree (hence the name “climbing” fibers). As cerebellar lobules form postnatally, the PC clusters disperse longitudinally into adult stripes and the CF terminals relocate with their targets and thus their terminal field becomes a long stripe in register with a specific PC stripe. Anatomical and physiological mapping has revealed that CF terminal fields from individual subnuclei of the inferior olivary complex form a stereotyped array of stripes throughout the cerebellar cortex. There is a strict and reproducible topographical relationship between subnuclei of the inferior olivary complex and specific PC stripes and zones in the cerebellar cortex. The first demonstration of alignment between striped CF terminal fields and zebrin II/AldoCþ/ PCs was by Gravel et al. (1987). The comparison was crude but did demonstrate the principle that such a reproducible alignment existed. The detailed mapping was taken up by Voogd and his colleagues, culminating in a systematic comparison of inferior olive projections and zebrin II/AldoCþ/ stripes (Voogd and Ruigrok, 2004). The CFs also project collaterals to specific targets in the deep cerebellar and lateral vestibular nuclei such that the cerebellar circuit becomes an array of inferior oliveePC-deep cerebellar nuclear modules (reviewed in Apps et al., 2018) (Fig. 11.4).

248 PART | I Circuit development

(A)

(B)

FIGURE 11.4 Coronal sections of an adult rat cerebellum, illustrating the matching of the olivocerebellar projection and the modular arrangement of PCs. The arrows point to the precise coincidences between the limits of parasagittally disposed bands of zebrin I positive PCs and the border of bands with labeled climbing fibers. The bar represents 115 mm. From Wassef, M., Chedotal, A., Cholley, B., Thomasset, M., Heizmann, C.W., Sotelo, C., 1992. Development of the olivocerebellar projection in the rat: I. Transient biochemical compartmentation of the inferior olive. J. Comp. Neurol. 323, 519e536; Wassef, M., Angaut, P., Arsenio-Nunes, M.L., Bourrat, F., Sotelo, C. Purkinje cell heterogeneity its role in organizing the topography of the cerebellar cortex connections. In: Llinas, R., Sotelo, C. (Eds.), The Cerebellum Revisited. Springer, New York, pp. 5e21.

Synaptic connections between CFs and PCs are present as early as E19 (Mason, 1987; Wassef et al., 1992) but most synaptogenesis occurs postnatally. Cajal (1911) outlined three stages in the postnatal development of CF innervation. First, at the “pericellular nest” stage (P0eP5) that follows the “creeper stage” (see above), CFs from several olivary neurons form synaptic connections with PC somata. At the “capuchon” (hooded) stage (wP8), CFs are first seen “climbing” the PC dendrites and multiple CF terminals form on the apical PC soma and the proximal dendrites. Finally, during the “young climbing fiber” stage (P9eP15) the CF synapses translocate to contact the stubby thorns emerging from branches of the proximal dendritic compartment (reviewed in Sotelo and Chédotal, 2005; Watanabe and Kano, 2011). In the adult there is a remarkable numerical matching such that while each inferior olivary cell innervates multiple PCs, each PC is innervated by only one CF. At birth CFs emerging from several inferior olivary neurons innervate each PC (the transient phase of multiple innervation: Crépel et al., 1976). However, in the adult each PC is innervated by only one CF (Cajal, 1911; Eccles et al., 1966a,b). This one-to-one relationship is achieved by the elimination of supernumerary CFs (e.g., Crépel et al., 1976). The process of synapse elimination during the transition from multiple innervated PCs to single innervated PC resembles the changes taking place during the development of neuromuscular junctions (Sanes and Lichtman, 1999). First a single dominant CF is selected, mediated by PC Ca2þ influx, resulting in one “strong” input and other “weaker” ones (Hashimoto and Kano, 2003). Subsequently, mainly during P7eP14, the strong CF translocates from the soma to the dendritic arbor and the weaker inputs are eliminated (Hashimoto et al., 2009a). The elimination of excess CFs is activity-dependent: for example, altered neuronal activity either in the inferior olive (Andjus et al., 2003) or the PCs (Lorenzetto et al., 2009) both impede elimination and result in persistent multiple CF innervation. It is also proven that granule cell-PC interactions play a role in CF elimination. For example, cerebella deficient in granule cells preserve CF multi-innervation in adulthood (e.g., Mariani et al., 1977). The mechanism(s) of synapse elimination involves glutamate receptors on PC dendritic spines since the deletion of glutamate receptor d2 subunit (GluRd2), a receptor present in tertiary spines, target of parallel fibers, but absent from the stubby thorns contacted by CFs (Landsend et al., 1997), allows the single dominant CF to be selected normally but subsequently the process stalls (Hashimoto et al., 2009b). In addition, the GluRd2/ mouse is deficient in parallel fiber-

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

249

PC dendrite synapses (Kurihara et al., 1997), which led Hashimoto et al. (2009b) to conclude that CF synapse elimination is initially independent of parallel fiber->PC synapse formation but later becomes dependent on it as parallel fibers and CFs compete for postsynaptic territory. There may also be a role for GABAergic inhibition, as multiple innervation persists in mutant mice with diminished GABAergic transmission, an effect that is reversed by enhancing of GABA-A receptor sensitivity by the topical application of diazepam (Nakayama et al., 2012). The topographical relationship between subnuclei of the inferior olivary complex and specific PC stripes and zones in the cerebellar cortex develops because ingrowing CFs recognize their target cells by positional cues intrinsic to the PC cluster architecture (the “matching hypothesis”: e.g., Sotelo and Wassef, 1991; Sotelo and Chédotal, 2005). Consistent with this proposal, in vitro studies demonstrated that experimental rotation of the cerebellar target before CF innervation produced a corresponding rotation of the CF projection, strongly suggesting that the alignment of PC stripes and CF terminal fields is achieved by direct chemospecific interactions (Chédotal et al., 1997). The specific molecular interactions involved are unclear. One candidate is the cell adhesion molecule BEN, which is expressed transiently in the embryonic chick cerebellum in a subset of inferior olive neurons and their corresponding PC target clusters (Chédotal et al., 1996). Another possibility is Eph-Ephrin signaling: EphA tyrosine kinase receptors are expressed in specific domains of the inferior olive, and their CF axons target PC clusters that transiently express ephrinA, and ephrinA2 overexpression disrupts the topography of the olivocerebellar map (Nishida et al., 2002). The matching hypothesis does not eliminate the possibility that during the phase of membrane recognition, at the initiation of synaptogenesis, the presynaptic afferent fibers might influence postsynaptic neuron genes that are not expressed cell autonomously. If the heterogeneity of olivocerebellar fibers could regulate, for instance, the expression of zebrin I in postmitotic PCs, or to change other molecules that could be operative during the membrane recognition process required for the matching hypothesis, the heterogeneity of PCs would be induced and not genetically programmed. Lesion experiments in newborn animals, particularly pedunculotomies, interrupting the arrival of CFs and MFs, have been carried out in newborn animals and no effects have been noted (zebrin IdLeclerc et al., 1988). However, since CF-PC synaptogenesis starts in fetal life, the investigation of the role of prenatal interactions was necessary. Such a study was realized by using transplantations of E12 and E15 rat cerebellar anlagen (before the arrival of extracerebellar fibers), either into the anterior chamber of the eye (in oculo) or into a cavity in the cerebral cortex (in cortico) of adult host rats. The grafts were maintained till an advanced stage of development, when zebrin I is already expressed in PCs that developed normally. Zebrin Iþ clusters of PCs were found alternating with zebrin Idones showing that normal afferent input is not a prerequisite for zebrin I expression (Wassef et al., 1990) (Fig. 11.5). The matching hypothesis can be also applied to CF regeneration, which can occur in both immature and adult animals. For example, unilateral transection of the inferior cerebellar peduncle in newborn rats provokes the complete degeneration of the contralateral olive, with the subsequent loss of CFs in the ipsilateral cortex. Under such circumstances, axons from the remaining inferior olive recross the cerebellar midline and partially innervate the deprived hemicortex. The astonished result is that the organization of the compensatory projections emerging from the sprouting of intact olivary axons also give rise to CFs, which end in the molecular layer and form the typical parasagittal stripes of the olivocerebellar projections (Angaut et al., 1985). A similar study was carried out by Zagrebelsky et al. (1997). These authors not only corroborated the previous results but were able to demonstrate that the newly formed CF stripes are restricted at zebrin II/AldoCþ/ PC boundaries, as in normal development. Similarly, in adult rats, inferior olivary neurons can be killed by the intraperitoneal injection of the neurotoxin 3-acetylpyridine (3AP). This killing is dose-dependent, making possible the survival of a small percentage of olivary neurons. Several months after the 3AP injection, the few remaining inferior olivary neurons were able to sprout at the origin of CF collateral reinnervation. The latter was found to respect the topography of the PC compartments. Other good examples come from homotypic transplants of embryonic explants into the adult cerebellum. In these instances, either adult noninjured CFs or regenerating transected CFs are able to provide either collaterals or regenerative sprouts that enter the transplanted cerebellum and innervate the embryonic PCs. These fibers end in irregular stripes that mimic the olivocerebellar projections in control rats (Armengol et al., 1989). Within the transplants the embryonic PCs pursue an apparently normal progression and can create alternating zebrin II/AldoCþ/ PC clusters (Rouse and Sotelo, 1990). Finally, the reestablishment of topographic organized projections has been obtained with solid pieces of embryonic cerebellum transplanted adjacent to a lesion of the white matter of an adult rat cerebellum. The transected CFs regenerate into the transplants and form discrete patches comprising several terminal arbors impinging upon PC dendrites. The majority (96%) of CF terminal fields exclusively innervate PCs of a single phenotype, and were restricted at the border of the zebrin II/AldoCþ/ PC clusters (Rossi et al., 2002). It does not appear that CF synapse elimination plays a prominent role in sculpting olivocerebellar topography (even if such sculpting is commonplace elsewhere in the nervous system: e.g., Changeux and Danchin, 1976). Rather, perhaps perinatal multi-innervation of PCs by CFs is a form of redundancy that assures that every PC receives a CF synapse.

250 PART | I Circuit development

(A)

(B)

(C)

FIGURE 11.5 Modular reorganization of climbing fibers sprouting and/or regeneration in developing and mature cerebella. (A) Darkfield low power photomicrograph of a cerebellar section of a rat subjected to sectioning of the left cerebellar peduncles within the first 48 h after birth, and injected with tritiated leucine (60 mCi/mL) as an anterograde axonal marker in the remaining left inferior olive, 3 months later. The animal was intracardially perfused 24 h after the injection. The remaining inferior cerebellar peduncle (ICP) is heavily labeled. Labeled axons are seen crossing the midline (arrow) and terminal labeled CFs are distributed over sagittal stripes in both the intact (right) and the deprived (left) cortices. Note the mirror images between normal CFs and compensatory innervation. Magnification 16. (B and C) adult rat cerebellum that had received a vermal injection of kainic at 2 months of age. One month later, the rat received a vermal graft of an isogenic piece of E14 cerebellar anlage. Two months after grafting, the rat received 200 nL of tritiated leucine in the inferior olivary complex, and was perfused 24 h later. The solid graft evolved into a minicerebellum (C) and radioautographic silver grains are present with a patchy distribution throughout the molecular layer of the graft (arrows in B), indicating that the new CFs are innervating only a selective population of PCs. Magnification 30. (A) Taken from Angaut, P., Alvarado-Mallart, R.M., Sotelo, C., 1985. Compensatory climbing fiber innervation after unilateral pedunculotomy in the newborn rat: origin and topographic organization. J. Comp. Neurol. 236 (2), 161e178. (B and C) Taken from Armengol, J.A., Sotelo, C., Angaut, P., Alvarado-Mallart, R.M., 1989. Organization of host afferents to cerebellar grafts implanted into kainate lesioned cerebellum in adult rats. Eur. J. Neurosci. 1 (1), 75e93.

11.6 Development and refinement of mossy fiber projections Mossy fibersdfor example, the spinocerebellar projectiondalso terminate in stripes in the cerebellar cortex (Brodal and Grant, 1962). For example, spinal MFs end almost exclusively in the granular layer of the vermis, where they are distributed within two distinct areas: lobules I to V, and lobules VIII/IX. The spinocerebellar projection is segregated into five parallel, parasagittal stripes varying in width from 200 to 480 mm, and separated by terminal-free intervals of 600e800 mm (which contain the terminal fields, inter alia, of the external cuneate projections: Ji and Hawkes, 1994). Striped terminal field projections are also reported for other classes of MF projections, independent of their origin (e.g., pontine projectionsdSerapide et al., 2001). The first comparison with PC stripes was for spinocerebellar MF projections to the vermis (Gravel and Hawkes, 1990; Ji and Hawkes, 1994). Since then, numerous studies have generated increasingly precise hodological maps (e.g., Cerminara et al., 2013). The same principle has emerged as for the CFsdthere is a consistent topographic relationship between MF terminal fields and PC stripes (Fig. 11.6). MFs of different origins arrive at different times during development. The earliest MFs to enter the cerebellum (E13) are the primary vestibular fibers (reviewed in Rahimi-Balaei et al., 2015). By E14 using anterograde tracing from the murine caudal cervical spinal cord, labeled spinocerebellar fibers are also revealed in the rostrolateral part of the

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

(A)

251

(D)

(B)

(C)

FIGURE 11.6 Development of the mossy fiber projection. Heterologous synapses in agranular cerebellum. (AeC) Topographic distribution of spinocerebellar afferents in the anterior lobe of rat pups aged P5 and P7 (tracer, WGA-HRP, was injected 24 h before fixation). Micrographs of coronal sections of pups taken with polarized illumination. (A) By P5 the columns begin to be formed (protocolumnar organization). Abundant fiber dispersion is present between the forming columns. By P7 (B) the intercolumnar dispersion decreases, and the adult columnar organization is already apparent. (C) Brightfield photomicrograph illustrating the central band in (B); the label is distributed within the white matter (WM) and the granular layer (GL), but is absent from the molecular layer (ML) and external granular layer (EGL). 140. Magnification (AeC) 210. (D) Electron micrograph of the upper molecular layer of a 22-day-old weaver mouse. PC dendritic profiles (PCD) are seen with numerous spines receiving either free postsynaptic differentiations (arrowheads) or establishing heterologous synaptic contacts (arrows) with a mossy fiber, 21,000. (AeC) Taken from Arsenio-Nunes, M.L., Sotelo, C., 1985. Development of the spinocerebellar system in the postnatal rat. J. Comp. Neurol. 237, 291e306. (D) Taken from Sotelo, C., 1975. Anatomical, physiological and biochemical studies of the cerebellum from mutant mice. II. Morphological study of cerebellar cortical neurons and circuits in the weaver mouse. Brain Res. 94 (1), 19e44.

cerebellum. By E17/18 labeled fibers were observed not only on the cerebellar cortex but also in the deep nuclei (Ashwell and Zhang, 1992). In the rat, trigeminocerebellar neurons from the interpolaris subnucleus of the nucleus of the trigeminal spinal tract, and neurons of the lateral reticular nucleus start to be labeled by E22 and P0, respectively (Ashwell and Zhang, 1992). Finally, the last neurons to be labeled, only after birth, are those of the pontine nuclei (Ashwell and Zhang, 1992). However, the mature striped organization of the projection is only qualitatively attained by P7. From P0 to P3, most axons are in a waiting period, disposed within the prospective white matter. As the granule cells arrive in number from the external granular layer (wP3), they invade the nascent granular layer. MF terminals have a tendency to cluster within their mature target PC clusters/stripes (the so-called protocolumn stage), but there are still numerous ectopic projections into the intervening spaces. The characteristic striped segregation of the adult axons is not attained until between P7 and P30. This timetable is consistent with ultrastructural examination. Early synapses between MFs and their postsynaptic granule cell dendrites are identified by P5 (Arsenio-Nunes and Sotelo, 1985). The spinal mossy rosettes appear as axonal varicosities partially covered by one or two postsynaptic dendrites, corresponding to the “primitive stage” of MF synaptogenesis. By P7 the number of mossy rosettes has increased and maturation has progressed and most are in the “cup stage” and may even be seen intermixed with terminals already in the mature “claw stage” (Fig. 11.7). The evidence that PCs organize the MF projection maps comes from multiple sources. First, transient ectopic synapses are found between PCs and MFs in the perinatal cerebellum (Mason and Gregory, 1984). These transient synapses are functional, as shown by Takeda and Maekawa (1989) for vestibular MFs. Bone morphogenetic protein 4 (BMP4) appears to act as a retrograde PC-derived signal that negatively regulates MF-PC contacts and their synaptic specificity. The specific genetic ablation of BMP4 from PCs led to MF growth and exuberant MF-PC interactions (Kalinovsky et al., 2011). Subsequently, as the granular layer matures, the MF-PC synapses are displaced by competition from the newly migrated granule cells. In the contest of the chemoaffinity hypothesis, the formation of ectopic synapses is interpreted as a wide

252 PART | I Circuit development

FIGURE 11.7 PCs are the organizers of the distribution od spinocerebellar MFs, evidence obtained from the study of the weaver (AeC) and staggerer (DeF) cerebella. (A) Electron micrograph of the neuropil of the molecular layer of an adult weaver (wv/wv) vermal cortex showing a PC dendritic profile (PCD) surrounded by numerous free spines (S) provided with postsynaptic differentiations facing glial profiles. Note the lack of parallel fibers. 59,000. (D) Molecular layer of a 28-day-old staggerer mouse. The neuropil contains bundles of parallel fibers separated by thin protoplasmic lamellae of astrocytes. The abundant parallel fibers are alternating with presynaptic varicosities (asterisk), free of postsynaptic elements, and directly apposed to glial processes (arrowhead). Note the complete absence of PC dendritic spines. 31,000. (B, C, E, and F) Photomicrographs using polarized light. Coronal sections at different levels of the cerebellum. Comparison of the projection pattern of the spinal axons in the posterior (B) and the anterior (C) regions of the rostral zone of the vermis in weaver mice. Note that despite the absence of granule cells the mossy fibers are organized in parasagittal bands. In the posterior region (B), in addition to their concentration in the midsagittal band, a few spinal fibers project in a peripheral bilateral area where they assume a roughly columnar distribution. In the anterior region, the spinal axons are distributed within four pairs of symmetrically disposed parallel bands (numbers). (E) The staggerer vermis is atrophic, but lobulated with disorganized PC bodies, and short zones depleted of them (E arrows). At the center of the cerebellum spinocerebellar axons are densely packed, forming a central cluster with a balllike form (E) and much more dispersed in more lateral zones. (F) Higher magnification of the central spinocerebellar axons. The bars in (C and F) represent 90 mm. (A) From Sotelo, C., 1975. Anatomical, physiological and biochemical studies of the cerebellum from mutant mice. II. Morphological study of cerebellar cortical neurons and circuits in the weaver mouse. Brain Res. 94 (1), 19e44. (D) From Sotelo, C., 1990. Cerebellar synaptogenesis: what we can learn from mutant mice. J. Exp. Biol. 153, 225e249; Arsenio-Nunes et al. (1988) (B, C wv; E,F sg). Wassef, M., Sotelo, C., Cholley, B., Brehier, A., Thomasset, M., 1987. Cerebellar mutations affecting the postnatal survival of Purkinje cells in the mouse disclose a longitudinal pattern of differentially sensitive cells. Dev. Biol. 124, 379e389.

range of graded preferences between synaptic partners rather than an exclusive or even a narrow range of synaptic affinities (see Sotelo, 2004). Thus, we see the persistence of MF-PC synapses in the absence of competition with parallel fibers, when the PCs remain in a severely synaptic “non-saturated” state, as is the case in granuloprival cerebella. The first electrophysiological and ultrastructural evidence for the persistence of functional MF-PC synapses in the adult cerebellar

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

253

cortex was obtained in ferrets after infection with panleukopenia virus (Llinás et al., 1973). Similar results were obtained in weaver (Kcnj6wv) mutant mice (Sotelo, 1975). In this mutant, the granule cells carry a missense mutation in a G-protein coupled inwardly rectifying potassium channel (GIRK2: Patil et al., 1995); the spinocerebellar projection pattern is roughly normal despite the rarity of granule cells (Arsenio-Nunes et al., 1988). In contrast, in the staggerer mouse (sg/sg Rorasg), in which the molecular target of the mutation is the PC retinoid-related orphan receptor a (Hamilton et al., 1996), which prevents PC development in a cell-autonomous fashion, resulting in a large decrease in their number and altered function and immature morphology, the spinocerebellar projection is severely altereddthe MF stripes are missing and the MF terminals are distributed more or less uniformly within the cortex (Arsenio-Nunes et al., 1988). Thus granule cells are neither necessary to attract MFs to their proper terminal domains nor for their parasagittal organization. PCs, on the contrary, are required for the proper location of these fibers, and thus appear to be the embryonic organizers of the cerebellar projection maps (Sotelo and Wassef, 1991). Consistent with this interpretation, surgical lesion of the spinocerebellar projection does not result in the expansion of the terminal fields of the neighboring cuneocerebellar projection (Ji and Hawkes, 1995).

11.7 Cerebellar interneurons in the cerebellar circuit The cerebellar cortex contains at least seven different types of interneurons: two types of excitatory interneuronsdthe granule cells and the UBCs, which originate from the upper rhombic lip, and five types of inhibitory interneuronsdbasket, stellate, Golgi, Lugaro, and candelabrumdthat arise from the ventricular zone of the fourth ventricle (Fig. 11.8).

11.7.1 Excitatory interneurons 11.7.1.1 Granule cells Granule cells are the most numerous neurons of the mammalian brain. They receive their primary afferent input from MFs and project in turn to the PC dendritic arbor as parallel fibers. During development, granule cell precursors leave the upper rhombic lip at around E13, and via tangential migration invade the superficial region of the cerebellar cortex to establish a main proliferative zone in the so-called external granule cell layer that by E16 covers the whole cerebellar surface, and persists until the third postnatal week (Miale and Sidman, 1961). After their proliferative phase that occurred mainly during the first 2-weeks postnatal, they start their migration, first with a tangential migration (Komuro et al., 2001) followed by a second radial migration through the molecular layer along Bergmann radial fibers to their final laminar position in the maturing granular layer. Their cell bodies with twoeseven short dendrites, although 62% of them have four dendrites, are localized in the granule cell layer. They receive no synapses at their somatic surface but the terminal dendritic branches end in an enlargement called a “dendritic claw” or “dendritic digit” that encloses the excitatory MF terminal in a glomerulus. The complex structural organization of the MF->granule cell glomerulus has been described in several species (see references in Palay and Chan-Palay, 1974). It comprises glial cells and two presynaptic elements, the excitatory MFs and the inhibitory Golgi axons varicosities, that make multiple synapses on the dendritic claws of the granule cells. A single MF makes excitatory synaptic contact en passant with the dendrites of several cerebellar granule cells. It is estimated that the ratio of postsynaptic granule cells per MF is in the range of 1:10 to 1:100, indicating a high degree of synaptic divergence (Billing et al., 2014). Cadherin-7 (Cad-7), a class II cadherin cell adhesion molecule, mediates the specific innervation. Both granule cells and MFs express Cad-7, and a Cad-7 homophilic interaction regulates both MF axonal growth termination and specific synapse formation. Cad-7 knockdown in MFs impairs their connectivity with granule cells (Kuwako et al., 2014). Activation of granule cells by MFs initiates feedforward-excitation of PCs by parallel fibers, the axons of the granule cell. Each parallel fiber makes a single synapse with a given PC on a dendritic spine, one parallel fiber may contact up to a few hundreds of PCs, and one PC can receive up to 200,000 granule cell synapses (Harvey and Napper, 1991). The molecular mechanism that controls the specificity of parallel fiber->PC synapses is mediated by cerebellin-1 (Cbln1), a secreted protein expressed by granule cells, that functions as a trans-synaptic cell adhesion molecule by binding to presynaptic neurexins and postsynaptic GluRd2 localized at PC dendritic spines (Uemura et al., 2010). Deletion of Cbln1 in mice induces severe ataxia closely resembling that observed in GluRd2 KO mice (Uemura et al., 2010). Granule cells, like PCs, are not homogeneous. It is likely that the cerebellar granular layer has multiple lineage histories and derives from multiple distinct precursor pools within the rhombic lip. Furthermore, these lineages align with PC antigenic boundaries (Hawkes et al., 1999) and with lineage compartments seen in murine chimeras (e.g., Swanson and Goldowitz, 2011). Consistent with this, genetic fate mapping shows that early-born granule cell progenitors

254 PART | I Circuit development

FIGURE 11.8 Morphological feature and laminar localization of cerebellar interneurons. (AeH) Original drawing of classical cerebellar interneurons by Ramón y Cajal (A ¼ basket, D ¼ stellate, G ¼ granule, H ¼ Golgi cells), and of post-Cajal new classes of interneurons, drawings by Herbert Axelrad (B ¼ Lugaro, C ¼ candelabrum, E ¼ unipolar brush, F ¼ globular cells). (I) Schematic representation of the main interneuron types and their laminar localization. Axons are drawn with a thinner line than the dendrites. Note the specific localization of axon and dendrite arborization for each type of interneuron in the granule cell layer (GCL), Purkinje Cell Layer (PCL), or Molecular Layer (ML). Sotelo, C., 1975. Anatomical, physiological and biochemical studies of the cerebellum from mutant mice. II. Morphological study of cerebellar cortical neurons and circuits in the weaver mouse. Brain Res. 94 (1), 19e44; Arsenio-Nunes, M.L., Sotelo, C., Werhrlé, R., 1988. Organization of spinocerebellar projection map in three types of agranular cerebellum: Purkinje cells vs. granule cells as organizer element. J. Comp. Neurol. 273, 120e136; Sotelo, C., 1990. Cerebellar synaptogenesis: what we can learn from mutant mice. J. Exp. Biol. 153, 225e249.

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

255

(E12.5eE15.5) migrate preferentially into the anterior vermis, whereas later-born ones distribute more evenly along the anteroposterior axis, and late-born ones (wE17) primarily occupy the vestibulocerebellum. These late-born granule cells are also unique in the expression of Lmx1 (Chizhikov et al., 2010). Similar restriction boundaries are seen in the phenotypes of many mutant mice (e.g., NeuroD1/ mutant: Miyata et al., 1999). Finally, differential gene expression patterns reveal boundaries that divide the granular layer into both transverse zones and parasagittal stripes, all of which align with the PC map (e.g., neuronal nitric oxide synthase). These granule cell phenotypes and expression boundaries are sometimes lineage restricted and other times reflect secondary interactions with the local environment. Granular layer heterogeneity and patterning is reviewed in more detail in Armstrong and Hawkes (2014).

11.7.1.2 Unipolar brush cells (UBCs) Although UBCs and granule cells both originate from Math-1þ precursors in the rhombic lip, they are clearly from distinct lineages. Unlike granule cells, they express the transcription factor Tbr2 and migrate from the external granular layer to the granular layer between P0 to P10 before the completion of granule cell migration (Englund et al., 2006). This recently discovery of glutamatergic interneurons being concentrated primarily in the vestibulocerebellum (see review in Mugnaini et al., 2011) is another example of the patterned organization of the cerebellar cortex (e.g., Chung et al., 2009a). UBC axons ramify within the granular layer, terminating in large MFelike varicosities with granule cell dendrites and the terminal dendrioles of other UBCs (see Mugnaini et al., 2011). Like granule cells, the UBC dendritic brushes are also contacted by extrinsic MFs, but with a constant one-to-one relation between an MF varicosity and a unique UBC dendrite, which ensures the individuality of the MF input pathway. Inhibitory synapses have been also observed in contact with the UBC dendritic brushes and rarely with the cell body of UBCs (see Mugnaini et al., 2011). These inputs display reactivity for GABA and glycine and therefore are attributed to Golgi cells (reviewed in Mugnaini et al., 2011). In addition to the UBCs lobular restriction reported above, unipolar brush cells of different phenotypes are also aligned in the granular layer beneath specific PC stripes (Chung et al., 2009a,b). This restriction is more evident in the Ebf2 null mouse cerebellum where, in addition to some PC death, there is a failure to repress the zebrin IIþ phenotype in the normally zebrin II PC population. This dislocation of the PC patterning is accompanied by UBC invasion of the transdifferentiated anterior vermis, and speaks to the PC architecture as the element that restricts interneuron distributions (Chung et al., 2009a).

11.7.2 Inhibitory interneurons 11.7.2.1 Granule cell layer interneurons Camillo Golgi (1874) with his “black reaction” method described the two types of large inhibitory interneurons in the granular layer of the cerebellum. The first, with a long fusiform soma located just below the PC layer, was most probably rediscovered years later by Ernesto Lugaro (1894) and thereafter named the “Lugaro” cell. The second type with polygonal cell somata located throughout the granular layer were called “the Golgi cells of the cerebellum” by Retzius, a terminology followed by Cajal (1911). 11.7.2.1.1 Golgi cells Five distinct subpopulations of Golgi cells are identified based on neurochemical phenotype (e.g., expression of mGluR2, neurogranin, GAD67 or glycine transporter type 2 (GlyT2)), and cell shape, size, and location in the granular layer (Simat et al., 2007). In addition, at least in cat and human, a subpopulation of Golgi cells is cholinergic. This population (about 5%) is concentrated in the vestibulocerebellum (Illing, 1990). However, despite molecular differences all Golgi cells share a similar morphologyda characteristic dendritic arbor that extends into the molecular layer, often reaching the pial surface (Golgi, 1874), and an axonal arbor restricted to the granular layer (Palay and Chan-Palay, 1974). Golgi cells are the sole inhibitory inputs to UBCs (reviewed in Mugnaini et al., 2011), and are the major source of inhibitory synapses on granule cells claws in the glomeruli (Eccles et al., 1967). While the majority of Golgi cells use both GABA and glycine as neurotransmitters, a single Golgi cell will differentially mediate GABAergic or glycinergic transmission depending on its target. Granule cells and UBCs display GABAergic and glycinergic currents, respectively, upon activation of Golgi cells (Dugué et al., 2005), mainly due to cell-type-specific postsynaptic receptor expression and/or trafficking to the synapse since Golgi cells corelease both GABA and glycine at each individual bouton. Golgi cells receive feedforward and feedback excitatory inputs from MFs in the granular layer and parallel fibers in the molecular layer (see references in Kanichay and Silver, 2008). However, the identity of inhibitory inputs to Golgi cell

256 PART | I Circuit development

FIGURE 11.9 Role of Ptf1a and Atoh-1 in neural specification in the cerebellum. Schematic representation of embryonic cerebellar development (E10.5 to E13.5). Ptf1a-expressing progenitors from the ventricular zone produce Deep Cerebellar Nuclei (DCN) inhibitory neurons starting at E10.5. At later stages, they will also produce Purkinje cells (E11.5) and Pax-2 precursors (E12.5) give rise to all inhibitory interneurons (Golgi, Lugaro, Globular, Candelabrum, Basket, and Stellate Cells). In parallel, Atoh-1-expressing progenitors from the rhombic lip produce all excitatory neurons, including DCN excitatory neurons, granule cells, and UBC. Original schema from Fabrice Ango.

dendrites in the molecular layer has been controversial. While it was assumed that molecular layer interneurons (i.e., stellate and basket cells) make GABAergic synapses on Golgi cell dendrites based on electron microscopy (Palay and Chan-Palay, 1974) and electrophysiological evidence (Dumoulin et al., 2001), recent studies argue against the presence of such functional synapses (Eyre and Nusser, 2016). Rather, only a minority of the inhibitory inputs to Golgi cells in the molecular layer originate from local interneurons, and the majority of their inhibitory inputs exclusively release GABA. Instead, GABAergic cells in the DCN also innervate Golgi cells, suggesting that Golgi cells receive major feedback inhibitory inputs from the deep cerebellar nuclei to regulate local inhibitory networks (Ankri et al., 2015). Recently it has been also established that some Golgi cells, in regions that regulate eye movements and vestibulocerebellum, receive tonic inhibitory inputs from PC axon collaterals that create a feedback mechanism to mediate temporal integration (Guo et al., 2016). In addition, Golgi cells also receive inhibitory glycinergic innervation from Lugaro cells (Dumoulin et al., 2001) (Fig. 11.9). 11.7.2.1.2 Lugaro cells Lugaro cells are characterized by a fusiform cell body with thick proximal dendrites extending horizontally into the PC layer and axons that project more or less directly into the molecular layer. On the basis of their cell soma morphology, Lugaro cells are divided into three subgroups, large fusiform, triangular, and small fusiform (Lainé and Axelrad, 2002; Simat et al., 2007). There may also be a fourth subtyped the deep Lugaro celldwith globular somata located throughout the granular layer and dendrites extending into the molecular layer, similar to the Golgi cell dendrites but with axon targets also confined to the molecular layer, a hallmark of Lugaro cells (Lainé and Axelrad, 2002). Lugaro cells receive synaptic input from the recurrent collaterals of PC axons in the subganglionic plexus and form inhibitory GABAergic and glycinergic synapses on stellate, basket, and particularly Golgi cell dendrites (Dumoulin et al., 2001). They are excited by monoaminergic inputs and, notably, serotonin induces strong spiking activity that generates large-amplitude inhibitory postsynaptic potentials in Golgi cells (Dieudonné and Dumoulin, 2000). To date, no other excitatory input has been identified on Lugaro cells. Finally, Lugaro cells selectively innervate other inhibitory interneurons, via their parasagittal-oriented axons that preferentially contact stellate/basket cells (Lainé and Axelrad, 1998) and their transverse fibers that contact Golgi cells (Dumoulin et al., 2001). Golgi and Lugaro cells both go through their last mitosis between E13.5 and P 4 (see references in Leto et al., 2016). They start out in the Ptf1aþ domain of the ventricular zone that earlier generated the PCs, and migrate through the prospective white matter where they continue to divide (so-called “transit amplification”: see Leto et al., 2016). From the prospective white matter, they migrate to the nascent granular layer where they settle and differentiate.

11.7.2.2 Purkinje cell layer interneurons 11.7.2.2.1 Candelabrum cells The somata of candelabrum cells lie interspersed between the PC bodies and are smaller than the latter (Lainé and Axelrad, 1994). The dendritic arbors emerge from their apical poles in the form of chandelierlike processes that spread throughout

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

257

the molecular layer. Their axons run horizontally above the PC layer, giving off vertically oriented beaded branches that ascend through the major part of the molecular layer. It has been hypothesized that candelabrum interneurons are GABAergic (Sudarov et al., 2011) because they are marked in the transgenic Ascl1-CreER BAC line, which marks all other GABAergic cerebellar interneurons. The candelabrum cells remain one of the most enigmatic interneurons of the cerebellar cortex, and neither inputs nor synaptic targets are well understood. However, the work of Sudarov et al. (2011) together with the morphology described by Lainé and Axelrad (1994) suggests that PCs are the likely target of candelabrum cells, and that their axons might contact the PC dendrites for all their extent. Regarding the development of candelabrum interneurons, it is known that all other cerebellar GABAergic neurons originate from the ventricular zone of the cerebellar primitive neuroepithelium. Candelabrum cell birthdays seem to be early postnatal, at ages between the basket cells (the oldest) and the stellate cells (the youngest), but we have no details. Also, Sudarov et al. (2011) concluded that their final positions depend on when they first express the transcription factor Ascl1, thereby creating an inside-outside pattern.

11.7.2.3 Molecular layer interneurons The molecular layer basket and stellate cells were originally classified as distinct cell types based on their axonal morphology and the location of their somata (Cajal, 1911). However, the lack of a specific genetic marker to identify them during development, their similar morphological features, and somatic recombination-based clonal analyses that suggest that basket and stellate cells belong to the same lineage sustain a long-lasting debate about whether they belong to the same cell population that differentiate differently according to the local environments in which they settle (Rakic, 1972; Sultan and Bower, 1998). Basket interneurons are generated in the white matter by transit amplification between P3 and P9 and stellate interneurons are born mainly between P5 and P12 (Sudarov et al., 2011) suggesting that temporal information in the neuroepithelium may be involved in their different specification. Similarly, by making use of public gene expression resources Schilling and Oberdick (2009) found evidence that basket and stellate cells may be molecularly distinct. From the prospective white matter to the molecular layer, precursors of molecular layer interneurons receive GABAergic and glutamatergic synaptic inputs that modulate their migratory speed and directionality in the granular layer. Putative GABAergic synapses between immature molecular layer interneurons have been identified as early as P4 and could regulate molecular layer interneuron migration (Simat et al., 2007); however, the nature of excitatory input is currently unknown. The migration of basket and stellate cell progenitors in the molecular layer exhibits four distinct phases in an organotypic slice model (Cameron et al., 2009). Rostrocaudal tangential migration across the surface of the molecular layer is followed by a second phase in which the cell changes its orientation from horizontal to vertical. Next comes a radial migration toward the PC layer and finally, in the fourth phase, a second tangential migration occurs in the middle of the molecular layer. Whether both basket and stellate cells go through these sequential migration phases is not known. Indeed, since most basket cells occupy the deeper part of the molecular layer while stellate cells are enriched in the more superficial part, one can argue that they might use different migration routes. Finally, both basket and stellate cells are restricted at PC stripe boundaries (the topic is reviewed in Leto et al., 2016) (Fig. 11.10). 11.7.2.3.1 Basket cells During development, basket cell axons are the first GABAergic terminals to synapse on PCs (by P6: Ango et al., 2004). Multiple dendrites extend from the cell body and radiate into the molecular layer where they receive inhibitory synaptic input from other molecular layer interneurons and excitatory inputs mainly from parallel fibers (see in Palay and Chan-Palay, 1974). The axon of the basket cell usually runs rostrocaudally above the PC layer and sends several descending collaterals toward the PC somata. The descending collaterals form a pericellular nest around the PC soma and continue to grow toward the initial axon segment to form an axo-axonic synapsedthe so-called “pinceau”dconsisting of multiple basket cell axon branchlets that completely enwrap the PC initial axon segments (Cajal, 1911; Ango et al., 2004; Sotelo, 2008). During development, once basket axon collaterals reach the PC somata, a subcellular gradient of NF186 (a cell adhesion molecule of the L1-CAM family) directs basket axon terminals to the PC initial axon segment. The ankyrin-Gassociated form of NF186 at the initial axon segment is necessary for pinceau formation and/or stabilization since knockdown of the scaffolding protein Ankyrin-G or its binding partner neurofascin (NF186) disrupts pinceau synapse stabilization, even in adult mice (Ango et al., 2004; Buttermore et al., 2012). Recently, the presynaptic binding partner of neurofascin 1 was identified as the axon guidance receptor neuropilin 1. It is expressed in basket cell axons and directly interacts in trans with postsynaptic neurofascin 1 during target recognition (Telley et al., 2016). Notably Sema3A, the canonical ligand of neuropilin 1, is only expressed by PCs during cerebellar development (Saywell et al., 2014), where it

258 PART | I Circuit development

FIGURE 11.10 Molecular and cellular mechanisms of specific synapse formation on Purkinje cells subcellular domains. (A) Granule cells make specific synaptic contact on Purkinje dendritic spines. The formation of these synapses is achieved by transsynaptic interaction between the postsynaptic receptor GluD1/D2 and the presynaptic neurexin together with C1q proteins. (B) Basket cells innervate the soma and AIS of Purkinje cells. The axon guidance molecule Sema3A binds and stabilizes the presynaptic NRP1 that interacts with the postsynaptic neurofascin to trigger synapse formation. (C) Stellate cells make specific synaptic contacts with the dendritic shaft of Purkinje cells. Close Homologue of L1 (CHL1) is localized to Bergmann glial fibers, and contributes to the organization of stellate axons along BG fibers and to the innervation of Purkinje cell dendrites. Note that the pre and postsynaptic partners are not yet identified. Original schemas from Fabrice Ango.

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

259

triggers basket cell axon guidance toward the PC somata. At the PC somata, Sema3A stabilizes neurofascin 1 and facilitates its interaction with neuropilin 1 at the initial axon segment (Telley et al., 2016). Finally, Sema3A induces basket axon terminal branching through a Fyn-dependent mechanism (Cioni et al., 2013). As for other interneurons, basket cells have multiple synaptic targets and a few basket axon collaterals even grow upward toward the PC dendrites, suggesting that distinct mechanisms control basket cell axon architecture and their target innervation pattern. One intriguing aspect of the pinceau is that it is largely devoid of chemical synapses and gap junctions (Sotelo, 2008), even if the latter are present in “empty baskets” after degeneration of PCs, as is the case in the adult cerebellum of nervous mutant mice (Sotelo and Triller, 1979). This situation questions the function of pinceau in PC inhibition. Molecular and anatomical evidence shows that pinceau develop septatelike junctions, with high expression of voltage-dependent potassium channels and a lack of voltage-dependent sodium channels (Sotelo, 2008), which suggest electrical inhibition. An elegant study by Blot and Barbour (2014) confirmed this hypothesis and demonstrated that pinceau reduce the PC firing rate by ephaptic inhibition, an ultrafast feedforward inhibition that allows granule cells to inhibit PCs without a preceding excitation phase. 11.7.2.3.2 Stellate cells In contrast to basket cells, whose somata are enriched in the deeper third of the molecular layer, stellate cell somata preferentially occupy the superficial two-thirds of the molecular layer. They extend several diverging dendrites with an axon that develop in the parasagittal plane with an extensive ramification of ascending and descending collaterals covered with varicosities. The main functional difference between basket and stellate cells is their innervation of the PCs. While basket cells innervate the somata and initial axon segments of PCs, stellate cells form synapses mainly on the dendritic shafts. Interestingly, basket and stellate cells use completely different cellular and molecular mechanisms to achieve dendritic innervation. Stellate cells grow multiple collaterals that associate with glia during development and extend their axons that contact the Bergmann glial process fibers (Ango et al., 2008) via an interaction with the L1-CAM superfamily cell adhesion molecule, CHL1. Knocking-down CHL1 expression is sufficient to induce aberrant stellate axon organization and to decrease synaptic contact with PC dendrites.

Acknowledgment This work was supported by grants from the Agence Nationale de la Recherche (ANR-14-CE13-0004-01).

References Ahn, A.H., Dziennis, S., Hawkes, R., Herrup, K., 1994. The cloning of zebrin II reveals its identity with aldolase C. Development 120, 2081e2090. Andjus, P.R., Zhu, L., Cesa, R., Carulli, D., Strata, P., 2003. A change in the pattern of activity affects the developmental regression of the Purkinje cell polyinnervation by climbing fibers in the rat cerebellum. Neuroscience 121 (3), 563e572. Angaut, P., Alvarado-Mallart, R.M., Sotelo, C., 1985. Compensatory climbing fiber innervation after unilateral pedunculotomy in the newborn rat: origin and topographic organization. J. Comp. Neurol. 236 (2), 161e178. Ango, F., Di Cristo, G., Higashiyama, H., Bennett, V., Wu, P., Huang, Z.J., 2004. Ankyrin-based subcellular gradient of neurofascin, an immunoglobulin family protein, directs GABAergic innervation at Purkinje axon initial segment. Cell 119, 257e272. Ango, F., Wu, C., van der Want, J.J., Wu, P., Schachner, M., Huang, Z.J., 2008. Bergmann glia and the recognition molecule CHL1 organize GABAergic axons and direct innervation of Purkinje cell dendrites. PLoS Biol. 6 (4), e103. https://doi.org/10.1371/journal.pbio.0060103. Ankri, L., Husson, Z., Pietrajtis, K., Proville, R., Léna, C., Yarom, Y., Dieudonné, S., Uusisaari, M.Y., 2015. A novel inhibitory nucleo-cortical circuit controls cerebellar Golgi activity. eLIFE 4, e06262. Apps, R., Hawkes, R., 2009. Cerebellar cortical organization: a one-map hypothesis. Nat. Rev. Neurosci. 10, 670e681. Apps, R., Hawkes, R., Aoki, S., Bengtsson, F., Brown, A.M., Chen, G., Ebner, T.J., Isope, P., Jörntell, H., Lackey, E.P., Lawrenson, C., Lumb, B., Schonewille, M., Sillitoe, R.V., Spaeth, L., Sugihara, I., Valera, A., Voogd, J., Wylie, D.R., Ruigrok, T.J.H., 2018. Cerebellar modules and their role as operational cerebellar processing units: a consensus paper. Cerebellum. https://doi.org/10.1007/s12311-018-0952-3. Armengol, J.A., Sotelo, C., Angaut, P., Alvarado-Mallart, R.M., 1989. Organization of host afferents to cerebellar grafts implanted into kainate lesioned cerebellum in adult rats. Eur. J. Neurosci. 1 (1), 75e93. Armstrong, C.L., Hawkes, R., 2014. Pattern formation in the cerebellum. In: Colloquium Digital Library of Life Sciences. Morgan & Claypool, p. 138. Arsenio-Nunes, M.L., Sotelo, C., 1985. Development of the spinocerebellar system in the postnatal rat. J. Comp. Neurol. 237, 291e306. Arsenio-Nunes, M.L., Sotelo, C., Werhlé, R., 1988. Organization of spinocerebellar projection map in three types of agranular cerebellum: Purkinje cells versus granule cells as organizer element. J. Comp. Neurol. 273, 120e136. Ashwell, K.W., Zhang, L.L., 1992. Ontogeny of afferents to the fetal rat cerebellum. Acta Anat. 145 (1), 17e23.

260 PART | I Circuit development

Azevedo, F.A.C., Carvalho, L.R.B., Grinberg, L.T., Farfel, J.M., Ferreti, R.E.L., Leite, R.E.P., Jacob filho, W., Lent, R., Herculano-Houzel, S., 2009. Equal number of neuronal and nonneuronal cells make the human brain an isometrivally sealed-up primate brain. J. Comp. Neurol. 513, 532e541. Billing, G., Piasini, E., LTrincz, A., Nusser, Z., Silver, R.A., 2014. Network structure within the cerebellar input layer enables lossless sparse encoding. Neuron 83 (4), 960e974. Blot, A., Barbour, B., 2014. Ultra-rapid axon-axon ephaptic inhibition of cerebellar Purkinje cells by the pinceau. Nat. Neurosci. 17 (2), 289e295. Brochu, G., Maler, L., Hawkes, R., 1990. Zebrin II: a polypeptide antigen expressed selectively by Purkinje cells reveals compartments in rat and fish cerebellum. J. Comp. Neurol. 291, 538e552. Brodal, A., Grant, U., 1962. Morphology and temporal course of degeneration in cerebellar mossy fibers following transection of spinocerebellar tracts in the cat. An experimental study with silver methods. Exp. Neurol. 5, 67e87. Buttermore, E.D., Piochon, C., Wallace, M.L., Philpot, B.D., Hansel, C., Bhat, M.A., 2012. Pinceau organization in the cerebellum requires distinct functions of neurofascin in Purkinje and basket neurons during postnatal development. J. Neurosci. 32 (14), 4724e4742. Cajal, S.R., 1911. Histologie du Système Nerveux de l’Homme et des Vertébrés, vol. 2. Maloine, Paris. Cameron, D.B., Kasai, K., Jiang, Y., Hu, T., Saeki, Y., Komuro, H., 2009. Four distinct phases of basket/stellate cell migration after entering their final destination (the molecular layer) in the developing cerebellum. Dev. Biol. 332 (2), 309e324. Cerminara, N.L., Aoki, H., Loft, M., Sugihara, I., Apps, R., 2013. Structural basis of cerebellar microcircuits in the rat. J. Neurosci. 33 (42), 16427e16442. Changeux, J.P., Danchin, A., 1976. Selective stabilization of developing synapses as a mechanism for the specification od neuronal networks. Nature 264 (5588), 705e771. Chédotal, A., Sotelo, C., 1993. The “creeper stage” in cerebellar climbing fiber synaptogenesis precedes the “pericellular nest”: ultrastructural evidence with parvalbumin immunocytochemistry. Dev. Brain Res. 76 (2), 207e220. Chédotal, A., Pourquié, O., Ezan, F., San Clemente, H., Sotelo, C., 1996. BEN as a presumptive target recognition molecule during the development of the olivocerebellar system. J. Neurosci. 16 (10), 3296e3310. Chédotal, A., Bloch-Gallego, E., Sotelo, C., 1997. The embryonic cerebellum contains topographic cues that guide developing inferior olivary axons. Development 124 (4), 861e870. Chizhikov, V.V., Lindgren, A.G., Mishima, Y., Roberts, R.W., Aldinger, K.A., Miesegaes, G.R., Currle, D.S., Monuki, E.S., Millen, K.J., 2010. Lmx-1a regulates fates and location of cells originated from the cerebellar rhombic lip and telencephalic cortical hem. Proc. Natl. Acad. Sci. U.S.A. 107 (23), 10725e10730. Chung, S.H., Marzban, H., Croci, L., Consalez, G.G., Hawkes, R., 2008. Purkinje cell subtype specification in the cerebellar cortex: early B-cell factor 2 acts to repress the zebrin II-positive Purkinje cell phenotype. Neuroscience 153 (3), 721e732. Chung, S.H., Marzban, H., Watanabe, M., Hawkes, R., 2009. Phospholipase Cbeta4 expression identifies a novel subset of unipolar brush set in the adult mouse cerebellum. Cerebellum 8, 267e276. Chung, S.H., Sillitoe, R.V., Croci, L., Badaloni, A., Consalez, G., Hawkes, R., 2009. Purkinje cell phenotype restricts the distribution of unipolar brush cells. Neuroscience 164 (4), 1496e1508. https://doi.org/10.1016/j.neuroscience.2009.09.080. Cioni, J.M., Telley, L., Saywell, V., Cadihac, C., Jourdan, C., Huber, A.B., Huang, J.Z., Jahannault-Talignani, C., Ango, F., 2013. SEMA3A signaling controls layer-specific interneuron branching in the cerebellum. Curr. Biol. 23, 850e861. Crépel, F., Mariani, J., Delhaye-Bouchaud, N., 1976. Evidence for a multiple innervation of Purkinje cells by climbing fibers in the immature cerebellum. J. Neurobiol. 7 (6), 567e578. Dastjerdi, F.V., Consalez, G.G., Hawkes, R., 2012. Pattern formation during development of the embryonic cerebellum. Front. Neuroanat. 6, 10. https:// doi.org/10.3389/fnana.2012.00010. De Camilli, P., Miller, P.E., Levitt, P., Walter, U., Greengard, P., 1984. Anatomy of cerebellar Purkinje cells in the rat determined by a specific immunohistochemical marker. Neuroscience 11, 761e817. Dieudonné, S., Dumoulin, A., 2000. Serotonin-driven long range inhibitory connections in the cerebellar cortex. J. Neurosci. 20 (5), 1837e1848. Dugué, G.P., Dumoulin, A., Triller, A., Dieudonné, S., 2005. Target-dependent use of co-released inhibitory transmitters at central synapses. J. Neurosci. 25, 6490e6498. Dumoulin, A., Triller, A., Dieudonné, S., 2001. IPSC kinetics at identified GABAergic and mixed GABAergic and glycinergic synapses onto cerebellar Golgi cells. J. Neurosci. 21 (16), 6045e6057. D’Arcangelo, G., 2014. Reelin in the years: controlling neuronal migration and maturation in the mammalian brain. Adv. Neurosci., 597395 https:// doi.org/10.1155/2014/597395. Eccles, J.C., Llinás, R., Sasaki, K., 1966. The excitatory synaptic action of climbing fibers on the Purkinje cells of the cerebellum. J. Physiol. 182 (2), 268e296. Eccles, J.C., Llinás, R., Sasaki, K., 1966. The mossy fibre-granule cell relay of the cerebellum and its inhibitory control by Golgi cells. Exp. Brain Res. 1 (1), 82e101. Eccles, J.C., Sasaki, K., Strata, P.A., 1967. Comparison of the inhibitory actions of Golgi cells and of basket cells. Exp. Brain Res. 3 (1), 81e94. Englund, C., Kowalczyk, T., Daza, R.A., Dagan, A., Lau, C., Rose, M.F., Hevner, R.F., 2006. Unipolar brush cells of the cerebellum are produced in the rhombic lip and migrate through developing white matter. J. Neurosci. 26, 9184e9195. Eyre, M.D., Nusser, Z., 2016. Only a minority of inhibitory inputs to cerebellar Golgi cells originates from local GABAergic cells. eNeuron, 0055e16.2016. Golgi, C., 1874. Sulla fina anatomia del cervelletto umano. In: Archivio Italiano per le Malattie Nervose, vol. 2, pp. 90e107.

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

261

Graus-Porta, D., Blaess, S., Senften, M., Littlewood-Evans, A., Damsky, C., Huang, Z., Orban, P., Klein, R., Schittny, J.C., Müller, U., 2001. Beta 1-class integrins regulate the development of the laminae and folia in the cerebral and cerebellar cortex. Neuron 31 (3), 367e379. Gravel, C., Hawkes, R., 1990. Parasagittal organization of the rat cerebellar cortex: direct comparison of Purkinje cell compartments and the organization of the spinocerebellar projection. J. Comp. Neurol. 291 (1), 79e102. Gravel, C., Eisenman, L., Sasseville, R., Hawkes, R., 1987. Parasagittal organization of the rat cerebellar cortex: a direct correlation between antigenic bands revealed by mabQ113 and the topography of the olivocerebellar projection. J. Comp. Neurol. 265, 294e310. Guo, C., Witter, L., Rudolph, S., Elliot, H.L., Ennis, K.A., Regehr, W.G., 2016. Purkinje cells directly inhibit granule cells in specialized regions of the cerebellar cortex. Neuron 91 (6), 1330e1341. Hamilton, B.A., Frankel, W.N., Kerrebrock, A.W., Hawkins, T.L., FitzHugh, W., Kusumi, K., Russell, L.B., Mueller, K.L., van Berkel, V., Birren, B.W., Kruglyak, L., Lander, E.S., 1996. Disruption of the nuclear hormone receptor RORalpha in staggerer mice. Nature 379 (6567), 736e739. Harvey, R.J., Napper, R.M., 1991. Quantitative studies on the mammalian cerebellum. Prog. Neurobiol. 36 (6), 437e463. Hashimoto, K., Kano, M., 2003. Functional differentiation of multiple climbing fiber inputs during synapse elimination in the developing cerebellum. Neuron 38 (5), 785e796. Hashimoto, K., Ichikawa, R., Kitamura, K., Watanabe, M., Kano, M., 2009. Translocation of a “winner” climbing fiber to the Purkinje cell dendrite and subsequent elimination of “losers” from the soma in developing cerebellum. Neuron 63 (1), 106e118. Hashimoto, K., Yoshida, T., Sakimura, K., Mishina, M., Watanabe, M., Kano, M., 2009. Influence of parallel fiber-Purkinje cell synapse formation on postnatal development of climbing fiber-Purkinje cell synapses in the cerebellum. Neuroscience 162 (3), 601e611. Hawkes, R., Belerbach, E., Tan, S.S., 1999. Granule cell dispersion is restricted across transverse boundaries in mouse chimeras. Eur. J. Neurosci. 11 (11), 3800e3808. Hiesberger, T., Trommsdorff, M., Howell, B.W., Goffinet, A., Mumby, M.C., Cooper, J.A., Herz, J., 1999. Direct binding of Reelin to VLDL receptor and ApoE receptor 2 induces tyrosine phosphorylation of disabled-1 and modulates tau phosphorylation. Neuron 24 (2), 481e489. Howell, B.W., Hawkes, R., Soriano, P., Cooper, J.A., 1997. Neuronal position in the developing brain is regulated by mouse disabled-I. Nature 389 (6652), 373e377. Illing, R.B., 1990. A subtype of cerebellar Golgi cells may be cholinergic. Brain Res. 522, 267e274. Jansen, J., Brodal, A., 1940. Experimental studies on the intrinsic fibers of the cerebellum. II. The cortico-nuclear projection. J. Comp. Neurol. 73, 267e321. Ji, Z., Hawkes, R., 1994. Topography of Purkinje cell compartments and mossy fiber terminal fields in lobules II and III of the rat cerebellar cortex: spinocerebellar and cuneocerebellar projections. Neuroscience 61 (4), 935e954. Ji, Z., Hawkes, R., 1995. Developing mossy fiber terminal fields in the rat cerebellar cortex may segregate because of Purkinje cell compartmentation and not competition. J. Comp. Neurol. 359 (2), 197e212. Kalinovsky, A., Boukhtouche, F., Blazeski, R., Bornmann, C., Suzuki, N., Mason, C.A., Scheiffele, P., 2011. Development of axon-target specificity of ponto-cerebellar afferents. PLoS Biol. 9 (2), e1001013. Kanichay, R.T., Silver, R.A., 2008. Synaptic and cellular properties of the feedforward inhibitory circuit within the input layer of the cerebellar cortex. J. Neurosci. 28 (36), 8955e8967. Komuro, H., Yacubova, E., Yacubova, E., Rakic, P., 2001. Mode and tempo of tangential cell migration in the cerebellar external granular layer. J. Neurosci. 21, 527e540. Kurihara, H., Hashimoto, K., Kano, M., Takayama, C., Sakimura, K., Mishina, M., Inoue, Y., Watanabe, M., 1997. Impaired parallel fiber–>Purkinje cell synapse stabilization during cerebellar development of mutant mice lacking the glutamate receptor delta2 subunit. J. Neurosci. 17 (24), 9613e9623. Kuwako, K.I., Nishimoto, Y., Kawase, S., Okano, H.J., Okano, H., 2014. Cadherin-7 regulates mossy fiber connectivity in the cerebellum. Cell Rep. 9 (1), 311e323. Lainé, J., Axelrad, H., 1994. The candelabrum cell: a new interneuron in the cerebellar cortex. J. Comp. Neurol. 339 (2), 159e173. Lainé, J., Axelrad, H., 1998. Lugaro cells target basket and stellate cells in the cerebellar cortex. Neuroreport 9 (10), 2399e2403. Lainé, J., Axelrad, H., 2002. Extending the cerebellar Lugaro cell class. Neuroscience 115 (2), 363e374. Landsend, A.S., Amiry-Moghaddam, M., Matsubara, A., Bergersen, L., Usami, S., Wenthold, R.J., Ottersen, O.P., 1997. Differential localization of delta glutamate receptors in the rat cerebellum: coexpression with AMPA receptors in parallel fiber-spines synapses and absence from climbing fiber-spine synapses. J. Neurosci. 17 (2), 834e842. Larouche, M., Hawkes, R., 2006. From clusters to stripes: the developmental origins of adult cerebellar compartmentation. Cerebellum 5 (2), 77e88. Larouche, M., Beffert, U., Herz, J., Hawkes, R., 2008. The Reelin receptors Apoer2 and Vldlr coordinate the patterning of Purkinje cell topography in the developing mouse cerebellum. PLoS One 3, e1653. https://doi.org/10.1371/journal.pone.0001653. Leclerc, N., Gravel, C., Hawkes, R., 1988. Development of parasagittal zonation in the rat cerebellar cortex: MabQll3 antigenic bands are created postnatally by the suppression of antigen expression in a subset of Purkinje cells. J. Comp. Neurol. 273, 399e420. Leto, K., Arancillo, M., Becker, E.B., Buffo, A., Chiang, C., Ding, B., Dobyns, W.B., Dusart, I., Haldipur, P., Hatten, M.E., Hoshino, M., Joyner, A.L., Kano, M., Kilpatrick, D.L., Koibuchi, N., Marino, S., Martinez, S., Millen, K.J., Millner, T.O., Miyata, T., Parmigiani, E., Schilling, K., Serkerková, G., Sillitoe, R.V., Sotelo, C., Uesaka, N., Wefers, A., Wingate, R.J., Hawkes, R., 2016. Consensus paper. Cerebellar development. Cerebellum 15 (6), 789e828. Llinás, R., Hillman, D.E., Precht, W., 1973. Neuronal circuit reorganization in mammalian agranular cerebellar cortex. J. Neurobiol. 4, 69e94. Lorenzetto, E., Caselli, L., Feng, G., Yuan, W., Nerbonne, J.M., Sanes, J.R., Buffelli, M., 2009. Genetic perturbation of postsynaptic activity regulates synapse elimination in developing cerebellum. Proc. Natl. Acad. Sci. U.S.A. 106 (38), 16475e16480.

262 PART | I Circuit development

Lugaro, E., 1894. Sulle connessioni tra gli elementi nervosi della corteccia cerebellare con considerazioni generali sul significato fisiologico dei rapporti tra gli elementi nervosi. Riv. Sper. Freniatr. Med. Leg. 20, 297e331. Mariani, J., Crepel, F., Mikoshiba, K., Changeux, J.P., Sotelo, C., 1977. Anatomical, physiological and biochemical studies of the cerebellum from reeler mutant mouse. Philos. Trans. R. Soc. Lond. B Biol. Sci. 281, 1e28. Mason, C.A., 1987. The development of cerebellar mossy fibers and climbing fibers: embryonic and postnatal features. In: King, J.S. (Ed.), New Concepts in Cerebellar Neurobiology. Alan R. Liss, New York, pp. 57e88. Mason, C.A., Gregory, E., 1984. Postnatal maturation of cerebellar mossy and climbing fibers: transient expression of dual features on single axons. J. Neurosci. 4, 1715e1735. Miale, I., Sidman, R.L., 1961. An autoradiographic analysis of histogenesis in the mouse cerebellum. Exp. Neurol. 4, 277e296. Miyata, T., Maeda, T., Lee, J.E., 1999. NeuroD is required for the differentiation of the granule cells in the cerebellum and hippocampus. Genes Dev. 3 (13), 1647e1652. Mugnaini, E., Sekerková, G., Martina, M., 2011. The unipolar brush cell a remarkable neurons finally receiving deserved attention. Brain Res. Rev. 66 (1e2), 220e245. Nakayama, H., Miyazaki, T., Kitamura, K., Hashimoto, K., Yanagawa, Y., Obata, K., Sakimura, K., Watanabe, M., Kano, M., 2012. GABAergic inhibition regulates developmental synapse elimination in the cerebellum. Neuron 74 (2), 384e396. Nishida, K., Flanagan, J.G., Nakamoto, M., 2002. Domain-specific olivocerebellar projection regulated by the EphA-ephrin-A interaction. Development 129 (24), 5647e5658. Palay, S.L., Chan-Palay, V., 1974. Cerebellar Cortex Cytology and Organization. Springer, Berlin. Paradies, M.A., Eisenman, L.M., 1993. Evidence of early topographic organization in the embryonic olivocerebellar projection: a model system for the study of pattern formation processes in the central nervous system. Dev. Dynam. 197 (2), 125e145. Patil, N., Cox, D.R., Bhat, D., Faham, M., Myers, R.M., Peterson, A.S., 1995. A potassium channel mutation in weaver mice implicates membrane excitability in granule cell differentiation. Nat. Genet. 11, 126e129. Rahimi-Balaei, M., Afsharinezhad, P., Bailey, K., Buchok, M., Yeganeh, B., Marzban, H., 2015. Embryonic stages in cerebellar afferent development. Cerebellum Ataxias. https://doi.org/10.1186/s40673-015-0026-y. Rakic, P., 1972. Extrinsic cytological determinants of baskets and stellate cells dendritic pattern in the cerebellar molecular layer. J. Comp. Neurol. 146 (3), 335e354. Redies, C., Neudert, F., Lin, J., 2011. Cadherins in cerebellar development: translation of embryonic patterning into mature functional compartmentalization. Cerebellum 10 (3), 393e408. Rossi, F., Saggiorato, C., Strata, P., 2002. Target-specific innervation of embryonic cerebellar transplants by regenerating olivocerebellar axons in the adult rat. Exp. Neurol. 173 (2), 205e212. Rouse, R.V., Sotelo, C., 1990. Grafts of dissociated cerebellar cells containing Purkinje cell precursors organize into zebrin I defined compartments. Exp. Brain Res. 82 (2), 401e407. Sanes, J.R., Lichtman, J.W., 1999. Development of the vertebrate neuromuscular junction. Annu. Rev. Neurosci. 22, 389e442. Saywell, V., Cioni, J.M., Ango, F., 2014. Developmental gene expression profile of axon guidance cues in Purkinje cells during cerebellar circuit formation. Cerebellum 13 (3), 307e317. Schilling, K., Oberdick, J., 2009. The treasury of the commons: making use of public gene expression resources to better characterize the molecular diversity of inhibitory interneurons in the cerebellar cortex. Cerebellum 8 (4), 477e489. Serapide, M.F., Pantó, M.R., Parenti, R., Zappalà, A., Cicirata, F., 2001. Multiple zonal projections of the basilar pontine nuclei to the cerebellar cortex of the rat. J. Comp. Neurol. 430, 471e484. Shinoda, Y., Sugihara, I., Wu, H.S., Sugiuchi, Y., 2000. The entire trajectory of single climbing and mossy fibers in the cerebellar nuclei and cortex. Prog. Brain Res. 124, 173e186. Simat, M., Parpan, F., Fritschy, J.M., 2007. Heterogeneity of glycinergic and gabaergic interneurons in the granular cell layer of mouse cerebellum. J. Comp. Neurol. 500 (1), 71e83. Sotelo, C., 1975. Anatomical, physiological and biochemical studies of the cerebellum from mutant mice. II. Morphological study of cerebellar cortical neurons and circuits in the weaver mouse. Brain Res. 94 (1), 19e44. Sotelo, C., 2004. Cellular and genetic regulation of the development of the cerebellar system. Prog. Neurobiol. 72 (5), 295e339. Sotelo, C., 2008. Development of “pinceaux” formations and dendritic translocation of climbing fibers during the acquisition of the balance between glutamatergic and gamma-aminobutyric acidergic inputs in developing Purkinje cells. J. Comp. Neurol. 506 (2), 240e262. Sotelo, C., Chédotal, A., 2005. Development of the olivocerebellar system: migration and formation of cerebellar maps. Prog. Brain Res. 148, 1e20. Sotelo, C., Triller, A., 1979. Fate of presynaptic afferents to Purkinje cells in the adult nervous mutant mouse: a model to study presynaptic stabilization. Brain Res. 175 (1), 11e36. Sotelo, C., Wassef, M., 1991. Cerebellar development: afferent organization and Purkinje cell heterogeneity. Philos. Trans. R. Soc. Lond. B Biol. Sci. 331 (1261), 307e313. Strasser, V., Fasching, D., Hauser, C., Mayer, H., Bock, H.H., Hiesberger, T., Herz, J., Weeber, E.J., Sweatt, J.D., Pramatarova, A., Howell, B., Schneider, W.J., Nimpf, J., 2004. Receptor clustering is involved in Reelin signaling. Mol. Cell Biol. 24, 1378e1386. Sudarov, A., Turnbull, R.K., Kim, E.J., Lebel-Potter, M., Guillemot, F., Joyner, A.L., 2011. Ascl1 genetics reveals insights into cerebellum local circuit assembly. J. Neurosci. 31 (30), 11055e11069.

Development of the neuronal circuitry of the cerebellar cortex Chapter | 11

263

Sultan, F., Bower, J.M., 1998. Quantitative Golgi study of the rat molecular layer interneurons using principal components analysis. J. Comp. Neurol. 393 (3), 353e373. Swanson, D.J., Goldowitz, D., 2011. Experimental Sey mouse chimeras reveal the developmental deficiencies of Pax6-null granule cells in the postnatal cerebellum. Dev. Biol. 351 (1), 1e12. Takeda, T., Maekawa, K., 1989. Transient direct connection of vestibular mossy fibers to the vestibulocerebellar Purkinje cells in early postnatal development of kittens. Neuroscience 32, 99e111. Telley, L., Cadihac, C., Cioni, J.M., Jahannault-Talignani, C., Huettl, R.E., Sarrailh-Faivre, C., Dayer, A., Huber, A.B., Ango, F., 2016. Dual function of NRP1 in axon guidance and subcellular target recognition. Neuron 91 (6), 1276e1291. Trommsdorff, M., Gotthardt, M., Hiesberger, T., Shelton, J., Stockinger, W., Nimpf, J., Hammer, R.E., Richardson, J.A., Herz, J., 1999. Reeler/Disabledlike disruption of neuronal migration in knockout mice lacking the VLDL receptor and ApoE receptor 2. Cell 97 (6), 689e701. Uemura, T., Lee, S.J., Yasumura, M., Takeuchi, T., Yoshida, T., Ra, M., Taguchi, R., Sakimura, K., Mishina, M., 2010. Trans-synaptic interaction of GluRdelta2 and Neurexin through Cbln1 mediates synapse formation in the cerebellum. Cell 141 (6), 1068e1079. Voogd, J., Ruigrok, T.J.H., 2004. The organization of the corticonuclear and olivocerebellar climbing fiber projections to the rat cerebellar vermis: the congruence of projection zones and the zebrin pattern. J. Neurocytol. 33 (1), 5e21. Wassef, M., Sotelo, C., 1984. Asynchrony in the expression of guanosine 30 :50 -phosphate-dependent protein kinase by clusters of Purkinje cells during the perinatal development of rat cerebellum. Neuroscience 13, 1217e1241. Wassef, M., Sotelo, C., Thomasset, M., Granholm, A.C., Leclerc, N., Rafrafi, J., Hawkes, R., 1990. Expression of compartmentation antigen zebrin I in cerebellar transplants. J. Comp. Neurol. 294 (2), 223e234. Wassef, M., Chedotal, A., Cholley, B., Thomasset, M., Heizmann, C.W., Sotelo, C., 1992. Development of the olivocerebellar projection in the rat: I. Transient biochemical compartmentation of the inferior olive. J. Comp. Neurol. 323, 519e536. Watanabe, M., Kano, M., 2011. Climbing fiber synapse elimination in cerebellar Purkinje cells. Eur. J. Neurosci. 34 (10), 1697e1710. Zagrebelsky, M., Strata, P., Hawkes, R., Rossi, F., 1997. Reestablishment of the olivocerebellar projection map by compensatory transcommissural reinnervation following unilateral transection of the inferior cerebellar peduncle in the newborn rat. J. Comp. Neurol. 379 (2), 283e299.

Chapter 12

Introduction to cognitive development from a neuroscience perspective Helen Tager-Flusberg Department of Psychological & Brain Sciences, Boston University, Boston, MA, United States

Chapter outline 12.1. Introduction 12.2. Frameworks and methods 12.2.1. Conceptual frameworks 12.2.2. Eye-tracking 12.2.3. Electrophysiology

267 267 267 268 268

12.2.4. MRI and other imaging methods 12.2.5. Summary 12.3. Overview References

269 269 270 271

12.1 Introduction Newborns depend for their survival on caregivers, almost always mothers, who provide them with food, warmth, and comfort. They come into this world equipped with a set of reflexes, basic sensory capacities and preferences, and primitive means for signaling their needs. Over the first 2 years of life infants undergo rapid developments in cognition and behavior and their brains undergo exponential growth and change. During these early years infants are transformed into toddlers capable of goal-directed actions and independent mobility; they develop sophisticated concepts of the world around them, acquire the basic use and understanding of language, and become equipped to navigate their social environment with ease. Nevertheless, at the age of two toddlers still have a great deal to learn and developmental changes continue, at a steadier rate, over the next two decades. In this chapter, we follow this remarkable journey in human development with particular emphasis on the early years. Studies of child development began almost a century ago; however, our conceptual frameworks and knowledge about what children know as they grow have shifted over the years, driven largely by the introduction of new methods and technology that are now opening up the possibility of asking key questions about the neural and cognitive mechanisms that drive developmental change. This chapter serves as a roadmap to the basic frameworks and methods that have guided the field of child development over the past few decades.

12.2 Frameworks and methods 12.2.1 Conceptual frameworks The scientific field of cognitive development began with the seminal work of Jean Piaget, though its intellectual roots go back to philosophical questions about the origins of knowledge and mind raised by Plato, Aristotle, and others in ancient Greece. Piaget, who was a biologist by training, was the first to develop a comprehensive theoretical framework for how children acquire core concepts that are the foundation of human thought. His “constructivist” theory claimed that fundamental concepts, such as space, time, and causality, are acquired by the baby operating on the environment. Through very basic processes, complex abstract concepts are built up through actions and interactions with objects and people. In Piaget’s theory, children develop through a series of qualitatively distinct stages, each defined by a new form of mental

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00012-3 Copyright © 2020 Elsevier Inc. All rights reserved.

267

268 PART | II Cognitive development

representation. He was committed to a strong view on the biological contributions to cognitive development by arguing for the importance of evolution, maturation, and adaptation for understanding change over time; however, his empirical work exclusively focused on observable behavior, which clearly contributed to the significant role he attributed to action, especially as the foundation of representational knowledge (Flavell, 1996). Piaget’s work lies between the extremes of nativism and empiricism. He began writing in the early part of the 20th century, but his work did not become well known in the English-speaking scientific community until the 1970s, when psychologists became more open to ideas that were grounded in a more complex view of how biology intersects with experience. Nevertheless, even though Piaget firmly argued that developmental changes took place in children’s cognition, he was not able to articulate how such changes were accomplished; instead, he focused more on generating the important questions to be addressed by developmental scientists that remain central to the field today. Over the last few decades developmental science has blossomed, as it has taken on questions about what the starting point is for newborns, what the child knows at different ages, how this knowledge is organized and changes over time, and which factors and basic learning mechanisms contribute to these changes. These new questions are being addressed using new tools that provide a more direct window into the minds and brains of babies and young children than simply observing their behavior. While developmental scientists are informed by parallel research on other species, the unique social and linguistic abilities of humans transform the child’s perceptions of the world in ways that require different frameworks and approaches for investigating developmental processes (cf. Vygostky, 1978).

12.2.2 Eye-tracking For Piaget, infants’ actions on the world revealed their underlying knowledge; however, motor development is a slow process in humans and thus provides a very indirect assessment of their cognitive capacities. Beginning with the seminal work of Robert Fantz (1958), researchers have been using eye gaze patterns to provide a more direct approach to what infants see, discriminate, prefer, remember, and anticipate. Because they have relatively good ocular-motor control, babies’ eye movements are a reliable and valid way of revealing mental processing. Studies using eye gaze patterns, including measures of first fixations, time spent looking at images, and anticipatory looking patterns, have demonstrated that even newborns are not simply a bundle of reflexes with basic sensory capacities. Instead, it is clear that infants have perceptual preferences that are biased toward attending to particular events and entities in their environment and rapid changes over the first few months of life consolidate and expand on these initial biases. For many years, researchers relied either on manual on-line coding of eye gaze patterns or video-taping of infants’ looking patterns and later laborious coding of their eye movements by hand. It is the most common behavioral method in use today for studies on infant perception and cognitive, social, and language development (Aslin, 2007). The advent of automated eye-trackers, particularly ones that do not require head mounted cameras or complex calibration procedures, led to changes in our ability to capture the microstructure of eye gaze patterns in infants, including much finer temporal and spatial resolution, without requiring manual coding that has the potential for introducing error or bias (Perez et al., 2018). Eye-trackers can also be used to extract other ocular measures including pupil dilation, which has been used as an index of task difficulty and mental effort and eyeblink rate, which correlates with dopamine levels, and is interpreted as an index of learning and goal-directed behavior (Eckstein et al., 2017). Taken together, these various measures that are captured by eye-trackers have significantly enhanced research on the early development of learning, cognitive processes, and plasticity.

12.2.3 Electrophysiology If eye-tracking is the method of choice for investigating the minds of babies, then electrophysiology has been the method of choice for investigating their brain function. The broader field of cognitive neuroscience has relied on a range of technologies to probe brain structure and function in people; however, not all of them are easily adapted for use with infants and young children. Electrophysiological recordings, including electroencephalography (EEG) and event-related potentials (ERP), provide good measures of temporal processing and dynamics of cognitive processes. The primary advantages for their use in studying infants and young children include safety, ease of use, and tolerance of some movement. The EEG signals collected from electrodes placed over the scalp reflect ongoing electrical activity in the brain; ERPs are signals that occur in response to a specific stimulus and are most widely used as a neural assessment for a range of cognitive processes. Spontaneous or time-locked EEG activity can be decomposed into constituent frequencies to quantify power in specific bandwidths, each of which reflects different aspects of neural processing related to cognition. ERPs are thought to detect postsynaptic potentials from pyramidal cells summed over a large number of neurons. Invariant stimulusrelated electrical activity is extracted through an averaging process over a large number of trials in order to reduce the noise due to random components. The average ERP is then analyzed temporally, as a series of positive and negative components

Introduction to cognitive development from a neuroscience perspective Chapter | 12

269

characterized by their polarity, peak latency, amplitude, and distribution over the scalp (Csibra et al., 2008). The introduction of high-density arrays, which include a large number of electrodes embedded in a net that is easily fitted over the scalp and well tolerated by babies, provide more complete spatial coverage and have led to new methods of analysis with better spatial and temporal resolution. For example, recent advances have been made in methods for identifying source localization, providing more information about the neural origins of the EEG signal, and for computing functional connectivity (e.g., Xie et al., 2018). Different components in the ERP signal have been linked to specific cognitive processes including, for example, attention, face processing, memory, and a range of linguistic processes from speech to syntax. Much of what is known about the cognitive processes associated with specific ERP components comes from studies of adults. There are significant developmental changes in the latency, morphology, and topography in the known ERP components, and while the precise roots of these changes are not yet fully understood, they are thought to reflect a combination of cognitive advances and more efficient neural processing. While electrophysiological methods are widely used now in human developmental neuroscience, they do have some limitations including their limited spatial resolution and lack of sensitivity to processing in subcortical brain areas.

12.2.4 MRI and other imaging methods Cognitive neuroscience has relied most extensively on magnetic resonance imaging (MRI) as the single most effective, noninvasive, and safe method for examining in vivo human brain structure and function. MRI provides high-resolution spatial information and its application in developmental science has provided detailed information about volumetric changes in regional gray and white matter. The advent of diffusion tensor imaging (DTI), a variant of conventional MRI, and novel methods of analyses have led to advances in the ability to identify and characterize developmental changes in white matter pathways (Deoni et al., 2011; Dubois et al., 2014). Functional aspects of brain development have been tracked using functional MRI (fMRI), which measures changes in the blood oxygenation levels (BOLD response), an indirect measure of increases in regional neuronal activity. fMRI provides excellent spatial resolution when local differences in the BOLD response are analyzed between tasks or groups. Advances in analytic methods such as connectivity analyses have led to an increase in studies tracking developmental changes in systems level cortical representations (Aslin et al., 2015). Despite their importance in developmental cognitive neuroscience, there are several challenges in the use of MRI and associated technologies with infants and young children, not least of which is the requirement to lie completely still in a noisy enclosed tunnel for relatively long time periods while the scan data are collected (Power et al., 2012). This requires a good deal of cooperation from an awake child, which is possible in children who have been carefully prepared, or the use of alternative approaches such as scanning during sleep or sedation (cf. Denisova, 2019). New methods for handling motion in scanning sequences and analyses continue to be developed which makes the use of MRI more feasible with young children (e.g., Siegel et al., 2014; Zaitsev et al., 2017). While we are still learning how developmental changes in brain morphology and metabolism may influence the BOLD response, significant progress has been made in fMRI research on young infants and children (e.g., Deen et al., 2017; Zhang et al., 2019). Other technologies that have been introduced to investigate functional brain development include magnetoencephalography (MEG), which has parallels to EEG and MRI, and functional Near Infra-Red Spectroscopy (fNIRS), which, like fMRI, measures changes in hemodynamic responses that are assumed to be related to regional brain activity. Advances in MEG technology have begun to make their way into studies of young infants and children (for review, see Kao and Zhang, 2019), though there are still significant challenges similar to those described for MRI. For fNIRS, optical probes placed on the scalp measure changes in blood oxygenation levels that reflect surface cortical activity, offering moderately good spatial resolution, depending on the number and location of the probes. Considerable methodological and technological advances have been made in recent years and the number of studies using NIRS to investigate functional activity in the brains of very young infants is increasing each year (cf. Bulgarelli et al., 2018). It has been used to probe many aspects of cognitive development and is well suited to studying in vivo social interaction, which makes this a promising tool for investigating, for example, parent-child interaction (McDonald and Perdue, 2018). However, it does not detect neural responses generated in structures that lie below the cortical surface which limits its use in studying key aspects of memory, emotion, and social perception.

12.2.5 Summary Significant and rapid advances in developmental cognitive neuroscience have been made based on the introduction of new conceptual frameworks and methods for probing cognition and brain processes. We are now able to link behavioral and

270 PART | II Cognitive development

brain changes in ways that allow us to test specific theoretically grounded hypotheses about the neural bases of cognitive development. Yet, progress can only be made if our methods and technologies are used in the context of well-designed experiments and an appreciation of the limitations in the application and interpretation of findings from each available method. As noted, challenges remain in using all the methods surveyed here with pediatric populations: they all require a considerable amount of cooperation and minimal movement. While young infants and older children can tolerate the requirements for most of these methodologies, toddlers and preschoolers are far more active than infants and less compliant than school-aged children and therefore relatively less is known about development during these critical years. Similar challenges are faced by researchers studying children with neurodevelopmental disorders. It is expected that technological innovations in the coming years will help to fill in these gaps to provide a more complete picture of cognitive development from birth through adolescence in both typical and atypical development.

12.3 Overview This chapter provides comprehensive and detailed coverage of the current state of research on cognitive development including both behavioral and neuroscience perspectives in the field. The field is still young and studies that attempt to address the developmental relationship between cognitive and neural processes have really only begun in earnest during the past decade. Mark Johnson discusses the theories that have dominated the newly emerging field of developmental cognitive neuroscience. He highlights the importance of having theoretically driven research and evaluates the predictions and evidence from each using research from different cognitive domains on both typical and atypical children. He concludes that the theory of interactive specialization, which has its roots in Piagetian theory, offers the most promise of a framework that captures biological developmental change that is grounded in experience. Bodison, Colby, and Sowell provide a detailed summary of what is currently known about postnatal structural changes in the brain based primarily on studies using MRI and DTI. They describe the very different trajectories in graey and white matter development, emphasizing the significance of developmental timing and key genetic and experience-dependent factors driving the dynamic processes of brain development that continue through late adolescence. Thompson, Aguero, and Lany take on the central question about how infants learn; in particular, how they can acquire abstract and complex cognitive structure based on inputs from the environment. They focus on infants’ capacities to extract different types of statistical regularities from the perceptual information that are central to the formation of linguistic and object categories. Evidence from studies conducted over the past decade suggests that infants actively use probability distributions, sequential structure, correlations and association in auditory and visual inputs beginning early in the first year of life. These active learning mechanisms, which are presumably basic capacities that evolve via experience, are the building blocks that drive cognitive development. For humans, vision is arguably the most important perceptual system that drives conceptual development, particularly for growth in understanding objects and events. It is also the system that is most well studied and understood across different species. Scott Johnson discusses how infants come to experience a world of stable objects beginning with limited but organized vision at birth that undergoes rapid functional changes driven by brain maturation coupled with learning from experience and manual explorations during the first year of life. More advanced aspects of visuospatial abilities are taken up by Stiles, Akshoomof, and Haist, who discuss the development of ventral and dorsal neurocognitive processes including global and local pattern perception, spatial construction, localization attention, and manipulation. Their study describes these processes, particularly the vulnerability of the dorsal stream in visuospatial development in children who have experienced brain injury or with neurogenetic syndromes. Human memory serves a range of important functions that are carried out by different systems. Bauer and Dugan describe the major forms of memory each of which follows its distinct developmental trajectory. They summarize studies that demonstrate the emergence of both declarative and nondeclarative memory systems during the first 2 years of life, which is far earlier than had previously been thought. At the same time, they argue that some aspects of declarative memory have a more protracted course not reaching maturity until middle childhood, paralleling what is known about the development of the underlying neurobiology of memory processing systems as demonstrated by studies of ERPs in children. Our earliest autobiographical memories are sparse and usually for events that took place during the preschool years. One reason that has been offered for this phenomenon is that encoding and retrieving these memories depend on language. By the time most children are 3 years old, they have mastered the fundamentals of language from sounds to words to grammar which allows them to form narrative memories of their lives. Tager-Flusberg and Finch describe the development of language covering the major milestones as well as how language intersects with aspects of motor, conceptual, and social cognitive development. Much is known about the importance of left hemisphere frontal/temporal cortical systems in adult language processing. These left lateralized systems emerge during the first year of life in typically developing children and

Introduction to cognitive development from a neuroscience perspective Chapter | 12

271

continue to mature through middle childhood. Failure to lateralize language functions to the left hemisphere during this early sensitive period appears to be one hallmark finding among children with language disorders. Social-emotional development has become a very active area of research in developmental science, in part because of the complexity and richness of human social lives, in part because of the important consequences when development in this broad domain is impaired in a range of genetic syndromes, early adversity, or forms of psychopathology. Bayat and Nelson focus on the development of face processing skills. The authors describe the initial rapid developments that take place during the first year of life in the foundational cognitive processes for identifying faces and facial expressions that depend on dedicated neural architecture in the fusiform region of the temporal cortex. At the same time, these advances are followed by a more protracted period of development during which time there are both volumetric changes in the fusiform “face area” as well as temporal and morphological changes in the signature ERP signal that is elicited by faces, the N170. The development of the neural systems underlying more complex aspects of social cognition is covered by Richardson and Saxe. In particular, they focus on the development of children’s theory of mind: the ability to reason about people’s actions based on their more abstract mental states such as thoughts or beliefs. While the classic studies in this area concluded that theory of mind emerges at about the age of 4 based on behavioral and ERP studies that used a task to evaluate a child’s understanding of false beliefs, later eye tracking studies suggest that this understanding may already be in place in at least an implicit form by around 18 months. At the same time, fMRI studies indicate that the brain region that is crucially involved in theory of mind processing, the right temporal-parietal junction, continues to show functional developmental changes through middle childhood. Decety and Michalska focus more on the affective components of social behavior: the development of the capacity to respond to another person’s distress, or empathy. They summarize the normal course of development in the psychological processes and neurobiological mechanisms that drive empathic behavior, and then how these might go awry in children with conduct disorder or related forms of psychopathy. Rueda and Conejero focus on the development from early infancy through middle childhood of the complex set of networks that are involved in executive attentional processes and self-regulation. An exciting recent advance in this area is research that finds associations between common genetic variants with individual differences in specific attentional components; in turn, these differences influence environmental experiences of children as mediated, for example, by parenting behaviors. Buzzell, Lahat, and Fox explore the development of two aspects of cognitive control that play important roles in decision making and social behavior: inhibitory control and self-monitoring. The neural substrates for these advanced cognitive systems as indexed by fMRI and ERP measures depend on areas in prefrontal cortex that do not reach the end point of development until late adolescence. Hughes provides a summary of research on classic executive function measures in both typical and atypical children. She argues that because the executive functions and their neural substrates, which encompass multiple brain regions, develop incrementally over the entire period of childhood and adolescence, they are more susceptible to environmental influences. She describes examples of such influences including studies of training, parent-child interaction, and clinical populations. Norona, Doom, Davis, and Gunnar focus on the effects of stress, drawing heavily on what is known from animal models to investigate whether the findings from that body of literature can be extended to current work on prenatal and postnatal stress responses in human development. Beltz, Kelly, and Berenbaum take up the issue of sex differences in development, which has important implications for understanding many neurodevelopmental disorders that differentially affect males and females for reasons that are still not well understood. The cognitive neuroscience of human development has become a very active area of research. The exponential growth of this field, as evidenced by the numbers of papers, journals, and books published over the past decade, has largely been driven by methodological advances that allow us to view more directly the minds and brains of babies and children and observe how they change over time. As these methods develop further, it is expected that greater progress will be made by asking broader questions about the significance of the developmental trajectories seen within and across cognitive domains, the extent and constraints that operate on individual variation and developmental plasticity, and the precise ways in which biological and nonbiological factors influence the developmental course of brain and cognitive development.

References Aslin, R., 2007. What’s in a look? Dev. Sci. 10, 48e53. Aslin, R., Shukla, M., Emberson, L., 2015. Hemodynamic correlates of cognition in human infants. Annu. Rev. Psychol. 66, 349e379. Bulgarelli, C., Blasi, A., Arridge, S., et al., 2018. Dynamic causal modelling on infant fNIRS data: a validation study on a simultaneously recorded fNIRS-fMRI dataset. Neuroimage 175, 413e424. Csibra, G., Kushnerenko, E., Grossmann, T., 2008. Electrophysiological methods in studying infant cognitive development. In: Nelson, C.A., Luciana, M. (Eds.), Handbook of Developmental Cognitive Neuroscience, second ed. MIT Press, Cambridge, MA, pp. 247e262.

272 PART | II Cognitive development

Deen, B., Richardson, H., Dilks, D., et al., 2017. Organization of high-level visual cortex in human infants. Nat. Commun. https://doi.org/10.1038/ ncomms13995. Denisova, K., 2019. Neurobiology, not artifacts: challenges and guidelines for imaging the high risk infant. Neuroimage 185, 624e640. Deoni, S., Mercure, E., Blasi, A., et al., 2011. Mapping infant brain myelination with magnetic resonance imaging. J. Neurosci. 31, 784e791. Dubois, J., Dehaene-Lambertz, G., Kulikova, S., Poupon, C., Huppi, P., Hertz-Pannier, L., 2014. The early development of brain white matter: a review of imaging studies in fetuses, newborns and infants. Neuroscience 276, 48e71. Eckstein, M., Guerra-Carrillo, B., Miller Singley, A., Bunge, S., 2017. Beyond eye gaze: what else can eyetracking reveal about cognition and cognitive development? Dev. Cognit. Neurosci. 25, 69e91. Fantz, R.L., 1958. Pattern vision in young infants. Psychol. Rec. 8, 43e47. Flavell, J., 1996. Piaget’s legacy. Psychol. Sci. 7, 200e203. Kao, C., Zhang, Y., 2019. Magnetic source imaging and infant MEG: current trends and technical advances. Brain Sci. https://doi.org/10.3390/ brainscie9080181. McDonald, N., Perdue, K., 2018. The infant brain in the social world: moving toward interactive social neuroscience with functional near-infrared spectroscopy. Neurosci. Biobehav. Rev. 87, 38e49. Perez, D., Radkowska, A., Raczaszek-Leonardi, J., Tomalska, P., The Talby Study Team, 2018. Beyond fixation durations: recurrence quantification analysis reveals spatiotemporal dynamics of infant visual scanning. J. Vis. 18. https://doi.org/10.1167/18.13.5. Power, J., Barnes, K., Snyder, A., Schlaggar, B., Peterson, S., 2012. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage 59, 2142e2154. Siegel, J., Power, J., Dubis, J., Vogel, A., Church, J., Schlaggar, B., Peterson, S., 2014. Statistical improvements in functional magnetic resonance imaging analyses produced by censoring high-motion data points. Hum. Brain Mapp. 35, 1981e1996. Vygostky, L., 1978. Mind in Society. Harvard University Press, Cambridge, MA. Xie, W., Mallin, B., Richards, J., 2018. Development of infant sustained attention and its relation to EEG oscillations: an EEG and cortical source analysis study. Dev. Sci. 21. https://doi.org/10.1111/desc.12562. Zaitsev, M., Akin, B., LeVan, P., Knowles, B., 2017. Prospective motion correction in functional MRI. Neuroimaging 154, 33e42. Zhang, H., Shen, D., Lin, W., 2019. Resting-state functional MRI studies on infant brains: a decade of gap-filling efforts. Neuroimage 185, 664e684.

Chapter 13

Theories in developmental cognitive neuroscience Mark H. Johnson Department of Psychology, University of Cambridge, Cambridge, United Kingdom

Chapter outline 13.1. Introduction 273 13.2. Why do we need theories? 274 13.3. Frameworks for understanding human functional brain development 275 13.3.1. Maturational viewpoint 276 13.3.2. Interactive specialization 276 13.3.3. Skill learning 276 13.4. Assumptions underlying the three frameworks 277 13.4.1. Deterministic vs. probabilistic epigenesis 277 13.4.2. Static vs. dynamic mapping 277 13.4.3. Plasticity 278

13.5. 13.6. 13.7. 13.8.

Predictions and evidence Functional brain imaging Critical or sensitive periods Atypical development: from genetics to behavior in developmental cognitive neuroscience 13.9. Interactive specialization: future challenges 13.10. Summary, conclusions, and recommendations Acknowledgments References

278 279 281 282 283 285 286 286

13.1 Introduction Over the past two decades a new field of science has emerged at the interface between developmental neuroscience and developmental psychology. This field has become known as developmental cognitive neuroscience (DCN) (Johnson and de Haan, 2015). The exciting mix of these previously separate fields has led to a burgeoning of new observations, methods, and ideas. However, a characteristic shared with most newly emerging interdisciplinary areas in the biological sciences is the feeling that the knowledge acquired to date is rather fragmentary, with many intuitively surprising observations remaining unexplained. How are we to come to understand these phenomena, and to interpret and explain them within a broader context of other findings with different methods and populations? To do this we need to develop theories, and, I will argue, these theories need to be of a particular kind to be useful in progressing knowledge. As a scientific discipline, DCN sits at the convergence of two of the oldest philosophical and scientific debates of mankind. The first of these questions concerns the relation between mind and body, and specifically between the physical substrate of the brain and the mental processes it supports. This issue is fundamental to the scientific discipline of cognitive neuroscience. The second debate concerns the origin of organized biological structures, such as the highly complex structure of the adult human brain. This issue is fundamental to the study of development. Light can be shed on these two fundamental issues by tackling them both simultaneously, and specifically by focusing on the relation between the postnatal development of the human brain and the cognitive processes it supports. The second of the two debates above: that of the origins of organized biological structure can be posed in terms of phylogeny or ontogeny. The phylogenetic (evolutionary) version of this question concerns the origin of the characteristics of species, and has been addressed by Charles Darwin and many others since. The ontogenetic version of this question concerns individual development within a life span. The ontogeny question has been somewhat neglected relative to phylogeny, since some influential scientists have held the view that once a particular set of genes has been selected by

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00013-5 Copyright © 2020 Elsevier Inc. All rights reserved.

273

274 PART | II Cognitive development

evolution, ontogeny is simply a process of executing the instructions coded for by those genes. By this view, the ontogenetic question essentially reduces to phylogeny (a position sometimes termed “nativism”). In contrast to this view, many now agree that ontogenetic development is an active process through which biological structure is constructed afresh in each individual by means of complex and variable interactions between genes and their environments. By this view, the information is not in the genes, but emerges from the constructive interaction between genes and their environment (see also Oyama, 2000). What actually is ontogenetic development? Many introductory biology textbooks define development in terms of an increasing restriction of fate. This refers to the basic observation that as the biological development of an individual organism (ontogeny) proceeds, the range of options for further specification or specialization available to the organism at that stage decreases. Structural or functional specialization is an end state in which there are few or no options left to the organism. From this perspective, plasticity can be defined as a developmental stage in which there are still options available for alternative developmental pathways (Thomas and Johnson, 2008). Another dimension of ontogenetic development is that it involves the construction of increasingly complex hierarchical levels of biological organization, including the brain and the cognitive processes it supports. As we will see later in this chapter, organizational processes at one level, such as cellular interactions, can establish new emergent functions at a higher level, such as those associated with overall brain structure. This characteristic of ontogeny means that a full picture of developmental change requires different levels of analysis to be investigated simultaneously. The developmentalist, I argue, needs to go beyond statements such as a psychological change being due to maturation, and actually provide an account of the processes causing the change at cellular and molecular levels. Thus, in contrast to most other areas of psychology and cognitive science, a complete account of developmental change specifically requires an interdisciplinary approach. Given the above considerations, it is perhaps surprising that only recently has there been renewed interest in examining relations between brain and cognitive development. Although the field of developmental psychology was originally founded by biologists (such as Darwin and Piaget), biological approaches to human behavioral development fell out of favor at the end of the last century for a variety of reasons, including the widely held belief among cognitive psychologists in that period that the software of the mind is best studied without reference to the hardware of the brain (see later for the details of this argument). However, the recent explosion of basic knowledge on mammalian brain development makes the task of relating brain to behavioral changes considerably more viable than previously. In parallel, molecular and cellular methods, along with theories of self-organizing dynamic networks, have led to great advances in our understanding of how vertebrate brains are constructed during ontogeny. These advances, along with those in noninvasive structural and functional neuroimaging, have contributed to the recent emergence of DCN.

13.2 Why do we need theories? Some might argue that, as a subfield of biology, investigators in DCN should proceed with their empirical investigations of human development unbiased by any perspective or theory, or simply wait until enough data is in before speculating on its significance. However, even a cursory read of philosophy of science tells you that this view is, at best, somewhat naïve. While Victorian naturalists simply made descriptive observations about the animals and plants they studied, when we move to a modern science led by experiments the rules change. The experiments we choose to conduct (out of the many millions of potential experiments we could do) are inevitably guided by implicit assumptions and biases. For example, many of the earliest functional MRI experiments with children professed to be neutral and exploratory, while actually assuming that we would see new brain regions becoming activated (coming “on-line”) as development proceeds. As we will see later, in many cases this basic starting assumption was violated. In addition, theories in science do not just attempt to explain sets of data posthoc, but they should actually generate predictions and, ideally, direct whole lines of empirical investigation. The need for specific testable theory is greatly enhanced in the current attempts to improve scientific practice following the “replication crisis” in psychology and neuroscience. In this climate, preregistration of a data analysis plan which lays out precisely how specific hypotheses will be tested is becoming the norm. The clarity of hypotheses and corresponding data analyses requires a greater emphasis on the theories that generate the hypotheses. Having emphasized the importance of theories, we must also be sure to achieve an appropriate balance with data gathered from a variety of different sources and methods. In one of the parent disciplines of DCNdcognitive developmentdseveral grand theoretical castles have been built on the shaky sands of just one particular type of behavioral test (e.g., habituation to a repeated visual event). In most of the biological sciences, confidence in an observation or conclusion is greatly increased by seeking multiple different sources of converging evidence. Interdisciplinary fields such as DCN face a formidable challenge in the development of adequate theories since scientists are required to develop theories that not only cross different levels of observation (such as genetic, neural, and

Theories in developmental cognitive neuroscience Chapter | 13

275

behavioral), but that also relate those different levels together in some way. Ideally, DCN theories should relate evidence from different levels of observation in terms of one level of explanation. As mentioned earlier, for several decades in the field of cognitive development it was generally considered inappropriate to attempt to relate different levels of explanation. Rather, the aim was to explain one level of observation (change in behavior) in terms of one level of explanation (cognitive). This widespread view was taken for a variety of reasons, but one influential source was the work of David Marr (1982). Marr argued that, because the same computation can, in principle, be implemented on different computer or neural architectures, a computational account of cognition could, and should, be constructed independently of the details of its implementation on hardware. This influential argument led to the view that considering the role of the brain in cognitive development was reductionist in the sense that molecular and cellular processes could never provide an adequate explanation of cognitive processes. While the case against a simple reductionist view is clearly correct, as stated earlier theories of biological development critically need to explain the reverse process (to reductionism) of the emergence of higher order structures of organization. Thus, while not denying Marr’s antireductionist point, constructing a specific type of neural computer hardware will constrain the range of possible computations that could be supported. With these considerations in mind, Mareschal et al. (2007), among others, argue that there are constraints on computation imposed by its detailed implementation. Further, when attempting to bridge levels of explanation, mechanistic accounts of processes of computation and developmental change should be consistent across different levels, i.e., there is a need for isomorphism between levels of description. Since the goal of DCN is to relate the genetic, neural, cognitive, and behavioral accounts of human development, devising theories that relate the different levels of observation seems crucial. Mareschal et al. (2007) describe it thus: We would argue that strong pragmatic considerations mean that what is achievable in real time at one level of description strongly constrains the appropriate theories of what is going on at other levels. What is more easily achievable in the brain . .will strongly constrain the character of cognition. More specifically, the brain can implement much more readily certain representational states and transformations more than others. It is these primitives that the researcher should initially use to construct theories of cognition. (p. 209)

Theories come in different shapes and sizes. Specifically, the amount and range of DCN data accounted for can independently vary along at least two dimensions: (a) how domain-specific or domain-general (domain here is used in a general sense to refer to an aspect of cognition) a theory is, and (b) how many levels of explanation/observation are incorporated or integrated. It is a defining feature of DCN, as opposed to traditional cognitive development, that multiple levels of observation are considered and related in terms of a single process or causal mechanism. One reason for this is that the parent discipline of cognitive development had been built on the strategy of explaining changes in behavior during development in terms of cognitionda level of explanation that is not itself directly observable. While the scientific strategy of theorizing at a level that is not directly observable is not unique to cognitive psychology, constraining theories of this kind by only one level of explanation is high-risk due to the lack of constraints it imposes. In other words, a very wide variety of theories can successfully account for data at one observable level only. I propose that a better strategy is to sandwich a nonobservable level of explanation (such as cognition) between two levels that are directly observable, such as those of brain and behavior. Theories in DCN could potentially vary enormously in the scope of data that they account for, from a single cognitive domain in a single population, to an account that crosses domains of cognition and populations (typical and atypical development). Often, in biology the broader the scope of a theory, the less clearly it makes detailed domain-specific predictions. Thus, some have referred to such broad-scope theories as frameworks (Kuhn, 1996). Put simply, frameworks are ways of thinking about, or viewpoints on, a large body of data. Frameworks may have some testable elements, but primarily serve as a coherent set of assumptions that, taken together, offer an account of a wide range of phenomena. In addition, within a framework more specific and detailed theories (and thence hypotheses) can be constructed. Further, they guide lines of research and the kinds of hypotheses that are explored. In a young and newly emerged field, I suggest the first priority should be to develop appropriate and useful frameworks, since adopting the wrong framework could be an expensive diversion in terms of both time and money.

13.3 Frameworks for understanding human functional brain development A review of the literature reveals the three different frameworks on human postnatal functional brain development that provide a background and motivation for studies in the field (Johnson and de Haan, 2015).

276 PART | II Cognitive development

13.3.1 Maturational viewpoint According to the maturational viewpoint, newly emerging sensory, motor, and cognitive functions during human postnatal development are related to the maturation of particular areas of the brain, usually regions of cerebral cortex. Much of the early research attempting to relate brain to behavioral development in humans took this approach. Evidence concerning the differential neuroanatomical development of brain regions was used to estimate the age when a particular brain region was likely to show adult functionality. Success in a new behavioral task or domain of cognition at a given age was then attributed to the maturation of relevant brain regions. Functional brain development is in this sense depicted as the reverse of adult neuropsychological studies of patients with brain damage, with specific brain regions being added-in during development (with the converse effects from being deleted by damage). Despite the intuitive appeal and attractive simplicity of the maturational approach, it failed to successfully explain many observations on human functional brain development. For example, evidence discussed later shows that some of the regions that are latest to develop by neuroanatomical criteria can be activated shortly after birth and clearly support cognitive functions well before they would be considered anatomically mature. Thus, the emergence of new behaviors is not usefully linked to the onset of mature functioning, but rather we need an explanation of how partial or immature functioning of regions influences behavior. Further, where functional activity has been assessed by fMRI during a behavioral transition, multiple cortical and subcortical areas appear to change their response pattern (e.g., Luna et al., 2001; Supekar et al., 2009) rather than a few specific areas becoming active. Another difficulty for the maturational viewpoint is that associations between neural and cognitive changes based on age of onset are theoretically weak due to the great variety of neuroanatomical and neurochemical measures that change at different times in different regions of the brain. Thus, as the brain is continuously developing until the teenage years, it is nearly always possible to find a potential neural correlate for any behavioral change in development.

13.3.2 Interactive specialization In contrast to the maturational viewpoint, the interactive specialization framework assumes that postnatal functional brain development, at least within cerebral cortex, involves a process of organizing patterns of interregional interactions to tune functional specificity within a region (Johnson, 2011). According to this view, the response properties of a specific cortical region are partly determined by its patterns of connectivity to other regions, and their patterns of activity. During postnatal development changes in the response properties of cortical regions occur as they interact and compete with each other to acquire their role in new computational abilities. From this perspective, some cortical regions may begin with poorly defined broad functions, and consequently are partially activated in a wide range of different stimuli and task contexts. During development, activity-dependent interactions between regions sharpens up the functions and response properties of cortical regions such that their activity becomes restricted to a narrower set of circumstances (e.g., a region originally activated by a wide variety of visual objects may come to confine its response to upright human faces). The onset of new behavioral competencies during infancy will therefore be associated with changes in activity over several different regions, and not just by the maturation of one or more additional region(s).

13.3.3 Skill learning A third perspective on human functional brain development, skill learning, involves the proposal that the brain regions active in infants during the onset of new perceptual or motor abilities are similar, or even identical to, those involved in complex skill acquisition in adults. For example, Gauthier and colleagues have shown that extensive training of adults to identify individual artificial objects (called “Greebles”) eventually results in activation of a cortical region previously preferentially activated by faces, the fusiform face area (Gauthier et al., 1999). This indicates that this region is normally activated by faces in adults, not because it is prespecified to do so, but due to our extensive expertise with that class of stimulus. Extended to development, this view would argue that development of face processing during infancy and childhood could proceed in a similar manner to acquisition of perceptual expertise for a novel visual category in adults (see Gauthier and Nelson, 2001). While the degree to which parallels can be drawn between adult expertise and infant development remains unclear to the extent that the skill learning hypothesis is correct, it presents a clear view of a continuity of mechanisms of learning and plasticity throughout the life span. Recent variants of “Bayesian” computational models in which predictions become finetuned by learning have begun to be applied to human cognitive development. However, currently it is only in atypical development that links between these models and brain systems have been made (e.g., Lawson et al., 2015).

Theories in developmental cognitive neuroscience Chapter | 13

277

13.4 Assumptions underlying the three frameworks Underlying these frameworks are differing sets of key assumptions.

13.4.1 Deterministic vs. probabilistic epigenesis Gottlieb (1992) distinguished between two approaches to the study of development: deterministic epigenesis in which it is assumed that there is a unidirectional causal path from genes to structural brain changes and then to psychological function; and probabilistic epigenesis in which interactions between genes, structural brain changes, and psychological function are viewed as bidirectional, dynamic, and emergent. In many ways it is a defining feature of the maturational approach that it assumes deterministic epigenesis; region-specific gene expression is assumed to effect changes in intraregional connectivity that, in turn, allows new functions to emerge. A related assumption commonly made within the maturational approach is that there is a one-to-one mapping between brain and cortical regions and particular cognitive functions, such that specific computations come online following that maturation of circuitry intrinsic to the corresponding cortical region. In some respects, this view parallels mosaic development at the cellular level in which simple organisms (such as Caenorhabditis elegans) are constructed through cell lineages that are largely independent of each other (see Elman et al., 1996, for discussion). Similarly, different cortical regions are assumed to have different maturational timetables, thus enabling new cognitive functions to emerge at different ages. In contrast to the maturational approach, Interactive Specialization (IS) (Johnson, 2011) has a number of different underlying assumptions. Specifically, a probabilistic epigenesis assumption is coupled with the view that cognitive functions are the emergent product of interactions between different brain regions, rather than the product of a single region. In this respect, IS follows trends in adult functional neuroimaging. For example, Friston and Price (2001) point out that the error in assuming that particular functions can be localized within individual cortical regions. Rather, they argue, the response properties of a region are determined by its patterns of connectivity to other regions, as well as by those other regions’ current activity states. By this view, “the cortical infrastructure supporting a single function may involve many specialized areas whose union is mediated by the functional integration among them” (p.276). Similarly, in discussing the design and interpretation of adult functional MRI studies, Carpenter and collaborators have argued that: “In contrast to a localist assumption of a one-to-one mapping between cortical regions and cognitive operations, an alternative view is that cognitive task performance is subserved by large-scale cortical networks that consist of spatially separate computational components, each with its own set of relative specializations, that collaborate extensively to accomplish cognitive functions” (2001, p. 360). Extending these ideas to development, the IS framework emphasizes changes in interregional connectivity, as opposed to the maturation of intraregional connectivity. While the maturational approach may be analogous to mosaic cellular development, the IS view corresponds to the regulatory development seen in higher organisms in which cell-cell interactions are critical in determining developmental fate. While mosaic development can be faster than regulatory development, the latter has several advantages. Namely, regulatory development is more flexible and better able to respond to damage, and it is more efficient in terms of genetic coding. In regulatory development genes do not code directly, but need only orchestrate cellular-level interactions to yield more complex structures (see Elman et al., 1996).

13.4.2 Static vs. dynamic mapping As well as the mapping between structure and function at one age, we can also consider how this mapping might change during development. When discussing functional imaging of developmental disorders, many laboratories have assumed that the relation between brain structure and cognitive function is unchanging during development (Johnson et al., 2002; Karmiloff-Smith, 2002). This assumption is partly why it is sometimes considered acceptable to study developmental disorders in adulthood and to then extrapolate back in time to early development. Contrary to this view, the IS approach suggests that when a new computation or skill is acquired, there is a reorganization of the functional connectivity between different brain structures and regions. This reorganization process could even change how previously acquired cognitive functions are represented in the brain. Thus, the same behavior could potentially be supported by different neural substrates at different ages during development. Stating that structure-function relations can change with development is all very well, but it lacks the specificity required to make all but the most general predictions. Fortunately, the view that there is competitive specialization of regions during development gives rise to expectations about the types of changes in structure-function relations that should be observed. Specifically, as regions become increasingly selective in their response properties during development, patterns of cortical activation during behavioral tasks may therefore be more extensive than those observed in adults, and

278 PART | II Cognitive development

involve different patterns of activation. Additionally, within broad constraints, successful behavior in the same tasks can be supported by different patterns of cortical activation in infants and adults. Evidence in support of this view will be discussed later. The basic assumption underlying the skill learning approach is that there is a continuity of the circuitry underlying skill acquisition from birth through to adulthood (see Poldrack, 2002, and Ungerleider et al., 2002, for review of the neural systems involved in perceptual and motor skill learning in adults). These circuits are likely to involve a network of structures that retains the same basic function across developmental time (a static brain-cognition mapping). However, other brain regions may respond to learning with dynamic changes in functionality similar or identical to those hypothesized within the IS framework. For example, neuroimaging studies of adults acquiring the skill of mirror reading show both increases and decreases of cortical activity over widespread regions during learning: unskilled performance is associated with activation in bilateral occipital, parietal, and temporal lobes and cerebellum, with acquisition of skill leading to decreases in bilateral occipital and right parietal activation and to increases in inferior temporal lobe and caudate nucleus activation (Poldrack, 2002). According to the skill learning view, similar dynamic changes in brain activation would occur as skills emerge during development.

13.4.3 Plasticity Another way in which the three frameworks differ is with regard to the mechanisms of, and assumptions about, plasticity. Plasticity in brain development is a phenomenon that has generated much controversy, with several different conceptions and definitions having been presented. According to the maturational framework, plasticity is a specialized mechanism that is activated following brain injury. According to the IS approach, plasticity is simply the state of having a region or system’s function not yet fully specialized. That is, there is still remaining scope for developing more finely tuned responses, and a reduction in plasticity is simply a byproduct of the processes of development. As mentioned earlier, this definition corresponds well with the view of developmental biologists that development involves the increasing restriction of fate. Finally, according to the skill learning hypothesis view, the functional plasticity present in early development share characteristics and processes in common with the plasticity underlying acquisition and retention of skills in adults (Karni and Bertini, 1997). Thus, unlike the IS approach, plasticity does not necessarily reduce during development. To summarize, the maturational view is characterized by the interrelated assumptions that (1) cortical areas are a mosaic of regions with independent developmental timetables, (2) deterministic epigenesis means that inherent structural development in a region causes or allows functional changes, (3) there is fixed regional structure/function mapping, and (4) plasticity involves specialized mechanisms triggered by injury. In contrast, the Interactive Specialization approach is based on the assumptions that (1) cortical areas are inextricably linked through dense patterns of interconnections that contribute to coordinated sequences of development, (2) probabilistic epigenesis gives a vital role to intrinsically and extrinsically generated activity in sculpting anatomical development, (3) combinations of cortical regions may support similar or identical behaviors in different ways during the course of development, and (4) plasticity is the inherent state of an unspecialized neural system. The third perspective, Skill Learning, is based on the assumption that (1) specific cortical areas or networks are specialized for perceptual and motor skill acquisition from early on, (2) the acquisition of these skills shapes the response functions of the same or other cortical regions, (3) dynamic changes occur in the neural substrate of a skill as the brain becomes expert, and (4) plasticity is retained at fairly constant level throughout development.

13.5 Predictions and evidence Frameworks are useful for a variety of reasons, but particularly so when they help to generate predictions that direct research (albeit that they will not always make opposing predictions), and when they offer coherent explanations of previously puzzling observations. In the sections that follow, I will overview three of the sources of evidence that an adequate account of human functional brain development will need to address. Before this, however, we review the types of predictions that arise from the three frameworks outlined above. The first set of predictions concerns the neural correlates of the onset of new cognitive abilities during human postnatal development. According to a maturational view, new behavioral abilities are mediated by new computational units that underlie cognition that are, in turn, enabled or allowed by the structural anatomical development of one or more brain regions. A consequence of this is an expected general increase in the number of brain structures and pathways that will be activated in response to given task demands with development. In contrast to these predictions, in the IS approach we anticipate patterns of functional brain activation consistent with the tuning up of a subset of a network of regions for processing particular stimulus in relation to task demands. Turning to the skill learning approach, here we anticipate that

Theories in developmental cognitive neuroscience Chapter | 13

279

the onset of new abilities will often be associated with activation of a particular fixed network of skill acquisition areas. We would expect the same network to be activated at the acquisition of new skills, but as the new behavioral ability is acquired a different network of brain regions may become involved. With regard to the issue of how cognitive functions are mapped on to patterns of brain activity, according to the maturational approach if we compare two age groups in a task for which they show identical behavioral performance, we should also expect to see identical patterns of brain activation (fixed mapping). This is not necessarily the case for the IS approach, since the exact patterns of brain activation that support a function will change according to the degree of specialization of component regions within their supporting interactive network. Indeed, the IS approach predicts that the patterns of regional brain activation supporting a function will change during development, both in terms of localization and specialization. In this context, localization refers to the extent to which a given function is associated with a region or area of cortex. Specifically, the extent (quantity) of cortex activated following the presentation of a given task or perceptual stimulus may change during development. Specialization refers to the degree of specificity of function of a given region or area of cortex. Functions may be finely tuned, such as an area that is activated only by a restricted category of visual objects or under a very narrow range of task circumstances, or broadly tuned, in that they are activated under a wide range of circumstances. According to the IS framework, the issues of localization and specialization are two sides of the same coin, and are both consequences of the same common underlying mechanisms. The skill learning view invokes the reactivation of one or more skill -learning circuits at the onset of a task, followed by a different pattern of activation after the skill is acquired. In this case, in many comparisons between age groups, the younger group will have acquired the skill in less depth than the older group, giving rise to different patterns of underlying brain activation. Interestingly, however, patterns of changing brain activation while adults acquire new skills should mirror the changes seen during development as infants and children acquire simpler skills.

13.6 Functional brain imaging Elsewhere I have reviewed in some detail evidence from functional imaging pertaining to the development of face perception and social cognition (Cohen-Kadosh and Johnson, 2007; Johnson, 2011). In this section evidence is reviewed from the functional neuroimaging of typical development relating to the three perspectives on functional brain development described above. I suggest that the evidence currently available does not offer much support to the maturational view, at least not without substantial modification. Instead, behavioral change in development often seems to be accompanied by large-scale dynamic changes in the interactions between regions. In other words, networks of regions compete and adapt, driving increased regional specialization during development. It is worth noting that the different assumptions underlying the frameworks discussed above has influenced the types of experiments that are conducted and the way that they are subsequently analyzed. For example, when a maturational approach is taken and expectation is that a particular area will become additionally active during development for a particular task, then brain imaging analysis may be focused on particular regions in detail rather than on the whole brain and network dynamics. As a consequence, any possible changes in more distant brain structures would not be detected. By contrast, if one adopts the interactive specialization approach, then the importance of whole-brain imaging and network connectivity is clearly apparent. A number of authors have described developmental changes in the patterns of cortical activation associated with improvements in behavioral and cognitive skills during postnatal life (see Johnson, 2011, for review). Some of these studies used event-related potentials, indicating both for word learning (Neville, 1991) and face processing (de Haan et al., 2002), there is increasing localization of processing with age and experience of a stimulus class. That is, electrophysiological recordings reveal a wider area of processing for words or faces in younger infants than in older ones whose processing has become more specialized and localized. From the interactive specialization framework, such developmental changes are accounted for in terms of more pathways being partially activated in younger infants prior to experience with a class of stimuli. With increasing experience, the specialization of one or more of those pathways occurs over time. Taking the example of word recognition, ERP activity differentiating between comprehended and noncomprehended words is initially found over widespread cortical areas. This narrows to left temporal leads after children’s vocabularies have reached a certain level, irrespective of maturational age (Mills et al., 1993). Changes in the extent of localization can be viewed as a direct consequence of specialization. Initially multiple pathways are activated for most stimuli. With increasing experience, fewer pathways become activated by each specific class of stimulus. Pathways become tuned to specific functions and are therefore no longer engaged by the broad range of stimuli, as was the case earlier in development. Additionally, there may be inhibition from pathways that are becoming increasingly specialized for that function. In this

280 PART | II Cognitive development

sense, then, there is competition between pathways to recruit functions, with the pathway best suited for the function (by virtue of its initial biases) usually winning out. Further evidence to support the interactive specialization view comes from functional MRI studies in children. We illustrate this with the example of a specialized subnetwork implicated in face processing: the core face network. These studies can potentially inform us about changes in the degree of localization and functional specialization of the cortical face network. All of the studies conducted to date show that certain regions of the cortex show reliable activation to faces from at least mid-childhood, and most studies report dynamic changes in the extent of cortical tissue activated between children and adults (including activation of additional areas in children that are not typically found in the mature adult brain). Some studies also provided evidence for increasing functional specialization (degree of face specificity) with age or experience. For example, Scherf et al. (2007) used naturalistic movies of faces, objects, buildings, and navigation scenes in a passive viewing task with children (5e8 years), adolescents (11e14 years), and adults. They found that the children exhibited similar patterns of activation of the face processing areas commonly reported in adults (such as the fusiform face areadFFA). However, this activation was not selective for the category of face stimuli; the regions were equally strongly activated by objects and landscapes. In a similar study, Golarai et al. (2007) tested children (7e11 years), adolescents (12e16 years), and adults with static object categories (faces, objects, places, and scrambled abstract patterns). They found substantially larger right FFA and left parahippocampal volumes of selective activation to faces in adults than in children. The developmental changes observed in these fMRI studies thus provide strong support for the gradual emergence of cortical tissue specifically tuned functions within the cerebral cortex (Johnson, 2011). While experiments such as these provide evidence for the increasing specialization of individual regions of cortex during human postnatal development, it is clear that the next step is to understand how networks of different regions, each with their own different specializations, emerge. As mentioned above, face processing is a good test domain as a “core face network” of cortical regions has been well established in adults, and activity in this network is modulated by task demands in adults (Cohen Kadosh et al., 2010). A Region-of-Interest (ROI) analysis of the core face network (anterior fusiform gyrus [FG], inferior occipital gyrus [OG], and the superior temporal sulcus [STS]) in adults revealed consistent patterns of coactivation suggesting that these regions are part of a common processing network that responds flexibly in the context of different processing strategies or stimulus changes. Cohen Kadosh et al. (2010) examined the emergence of the network underlying face processing in younger (7e8 years old) and older (10e11 years old) school-age children as well as young adults, and found that children showed substantially weaker functional connectivity (coactivation) within the face network. More notably, no evidence was found for the influence of task-demands on the effective connectivity within the network in the two children groups. Thus, while both child groups exhibited similar overall network structures, these weaker networks were not influenced by top-down task demands. In terms of the IS framework this would be interpreted as an initial network specialization to a stimulus class, then becoming extended to stimulus and task demand combinations. Another cortical area that has been intensively studied in terms of its potential emerging specialization is the “visual word form area” (VWFA). The visual word form area is a region in the left occipito-temporal cortex, centered on the midfusiform gyrus that preferentially responds to visually presented words in literate adults. Like the FFA, the VWFA appears to be involved in perceptual expertise, but in this case allowing words to be perceived and processed quickly and automatically in skilled readers. Important changes have been observed in activation of the VWFA over the years when children begin to learn to read, and are thus acquiring familiarity and expertise with words. Functional MRI studies have shown that the VWFA is typically bilateral in beginning readers, but shifts to the mature pattern of left-lateralization with age and increasing reading skill (Schlaggar et al., 2002). A prospective longitudinal fMRI study charted individual changes in cortical sensitivity to written words over a four-year period as a group of 7e12-year-olds acquired reading skills. There were age-related changes in children’s cortical sensitivity to word visibility in posterior left occipito-temporal sulcus around the anatomical location of the VWFA. The rate of increase in brain word sensitivity correlated with the rate of improvement in sight word reading efficiency (Ben-Shachar et al., 2011) suggests further that the evidence for neural specialization is associated with functional consequences, in this case increasing perceptual and cognitive skill. Another of the domains in which processes of cortical specialization have been studied in some detail is number cognition. Work with children and adults using neuroimaging indicates that the intraparietal cortical regions are activated during numerical processing. However, the pattern of cortical activation associated with numerical processing shows developmental changes with a general shift with development of reduced frontal and greater parietal activation resulting from basic and advanced numerical tasks (Ansari, 2008). For example, 8e12-year-olds differ from adults in the degree of recruitment of parietal and prefrontal networks while solving simple arithmetic tasks. Adults typically demonstrate activation primarily in the intraparietal sulcus (IPS), whereas children show significantly lower levels of activity in the IPS network but greater activation in frontal regions, areas responsible for attention and executive working memory (Kucian et al., 2005). Thus, evidence supports the view that functional specialization in the parietal cortex for mental arithmetic

Theories in developmental cognitive neuroscience Chapter | 13

281

increases with age, and is accompanied by a corresponding decrease of activity in prefrontal regions (Rivera et al., 2005; Ansari et al., 2005). These changes have been taken as evidence for Interactive Specialization of parietal functioning during ontogeny (Ansari and Dhital, 2006; Holloway and Ansari, 2010). The decreasing involvement of prefrontal areas, on the other hand, is assumed to reflect a developmental disengagement of domain-general processes related to executive control and working memory (Ansari et al., 2005; Rivera et al., 2005), a view consistent with elements of the skill learning framework. In sum, current evidence across several domains of cognition suggests that (1) new behavioral skills are accompanied by widespread changes across many regions of cortex, and (2) functional brain development involves the twin processes of increasing localization and increasing specialization.

13.7 Critical or sensitive periods The three perspectives on human functional brain development differ in their views as to the effects of early atypical experience. By the maturational view, differences in experience might influence the speed at which a function matures or the ultimate level of performance; in the skill learning view any atypical early experiences can potentially be compensated for later in development as the mechanisms for learning remain in operation in the same way; in the IS view atypical early experiences could have long-lasting effects because they could affect the specialization and localization of function, which may not be able to be altered later in life when there is less scope for plasticity. There have only been a limited number of studies examining the effects of atypical early experience in humans. Some studies have investigated the perception of facial information in children who experienced deprivation of patterned visual input in the early months of life due to bilateral, congenital cataracts. These patients were tested years after their cataracts were removed and they were fitted with contact lenses (i.e., years after visual input had been restored); thus, any effects of the few months of deprivation following birth would likely be absent or very minimal according to the maturation or skill learning views. However, investigation of these patients reveals persistent deficits in selective aspects of face processing. One study found that patients showed impairments in matching facial identity over changes in viewpoint (and tended to show an impairment in recognizing identity over changes in emotional expression), but performed normally on tests of lipreading, perception of eye gaze, and matching of emotional expressions (Geldart et al., 2002). A second study demonstrated that this difficulty in processing facial identity may be due to deficits in processing the spacing among facial features, since patients performed normally in discrimination of faces that differed only on individual features (e.g., mouth) but they were impaired in discrimination of faces that differ only in the spacing of the features (Le Grand et al., 2001). This was not due to a general impairment in perception of spacing of features, as they performed normally in discriminating nonface patterns, whether they differed by the shape of the features or the spacing of the features. The fact that these impairments persisted even after years of visual input to compensate for the early deprivation is not consistent with the maturational or skill learning views, but is consistent with the IS view that early atypical experience may have long-lasting consequences. The three perspectives we have discussed yield different types of predictions about the consequences of perinatal brain damage. According to the maturational view additional mechanisms of plasticity are activated following early damage. Specific additional explanations are then required to account for incidents of recovery of function. Also, it is not obvious why the extent of plasticity is greater earlier in life. From the IS perspective, there is a parsimonious explanation of recovery of function following perinatal damage, since the regional specialization of the remaining brain regions will be altered to compensate, particularly the corresponding regions in the other hemisphere. In cases of bilateral or extensive damage, recovery is less likely. From the skill learning perspective plasticity is a lifelong activity of the brain. Damage to the general circuits critical for skill acquisition will have long-lasting and widespread consequences, while damage to circuits specific to acquisition of particular skills or their retention may result in more isolated impairments. Of the three approaches, IS arguably gives the most parsimonious account of sensitive periods for plasticity, since plasticity is reduced when specialization of corresponding regions is achieved. In sum, with regard to the long-term effects of atypical early experience, or even variations of experience within the normal range, once again the three frameworks lead to different sets of expectations. From the skill-learning perspective variations in early experience will determine the extent of skills acquired. Early deprivation will be potentially reversible since the same mechanisms of skill acquisition are available later in life. From the IS perspective, long-term effects of atypical early experience can result from atypical patterns of regional specialization arising early in life. Such atypical patterns of specialization may be difficult to reverse once established. Finally, under the maturational view a primary variable influenced by the environment is the speed of maturation that may affect the level or maintenance of a skill. It is sometimes argued that early sensory deprivation may have a general slowing effect on the sequence of maturation.

282 PART | II Cognitive development

13.8 Atypical development: from genetics to behavior in developmental cognitive neuroscience We have seen in earlier sections that interactive specialization is a promising framework for understanding human functional brain development. While this level of explanation may be appropriate for characterizing the proximal causes of atypical cognition or behavior in developmental disorders such as autism or Williams syndrome, when development goes awry a satisfactory explanation of the disorder also requires an account of the causal mechanisms that initiate the atypical development trajectory (distal causes). This usually entails hypotheses about how atypicality at one level (such as genetics) causes or induces the onset of atypical development at other levels (brain, cognitive, behavior). An initial attempt to provide a framework for unraveling these complex causal pathways in developmental disorders came with the Causal Modeling approach of Morton and Frith (1995). According to its originators, Causal Modeling provided a theory-neutral system for modeling different theories about the paths of causation from a biological level to cognitive and behavioral levels. The models were represented in a graphical notation that allowed for easy comparison between competing explanations, but the framework itself did not involve the construction of computational or neural network models (although a theory represented in the notation could potentially be implemented this way). Causal Modeling is a useful way to compare different theories where these theories are based on the assumption of predetermined epigenesis (a one-way causal pathway from genetics to behavior), but it is a less natural format for capturing the complexities of probabilistic epigenesis (Gottlieb, 2007) in which cause can also run in reverse, e.g., sensory experience or internal states such as stress are known to effect gene expression profiles. In addition, Causal Modeling is intended as a notation for comparing different theories, and as such does not provide an explanation or generate testable hypotheses itself. A different perspective on identifying causal factors in development involves using implemented computational models that can capture complex nonlinear interactions that are hard to conceptualize with just schematic illustrations or verbal descriptions. Here the assumption is that multiple factors can interact in complex ways to determine outcome. In one of the several initial attempts to apply neural networks to developmental disorders, Oliver and colleagues charted the different ways in which the “normal” formation of structured representations in the cerebral cortex can go wrong (Oliver et al., 2000). These models were originally designed to investigate the mechanisms underlying the interactive specialization process discussed earlier. Several groups have used simple cortical matrix models to investigate the factors and mechanisms responsible for cortical specialization (e.g., Kerszberg et al., 1992; Shrager and Johnson, 1995; Oliver et al., 2000). In these artificial neural networks, connections between nodes are pruned according to variations of Hebbian learning: links between nodes that are often active together are strengthened, whereas links between nodes that are not often coactive get weaker and are pruned. In some of these models, the degree of pruning of connections during learning approximately matches that seen during the course of human brain development. During exposure to patterned input (roughly equivalent to sensory stimulation), nodes become more selective in their response properties, and under certain conditions clusters of nodes with similar response properties emerge. Thus, in these computational models selective pruning plays a role in the emergence of clusters of nodes (localization) that share common specific response properties (specialization). In order to explore developmental disorders, Oliver and colleagues made simulations with a simple cortical matrix model in which one or other of the parameters known to be important for the emergence of structured representations was deliberately changed. In this case, when the authors manipulated an aspect of the intrinsic structure of the network, the relative length of excitatory and inhibitory links, this initial state change totally disrupted the formation of structured representations. In other simulations, structured representations emerged, but were distorted in different ways relative to the typical development case. Oliver et al. (2000) aimed to generate a taxonomy of the ways that structured cortical representations could go awry in development, with the long-term aim that some of these artificial developmental disorders could potentially map onto those that occur in the real world. A criticism of the Oliver et al. (2000) approach, and several related models, is that while they could potentially provide an explanation at the level of brain structure and function, they do not capture the genetic, cognitive, and behavioral levels. More recent models have begun the ambitious task of simulating from the genetic to the behavioral levels, albeit within a restricted domain of cognition (past tense acquisition) (Thomas et al., 2013), and even simulating how given artificial neural networks behave within virtual environments. The application of dynamic neural field models in which the tuning up (specialization) of banks of neurons is simulated, combined with neuroimaging and behavioral analyses of infants and young children in executive function tasks, is a promising new avenue for more directly linking brain to cognition to understand developmental change within the IS framework (Perone et al., 2019).

Theories in developmental cognitive neuroscience Chapter | 13

283

13.9 Interactive specialization: future challenges To recapitulate on the basis of the IS view, it argues that early in postnatal development many areas begin with poorly defined functions, and consequently can be partially activated by a wide variety of sensory inputs and tasks. During development, activity-dependent interactions between regions result in modifications of the intraregional connectivity such that the activity of a given area becomes restricted to a narrower range of circumstances. As a result of becoming more finely tuned, small-scale functional areas become increasingly distinct from their surrounding cortical tissue, and this will be evident in functional imaging studies as increasing localization of function. In summary, according to the IS view, small-scale areas of cortex become tuned for certain functions as a result of a combination of factors, including (i) the suitability or otherwise of the biases within the large-scale region (e.g., transmitter types and levels, synaptic density, etc.), (ii) the information within the sensory inputs (sometimes partly determined by other brain systems), and (iii) competitive interactions with neighboring regions (so that functions are not duplicated). To date, majority of the research on the emergence of specialized functions in human cortex has focused on specific regions. However, it is clear from the IS viewpoint that the next step is to understand how networks involving different regions, each with their own different specializations, emerge. In other words, while we are beginning to understand functional brain development at the level of individual cortical regions, we are still in the dark about how the larger scale of cortical function in terms of networks of regions develops (Johnson and Munakata, 2005). In this section I review some initial evidence and theory that may begin to address this intriguing issue. Before considering the empirical evidence, we need to consider what makes a network of functional nodes more or less successful. A branch of mathematics called graph theory concerns itself with the relative efficiency of different kinds of networks. While it may seem at first that a lattice or grid pattern is the optimal design for a network, formal analysis of measures of local network connectivity and the average path length from one node to another show that so-called small world networks are the most efficient (Bassett and Bullmore, 2006). In contrast to the grid pattern of streets found in many American cities, small world networks are more like the clusters of small streets in a village that is then linked to other such villages by fast highways. Although the overall balance of the small local streets and highways can vary, most biological systems (and even the World Wide Web) are small world networks. Several studies have shown the regional interconnectivity of the adult brain is a highly efficient small world network, but how does this efficient network emerge? The first piece of the jigsaw comes from recent work by Fair et al. (2007, 2009) who used functional connectivity analyses in fMRI to study resting state control networks in school age children and adults. Their analysis allows them to infer the nature and strength of functional connections between 39 different cortical regions. They found that development entailed both segregation (i.e., decreased short range connectivity) and integration (i.e., increased long range connectivity) of brain regions that contribute to a network. In a similar study, the general developmental transition from more local connectivity to greater and stronger long-range network connectivity was confirmed using slightly different methods and 90 different cortical and subcortical regions (Supekar et al., 2009). The decrease in short-range interregional functional connectivity is readily explicable in terms of the IS view. As neighboring regions of cortical tissue become increasingly specialized for different functions (e.g., objects vs. faces), they will less commonly be coactivated. This process may also involve synaptic pruning and, as we heard in the last section, has been simulated in neural network models of cortex in which nodes with similar response properties cluster together spatially distinct from nodes with other response properties (Oliver et al., 1996). Thus, decreasing functional connectivity between neighboring areas of cortex is readily predicted by models implementing the IS view. More challenging from the current perspective is to account for the increase in long-range functional connections. A maturational explanation of the increase in long-range functional connectivity would suggest that this increase is due to the establishment or strengthening of the relevant fiber bundles. However, the increase in functional connectivity during development may occur after the relevant long-range fiber bundles are in place (see Fair et al., 2009 and Supekar et al., 2009 for discussion). While increased myelination is likely to be a contributory factor, (1) myelination itself can be a product of the activity/usage of a connection (Markham and Greenough, 2004) and (2) a general increase in myelin does not in itself account for the specificity of interregional activity into functional networks that support particular computations (but see Nagy et al., 2004). Thus, the strengthening and maintenance of long-range brain connections is likely to also be an activity-dependent aspect of brain development. This raises the question of why and how do particular anatomically distant brain regions begin to cooperate in a functional network? A key to answering this question may lie in scaling up the basic mechanisms of Hebbian learning. Instead of cells that fire together wire together we are seeing regions that tend to be coactivated in a given task context strengthening or

284 PART | II Cognitive development

maintaining the neural pathways between them. While each region is becoming individually specialized for a particular function, this intraregion change in tuning is modulated and influenced by its presence within an emerging network of coactivated structures. For example, in a task that requires visually guided action, a variety of visual and motor areas will be coactivated along with multimodal integration areas. If the task is repeated sufficiently often, then these patterns of coactivation will be strengthened, and specialization of individual regions will proceed within this context of overall patterns of activation. A second source of coactivation in the developing human brain is commonly overlookeddspontaneous activity during the resting state (with no task demands). Although there has been great interest in the resting state or default network in adults, only recently has this been studied using fMRI in children (although there is a long history of studying resting EEG in children). It seems likely that the oscillatory resting activity of the brain, which probably occupies more waking hours than those when the child is engaged in any specific tasks, may play a key role in strengthening and pruning the basic architecture of long-range connections. A third reason why anatomically distant regions may strengthen and maintain their connectivity relates to the fact that most of the long-range functional connections studied by Fair et al. (2007) involved links to parts of the prefrontal cortex. This part of the cortex is generally considered to have a special role during development in childhood and skill acquisition in adults (Thatcher, 1992; Gilbert and Sigman, 2007). Indeed, PFC may play a role in orchestrating the collective functional organization of other cortical regions during development. While there are several neural network models of PFC functioning in adults (e.g., O’Reilly, 2006), few if any of these have addressed development. However, another class of model intended to simulate aspects of development may be relevant both to PFC and to the issue of how networks of specialized regions come to coordinate their activity to support cognition. Knowledge-based cascade correlation (KBCC; Shultz et al., 2007) involves an algorithm and architecture that recruits previously learned functional networks when required during learning. Computationally, this dynamic neural network architecture has a number of advantages over other learning systems. Put simply, it can learn many tasks faster, or learn tasks that other networks cannot, because it can recruit the knowledge (computational abilities) of other self-contained networks as and when required. In a sense, it selects from a library of available computational systems to orchestrate the best combination for the learning problem at hand. While this class of model is not intended to be a detailed model of brain circuits (Shultz and Rivest, 2001; Shultz et al., 2007), it has been used to characterize frontal systems (Thivierge et al., 2005) and may capture important elements of the emerging interactions between PFC and other cortical regions at an abstract level. In addition, it offers initially attractive accounts of (1) why PFC is required for the acquisition of new skills, (2) why PFC is active from early development, but also shows prolonged developmental change, and (3) why early damage to PFC can have widespread effects over many domains. Although much work remains to be done to understand in more detail the factors that lead to the emergence of longrange networks, the graph theory analyses of changes during the school-age years are generating important insights. While, as described earlier, there are differences in the balance of short and long connections between children and adults, it is important to note that the network organization of children’s brains is as efficient as that of adults. In other words, while children’s brains are wired differently from those of adults, they are still optimally geared for the rapid and high-fidelity transmission of information. Whether the same is true in infancy and early childhood remains unknown. Aside from the shift from local to long-range connectivity, another change in network structure observed using graph theory analysis during development is in the hierarchical structure. Adult networks have a more hierarchical structure that is optimally connected to support top-down relations between one part of the network and another (Supekar et al., 2009). While hierarchical networks have a number of computational advantages to be discussed below, they are known to be less plastic and more vulnerable to damage or noise in the particular nodes at the top of the hierarchy. Thus, the network arrangement of children may be more flexible and plastic in response to unusual or atypical sensory input or environmental context. Further, the response to focal brain damage, particularly in the prefrontal cortex, may be more clearly understood in the light of these different network structures. One of the features of a hierarchical network is the capacity for one region to feedback highly processed sensory or motor input to the earlier stages of processing. In much the same way as we hypothesized that lateral interregional interactions help shape the intrinsic connectivity of areas to result in functional specialization, interactions between regions connected by feedback and feed-forward connections may also help shape the specialization of the areas involved. Top-down effects play an important role in sensory information processing in the adult brain (e.g., Siegel et al., 2000). For example, during perception, information propagates through the visual processing hierarchy from primary sensory areas to higher cortical regions, while feedback connections convey information in the reverse direction. In a neurocomputational model of feedback in visual processing in the adult brain, Spratling and Johnson (2004) demonstrated that a number of different phenomena associated with visual attention, figure/ground segmentation, and contextual cueing could all be

Theories in developmental cognitive neuroscience Chapter | 13

285

accounted for by a common mechanism underlying cortical feedback. Extending these ideas to development, there are potentially two important implications of feedback that will benefit from future exploration. The first of these will be to examine how the specialization of early sensory areas is shaped by top-down feedback, and vice versa, during development. The second topic for investigation will be to examine the consequences of relatively poor or diffuse cortical feedback in the immature cortex. Top-down feedback from PFC may also have a direct role in shaping the functional response properties of posterior cortical areas. In cellular recording studies from both humans and animals, evidence has accrued that the selectivity of response of neurons in areas such as the fusiform cortex may increase in real-time following the presentation of a stimulus. For example, McCarthy et al. (1997) measured local field potentials in face-selective regions of lateral fusiform cortex in human adults and found that responses of these neurons go from being face-selective at around 200 ms after stimulus presentation, to being face identity or emotion-selective at later temporal windows. This suggests that top-down cortical feedback pathways, in addition to their importance in attention and object processing (Spratling and Johnson, 2004, 2006), may increase the degree of specialization and localization in real-time, as well as in developmental time. Thus, some of the changes in functional specialization and localization seen in face-sensitive regions may reflect the increasing influence of interregional coordination with other regions, including the PFC. A final aspect of the transition from child brain network to the adult one is the greater connectivity between cortical and subcortical structures seen at younger ages (Supekar et al., 2009). This observation may be fundamental for our understanding of the emergence of the social brain and memory systems as it implies that the specialization of some cortical areas may be initially more dominated by structures such as the amygdala and hippocampus. As we approach adulthood, more networks become intrinsic to the cortex and develop a complex hierarchical structure more dominated by PFC.

13.10 Summary, conclusions, and recommendations Now that developmental cognitive neuroscience (DCN) has become established as an interdisciplinary field in its own right, it is time to evaluate and question the directions we are going in. One of the most common criticisms leveled at the newly emerging field is that it is primarily being driven forward by the powerful new methods for imaging brain structure and function in an infant and child-friendly way (as well as new techniques for genetic analyses), and that it lacks the theory-driven approach that characterizes much of the best work in cognitive development. Similar concerns are expressed, albeit less directly, by students who can be daunted by the somewhat fragmentary islands of data we have acquired to date about human functional brain development. Where is the overarching theory or framework within which they can make sense of disparate observations? A related concern sometimes expressed by those in cognitive science is that the hypotheses, which are presented in DCN, are reductionist, or otherwise impoverished as a cognitive explanation of infant or child behavior. In other words, this criticism is that what hypotheses and theories there are in the field are of the wrong type, and do not offer a satisfactory explanation of behavioral change in development. Starting with the criticism of a relative lack of theories in DCN, we have to acknowledge that, at least compared to the parent discipline of cognitive development, work in DCN has been generally less theory-driven (albeit with the exceptions discussed in this chapter). Why is this? A large part of the explanation I believe to be is due to the sudden increase in the volume and diversity of data available due to the new methods that have become available. Many theories that successfully accounted for sets of behavior observations in child development founder on the rocks when we also try account for neuroscience data relating to the same behavioral tasks. For one thing, when one more than doubles the quantity of data to be accounted for, then many previously successful theories will no longer offer a satisfactory explanation, simply because the chance of observing refuting evidence is much higher. Bringing powerful new methods into a field is analogous to a catastrophic environmental change during evolutiondthe majority of species (theories) simply cannot adapt and therefore die off. It takes generations for the better-adapted species to emerge. A second issue is that of accommodating new types of data. When you begin to study brain function directly, the first thing that hits you is the complexity of the processes involved. For example, neuroscience evidence indicates that the brain has at least three partially independent routes for executing eye movements. While these routes may have slightly different attributes, duplication of computations and (apparent) redundancy seems to be a basic feature of how the brain does things. Thus, at a sweep, simple single-route cognitive models appear less plausible. Add to this the complexity of feedback routes interacting with sensory-driven information, and the undoubted importance of temporal synchrony, and many existing theories of cognitive development look hopelessly simplistic. Of course, a common reaction to this is that theories of cognitive development are not intended to account for neuroscience datadthat is merely a matter of implementation. However, if you accept this argument I contend that you are not doing cognitive neuroscience (and I would argue that satisfactory explanations of development necessitate bridging between levels of observationdsee Johnson & de Haan (2015).

286 PART | II Cognitive development

This leads us to the second common criticism of theory in developmental science; the theories are of the wrong type to be of relevance for explaining the development of human behavior. Commonly, the view is expressed that theories in DCN are reductionist and therefore do not offer good explanations of cognitive change. As discussed earlier, it has been argued following Marr (1982), that cognition is a level of explanation independent from the underlying neuroscience. Recent directions in neuroscience suggest that, to the contrary, there is a large degree of interdependence between levels in real complex biological systems such as the brain. This has led to the proposal that we should be seeking theories that are consistent between different levels of explanation (see Mareschal et al., 2007, for a detailed discussion of this point). Ultimately, theories that are consistent with both behavioral and brain development evidence will have greater explanatory power than those confined to one level of observation. In considering the issues above, the current dearth of plausible theories in DCN seems unsurprising. After all, new fields in the biological sciences (in contrast to some physical sciences) often go through a natural history phase in which collection of basic data is the priority. However, in this chapter I have argued that we need to strive to bring more adequate and appropriate theories into the field. Thus, I offer three positive suggestions for hallmarks of a good theory in developmental cognitive neuroscience. 1. The theory advanced should genuinely relate neural observations to behavioral ones, and can be equally well tested (and refuted) by either neural or behavioral level observations. I suspect that a variety of different types of theories will emerge to serve this bridging function, but that they are unlikely to look like many existing cognitive development theories. Theories that have been developed purely on the basis of behavioral data are unlikely to naturally map on to brain imaging data, and there is a danger in seeking only confirmatory data. Ideally, we should develop theories of functional brain development that are equally compatible with brain and behavioral observations. 2. Theories in developmental science should involve mechanisms of change. This suggestion is not new (e.g., Mareschal and Thomas, 2007), but it is still surprisingly common to see theories that explain the state of affairs before and after a developmental transition, but that do not specify the mechanisms of the transition itself (other than using the terms such as maturation or learning). Theories of development need to be theories focused on change. 3. Given that theories in DCN are accounting for several levels of observation, and that they also need be compatible with undoubtedly complex and dynamic aspects of neural processing, we need to find ways to elucidate and present those theories so that they are both comprehensible and clarifying. This is the attraction and importance of formal computational modeling, be it symbolic, connectionist, or hybrid (see Marechal et al., 2007). While theories may initially develop as informal ideas, ultimately we should aim to implement them as computational models. Finally, I caution against being too prescriptive. In the long term it is probably good for the field to have a heterogeneous mix of different types of theories and let the data, and time, select those with the best fit to reality. After all, despite their prolonged domination, the dinosaurs did not inherit the globe.

Acknowledgments I thank several members of the Center for Brain & Cognitive Development, Birkbeck, for their comments on the initial version of this chapter. Financial support was provided by the UK Medical Research Council.

References Ansari, D., 2008. Effects of development and enculturation on number representation in the brain. Nat. Rev. Neurosci. 9 (4), 278. Ansari, D., Dhital, B., 2006. Age-related changes in the activation of the intraparietal sulcus during nonsymbolic magnitude processing: an event-related functional magnetic resonance imaging study. J. Cogn. Neurosci. 18 (11), 1820e1828. Ansari, D., Garcia, N., Lucas, E., Hamon, K., Dhital, B., 2005. Neural correlates of symbolic number processing in children and adults. Neuroreport 16 (16), 1769e1773. Bassett, D.S., Bullmore, E., 2006. Small-world brain networks. Neuroscientist 12, 512e523. Ben-Shachar, M., Dougherty, R.F., Deutsch, G.K., Wandell, B.A., 2011. The development of cortical sensitivity to visual word forms. J. Cogn. Neurosci. 23 (9), 2387e2399. Carpenter, P.A., Just, M.A., Keller, T.A., Cherkassky, V., Roth, J.K., Minshew, N.J., 2001. Dynamic cortical systems subserving cognition: fMRI studies with typical and atypical individuals. In: McClelland, J.L., Siegler, R.S. (Eds.), Mechanisms of Cognitive Development: Behavioral and Neural Perspectives. Erlbaum, Mahwah, NJ, pp. 353e383. Cohen Kadosh, K., Cohen Kadosh, R., Dick, F., Johnson, M.H., 2010. Developmental changes in effective connectivity in the emerging core face network. Cereb. Cortex 21 (6), 1389e1394. Cohen-Kadosh, K., Johnson, M.H., 2007. Developing a cortex specialized for face perception. Trends Cog. Sci. 11, 367e369.

Theories in developmental cognitive neuroscience Chapter | 13

287

de Haan, M., Humphries, K., Johnson, M.H., 2002. Developing a brain specialized for face processing: a converging methods approach. Develop. Psychob. 40, 200e212. Elman, J., Bates, E., Johnson, M.H., Karmiloff-Smith, A., Parisi, D., Plunkett, K., 1996. Rethinking Innateness: A Connectionist Perspective on Development. MIT Press, Cambridge, MA. Fair, D.A., Dosenbach, N.U.F., Church, J.A., Cohen, A.L., Brahmbhatt, S., Miezin, F., Barch, D.M., Raichle, M.E., Petersen, S.E., Schlaggar, B.L., 2007. Development of distinct control networks through segregation and integration. PNAS 104, 13507e13512. Fair, D.A., Cohen, A.L., Power, J.D., Dosenbach, N.U., Church, J.A., Miezin, F.M., Schlaggar, B.L., Petersen, S.E., 2009. Functional brain networks develop from a “local to distributed” organization. PLoS Comput. Bio. 5, e1000381. Friston, k.J., Price, C.J., 2001. Dynamic representations and generative models of brain function. Brain Res. Bull. 54, 275e285. Gauthier, I., Nelson, C., 2001. The development of face expertise. Curr. Op. Neurob. 11, 219e224. Gauthier, I., Tarr, M.J., Anderson, A.W., Skudlarski, P., Gore, J.C., 1999. Activation of the middle fusiform “face area” increases with expertise recognizing novel objects. Nat. Neurosc. 2 (6), 568e573. Geldart, S., Mondlach, C.J., Maurer, D., de Schonen, S., Brent, H.P., 2002. The effect of early visual deprivation on the development of face processing. Develop. Sci. 5, 490e501. Gilbert, C.D., Sigman, M., 2007. Brain state: top-down influences in sensory processing. Neuron 54, 677e696. Golarai, G., Ghahremani, D.G., Whitfield-Gabrieli, S., Reiss, A., Eberhardt, J.L., Gabrieli, J.D., Grill-Spector, K., 2007. Differential development of highlevel visual cortex correlates with category-specific recognition memory. Nat. Neurosci. 10 (4), 512. Gottlieb, G., 1992. Individual Development and Evolution. Oxford University Press, New York. Gottlieb, G., 2007. Probabilistic epigenesis. Dev. Sci. 10 (1), 1e11. Holloway, I.D., Ansari, D., 2010. Developmental specialization in the right intraparietal sulcus for the abstract representation of numerical magnitude. J. Cog. Neurosci. 22 (11), 2627e2637. Johnson, M.H., 2011. Interactive Specialization: a domain general framework for human functional brain development? Dev. Cog. Neurosci. 1, 7e21. Johnson, M.H., de Haan, M., 2015. Developmental Cognitive Neuroscience: An Introduction, fourth ed. John Wiley & Sons, Ltd. Johnson, M.H., Munakata, Y., 2005. Processes of change in brain and cognitive development. Trends Cog. Sci. 9, 152e158. Johnson, M.H., Halit, H., Grice, S., Karmiloff-Smith, A., 2002. Neuroimaging of Typical and Atypical Development: a perspective from multiple levels of analysis. Develop. Psychopath. 14, 521e536. Mareschal, D., Johnson, M.H., Sirois, S., Spratling, M., Thomas, M., Westermann, G., 2007. Neuroconstructivism. In: How the Brain Constructs Cognition, vol. I. Oxford University Press, Oxford, UK. Karmiloff-Smith, A., 2002. Development itself is the key to understanding developmental disorders. In: Johnson, M.H., Munakata, Y., Gilmore, R. (Eds.), Brain Development and Cognition: A Reader. Blackwell, Oxford, pp. 336e374. Karni, A., Bertini, G., 1997. Learning perceptual skills: behavioral probes into adult cortical plasticity. Curr. Op. Neurob. 4, 530e535. Kerszberg, M., Dehaene, S., Changeux, J., 1992. Stabilization of complex input-output functions in neural clusters formed by synapse selection. Neur. Networks (5), 403e413. Kucian, K., Loenneker, T., Dietrich, T., Martin, E., Von Aster, M., 2005. Gender differences in brain activation patterns during mental rotation and number related cognitive tasks. Psychol. Sci. 47 (1), 112. Kuhn, T.S., 1996. The Structure of Scientific Revolutions, third ed. University Of Chicago Press. Lawson, R.P., Friston, K.J., Rees, G., 2015. A more precise look at ’context’ in autism. Proc. Natl. Acad. Sci. USA 112, E5226. Le Grand, R., Mondloch, C.J., Maurer, D., Brent, H., 2001. Early visual experience and face processing. Nature 410e890. Luna, B., Thulborn, K.R., Munoz, P.D., Merriam, E.P., Garver, K.E., Minshew, N.J., Keshavan, M.S., Genovese, C.R., Eddy, W.F., Sweeney, J.A., 2001. Maturation of widely distributed brain function subserves cognitive development. NeuroImage 13 (5), 786e793. Mareschal, D., Thomas, M.S.C., 2007. Computational modeling in developmental psychology. IEEE Trans. Evolut. Comp. 11, 137e150. Markham, J.A., Greenough, W.T., 2004. Experience-driven brain plasticity: beyond the synapse. Neur. Glia Biol. 1, 351e363. Marr, D., 1982. Vision. W.H. Freeman, San Francisco. McCarthy, G., Puce, A., Gore, J.C., Allison, T., 1997. Face-specific processing in the human fusiform gyrus. J. Cog. Neurosc. 9 (5), 605e610. Mills, D.L., Coffey, S.A., Neville, H.J., 1993. Language acquisition and cerebral specialization in 20-month-old infants. J. Cog. Neurosc. 5, 317e334. Morton, J., Frith, U., 1995. Causal modelling: a structural approach to developmental psychopathology. In: Cicchetti, D., Cohen, D. (Eds.), Manual of Developmental Psychopathology, vol. 1. john Wiley, New York, pp. 357e390. Nagy, Z., Westerberg, H., Klingberg, T., 2004. Maturation of white matter is associated with the development of cognitive functions during childhood. J. Cog. Neurosc. 16, 1227e1233. Neville, H.J., 1991. Specialization of the language hemisphere?. In: Modularity and the Motor Theory of Speech Perception: Proceedings of a Conference to Honor Alvin M. Liberman. Psychology Press, p. 269. Oliver, A., Johnson, M.H., Shrager, J., 1996. The emergence of hierarchical clustered representations in a hebbian neural, model that simulates development in the neocortex. Network: Comp. Neur. Syst. 7, 291e299. Oliver, A., Johnson, M.H., Karmiloff-Smith, A., Pennington, B., 2000. Deviations in the emergence of representations: a neuroconstructivist framework for analysing developmental disorders. Dev. Sci. 3, 1e23. Oyama, S., 2000. The Ontogeny of Information: Developmental Systems and Evolution, second ed. Duke Univ. Press, Durham. O’Reilly, R.C., 2006. Biologically based computational models of high-level. Cog. Sci. 314, 91e94.

288 PART | II Cognitive development

Perone, S., Plebanek, D.J., Lorenz, M.G., Spencer, J.P., Samuelson, L.K., 2019. Empirical tests of a brain-based model of executive function development. Child Dev. 90 (1), 210e226. Poldrack, R.A., 2002. Neural systems for perceptual skill learning. Behav. Cogn. Neurosci. Rev. 1 (1), 76e83. Rivera, S.M., Reiss, A.L., Eckert, M.A., Menon, V., 2005. Developmental changes in mental arithmetic: evidence for increased functional specialization in the left inferior parietal cortex. Cereb. Cortex 15 (11), 1779e1790. Scherf, K.S., Behrmann, M., Humphreys, K., Luna, B., 2007. Visual category-selectivity for faces, places and objects emerges along different developmental trajectories. Dev. Sci. 10 (4), F15eF30. Schlaggar, B.L., Brown, T.T., Lugar, H.M., Visscher, K.M., Miezin, F.M., Petersen, S.E., 2002. Functional neuroanatomical differences between adults and school-age children in the processing of single words. Science 296 (5572), 1476e1479. Shrager, J., Johnson, M.H., 1995. Waves of growth in the development of cortical function: a computational model. In: Julesz, B., Kovacs, I. (Eds.), Maturational Windows and Adult Cortical Plasticity, the Santa Fe Institute Studies in the Sciences of Complexity, vol. XXIII. Addison-Wesley, New York, pp. 31e44. Shultz, T.R., Rivest, F., 2001. Knowledge-based cascade-correlation: using knowledge to speed learning. Conn. Sci. 13, 43e72. Shultz, T.R., Rivest, F., Egri, L., Thivierge, J.P., Dandurand, F., 2007. Could knowledge based neural learning Be useful in developmental robotics? The case of KBCC. Int. J. Hum. Rob. 4, 245e279. Siegel, M., Körding, K.P., König, P., 2000. Integrating top-down and bottom-up sensory processing by somato-dendritic interactions. J. Comput. Neurosc. 8, 161e173. Spratling, M., Johnson, M.H., 2004. A feedback model of visual attention. J. Cog. Neurosc. 16, 219e237. Spratling, M., Johnson, M.H., 2006. A feedback model of perceptual learning and categorization. Visual Cog. 13, 129e165. Supekar, K.S., Musen, M.A., Menon, V., 2009. Development of large-scale functional brain networks in children. NeuroImage S109. Thatcher, R.W., 1992. Cyclic cortical reoroganization during early childhood. Special issue: the role of frontal lobe maturation in cognitive and social development. Brain Cog. 20, 24e50. Thivierge, J.P., Titone, D., Shultz, T.R., 2005. Simulating frontotemporal pathways involved in lexical ambiguity resolution. Poster Proc. Cog. Sci. Soc. 2005, 2178e2183. Thomas, M., Johnson, M.H., 2008. New advances in understanding sensitive periods in brain development. Curr. Direct. Psycho. Sci. 17, 1e5. Thomas, M.S.C., Ronald, A., Forrester, N.A., 2013. Modelling social economic status effects on language development. Dev. Psychol. 49, 2325e2343. Ungerleider, L.G., Doyon, J., Karni, A., 2002. Imaging brain plasticity during motor skill learning. Neurob. Learn. Mem. 78 (3), 553e564.

Chapter 14

Structural brain development: birth through adolescence Stefanie C. Bodison1, 2, John B. Colby3 and Elizabeth R. Sowell2, 4 1

Chan Division of Occupational Science and Occupational Therapy, University of Southern California (USC), Los Angeles, CA, United States; 2Keck

School of Medicine of USC, Department of Pediatrics, Los Angeles, CA, United States; 3Department of Radiology and Biomedical Imaging, University of California, San Francisco, CA, United States; 4Developmental Cognitive Neuroimaging Laboratory, Children’s Hosiptal, Los Angeles, CA, United States

Chapter outline 14.1. Introduction 14.2. Postmortem studies and histology 14.2.1. Synaptogenesis and pruning 14.2.2. Myelination 14.2.3. Sex-specific differences 14.2.4. Summary 14.3. Magnetic resonance imaging volume analyses 14.3.1. Gray matter decreases in development 14.3.2. Regional and temporal dynamics 14.3.3. White matter increases in development 14.3.4. Sex differences 14.4. Magnetic resonance imaging brain mapping approaches 14.4.1. Voxel-based strategies 14.4.2. Cortical thickness 14.4.3. White matter

289 290 290 291 291 292 292 292 292 293 293 294 294 296 297

14.4.4. Sex differences 298 14.4.5. Summary 300 14.5. Diffusion magnetic resonance imaging 300 14.5.1. Diffusion tensor imaging theory 300 14.5.2. Diffusion parameters in development 300 14.5.3. Fiber tractography 302 14.5.4. Sex differences 303 14.5.5. Advanced diffusion magnetic resonance imaging techniques 303 14.5.6. Summary 304 14.6. Connecting different techniques 304 14.6.1. Multimodal imaging 304 14.6.2. Brainebehavior relationships 305 14.7. Conclusions and future directions 309 References 309

14.1 Introduction Human brain development is a dynamic process that begins in utero and continues prominently through childhood, adolescence, and young adulthood. While strongly influenced by genetic factors, the environment also prominently affects brain maturation by acting on the cellular and macroscopic levels. This experiential learning impacts both brain structure and function through forms of neuronal plasticity that continue throughout our lifetimes. However, despite the fact that investigating brain development is undoubtedly one of the keys to appreciating how we emerge as unique human beings and how this process can go awry in disease, our understanding of this important period has historically been hindered by two main factors. First, there has been a lack of reliable postmortem data, as thankfully children are generally healthy during development. Second, technological limitations of past methods such as positron emission tomography (PET) and computed tomography (CT) often imposed some modest risk of harm to the subject (e.g., ionizing radiation), which made the study of healthy typically developing children ethically questionable. The situation changed dramatically with the dissemination of magnetic resonance imaging (MRI) technology during the 1980s, which not only offers higher-quality images of the brain parenchyma than ultrasound, X-ray, CT, or PET but also does so in a way that is remarkably safe for the subject. Since its introduction over 30 years ago, MRI acquisition and analysis techniques have evolved to transform our understanding of the developing brain from birth through adolescence.

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00014-7 Copyright © 2020 Elsevier Inc. All rights reserved.

289

290 PART | II Cognitive development

This discussion will begin with a review of the historical postmortem and histological literature and will then move on to the groundbreaking in vivo neuroimaging investigations of the 1990s and mid-2000s that first examined brain development using 1.5 T (T) MRI technology. A collection of more detailed phenomena will then be examined, highlighting research findings using 3T MRI technology, with advanced brain mapping techniques that have come together as a set of classic features that characterize typical brain development. Finally, we will conclude with a discussion of the cutting-edge efforts being made to integrate these diverse observations within a more generalized “multimodal” imaging framework and relate them to advancements in cognitive development. A focus on prominent sex-specific, pubertal, regional, and temporal variations will be continually threaded throughout this discussion.

14.2 Postmortem studies and histology Although data sets were sparse, postmortem and histological studies were able to provide key insight into normal brain structure and development, as well as pathology, decades before the introduction of neuroimaging methods such as PET and MRI. Furthermore, the rich literature that developed from this early effort has provided a strong foundation of data against which newer imaging modalities can be validated. Compared with a modern imaging method such as MRI, there are several distinct advantages and disadvantages of these postmortem studies. Not only are data sets relatively small in postmortem samples, as mentioned previously, but also longitudinal studiesdvalued for their statistical power to detect changes over time within individuals among the highly variable populationdare impossible to conduct. Conversely, because postmortem methods can directly visualize the brain tissue, spatial resolution far exceeds even the best neuroimaging protocol, and there is less validation needed to ensure that the raw signal being measured faithfully represents the underlying neuronal architecture. Artifacts, however, are an important concern for either method. While postmortem methods may introduce artifacts due to cell death, fixation/staining procedures, and morphological changes due to osmotic pressure and mechanical damage, MRI data suffer artifacts from other sources such as magnetic susceptibility effects (signal loss in regions near large caverns of air), local image distortions caused by magnetic field inhomogeneities, and partial volume effects that occur when different structures fall within the same voxel. Many of these issues present less of a problem for the interpretation of larger data sets obtained with MRI, relative to postmortem data, as effects of artifacts generally become small as the number of samples becomes large. However, it should still be remembered that MRI only offers a wide-angle indirect view of tissue, which cannot reach down to the cellular level, and must be observed through the complex lens of magnetic resonance.

14.2.1 Synaptogenesis and pruning By the time an infant is born, the human brain already contains on the order of 100 billion neurons (Kandel et al., 2013). The period of rapid overall brain growth that began in utero continues after birth through the first years of life. Surprisingly, however, postmortem studies during the early part of the 20th century showed that total brain volume and weight actually plateau early and reach approximately 90% of their adult values by age 5 years (Dekaban and Sadowsky, 1978; Riddle et al., 2010; Vignaud, 1966). Even during this early period of pronounced overall growth, brain development is characterized as a dynamic process with both progressive and regressive changes that are influenced by complex genetic influences as well as experiencedependent plasticity due to environmental influences. As the infant brain grows in size, it also grows in complexity. Neurons undergo dendritic branching, forming an arbor of neural connections through synaptogenesis, and then ultimately refine this global brain network through the processes of myelination and synaptic pruning. Much of our understanding of the complex balance between synaptogenesis and synaptic pruning has evolved from the seminal histological work performed by Huttenlocher and colleagues, who mapped synaptic density in different areas of the brain throughout childhood. Overall synaptic density is comparable with the adult level at birth. It then rises even further through the first year of life to its peak at 12e18 months and then decreases during late childhood and young adulthood toward a stable adult plateau of w1 billion synapses*mm 3 (Huttenlocher, 1979). This has helped to form the theory that the flexible groundwork laid through an initial overabundance of connections gives way to a reduceddbut more targeted and efficientdnetwork through experience-dependent synaptic pruning. Interesting regional variations were also observed during these studies, with primary visual and auditory cortex reaching their peak synaptic densities earlier than prefrontal cortex (Huttenlocher, 1979; Huttenlocher and Dabholkar, 1997; Huttenlocher et al., 1982). The extended period of synapse elimination also has regional variations, with pruning ending by age 12 years in the auditory cortex but continuing through midadolescence in the prefrontal cortex. This temporal pattern parallels concurrent gains in the cognitive domains that are thought to relate to these regions (Spear, 2000; Luna et al., 2004).

Structural brain development: birth through adolescence Chapter | 14

291

14.2.2 Myelination Myelination of axonal projections by oligodendroglia is also a prominent component of early brain development. This process begins in utero, continues rapidly through the first 5 years of life, and remarkably extendsdalthough at a slower ratedthrough young adulthood. Intracortical histological preparations by Kaes in 1907 were some of the first to demonstrate this prolonged trajectory of myelination and also its striking regional variability in timing (Kaes, 1907; Kemper, 1994). These slides not only demonstrated earlier trajectories in some areas (posterior temporal, precentral, and postcentral cortex) than others (superior parietal, anterior temporal, anterior frontal cortex) but also showed that regions with a more protracted developmental trajectory have more pronounced changes during older age. This has helped to form the “first-in-last-out” theory of aging (Davis et al., 2009), which suggests that higher-order cognitive manifestations (e.g., problem solving and logical reasoning)dsome of the last to develop (Luna et al., 2004)dare some of the first to degenerate in old age. Furthermore, the visible spread of myelin outward into the cortex results in an apparent cortical thinning, which suggests that normal developmental decreases in cortical thickness (discussed later) may be due, in part, to this progressive increase in myelin and not simply due to regressive changes such as synaptic pruning and cell loss. These initial observations in intracortical tissue were extended to the white matter in pioneering work performed by Yakovlev and Lecours in the 1960s. They demonstrated that white matter myelination begins in utero during the second trimester of pregnancy and continues throughout young adulthood (Yakovlev and Lecours, 1967). Additionally, they extended the earlier observations of regional variations in the timing of myelination and described a general posterior-toanterior trend in the timing of white matter myelination during development that has also been replicated in other samples (Kinney et al., 1988). Later independent research targeting the hippocampal formation has also noted striking increases in myelination, with a 95% increase observed in the extent of myelination relative to brain weight during the first two decades of life. Surprisingly, the authors noted that expanding myelination continued even through the fourth to sixth decades of life (Benes et al., 1994). Taken together, these observations suggest that structural white matter development, in the form of advancing myelination, proceeds in tune with overall cognitive developmentdwith areas involved in lower-order sensory and motor function myelinating earlier than areas involved with higher-order executive function. This correlated timing implies that there may be some relationship between advancing brain function and increased myelination; however, postmortem studies are limited from investigating this directly.

14.2.3 Sex-specific differences A pronounced sexual dimorphism in overall brain size emerges during the first 5 years of human brain development, with males having brains that are, on average, approximately 10% larger than females at their adult plateau (Dekaban and Sadowsky, 1978). This simple and widely reproducible observation has served as a catalyst for continued interest in the study of sex-specific differences during brain development to (1) map other detailed components of brain development that may also show sex-specific differences, (2) determine if there are any cognitive correlates with these findings (Kimura, 1996), (3) establish whatdif not total volume of brain matterdare the driving structural contributors to individual cognitive differences in areas such as language skills and overall intelligence, and, perhaps most importantly, (4) better understand and clinically address the range of neuropsychiatric disorders that tend to emerge during adolescence with prominent sex-specific affinities (Marsh et al., 2008). Interestingly, while some of this sex-specific variance in brain size can be attributed to height, which is consistent with broader trends across different mammalian species, there remains a significant sex-specific effect on brain size even when differences in body size are taken into account (Peters et al., 1998). Although the brains of adult males tend to be larger than adult females, this increase is actually smaller than what would be predicted based on differences in adult height alone. Histological findings indicating a 15% higher neuronal density in males than females are consistent with this (Rabinowicz et al., 2002), although conflicting reports from other studies prohibit firm conclusions on this point (Haug, 1987; Pakkenberg and Gundersen, 1997). A consideration of the fact that females actually tend to be taller than males during late childhood, perhaps due to faster pubertal maturation in girls, further weakens the idea of such a simple allometric relationship when age-matched males and females are compared (Giedd et al., 2006). These discrepancies highlight the diversity that exists among the postmortem literature on the topic of sex differences in brain development, which is also likely to be influenced by a variety of confounds (including cohort effects and observational bias) that have made interpretation challenging (Peters et al., 1998). Additionally, these reports are limited to either simple global measures, such as total brain volume or weight, or very local measures, such as neuronal density, and generally do not account for regional variations in measures such as cortical thickness and folding complexity (Luders et al., 2004; Sowell et al., 2006).

292 PART | II Cognitive development

14.2.4 Summary The central theme that emerges from this early postmortem work is that brain development from birth through adolescence is a uniquely dynamic process, encompassing both progressive and regressive events, with varying magnitudes and timing across different regions of the brain. In particular, the concurrent decrease in synaptic density and increase in white matter myelination is consistent with the principle of selective specialization, which has been postulated to be the driving force behind the creation of cognitive networks and thought to form the foundation for higher cognitive processes (Fuster, 2002; Post and Weiss, 1997; Tsujimoto, 2008). The initial overabundance of neurons and synapses during infancy is thought to provide a flexible substrate through which activity-dependent plasticity can fine-tune neural network activity, via processes such as synaptic pruning, which continue robustly through adolescence and in some form throughout life.

14.3 Magnetic resonance imaging volume analyses With the development of MRI, not only were clinicians provided with a superior technology for the diagnosis of brain injury and disease (Prager and Roychowdhury, 2007; Barkovich, 2006; Panigrahy and Blüml, 2009), but researchers were also provided with an unparalleled technology for the study of typical brain development in vivo. This, together with the expansion of computing technology during the 1980s, led to the first wave of structural neuroimaging studies using 1.5T MRI scanners aimed at extending previous postmortem results. Much of this early work utilized volumetric parcellation methods, whereby brain images are segmented according to different anatomical landmarks, and the volumes and tissue content (gray matter, white matter, cerebrospinal fluid) of these different regions are computed and compared between subject groups or throughout development. This parcellation step has been performed with a variety of methods, including the use of stereotactic coordinates (Jernigan et al., 1991a; Reiss et al., 1996), manually drawn regions of interest (ROIs) (Giedd et al., 1996c; Sowell et al., 2002b), and automated protocols (Giedd et al., 1996a). With the emergence of 3T scanners in the mid-2000s and expanding software and analyses techniques, the ability to gain refined insight into the volumetric properties of the brain was enhanced. This spurred additional studies to confirm and refine previous assumptions about longitudinal brain development, particularly related to variances in overall volume and gray matter changes over time as a function of sex and pubertal differences. As methods of statistical analysis have improved, challenges have arisen related to the complexity of associations across brain areas. For an excellent review of current methodological approaches and best practices, see Vijayakumar et al. (2017).

14.3.1 Gray matter decreases in development Given the previous postmortem observations of regional and temporal variations in synaptic density and myelination throughout the brain, the gray and white matter volume estimates extracted through these volumetric parcellation methods would be expected to show similar age-related developmental trajectories and regional differences. This was first demonstrated by Jernigan and Tallal, who observed that a group of children aged 8e10 years had significantly more cortical gray matter than a group of young adults, as well as a higher gray matter to white matter ratio (Jernigan and Tallal, 1990). A subsequent study extended these findings to confirm that the group differences were due to continuous age-related decreases in gray matter volume with timedindependent of brain sizedand localized these effects to superior frontal and parietal cortices (Jernigan et al., 1991b). These studies marked the first in vivo morphological evidence in support of the earlier postmortem histological work by Huttenlocher and colleagues. While not a direct measure of synaptic density, the volumetric MRI finding of decreased gray matter volume is consistent with the regressive synaptic pruning changes previously described and aligns with the theory that evolutionarily complex regions such as the frontal lobe show more protracted timing in their development than evolutionarily simpler regions such as primary motor/visual cortex. Even this early on, Jernigan and colleagues were also aware of the possible relationship between their in vivo MRI findings and the postmortem white matter myelination studies of Yakovlev and Lecours and suggested that an “apparent” cortical thinning could be due, in part, to progressing myelination. Thus, a component of these observed changes might not be a gray matter loss, per se, but a transition of unmyelinated “gray” matter into white matter, which, on MRI, would appear as a gray matter volume “loss” during the childhood and adolescent years.

14.3.2 Regional and temporal dynamics Since these initial observations of childhood and adolescent gray matter volume loss, other investigations have confirmed the general trend (Aubert-Broche et al., 2013; Caviness et al., 1996; Ducharme et al., 2015; Giedd et al., 1999a;

Structural brain development: birth through adolescence Chapter | 14

293

Ostby et al., 2009; Pfefferbaum et al., 1994; Reiss et al., 1996; Sowell et al., 2002b; Tamnes et al., 2017a; Vijayakumar et al., 2016; Wierenga et al., 2014a; Wilke et al., 2007) and extended these observations in several important ways. In a large cross-sectional sample of 161 subjects aged 3 months to 70 years, Pfefferbaum et al. (1994) and others were able to demonstrate the early rise and plateau in total brain volume by approximately age 5 years, as well as the late childhood and adolescent decline in cortical gray matter volume. The extended age range of the sample allowed them to characterize the trajectory of the gray matter volume decline as curvilinear, which suggested an overall inverted “U”-shaped curve consisting of early childhood gray matter increases followed by a relatively early peak, and then late childhood and adolescent reductions (Pfefferbaum et al., 1994). This general time course of cortical development is a feature that has gone on to become one of the hallmarks of structural brain development (Courchesne et al., 2000; Paus et al., 2001; Sowell et al., 2003). This was further replicated by Mills et al. (2016), who conducted a comprehensive analysis of volumetric changes across multiple longitudinal data sets using replicable statistical models. Their findings on 391 participants across 852 scans found consistent evidence of rising cortical gray matter volume in childhood followed by steadily decreasing gray matter volume into late adolescence and adulthood. Other studies have investigated the relative volume changes (controlling for global increases in total brain volume) more closely in broader age samples and in specific cortical and subcortical structures. In doing so, this work has demonstrated further heterogeneity in maturational timing and trajectory complexity across the brain. Importantly, the relative gray matter volume reduction during adolescence was confirmed to be most concentrated in the frontal and parietal lobes (Sowell et al., 2002b; Tanaka et al., 2012). Meanwhile, subcortical gray matter structures such as the basal ganglia also generally showed a relative volume reduction, although with a simpler linear trajectory than the cortex over the age range of late childhood to young adulthood (Ostby et al., 2009; Sowell et al., 2002b; Tazrouchi et al., 2009).

14.3.3 White matter increases in development Interestingly, while gray matter volume was observed to peak early, and researchers began to consistently observe that white matter volume continues to steadily increase roughly linearly from birth through adolescence and young adulthood (Aubert-Broche et al., 2013; Caviness et al., 1996; Lebel and Beaulieu, 2011; Mills et al., 2016; Paus et al., 2001; Pfefferbaum et al., 1994; Sowell et al., 2002b; Wilke et al., 2007). The timing of these changes shows a posterior-toanterior gradient, which generally parallels the overlying gray matter, and has led to continued investigation into the interaction between these processes (Barkovich et al., 1988). The white matter volume increase is consistent with the widespread reports of relative gray matter reductions during later childhood and adolescence, as a protracted increase in underlying white matter volume (due in part to increased oligodendroglial wrapping of axonal fibers) will increase total brain volume and therefore decrease the relative gray matter volumes of specific structures compared with this total. The midsagittal corpus callosum was one of the first white matter areas to be examined in more detail, with volumetric analyses showing robust increases in total area throughout adolescence (deBellis et al., 2001; Giedd et al., 1999b) and a surprising anterior-to-posterior trend in the timing of the growth curve when the corpus callosum was subdivided into seven distinct segments (Giedd et al., 1996a). This protracted nature of white matter development is a thread that we will see repeated in the following sections as imaging modalities and analysis techniques have advanced (Giedd et al., 1999a; Lebel et al., 2008b; Sowell et al., 2003), and one that has gained increasing interest as more attention is being focused on the network properties of the brain as potential important mediators for the late cognitive development seen in domains such as risk/ reward processing, cognitive control, and working memory (Spear, 2000).

14.3.4 Sex differences Sex-specific differences in brain structure were also extended with these structural imaging techniques. Initial studies revealed total brain volume to be approximately 10% larger in males than females at the plateau of overall brain volume that is reached during childhood (Caviness et al., 1996; Courchesne et al., 2000; Durston et al., 2001; Lenroot and Giedd, 2010; Gur et al., 2002), and the significant sex-specific effect appeared to remain even when height and weight were covaried (Giedd et al., 1996b). Recent studies have expanded these findings with males demonstrating larger overall brain volumes than females from childhood through adolescence and into adulthood (Giedd et al., 2012; Paus et al., 2017). Strong evidence suggests that there may be sexual dimorphism in the timing of the developmental trajectory in the cortex, such that males and females have similar overall trajectories of regional brain maturation (both with an inverted “U”shaped curve) but differing gender effects with time because of a difference in the timing of this trajectory (Giedd et al., 1999a). Specifically, there appears to be approximately a 1- to 2-year phase difference between girls and boys, with peaks in gray matter occurring earlier in girls than boys, and regional variations in both phase and the actual differences between

294 PART | II Cognitive development

the sexes (Lenroot et al., 2007). Because this is a temporally dynamic period of development (and not a static one, for example, comparing fully mature adults), assessing sex differences during childhood and adolescence has become a more complicated problem, which requires the dissociation of phase differences (particularly those caused by differences in age of pubertal onset) from sex differences in the maturational trajectories. Recent studies have sought to either control for overall brain size before comparing differences between the sexes (Paus et al., 2017) or to examine each sex separately and combine results only if similar trajectories are noted (Lebel and Beaulieu, 2011). Observations of sex differences in subcortical regions have also been investigated. In early studies, it appeared that, over the course of development, the amygdala increased in volume more in males, whereas the hippocampus more in females (Giedd et al., 1997, 1996b, 1996c, Wilke et al., 2007). This appeared to be in line with animal studies that have shown high densities of steroid hormone receptors in the medial temporal lobe (Sarkey et al., 2008) and also that sex steroids exert trophic effects on these structures (Zhang et al., 2008; Galea et al., 2006; Cooke, 2006). Findings by Herting et al. (2018), Raznahan et al. (2014), and Wierenga et al. (2014b) have further refined our understanding of typical longitudinal development of multiple subcortical structures including the thalamus, basal ganglia, hippocampus, amygdala, and cerebellum. Herting and colleagues assessed subcortical brain development within three distinct samples of children and young adults aged 8e22 years. For all individuals in the study, males were found to have significantly higher brain volumes than females, and significant sex differences were seen across ages over time. Wierenga and colleagues assessed the age- and puberty-related changes in volume of multiple subcortical structures and the cerebellum, finding significant age-related changes in volume for both males and females, with both increasing and decreasing patterns of volume with increasing age, as has been previously shown. Their biggest revelation was that self-reported pubertal status as reported by the Pubertal Development Scale (PDS) better correlated with changes in across various brain structures than did age (Wierenga et al., 2014b). In total, these longitudinal subcortical findings are particularly important in the context of adolescent brain development, as maturation of these processing centers, and their connections to areas such as the prefrontal cortex, may contribute to the dramatic changes seen in social and emotional domains during this period of development (Dahl, 2004; Steinberg, 2005). The caudate nucleus has also been shown to be relatively larger, controlling for total brain volume, in females across several distinct samples (Giedd et al., 1997, 1996b; Sowell et al., 2002b; Wilke et al., 2007). Put another way, the caudate has spared the reduction in volume that is typically shown by other structures in female brains. These observations of sexual dimorphism across multiple subcortical regions including the thalamus, basal ganglia, hippocampus, and amygdala are also important, as they may relate to the emergence of similar sex differences in the incidence of several neuropsychiatric disorders (e.g., attention-deficit hyperactivity disorder, Tourette’s syndrome) that are thought to involve these structures (Herting et al., 2018, Marsh et al., 2008, Raznahan et al., 2014, and Wierenga et al., 2014b).

14.4 Magnetic resonance imaging brain mapping approaches The early volumetric MRI imaging observations by Jernigan and others helped lay the foundation for the next wave of neuroimaging studies designed to further characterize the anatomical changes that occur during normal development. While the volumetric protocols were able to validate much of the classical postmortem literature, as well as provide further evidence for gray matter loss, white matter gain, and regional/temporal dynamics, they are unable to precisely localize where these maturational changes are taking place within the relatively large ROIs studied. Instead, these methods collapse entire regions of the brain down into one or several summary descriptive statistics that may not be characteristic of all functional and structural brain circuits within these large lobar regions. In contrast, newer methods such as voxel-based morphometry (VBM) and cortical thickness analysis are distinct in that they allow for statistical analysis at many points throughout the entire brain volume or at many points across the entire cortical surface and the creation of wholebrain “maps” to visual these data. These enhanced analysis modalities, together with the traditional methods discussed earlier, have contributed greatly to our understanding of normal brain development and provide an important context for the further study of neurodevelopmental and psychiatric disease (Eliez and Reiss, 2000; Marsh et al., 2008).

14.4.1 Voxel-based strategies In VBM, the local fractional gray matter volume is analyzed in the neighborhood around each voxel in the brain to generate whole-brain maps of gray matter “density” or “concentration” (Ashburner and Friston, 2000). Spatial normalization algorithms are applied to align the brains of individual subjects so that each voxel then can be compared throughout development or between groups. Consistent with the previous volumetric studies and postmortem examples, whole-brain mapping strategies utilizing VBM show decreasing gray matter density during later development. In line with the coarse

Structural brain development: birth through adolescence Chapter | 14

295

frontal and parietal lobar localizations of the earlier volumetric reports, the regions showing the most protracted changes in these new analyses include clusters in the dorsal frontal and parietal cortices during the transition from childhood to adolescence (Sowell et al., 1999a), as well as a distinct grouping of dorsal, medial, and orbital frontal cortical areas during the later transition from adolescence into young adulthood (see Fig. 14.1) (Sowell et al., 1999b). The relative specificity of these later changes to the frontal lobes is consistent with the similarly protracted time course of cognitive development in executive function domains, which are also typically thought to involve these frontal regions (Casey et al., 2005; Luna et al., 2004, 2010; Spear, 2000). This notion of gray matter density was extended to allow for analysis on the cortical surface through the method of cortical pattern matching, where sulcal landmarks are manually identified and used to drive accurate nonlinear spatial normalization into a common template, while helping to account for regional, gender, and individual variability (Ashburner et al., 2003; Luders et al., 2004). Using this technique, protracted postadolescent gray matter density decreases were again demonstrated in dorsal frontal cortex (Gogtay et al., 2004; Sowell et al., 2001b) and, for the first time, shown to correlate significantly with underlying relative brain growth in these regions (Sowell et al., 2001b). This suggests the combined influences of both regressive processes such as synaptic pruning, as well as progressive processes such as myelination, are acting in these areas. Using similar gray matter density measurement techniques, and a powerful longitudinal design that tracked individuals prospectively for 8e10 years, Gogtay et al. provided further evidence that lower-order somatosensory and visual areas develop earlier than the higher-order association cortices that integrate these processes and also that phylogenetically older areas develop earlier than younger areas (Gogtay et al., 2004). Surprisingly, however, gray matter density increases were actually observed in bilateral perisylvian regions during the transition from adolescence to adulthood and shown to correlate with both lateralized differences in sylvian fissure morphology and concomitant local brain growth (Sowell et al., 2001b, 2002a). This suggests a particularly extended developmental trajectory in these gray matter regions beyond that in the dorsal frontal lobe and perhaps implies a unique position for these canonical language areas in the developmental landscapedwith the typical inverse correlation between density and volume (decreasing density,

Child to adolescent

Adolescent to adult FIGURE 14.1 Gray matter density maturation. Voxel-based morphometry (VBM) measurements of fractional gray matter density/concentration show typical decreases during development. Colored volumes within a transparent cortical surface rendering represent the extent of significant decreases in gray matter density during the transition from childhood to adolescence (top panel) and adolescence to adulthood (bottom panel). Color coding indicates which changes occurred in the frontal lobe (purple), parietal lobe (red), occipital lobe (yellow), temporal lobe (blue), and subcortical regions (green) (Sowell et al., 1999a, 1999b).

296 PART | II Cognitive development

increasing volume) reversed to give a direct relationship (increasing density, increasing volume) in these areas during this age range (Sowell et al., 2003). Taken together, these findings again highlight both the regional and temporal complexities to the normal developmental sequence of brain structure.

14.4.2 Cortical thickness The investigation of apparent cortical gray matter decreases during development reached full stride with the development of MRI cortical thickness measurement techniques. These automated algorithms extract mesh models of the white matter/ gray matter boundary surface and the pial (i.e., cortical) surface and then directly calculate the cortical thickness at many points throughout the cortical sheet (see Fig. 14.2) (Fischl and Dale, 2000). Unlike the rather abstract interpretation of fractional gray matter “density” estimates, cortical thickness estimates are in physical units of millimeters and validate exceptionally well against the historical postmortem cortical thickness mapsdwith average measurements in children ranging from 1.5 mm in occipital cortex to 5.5 mm in dorsomedial frontal cortex (see Fig. 14.2) (Sowell et al., 2004a; Von Economo, 1929). In addition to their strong agreement with postmortem data in terms of absolute thickness estimates, reports using this method are also in line with the mounting evidence from postmortem, volumetric, and VBM density measurements, which supports the picture that gray matter thickness peaks early and then declines due to a combination of progressive events such as enhanced myelin penetration into the cortical neuropil and regressive events such as continued synaptic pruning (O’Donnell et al., 2005; Shaw et al., 2008; Sowell et al., 2004a; Wierenga, 2014a; Tamnes, 2010). In a longitudinal study of 45 typically developing children aged 5e11 years, who were scanned 2 years apart, these techniques were able to demonstrate gray matter thinning of w0.15e0.30 mm per year coupled to relative brain volume increases in right frontal and bilateral parietooccipital regions (see Fig. 14.3). This study was also able to reproduce the surprising earlier findings of

(A)

(D) 5.5 5.0 4.5 4.0 3.5 3.0 2.5 2.0 1.5 1.0

(B) (E)

(C)

6.0 5.0 4.0 3.0 2.0

< 2 mm 2.0–2.5 mm 2.6–3.0 mm 3.0–3.5 mm 3.6–4.0 mm

1.0 0.0

FIGURE 14.2 Cortical thickness analysis. Panels AeC show a single representative slice for one subject: (A) Raw T1-weighted anatomical MRI scan. (B) Gray/white matter tissue segmentation. (C) Gray matter thickness image, with thickness (mm) coded by color (warmer colors overlie the areas with the thickest cortex). (D) An in vivo average cortical thickness map generated by performing this analysis on a cross-sectional sample of 45 subjects. The brain surface rendering is color-coded according to the underlying cortical thickness (mm) and the color bar at right. The regional variations in cortical thickness can be compared with an adapted version of the classical Von Economo postmortem cortical thickness map (E), which has been color-coded in a similar manner over the original stippling pattern to highlight the similarity between the two maps (Sowell et al., 2004a; Von Economo, 1929).

Structural brain development: birth through adolescence Chapter | 14

Left

297

Right

Perisylvian thickness increase Left inferior frontal thickness increase

3.0 2.5 2.0 1.5 1.0 0.5 0.0 0.5 1.0 1.5 2.0 2.5 3.0

0.05

TD

0.01

TD

0.001

TD

0.05

TI

FIGURE 14.3 Gray matter thickness maturation. Statistical maps showing the significance of cortical thickness change in a longitudinal sample of 45 children scanned twice between the ages of 5 and 11 years. Areas showing significant thickness decrease (TD) are displayed in red, and areas showing significant thickness increase (TI) are displayed in white (see color bar and significance thresholds at right). Nonsignificant areas are coded by their t-statistic according to the left rainbow color bar. Arrows highlight the relative specificity of thickness increases during this age range to canonical language areas in the left inferior frontal gyrus (Broca’s area) and perisylvian region (Wernicke’s area) (Sowell et al., 2004a).

gray matter increases in bilateral perisylvian language areas (Wernicke’s area) and extended these observations to the left inferior frontal gyrusdanother language area (Broca’s area) (Sowell et al., 2004a). Cortical thickening was estimated to be at a rate of 0.10e0.15 mm per year in these areas. This unique pattern of cortical thickening in the canonical language regions could be related to parallel gains in language processing made during this period of development. In another large longitudinal study of 375 children and young adults, changes in cortical thickness were modeled with a low-order polynomial basis set to investigate regional differences in the complexity of the developmental trajectory (Shaw et al., 2008). Patterns of varying complexity were found to parallel the established histological maps of cytoarchitectonic complexity and agree with the previous literature (Gogtay et al., 2004; Sowell et al., 2004a)dwith simpler laminar areas, such as the limbic cortex having simpler trajectories, and more complex laminar areas, such as association cortex having more complex trajectories. Although cortical thinning during adolescence reflects developmental processes such as myelination and synaptic pruning (Brown et al., 2012; Raznahan et al., 2011; van Solen et al., 2012; Zhou et al., 2015) it is important to note that cortical thinning also continues in some form throughout the rest of the life span (Sowell et al., 2003; Zhou et al., 2015). This likely belies a shift in etiology to degenerative changes associated with aging (Sowell et al., 2004b), and recent work has sought to delineate this inflection point more precisely. By analyzing local gray and white matter signal intensities in the context of cortical thinning, the timing of the developmental peak was found to range from 8 to 30 years of age in different regions of the cortex, with the regional pattern following the general posterior-to-anterior gradient discussed before (Westlye et al., 2010). In addition to an overall thinning effect, research by Aleman-Gomez et al. (2013) found that the cortical surface of the brain globally flattens during adolescence, with significant sulcal widening and decreased sulcal depth cooccurring in the frontal and occipital lobes. This flattening effect illustrates a direct connection between cortical thinning and sulcal widening and helps to account for the concurrent presence of increasing white matter thickness coupled with cortical thinning.

14.4.3 White matter Even before the widespread adoption of diffusion imaging, which will be discussed in the next section, researchers were able to adapt traditional anatomical MRI analysis techniques to study white matter development (Wozniak and Lim, 2006). Magnetization transfer ratio imaging is sensitive to the “bound” protons found on the phospholipids of myelin (Wolff and Balaban, 1989) and reflects the increasing myelination during early development (Engelbrecht et al., 1998) as well as the posterior-to-anterior trend in the timing of this process (van Buchem et al., 2001). T2 relaxometry, which estimates the fraction of water in the brain that is associated with the phospholipid bilayer of myelin (MacKay et al., 1994), has also been used to demonstrate the caudal-to-rostral wave of myelination (Lancaster et al., 2003). In an application of the VBM technology to the white matter, Paus and colleagues were able to powerfully interrogate the rather general “global white matter increases” observation previously described to obtain a much richer localization of the precise anatomical regions

298 PART | II Cognitive development

7–11 years (boy)

6–7 years (girl) +80% +60% 8–12 years (girl)

+40% +20% 0% −20%

9–13 years (girl)

Local growth

11–15 years (boy)

3–6 years (girl)

FIGURE 14.4 Corpus callosum maturation. Maps of the local volume changes in the corpus callosum are shown for six individuals aged 3e15 years, who were scanned twice longitudinally with an interval of up to 4 years. Maturation includes outward tissue expansion (warmer colors), with a dynamic wave in timing such that more frontal regions show prominent change early, and more posterior regions show prominent change later (Thompson et al., 2000).

involved. In an 88 subject sample of children aged 4e17 years, they observed a prominent increase in white matter density in the internal capsule bilaterally, as well as the left arcuate fasciculus, suggesting continued maturation of corticospinal and frontotemporal fibers through this age range (Paus et al., 1999). This work agrees with the postmortem data from Yakovlev and Lecours and demonstrates the unique progressive changes that are occurring in the white matter while the cortex has shifted to undergo predominantly regressive events. Confirming the surprising corpus callosum results of the classical volumetric study by Giedd et al. that was discussed earlier, Thompson and colleagues applied a continuum mechanics approach to obtain maps of local tissue deformation in the corpus callosum during development. Their longitudinal design studied six children aged 3e11 years with a follow-up interval of up to 4 years and again demonstrated an anterior-to-posterior wave in the timing of maximal local growth (see Fig. 14.4) (Thompson et al., 2000). This contrasts with the general posterior-to-anterior trend that has been observed in gray matter cortical regions and suggests a unique pattern of development in this region of interhemispheric fiber connectivity.

14.4.4 Sex differences Continuing the trend from volumetric results, VBM gray matter density and cortical thickness observations of sex-specific effects during development have also been variable (Wilke et al., 2007). However, this topic remains a critical issue, as sex-specific differences in brain development are likely to contribute to the sexually dimorphic susceptibilities to a variety

Structural brain development: birth through adolescence Chapter | 14

299

of psychiatric disordersdsuch as schizophrenia and major depressiondthat emerge during adolescence (Durston et al., 2001; Lenroot and Giedd, 2010). Returning to the issue of gender differences in development, Sowell et al. analyzed cortical thickness and local brain size (taken as the distance from the center of the brain) in a large sample of 176 healthy subjects aged 7e87 years. In line with previous studies, male brains were larger than females at all locations. Strikingly, however, absolute cortical thickness was greater in females in right inferior parietal and posterior temporal regions even without accounting for the smaller overall size of female brains. This finding was not significantly modulated by age and was demonstrated even more robustly across broad right temporal and parietal regions when an age- and brain volumematched subset of 18 males and 18 females was evaluated (See Fig. 14.5). These findings suggest that there are both regionally and sex-specific differences in cortical thickness that appear relatively early in childhood and support earlier reports of selective relative increases in gray matter volumes in females (Gur et al., 2002; Goldstein et al., 2001; Nopoulos et al., 2000; Allen et al., 2003; Im et al., 2006; Mutlu et al., 2013; Sowell et al., 2002b). Although the corpus callosum is also a frequent target for brain mapping research into sex-specific effects on brain development, no consensus has been reached, and the topic remains frequently debated (Giedd et al., 2006). Because of the overall smaller brain volume in females, it has also been proposed that there may be evolutionary pressure to develop other compensatory mechanisms. Through sulcal delineation and cortical pattern matching techniques, it has been shown that females tend to develop a greater degree of cortical “complexity” by young adulthood (Luders et al., 2004). This suggests that there is more cortical surface per unit volume in females and may be one mechanism through which female brains have become optimized for their smaller size. An increasing focus is also being shifted away from sex-specific differences, per se, to the known differences in pubertal timing and sex steroid levels that are likely to be major contributors to observe sex-specific effects and their frequently observed modulation by age (Giedd et al., 2006; Lenroot et al., 2007). The emerging picture suggests that

(A) 0.30 0.25 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10 −0.15 −0.20 −0.25 −0.30 −0.35 −0.40 −0.45

(B) 0.20 0.15 0.10 0.05 0.00 −0.05 −0.10 −0.15 −0.20

FIGURE 14.5 Sex-specific differences in cortical thickness. (A) Sex differences in cortical thickness (mm) among an age- and brain volumeematched sample of 18 males and 18 females. Warmer colors (0 on the color bar at right) are regions where females have thinner cortex, relative to males. (B) Statistical maps showing the significance of these sex differences. Areas where the cortex is significantly thicker in females are shown in red and include right inferior parietal and posterior temporal, and left posterior temporal and ventral frontal regions. Areas where the cortex is significantly thinner in females are shown in white and are limited to small regions in the right temporal pole and orbitofrontal cortex. The correlation coefficient is mapped for nonsignificant regions according to the color bar at right.

300 PART | II Cognitive development

puberty and sex steroids do indeed have organizing effects on brain development (Peper et al., 2009a; Witte et al., 2010; Neufang et al., 2008; Bramen et al., 2010). One recent study of 107 9-year-old monozygotic and dizygotic twin pairs not only noted strong overall heritability in regional brain volumes but also demonstrated decreased frontal and parietal gray matter density among the subgroup of individuals who had begun to develop secondary sexual characteristics of puberty (Peper et al., 2009b). Further investigation among the same cohort revealed that the serum level of luteinizing hormone, one of the first indicators of puberty, is associated with both increased overall white matter volume and increased white matter density in the cingulum, middle temporal gyrus, and splenium of the corpus callosum (Peper et al., 2008). The splenium observation is particularly intriguing, as this is the same region shown to have maximal growth over the 9e13 age range in a different study (Thompson et al., 2000). These resultsdobserved between otherwise very well-matched groupsdsuggest that the onset of puberty and sex steroid levels may directly contribute to the decreases in gray matter and increases in white matter that are prominent features of normal brain development during late childhood and adolescence.

14.4.5 Summary Taken together, these structural imaging studies represent a powerful evolution in our understanding of brain development during childhood and adolescence. The overall picture remains one of early overall brain growth, followed by a transition around age 5 years to gray matter decreases coupled to persistent white matter increases. These processes continue through adolescence but relatively balance each other in magnitude. Thus, while overall net brain volume changes relatively little past the age of 5 years, adolescence remains a period of dynamic change beneath the pial surface.

14.5 Diffusion magnetic resonance imaging One of the remarkable discoveries to emerge from these developmental neuroimaging studies is the continued expansion of white matter volume well into adulthood (Sowell et al., 2003; Giedd et al., 1999a). This robust and protracted increase has rewritten the age range associated with brain development (Pujol et al., 1993) and has driven an increasing focus on the white matter and its network connectivity as a possible mediator for the late cognitive gains seen in executive function domains during typical development (Liston et al., 2005), as well as a possible mechanism for neuropathology (Le Bihan (2003)) and training-induced increases in performance (Bengtsson et al., 2005; Carreiras et al., 2009).

14.5.1 Diffusion tensor imaging theory Simultaneously with this growing interest in studying the white matter, as it relates to connectivity between still-maturing brain regions and cognitive function, diffusion imaging was maturing as an MRI variation specifically tuned to examine the white matter (Basser et al., 1994; Le Bihan et al., 1986; Pierpaoli et al., 1996). Since the diffusion properties of water within neural tissue are affected by the geometry of the neuronal microenvironment, it is intuitive that diffusion imaging can provide a sensitive lens through which the microstructural properties of the white matter can be investigated. Specifically, differences in microstructural properties such as fiber coherence, axon packing, and myelination have all been shown to manifest as changes in the diffusion MRI signal (Beaulieu, 2002). By viewing this diffusion landscape within the brain from multiple angles, a more complete “tensor” model of diffusion can be generated for each voxel (Basser et al., 1994). This can be thought of geometrically as a diffusion ellipsoid, with diffusion components in the radial (RD, radial diffusivity) and axial (AD, axial diffusivity) directions (see Fig. 14.6). The size of this ellipsoid corresponds to the overall mean diffusivity (MD). The shape of the ellipsoid corresponds to the directionality of diffusion and is termed fractional anisotropy (FA). It can vary from 0, for perfectly isotropic diffusion, to 1, for perfectly anisotropic diffusion (e.g., the ventricles have low FA, whereas the corpus callosum has high FA). Because it has been shown to be sensitive to myelination, this FA metric has received considerable attention as a way to track the developmental maturation within the white matter and investigate disease. See Le Bihan (2003) and Tamnes, Roalf et al. (2017b) for an excellent review of the methodological progress in this area.

14.5.2 Diffusion parameters in development Using this unique framework, there has been a surge in research aimed at more deeply characterizing the normal developmental processes in these important regions of connectivity that were previously obscured by low contrast within the white matter on traditional T1-weighted anatomical MRI. Similar to the general description in the overlying gray matter,

Structural brain development: birth through adolescence Chapter | 14

(B)

AD

(A)

301

(C)

↓FA ↓MD ↓AD

RD

↓FA ↑MD ↑RD

↑FA ↑MD ↑AD

(D)

↑FA ↓MD ↓RD

FIGURE 14.6 Diffusion tensor imaging (DTI) metrics. DTI metrics include fractional anisotropy (FA), which is a unitless measure of the directionality of diffusion, and mean diffusivity (MD), which is the overall magnitude of diffusion. The center panel shows a cross section of the DTI ellipsoid model of diffusion, which is assumed to be oriented along the fiber axis (shown here as a cylinder). Individual diffusion components along the axial (AD) and radial (RD) directions contribute to the FA and MD values at each point in the brain. Panels (AeD) show different changes in the individual diffusion parameters and their varying effects on FA and MD. Note that changes in different diffusion components (AD or RD) can lead to the same effect on one diffusion metric but have opposite effects on the other. Panel D represents the prevailing regime during development, where decreasing RDddue, in part, to advancing myelinationdleads to increasing FA (a more pointed ellipsoid) and decreasing MD (a smaller ellipsoid).

the developmental trajectory within the white matter is a nonlinear function of time and has prominent regional variations (Lebel et al., 2008b; Mukherjee et al., 2001; Snook et al., 2005). From birth, there is a rapid rise in diffusion directionality (FA; see Fig. 14.7), coupled to a decrease in overall diffusivity (MD) (Bava et al., 2010; Colby et al., 2011; Engelbrecht et al., 2002; Hüppi et al., 1998; Huang et al., 2015; Krogsrud et al., 2016; Löbel et al., 2009; Morriss et al., 1999; Mukherjee et al., 2001; Neil et al., 1998; Schmithorst and Yuan, 2010; Schneider et al., 2004). In an interesting contrast to this general pattern within the white matter, gray matter cortical regions actually have been observed to have decreasing FA in a sample of preterm infants (McKinstry et al., 2002). This could reflect the fact that changes in FA are not highly specific for myelination and may also occur in response to cortical maturational processes such as synaptogenesis. Furthermore, these observations may be related to the perinatal period of selective vulnerability in neural tissue, which has been demonstrated in animal studies and confirmed in humans through MRI (Miller and Ferriero, 2009). The white matter pattern of increasing FA and decreasing overall diffusion, although not universally reported in later development (Schneiderman et al., 2007), generally continues in a decelerating fashion throughout childhood and adolescence and, in some areas, into adulthood (Bonekamp et al., 2007; Klingberg et al., 1999; Schmithorst et al., 2002; Zhang et al., 2007). There is a relatively stable plateau of these parameters during adulthood and then eventual declines later in life (Davis et al., 2009; Salat et al., 2005). Accordingly, the developmental rising portion of this arc has been modeled as a linear (Snook et al., 2005), polynomial (Hsu et al., 2010), or exponential function (Lebel et al., 2008b; Mukherjee et al., 2001; Schneider et al., 2004). The earliest reports utilized an ROI approach to look at diffusion properties averaged across specific anatomical locations and were able to reproduce the “increasing FA, decreasing MD” pattern across a broad variety of regions within the brain and during different periods of development. In one example, Suzuki and colleagues examined ROIs placed bilaterally in the frontal and parietal white matter of 16 children and young adults. They observed increased FA and decreased overall diffusivity with age but went on to make the important determination that the etiology of these changes in FA and MD was a primary decrease in both RD and AD components, with a larger decrease along the radial direction (Suzuki et al., 2003). This explains not only the overall decreased diffusivity that was observed (both components decreased) but also the increased diffusion directionality (one component decreased more than the other). The dominance of changes in RD during development is an important phenomenon that has been broadly replicated (Colby et al., 2011; Giorgio et al., 2008; Lebel et al., 2008b; Löbel et al., 2009; Qiu et al., 2008), although not universally (Giorgio et al., 2010; Ashtari et al., 2007), and is thought to relate to the primary role that extended myelination plays during this age range (Song et al., 2002). Paralleling the advancements made in the analysis of the cortex, methods have quickly adapted to include whole-brain mapping techniques that are able to examine the brain in a spatially continuous manner and better localize developmental changes. In general, these later efforts using VBM and similar techniques have both confirmed and extended the earlier

302 PART | II Cognitive development

Cingulum

0.65

Corticospinal tract

0.55 0.6

Genu

0.65

0.55

FA

FA

0.5 0.45

0.7

0.5 0.4

0.6

0.45 0.35

0.6

FA

0.4 5

0.5

10

15

20

Age (years)

25

30

5

10

15

20

25

30

FA

0.55

0.55

Age (years)

0.45

0.5

0.4

0.45 5

10

15

20

25

Splenium

0.65

5

30

10

0.6

Fornix

0.55

20

25

30

Inferior longitudinal

0.55

0.5

0.5

0.45

FA

FA

15

Age (years)

Age (years)

0.45

0.4

0.4

0.35

0.35 5

10

15

20

25

30

5

10

Age (years)

0.55

Superior fronto-occipital

20

25

30

Superior longitudinal

0.55

0.5

0.5

0.45

FA

FA

15

Age (years)

0.4

0.45 0.4

0.35

0.35 5

10

15

20

25

5

30

10

0.55

15

20

25

30

Age (years)

Age (years)

Uncinate

Inferior fronto-occipital

0.6

0.5 0.55

FA

FA

0.45 0.5

0.4 0.45 0.35 0.4 0.3 5

10

15

20

A e

ears

25

30

5

10

15

20

25

30

Age (years)

FIGURE 14.7 White matter maturation. Diffusion tensor imaging (DTI) tractography was used to identify 10 major white matter tracts in 202 individuals aged 5e30 years (center panels show the extracted tracts for a representative subject). Broad age-related increases in fractional anisotropy (FA), a DTI index of white matter maturation that is sensitive to myelination, were observed across all tracts. Maturational trajectories generally followed an exponential rise, with regional variations in mean FA as well as developmental timing. The surrounding scatterplots demonstrate these relationships and are color-coded according to the tracts in the center panels (Lebel et al., 2008b).

ROI findings of broadly increasing FA and decreasing MD (Snook et al., 2007). Tract-based spatial statistics (TBSS) is an evolution of these methods that is tailored specifically to the analysis of diffusion tensor imaging (DTI) data and has been used successfully to demonstrate age-related changes in diffusion imaging parameters (Colby et al., 2012; Bava et al., 2010; Burzynska et al., 2010; Giorgio et al., 2008, 2010). By projecting the imaging data onto a tract “skeleton” consisting of the cores of the white matter tracts, TBSS avoids some of the alignment problems that arise when the high-contrast FA maps are compared using traditional voxel-by-voxel techniques (Smith et al., 2006, 2007). In a sample of 75 children through young adults that were analyzed using this approach, widespread FA increases and diffusivity decreases were again demonstrated spanning the frontal, temporal, and parietal lobes and the cerebellum (Qiu et al., 2008). Recognizing the need to synthesize these reports into a normative reference standard against which to judge clinical abnormalities, effort has also been directed toward generating developmental brain atlases that integrate this diverse set of information (Hermoye et al., 2006; Löbel et al., 2009; Mori et al., 2008; Verhoeven et al., 2010).

14.5.3 Fiber tractography By making the assumption that the direction of the diffusion ellipsoid (i.e., the direction of principle diffusion) is pointing in the same direction as the neuronal fiber axis, streamlines can be generated passing from voxel to voxel along the path of

Structural brain development: birth through adolescence Chapter | 14

303

principle diffusion. In this manner, the DTI technology has been extended to allow for in vivo fiber tractography (Behrens et al., 2003; Colby et al., 2012; Conturo et al., 1999; Mori et al., 1999; Catani et al., 2002). This allows for individualized measurements to be made that are tailored to each subject’s anatomy, which circumvents many of the problems associated with attempting to register a diverse set of brains to a single template. Although these algorithms have generally validated well against postmortem dissections for many major white matter tracts, specific limitations related to issues such as partial volume averaging and complex fiber geometries must be considered (Pierpaoli et al., 1996). Using this technology, together with standardized protocols for delineating the major white matter tracts of interest (Wakana et al., 2007), researchers have mapped the development of white matter fiber connectivity from before birth (Huang, 2006, 2010; Huang et al., 2009), through childhood, adolescence, and adulthood (Behrens et al., 2003; Brouwer et al., 2012; Herting et al., 2017; Liu et al., 2010; Mishra et al., 2013; Wakana et al., 2004), and even through evolution (Rilling et al., 2008). Like other developmental neuroimaging efforts, these data provide important insight into human brain development in their own right and additionally serve as important normative markers against which pathology can be judged (Adams et al., 2010; Lebel et al., 2008a; Thomas et al., 2009). In a seminal report on the typical developmental trajectories within 10 major white matter tracts in a large sample of 202 subjects aged 5e30 years, Lebel et al. observed continually increasing FA in all regions (generally approximated well by an exponential function) but regional variations in timing such that the time to reach 90% of the adult plateau varied from approximately 7 years old in the inferior longitudinal fasciculus to beyond 25 years old in the cingulum and uncinate fasciculus (see Fig. 14.7) (Lebel et al., 2008b). Overall, they note that frontotemporal connections were the slowest to develop. In a representative example of the degree of intersubject diversity that exists even within tracts, DTI tractography has been used to demonstrate lateralization of different white matter tracts (Bonekamp et al., 2007). In one particular study, left lateralization was shown for the arcuate fasciculus (temporoparietal part of the superior longitudinal fasciculus), with higher FA and more streamlines in the left hemisphere (Lebel and Beaulieu, 2009). These findings are in line with previous observations of left lateralization of perisylvian regions (Geschwind and Levitsky, 1968; Pujol et al., 2002) and are thought to relate to the left hemisphere language dominance that exists in the majority of the population. Interestingly, this same pattern has been demonstrated even in neonates, suggesting that the structural basis of left hemisphere language dominance is present long before the development of speech (Liu et al., 2010). Previous morphometric findings of local volume increases within the corpus callosum (Giedd et al., 1996a; Thompson et al., 2000) have also been explored with tractography. In a large sample of 315 subjects aged 5e59 years, Lebel and others demonstrated the typical trajectory of increasing FA and decreasing MD in the fiber tracts leading from all midsagittal sections of the corpus callosum (Lebel et al., 2010). They also observed an “outer-to-inner” trend in the timing of these maturational arcs, which contrasts with the anterior-to-posterior volumetric trend observed on T1-weighted MRI (Thompson et al., 2000) and highlights the additional insight that can be uncovered when the full extent of a tract is considered.

14.5.4 Sex differences Diffusion imaging also reveals sex-specific structural differences within the white matter (Chahal et al., 2018; Lenroot and Giedd, 2010; Schmithorst et al., 2008; Simmonds et al., 2014; Wang et al., 2012) and pubertal influences on changes in white matter over time (Brouwer et al., 2012; Herting et al., 2017; Peters et al., 2012). In one tractography study of 114 children, adolescents, and young adults, Asato et al. found generally decreasing RD, and protracted maturation past adolescence, in projection and association fibers that included connections between the prefrontal cortex and the striatum. Furthermore, they observed that white matter microstructural maturation proceeded in parallel with pubertal changes, with females having overall earlier maturation of white matter tracts than males (Asato et al., 2010). This suggests that there may be hormonal influences on white matter maturation and that, by considering these aspects, one may obtain a more appropriate estimate of developmental progress than by only considering chronological age. This notion is supported by concurrent findings with structural MRI that demonstrate white matter volume increases during adolescence, especially in boys, are affected by testosterone levels and androgen receptor genes (Menzies et al., 2015; Paus et al., 2010; Perrin et al., 2008).

14.5.5 Advanced diffusion magnetic resonance imaging techniques While DTI has been the most commonly applied analysis technique of white matter microstructure, it has several limitations related to the examination of the direction of fiber diffusivity and interpretation of white matter complexity, especially when fiber tracts intersect or overlap (Jones and Cercignami, 2010; Tamnes et al., 2017b; Tournier et al., 2011). To better understand the nuances of white matter architecture specifically related to directionality, three additional diffusion

304 PART | II Cognitive development

MRI procedures have been developed: high angular resolution diffusion imaging (HARDI), diffusion kurtosis imaging (DKI), and neurite orientation dispersion and density imaging (NODDI). HARDI methods seek to measure the diffusionweighted signal using a larger number of uniformly distributed gradient directions to attempt to capture higher angular frequencies bot adequately modeled by a single diffusion tensor (Chiang et al., 2008; Re et al., 2017). With DKI analyses, the water diffusion and the common problems associated with the tissue structure are quantified and estimated using nonGaussian techniques. With this method, images must be acquired for at least 3 b-values in 15 diffusion directions, but it is hypothesized that the diffusional kurtosis computed here is sensitive to account for the diffusional heterogeneity often seen in neuropathologies (Jensen and Helpern, 2010). NODDI techniques propose combining a three-compartment tissue model with a two-shell HARDI protocol to disentangle FA analyses related to neurite density and orientation (Zhang et al., 2012). This method allows for more precise characterization of neurites, including orientation dispersion. While each of these advanced diffusion MRI techniques holds promise to enhance our understanding of white matter architecture and elucidate greater complexity about developmental trends, it is important to note that a primary disadvantage in utilizing these evolving techniques to date is the substantial increase in acquisition time when compared with DTI procedures. Future researchers will need to carefully consider the feasibility of extended scan times when imaging young children and hope that advancing computing technologies and statistical methods will assist in decreasing acquisition times of these advanced diffusion MRI methods.

14.5.6 Summary Taken together, diffusion imaging studies generally show increasing diffusion directionality (FA) and decreasing overall diffusion (MD) during development. These changes are predominantly due to decreasing RD from the fiber axis, which suggests a primary role for myelination in this process. These changes progress rapidly from birth, through childhood, and eventually level off to a relatively stable adult plateau. Paralleling what has been observed in the cortex and through volumetric observations, there are regional variations in the timing of this developmental trajectory that follow a roughly posterior-to-anterior trend. Sexual dimorphism is also present, with females exhibiting earlier white matter maturation than malesda trend that mimics their differences in pubertal timing.

14.6 Connecting different techniques 14.6.1 Multimodal imaging Although the development of cortical gray matter and the development of white matter microstructure have been investigated independently, one needs to consider their dynamics jointly to determine what relationships exist between them. This challenge returns to one of the original questions that stemmed from the postmortem histological findingsdthat is, “To what degree do myelination and synaptic pruning (and other cellular processes) contribute to the decreasing gray matter and increasing white matter that is found during brain development?” While these phenomena are undoubtedly linked, it remains unclear which is dominant and exactly how they interact. The maturation of DTI and structural MRI analysis techniques has now made it possible to investigate these questions using in vivo imaging data; however, in the end, it will likely be necessary to complete the circle and validate these observations back in histological preparations. Early on in these investigations, Giorgio et al. began by using the TBSS method, discussed before, to demonstrate broad increases in FA that were driven predominantly by decreases in RD. They then made an important and innovative step by incorporating both DTI tractography and gray matter VBM to show that the putative fibers leading from the white matter regions with the strongest developmental effects connect with regions showing significantly decreased gray matter density in the cortex. Furthermore, they, and others, observed that the gray matter density decreases were significantly correlated with the FA increases in the connected white matter (Giorgio et al., 2008; Jeon et al., 2015). By following the structural connectivity present in the actual data, and using these patterns to guide their comparisons, this protocol links the concurrent phenomena of white matter FA increases and gray matter density decreases more convincingly than that was possible with previous qualitative visual inspections. Tamnes et al. later investigated this same general question in a different manner by integrating cortical thickness, volumetric, and DTI measurements derived from a single sample of 168 participants aged 8e30 years (see Fig. 14.8) (Tamnes et al., 2010). As expected, they were able to demonstrate a combination of the phenomena seen in earlier individual studies, including broad cortical thickness decreases, white matter volume increases, FA increases (predominantly decreases in radial diffusion), and MD decreases. Most importantly, however, they were able to go on to demonstrate that, of the three measures, cortical thickness had the strongest relationship with age. Furthermore, although the DTI and volume measures explained some of the variance in cortical

Structural brain development: birth through adolescence Chapter | 14

Relative change: Cortical thickness

Relative change: WM volume

Relative change: FA

Relative change: MD

305

+ 20® % 15–20% 5–15% 0–5% 5–15% 15–20% 20® %

Relative change: DA

Relative change: DR



FIGURE 14.8 Multimodal imaging: volumes, cortical thickness, and diffusion tensor imaging (DTI). Concurrent volumetric, cortical thickness, and DTI analyses were performed in the same sample of 168 participants aged 8e30 years. The percent changes in cortical thickness, white matter volume, fractional anisotropy (FA), mean diffusivity (MD), axial diffusion component (DA), and radial diffusion component (DR) are mapped by region and colorcoded according to the color bar at right. Medial structures and corpus callosum are masked out (Tamnes et al., 2010).

thickness and each other, none of the measures were redundant. This implies that each may be sensitive to different microstructural processes and that all are useful indicators of brain development and microstructural integrity (Fjell et al., 2008). This reiterates the likely mixed regime of both synaptic pruning within the cortex and advancing myelination at the grayewhite cortical interface, which is contributing to brain morphological changes seen during adolescence. In another example, Choi et al. examined a completely different topicdgeneral intelligencedbut were able to gain similar benefits by integrating multiple imaging modalities. They observed that intelligence was generally related to cortical thickness and functional MRI (fMRI) blood flow response during a reasoning task. Because both sets of scans were performed on the same sample of subjects, however, the authors were able to go a step furtherdbeyond this simple generalizationdto observe that the crystallized component of intelligence was more strongly related to cortical thickness, whereas the fluid component of intelligence was more strongly related to functional blood flow response (Choi et al., 2008). At present, there are now more readily available large data sets that include different neuroimaging modalities to assess longitudinal structural brain development than ever before. Brown et al. (2012) assessed the cross-sectional age effects across different neuroimaging modalities and found that age contributions varied widely across measure type and within different neuroanatomical structures (see Fig. 14.9). Their findings illuminate that much more work is needed to help determine the associations across these varying neuroimaging modalities. Given its wide availability, maybe the ABCD study will provide opportunities for the next wave of research to assess the interactions between imaging modality and neuroanatomical structure.

14.6.2 Brainebehavior relationships While important neuroanatomical insight can be gleaned from these structural brain mapping observations, perhaps the most significant outgrowth of this research has been an expanded understanding of the cognitive and behavioral changes that accompany this underlying maturation of brain structure. There has been a long tradition of investigation into the cognitive correlates of brain structure, but unfortunately many of the early findingsdwhich commonly focused on differences between ethnic or social groupsdare unreliable because of data collection and analysis bias (Gould, 1978, 1981). With the advent of MRI, however, volumetric measurements of total brain size have shown a modest but reproducible correlation with general intelligence that emerges over the course of development (Peters et al., 1998; Reiss et al., 1996; Willerman et al., 1991; Witelson et al., 2005), but ongoing research indicates that brainecognition relationships are not constant across development (Fjell et al., 2012), supporting the need to consider different aspects of white matter brain

306 PART | II Cognitive development

FIGURE 14.9 Age-varying contributions of different imaging measures to the prediction of age. At the youngest ages, from about 3 to 11 years old, measures of T2 signal intensity within subcortical regions of interest (ROIs) were by far the strongest predictors of developmental phase, declining in importance through the early teens. Diffusion measures within white matter fiber tracts, in comparison, were consistently strong predictors across the age range, becoming the highest contributor during the middle ages of about 12e15 years. T1-derived morphological measures varied, with cortical thickness and subcortical volumes contributing more than cortical area, which was consistently the weakest predictor over age. Interestingly, diffusivity measures within subcortical ROIs increased sharply at about age 14 years and were the strongest maturational predictors at the oldest ages, from about 17 to 20 years old (Brown, 2012).

maturation over time. Moreover, the correlational nature of these findings does not at all suggest that groups with different brain sizes, like males and females, will have different intelligence. Indeed, independent of the possible relationships with neuroanatomy, it remains exceptionally controversial whether there is even any overall gender effect on intelligence (Blinkhorn, 2005; Hedges and Nowell, 1995; Irwing and Lynn, 2006; Jorm et al., 2004; Lynn and Irwing, 2004; Neisser et al., 1996), and if so, whether the small effect magnitudes that have been reported are relevant given the possible biases that may have contributed. An important additional phenomenon to consider is that both brain structure and intelligence are highly heritable (Shaw, 2007; Thompson et al., 2001). Both are further impacted by environmental influences in a process that begins in utero, continues throughout life, and contributes to individual variations in structural brain development and cognitive function that exist even among monozygotic twins. Although not exclusive, the orchestration of structural brain development by these genetic and environmental factors is one way in which they can converge to influence cognitive development (Toga and Thompson, 2005). Since there is evidence that brain development takes place through selective elimination and connectivity optimization, with prominent regional and temporal variability, it is not surprising that a global measure like total brain volume may not be the optimal choice for investigating the structural basis of cognitive development. Fortunately, the brain mapping strategies, discussed before, have had more success examining brain regionespecific relationships between structure and function. This work has supported many of the classical structureefunction relationships discovered through lesion studiesdfor example, that the prefrontal cortex is related to cognitive control (Damasio et al., 1994)dand also has extended these findings by (1) providing more detail, (2) including more normative subjects without pathology, and (3) allowing for broader investigation in the pediatric population. In this way, these modern neuroanatomical imaging studies, together with complementary results from functional neuroimaging (fMRI) methods that can measure task-dependent blood flow response within the brain (Casey et al., 1995; Luna et al., 2010), have formed a powerful framework to investigate how brain development relates to cognitive function during childhood and adolescence. In this vein, continued investigation into the structural basis of general intelligence has revealed ageevariable relationships between IQ and regional brain structure. In line with the total brain volume results, a correlation between IQ and gray matter volume develops by adulthood (Wilke et al., 2003). However, regional relationships between IQ and gray matter structural measures appear earlier and have been reported to include the anterior cingulate during childhood (Wilke et al., 2003), the

Structural brain development: birth through adolescence Chapter | 14

307

orbitofrontal cortex during adolescence (Frangou et al., 2004), and the frontal lobedparticularly the prefrontal cortexdby adulthood (Haier et al., 2004; Reiss et al., 1996; Thompson, et al., 2001). Interestingly, these regional relationships between gray matter development and IQ appear to be modulated by sex, although the specific regions reported to be most associated with IQ for each sex have been variable (Haier et al., 2005; Narr et al., 2006). In one important study, which investigated the relationship between cortical thickness maturation and IQ in a large longitudinal sample of 307 children and adolescents, IQ was observed to correlate most closely not with cortical thickness, per se, but rather with the shape of the developmental trajectory in cortical thickness change (see Fig. 14.10) (Shaw et al., 2006). The subjects that had the highest IQs tended to have the most dynamic cortical maturation, with more rapid cortical thickening during early childhood and more rapid cortical thinning during late childhood and adolescence. However, in terms of absolute thickness, the superior intelligence group actually had thinner cortex at the start of the age range studied (approximately age 7 years), peaked later, and then had relatively equal thickness to the others by the end of the age range (approximately age 19 years). The relationship between cortical thinning and higher scores on measures of intelligence has been observed in other studies as well (Poerter et al., 2011; Squeglia et al., 2013; Schnack et al., 2015). This observation highlights the notion that, like the pattern of structural maturation itself, the relationships between brain structure and cognitive ability are complicated by their dependency on age during the course of development. While the specific pattern and methodologies of these studies have varied widely, the common pattern that has emerged is a relationship between frontal lobe structural brain development and general intellectual ability. Other studies have investigated more specific cognitive functions and their relation to gray matter structure. In the same longitudinal sample of 45 typically developing children that was described previously, we observed inverse correlations between performance on the vocabulary subtest of the Weschler Intelligence Scale for Children (Wechsler, 2003)da test of general verbal intellectual functioningdand gray matter thickness in left dorsolateral frontal and lateral parietal regions (see Fig. 14.11) (Sowell et al., 2004a). This is consistent with the language dominance of the left hemisphere and suggests

t-statistic

2.6 Superior intelligence (S) High intelligence (H) Average intelligence (A)

5.00 4.75 4.50 S vs H,S vs A P < 0.0001

4.25

H vs A P = 0.56

4.00 7

9

11

Cortical thickness (mm)

(A)

13 15 Age (yr)

17

19

4.50 4.25 4.00 3.75

(C)

S vs H, S vs A P < 0.0001 H vs A P = 0.99

7

9

11

13 15 Age (yr)

Right superior/medial prefrontal gyral cluster 5.00 4.75 4.50 4.25 4.00

S vs A P < 0.001, S vs H P = 0.0005 H vs A P = 0.56

7

9

11

(B)

Left superior/medial prefrontal gyral cluster

4.75

Cortical thickness (mm)

Trajectories at indicated cortical point (MNI coordinates x = 10, y = 44, z = 48)

17

19

Cortical thickness (mm)

Cortical thickness (mm)

5.0

13 15 Age (yr)

17

19

Left middle temporal gyral cluster 5.00 4.75 4.50 4.25 4.00

(D)

S vs A P = 0.0005, S vs H P = 0.003 H vs A P = 0.14

7

9

11

13 15 Age (yr)

17

19

FIGURE 14.10 Trajectory of cortical thickness change versus IQ. Higher IQ was associated with a more dynamic trajectory (more rapid thickening and thinning) in cortical thickness maturation among a sample of 307 children and adolescents scanned longitudinally. The center panel shows regions where there was a significant interaction between IQ group (superior, high, or average) and a cubic age3 term in the regression analysis, which implies a varying trajectory shape in these regions. These individual trajectories are plotted in panels (AeD) and are color-coded according to intelligence group. Arrows indicate the age at peak cortical thickness for each trajectory (Shaw et al., 2006).

308 PART | II Cognitive development

p = 0.1–0.05 p = 0.05–0.01 p < 0.01

FIGURE 14.11 Cortical thickness versus language functioning. Statistical maps showing the significance of the relationship between changes in cortical thickness and changes in vocabulary scores in a longitudinal sample of 45 children scanned twice between the ages of 5 and 11 years. Areas with a significant negative relationship (cortical thinning was associated with improved language performance) are color-coded according to their P value, with the significance thresholds shown in the color bar at right. No positive correlations were observed (Sowell et al., 2004a).

a possible relationship between these concurrent structural and cognitive developmental processes. While originally interpreted as possibly relating to developmental cortical thinning, the results of the Shaw et al. (2006) study suggest that the individuals with the greatest verbal intellectual function here may still have been on the upstroke of their developmental arc in our much younger sample (age 5e11 years) and simply had thinner cortex at the time sampled. This nuance is also reflected in another study, which had an older sample (age 6e18 years) during the later period of development where increased cortical thickness is associated with higher IQ (Karama et al., 2009). Further studies, again in the young sample of 5- to 11-year-olds, have investigated even more targeted cognitive subtests, including phonological processing, and motor speed and dexterity. Structural development in the inferior frontal gyrus (a phylogenetically more complex area that matures slower and is still on the upward stroke of cortical thickening) was expected to relate to advances in phonological processing, which has been shown to involve this area on functional imaging studies (Bookheimer, 2002), but not to relate to advances in motor processing. Conversely, structural development in the hand motor region (a phylogenetically simpler area that matures earlier and is already experiencing cortical thinning) was expected to relate to advances in motor processing but not phonological processing. This predicted double dissociation was demonstrated as expected, which not only illustrates a specific alignment between language development and structural development in the inferior frontal gyrus but also reiterates the regionally specific definition of “structural development” during childhooddwith some cortical regions thinning, but some relatively specific language areas still exhibiting thickening. A similar analysis has also revealed relationships between cortical thinning and both delayed verbal recall functioning and visuospatial memory ability, which is again consistent with the functional neuroimaging literature that suggests the dorsolateral prefrontal cortex is involved with memory recall (Casey et al., 1995; Sowell et al., 2001a). The relationship between cognitive development and structural brain development is further supported by intervention/training studies, which suggest that even relatively short periods of cognitive or motor training can be associated with at least short-term morphological changes in brain structure (Draganski et al., 2004). Diffusion imaging indicators of white matter development also relate to cognitive function. In a sample of 23 children and adolescents, there was a significant direct relationship between diffusion characteristics (FA) and working memory ability in inferior frontal and temporooccipital regions, and the genu of the corpus callosum (Nagy et al., 2004). This relationship existed above and beyond the correlation of each individual measure with age, which suggests that the maturation of the white matter in specific areasdas indexed by FAdmay play a role in the development of (or simply reflect the development of) specific cognitive domains. In a similar design, others have shown correlations between Chinese reading score and FA in the anterior limb of the left internal capsule and English reading score and FA in the corona radiata (Qiu et al., 2008). In the arcuate fasciculus lateralization tractography study discussed earlier, greater leftward lateralization was associated with better performance on cognitive tests of receptive vocabulary and phonological processing (Lebel and Beaulieu, 2009). Response inhibition, which is characterized by cognitive control of thoughts, actions, and emotions, is hypothesized to be mediated by specific white matter networks of the inferior frontal gyrus, presupplementary motor cortex, and subthalamic nuclei. In a study by Madsen et al. (2010), stop-signal reaction time, which serves as a proxy for response inhibition, was significantly, negatively associated with the inferior frontal gyri

Structural brain development: birth through adolescence Chapter | 14

309

networks. These studies suggest not only that diffusion imaging is a useful technique for tracking normal anatomical maturation within the white matter but also that regional DTI metrics can provide reflections of cognitive development in specific domains.

14.7 Conclusions and future directions Our understanding of human brain development has accelerated over the past 30 years through the use of MRI and in vivo human brain mapping. Postmortem and histological studies have demonstrated that brain maturation, on the cellular level, encompasses both progressive and regressive events. These include synaptic pruning and protracted myelination, which continue to shape the underlying neural microstructure and regional brain morphology long after overall brain volume begins to plateau around age 5 years. Brain development, in general, can be characterized as both nonlinear with respect to time and also variable with respect to region. The hallmark of structural brain development during childhood is a striking change in the relative proportions of gray and white matterdwith a peak and then decline in gray matter volume and cortical thickness but a relatively sustained increase in white matter beyond adolescence. Across these different regions, there is a general posterior-to-anterior and inferior-to-superior trend in the timing of maturation, such that primary somatosensory and phylogenetically older areas of the brain tend to mature earlier than higher-order association corticesdparticularly areas in the frontal lobe. Within the white matter, diffusion imaging indicators show decreasing diffusivity (MD) and increasing directionality (FA), which suggests that myelination continues through young adulthood and perhaps even beyond. Performance across a variety of cognitive domains has also been shown to relate to these structural changes, with the specificity of these relationships generally in line with classic functional neuroanatomical localizations. Although the complexity of the regional and temporal patterns of structural brain development makes investigating and interpreting these brainebehavior relationships challenging, future work should continue to focus on the possible functional manifestations of structural brain development. Particularly, by integrating different structural and functional imaging modalities with thorough cognitive assessments, we can investigate the ways in which these processes interact with each other within a more inclusive framework that more realistically encompasses the full developmental landscape. With the increasingly broad array of radiological features of development that have been characterized, there is additionally a growing need to reintegrate a firm neurobiological understanding of the cellular mechanisms that facilitate these changes. Finally, effort should continue to be directed toward uncovering the ways in which this basic neuroscientific knowledge concerning human brain development can be translated into a better context for the understanding and clinical treatment of neurodevelopmental disorders.

References Adams, E., Chau, V., Poskitt, K.J., Grunau, R.E., Synnes, A., Miller, S.P., 2010. Tractography-based quantitation of corticospinal tract development in premature newborns. J. Pediatr. 156 (6), 882e888. Aleman-Gomez, Y., Janssen, J., Schnack, H., Balaban, E., Pina-Camacho, L., Alfaro-Almagro, F., et al., 2013. The human cerebral cortex flattens during adolescence. J. Neurosci. 33 (38), 15004e15010. https://doi.org/10.1523/JNEUROSCI.1459-13.2013. Allen, J.S., Damasio, H., Grabowski, T.J., Bruss, J., Zhang, W., 2003. Sexual dimorphism and asymmetries in the grayewhite composition of the human cerebrum. Neuroimage 18 (4), 880e894. Asato, M.R., Terwilliger, R., Woo, J., Luna, B.S., 2010. White matter development in adolescence: a DTI study. Cerebr. Cortex 20 (9), 2122e2131. Ashburner, J., Friston, K.J., 2000. Voxel-based morphometrydthe methods. Neuroimage 11 (6), 805e821. Ashburner, J., Csernansk, J.G., Davatzikos, C., Fox, N.C., Frisoni, G.B., Thompson, P.M., 2003. Computer-assisted imaging to assess brain structure in healthy and diseased brains. Lancet Neurol. 2 (2), 79e88. Ashtari, M., Cervellione, K.L., Hasan, K.M., Wu, J., McIlree, C., Kester, H., et al., 2007. White matter development during late adolescence in healthy males: a cross-sectional diffusion tensor imaging study. Neuroimage 35 (2), 501e510. Aubert-Broche, B., Fonov, V.S., Garcia-Lorenzo, D., Mouiha, A., Guizard, N., Coupe, P., et al., 2013. A new method for structural volume analysis of longitudinal brain MRI data and its application in studying the growth trajectories of anatomical brain structures in childhood. Neuroimage 82, 393e402. https://doi.org/10.1016/j.neuroimage.2013.05.065. Barkovich, A.J., Kjos, B.O., Jackson Jr., D.E., Norman, D., 1988. Normal maturation of the neonatal and infant brain: MR imaging at 1.5 T. Radiology 166 (1), 173e180. Barkovich, A.J., 2006. MR imaging of the neonatal brain. Neuroimaging Clin. 16 (1), 117e135. Basser, P.J., Mattiello, J., LeBihan, D., 1994. MR diffusion tensor spectroscopy and imaging. Biophys. J. 66 (1), 259e267. Bava, S., Thayer, R., Jacobus, J., Ward, M., Jernigan, T.L., Tapert, S.F., 2010. Longitudinal characterization of white matter maturation during adolescence. Brain Res. 1327, 38e46.

310 PART | II Cognitive development

Beaulieu, C., 2002. The basis of anisotropic water diffusion in the nervous systemea technical review. NMR Biomed. 15 (7-8), 435e455. Behrens, T.E., Johansen-Berg, H., Woolrich, M.W., Smith, S.M., Wheeler-Kingshott, C.A.M., Boulby, P.A.,., Thompson, A.J., 2003. Non-invasive mapping of connections between human thalamus and cortex using diffusion imaging. Nat. Neurosci. 6 (7), 750. Benes, F.M., Turtle, M., Khan, Y., Farol, P., 1994. Myelination of a key relay zone in the hippocampal formation occurs in the human brain during childhood, adolescence, and adulthood. Arch. Gen. Psychiatr. 51 (6), 477e484. Bengtsson, S.L., Nagy, Z., Skare, S., Forsman, L., Forssberg, H., Ullén, F., 2005. Extensive piano practicing has regionally specific effects on white matter development. Nat. Neurosci. 8 (9), 1148. Blinkhorn, S., 2005. Intelligence: a gender bender. Nature 438 (7064), 31. Bonekamp, D., Nagae, L.M., Degaonkar, M., Matson, M., Abdalla, W.M., Barker, P.B., et al., 2007. Diffusion tensor imaging in children and adolescents: reproducibility, hemispheric, and age-related differences. Neuroimage 34 (2), 733e742. Bookheimer, S., 2002. Functional MRI of language: new approaches to understanding the cortical organization of semantic processing. Annu. Rev. Neurosci. 25 (1), 151e188. Bramen, J.E., Hranilovich, J.A., Dahl, R.E., Forbes, E.E., Chen, J., Toga, A.W., et al., 2010. Puberty influences medial temporal lobe and cortical gray matter maturation differently in boys than girls matched for sexual maturity. Cerebr. Cortex 21 (3), 636e646. Brouwer, R.M., Mandl, R.C., Schnack, H.G., van Soelen, I.L., van Baal, G.C., Peper, J.S., et al., 2012. White matter development in early puberty: a longitudinal volumetric and diffusion tensor imaging twin study. PloS One 7 (4), e32316. https://doi.org/10.1371/journal.pone.0032316. Brown, T.T., Kuperman, J.M., Chung, Y., Erhart, M., McCabe, C., Hagler Jr., D.J., et al., 2012. Neuroanatomical assessment of biological maturity. Curr. Biol. 22 (18), 1693e1698. Burzynska, A.Z., Preuschhof, C., Bäckman, L., Nyberg, L., Li, S.C., Lindenberger, U., Heekeren, H.R., 2010. Age-related differences in white matter microstructure: region-specific patterns of diffusivity. Neuroimage 49 (3), 2104e2112. Carreiras, M., Seghier, M.L., Baquero, S., Estévez, A., Lozano, A., Devlin, J.T., Price, C.J., 2009. An anatomical signature for literacy. Nature 461 (7266), 983. Casey, B.J., Cohen, J.D., Jezzard, P., Turner, R., Noll, D.C., Trainor, R.J., et al., 1995. Activation of prefrontal cortex in children during a nonspatial working memory task with functional MRI. Neuroimage 2 (3), 221e229. Casey, B.J., Tottenham, N., Liston, C., Durston, S., 2005. Imaging the developing brain: what have we learned about cognitive development? Trends Cognit. Sci. 9 (3), 104e110. Catani, M., Howard, R.J., Pajevic, S., Jones, D.K., 2002. Virtual in vivo interactive dissection of white matter fasciculi in the human brain. Neuroimage 17 (1), 77e94. Caviness Jr., V.S., Kennedy, D.N., Richelme, C., Rademacher, J.F.P.A., Filipek, P.A., 1996. The human brain age 7e11 years: a volumetric analysis based on magnetic resonance images. Cerebr. Cortex 6 (5), 726e736. Chahal, R., Vilgis, V., Grimm, K.J., Hipwell, A.E., Forbes, E.E., Keenan, K., Guyer, A.E., 2018. Girls’ pubertal development is associated with white matter microstructure in late adolescence. Neuroimage 181, 659e669. Chiang, M.C., Leow, A.D., Klunder, A.D., Dutton, R.A., Barysheva, M., Rose, S.E., et al., 2008. Fluid registration of diffusion tensor images using information theory. IEEE Trans. Med. Imag. 27 (4), 442e456. Choi, Y.Y., Shamosh, N.A., Cho, S.H., DeYoung, C.G., Lee, M.J., Lee, J.M., et al., 2008. Multiple bases of human intelligence revealed by cortical thickness and neural activation. J. Neurosci. 28 (41), 10323e10329. Colby, J.B., Van Horn, J.D., Sowell, E.R., 2011. Quantitative in vivo evidence for broad regional gradients in the timing of white matter maturation during adolescence. Neuroimage 54 (1), 25e31. https://doi.org/10.1016/j.neuroimage.2010.08.014. Colby, J.B., Soderberg, L., Lebel, C., Dinov, I.D., Thompson, P.M., Sowell, E.R., 2012. Along-tract statistics allow for enhanced tractography analysis. Neuroimage 59 (4), 3227e3242. https://doi.org/10.1016/j.neuroimage.2011.11.004. Conturo, T.E., Lori, N.F., Cull, T.S., Akbudak, E., Snyder, A.Z., Shimony, J.S., et al., 1999. Tracking neuronal fiber pathways in the living human brain. Proc. Natl. Acad. Sci. U.S.A. 96 (18), 10422e10427. Cooke, B.M., 2006. Steroid-dependent plasticity in the medial amygdala. Neuroscience 138 (3), 997e1005. Courchesne, E., Chisum, H.J., Townsend, J., Cowles, A., Covington, J., Egaas, B., et al., 2000. Normal brain development and aging: quantitative analysis at in vivo MR imaging in healthy volunteers. Radiology 216 (3), 672e682. Dahl, R.E., 2004. Adolescent brain development: a period of vulnerabilities and opportunities. Keynote address. Ann. N. Y. Acad. Sci. 1021 (1), 1e22. Damasio, H., Grabowski, T., Frank, R., Galaburda, A.M., Damasio, A.R., 1994. The return of Phineas Gage: clues about the brain from the skull of a famous patient. Science 264 (5162), 1102e1105. Davis, S.W., Dennis, N.A., Buchler, N.G., White, L.E., Madden, D.J., Cabeza, R., 2009. Assessing the effects of age on long white matter tracts using diffusion tensor tractography. Neuroimage 46 (2), 530e541. De Bellis, M.D., Keshavan, M.S., Beers, S.R., Hall, J., Frustaci, K., Masalehdan, A., et al., 2001. Sex differences in brain maturation during childhood and adolescence. Cerebr. Cortex 11 (6), 552e557. Dekaban, A.S., Sadowsky, D., 1978. Changes in brain weights during the span of human life: relation of brain weights to body heights and body weights. Ann. Neurol. 4 (4), 345e356. Draganski, B., Gaser, C., Busch, V., Schuierer, G., Bogdahn, U., May, A., 2004. Neuroplasticity: changes in grey matter induced by training. Nature 427 (6972), 311. Durston, S., Pol, H.E.H., Casey, B.J., Giedd, J.N., Buitelaar, J.K., Van Engeland, H., 2001. Anatomical MRI of the developing human brain: what have we learned? J. Am. Acad. Child Adolesc. Psychiatr. 40 (9), 1012e1020.

Structural brain development: birth through adolescence Chapter | 14

311

Ducharme, S., Albaugh, M.D., Nguyen, T.V., Hudziak, J.J., Mateos-Perez, J.M., Labbe, A., et al., 2015. Trajectories of cortical surface area and cortical volume maturation in normal brain development. Data Brief 5, 929e938. https://doi.org/10.1016/j.dib.2015.10.044. Eliez, S., Reiss, A.L., 2000. Annotation: MRI neuroimaging of childhood psychiatric disorders: a selective review. J. Child Psychol. Psychiatry Allied Discip. 41 (6), 679e694. Engelbrecht, V., Rassek, M., Preiss, S., Wald, C., Mödder, U., 1998. Age-dependent changes in magnetization transfer contrast of white matter in the pediatric brain. Am. J. Neuroradiol. 19 (10), 1923e1929. Engelbrecht, V., Scherer, A., Rassek, M., Witsack, H.J., Mödder, U., 2002. Diffusion-weighted MR imaging in the brain in children: findings in the normal brain and in the brain with white matter diseases. Radiology 222 (2), 410e418. Fischl, B., Dale, A.M., 2000. Measuring the thickness of the human cerebral cortex from magnetic resonance images. Proc. Natl. Acad. Sci. U.S.A. 97 (20), 11050e11055. Fjell, A.M., Westlye, L.T., Greve, D.N., Fischl, B., Benner, T., van der Kouwe, A.J., et al., 2008. The relationship between diffusion tensor imaging and volumetry as measures of white matter properties. Neuroimage 42 (4), 1654e1668. Fjell, A.M., Walhovd, K.B., Brown, T.T., Kuperman, J.M., Chung, Y., Hagler Jr., D.J., et al., 2012. Multimodal imaging of the self-regulating developing brain. Proc. Natl. Acad. Sci. U.S.A. 109 (48), 19620e19625. https://doi.org/10.1073/pnas.1208243109. Frangou, S., Chitins, X., Williams, S.C., 2004. Mapping IQ and gray matter density in healthy young people. Neuroimage 23 (3), 800e805. Fuster, J.M., 2002. The Prefrontal Cortex: Anatomy, Physiology, and Neuropsychology of the Frontal Lobe. Lippincott-Raven, Philadelphia. Galea, L.A., Spritzer, M.D., Barker, J.M., Pawluski, J.L., 2006. Gonadal hormone modulation of hippocampal neurogenesis in the adult. Hippocampus 16 (3), 225e232. Geschwind, N., Levitsky, W., 1968. Human brain: left-right asymmetries in temporal speech region. Science 161 (3837), 186e187. Giedd, J.N., Rumsey, J.M., Castellanos, F.X., Rajapakse, J.C., Kaysen, D., Vaituzis, A.C., et al., 1996. A quantitative MRI study of the corpus callosum in children and adolescents. Dev. Brain Res. 91 (2), 274e280. Giedd, J.N., Snell, J.W., Lange, N., Rajapakse, J.C., Casey, B.J., Kozuch, P.L., et al., 1996. Quantitative magnetic resonance imaging of human brain development: ages 4e18. Cerebr. Cortex 6 (4), 551e559. Giedd, J.N., Vaituzis, A.C., Hamburger, S.D., Lange, N., Rajapakse, J.C., Kaysen, D., et al., 1996. Quantitative MRI of the temporal lobe, amygdala, and hippocampus in normal human development: ages 4e18 years. J. Comp. Neurol. 366 (2), 223e230. Giedd, J.N., Castellanos, F.X., Rajapakse, J.C., Vaituzis, A.C., Rapoport, J.L., 1997. Sexual dimorphism of the developing human brain. Prog. Neuro Psychopharmacol. Biol. Psychiatr. 21 (8), 1185e1201. Giedd, J.N., Blumenthal, J., Jeffries, N.O., Castellanos, F.X., Liu, H., Zijdenbos, A., et al., 1999. Brain development during childhood and adolescence: a longitudinal MRI study. Nat. Neurosci. 2 (10), 861. Giedd, J.N., Blumenthal, J., Jeffries, N.O., Rajapakse, J.C., Vaituzis, A.C., Liu, H., et al., 1999. Development of the human corpus callosum during childhood and adolescence: a longitudinal MRI study. Prog. Neuro Psychopharmacol. Biol. Psychiatr. 23 (4), 571e588. Giedd, J.N., Clasen, L.S., Lenroot, R., Greenstein, D., Wallace, G.L., Ordaz, S., et al., 2006. Puberty-related influences on brain development. Mol. Cell. Endocrinol. 254, 154e162. Giedd, J.N., Raznahan, A., Mills, K.L., Lenroot, R.K., 2012. Review: magnetic resonance imaging of male/female differences in human adolescent brain anatomy. Biol. Sex Differ. 3 (19). https://doi.org/10.1186/2042-6410-3-19. Giorgio, A., Watkins, K.E., Douaud, G., James, A.C., James, S., De Stefano, N., et al., 2008. Changes in white matter microstructure during adolescence. Neuroimage 39 (1), 52e61. Giorgio, A., Watkins, K.E., Chadwick, M., James, S., Winmill, L., Douaud, G., et al., 2010. Longitudinal changes in grey and white matter during adolescence. Neuroimage 49 (1), 94e103. Gogtay, N., Giedd, J.N., Lusk, L., Hayashi, K.M., Greenstein, D., Vaituzis, A.C., et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. U.S.A. 101 (21), 8174e8179. Goldstein, J.M., Seidman, L.J., Horton, N.J., Makris, N., Kennedy, D.N., Caviness Jr., V.S., et al., 2001. Normal sexual dimorphism of the adult human brain assessed by in vivo magnetic resonance imaging. Cerebr. Cortex 11 (6), 490e497. Gould, S.J., 1978. Morton’s ranking of races by cranial capacity. Unconscious manipulation of data may be a scientific norm. Science 200 (4341), 503e509. Gould, S., 1981. The Mismeasure of Man, first ed. Norton, New York. Gur, R.C., Gunning-Dixon, F., Bilker, W.B., Gur, R.E., 2002. Sex differences in temporo-limbic and frontal brain volumes of healthy adults. Cerebr. Cortex 12 (9), 998e1003. Haier, R.J., Jung, R.E., Yeo, R.A., Head, K., Alkire, M.T., 2004. Structural brain variation and general intelligence. Neuroimage 23 (1), 425e433. Haier, R.J., Jung, R.E., Yeo, R.A., Head, K., Alkire, M.T., 2005. The neuroanatomy of general intelligence: sex matters. Neuroimage 25 (1), 320e327. Haug, H., 1987. Brain sizes, surfaces, and neuronal sizes of the cortex cerebri: a stereological investigation of man and his variability and a comparison with some mammals (primates, whales, marsupials, insectivores, and one elephant). Am. J. Anat. 180 (2), 126e142. Hedges, L.V., Nowell, A., 1995. Sex differences in mental test scores, variability, and numbers of high-scoring individuals. Science 269 (5220), 41e45. Hermoye, L., Saint-Martin, C., Cosnard, G., Lee, S.K., Kim, J., Nassogne, M.C., et al., 2006. Pediatric diffusion tensor imaging: normal database and observation of the white matter maturation in early childhood. Neuroimage 29 (2), 493e504. Herting, M.M., Kim, R., Uban, K.A., Kan, E., Binley, A., Sowell, E.R., 2017. Longitudinal changes in pubertal maturation and white matter microstructure. Psychoneuroendocrinology 81, 70e79. https://doi.org/10.1016/j.psyneuen.2017.03.017.

312 PART | II Cognitive development

Herting, M.M., Johnson, C., Mills, K.L., Vijayakumar, N., Dennison, M., Liu, C., et al., 2018. Development of subcortical volumes across adolescence in males and females: a multisample study of longitudinal changes. Neuroimage 172, 194e205. https://doi.org/10.1016/j.neuroimage.2018.01.020. Hsu, J.L., Van Hecke, W., Bai, C.H., Lee, C.H., Tsai, Y.F., Chiu, H.C., et al., 2010. Microstructural white matter changes in normal aging: a diffusion tensor imaging study with higher-order polynomial regression models. Neuroimage 49 (1), 32e43. Huang, H., Zhang, J., Wakana, S., Zhang, W., Ren, T., Richards, L.J., et al., 2006. White and gray matter development in human fetal, newborn and pediatric brains. Neuroimage 33 (1), 27e38. Huang, H., Xue, R., Zhang, J., Ren, T., Richards, L.J., Yarowsky, P., et al., 2009. Anatomical characterization of human fetal brain development with diffusion tensor magnetic resonance imaging. J. Neurosci. 29 (13), 4263e4273. Huang, H., 2010. Structure of the fetal brain: what we are learning from diffusion tensor imaging. Neuroscientist 16 (6), 634e649. Huang, H., Shu, N., Mishra, V., Jeon, T., Chalak, L., Wang, Z.J., et al., 2015. Development of human brain structural networks through infancy and childhood. Cerebr. Cortex 25 (5), 1389e1404. https://doi.org/10.1093/cercor/bht335. Hüppi, P.S., Maier, S.E., Peled, S., Zientara, G.P., Barnes, P.D., Jolesz, F.A., Volpe, J.J., 1998. Microstructural development of human newborn cerebral white matter assessed in vivo by diffusion tensor magnetic resonance imaging. Pediatr. Res. 44 (4), 584. Huttenlocher, P.R., 1979. Synaptic density in human frontal cortex-developmental changes and effects of aging. Brain Res. 163 (2), 195e205. Huttenlocher, P.R., De, C.C., Garey, L.J., der Loos Van, H., 1982. Synaptic development in human cerebral cortex. Int. J. Neurol. 16, 144e154. Huttenlocher, P.R., Dabholkar, A.S., 1997. Regional differences in synaptogenesis in human cerebral cortex. J. Comp. Neurol. 387 (2), 167e178. Im, K., Lee, J.M., Lee, J., Shin, Y.W., Kim, I.Y., Kwon, J.S., Kim, S.I., 2006. Gender difference analysis of cortical thickness in healthy young adults with surface-based methods. Neuroimage 31 (1), 31e38. Irwing, P., Lynn, R., 2006. Intelligence: is there a sex difference in IQ scores? Nature 442 (7098), E1. Jeon, T., Mishra, V., Ouyang, M., Chen, M., Huang, H., 2015. Synchronous changes of cortical thickness and corresponding white matter microstructure during brain development accessed by diffusion MRI tractography from parcellated cortex. Front. Neuroanat. 9, 158. https://doi.org/10.3389/ fnana.2015.00158. Jensen, J.H., Helpern, J.A., 2010. MRI quantification of non-Gaussian water diffusion by kurtosis analysis. NMR Biomed. 23 (7), 698e710. Jernigan, T.L., Tallal, P., 1990. Late childhood changes in brain morphology observable with MRI. Dev. Med. Child Neurol. 32 (5), 379e385. Jernigan, T.L., Archibald, S.L., Berhow, M.T., Sowell, E.R., Foster, D.S., Hesselink, J.R., 1991. Cerebral structure on MRI, Part I: localization of agerelated changes. Biol. Psychiatr. 29 (1), 55e67. Jernigan, T.L., Trauner, D.A., Hesselink, J.R., Tallal, P.A., 1991. Maturation of human cerebrum observed in vivo during adolescence. Brain 114 (5), 2037e2049. Jones, D.K., Cercignani, M., 2010. Twenty-five pitfalls in the analysis of diffusion MRI data. NMR Biomed. 23 (7), 803e820. Jorm, A.F., Anstey, K.J., Christensen, H., Rodgers, B., 2004. Gender differences in cognitive abilities: the mediating role of health state and health habits. Intelligence 32 (1), 7e23. Kaes, T., 1907. Die Grosshirnrinde des Menschen in ihern Massen und in iherm Fasergehalt. Gustav Fischer, Jena, Germany. Kandel, E.R., Schwartz, J.H., Jessell, T.M., Siegelbaum, S.A., Hudspeth, A.J., 2013. Principles of Neural Science, fifth ed. McGraw-Hill, New York. Karama, S., Ad-Dab’bagh, Y., Haier, R.J., Deary, I.J., Lyttelton, O.C., Lepage, C., et al., 2009. Erratum to “Positive association between cognitive ability and cortical thickness in a representative US sample of healthy 6 to 18 year-olds”. Intelligence 37 (4), 432e442. Kemper, T.L., 1994. Neuroanatomical and neuropathological changes during aging and dementia. In: Albert, M.L., Knoefel, J.E. (Eds.), Clinical Neurology of Aging. Oxford University press, New York, NY, pp. 8e9. Kimura, D., 1996. Sex, sexual orientation and sex hormones influence human cognitive function. Curr. Opin. Neurobiol. 6 (2), 259e263. Kinney, H.C., Brody, B.A., Kloman, A.S., Gilles, F.H., 1988. Sequence of central nervous system myelination in human infancy: II. Patterns of myelination in autopsied infants. J. Neuropathol. Exp. Neurol. 47 (3), 217e234. Klingberg, T., Vaidya, C.J., Gabrieli, J.D., Moseley, M.E., Hedehus, M., 1999. Myelination and organization of the frontal white matter in children: a diffusion tensor MRI study. Neuroreport 10 (13), 2817e2821. Krogsrud, S.K., Fjell, A.M., Tamnes, C.K., Grydeland, H., Mork, L., Due-Tonnessen, P., et al., 2016. Changes in white matter microstructure in the developing brain–A longitudinal diffusion tensor imaging study of children from 4 to 11years of age. Neuroimage 124 (Pt A), 473e486. https:// doi.org/10.1016/j.neuroimage.2015.09.017. Lancaster, J.L., Andrews, T., Hardies, L.J., Dodd, S., Fox, P.T., 2003. Three-pool model of white matter. J. Magn. Reson. Imag. 17 (1), 1e10. Le Bihan, D., Breton, E., Lallemand, D., Grenier, P., Cabanis, E., Laval-Jeantet, M., 1986. MR imaging of intravoxel incoherent motions: application to diffusion and perfusion in neurologic disorders. Radiology 161 (2), 401e407. Le Bihan, D., 2003. Looking into the functional architecture of the brain with diffusion MRI. Nat. Rev. Neurosci. 4 (6), 469. Lebel, C., Rasmussen, C., Wyper, K., Walker, L., Andrew, G., Yager, J., Beaulieu, C., 2008. Brain diffusion abnormalities in children with fetal alcohol spectrum disorder. Alcohol Clin. Exp. Res. 32 (10), 1732e1740. Lebel, C., Walker, L., Leemans, A., Phillips, L., Beaulieu, C., 2008. Microstructural maturation of the human brain from childhood to adulthood. Neuroimage 40 (3), 1044e1055. Lebel, C., Beaulieu, C., 2009. Lateralization of the arcuate fasciculus from childhood to adulthood and its relation to cognitive abilities in children. Hum. Brain Mapp. 30 (11), 3563e3573. Lebel, C., Caverhill-Godkewitsch, S., Beaulieu, C., 2010. Age-related regional variations of the corpus callosum identified by diffusion tensor tractography. Neuroimage 52 (1), 20e31.

Structural brain development: birth through adolescence Chapter | 14

313

Lebel, C., Beaulieu, C., 2011. Longitudinal development of human brain wiring continues from childhood into adulthood. J. Neurosci. 31 (30), 10937e10947. https://doi.org/10.1523/JNEUROSCI.5302-10.2011. Lenroot, R.K., Gogtay, N., Greenstein, D.K., Wells, E.M., Wallace, G.L., Clasen, L.S., et al., 2007. Sexual dimorphism of brain developmental trajectories during childhood and adolescence. Neuroimage 36 (4), 1065e1073. Lenroot, R.K., Giedd, J.N., 2010. Sex differences in the adolescent brain. Brain Cognit. 72 (1), 46e55. Liston, C., Watts, R., Tottenham, N., Davidson, M.C., Niogi, S., Ulug, A.M., Casey, B.J., 2005. Frontostriatal microstructure modulates efficient recruitment of cognitive control. Cerebr. Cortex 16 (4), 553e560. Liu, Y., Balériaux, D., Kavec, M., Metens, T., Absil, J., Denolin, V.,., Aeby, A., 2010. Structural asymmetries in motor and language networks in a population of healthy preterm neonates at term equivalent age: a diffusion tensor imaging and probabilistic tractography study. Neuroimage 51 (2), 783e788. Löbel, U., Sedlacik, J., Güllmar, D., Kaiser, W.A., Reichenbach, J.R., Mentzel, H.J., 2009. Diffusion tensor imaging: the normal evolution of ADC, RA, FA, and eigenvalues studied in multiple anatomical regions of the brain. Neuroradiology 51 (4), 253e263. Luders, E., Narr, K.L., Thompson, P.M., Rex, D.E., Jancke, L., Steinmetz, H., Toga, A.W., 2004. Gender differences in cortical complexity. Nat. Neurosci. 7 (8), 799. Luna, B., Garver, K.E., Urban, T.A., Lazar, N.A., Sweeney, J.A., 2004. Maturation of cognitive processes from late childhood to adulthood. Child Dev. 75 (5), 1357e1372. Luna, B., Padmanabhan, A., O’Hearn, K., 2010. What has fMRI told us about the development of cognitive control through adolescence? Brain Cognit. 72 (1), 101e113. Lynn, R., Irwing, P., 2004. Sex differences on the progressive matrices: a meta-analysis. Intelligence 32 (5), 481e498. Mackay, A., Whittall, K., Adler, J., Li, D., Paty, D., Graeb, D., 1994. In vivo visualization of myelin water in brain by magnetic resonance. Magn. Reson. Med. 31 (6), 673e677. Madsen, K.S., Baare, W.F., Vestergaard, M., Skimminge, A., Ejersbo, L.R., Ramsoy, T.Z., et al., 2010. Response inhibition is associated with white matter microstructure in children. Neuropsychologia 48 (4), 854e862. https://doi.org/10.1016/j.neuropsychologia.2009.11.001. Marsh, R., Gerber, A.J., Peterson, B.S., 2008. Neuroimaging studies of normal brain development and their relevance for understanding childhood neuropsychiatric disorders. J. Am. Acad. Child Adolesc. Psychiatr. 47 (11), 1233e1251. McKinstry, R.C., Mathur, A., Miller, J.H., Ozcan, A., Snyder, A.Z., Schefft, G.L., et al., 2002. Radial organization of developing preterm human cerebral cortex revealed by non-invasive water diffusion anisotropy MRI. Cerebr. Cortex 12 (12), 1237e1243. Menzies, L., Goddings, A.L., Whitaker, K.J., Blakemore, S.J., Viner, R.M., 2015. The effects of puberty on white matter development in boys. Dev. Cogn. Neurosci. 11, 116e128. https://doi.org/10.1016/j.dcn.2014.10.002. Miller, S.P., Ferriero, D.M., 2009. From selective vulnerability to connectivity: insights from newborn brain imaging. Trends Neurosci. 32 (9), 496e505. Mills, K.L., Goddings, A.L., Herting, M.M., Meuwese, R., Blakemore, S.J., Crone, E.A., et al., 2016. Structural brain development between childhood and adulthood: convergence across four longitudinal samples. Neuroimage 141, 273e281. https://doi.org/10.1016/j.neuroimage.2016.07.044. Mishra, V., Cheng, H., Gong, G., He, Y., Dong, Q., Huang, H., 2013. Differences of inter-tract correlations between neonates and children around puberty: a study based on microstructural measurements with DTI. Front. Hum. Neurosci. 7, 721. https://doi.org/10.3389/fnhum.2013.00721. Mori, S., Crain, B.J., Chacko, V.P., Van Zijl, P.C., 1999. Three-dimensional tracking of axonal projections in the brain by magnetic resonance imaging. Ann. Neurol. 45 (2), 265e269. Mori, S., Oishi, K., Jiang, H., Jiang, L., Li, X., Akhter, K., et al., 2008. Stereotaxic white matter atlas based on diffusion tensor imaging in an ICBM template. Neuroimage 40 (2), 570e582. Morriss, M.C., Zimmerman, R.A., Bilaniuk, L.T., Hunter, J.V., Haselgrove, J.C., 1999. Changes in brain water diffusion during childhood. Neuroradiology 41 (12), 929e934. Mukherjee, P., Miller, J.H., Shimony, J.S., Conturo, T.E., Lee, B.C., Almli, C.R., McKinstry, R.C., 2001. Normal brain maturation during childhood: developmental trends characterized with diffusion-tensor MR imaging. Radiology 221 (2), 349e358. Mutlu, A.K., Schneider, M., Debbane, M., Badoud, D., Eliez, S., Schaer, M., 2013. Sex differences in thickness, and folding developments throughout the cortex. Neuroimage 82, 200e207. https://doi.org/10.1016/j.neuroimage.2013.05.076. Nagy, Z., Westerberg, H., Klingberg, T., 2004. Maturation of white matter is associated with the development of cognitive functions during childhood. J. Cognit. Neurosci. 16 (7), 1227e1233. Narr, K.L., Woods, R.P., Thompson, P.M., Szeszko, P., Robinson, D., Dimtcheva, T., et al., 2006. Relationships between IQ and regional cortical gray matter thickness in healthy adults. Cerebr. Cortex 17 (9), 2163e2171. Neil, J.J., Shiran, S.I., McKinstry, R.C., Schefft, G.L., Snyder, A.Z., Almli, C.R., et al., 1998. Normal brain in human newborns: apparent diffusion coefficient and diffusion anisotropy measured by using diffusion tensor MR imaging. Radiology 209 (1), 57e66. Neisser, U., Boodoo, G., Bouchard Jr., T.J., Boykin, A.W., Brody, N., Ceci, S.J., et al., 1996. Intelligence: knowns and unknowns. Am. Psychol. 51 (2), 77. Neufang, S., Specht, K., Hausmann, M., Güntürkün, O., Herpertz-Dahlmann, B., Fink, G.R., Konrad, K., 2008. Sex differences and the impact of steroid hormones on the developing human brain. Cerebr. Cortex 19 (2), 464e473. Nopoulos, P., Flaum, M., O’Leary, D., Andreasen, N.C., 2000. Sexual dimorphism in the human brain: evaluation of tissue volume, tissue composition and surface anatomy using magnetic resonance imaging. Psychiatr. Res. Neuroimaging 98 (1), 1e13. O’donnell, S., Noseworthy, M.D., Levine, B., Dennis, M., 2005. Cortical thickness of the frontopolar area in typically developing children and adolescents. Neuroimage 24 (4), 948e954.

314 PART | II Cognitive development

Østby, Y., Tamnes, C.K., Fjell, A.M., Westlye, L.T., Due-Tønnessen, P., Walhovd, K.B., 2009. Heterogeneity in subcortical brain development: a structural magnetic resonance imaging study of brain maturation from 8 to 30 years. J. Neurosci. 29 (38), 11772e11782. Pakkenberg, B., Gundersen, H.J.G., 1997. Neocortical neuron number in humans: effect of sex and age. J. Comp. Neurol. 384 (2), 312e320. Panigrahy, A., Blüml, S., 2009. Neuroimaging of pediatric brain tumors: from basic to advanced magnetic resonance imaging (MRI). J. Child Neurol. 24 (11), 1343e1365. Paus, T., Zijdenbos, A., Worsley, K., Collins, D.L., Blumenthal, J., Giedd, J.N., et al., 1999. Structural maturation of neural pathways in children and adolescents: in vivo study. Science 283 (5409), 1908e1911. Paus, T., Collins, D.L., Evans, A.C., Leonard, G., Pike, B., Zijdenbos, A., 2001. Maturation of white matter in the human brain: a review of magnetic resonance studies. Brain Res. Bull. 54 (3), 255e266. Paus, T., Nawaz-Khan, I., Leonard, G., Perron, M., Pike, G.B., Pitiot, A., et al., 2010. Sexual dimorphism in the adolescent brain: role of testosterone and androgen receptor in global and local volumes of grey and white matter. Horm. Behav. 57 (1), 63e75. Paus, T., Wong, A.P.-Y., Syme, C., Pausova, Z., 2017. Sex differences in the adolescent brain and body: findings from the saguenay youth study. J. Neurosci. Res. 95, 362e370. https://doi.org/10.1002/jnr.23825. Peper, J.S., Brouwer, R.M., Schnack, H.G., van Baal, G.C.M., van Leeuwen, M., van den Berg, S.M., et al., 2008. Cerebral white matter in early puberty is associated with luteinizing hormone concentrations. Psychoneuroendocrinology 33 (7), 909e915. Peper, J.S., Brouwer, R.M., Schnack, H.G., van Baal, G.C., van Leeuwen, M., van den Berg, S.M., et al., 2009. Sex steroids and brain structure in pubertal boys and girls. Psychoneuroendocrinology 34 (3), 332e342. Peper, J.S., Schnack, H.G., Brouwer, R.M., Van Baal, G.C.M., Pjetri, E., Szekely, E., et al., 2009. Heritability of regional and global brain structure at the onset of puberty: a magnetic resonance imaging study in 9-year-old twin pairs. Hum. Brain Mapp. 30 (7), 2184e2196. Perrin, J.S., Hervé, P.Y., Leonard, G., Perron, M., Pike, G.B., Pitiot, A., et al., 2008. Growth of white matter in the adolescent brain: role of testosterone and androgen receptor. J. Neurosci. 28 (38), 9519e9524. Peters, M., Jäncke, L., Staiger, J.F., Schlaug, G., Huang, Y., Steinmetz, H., 1998. Unsolved problems in comparing brain sizes in Homo sapiens. Brain Cognit. 37 (2), 254e285. Peters, B.D., Szeszko, R.R., Radua, J., et al., 2012. White matter development in adolescence: diffusion tensor imaging and meta-analytic results. Schizophr. Bull. 38 (6), 1308e1317. Pfefferbaum, A., Mathalon, D.H., Sullivan, E.V., Rawles, J.M., Zipursky, R.B., Lim, K.O., 1994. A quantitative magnetic resonance imaging study of changes in brain morphology from infancy to late adulthood. Arch. Neurol. 51 (9), 874e887. Pierpaoli, C., Jezzard, P., Basser, P.J., Barnett, A., Di Chiro, G., 1996. Diffusion tensor MR imaging of the human brain. Radiology 201 (3), 637e648. Porter, J.N., Collins, P.F., Muetzel, R.L., Lim, K.O., Luciana, M., 2011. Associations between cortical thickness and verbal fluency in childhood, adolescence, and young adulthood. Neuroimage 55 (4), 1865e1877. https://doi.org/10.1016/j.neuroimage.2011.01.018. Post, R.M., Weiss, S.R., 1997. Emergent properties of neural systems: how focal molecular neurobiological alterations can affect behavior. Dev. Psychopathol. 9 (4), 907e929. Prager, A., Roychowdhury, S., 2007. Magnetic resonance imaging of the neonatal brain. Indian J. Pediatr. 74 (2), 173e184. Pujol, J., Vendrell, P., Junqué, C., Martí-Vilalta, J.L., Capdevila, A., 1993. When does human brain development end? Evidence of corpus callosum growth up to adulthood. Ann. Neurol. 34 (1), 71e75. Pujol, J., López-Sala, A., Deus, J., Cardoner, N., Sebastián-Gallés, N., Conesa, G., Capdevila, A., 2002. The lateral asymmetry of the human brain studied by volumetric magnetic resonance imaging. Neuroimage 17 (2), 670e679. Qiu, D., Tan, L.H., Zhou, K., Khong, P.L., 2008. Diffusion tensor imaging of normal white matter maturation from late childhood to young adulthood: voxel-wise evaluation of mean diffusivity, fractional anisotropy, radial and axial diffusivities, and correlation with reading development. Neuroimage 41 (2), 223e232. Rabinowicz, T., Petetot, J.M.C., Gartside, P.S., Sheyn, D., Sheyn, T., de Courten-Myers, G.M., 2002. Structure of the cerebral cortex in men and women. J. Neuropathol. Exp. Neurol. 61 (1), 46e57. Raznahan, A., Lerch, J.P., Lee, N., Greenstein, D., Wallace, G.L., Stockman, M., et al., 2011. Patterns of coordinated anatomical change in human cortical development: a longitudinal neuroimaging study of maturational coupling. Neuron 72 (5), 873e884. https://doi.org/10.1016/j.neuron.2011.09.028. Raznahan, A., Shaw, P.W., Lerch, J.P., Clasen, L.S., Greenstein, D., Berman, R., et al., 2014. Longitudinal four-dimensional mapping of subcortical anatomy in human development. Proc. Natl. Acad. Sci. U.S.A. 111 (4), 1592e1597. https://doi.org/10.1073/pnas.1316911111. Re, T.J., Levman, J., Lim, A.R., Righini, A., Grant, P.E., Takahashi, E., 2017. High-angular resolution diffusion imaging tractography of cerebellar pathways from newborns to young adults. Brain Behav 7 (1), e00589. https://doi.org/10.1002/brb3.589. Reiss, A.L., Abrams, M.T., Singer, H.S., Ross, J.L., Denckla, M.B., 1996. Brain development, gender and IQ in children: a volumetric imaging study. Brain 119 (5), 1763e1774. Riddle, W.R., DonLevy, S.C., Lee, H., 2010. Modeling brain tissue volumes over the lifespan: quantitative analysis of postmortem weights and in vivo MR images. Magn. Reson. Imag. 28 (5), 716e720. Rilling, J.K., Glasser, M.F., Preuss, T.M., Ma, X., Zhao, T., Hu, X., Behrens, T.E., 2008. The evolution of the arcuate fasciculus revealed with comparative DTI. Nat. Neurosci. 11 (4), 426e428. Salat, D.H., Tuch, D.S., Hevelone, N.D., Fischl, B., Corkin, S., Rosas, H.D., Dale, A.M., 2005. Age-related changes in prefrontal white matter measured by diffusion tensor imaging. Ann. N. Y. Acad. Sci. 1064 (1), 37e49. Sarkey, S., Azcoitia, I., Garcia-Segura, L.M., Garcia-Ovejero, D., DonCarlos, L.L., 2008. Classical androgen receptors in non-classical sites in the brain. Horm. Behav. 53 (5), 753e764.

Structural brain development: birth through adolescence Chapter | 14

315

Schmithorst, V.J., Wilke, M., Dardzinski, B.J., Holland, S.K., 2002. Correlation of white matter diffusivity and anisotropy with age during childhood and adolescence: a cross-sectional diffusion-tensor MR imaging study. Radiology 222 (1), 212e218. Schmithorst, V.J., Holland, S.K., Dardzinski, B.J., 2008. Developmental differences in white matter architecture between boys and girls. Hum. Brain Mapp. 29 (6), 696e710. Schmithorst, V.J., Yuan, W., 2010. White matter development during adolescence as shown by diffusion MRI. Brain Cognit. 72 (1), 16e25. Schnack, H.G., van Haren, N.E., Brouwer, R.M., Evans, A., Durston, S., Boomsma, D.I., et al., 2015. Changes in thickness and surface area of the human cortex and their relationship with intelligence. Cerebr. Cortex 25 (6), 1608e1617. https://doi.org/10.1093/cercor/bht357. Schneider, J.F., Il’Yasov, K.A., Hennig, J., Martin, E., 2004. Fast quantitative diffusion-tensor imaging of cerebral white matter from the neonatal period to adolescence. Neuroradiology 46 (4), 258e266. Schneiderman, J.S., Buchsbaum, M.S., Haznedar, M.M., Hazlett, E.A., Brickman, A.M., Shihabuddin, L., et al., 2007. Diffusion tensor anisotropy in adolescents and adults. Neuropsychobiology 55 (2), 96e111. Shaw, P., Greenstein, D., Lerch, J., Clasen, L., Lenroot, R., Gogtay, N.E.E.A., et al., 2006. Intellectual ability and cortical development in children and adolescents. Nature 440 (7084), 676. Shaw, P., 2007. Intelligence and the developing human brain. Bioessays 29 (10), 962e973. Shaw, P., Kabani, N.J., Lerch, J.P., Eckstrand, K., Lenroot, R., Gogtay, N., et al., 2008. Neurodevelopmental trajectories of the human cerebral cortex. J. Neurosci. 28 (14), 3586e3594. Simmonds, D.J., Hallquist, M.N., Asato, M., Luna, B., 2014. Developmental stages and sex differences of white matter and behavioral development through adolescence: a longitudinal diffusion tensor imaging (DTI) study. Neuroimage 92, 356e368. https://doi.org/10.1016/ j.neuroimage.2013.12.044. Smith, S.M., Jenkinson, M., Johansen-Berg, H., Rueckert, D., Nichols, T.E., Mackay, C.E., et al., 2006. Tract-based spatial statistics: voxelwise analysis of multi-subject diffusion data. Neuroimage 31 (4), 1487e1505. Smith, S.M., Johansen-Berg, H., Jenkinson, M., Rueckert, D., Nichols, T.E., Miller, K.L.,., Behrens, T.E., 2007. Acquisition and voxelwise analysis of multi-subject diffusion data with tract-based spatial statistics. Nat. Protoc. 2 (3), 499. Snook, L., Paulson, L.A., Roy, D., Phillips, L., Beaulieu, C., 2005. Diffusion tensor imaging of neurodevelopment in children and young adults. Neuroimage 26 (4), 1164e1173. Snook, L., Plewes, C., Beaulieu, C., 2007. Voxel based versus region of interest analysis in diffusion tensor imaging of neurodevelopment. Neuroimage 34 (1), 243e252. Song, S.K., Sun, S.W., Ramsbottom, M.J., Chang, C., Russell, J., Cross, A.H., 2002. Dysmyelination revealed through MRI as increased radial (but unchanged axial) diffusion of water. Neuroimage 17 (3), 1429e1436. Sowell, E.R., Thompson, P.M., Holmes, C.J., Batth, R., Jernigan, T.L., Toga, A.W., 1999. Localizing age-related changes in brain structure between childhood and adolescence using statistical parametric mapping. Neuroimage 9 (6), 587e597. Sowell, E.R., Thompson, P.M., Holmes, C.J., Jernigan, T.L., Toga, A.W., 1999. In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nat. Neurosci. 2 (10), 859. Sowell, E.R., Delis, D., Stiles, J., Jernigan, T.L., 2001. Improved memory functioning and frontal lobe maturation between childhood and adolescence: a structural MRI study. J. Int. Neuropsychol. Soc. 7 (3), 312e322. Sowell, E.R., Thompson, P.M., Tessner, K.D., Toga, A.W., 2001. Mapping continued brain growth and gray matter density reduction in dorsal frontal cortex: inverse relationships during postadolescent brain maturation. J. Neurosci. 21 (22), 8819e8829. Sowell, E.R., Thompson, P.M., Rex, D., Kornsand, D., Tessner, K.D., Jernigan, T.L., Toga, A.W., 2002. Mapping sulcal pattern asymmetry and local cortical surface gray matter distribution in vivo: maturation in perisylvian cortices. Cerebr. Cortex 12 (1), 17e26. Sowell, E.R., Trauner, D.A., Gamst, A., Jernigan, T.L., 2002. Development of cortical and subcortical brain structures in childhood and adolescence: a structural MRI study. Dev. Med. Child Neurol. 44 (1), 4e16. Sowell, E.R., Peterson, B.S., Thompson, P.M., Welcome, S.E., Henkenius, A.L., Toga, A.W., 2003. Mapping cortical change across the human life span. Nat. Neurosci. 6 (3), 309. Sowell, E.R., Thompson, P.M., Leonard, C.M., Welcome, S.E., Kan, E., Toga, A.W., 2004. Longitudinal mapping of cortical thickness and brain growth in normal children. J. Neurosci. 24 (38), 8223e8231. Sowell, E.R., Thompson, P.M., Toga, A.W., 2004. Mapping changes in the human cortex throughout the span of life. Neuroscientist 10 (4), 372e392. Sowell, E.R., Peterson, B.S., Kan, E., Woods, R.P., Yoshii, J., Bansal, R., et al., 2006. Sex differences in cortical thickness mapped in 176 healthy individuals between 7 and 87 years of age. Cerebr. Cortex 17 (7), 1550e1560. Spear, L.P., 2000. The adolescent brain and age-related behavioral manifestations. Neurosci. Biobehav. Rev. 24 (4), 417e463. Squeglia, L.M., Jacobus, J., Sorg, S.F., Jernigan, T.L., Tapert, S.F., 2013. Early adolescent cortical thinning is related to better neuropsychological performance. J. Int. Neuropsychol. Soc. 19 (9), 962e970. https://doi.org/10.1017/S1355617713000878. Steinberg, L., 2005. Cognitive and affective development in adolescence. Trends Cognit. Sci. 9 (2), 69e74. Suzuki, Y., Matsuzawa, H., Kwee, I.L., Nakada, T., 2003. Absolute eigenvalue diffusion tensor analysis for human brain maturation. NMR Biomed. 16 (5), 257e260. Tamnes, C.K., Ostby, Y., Walhovd, K.B., Westlye, L.T., Due-Tonnessen, P., Fjell, A.M., 2010. Neuroanatomical correlates of executive functions in children and adolescents: a magnetic resonance imaging (MRI) study of cortical thickness. Neuropsychologia 48 (9), 2496e2508. https://doi.org/ 10.1016/j.neuropsychologia.2010.04.024.

316 PART | II Cognitive development

Tamnes, C.K., Herting, M.M., Goddings, A.L., Meuwese, R., Blakemore, S.J., Dahl, R.E., et al., 2017. Development of the cerebral cortex across adolescence: a multisample study of inter-related longitudinal changes in cortical volume, surface area, and thickness. J. Neurosci. 37 (12), 3402e3412. https://doi.org/10.1523/JNEUROSCI.3302-16.2017. Tamnes, C.K., Roalf, D.R., Goddings, A.L., Lebel, C., 2017. Diffusion MRI of white matter microstructure development in childhood and adolescence: methods, challenges and progress. Dev. Cogn. Neurosci. https://doi.org/10.1016/j.dcn.2017.12.002. Tanaka, C., Matsui, M., Uematsu, A., Noguchi, K., Miyawaki, T., 2012. Developmental trajectories of the fronto-temporal lobes from infancy to early adulthood in healthy individuals. Dev. Neurosci. 34 (6), 477e487. https://doi.org/10.1159/000345152. Thomas, C., Avidan, G., Humphreys, K., Jung, K.J., Gao, F., Behrmann, M., 2009. Reduced structural connectivity in ventral visual cortex in congenital prosopagnosia. Nat. Neurosci. 12 (1), 29. Thompson, P.M., Giedd, J.N., Woods, R.P., MacDonald, D., Evans, A.C., Toga, A.W., 2000. Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature 404 (6774), 190. Thompson, P.M., Cannon, T.D., Narr, K.L., Van Erp, T., Poutanen, V.P., Huttunen, M., et al., 2001. Genetic influences on brain structure. Nat. Neurosci. 4 (12), 1253. Toga, A.W., Thompson, P.M., 2005. Genetics of brain structure and intelligence. Annu. Rev. Neurosci. 28, 1e23. Tournier, J.D., Mori, S., Leemans, A., 2011. Diffusion tensor imaging and beyond. Magn. Reson. Med. 65 (6), 1532e1556. Tsujimoto, S., 2008. The prefrontal cortex: functional neural development during early childhood. Neuroscientist 14 (4), 345e358. Tzarouchi, L.C., Astrakas, L.G., Xydis, V., Zikou, A., Kosta, P., Drougia, A., Andronikou, S., Argyropoulou, M.I., 2009. Age-related grey matter changes in preterm infants: an MRI study. Neuroimage 47, 1148e1153. van Buchem, M.A., Steens, S.C., Vrooman, H.A., Zwinderman, A.H., McGowan, J.C., Rassek, M., Engelbrecht, V., 2001. Global estimation of myelination in the developing brain on the basis of magnetization transfer imaging: a preliminary study. Am. J. Neuroradiol. 22 (4), 762e766. van Solen, I.L., Brouwer, R.M., van Baal, G.C., Schnack, H.G., Peper, J.S., Collins, D.L., et al., 2012. Genetic influences on thinning of the cerebral cortex during development. Neuroimage 59 (4), 3871e3880. https://doi.org/10.1016/j.neuroimage.2011.11.044. Verhoeven, J.S., Sage, C.A., Leemans, A., Van Hecke, W., Callaert, D., Peeters, R., et al., 2010. Construction of a stereotaxic DTI atlas with full diffusion tensor information for studying white matter maturation from childhood to adolescence using tractography-based segmentations. Hum. Brain Mapp. 31 (3), 470e486. Vignaud, J., 1966. Development of the nervous system in early life. Part III. Radiological study of the normal skull in premature and newborn infants. In: Falkner, F.T. (Ed.), Human Development. WB Saunders Co, Philadelphia. Vijayakumar, N., Allen, N.B., Youssef, G., Dennison, M., Yucel, M., Simmons, J.G., Whittle, S., 2016. Brain development during adolescence: a mixedlongitudinal investigation of cortical thickness, surface area, and volume. Hum. Brain Mapp. 37 (6), 2027e2038. https://doi.org/10.1002/hbm.23154. Vijayakumar, N., Mills, K.L., Alexander-Bloch, A., Tamnes, C.K., Whittle, S., 2017. Structural brain development: a review of methodological approaches and best practices. Dev. Cogn. Neurosci. https://doi.org/10.1016/j.dcn.2017.11.008. Von Economo, C., 1929. The Cytoarchitectonics of the Human Cerebral Cortex. H. Milford Oxford University Press. Wakana, S., Jiang, H., Nagae-Poetscher, L.M., Van Zijl, P.C., Mori, S., 2004. Fiber tractebased atlas of human white matter anatomy. Radiology 230 (1), 77e87. Wakana, S., Caprihan, A., Panzenboeck, M.M., Fallon, J.H., Perry, M., Gollub, R.L., et al., 2007. Reproducibility of quantitative tractography methods applied to cerebral white matter. Neuroimage 36 (3), 630e644. Wang, Y., Adamson, C., Yuan, W., Altaye, M., Rajagopal, A., Byars, A.W., Holland, S.K., 2012. Sex differences in white matter development during adolescence: a DTI study. Brain Res. 1478, 1e15. https://doi.org/10.1016/j.brainres.2012.08.038. Wechsler, D., 2003. Wechsler Intelligence Scale for Children, fourth ed. The Psychological Corporation, New York. Westlye, L.T., Walhovd, K.B., Dale, A.M., Bjørnerud, A., Due-Tønnessen, P., Engvig, A., et al., 2010. Differentiating maturational and aging-related changes of the cerebral cortex by use of thickness and signal intensity. Neuroimage 52 (1), 172e185. Wierenga, L., Langen, M., Ambrosino, S., van Dijk, S., Oranje, B., Durston, S., 2014. Typical development of basal ganglia, hippocampus, amygdala and cerebellum from age 7 to 24. Neuroimage 96, 67e72. https://doi.org/10.1016/j.neuroimage.2014.03.072. Wierenga, L.M., Langen, M., Oranje, B., Durston, S., 2014. Unique developmental trajectories of cortical thickness and surface area. Neuroimage 87, 120e126. https://doi.org/10.1016/j.neuroimage.2013.11.010. Wilke, M., Sohn, J.H., Byars, A.W., Holland, S.K., 2003. Bright spots: correlations of gray matter volume with IQ in a normal pediatric population. Neuroimage 20 (1), 202e215. Wilke, M., Krägeloh-Mann, I., Holland, S.K., 2007. Global and local development of gray and white matter volume in normal children and adolescents. Exp. Brain Res. 178 (3), 296e307. Willerman, L., Schultz, R., Rutledge, J.N., Bigler, E.D., 1991. In vivo brain size and intelligence. Intelligence 15 (2), 223e228. Witelson, S.F., Beresh, H., Kigar, D.L., 2005. Intelligence and brain size in 100 postmortem brains: sex, lateralization and age factors. Brain 129 (2), 386e398. Witte, A.V., Savli, M., Holik, A., Kasper, S., Lanzenberger, R., 2010. Regional sex differences in grey matter volume are associated with sex hormones in the young adult human brain. Neuroimage 49 (2), 1205e1212. Wolff, S.D., Balaban, R.S., 1989. Magnetization transfer contrast (MTC) and tissue water proton relaxation in vivo. Magn. Reson. Med. 10 (1), 135e144. Wozniak, J.R., Lim, K.O., 2006. Advances in white matter imaging: a review of in vivo magnetic resonance methodologies and their applicability to the study of development and aging. Neurosci. Biobehav. Rev. 30 (6), 762e774.

Structural brain development: birth through adolescence Chapter | 14

317

Yakovlev, P.I., Lecours, A.R., 1967. The myelogenetic cycles of regional maturation of the brain. In: Minkowski, A. (Ed.), Regional Development of the Brain Early in Life. Blackwell Scientific Publications Inc, pp. 3e70. Zhang, H., Schneider, T., Wheeler-Kingshott, C.A., Alexander, D.C., 2012. NODDI: practical in vivo neurite orientation dispersion and density imaging of the human brain. Neuroimage 61 (4), 1000e1016. Zhang, J., Evans, A., Hermoye, L., Lee, S.K., Wakana, S., Zhang, W., et al., 2007. Evidence of slow maturation of the superior longitudinal fasciculus in early childhood by diffusion tensor imaging. Neuroimage 38 (2), 239e247. Zhang, J.M., Konkle, A.T., Zup, S.L., McCarthy, M.M., 2008. Impact of sex and hormones on new cells in the developing rat hippocampus: a novel source of sex dimorphism? Eur. J. Neurosci. 27 (4), 791e800. Zhou, D., Lebel, C., Treit, S., Evans, A., Beaulieu, C., 2015. Accelerated longitudinal cortical thinning in adolescence. Neuroimage 104, 138e145. https:// doi.org/10.1016/j.neuroimage.2014.10.005.

Chapter 15

Statistical learning mechanisms in infancy Abbie Thompson1, Ariel Aguero2 and Jill Lany3 1

Valparaiso University, Valparaiso, IN, United States; 2University of Notre Dame, Notre Dame, IN, United States; 3University of Liverpool,

Liverpool, United Kingdom

Chapter outline 15.1. Learning probability distributions 15.2. Learning co-occurrence statistics 15.2.1. Learning co-occurrence statistics in speech: word segmentation 15.2.2. Do infants learn words from co-occurrence statistics? 15.2.3. Learning co-occurrence statistics in the visual domain 15.2.4. Learning co-occurrence statistics in speech: word segmentation

319 320 321 322 323 324

15.2.5. Learning co-occurrence relations between words and referents: cross-situational learning 15.3. Linking individual differences in statistical learning to language development 15.4. Statistical learning in individuals with language delays and disorders 15.5. Scaling statistical learning to real-world challenges 15.6. Conclusions Acknowledgment References

326 327 328 329 330 331 331

15.1 Learning probability distributions While psychologists long imagined that infants experience the world as a bewildering array of sights, sounds, and sensations (James, 1890), research since the 1950s has shown that infants have more advanced perceptual and cognitive abilities than was once thought. For example, infants’ visual, auditory, tactile, and vestibular systems are functional at birth (and even beforehand), and their discrimination abilities in these areas, while not adultlike, are nonetheless remarkably acute (Kellmann and Arterberry, 2000). Perhaps, then, it should not be surprising that very young infants show evidence of learning from stimuli in their environment. In this chapter we will review recent research focused on infants’ ability to learn several kinds of statistical structure, such as probability distributions, sequential structure, and associations between properties within and across instances and modalities. These types of statistical structures appear to play a role in many aspects of development, including auditory and visual perception, language development, object perception, and event processing. We focus much of our discussion on the role of these learning mechanisms in language acquisition, though we also make connections between language learning and learning in other domains. In addition, we consider several controversial questions, such as what the mechanistic underpinnings of statistical learning may be, how much infants can really learn about things like words, objects, and events from statistics, whether they can track statistics outside of simple lab tasks, and what individual differences in statistical learning ability can tell us about language development.A basic form of learning available to infants is sensitivity to the frequencies with which items or events occur in the environment. Early research in infant perceptual development revealed that infants can differentiate between frequently and infrequently seen items: when presented with successive trials of two pictures side by side, in which one is repeated while the other changes, infants spend more time looking at the novel picture (Fantz, 1964). However, infants can do something potentially much more powerful than tracking frequencies of individual items or instances: they can track the relative frequencies of items within a set, or probability distributions. An example of infants’ sensitivity to probability distributions comes from research on the development of speech perception. Differences between speech sounds that correspond to differences in the meanings of words are called

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00015-9 Copyright © 2020 Elsevier Inc. All rights reserved.

319

320 PART | II Cognitive development

“phonetic contrasts.” For example, the English sounds /r/ and /l/ are considered to be different phonemes, representing a phonetic contrast, because a switch from one to the other changes the meaning of a word (i.e., “right” vs. “light”). However, while /r/ and /l/ are perceived differently by adult English speakers, they are not perceived differently by native speakers of Japanese, as these speech sounds do not correspond to a meaningful difference in their language. The fact that adults’ ability to discriminate among speech sounds reflects the phonetic organization of their native language strongly suggests that experience with language results in phonetic tuning. Indeed, very young infants readily discriminate among almost all of the speech sounds of the world, regardless of the language spoken by the adults in their community (e.g., Eimas et al., 1971; for review see Aslin et al., 1998). By just 6 months of age, infants’ phonetic discrimination patterns begin to reflect the phonetic contrasts relevant in their native language (Werker and Tees, 1984; Kuhl et al., 1992). Learning about probability distributions, and more specifically registering the relative frequency with which sounds occur, appears to be a mechanism that underlies phonetic tuning. Within a given language, the distribution of speech sounds along a particular continuum reflects phonetic category information (e.g., Lisker and Abramson, 1964). Specifically, every production of a speech sound differs on a variety of acoustic dimensions, influenced by the characteristics like individual speaker’s vocal tract, speech rate, coarticulation, and phrase or sentence-level prosody. However, the values of acoustic dimensions that are critical to perceived differences between phonemes tend to occur in nonuniform distributions, such that some values along a dimension are much more frequent than others (i.e., forming a bimodal or trimodal distribution). Maye et al. (2002) tested whether experience with unevenly distributed sound profiles influences phonetic perception in English-learning infants at 6 and 8 months of age. To do so, they created a set of eight speech sounds forming a continuum between [da] and [ta]. Infants in the Unimodal condition heard a distribution in which the intermediate values occurred most frequently, with decreasing frequencies toward the tails of the distribution. For infants in the Bimodal condition, tokens at the two endpoints of the distribution occurred with high frequency, while the tokens at the midpoint and endpoints were relatively less frequent. Only infants in the bimodal condition showed discrimination of the endpoints, suggesting that infants’ experience with unimodal distributions along an acoustic continuum plays a key role in the loss of sensitivity to phonetic contrasts not relevant in their language. Subsequent work revealed that experience with distributional variation in speech input can also result in an enhancement or sharpening of infants’ sensitivity to difficult contrasts (Maye et al., 2008) and that at 10e12 months of age, just after the period of perceptual reorganization, infants’ speech sound discrimination abilities are less readily influenced by experience with distributional information. Specifically, infants need more extensive experience with a bimodal distribution to show evidence of discriminating between speech sounds that are not contrastive in their native language (Yoshida et al., 2010). Kuhl (2004) has proposed that infants’ experience with acoustic and distributional cues work in concert with maturational changes in neural development to produce such changes. Data from neurophysiological recordings suggest that there are neural changes in infants’ processing of nonnative phonetic contrasts that map onto performance on behavioral discrimination tasks (Cheour et al., 1998), and that neural signatures indexing changes in discrimination predict later vocabulary development (Rivera-Gaxiola et al., 2005). In sum, experience with distributional information plays an important role in the dramatic changes in speech sound discrimination that takes place between 8 and 10 months of age. While infants remain sensitive to distributional cues marking nonnative phonemic contrasts after the period of reorganization, these findings suggest that neural changes, increased experience with the distributional properties of one’s native language, or a combination of these factors reduce their impact on discrimination. It is important to note that the social context in which infants experience nonnative language distributional patterns may also modulate these effectsdinfants show greater sensitivity to different nonnative contrasts after listening to a native speaker in person than after watching and listening to a native speaker on a video tape (Kuhl et al., 2003). These findings suggest that although reorganization takes place early, the capacity to learn novel contrasts is maintained, and can be facilitated through social interaction.

15.2 Learning co-occurrence statistics Within a given domain, the reliable co-occurrence of elements is typically a surface manifestation of a meaningful relationship between those elements. Thus, co-occurrence information has the potential to play a pervasive role in learning, as long as an individual is sensitive to it. Despite the seemingly simple nature of such associative relationships, tracking sequential associations in most environmental patterns can be quite challenging. It entails encoding a stimulus, as well as those that precede and follow it, and maintaining memory representations across multiple occurrences of each stimulus. This might not seem difficult to do for one item that occurs in a reliable context, but consider tracking this information for a set of 100, or even 1000, different items that occur in highly variable contexts.

Statistical learning mechanisms in infancy Chapter | 15

321

15.2.1 Learning co-occurrence statistics in speech: word segmentation Natural language provides a compelling illustration of how challenging it can be to track patterns of co-occurrence. Given that the sounds comprising words and sentences unfold over time, as opposed to being expressed simultaneously, spoken language is rich with sequential structuredand thus with regularities in the co-occurrence of sounds across time. Language is hierarchically organized, consisting of patterns at both very fine-grained and larger-grained levels, and there are sequential regularities at each level of structure. Because of the high demands that tracking such complex sequential structure would place on infants, it is critical to determine whether infants have sufficient computational resources to track such complex information, and the extent to which such learning mechanisms could support language acquisition. A key sequential regularity in language pertains to how syllables are combined to form words. Syllables that reliably co-occur often belong to the same word, while syllables that rarely co-occur are more likely to span word boundaries (e.g., Swingley, 2005). Because of this feature of language, transitional probabilities between syllables tend to be higher within words than between words. Formally, the transitional probability (TP) of a co-occurrence relationship between two elements, X and Y, is determined by dividing the frequency of XY by the frequency of X. This yields the probability that if X occurs, Y will also occur. Saffran et al. (1996) tested whether 8-month-old infants, who are just beginning to learn their first words, are able to learn TPs. In their task, infants first listened to a stream of synthesized speech in which the only potential cue to word boundaries were the TPs between adjacent syllables (they were 1.0 within words and 0.33 across word boundaries). Infants listened to the stream for about 2 min, and were then tested on their ability to discriminate between the syllable sequences with high TPs of 1.0 (“words”) and novel combinations of the familiar syllables (“nonwords”), with TPs of 0. Infants listened longer to the strings containing nonwords, suggesting they distinguished between syllable sequences that were attested in their input and sequences that had never occurred. A second group was familiarized in the same manner and tested on sequences with high TP “words” and “partwords,” or sequences of syllables that had occurred across word boundaries (e.g., the final syllable of one word and the first two syllables of another word), which had TPs of 0.33. Infants listened longer to the partwords, suggesting that they can also discriminate sequences that contain highly reliable transitions from those that contain less reliable ones. Subsequent studies have probed the kinds of sequential statistics that infants are sensitive to in these studies, such as the frequency of a sequence versus its transitional probability. The frequency of a sequence is simply the number of times it occurs, while the TP of a co-occurrence relationship provides a measure of how tightly linked or connected X and Y are, controlling for the raw frequency of X. If X occurs many times without Y, then no matter how many times it occurs with Y, the TP will be relatively low. Conversely, a sequence can have low frequency but a high TP. In the Saffran et al. (1996) study, the syllable transitions within words were both more frequent and had higher TPs than the syllable transitions in partwords. Aslin et al. (1998) thus investigated whether infants can track TPs, or just frequencies. Specifically, they modified the design of Saffran et al. (1996) such that two of the words occurred twice as often as the other words. This manipulation permitted a design in which the four test itemsdtwo words and two partwordsdwere equally frequent; however, the TPs within words were 1.0, and the TPs spanning the word boundaries were still 0.33. Infants showed discrimination between the words and partwords, despite the fact that the syllable sequences occurred equally often in both types of test items. This suggests that by 8 months of age, infants have a powerful mechanism for tracking co-occurrence relationships, and for distinguishing potentially spurious co-occurrences from ones in which there is a very tight connection. In subsequent work, infants even younger than 8 months have shown behavioral evidence of tracking TPs in speech (Johnson and Tyler, 2010). In fact, even newborn infants appear to be sensitive to this kind of statistical structure (Teinonin et al., 2009). In this study, sleeping newborns were familiarized with a stream of syllables in which potential word boundaries were marked by TPs. The newborns wore EEG caps, and ERP amplitudes to the stimuli were recorded. There was a stronger negative response to word-initial syllables than to later ones, which suggests that infants recognized the onsets of the words, and processed them differently than the word-internal syllables. The observed negative-going waveform may reflect the degree to which the syllables could be predicted by the preceding syllables: Within words the syllable sequences were entirely predictable, but were relatively unpredictable across the word boundaries, and the greater negativity could reflect the relative unexpectedness of the word-initial syllable. It is important to note that the TPs in these studies have primarily been described in terms of prospective relationships, or the probability that, given the occurrence of a syllable, another syllable will follow (i.e., the probability that the syllable “by” will follow the syllable “ba,” as in the word baby). However, infants could also be sensitive to the probabilities of retrospective relationships (i.e., the probability that the syllable “ba” will precede the syllable “by”). In the studies of sensitivity to TPs reviewed so far, the words contained prospective, or forward TPs (FTPs) of 1.0, and also retrospective, or backward TPs (BTPs) of 1.0. This means that infants could have used either one (or both) to discriminate words from

322 PART | II Cognitive development

partwords. Pelucchi et al. (2009a) thus familiarized 8-month-old infants with a corpus in which words and partwords differed in their BTPs, but both had FTPs of 1.0. Thus, infants could only use BTPs to distinguish between the high TP (HTP) and low TP (LTP) words. Infants discriminated between words and partwords, suggesting that they were sensitive to the BTPs of the syllable sequences. While FTPs may facilitate anticipating upcoming stimuli, infants’ sensitivity to BTPs may help them to remember what came before a particular element. Thus, infants’ sensitivity to both types of relationships potentially enhances word segmentation and other tasks that require sequential learning and processing. There is evidence, however, that infants’ tendency to privilege forward versus backward TPs in speech across development is shaped by native-language experience (Thiessen et al., 2019). In Korean, word-order patterns are most predictable in the forward direction, and in English they are more predictable in the backward direction. At 7 months neither infants learning English as their native language, nor those learning Korean, appear to be biased to track TPs in one direction over the other. However, by 13 months, infants learning Korean are more likely to track prospective TPs, and those learning English are more likely to track retrospective ones. This bias is maintained into adulthood (Onnis and Thiessen, 2013). These findings suggest that learners’ sensitivity to statistical structure in their environment is shaped by their prior experiences.

15.2.2 Do infants learn words from co-occurrence statistics? It is clear from the studies we have just described, which predominantly used auditory discrimination paradigms, that infants are influenced by TPs in artificial language materials. But, there is debate about whether that sensitivity to TPs leads to recognizing words versus just coherent generic sequences. In other words, the fact that infants discriminate between HTP and LTP sequences does not tell us whether infants actually learn “words” from statistics (Endess and Mehler, 2009). Infants readily track TPs in nonlinguistic auditory streams (Saffran et al., 1999) as well as in visual materials, as we discuss below, and thus it seems unlikely that they represent all of those sequences as words. One line of research that bears on this question tested whether infants have an advantage in mapping sequences with high TPs (or HTPs) to referents. Graf Estes and colleagues reasoned that if these sequences are represented as potential words, rather than as generic sound sequences, they should be more readily learned as object labels (Graf Estes et al., 2007). In this study 17-month-olds were first familiarized with an artificial language in which the only cues to potential word boundaries were the TPs between adjacent syllables. Infants were then habituated to two label-object pairs. For some infants, the labels corresponded to the “words” (TP ¼ 1.0), and for the other infants the labels were equally frequent “partwords” (TP ¼ 0.5) from the speech stream. Only infants who were trained on pairings between the HTP sequences and objects showed evidence of learning the mappings, suggesting that learning labels composed of syllables with high TPs is easier than learning labels with low TPs (or LTPs). These findings were recently replicated using natural language materials (Hay et al., 2011). In this study, 17-month-old infants learning English as their native language were familiarized to a corpus of Italian sentences. Across the sentences there were HTP words (i.e., these words had TPs of 1.0, because their syllables did not occur anywhere else within the corpus). There were also equally frequent LTP words (i.e., these words had TPs of 0.33, because the syllables that comprised them occurred in other words throughout the corpus). As in the Graf Estes study, infants were then trained with either HTP or LTP words used as labels for potential referents. Infants showed evidence of mapping the HTP words to referents, but failed to do so for LTP words. These findings suggest that infants are sensitive to TPs in more natural language materials that were much more complex than the artificial languages often used in studies of statistical learning, and lend additional support to the suggestion that sensitivity to TPs can result in wordlike representations. However, there is another potential explanation for the findings that infants are more likely to learn associations between HTP sequences and objects than associations between LTP sequences and objects. Specifically, it is possible that HTP sequences are better learned than LTP sequences simply because they are easier to process, and there is a general boost to learning them. It is hard to disentangle whether benefits arise from sequences’ high internal coherence versus from having “lexical status.” One way to begin to tear these explanations apart is by testing whether any HTP speech sequence is better learned as a label, or whether learning HTP sequences that are dissimilar to infants’ native language begins to suffer around the time at which infants become narrower in the kinds of stimuli they will accept as labels. Initially infants are fairly flexible in the forms that they will learn as labels, learning gestures, nonlinguistic sounds, etc., as readily as spoken words. But, as they approach age 2, infants begin to resist learning these nonword forms, especially those with larger native-language vocabularies. Interestingly, while at 17 months English-learning infants show an advantage in learning HTP over LTP sequences; this seems to be changing at 21 months. By 21 months, only infants with smaller English vocabularies showed this effect, learning HTP over LTP sequences as words, while those with larger vocabularies failed to learn them (Shoaib et al., 2018). Infants both younger and older than 17 months continue to differentiate between HTP and

Statistical learning mechanisms in infancy Chapter | 15

323

LTP words (Hay et al. (2011); Karaman and Hay (2018)), suggesting that infants segmented the HTP words, but resisted learning them as labels. Importantly, while they resisted learning them, these findings suggest that they did so because they are represented as potential words. These findings may be relevant to questions about the mechanisms by which infants come to distinguish between sequences with high versus low statistical coherence (see Thiessen, 2017 and Thiessen et al., 2013 for reviews). One possibility is that the mechanism actually involves computing co-occurrence likelihoods across input sequences, and that as a result, learners use TPs to identify boundaries within streams of speech. This process has been successfully modeled using connectionist networks, such as recurrent networks (e.g., Christiansen et al., 1998). Another possibility is that infants encode and remember chunks from their input. There are different formalizations of chunking models, but, following Thiessen (2017) we will use the PARSER model (Perruchet and Vintner, 1998) to illustrate. This account emphasizes the fact that syllables that comprise HTP sequences are more likely to recur in language input in that configuration (i.e., the component syllables are relatively likely to reappear together). In contrast, the components of chunks with lower TPs are more likely to occur with other syllables. Exposure to repeated chunks (or HTP sequences) strengthens representations of the entire “chunk,” while exposure to syllable sequences that contain recombinations of elements present in previous chunks reduces the strength. Boundary finding models predict increasing sensitivity to TPs within units with accumulating exposure to them, but chunking models predict that as chunks are strengthened, sensitivity to internal TPs actually diminishes. There is evidence from adults that as wordlike units become more familiar, sensitivity to internal TPs diminishes (Giroux and Rey, 2009). Likewise, evidence that sequences with high TPs are represented as “words,” or as meaningful units, is in line with chunking accounts. However, there is still much work to be done to more fully describe the learning mechanisms that result in sensitivity to statistical regularities.

15.2.3 Learning co-occurrence statistics in the visual domain The studies described so far suggest that infants’ sensitivity to co-occurrence regularities could play a role in learning language structure in speech. But far from being specific to learning in the auditory domain, infants’ sensitivity to cooccurrence supports learning across perceptual modalities/domains. For example, in the visual domain, elements that reliably co-occur across time and space tend to belong to the same object, and thus provide information about object boundaries. Thus, just as in the case of segmenting words from continuous speech, the co-occurrence of features in the visual field is potentially a powerful cue that could be used to learn which visual features form objects. However, in the case of objects, the co-occurrence relations are not sequential, but rather spatial. To test whether infants can track such conditional probabilities in complex visual displays, Fiser and Aslin (2002) showed 9-month-old infants scenes containing three elements. In each scene, three discrete shapes were presented simultaneously in a 2X2 grid. Two of the elements formed a “base pair,” such that those objects always occurred together in a consistent spatial orientation. The probability of co-occurrence, or TP, was 1.0 for those objects. The two elements of each base pair also occurred with a third element, but its position relative to the base pair varied (TP ¼ 0.25). Infants were then tested on scenes containing two elementsdeither base pair elements in their proper configuration or two elements that did not form a base pair. Infants were able to discriminate between base pairs and nonbase pairs when they differed in the frequency with which they appeared during habituation, and also when they were equally frequent but differed in their TPs. This suggests that sensitivity to statistical information facilitates object perception, and raises the question of whether such learning can support object perception in younger infants. Infants also track sequential statistics in visual sequences, which may be relevant for perceiving events that unfold across time. For example, Kirkham et al. (2002) tested infants’ ability to detect sequential relationships in a series of shapes. In these experiments, 2-, 5-, and 8-month-old infants were habituated to a series of six shapes that appeared sequentially on a screen. The objects were grouped into three sets of sequential pairs. Within a pair, one shape was reliably followed by another shape; the TPs between objects within a pair were 1.0, while the TPs at pair boundaries were 0.33. At all of the ages tested, infants discriminated between strings that preserved the statistical regularities and sequences in which they were violated. Impressively, even newborn infants appear to learn these statistics, though only when the number of pairs is reduced to 2 (Bulf et al., 2011). Furthermore, while the infants in these experiments could have used either frequency of co-occurrence or TPs to distinguish between the test strings, subsequent work revealed that by 4 months of age, infants can separately track frequency of shape co-occurrence, as well as the conditional probability of shape pairs (Marcovitch and Lewkowicz, 2009). These findings suggest that the ability to track sequential co-occurrence relationships in the visual domain emerges early in development. These studies tested sensitivity to co-occurrence statistics within a series of shapes that appear one at a time, such that the key co-occurrence relations were highly temporal. However, many sequential structures also involve spatial

324 PART | II Cognitive development

information. Tummeltshammer et al. (2017) recently tested whether there are parallels in 8-month-olds’ ability to track sequential visual structure and spatial regularities in dynamic events using a closely matched set of materials. They also tested whether infants track forward and backward TPs in both kinds of events, just as they do in spoken language materials. In the materials testing sequential learning, objects appeared one at a time in a central location, as they did in Kirkham et al. (2002). In the materials testing spatial relationships, the same objects were used, but the shape sequence streamed across the screen, such that multiple elements were present simultaneously. In both sets of materials, there were shapes sequences that had high forward TPs, and also sequences that had high backward TPs. Learning was assessed by measuring infants’ attention to streams with high and low TPs. Infants showed evidence of tracking the TPs in both streams, but stronger sensitivity to backward TPs in temporal versus spatial stream. These results suggest that infants can track structure in dynamic events that have a spatial component, in addition to being able to track temporal structure. However, the differences in learning across the two kinds of streams raise the possibility that different mechanisms are involved in learning temporal versus spatial visual regularities. Infants can also track statistics in more complex spatiotemporal events, such as a stream of continuous human action. Considering everyday events, such as brushing one’s teeth, making coffee, or opening a toy container, the subcomponents of the action tend to co-occur in reliable sequences, and thus tracking sequential structure in the visual domain may play a role in both detecting event boundaries and in learning their internal structure. Adults can use statistical information to learn regularities in dynamic action sequences (Baldwin et al., 2008), and this ability appears to be present by at least 7e9 months (Roseberry et al., 2011). In this study, infants were familiarized to a video of a man performing a continuous series of two hand movements per second. Some movements were reliably paired, such that the TPs between the movements were 1.0, while the TPs were 0.5 across movements that spanned the pairs. Infants were able to differentiate between sequences with high versus low TPs. Subsequent work showed that infants can track TPs within a stream of human-produced actions even when “coarticulary” cuesddifferences on the way one action is performed based on what will be done next, which can serve as a cue to what will happen nextdare removed (Stahl et al., 2014). These findings suggest that infants can use statistics to help them perceive event structure and coherence. A key dimension of many event sequences is causality. Specifically, when one event causes another, there tends to be a reliable sequential relation, with the cause preceding the effect. Recent evidence suggests that by 24 months of age, toddlers can use statistical structure (specifically sequential co-occurrence) to infer causality (Waismeyer et al., 2015). In this study, toddlers watched as an adult placed two different objects onto a box repeatedly, one at a time. Only one of the objects, when placed on the box, resulted in a marble being dispensed from an apparatus located a short distance away. Thus, there were no visual cues indicating direct physical causality for either objectdthe only cue was the co-occurrence likelihood of the marble being dispensed when each of the objects was placed on the box. The toddlers were then given a chance to interact with the objects themselves, and more often put the “causal” object on the box. They even did so when the outcome was only probabilistically related rather than deterministically, and when frequency with which each object resulted in the marble dispensation was equated. These results suggest that toddlers can use statistical information about how reliably (rather than just how frequently) events co-occur in a sequence to determine causality. Of course, reliable co-occurrence alone is not an infallible cue to causality, and it is an open question how the causal relations that underlie the statistical regularities in such sequences are discovered. One clue may come from a recent study by Monoy et al. (2017), who found evidence that infants between 8 and 11 months of age were able to track TPs in action sequences in which a hand reached in and acted on a set of objects affixed in a circle to board. There were two reliable “action pairs” (with TPs on 1.0) and nonpair sequences that were less reliable (TPs ¼ .167). The completion of only one of those action pairs was followed by visually salient outcome (i.e., a star at the center of the board lights up after the occurrence of that action pair). Infants showed anticipatory looking to upcoming locations during learning only for the pair followed by the salient outcome, suggesting it was best learned. EEG recordings also suggested that these pairs had some special status: a front-central Nc (reflecting visual attention) present when the pair followed by the salient outcome was disrupted, but not when the other pair was disrupted. At minimum these findings suggest that salient outcomes can facilitate tracking TPs. But these findings also suggest that one way infants begin to identify causality from co-occurrence may be by noting salient outcomes.

15.2.4 Learning co-occurrence statistics in speech: word segmentation Thus far we have reviewed evidence suggesting that infants’ sensitivity to co-occurrence structure in the environment may be relevant to identifying words and objects, and perceiving coherent events. Sensitivity to co-occurrence relationships appears to play a role in learning grammatical patterns, such as learning how words can be combined into phrases and sentences. For example, several studies using auditory preference paradigms suggest that infants can track TPs not just

Statistical learning mechanisms in infancy Chapter | 15

325

between syllables but between words. For example, Gómez and Gerken (1999) tested whether 12-month-old infants can learn TPs within multiword “sentences.” They created an artificial language in which nonsense words were combined to form sentences in which there were probabilistic regularities in the ordering of words. Infants were then tested on novel grammatical and ungrammatical strings that contained familiar words. Infants successfully discriminated between grammatical and ungrammatical strings, providing evidence that they learned the probabilistic co-occurrence relationships between words. A critical property of grammatical structure is that it pertains not just to how individual words are combined (e.g., “the cat” is grammatical, but “cat the” is not), but also to how words from different grammatical categories can be combined (e.g., determiners like “the” and “a” precede nouns rather than follow them). The ability to track co-occurrence patterns among words, and specifically the extent to which words share co-occurrence patterns, is useful for learning word-order patterns (Wonnacott et al., 2008). Nonetheless, learning grammatical categories based on shared co-occurrence alone (also referred to as a distributional cue) can be very difficult (Smith, 1969; Braine, 1987). Importantly, in natural languages, words from different syntactic categories are distinguished both by distributional (Cartwright and Brent, 1997; Mintz, 2003; Mintz et al., 2002; Monaghan et al., 2005; Redington et al., 1998) and phonological cues (Farmer et al., 2006; Kelly, 1992; Monaghan et al., 2005). In other words, words from the same grammatical category share correlated distributional and phonological features. For example, nouns both tend to occur after “a” and “the,” and to have a strong-weak stress pattern. Gómez and Lakusta (2004) found that 12-month-olds can learn correlations between distributional and phonological features, suggesting that category learning may initially involve tracking the correlations between distributional and phonological features of words. Infants can even capitalize on experience with such cues to facilitate learning the meanings of words within such categories (Lany and Saffran, 2010), even when the relations are probabilistic (Lany, 2014). These studies suggest that infants readily track co-occurrence relationships between adjacent words. However, in natural languages, co-occurrence relationships can also occur between nonadjacent elements. For example, grammatical dependencies marking tense are often nonadjacent, as in the relationship between auxiliaries such as “is” and the progressive inflection “ing,” as they are necessarily separated by a verb (e.g., “is running,” “is eating,” “is talking”). Likewise, a plural noun predicts plural marking on the subsequent verb, but the noun and verb can be separated by modifiers, as in “The kids who were late to school are in trouble.” Tracking nonadjacent dependencies is likely to place greater demands on memory than tracking adjacent dependencies, as elements must be remembered long enough to be linked to other elements occurring later in time. In addition, because nonadjacent dependencies can be separated by several word elements, there are many potentially irrelevant relationships for the learner to track, presenting a considerable computational burden. Gómez (2002) investigated the conditions that promote sensitivity to nonadjacent relationships. In particular, she hypothesized that there is a relationship between the presence of salient adjacent structure and the tendency to track nonadjacent structure. Because tracking adjacent relationships appears to be relatively easy for infants, they are likely to focus on adjacent structure as long as it is reliable. However, when the variability in adjacent relationships is high (i.e., when adjacent TPs are low), infants may be less likely to track those relationships and more likely to track reliable nonadjacent regularities. To test this hypothesis, Gómez (2002) exposed 18-month-olds to artificial language strings containing nonadjacent dependencies with TPs of 1.0. Critically, these dependencies were separated by an intervening element drawn from a set of 3, 12, or 24 different elements, and thus the adjacent TPs between words varied across conditions: relatively high (0.33, set size 3), medium (0.08, set size 12), or very low (less than 0.01, set size 24). Only infants in the set size 24 condition, in which the predictability of adjacent sequences was very low, successfully discriminated between familiar grammatical strings and strings that violated the nonadjacent relationships. In subsequent research, Gómez and Maye (2005) found that 15-month-old infants also track nonadjacent dependencies under conditions of high variability, but that 12-month-olds fail to do so. However, Lany and Gómez (2008) found that prior learning about adjacent relationships facilitates infants’ detection of nonadjacent patterns. In particular, when 12-month-old infants were first given experience with adjacent relationships between word categories, they then successfully detected novel nonadjacent relationships between words from those categories. However, this effect appeared to be restricted to females, and similar sex differences have been found in other studies of nonadjacent dependency learning. For example, much younger infants do show evidence of tracking TPs when looking at ERPs to their violations, but females show a more mature neural signature indicating discrimination (Mueller et al., 2012). This difference has also been observed in toddlers, with females more likely to generalize to novel instances of nonadjacent relations than males (Willits et al., 2017). These results may suggest that there may be differences in the development of the mechanisms by which males and females learn nonadjacent dependencies. They also raise the possibility of using individual differences in performance on statistical learning tasks to better understand how these learning processes support development. We will return to this issue in a later section.

326 PART | II Cognitive development

15.2.5 Learning co-occurrence relations between words and referents: cross-situational learning A central problem facing infant language learners is that of learning relations between words and their referents. Even the occurrence of a word with a very concrete and observable meaning, such as “rabbit,” will coincide with many objects and events in the environment, sometimes but not always including an actual rabbit (e.g., Quine, 1960). Thus, establishing the referent of a novel word poses a formidable challenge. Even if the infant is fixating on a rabbit, she still must determine that the label refers to the animal itself, as opposed to its floppy ears, or to its color. Recent research suggests infants’ sensitivity to the reliability of co-occurrence relationships between words and particular aspects of the environment, also referred to as “cross-situational learning,” may facilitate learning in such complex environments. Smith and Yu (2008) tested whether infants in the early stages of word learning can capitalize on these relationships using a cross-situational learning paradigm. They presented 12- and 14-month-old infants with six words whose referents were embedded in complex scenes. On each trial, infants heard a label (e.g., “bosa”) while viewing a scene consisting of the referent along with a distractor object. Given just a single trial of this nature, infants would be unable to determine which object was the referent. However, on other trials in which “bosa” was presented, one of the same objects was presented along with a different distractor. On other trials, the label “manu” was presented, and the object that consistently occurred with it served as a distractor on other trials. Thus, across trials, each label consistently occurred with one object. After infants were familiarized with the label-object training trials, they were tested using a preferential looking procedure. On each trial, infants were shown two objects simultaneously while the label for one of them was repeated. Infants looked significantly longer to the object that matched the label at both ages, with stronger effects for the 14-month-olds than for the 12-month-olds. These findings suggest that infants are sensitive to the reliability of label-object pairings across occurrences, and can integrate information across trials to settle on the most probable referent. While infants as young as 12 months are able to learn word-referent pairing in cross-situational learning tasks, their ability to succeed in these tasks appears to be related to their attention and memory abilities (Smith and Yu, 2013; Vlach and Johnson, 2013; Yu and Smith, 2011). Yu and Smith (2011) used a cross-situational learning paradigm to teach 12- and 14-month-old infants six novel word-referent mappings over a series of trials. While the infants were able to learn the word-mappings through cross-situational statistics, not all of the infants learned the words equally well. The majority of the infants learned at least half of the words, but some infants only learned one or two of them. Analysis of infants’ looking to referents during training revealed that two factors were associated with better word-referent learning; longer fixations to referents and producing fewer gaze-shifts between the referents. Longer fixations and reduced gaze-shifting may lead to better learning by strengthening connections between words and referents across trials, whereas short fixations and many gaze-shifts may reduce the strength of the associations, and lead to weaker learning. Smith and Yu (2013) tested whether novelty drives infants’ attention during the word-learning process. They used the stimuli from their previous study (Yu and Smith, 2011), but the trials were rearranged into a blocked design so that one object, the stable object, remained in the same location throughout a block of five trials and the object in the other location, the changing object, displayed a new referent on each trial within the block. This design allows examination of novelty driven attention because the nonrepeating, changing, object on each trial of a block was thought to be more salient than the repeating, stable, object. Analysis of infants’ looking times during testing revealed that only 19 of the 48 infants in the study looked longer at the correct referent than the distractor. The infants who looked longer at the correct referent during testing were classified as learners, and the remaining infants were considered nonlearners. Examination of infants’ looking behavior during training revealed that nonlearners tended to fixate on the changing object, while learners tended to look at both the stable and changing objects during each trial. The results suggest that novelty driven attention is not supportive of cross-situational word learning, and that more flexible attention is beneficial for learning. This finding could be interpreted to be inconsistent with the results of Yu and Smith (2011), in which longer fixations and fewer gaze-shifts were associated with learning word-referent mappings. However, these two studies differ in the saliency of the objects across learning. Taken together the results suggest that when possible referents for labels are equally novel, fixed visual attention is adaptive for learning words. However, when one object is more salient, more flexible attention benefits learning. Infants in the real world are often presented learning events for word-object pairings with temporal gaps between the presentations of the word-object pairings. Thus, Vlach and Johnson (2013) examined infants’ cross-situational learning to test infants’ learning at different timescales. Sixteen- and 20-month-old infants were trained on 12 cross-situational wordreferent pairings; 6 of the word-referent pairings were presented in the massed schedule and 6 were presented in the interleaved schedule. The training trials were divided into six blocks so that one massed referent was present on all six trials in a block, and each of the interleaved referents appeared once in each of the blocks. Importantly, the interleaved referents appeared in a fixed order across blocks, so that the interleaved object that appeared in the first trial of each block

Statistical learning mechanisms in infancy Chapter | 15

327

was the same. The side of presentation for the massed and interleaved objects was random across trials and blocks. Thus, this study differed from that of Smith and Yu (2013) in that the massed, or stable, referent changed location across training, and the interleaved, or changing, object occurred in a fixed order across blocks so the same number of trials intervened between the presentations of each interleaved object. The 16- and 20-month-old infants looked for similar amounts of time during training, suggesting that overall levels of attention did not differ as a function of age. However, there were differences in infants’ selective attention. Older infants tended to look equally at the massed and interleaved objects across trials within a block, while the younger infants initially looked more at the massed objects in the beginning of a block of trials and then switched to looking more at the interleaved objects at the end of the block. Furthermore, the 16-month-olds only learned the massed word-referent mappings, whereas the 20-month-olds learned both the massed and interleaved word-referent mappings. The results of this study provides additional support for the view that flexibility in visual attention is important for language development. The fact that the 16-month-olds were able to learn the massed, but not the interleaved, referents suggests that differences in memory abilities may also explain differences in infants’ cross-situational learning. The 16-month-olds’ inability to learn the interleaved referents may suggest that their memory systems are not yet capable of integrating information about word-referent co-occurrences over long delays (Vlach and Johnson, 2013). Recent work with toddlers and preschoolers suggests that children’s memory abilities are related to their cross-situational statistical learning (Vlach and DeBrock, 2017). Children’s cross-situational statistical learning was predicted by their word recognition memory, object recognition memory, and their word-object binding memory. Importantly children’s memory predicted children’s crosssituational statistical learning even after accounting for their age. These results suggest that memory abilities support cross-situational learning processes.

15.3 Linking individual differences in statistical learning to language development Determining whether there are individual differences in statistical learning, and the extent to which such differences are related to language development, has both theoretical and practical implications for accounts of language learning. For example, if statistical learning mechanisms play a substantive role in language development, then it should be possible to link individual differences in language development to statistical learning processes. Studying individual differences in statistical learning may also shed light on the question of whether statistical learning is domain general, or whether statistical learning is carried out independently in different sensory modalities, or for specific aspects of language development. In addition, such work has potentially shed light on whether difficulties with statistical learning could explain some aspects of delayed language development in individuals with clinical diagnoses, such as Specific Language Impairment (SLI) or Autism Spectrum Disorder (ASD). There has been debate about the extent to which there is meaningful individual variation in statistical learning in the first place. For example, Reber (1993) suggested that statistical learning is a more implicit and primitive form of learning, and thus that it ought to be relatively invariant across individuals, as well as independent from factors typically associated with explicit learning (e.g., learners’ age, intelligence, and language disorder). In support of such predictions Reber et al. (1991) found that individuals’ performance on an implicit learning task requiring classification of a subsequent letter in a string generated from an artificial grammar was significantly less variable than their performance on a task requiring explicit access to letter-ordering rules. However, these findings do not directly tell us whether statistical learning ability itself varies across individuals, and whether such variation relates to language proficiency. In fact, recent studies suggest that there may be meaningful variation in statistical learning processes. For example, Misyak and Christiansen (2012) found that individuals who performed better on artificial language-learning tasks involving adjacent and nonadjacent dependencies also performed better on a reading comprehension task. This correlation held even after controlling for verbal working memory and vocabulary size. However, verbal working memory was correlated with both statistical learning and reading comprehension, and was no longer correlated with reading comprehension after the effect of statistical learning was accounted for. This may suggest that the relation between statistical learning and language skills reflect the operation of explicit learning systems (i.e., working memory). Importantly however, the link between working memory and statistical learning has not always held across studies (e.g., Kidd and Arciuli, 2016). Further research is therefore necessary to better assess whether verbal working memory underlies the link between statistical learning and language outcome. In another recent work, variation in adults’ statistical learning ability has also been linked with factors related to language comprehension and use, such as vocabulary size, reading pace, speech perception, and learning the orthography of a second language (Misyak and Christiansen, 2012; Misyak et al., 2010; Conway et al., 2010; Frost et al., 2013). Taken together, these findings suggest that statistical learning ability may contribute to variation in language skills.

328 PART | II Cognitive development

While not extensive, there is some research addressing the extent to which there are individual differences in infants’ sensitivity to statistical structure, and linking these differences to specific aspects of language development. Interestingly, several studies have assessed this using visual statistical learning tasks. For example, in a study by Shafto et al. (2012) infants were shown sets of three geometric objects, one at a time, at three separate locations. Critically, the second object always predicted the location of the third object. Thus, infants could anticipate the location of the third object, which would be evidenced by more rapid shifting of attention to that location over the course of the experiment. As expected, infants shifted their attention more quickly to the location of the third object in anticipation of its appearance, and this anticipatory measure correlated with vocabulary size at 8.5 months of age. Variation in visual statistical learning is also correlated with language development later in childhood (Kidd and Arciuli, 2016). In this study, 7- to 8-year-olds were given a statistical learning task with triplet sequences of “alien” characters as well as a picture-pointing task testing comprehension of phrases embedded with passives and object relative clauses. These particular syntactic structures were chosen given their relatively low frequency in spoken English, and the relatively large variability in how well children comprehend them. As expected, performance on the visual statistical learning task correlated with comprehension of these constructions, but not with comprehension of simpler sentence structures. In contrast, performance on the statistical learning task was not related to vocabulary size. These findings suggest that the relation between statistical learning and language development is not static, and specifically that statistical learning appears to matter more for word learning in early childhood compared to later childhood, at which point factors such as formal literacy instruction begins to play a larger role in lexical development. On the other hand, greater sensitivity to probabilistic structure may continue to support syntactic development across childhood. The results of both studies may have implications for questions about whether the learning mechanism involved in tracking statistical regularities in vision and audition are overlapping versus distinct. On the one hand, these could be taken to suggest that the same mechanism is involved in tracking regularities within sequences of shapes and within sequences of sounds. Alternatively, they could be taken to suggest that visually based statistical regularities are relevant to learning words and grammatical patterns. Infants’ performance on several tasks assessing the ability to learn statistical regularities in the auditory domain has also been linked to their native-language development. For example, Estes et al. (2016) found that 18- and 19-month-olds who performed worse on a statistical learning task using nonnative phonotactics tended to have larger vocabularies, as assessed by a parent-report survey. These results suggest that mastery of native phonotactics interfered with infants’ ability to learn patterns that violate native phonotactics. These results are in line with evidence that enhanced discrimination of phonemic contrasts relevant to one’s native language comes at the cost of discriminating nonnative contrasts (Kuhl et al., 1992), but that this specialization supports later lexical development (Kuhl et al., 2005). Infants’ ability to learn statistics in spoken language materials has been linked not just to vocabulary size, but also to early grammatical development. For example, Lany (2014) found that infants who are better able to capitalize on statistical regularities marking word categories in the service of learning words (in the task used by Lany and Saffran, 2010, described above) are more likely to combine words in their own speech. Likewise, Lany and Shoaib (2020) found that infants who are better able to learn nonadjacent dependencies at 15 months in the artificial language task used by Gómez (2002) are more sensitive to native-language nonadjacent dependencies at 18 months, though this relation only emerged for female infants. Another way that statistical learning may be related to language development is by promoting the ability to encode speech in real time. Spoken language is fast and ephemeral, something that is painfully clear when listening to a language we are just beginning to learn. Thus, attaining a reasonable understanding of speech in real time is a major developmental achievement. Infants who perform better on two classic statistical learning tasks, one testing the ability to use statistical regularities to find words in fluent speech, and the other testing the ability to learn how words can be combined into grammatical patterns, are also faster to comprehend simple sentences in their native language (Lany et al., 2018). Altogether, these findings suggest that there is meaningful variation in statistical learning ability, and that it contributes to their language proficiency. However, much more work needs to be done to determine the relative importance of statistical learning in language development, and how it is related to other processes that support language learning.

15.4 Statistical learning in individuals with language delays and disorders Examining statistical learning abilities in individuals diagnosed with language delays or disorders can shed light on the degree to which it contributes to language-learning difficulties characteristic of different language disorders. For example, individuals diagnosed with developmental dyslexia, a disorder marked by heightened difficulty with phonological processing, tend to perform worse on statistical learning tasks that involve detecting the reliable co-occurrence of syllables or tones (Gabay et al., 2015). Such findings suggest that difficulties with statistical learning are relevant to language outcome in developmental dyslexia.

Statistical learning mechanisms in infancy Chapter | 15

329

Studies of statistical learning in cases of language delay and disorder may also shed light on the extent to which statistical learning plays a role in specific aspects of language development (e.g., learning vocabulary vs. syntax). For example, performance on visual statistical learning tasks has been linked with language outcomes in individuals diagnosed with SLI, but not in individuals with ASD. These findings may suggest that statistical learning plays a role in the languagelearning difficulties associated with SLI, but not in those characteristic of ASD (Zwart et al., 2017; Obeid et al., 2016). To elaborate, SLI is typically characterized by deficits in morphosyntax, vocabulary size, and motor control, all of which have been suggested to be supported by statistical learning processes (Ullman, Pierpont, 2005; Evans et al., 2009). Consistent with this suggestion, children diagnosed with SLI perform at chance levels on statistical learning tasks that entail tracking TPs among syllables and tones. Interestingly children with SLI performed above chance on the same statistical learning task when given increased exposure, the performance of which correlated with measures of receptive and expressive vocabulary. These results suggest that children with SLI are capable of learning TPs, and that statistical learning may play a role in their language development that is similar to the role it plays in typically developing children. However, they also suggest that children with SLI may require more experience with statistical regularities than typically developing children in order to learn them to the same degree. In contrast, children with ASD have deficits in semantic organization and pragmatic skills, but fairly typical patterns of vocabulary growth and grammatical development. A recent metaanalysis of 27 studies (13 on ASD, 14 on SLI) revealed no difference between statistical learning in children with ASD compared to controls, but did find evidence of poorer performance in children with SLI compared to controls (Obeid et al., 2016). Thus, children with ASD display atypical processing of semantic concepts but spared processing of syntax and sensitivity to sequential structure in language. On the other hand, children with SLI exhibit deficits in both vocabulary and grammatical development, which appear to be related to their performance on statistical learning tasks. A potential implication of these findings is that statistical learning may be more relevant for learning word forms and grammatical patterns, than for developing a lexicon with traditional semantic structure and organization. These conclusions are complicated by evidence that at least some children with ASD may extract probabilistic structure from the environment using different learning mechanisms than individuals who do not have ASD. A recent study found that children diagnosed with ASD, and who had high nonverbal IQ, allocated greater attention to unexpected sequences of shapes, as indexed by larger N1 responses, compared to children with ASD who were low on nonverbal IQ and typically developing controls (Jeste et al., 2015). Children with high nonverbal IQs also performed better on the statistical learning task. These findings may suggest that children with high nonverbal IQs used a compensatory process in the statistical learning task, and that children with ASD may in fact have some difficulty with statistical learning. Finally, these results highlight the importance of examining individual differences within children diagnosed with ASD, as the relation between statistical learning and nonverbal IQ in children with ASD was not found when averaging performance across children with ASD. More research is therefore necessary to generate an accurate picture of the statistical learning abilities of children with ASD, and whether they are related to language-learning difficulties. Here we have outlined the evidence that there is a link between statistical learning and at least some forms of atypical language development. In particular, children with SLI have been shown to perform consistently worse on statistical learning tasks, which is predictive of characteristic impairments in language for these children. However, there is less evidence that children with ASD have difficulty with statistical learning, suggesting that the social and conceptual difficulties observed in this population may not be related to statistical learning processes.

15.5 Scaling statistical learning to real-world challenges How do infants’ statistical learning abilities stand up to the challenges they face when learning language in the real world? Several studies have tested whether infants track statistics in more complex learning environments. One example comes from studies of infants’ ability to track TPs in speech. Infants as young as 8 months readily track TPs in laboratory settings distinguishing between HTP and LTP sequences in artificial languages (e.g., Aslin et al., 1998; Saffran et al., 1996). Importantly, artificial languages are nowhere near as complex as natural languages. Infants may be able to track statistics in highly simplified situations, but are they also capable of doing so in natural languages? Pelucchi et al. (2009a) tested whether infants can track TPs in a natural language. English-learning 8-month-olds were presented with Italian sentences that contained two words with HTPs and two words with LTPs. Following the 2-minute familiarization period, infants were tested on their ability to discriminate the HTP and LTP words. Infants differentiated the HTP and LTP words, suggesting that infants are able to track TPs in complex natural languages. Additional studies have found that 8-month-old infants can track backward TPs in Italian (Pelucchi et al., 2009b) and that 17-month-old infants will map HTP sequences as labels to referents (Hay et al., 2011). These results suggest that infants are capable of tracking TPs in natural languages, and using them in the service of learning words.

330 PART | II Cognitive development

However, infants need to be able to generalize word forms so that they can recognize that a word produced by different individuals is still the same word, even if it sounds a little different. The majority of statistical learning studies expose infants to a stream of speech produced by a single individual and test infants with tokens produced by the same individual. Thus, it is unknown if infants are able to track TPs in speech produced by multiple speakers and generalize the tokens to a new speaker, a task that infants face in the real world. Graf Estes and Lew-Williams (2015) tested whether infants can track TPs in an artificial language produced by two different speakers, and also by eight different speakers. Eight- and 10-monthold infants who heard the artificial language produced by eight different speakers discriminated HTP and LTP words at test. Moreover, these infants were able to generalize their learning and discriminate HTP and LTP words at test when they were produced by a new speaker. However, when the speech stream was produced by only two different speakers, infants did not discriminate HTP and LTP words at test, and were not able to generalize to a new speaker. This suggests that infants are able to use statistical learning cues to segment speech in situations with high variability, but not low variability. High variability may promote detection of underlying commonalities between the speakers whereas low variability may highlight differences between the speakers. Thus, infants are able to track TPs in speech produced by multiple speakers, when the variability is high, or by a single speaker, but have difficulty in instances of low variability, when the speech is produced by two speakers. Given that infants typically hear speech from more than two individuals, these findings suggest that infants can track statistical regularities in situations that mirror their real-world learning environment. However, they also suggest that infants’ experience with speech from a single, highly familiar primary caregiver can contribute to their ability to learning the statistical regularities of their native language. Another question about how well statistical learning abilities scale up to real-world challenges pertains to infants’ ability to remember statistical structure across delays. Specifically, relatively little is known about how well infants remember words they have segmented through statistical learning. In the studies reviewed so far, infants were tested immediately after a training period, and only a handful of studies to date have incorporated retention tests. In one such study, Karaman and Hay (2018) tested 8-month-olds’ retention of HTP and LTP words after a 10-minute delay. Infants failed to discriminate HTP and LTP words at test after the delay, suggesting memory for the HTP sequences was not very strong. However, when infants were given additional exposure to both the isolated HTP and LTP words immediately after familiarization, they were able to discriminate the HTP and LTP words at test after a 10-minute delay. Simon et al. (2017) tested 6.5-month-old infants’ retention of HTP and LTP words over a delay that corresponded to a period of sleep or wakefulness. Infants were time-matched for the delay between learning and test for the sleep and wakefulness groups. Infants who slept discriminated HTP and LTP words, while infants who remained awake did not. Similar benefits of sleep have been found with 15-month-olds’ retention of nonadjacent dependencies. Fifteen-monthold infants are able to remember nonadjacent dependencies after a 4-hour delay between familiarization and test; however, only infants who napped during the 4-hour delay were able to abstract the statistical patterns and generalize them to a similar pattern (Gómez et al., 2006). In a follow-up study, 15-month-old infants who napped in the 4 h following familiarization of nonadjacent dependencies were able to abstract and generalize the statistical patterns to new stimuli 24 hours later (Hupbach et al., 2009). Taken together, these studies suggest that infants are able to remember statistical structure, and that sleep promotes retention and generalization of such structure. In sum, there is a growing body of evidence to suggest that infants’ statistical learning abilities are up to the challenges they face in their everyday lives. Infants are able to track statistics in natural languages and across speakers as well as remember statistical patterns over delays. However, these are just a few of the obstacles infants face, and it will be important to further investigate how infants’ statistical learning abilities scale up to the challenges infants face in their natural environment.

15.6 Conclusions The research reviewed in this chapter suggests that infants’ sensitivity to statistical structure is powerful and nuanced, and plays a role in diverse aspects of development. Infants’ environments are rich with meaningful statistical structure, and infants readily track it. Far from being limited to simple associative learning, such as the stimulus-response learning of classical learning theory, the studies reviewed in this chapter reveal a sensitivity to statistical structure that is powerful and nuanced, capable of tracking fine-grained detail but also forming generalizations. Advances in the study of statistical learning mechanisms have shed new light on many aspects of development, such as auditory and visual perception, language development, and event processing. Continued study of infants’ statistical learning mechanisms holds great promise for advancing knowledge of cognitive development.

Statistical learning mechanisms in infancy Chapter | 15

331

Acknowledgment While psychologists long imagined that infants experience the world as a bewildering array of sights, sounds, and sensations (James, 1890), research since the 1950s has shown that infants have more advanced perceptual and cognitive abilities than was once thought. For example, infants’ visual, auditory, tactile, and vestibular systems are functional at birth (and even beforehand), and their discrimination abilities in these areas, while not adultlike, are nonetheless remarkably acute (Kellmann and Arterberry, 2000). Perhaps, then, it should not be surprising that very young infants show evidence of learning from stimuli in their environment. In this chapter we will review recent research focused on infants’ ability to learn several kinds of statistical structure, such as probability distributions, sequential structure, and associations between properties within and across instances and modalities. These types of statistical structures appear to play a role in many aspects of development, including auditory and visual perception, language development, object perception, and event processing. We focus much of our discussion on the role of these learning mechanisms in language acquisition, though we also make connections between language learning and learning in other domains. In addition, we consider several controversial questions, such as what the mechanistic underpinnings of statistical learning may be, how much infants can really learn about things like words, objects, and events from statistics, whether they can track statistics outside of simple lab tasks, and what individual differences in statistical learning ability can tell us about language development. Preparation of this chapter was supported by NSF BCS 1352443 to J.L.

References Aslin, R.N., Jusczyk, P.W., Pisoni, D.B., 1998. Speech and auditory processing during infancy: constraints on and precursors to language. In: Kuhn, D., Siegler, R. (Eds.), Handbook of Child Psychology: Cognition, Perception, and Language, vol. 2. Wiley, New York, pp. 147e254. Aslin, R.N., Saffran, J.R., Newport, E.L., 1998. Computation of conditional probability statistics by 8-month-old infants. Psychol. Sci. 9, 321e324. Baldwin, D., Andersson, A., Saffran, J.R., Meyer, M., 2008. Segmenting dynamic human action via statistical structure. Cognition 106, 1382e1407. Braine, M.D.S., 1987. What is learned in acquiring words classes: a step toward acquisition theory. In: MacWhinney, B. (Ed.), Mechanisms of Language Acquisition. Erlbaum, Hillsdale, NJ, pp. 65e87. Bulf, H., Johnson, S.P., Valena, E., 2011. Visual statistical learning in the newborn infant. Cognition 121, 127e131. Cartwright, T.A., Brent, M.R., 1997. Syntactic categorization in early language acquisition: formalizing the role of distributional analysis. Cognition 63, 121e170. Cheour, M., Ceponiene, R., Lehtokoski, A., Luuk, A., Allik, J., Alho, K., Naatanen, R., 1998. Development of language-specific phoneme representations in the infant brain. Nat. Neurosci. 1, 351e353. Christiansen, M.H., Allen, J., Seidenberg, M.S., 1998. Learning to segment speech using multiple cues: a connectionist model. Lang. Cogn. Process. 13 (2e3), 221e268. Conway, C.M., Bauernschmidt, A., Huang, S.S., Pisoni, D.B., 2010. Implicit statistical learning in language processing: word predictability is the key. Cognition 114 (3), 356e371. Eimas, P.D., Siqueland, E.R., Jusczy k, P., Vigorito, J., 1971. Speech perception in infants. Science 171, 303e306. Endress, A.D., Mehler, J., 2009. The surprising power of statistical learning: when fragment knowledge leads to false memories of unheard words. J. Mem. Lang. 60, 351e367. Estes, K.G., Gluck, S.C.W., Grimm, K.J., 2016. Finding patterns and learning words: infant phonotactic knowledge is associated with vocabulary size. J. Exp. Child Psychol. 146, 34e49. Evans, J.L., Saffran, J.R., Robe-Torres, K., 2009. Statistical learning in children with specific language impairment. J. Speech Lang. 52 (2), 321e335. Fantz, R.L., 1964. Visual experience in infants: decreased attention to familiar patterns relative to novel ones. Science 146, 668e670. Farmer, T.H., Christiansen, M.H., Monaghan, P., 2006. Phonological typicality influences on-line sentence comprehension. Proc. Natl. Acad. Sci. 103, 12203e12208. Fiser, Aslin, R.N., 2002. Statistical learning of new visual feature combinations by infants. Proc. Natl. Acad. Sci. 99, 15822e15826. Frost, R., Siegelman, N., Narkiss, A., Afek, L., 2013. What predicts successful literacy acquisition in a second language? Psychol. Sci. 24 (7), 1243e1252. Gabay, Y., Thiessen, E.D., Holt, L.L., 2015. Impaired statistical learning in developmental dyslexia. J. Speech Lang. Hear. Res. 58 (3), 934e945. Giroux, I., Rey, A., 2009. Lexical and sublexical units in speech perception. Cognitive Science 33 (2), 260e272. Gómez, R.L., 2002. Variability and detection of invariant structure. Psychol. Sci. 13, 431e436. Gómez, R.L., Bootzin, R.R., Nadel, L., 2006. Naps promote abstraction in language-learning infants. Psychol. Sci. 17 (8), 670e674. Gómez, R.L., Lakusta, L., 2004. A first step in form-based category abstraction in 12-month-old infants. Dev. Sci. 7, 567e580. Gómez, R.L., Maye, J., 2005. The developmental trajectory of nonadjacent dependency learning. Infancy 7, 183e206. Gómez, R.L., Gerken, L.A., 1999. Artificial grammar learning by one-year-olds leads to specific and abstract knowledge. Cognition 70, 109e135. Graf Estes, K., Evans, J.L., Alibali, M.W., Saffran, J.R., 2007. Can infants map meaning to newly segmented words? Psychol. Sci. 18, 254e260. Graf Estes, K., Lew-Williams, C., 2015. Listening through voices: infant statistical word segmentation across multiple speakers. Dev. Psychol. 51 (11), 1517e1528. Hay, J.F., Pelucchi, B., Graf Estes, K., Saffran, J.R., 2011. Linking sounds to meanings: infant statistical learning in a natural language. Cogn. Psychol. 63 (2), 93e106. Hupbach, A., Gómez, R.L., Bootzin, R.R., Nadel, L., 2009. Nap-dependent learning in infants. Dev. Sci. 12 (6), 1007e1012. James, W., 1890. The Principles of Psychology. Holt, New York.

332 PART | II Cognitive development

Jeste, S.S., Kirkham, N., Senturk, D., Hasenstab, K., Sugar, C., Kupelian, C., et al., 2015. Electrophysiological evidence of heterogeneity in visual statistical learning in young children with ASD. Dev. Sci. 18 (1), 90e105. Johnson, E.K., Tyler, M.D., 2010. Testing the limits of statistical learning for word segmentation. Dev. Sci. 13, 339e345. Karaman, F., Hay, J.F., 2018. The longevity of statistical learning: when infant memory decays, isolated words come to the rescue. J. Exp. Psychol. Learn. Mem. Cogn. 44 (2), 221e232. Kellmann, P.J., Arterberry, M.E., 2000. The Cradle of Knowledge. The MIT Press, Cambridge, MA. Kelly, M.H., 1992. Using sound to solve syntactic problems: the role of phonology in grammatical category assignments. Psychol. Rev. 99 (2), 349. Kidd, E., Arciuli, J., 2016. Individual differences in statistical learning predict children’s comprehension of syntax. Child Dev. 87 (1), 184e193. Kirkham, N.Z., Slemmer, J.A., Johnson, S.P., 2002. Visual statistical learning in infancy: evidence for a domain general learning mechanism. Cognition 83, B35eB42. Kuhl, P.K., Conboy, B.T., Padden, D., Nelson, T., Pruitt, J., 2005. Early speech perception and later language development: implications for the" critical period. Lang. Learn. Dev. 1 (3e4), 237e264. Kuhl, P.K., Williams, K.A., Lacerda, F., Stevens, K.N., Lindblom, B., 1992. Linguistic experience alters phonetic perception in infants by 6 months of age. Science 255, 606e608. Kuhl, P.K., 2004. Early language acquisition: cracking the speech code. Nat. Rev. Neurosci. 5, 831e843. Kuhl, P.K., Tsao, F.M., Liu, H.M., 2003. Foreign-language experience in infancy: effects of short-term exposure and social interaction on phonetic learning. Proc. Natl. Acad. Sci. USA 100, 9096e9101. Lany, J., 2014. Judging words by their covers and the company they keep: probabilistic cues support word learning. Child Dev. 85 (4), 1727e1739. Lany, J., Gómez, R.L., 2008. Twelve-month-old infants benefit from prior experience in statistical learning. Psychol. Sci. 19, 1247e1252. Lany, J., Saffran, J.R., 2010. From statistics to meaning: infant acquisition of lexical categories. Psychol. Sci. 21, 284e291. Lany, J., Giglio, M., Oswald, M., 2018. Infants’ lexical processing efficiency is related to vocabulary size by one year of age. Infancy 23 (3), 342e366. Lany, J., Shoaib, A., 2020. Individual differences in non-adjacent statistical dependency learning in infants. Journal of Child Language 47 (2), 483e507. Lisker, L., Abramson, A.S., 1964. A cross-language study of voicing in initial stops: acoustical measurements. Word 20, 384e482. Marcovitch, S., Lewkowicz, D.L., 2009. Sequence learning in infancy: the independent contributions of conditional probability and pair frequency information. Dev. Sci. 12, 1020e1025. Maye, J., Weiss, D.J., Aslin, R.N., 2008. Statistical phonetic learning in infants: facilitation and feature generalization. Dev. Sci. 11, 122e134. Maye, J., Werker, J.F., Gerken, L., 2002. Infant sensitivity to distributional information can affect phonetic discrimination. Cognition 82, B101eB111. Mintz, T.H., 2003. Frequent frames as a cue for grammatical categories in child directed speech. Cognition 90, 91e117. Mintz, T.H., Newport, E.L., Bever, T.G., 2002. The distributional structure of grammatical categories in speech to young children. Cogn. Sci. 26, 393e424. Misyak, J.B., Christiansen, M.H., 2012. Statistical learning and language: an individual differences study. Lang. Learn. 62 (1), 302e331. Misyak, J.B., Christiansen, M.H., Tomblin, J.B., 2010. On-line individual differences in statistical learning predict language processing. Front. Psychol. 1, 31. Monaghan, P., Chater, N., Christiansen, M.H., 2005. The differential role of phonological and distributional cues in grammatical categorization. Cognition 96, 143e182. Monroy, C.D., Gerson, S.A., Domínguez-Martínez, E., Kaduk, K., Hunnius, S., Reid, V., 2019. Sensitivity to structure in action sequences: an infant event-related potential study. Neuropsychologia 126, 92e101. Mueller, J.L., Friederici, A.D., Männel, C., 2012. Auditory perception at the root of language learning. Proc. Natl. Acad. Sci. 109 (39), 15953e15958. Obeid, R., Brooks, P.J., Powers, K.L., Gillespie-Lynch, K., Lum, J.A., 2016. Statistical learning in specific language impairment and autism spectrum disorder: a meta-analysis. Front. Psychol. 7, 1245. Onnis, L., Thiessen, E., 2013. Language experience changes subsequent learning. Cognition 126 (2), 268e284. Pelucchi, B., Hay, J.F., Saffran, J.R., 2009. Learning in reverse: eight-month-old infants track backwards transitional probabilities. Cognition 11, 244e247. Pelucchi, B., Hay, J.F., Saffran, J.R., 2009. Statistical learning in a natural language by 8-month-old infants. Child Dev. 80, 674e685. Perruchet, P., Vinter, A., 1998. PARSER: a model for word segmentation. J. Mem. Lang. 39 (2), 246e263. Quine, W.V.O., 1960. Word and Object. MIT Press, Cambridge, MA. Reber, A.S., 1993. Implicit learning and tacit knowledge: an essay on the cognitive unconscious. In: Oxford Psychology Series, No 19. Oxford University Press, New York ; Oxford. Reber, A.S., Walkenfeld, F.F., Hernstadt, R., 1991. Implicit and explicit learning: individual differences and IQ. J. Exp. Psychol. Learn. Mem. Cogn. 17 (5), 888e896. Redington, M., Chater, N., Finch, S., 1998. Distributional information: A powerful cue for acquiring syntactic categories. Distributional information: A powerful cue for acquiring syntactic categories. Cognitive Science 22 (4), 425e469. Rivera-Gaxiola, M., Silva-Pereyra, J., Kuhl, P.K., 2005. Brain potentials to native and non-native speech contrasts in 7- and 11-month-old American infants. Dev. Sci. 8, 162e172. Roseberry, S., Richie, R., Hirsh-Pasek, K., Golinkoff, R.M., Shipley, T.F., 2011. Babies catch a break: 7-to 9-month-olds track statistical probabilities in continuous dynamic events. Psychol. Sci. 22 (11), 1422e1424. Saffran, J.R., Aslin, R.N., Newport, E.L., 1996. Statistical learning by 8-month-old infants. Science 274, 1926e1928.

Statistical learning mechanisms in infancy Chapter | 15

333

Saffran, J.R., Johnson, E.K., Aslin, R.N., Newport, E.L., 1999. Statistical learning of tone sequences by human infants and adults. Cognition 70 (1), 27e52. Shafto, C.L., Conway, C.M., Field, S.L., Houston, D.M., 2012. Visual sequence learning in infancy: domain-general and domain-specific associations with language. Infancy 17 (3), 247e271. Shoaib, A., Wang, T., Hay, J.F., Lany, J., 2018. Do infants learn words from statistics? Evidence from English-learning infants hearing Italian. Cogn. Sci. 42 (8), 3083e3099. Simon, K.N., Werchan, D., Goldstein, M.R., Sweeney, L., Bootzin, R.R., Nadel, L., Gómez, R.L., 2017. Sleep confers a benefit for retention of statistical language learning in 6.5 month old infants. Brain Lang. 167, 3e12. Smith, L.B., Yu, C., 2008. Infants rapidly learn word-referent mappings via cross-situational statistics. Cognition 106, 1558e1568. Smith, L.B., Yu, C., 2013. Visual attention is not enough: individual differences in statistical word-referent learning in infants. Lang. Learn. Dev. 9, 25e49. Smith, K.H., 1969. Learning co-occurrence restrictions: rule learning or rote learning? J. Verb. Behav. 8, 319e321. Stahl, A.E., Romberg, A.R., Roseberry, S., Golinkoff, R.M., Hirsh-Pasek, K., 2014. Infants segment continuous events using transitional probabilities. Child Dev. 85 (5), 1821e1826. Swingley, D., 2005. Statistical clustering and the contents of the infant vocabulary. Cognitive Psychology 50 (1), 86e132. Teinonen, T., Fellman, V., Näätänen, R., Alku, P., Huotilainen, M., 2009. Statistical language learning in neonates revealed by event-related brain potentials. BMC Neurosci. 10, 21. Thiessen, E.D., Onnis, L., Hong, S.J., Lee, K.S., 2019. Early developing syntactic knowledge influences sequential statistical learning in infancy. J. Exp. Child Psychol. 177, 211e221. Thiessen, E.D., Kronstein, A.T., Hufnagle, D.G., 2013. The extraction and integration framework: a two-process account of statistical learning. Psychol. Bull. 139 (4), 792. Thiessen, E.D., 2017. What’s statistical about learning? Insights from modelling statistical learning as a set of memory processes. Phil. Trans. R. Soc. B 372 (1711), 20160056. Tummeltshammer, K., Amso, D., French, R.M., Kirkham, N.Z., 2017. Across space and time: infants learn from backward and forward visual statistics. Dev. Sci. 20 (5), e12474. Ullman, M.T., Pierpont, E.I., 2005. Specific language impairment is not specific to language: the procedural deficit hypothesis. Cortex 41 (3), 399e433. Vlach, H.A., DeBrock, C.A., 2017. Remember dax? Relations between children’s cross-situational word learning, memory, and language abilities. J. Mem. Lang. 93, 217e230. Vlach, H.A., Johnson, S.P., 2013. Memory constraints on infants’ cross-situational statistical learning. Cognition 127, 375e382. Waismeyer, A., Meltzoff, A.N., Gopnik, A., 2015. Causal learning from probabilistic events in 24-month-olds: an action measure. Dev. Sci. 18 (1), 175e182. Werker, J.F., Tees, R.C., 1984. Cross-language speech perception: evidence for perceptual reorganization during the first year of life. Infant Behav. Dev. 7, 49e63. Willits, J., Saffran, J.R., Lany, J., 2017. Toddlers Can Use Semantic Cues to Learn Difficult Nonadjacent Dependencies. PsyArXiv. Wonnacott, E., Newport, E.L., Tanenhaus, M.K., 2008. Acquiring and processing verb argument structure: distributional learning in a miniature language. Cogn. Psychol. 56, 165e209. Yoshida, K.A., Pons, F., Maye, J., Werker, J.F., 2010. Distributional phonetic learning at 10 Months of age. Infancy 15 (4), 420e433. Yu, C., Smith, L.B., 2011. What you learn is what you see: using eye movements to study infant cross-situational word learning. Dev. Sci. 14, 165e180. Zwart, F.S., Vissers, C.T.W., Kessels, R.P., Maes, J.H., 2017. Procedural learning across the lifespan: a systematic review with implications for atypical development. J. Neuropsychol. https://doi.org/10.1111/jnp.12139.

Chapter 16

Development of the visual system Scott P. Johnson University of California, Los Angeles, CA, United States

Chapter outline 16.1. Classic theoretical accounts 16.1.1. Piagetian theory 16.1.2. Gestalt theory 16.2. Prenatal development of the visual system 16.2.1. Development of structure in the visual system 16.2.2. Prenatal visual function 16.3. Visual perception in the newborn 16.3.1. Visual organization at birth 16.3.2. Visual behaviors at birth 16.3.3. Faces and objects 16.4. Postnatal visual development 16.4.1. Visual physiology 16.4.2. Critical periods 16.4.3. Development of visual attention 16.4.4. Cortical maturation and oculomotor development

336 336 337 337 338 338 339 340 340 340 340 342 342 343 343

16.4.5. 16.4.6. 16.4.7. 16.4.8. 16.4.9.

Development of visual memory 345 Development of visual stability 345 Object perception 346 Face perception 347 Critical period for development of holistic perception 347 16.5. How infants learn about objects 349 16.5.1. Learning from targeted visual exploration 349 16.5.2. Learning from associations between visible and occluded objects 350 16.5.3. Learning from visual-manual exploration 350 16.5.4. Hormonal and environmental influences on object perception 352 16.6. Summary and conclusions 353 References 355

The purpose of vision is to obtain information about the surrounding environment so that we may plan appropriate actions. Consider, for example, the view through the windshield when driving (Fig. 16.1). The driver must detect and react to the road and any possible obstacles, accommodating changes of direction and avoiding objects in the path; thus visual information helps guide decisions about where to steer and when to accelerate or brake. To remain safe, therefore, the driver must know what the risks are, and this invariably involves knowing what objects there are in the visual scene. The importance of accurate perception of our surroundings is attested by the allotment of cortical tissue devoted to vision: By some estimates, over 50% of the cortex of the macaque monkey (a phylogenetically close cousin to Homo sapiens) is involved in visual perception, and there are perhaps 30 distinct cortical areas that participate in visual or visuomotor processing (Felleman and Van Essen, 1991; Van Essen et al., 1992). This chapter will review theory and data concerning development of the human visual system with an emphasis on object perception. As we will see, infants are prepared to see objects and understand many of their properties (e.g., permanence, coherence) well in advance of locomotion, so that by the time infants begin to crawl and walk, they have a good sense of what and where obstacles might be, even if the hazards these objects pose remain unknown. There is much else for infants to learn. Visual scenes, for example, tend to be very complex: a multitude of overlapping and adjacent surfaces with distinct shapes, colors, textures, and depths relative to the observer. Yet our visual experience as adults is not one of incomplete fragments of surfaces, but instead one of objects, most of which have a shape that can be inferred from partial views and incomplete information. Is the infant’s visual system sufficiently functional and organized to make sense of the world from the onset of visual experience at birth, able to bind shapes, colors, and textures into coherent forms, and to perceive objects as regular and predictable and complete across space and time? Or does the infant’s visual system require a period of maturation and experience within which to observe and learn about the world?

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00016-0 Copyright © 2020 Elsevier Inc. All rights reserved.

335

336 PART | II Cognitive development

FIGURE 16.1 Incoming visual information helps us guide our actions.

These “nature versus nurture” questions begin to lose their force when the details of visual development are examined and explained, because visual development stems from growth, maturation, and experience from learning and action; all happen simultaneously and all influence one another. Infants free of disability or developmental delay are born with a functional visual system that is prepared to contribute in important ways to learning, but incapable of perceiving objects in an adultlike fashion. Developmental processes that lead to mature perception and interpretation of the visual world as coherent, stable, and predictable constitute an area of active investigation and are beginning to be understood.

16.1 Classic theoretical accounts Discussions of the nature versus nurture of cognitive development are entrenched and persistent. Such discussions are particularly vigorous when concerning infant cognition, and have tended to be long on rhetoric but short on evidence, in part because the evidence has been, until recently, relatively sparse. Research on visual development, in contrast, has tended to focus on developmental changes in neural mechanisms, with much of the evidence coming from animal models (Kiorpes and Movshon, 2004; Teller and Movhson, 1986). Research on human infants’ visual development has often been motivated by two theoretical accounts, each of which considers seriously both the starting point for postnatal development and the mechanisms of change that yield stable, mature object perception: Piagetian theory and Gestalt theory.

16.1.1 Piagetian theory The first systematic study of infants’ perception and knowledge of objects was conducted by Jean Piaget in the 1920s and 1930s (Piaget, 1952/1936, 1954/1937). According to Piaget, knowledge of objects and space developed in parallel, and were interdependent: one cannot perceive or act on objects accurately without awareness of their position in space relative to other objects and to the observer. Knowledge of the self and of external objects as distinct, coherent, and permanent entities grew from active manual search, initiated by the child. When the child experiences her own movements, she comes to understand them as movements of objects through space, and induces the same knowledge to movements of other objects.

Development of the visual system Chapter | 16

337

Initially, prior to any manual action experience, infants understand the world as a “sensory tableaux” in which images shift unpredictably and lack permanence or substance; in an important sense, the world of objects that we take for granted does not yet exist. Active search behavior emerges only after 4 months, and marks the beginnings of “true” object knowledge. Over the next few months, infants reveal this knowledge, for example, by following the trajectory of thrown or dropped objects, and by retrieval of a desired object from under a cover when it had been seen previously. Later in infancy infants are able to search accurately for objects even when there are multiple potential hiding places, marking the advent of full “object permanence.” Piaget placed more emphasis on manual search than visual skills as holding an important role for developmental changes in object perception, yet the lessons from his theory for questions of development of the visual system could not be more relevant. Upon the infant’s first exposure to patterned visual input, he does not inhabit a world of objects, but rather a world of disconnected images devoid of depth, coherence, and permanence. Building coherent things from these disconnected images comes from action and experience with objects over time.

16.1.2 Gestalt theory Piagetian theory can be contrasted with a coeval, competing account. The Gestalt psychologists, unlike Piaget, were not strictly developmentalists, but they did have much to say about how visual experience might be structured in the immature visual system. They suggested that subjective experience corresponds to the simplest and most regular interpretation of a particular visual array in accord with a general “minimum principle,” or Prägnanz (Koffka, 1935). The relatively basic shapes of most objects are more coherent, regular, and simple than disconnected and disorganized forms. The minimum principle and Prägnanz were thought to be rooted in the tendency of neural activity toward minimum work and energy, which impel the visual system toward simplicity (Koffka, 1935). The minimum principle is a predisposition inherent in the visual system, and so it follows that young infants should experience the visual environment as do adults. In one of the few sections of Gestalt writings to focus on development, a “primitive mentality” was attributed to the human infant (Koffka, 1959/1928; Köhler, 1947), implying that one’s perceptual experience is never one of disorganized chaos, no matter what one’s position in the lifespan. Hebb (1949) noted, in addition, that the newborn infant’s electroencephalogram (i.e., recording of continuous brain activity) was organized and somewhat predictable, perhaps reflecting organized sensory systems at birth and serving as a stable foundation for subsequent perceptual development. Gibson (1950) suggested that visual experience begins with “embryonic meanings,” a position echoed by Zuckerman and Rock (1957), who argued that an organized world could not arise from experience in the form of memory for previously encountered scenes and objects, because experience cannot operate in an organized fashion over inherently disorganized inputs. Necessarily, therefore, the starting point of visual organization is inherently organized. Like Piaget, Gestalt psychologists proposed that development of object perception per se involved active manual exploration, which imparts additional information about specific object kinds (Koffka, 1959), but the starting point for visual experience is necessarily quite different on the two accounts. On the Gestalt view, perceptual organization precedes object knowledge; on the Piagetian view, object knowledge and perceptual organization develop in tandem. Piagetian and Gestalt accounts each specify a starting point for postnatal development, and each has particular views about how development of the infant’s visual world might proceed. Neither account is wholly on one side of the naturenurture issue, and both accounts have offered testable predictions that have guided subsequent research; and as we will see later in this chapter, both accounts have influenced important research on object perception in infants. Yet neither can be taken as complete, in part because neither took a sufficiently comprehensive approach to vision. A quote from Gibson (1979) helps explain why this is: the visual system comprises “the eyes in the head on a body supported by the ground, the brain being only the central organ of a complete visual system. When no constraints are put on the visual system, we look around, walk up to something interesting, and move around it so as to see it from all sides, and go from one vista to another” (p. 1). Vision is not passive, even in infancy; at no point in development are infants simply inactive recipients of visual stimulation. Instead, they are active perceivers, and active participants in their own development, even from before birth (von Hofsten, 2004). Young infants do not have all the action systems implied by Gibson’s quote at their disposal, but eye movements are a notable exception, and as we will see in later sections, there are strong reasons to suspect a critical role for oculomotor behavior as a means of cognitive development.

16.2 Prenatal development of the visual system The mammalian visual system, like other sensory and cortical systems, begins to take shape early in prenatal development. For example, in humans the retina starts to form around 40 days postconception and is thought to have a relatively

338 PART | II Cognitive development

complete set of cells by 160 days, though the growth of individual cells and their latticelike organization characteristic of mature structure continue to mature well past birth (Finlay et al., 2003). The distinction between foveal and extrafoveal regions (viz., what will become thalamus and cortex) is present early; like the retina, the topology and patterning of receptors and neurons continue to change throughout prenatal development and the first year after birth. Foveal receptors are overrepresented in the cortical visual system, and detailed information about different parts of the scene is enabled by moving the eyes to different points (see Section 16.4.3). The musculature responsible for eye movements develops before birth in humans, as do subcortical systems (e.g., superior colliculus and brainstem) to control these muscles (Johnson, 2001; Prechtl, 2001). Many developmental mechanisms are common across mammalian species, including humans, though the timing of developmental events varies (Clancy et al., 2000; Finlay and Darlington, 1995). Data from humans are sparse, but the few cases where deceased embryos and fetuses are available demonstrate that many major structures (neurons, areas, and layers) in visual cortical and subcortical areas are in place by the end of the second trimester in utero (e.g., Zilles et al., 1986). Later developments consist of the physical growth of neurons and the proliferation and pruning of synapses, which is, in part, activity-dependent (Greenough et al., 1987; Huttenlocher et al., 1986).

16.2.1 Development of structure in the visual system The visual system consists of a richly interconnected yet functionally segregated network of areas, specializing in processing different aspects of visual scenes and visually guided behavior: contours, motion, luminance, color, objects, faces, approach versus avoidance, and so forth. Areal patterns are present in rudimentary form during the first trimester but the final forms continue to take shape well after birth; like synaptic pruning, developmental processes are partly the result of experience. Some kinds of experience are intrinsic to the visual system, as opposed to outside stimulation. Spontaneous prenatal activity in visual pathways contributes to retinotopic mapping, the preservation of sensory structure (Sperry, 1963). Spontaneous activity begins in the retina and extends through the thalamus, primary visual cortex, and higher visual areas. Waves of coordinated, spontaneous firing of retinal cells have been observed in chicks and ferrets (Wong, 1999). Waves travel across the retinal surface and are then systematically propagated through to the higher areas. This might be one way by which correlated inputs remain coupled and dissimilar inputs become dissociated, prior exposure to light. As soon as neurons are formed, find their place in cortex, and grow, they begin to connect to other neurons. There is a surge in synaptogenesis in visual areas around the time of birth and then a more protracted period in which synapses are eliminated, reaching adultlike levels at puberty (Bourgeois et al., 2000). This process is activity-dependent: synapses are preserved in active cortical circuits and lost in inactive circuits. Auditory cortex, in contrast, experiences a synaptogenesis surge several months earlier, which corresponds to its earlier functionality relative to visual cortex (viz., prenatally). Here, too, pruning of synapses extends across the next several years. (In other cortical areas, such as frontal cortex, there is a more gradual accrual of synapses without extensive pruning.) For the visual system, the addition and elimination of synapses, the onset of which coincides with the start of visual experience, provides an important mechanism by which the cortex tunes itself to environmental demands and the structure of sensory input.

16.2.2 Prenatal visual function The visual system is sufficiently mature by the last trimester of pregnancy to produce responses to light introduced into the womb, typically by flashing a bright light adjacent to the mother’s abdomen. Fetal brain responses, for example, can be recorded with functional magnetoencephalography (fMEG), which measures the magnetic fields generated by neuronal activity in the brain of the fetus, or with functional magnetic resonance imaging (fMRI), which measures changes in blood flow in the brain resulting from cortical activity. Both fMEG and fMRI have shown that the fetal visual system detects light (Dunn et al., 2015). Moreover, cortical and subcortical structures that control eye and head movements begin to develop prior to birth. One of the most important structures is the superior colliculus (SC), a midbrain structure that supports processing of visual, auditory, and somatosensory inputs and coordinates them with topographically ordered somatic and cortical representations, to orient the observer toward stimuli. The SC begins functioning in the third trimester, and its spatial layout and layered structure (Fig. 16.2) have been proposed to orient the neonate (and perhaps the fetus) preferentially to facelike stimuli (Pitti et al., 2013). (This topic is discussed in greater detail in Sections 16.3.3 and 16.4.8). This proposal was tested recently by placing a triangular, three-diode light source against the abdomens of pregnant women late in the term, while taking care to place the lights either in an upright or inverted orientation (Fig. 16.3) relative to the fetus’s viewpoint, which was assessed simultaneously with ultrasound imaging (Reid et al., 2017). As seen in Fig. 16.3, fetuses tended to move their heads toward the more facelike configuration, evidence that fetuses can control their behavior based on visual inputs.

Development of the visual system Chapter | 16

339

FIGURE 16.2 A model of fetal visual function in the superior colliculus (SC), a midbrain structure with connections to visual and somatic maps. If the spatial distribution of the neurons in the somatotopic map is preserved across layers, multimodal neurons in the SC may respond most strongly to visual patterns with a spatial configuration like eyes and mouth. This prediction was subsequently confirmed (Fig. 16.3). Reproduced from Pitti, A., Kuniyoshi, Y., Quoy, M., Gaussier, P., 2013. Modeling the minimal newborn’s intersubjective mind: The visuotopic-somatotopic alignment hypothesis in the superior colliculus. PLoS One 8, e69474.

FIGURE 16.3 Left: Fetuses were presented with three-dot configurations that were either upright or inverted schematic faces. Right: Fetuses tended to turn their heads more frequently to the dots when they formed an upright schematic face. Reproduced from Reid, V.M., Dunn, K., Young, R.J., Amu, J., Donovan, T., Reeissland, N., 2017. The human fetus preferentially engages with facelike visual stimuli. Curr. Biol. 27, 1825e1828.

16.3 Visual perception in the newborn Human infants are born with a functional visual system. The eye of the newborn is sensitive to light, and if motivated (i.e., awake and alert) the baby may react to visual stimulation with head and eye movements. Vision is relatively poor, however: acuity (detection of fine detail), contrast sensitivity (detection of different shades of luminance), color sensitivity, and sensitivity to direction of motion all undergo improvements after birth (Banks and Salapatek, 1983). The field of view is also relatively small, so that newborns often fail to detect targets too far distant or too far in the periphery. In addition, as far as we know neonates lack stereopsis, which is the perception of depth from binocular disparity (differences in the input to the two eyes). Maturation of the eye and cortical structures (see previous section) support developments in these visual functions, and learning plays an important role as well, discussed in greater detail in Section 16.5.

340 PART | II Cognitive development

16.3.1 Visual organization at birth Testing newborn infants is not for the faint of heart. Success is entirely dependent on the baby’s mood; this is at its most capricious early in postnatal life, and there is no predicting neonate behavior. Having said this, a number of patient scientists have conducted careful experiments with neonates; these experiments have revealed that despite relatively poor vision, neonates actively scan the visual environment. Early studies, summarized by Haith (1980), revealed systematic oculomotor behaviors that provided clear evidence of visual organization at birth. Newborns, for example, will search for patterned visual stimulation, tending to scan broadly until encountering an edge, at which point scanning narrows so that the edge can be explored. Such behaviors are clearly adaptive for investigating and learning about the visual world. In addition, newborn infants show consistent visual preferences. Fantz (1961) presented newborns with pairs of pictures and other two-dimensional patterns and recorded the member of the pair which attracted the infant’s visual attention, which he scored as proportion of fixation times per exposure. Infants typically looked longer at one member of the pair: bull’s eye versus stripes, or checkerboards versus solid forms, for example. Visual preferences have served as a method of choice ever since, in older infants as well as neonates. Slater (1995) described a number of newborns’ preferences: patterned versus unpatterned stimuli, curvature versus rectilinear patterns, moving versus static patterns, three-dimensional versus two-dimensional forms, and high versus low-contrast patterns, among others. In addition, perhaps due to the relatively poor visual acuity of the newborn visual system there is a preference for “global” form versus “local” detail in newborns (Macchi Cassia et al., 2002).

16.3.2 Visual behaviors at birth Fantz (1964) reported that repeated exposure to a single stimulus led to a decline of visual attention, and increased attention to a new stimulus, in 2- to 6-month-olds. A substantial number of subsequent investigations examined infants’ preferences for familiar and novel stimuli as a function of increasing exposure, and these in turn led to standardized methods for testing infant perception and cognition, such as habituation paradigms (Cohen, 1976), as well as a deeper understanding of infants’ information processing (Aslin, 2007; Hunter and Ames, 1989; Sirois and Mareschal, 2002). Neonates (and older infants) will habituate to repeated presentations of a single stimulus; habituation is operationalized as a decrement of visual attention across multiple exposures according to a predetermined criterion. Following habituation, infants generally show preferences for novel versus familiar stimuli, implying both discrimination of novel and familiar stimuli and memory for the stimulus shown during habituation. Neonates and older infants also recognize visual constancies or invariants, the identification of common features of a stimulus across some transformation, for instance shape, size, slant, and form (Slater et al., 1983). Recognition of invariants forms the basis for categorization.

16.3.3 Faces and objects Newborns prefer faces and facelike forms relative to other visual stimuli, and are thus well prepared to begin engaging in social interactions with conspecifics. Some have speculated that there is an innate representation for facial structure (Morton and Johnson, 1991); others have suggested that the preference stems from general-purpose visual biases that guide attention toward stimuli of a particular spatial frequency, with a prevalence of stimulus elements in the top portion, as seen in Fig. 16.4 (Turati et al., 2002; Valenza et al., 1996). As noted previously, fetuses, like newborns, also appear to orient preferentially to facelike patterns (Reid et al., 2017). Newborns’ object perception is not so precocious. Neonates perceive segregation of figure and ground (i.e., seeing objects as distinct from backgrounds), but there are limits in the ability to perceive object occlusion, as seen in Fig. 16.5A. Adults and older infants perceive this display as consisting of two objects, one moving back and forth behind the other (Kellman and Spelke, 1983). Neonates, however, seem to perceive this display as consisting of three disconnected parts (Slater et al., 1990). In these experiments, infants were habituated with the partly occluded rod display, followed by two test displays. One test display (Fig. 16.5B) consisted of the whole rod (no occluder), and the other consisted of two rod parts, separated by a gap in the space where the occluder was seen, corresponding to the visible rod portions in the habituation stimulus (Fig. 16.5C). For 4-month-olds, looking longer at the broken rod is taken as evidence that they perceived unity of the rod parts as unified behind the box, but for neonates looking longer at the complete rod implies perception of disjoint surfaces in similar displays. The developmental processes underlying this shift in perceptual abilities are discussed in Section 16.5.1.

16.4 Postnatal visual development As noted in Section 16.2, visual development begins prenatally; in this section I describe some of the ways it continues after birth. Both infants and adults scan visual scenes activelydon the order of two to four eye movements per second in

Development of the visual system Chapter | 16

STIMULI

Total fixation time

Number of discrete looks

53.86 s vs 37.62 s

10 vs 8.09

p < 0.03

p < 0.05

34.70 s vs 41.08 s

7.6 vs 8.3

p > 0.20

p > 0.30

44.15 s vs 22.89 s

10.43 vs 6.5

p < 0.003

p < 0.01

341

FIGURE 16.4 Facelike stimuli from experiments on neonates’ preferences. Reproduced from Turati, C., Simion, F., Milani, I., Umiltà, C., 2002. Newborns’ preference for faces: What is crucial? Dev. Psychol. 38, 875e882.

FIGURE 16.5 Rod-and-box displays from experiments on infants’ perception of partly occluded objects. (A) habituation stimulus. (B) and (C) test stimuli.

general (Johnson et al., 2004; Melcher and Kowler, 2001; van Renswoude et al., 2019)dbut visual function is relatively poor at birth in terms of processing and analyzing visual information. Functional visual development has been explained in terms of maturation of visual pathways in the brain and peripheral systems, such as the eyeball (Atkinson, 2000; Johnson, 1990, 2005). Acuity, for example, improves in infancy with a number of developments, all taking place in parallel: migration of receptor cells in the retina toward the center of the eye, elongation of the receptors to catch more incoming light, growth of the eyeball to augment the resolving power of the lens, myelination of the optic nerve and cortical neurons, and synaptogenesis and pruning.

342 PART | II Cognitive development

16.4.1 Visual physiology The visual system, like the rest of the brain, is organized modularly and hierarchically. Incoming light is transduced into neural signals by the retina, which passes information to the lateral geniculate nucleus (LGN, part of the thalamus), and then to primary visual area (V1) in cortex and higher visual areas. Successively higher visual areas code for visual attributes in larger portions of visual field and participate in more complex visual functions (see Fig. 16.6). For example, visual pathways extending from V5 (also known as area MT) through parietal cortex are largely responsible for coding motion. Infants younger than 2 months appear unable to discriminate different directions of motion until maturation of pathways extending to and originating in V5 (Johnson, 1990). For motion processing, therefore, development centers on a limited number of visual areas and a relatively small number of mechanisms (e.g., myelination, synaptic growth, and pruning). Object perception, in contrast, is far more complex, involving many areas, each of which is responsible for processing one or more of the many visual attributes that defines edges, surfaces, and objects.

16.4.2 Critical periods A critical period refers to a time in an individual’s ontogeny when some function or ability must be stimulated or it will be lost permanently (see Daw, 1995). This notion can be contrasted with a sensitive period, similar in concept but generally referring to scenarios in which effects of deprivation are not so severe. The formal study of critical periods was initiated by Wiesel and Hubel (1963), who covered or sutured one eye in kittens from birth for a period of 1e4 months and examined effects of visual deprivation by patching the unaffected eye and observing visual function of the affected eye alone. The deprived eye was effectively blind, as revealed by both behavioral and neural effects. Behavioral effects included an inability to navigate visually or respond to objects introduced by the experimenters, though the animals behaved normally under these circumstances when permitted to use the unaffected eye. Neural effects were examined by recording from single cells in visual cortex; in general, few cortical cells could be driven by the deprived eye in cortical regions normally responsive to input from both eyes, such as postlateral gyrus. Wiesel and Hubel also reported effects of eye closure in animals that were allowed some visual experience prior to deprivation, highlighting the distinction between critical and sensitive periods. The unaffected eye dominated activity of cells in visual cortex, but depended on both the extent of visual experience prior to deprivation and the duration of deprivation. Stereopsis, the detection of distance differences in near space (e.g., threading a needle), seems to emerge during a critical period. Stereopsis relies on slight differences in the inputs to the two eyes when they are directed to the same point,

PS

VIP SOMATO SENSORY

V2 V3

V3A

MOTOR

8

POa 7a

V4 V1 V

I S U A

VP

L

P M M S O AUDITORY T T L Y S E N S IT O R Y

PS

FIGURE 16.6 Cortical areas in the macaque monkey showing an outer view of the left hemisphere (upper left) and a flattened representation of sensory and motor regions. Visual areas are depicted in red, orange, and yellow. Reproduced from Van Essen, D.C., Maunsell, J.H.R., 1983. Hierarchical organization and the functional streams in the visual cortex. Trends Neurosci. 6, 370e375.

Development of the visual system Chapter | 16

343

also known as disparity. Cells in primary visual cortex are organized into “ocular dominance” columns that receive inputs from the two eyes and register the amount of disparity between them. These require binocular function early in lifedthe two eyes must be directed consistently at the same points and focus on them. This can be disrupted by amblyopia (poor vision in one eye) or strabismus (misalignment of the eyes). Normally, mature visual cortex contains cells responsive to both eyes, and a few to only one eye. Abnormal visual experience can produce a preponderance of cells responsive to only one or the other eye, but not to both. In typically developing infants, stereopsis emerges at about 4 months, as inputs from the two eyes into the ocular dominance columns become segregated (Held, 1985). (Prior to this time the inputs are more likely to be superimposed, which may result in frequent diplopia, or double vision, early in life.) The critical period in humans for development of stereopsis in humans is estimated to be 1e3 years (Banks et al., 1975).

16.4.3 Development of visual attention Visual attentiondeye movementsdis a combination of saccades and fixations. During a saccade, the point of gaze sweeps rapidly across the scene, and during a fixation, the point of gaze is stationary. Analysis of the scene is performed during fixations. Eye movements can also be smooth rather than saccadic, as when the head translates or rotates as the point of gaze remains stabilized on a single point in space (the eyes move to compensate for head movement), or when following a moving target. Visual attention in infancy has attracted a great deal of interest, because it is a behavior that is relatively mature, even at birth, and because it is relatively easy to observe (Johnson, 2005; Richards, 1998). Oculomotor behaviors that have been examined include detection of targets in the periphery, saccade planning, oculomotor anticipations, sustained versus transient attention, effects of spatial cuing, and eye/head movement integration; other tasks have examined inhibition of eye movements, such as disengagement of attention, inhibition of return, and spatial negative priming. Bronson (1990, 1994) examined scanning patterns as infants viewed simple geometric forms, and reported changes with development in attention to distributed visual features, including a greater tendency to scan between features, to direct saccades with greater accuracy, and in general to engage in more “volitional” scanning, starting at 2e3 months. There are important developments also in viewing complex scenes. In my lab, we recorded eye movements of infants and adults as they watched segments of an animated cartoon (A Charlie Brown Christmas) that was rich in social content (Frank et al., 2009). Three-month-olds’ attention was captured most by low-level image salience (variations in color, luminance, and motion) and by 9 months there was a stronger focusing of attention on faces. There were no reliable differences between age groups in measures such as mean saccade distance and fixation duration. One interpretation of these results is a developmental transition toward attentional capture by semantic contentdthe “meaning” inherent in social stimuli. In addition, developing control over visual attention facilitates infants’ ability to fixate stimuli of interest (Frank et al., 2014). Studies of real-world scene perception showed a similar trend: younger infants (3e4 months) tend to fixate the most visually salient regions of photographs, followed by a progression toward looking more at faces (Amso et al., 2014) and at objects in the scenes that were most often fixated by adults, presumably from semantic content (van Renswoude et al., 2019).

16.4.4 Cortical maturation and oculomotor development Gaze control in mature primates is accomplished with a coordinated system of both subcortical and cortical brain areas, as seen in Fig. 16.7. Control of eye movements originates in areas with outputs that are connected to the brainstem (including SC, discussed previously), which sends signals to the oculomotor musculature. Development of visual attention has often been interpreted as revealing development of cortical systems that control it. Visual attention has been suggested to be largely under subcortical control until the first few months after birth, after which there is increasing cortical control (Atkinson, 1984; Colombo, 2001; Johnson, 1990). For example, oculomotor smooth pursuit and perception of motion direction have been proposed to rely on a common cortical region, area V5, and the development of these visual functions in infancy has been tied to maturation of V5, as noted previously (Johnson, 1990). Smooth pursuit is maintenance of gaze on a moving target with nonsaccadic eye movements; motion direction perception is often tested with random dot displays, to control for the possibility that motion following is not simply a detection of change in position. Perceiving motion, and performing the computations involved in programming eye movements to follow motion, is thought to be founded on the same cortical structures (Thier and Ilg, 2005). This suggestion was tested empirically by Johnson et al. (2008), who observed infants between 58 and 97 days of age in both a smooth pursuit (Fig. 16.8, top panel) and a motion direction discrimination task (Fig. 16.8, center panel). Individual differences in performance on the two tasks were strongly correlated, and were also positively correlated with

344 PART | II Cognitive development

Anterior System Posterior System

MEF FEF

parietal lobe

occipital lobe superior colliculus brainstem

FIGURE 16.7 Subcortical and cortical structures involved in oculomotor control. FEF, frontal eye fields, MEF, medial eye fields. Reproduced from Schiller, P.H, Tehovnik, E.J., 2001. Look and see: How the brain moves your eyes about. Prog. Brain Res. 134, 127e142.

FIGURE 16.8 Top: Schematic depiction of stimuli used to examine smooth pursuit in young infants. A toy moved laterally at one of five speeds in one of five vertical positions on the screen. Only one toy was shown at a time. Center: Random-dot kinematograms used to examine motion direction discrimination in young infants. Dotted lines and dots, shown here to demarcate regions of motion, were not present in the stimulus. Bottom: Individual infant’s performance in smooth pursuit and direction discrimination tasks were correlated with age. Adapted from Johnson, S.P., Davidow, J., Hall-Haro, C., Frank, M.C., 2008. Development of perceptual completion originates in information acquisition. Dev. Psychol. 44, 1214e1224.

Development of the visual system Chapter | 16

345

age (Fig. 16.8, bottom panel), consistent with the maturation theory. Other visual functions in infancy that have been linked to cortical maturation include development of form and motion perception, stemming from maturation of parvocellular and magnocellular processing streams, respectively (Atkinson, 2000), and development of visual memory for object features and object locations, stemming from maturation of ventral and dorsal processing streams (Mareschal and Johnson, 2003).

16.4.5 Development of visual memory Memory for events, object features, and locations improves over the first several postnatal months (Rose et al., 2004). As noted previously, newborns will habituate to repeated presentations of a visual stimulus and recover interest to a novel one, clear evidence for a functional short-term visual memory store available at birth. Visual short-term memory in older infants has been examined with a “change-detection” task in which infants viewed a pair of displays side by side, each of which contained one or more shapes. On one side the object or objects underwent color changes every 250 ms (Ross-Sheehy et al., 2003). When there was one object per side, 4-month-olds looked longer toward the side with color changes, implying a short-term store of the color information across the 250 ms temporal gap. Visual short-term memory develops rapidly: Ten-month-olds retained color information across a set size of four different colors (Ross-Sheehy et al., 2003), and 7.5-month-olds retained information about color-location combinations (set size of three) across a 300 ms delay (Oakes et al., 2006). Studies of infant memory employing operant conditioning paradigms, in which infants are trained to kick their legs to move a mobile, have demonstrated long-term visual recognition stores that are available from at least 2 months under some conditions; the memories formed can last for several days or even weeks given sufficient training with the mobile and reminders (Rovee-Collier, 1999). Infants at 6 months can imitate observed behaviors after 24 h, and the retention interval is considerably longer in older infants (Barr et al., 1996). Developments in visual memory, like many other visual functions, have been proposed to stem from cortical development, in particular areas of the medial temporal lobe such as hippocampus, perirhinal and entorhinal cortices, and amygdala (Bauer, 2004; Nelson, 1995; Rose et al., 2004).

16.4.6 Development of visual stability Our gaze moves frequently from point to point in the visual scene, and our bodies move from place to place. Despite these continual disruptions and interruptions in visual input, we experience the visual world as an inherently stable place. Consider, for example, the difference in your visual experience when you read this text while shaking your head back and forth (as if you wanted to signify “no” to someone). Reading is not much compromised. Now, if possible, shake the text back and forth while holding your head steady. You will discover reading to be more difficult, yet the spatial relation between your head and the text in the two situations is similar. When you rotate your head, compensatory eye movements known as the vestibulo-ocular response, or VOR, allow the point of gaze to remain fixed or to continue moving volitionally as desired (as when reading). When the text moves, there is no such compensatory mechanism. Evidence from three paradigms suggests that visual stability emerges gradually across the first year after birth. First, young infants have difficulty discriminating optic flow patterns that simulate different directions of self-motion (Gilmore et al., 2004). Infants viewed a pair of random-dot displays in which the dots repeatedly expanded and contracted around a central point to simulate the effect of moving forward and backward under real-world conditions. On one side, the location of this point shifted periodically, which for adults specifies a change in heading direction; the location on the other side remained stationary. Under these circumstances adults detected a shift simulating a 5 degrees change in heading, but infants were insensitive to all shifts below 22 degrees, and sensitivity was unchanged between 3 and 6 months. Gilmore et al. speculated that optic flow sensitivity may be improved by self-produced locomotion after 6 months of age, or by maturation of the ventral visual stream. Second, young infants’ saccade patterns tend to be retinocentric, rather than body-centered, in a “double-step” tracking paradigm (Gilmore and Johnson, 1997). Retinocentric saccades are programmed without taking into account previous eye movements. Body-centered eye movements, in contrast, are programmed while updating the spatial frame of reference or coordinate system in which the behaviors occur. Infants first viewed a fixation point that then disappeared, followed in succession by the appearance and extinguishing of two targets on either side of the display. The fixation point was located at the top center of the display, and targets were located below it at the extreme left and right sides. As the infant viewed the fixation point and targets in sequence, there was an age-related transition in saccade patterns. Three-month-olds tended to direct their gaze downward from the first target, as if directed toward a target below the current point of gaze. In reality the second target was below the first locationdthe original fixation pointdnot the current point of gaze. Seven-month-old infants, in contrast, were more likely to direct gaze directly toward the second target. These findings imply that the young infants’ visual-spatial coordinate system, necessary to support perception of a stable visual world, may be insensitive to extraretinal information, such as eye and head position, in planning eye movements.

346 PART | II Cognitive development

Third, there are limits in the ability of infants younger than 2 months to switch attention flexibly and volitionally to consistently maintain a stable gaze. Movement of one’s body through the visual environment can produce an optic flow pattern, as can head movement while stationary (recall the head-shaking example). The two scenarios may produce similar visual inputs from optic flow, yet we readily distinguish between them. In addition, adult observers can generally direct attention to either moving or stationary targets, nearby or in the background, as desired. These are key features of visual stability, and four eye movement systems work in concert to produce it. Optokinetic nystagmus, or OKN, stabilizes the visual field on the retina as the observer moves through the environment. OKN is triggered by a large moving field, as when gazing out the window of a train: The eyes catch a feature, follow it with a smooth movement, and saccade in the opposite direction to catch another feature, repeating the cycle. The VOR, described previously, helps maintain a stable gaze to compensate for head movement. (OKN and the VOR are present and functional at birth, largely reflexive or obligatory, and are likely mediated by subcortical pathways; Atkinson and Braddick, 1981; Preston and Finocchio, 1983). The others are the saccadic eye movement system and smooth pursuit, to compensate for or cancel the VOR or OKN as appropriate. Aslin and Johnson (1994) observed suppression (cancellation) of the VOR to fixate a small moving target in 2and 4-month-olds, but not 1-month-olds, and Aslin and Johnson (1996) observed suppression of OKN to fixate a stationary target in 2-month-olds, but not in a younger group.

16.4.7 Object perception As noted previously in Section 16.3.3, “piecemeal” or fragmented perception of the visual environment appears to extend from birth through the first several months afterward, implying a fundamental shift in the infant’s perceptual experience. Because neonates and 4-month-olds appear to construe dynamic rod-and-box displays differentlydas disjoint surfaces and as occluded objects, respectivelydan important step in understanding development of perceptual completion is investigations of performance in 2-month-olds. In an initial investigation, 2-month-olds were found to show an “intermediate” pattern of performance (no reliable posthabituation preference), consistent with the possibility that spatial completion is developing at this point but not yet in final form (Johnson and Náñez, 1995). A followup study examined the hypothesis that 2-month-olds may perceive unity if given additional perceptual support. We simply increased the amount of visible rod surface revealed behind the occluder by reducing box height and by adding gaps in it, and under these conditions 2-month-olds provided evidence of unity perception (Johnson and Aslin, 1995). Adopting this approach with newborns, however, failed to reveal similar evidence: Even in “enhanced” displays, newborns seemed to perceive disjoint rather than unified rod parts (Slater et al., 1996; Slater et al., 1994). However, when newborns were tested with rod-and-box displays in which the rod parts appeared to “jump” from one location to the nextdthat is, apparent motion rather than smooth motiondthe infants appeared to perceive the rod parts as unified (Valenza and Bulf, 2011). These authors suggested that the smooth motion of rod parts (as tested by Slater et al., 1994, 1996) might be difficult to detect at birth, yet this explanation begs the question why newborns would construe rod parts undergoing smooth motion as disjoint surfaces. This question awaits further study. A number of studies have shown that young infants can maintain representations of the solidity and location of fully hidden objects across brief delays (e.g., Aguiar and Baillargeon, 1999; Spelke et al., 1992). Yet newborns provide little evidence of perceiving partly occluded objects, leading to the question of how perception of complete occlusion, or existence constancy, emerges during the first few months after birth. To address this question, experiments have examined infants’ responses to objects that move forward on a trajectory, disappear behind an occluder, reappear on the far side, and reverse direction, repeating the cycle (Fig. 16.9A). Following habituation to this display, infants viewed test displays consisting of continuous and discontinuous trajectories (Fig. 16.9B,C), analogous to the broken and complete test stimuli described previously. Four-month-olds appeared to treat the ball-and-box display depicted in Fig. 16.9A as consisting of

FIGURE 16.9 Ball-and-box displays from experiments on infants’ perception of existence constancy. (A) habituation stimulus. (B) and (C) test stimuli.

Development of the visual system Chapter | 16

347

two disconnected trajectories, rather than a single, partly hidden path (Johnson et al., 2003a,b), but by 6 months, infants perceived this trajectory as unitary. When occluder size was reduced, however, 4-month-olds’ posthabituation preferences (and thus, by inference, their percepts of spatiotemporal completion) were shifted, partway by an intermediate width, and fully by a narrow width, so narrow as to be only slightly larger than the ball itself. Reducing the spatial gap, therefore, supported perception of a complete trajectory in 4-month-olds. In 2-month-olds this manipulation appeared to have no effect, implying a lower age limit for trajectory completion (i.e., existence constancy) in infants, just as there may be for spatial completion. To account for these and related findings, Bremner et al. (2015) proposed a model in which younger infants’ perception of object persistence is subject to greater perceptual constraints compared with infants a few months older, and in which young infants require a combination of cues to perceive object persistence across occlusion.

16.4.8 Face perception As noted in Section 16.3.3, infants are better prepared to perceive faces than objects at birth, showing preferences for faces and facelike structures. Research on face perception in infants provides additional insights on mechanisms of recognition. In adults, face recognition is near ceiling when faces are upright, but when faces were inverted, performance is relatively poordthe inversion effect (Yin, 1969). This appears to be specific to faces; other visual configurations normally seen upright, such as houses, and are not vulnerable to the effect. These findings are thought to reflect a difference in implicit processing “strategies” when viewing upright versus inverted faces. When faces are upright, they are processed in terms of both the individual features and the spatial relations among features (viz., both piecemeal and holistic processing), but when inverted, these relations are more difficult to access, forcing greater reliance on only a single source of information for recognitiondthe featuresdand thus impairing performance. Carey and Diamond (1977) reported that children younger than 10 years of age do not show the inversion effect. This led to the suggestion that young children process faces according to features only, and that piecemeal to holistic processing develops during childhood, perhaps from experience viewing faces or maturation of the right cerebral hemisphere, implicated in complex visual-spatial tasks. Consistent with these findings, children’s discrimination of faces was impaired more by a mismatch in the spacing of features than by a mismatch in the features themselves (eyes, nose, and mouth) or faces’ outer contours, as seen in Fig. 16.10 (Mondloch et al., 2002), and there were dramatic improvements in performance from 6 years through adulthood to match identity of faces across changes of facial expression, orientation, and “lip reading” (mouthing different vowels), all of which require sensitivity to spatial relations among features (Mondloch et al., 2003). Other reports, however, provide evidence for a much earlier piecemeal-to-holistic shift in processing faces. First, Younger (1992) found that 10-month-old infants were sensitive to correlations among facial attributes in a face discrimination task; 7-month-olds provided evidence of discrimination from featural variations only. Second, evidence from a “switch” paradigm showed that 7-month-olds processed configurations of facial features that were disrupted by inversion (Cohen and Cashon, 2001). In the switch design, infants are habituated to a pair of distinct stimuli (in this case, faces); at test, selected features are switched from one to the other and infants are observed for recovery of interest to the new configuration. Another study using this design found a developmental progression toward processing configurations between 4 and 10 months (Schwarzer et al., 2007). Third, the inversion effect was found in face recognition tasks with 5-, 7-, and 9-month-olds, but when outer contours and inner facial features were inverted in separate experiments, only the older two age groups showed impairment from inversion, suggesting a greater flexibility in their processingdutilizing either internal or external features to recognize the faces (Rose et al., 2008).

16.4.9 Critical period for development of holistic perception Evidence for a critical period for holistic face perception comes from a study of individuals born with cataracts who underwent surgery to correct the problem (Le Grand et al., 2001). Each individual had at least 9 years of visual experience after surgery. The individuals were tested with face recognition tasks as described in the previous section, including tests of inversion effects, using some of the stimuli shown in Fig. 16.10. There was a specific deficit in recognition from configurational informationdthe spacing of featuresdbut not from featural information, where performance was not reliably different than controls. A particularly striking characteristic of these findings concerns the timing of cataract replacement, which for every patient was less than 7 months of agedand in a few cases, as little as 2e3 months. The critical period for development of holistic processing, therefore, appears to be exceedingly brief. Interestingly, infants at 2e3 months show no signs of the inversion effect (Cashon and Cohen, 2003) and sensitivity to some kinds of holistic information in faces is not adultlike until several years after this time, as noted previously.

348 PART | II Cognitive development

(A)

(B)

(C)

FIGURE 16.10 Stimuli used to test recognition of faces in which the spacing of features is varied (top row), the features (but not their spacing) are varied (center row), or the outer contours (but not features or spacing) are varied (bottom row). The faces in the leftmost positions of each row are identical; other faces in each row are variations of it. Reproduced from Mondloch, C.J., Le Grand, R., Maurer, D., 2002. Configural face processing develops more slowly than featural face processing. Perception 31, 553e566.

Some kinds of holistic object perception appear to be comprised by visual deprivation, but the evidence is complex. On the one hand, patients treated for cataracts showed no deficits, relative to controls, in identifying pictures of houses on the basis of both featural and configurational information, in contrast to face recognition (Robbins et al., 2008). And a case study of SRD, a woman who had cataracts removed at age 12, revealed few obvious deficits in object perception when tested 22 years later on shape matching, visual memory, and image segmentation tasks (Ostrovsky et al., 2006). Her performance at face recognition was impaired relative to controls, as expected from the Le Grand et al. (2001) study, but she was not tested explicitly for holistic object perception. On the other hand, a case study of MM, a man who lost his vision at 3.5 years and had cataract replacement nearly 40 years later, revealed marked deficits in object perception skills (Fine et al., 2003). Five months after surgery, MM was unable to detect transparency in overlapping forms, to see depth from perspective in a Necker cube, or to identify a shape defined by illusory contours (a Kanizsa square)dthe latter a paradigmatic instance of holistic processing, the binding of visual features across a spatial gap. Everyday objects were mostly unrecognizable, and he experienced difficulty discriminating faces and identifying emotional expression, reporting to rely on individual features rather than a “Gestalt” for these purposes. Cortical areas that give strong responses in normally sighted observers when viewing faces and objects (lingual and fusiform gyri) were largely inactive in MM. (Other visual functions were well preserved, such as contrast sensitivity, color perception, and motion perception, implying that they may have been more established and consequently robust to deprivation by the time MM was blinded in childhood). A study of illusory contour perception in cataract-replacement patients provides additional evidence for severe compromise in feature binding from early visual deprivation (Putzar et al., 2007). Patients were divided into two groups, one with cataract replacement prior to 6 months and the second after this time, and their performance was compared to controls. The patient group treated after 6 months showed elevated reaction times and greater miss rates when searching for illusory shapes among distracters, relative to real shapes; the other groups showed reliably less of a difference on these measures. Interviews conducted after testing revealed that the post-6-month patient group did not perceive the illusory figures at all, but rather adopted a strategy of finding regions in the scenes where the inducing elements pointed inward. Consistent with experiments on face perception described previously, these results point to the first several months after birth as a critical period for spatial integration of visual information.

Development of the visual system Chapter | 16

349

Notably, however, early visual deprivation does not impair all higher-level visual functions. Recent studies have shown that deprivation has little effect on susceptibility to the Ponzo and Müller-Lyer illusions, which are visual illusions involving geometric relations (Gandhi et al., 2015). In addition, effects of deprivation on categorizing faces versus nonface foils become attenuated during the first several months after cataract surgerydthat is, patients’ face categorization skills approach those of control participants with typical vision (Gandhi et al., 2017).

16.5 How infants learn about objects 16.5.1 Learning from targeted visual exploration Infants in the transition toward spatial completion in rod-and-box displaysd2 to 3 months of agedhave been observed for evidence that scanning patterns are associated with unity perception. These links are clear. Amso and Johnson (2006) and Johnson et al. (2004) observed 3-month-old infants in a perceptual completion task using the habituation paradigm described previously (Fig. 16.5). Infants’ eye movements were recorded with a corneal reflection eye tracker during the habituation phase of the experiment. We found systematic differences in scanning patterns between infants whose posthabituation test display preferences indicated unity perception and infants who provided evidence of perception of disjoint surfaces: “Perceivers” tended to scan more in the vicinity of the two visible rod segments, and to scan back and forth between them. In a younger sample (58e97 days), Johnson et al. (2008) found a reliable correlation between posthabituation preference (viz., our index of spatial completion) and targeted visual exploration, operationalized as the proportion of saccadic eye movements directed toward the moving rod parts, obviously the most relevant aspect of the stimulus for perception of completion. Spatial completion was not predicted by other measures of oculomotor performance, including mean number of fixations per second, mean saccade distance (to assess overall scanning activity), mean vertical position of each infant’s fixations (to assess a bias for the upper portion of the stimulus), and mean dispersion of visual attention (to assess the scanning of limited portions of the stimulus vs. scanning more broadly). Nor was spatial completion associated with another measure of oculomotor control, smooth pursuit. Rather, spatial completion was best predicted by saccades directed toward the vicinity of the moving rod parts. This can be a challenge for a developing oculomotor system, attested by the fact that targeted scans almost always followed the rod as it moved, rarely anticipating its position. Targeted visual exploration develops with time, stems from increasing endogenous control of oculomotor behavior, and consists of both selection of desired visual targets and inhibition of everything else in the visual scene. Evidence for development of selection comes from studies of orienting, discussed previously in Section 16.4.3. Evidence for development of its complement, inhibition, is relatively scarce. Newborns exhibit inhibition of return of the point of gaze to recently visited locations (Valenza et al., 1994), but inhibition of eye movements to covertly attended locations develops more slowly across the first year (Amso and Johnson, 2005, 2008). How selection and inhibition work together to maximize effective uptake of visual information is not yet known, but the experiments on spatial completion and eye movements begin to provide important insights. Very young infants’ ability to perceive occlusion may be precluded by insufficient access to visual information for unity: alignment, common motion, and other Gestalt cues such as similarity and interposition. An alternate view stressing developmental mechanisms that are independent of learning and experience might posit that emergence of spatial completion stems exclusively from maturation of neural structures responsible for object perception, and, as infants begin to perceive occlusion, their eye movement patterns support or confirm this percept. Amso and Johnson (2006) found that both spatial completion and scanning patterns were strongly related to performance in an independent visual search task in which targets were selected among distracters. This finding is inconsistent with the possibility that scanning patterns were tailored specifically to perceptual completion, and instead suggests that a general facility with targeted visual behavior leads to improvements across multiple tasksdprecisely the pattern of performance we observed. Scanning patterns during infants’ natural scene viewing were also found to be modulated by developments in selection and inhibiton (van Renswoude et al., 2019). How might developing object perception systems benefit from targeted scans? Eye movements may serve as a vital binding mechanism due to the relatively restricted visual field and poor acuity characteristic of infant vision. Visual information in the periphery is more difficult to access with a single glance, increasing the need to scan between features to ascertain their relations to one another. The developmental timing of targeted visual exploration in infants seems just right for another reason: the critical period for development of holistic object processing. It may be that motor feedback from scanning eye movements serves as a trigger for consolidation of neural circuits in areas that represent the stimulus, enabling association of the separate parts of an object seen on sequential fixations (Rodman, 2003). As an observer views salient object features, the point of gaze falls in rapid succession on components that will later be perceived as part of a coherent whole. Motor feedback signaling a series of sequential fixations within the central visual field could thus be a

350 PART | II Cognitive development

powerful cue to bind features together, a possibility consistent with close relations in adults between scan paths and pattern recognition (Noton and Stark, 1971; Rizzo et al., 1987) and scene perception (Henderson, 2003).

16.5.2 Learning from associations between visible and occluded objects By 6 months, infants’ short-term representations of unseen objects are sufficiently robust to guide reaching and oculomotor systems prospectively to intercept objects on hidden trajectories (Clifton et al., 1991; Johnson et al., 2003a; von Hofsten et al., 1998). At 4 months, prospective behaviordanticipations from eye and head movements to the place of reappearance of an object seen to move behind an occluderdis adapted to variations in occluder width and object speed, implying that under some conditions, infants may track with their “mind’s eye” (von Hofsten et al., 2007). Yet under other circumstances, 4-month-olds process partly occluded trajectories in terms of visible components only, not complete paths (Fig. 16.9). Representations of occluded objects in 4-month-olds, therefore, appear to be rather fragile and not completely established. To examine the possibility that learning can facilitate spatiotemporal completion, my colleagues and I presented ball-and-box displays to 4- and 6-month-olds as we recorded their eye movements (Johnson et al., 2003a). We reasoned that a representation of the object and its trajectory under occlusion would be reflected in a consistent pattern of anticipatory eye movements toward the place of reemergence, before the object’s appearance. The stimulus was identical to the displays used by Johnson et al. (2003b) to investigate spatiotemporal completion (Fig. 16.9A). Because 6-month-olds provided evidence of spatiotemporal completion in these displays when tested with a habituation paradigm, we predicted that oculomotor anticipations would be more frequent in the older age group. This prediction was supported. A higher proportion of 6-month-olds’ object-directed eye movements was classified as anticipatory (i.e., initiated prior to the ball’s emergence from behind the occluder, Fig. 16.11, top panel) relative to 4-month-olds (Fig. 16.11, center panel), corroborating the likelihood that spatiotemporal completion strengthens between 4 and 6 months. Evidence for learning as an important contributor to this developmental change came from a new group of 4-montholds in a “training” condition. These infants were first presented with an unoccluded, fully visible ball trajectory (no occluder) for 2 min followed by the ball-and-box display as per other conditions, and their eye movements were recorded. Here, the proportion of anticipations was reliably greater than that observed in the “baseline” conditions with untrained 4-month-olds, but not reliably different than that of untrained 6-month-olds (Fig. 16.11, bottom panel). In other words, 2 min of exposure led to behaviors characteristic of infants who are 2 months older. This rapid learning may stem from the ability to form associations between fully visible to partly or fully hidden objects. In the real world, infants are exposed to many different objects moving in different ways, presenting multiple opportunities for learning. For associative learning about occlusion to be a viable means of dealing with real-world events, associations between visible and partly occluded paths must be committed to memory. How long does such rapidly acquired associations last? To address this question, we replicated the Johnson et al. (2003a) methods with new groups of 4-month-olds, and observed a nearly identical pattern of anticipatory behaviors in baseline and training conditions (Johnson and Shuwairi, 2009). A third group received a half hour break between training and test, and performance reverted to baseline, implying that memory for the association was lost during the delay. But a fourth group, provided with a single “reminder” trial after an identical delay, showed a recovery of oculomotor anticipations equivalent to the no-delay training condition. (A fifth group, provided only a single training trial, showed no benefit in the form of anticipatory looking.) These findings suggest that accumulated exposure to occlusion events may be an important means by which existence constancy arises in infancy.

16.5.3 Learning from visual-manual exploration Spatial and spatiotemporal completion involve occlusion of far objects by nearer ones. Solid objects also occlude parts of themselves, meaning we cannot see the opposite surfaces from our present vantage point. Perceiving objects as solid in three-dimensional space constitutes 3D object completion, and we recently asked whether young infants perceive objects in this way (Soska and Johnson, 2008). Four- and 6-month-olds were habituated to a wedge rotating through 15 degrees around the vertical axis such that the far sides were never revealed (Fig. 16.10). Following habituation infants viewed two test displays in alternation, one an incomplete, hollow version of the wedge, and the other a complete, whole version, both undergoing a full 360 degree rotation revealing the entirety of the object shape. Four-month-olds showed no consistent posthabituation preference, but 6-month-olds looked longer at the hollow stimulus, indicating perception of the wedge during habituation as a solid, volumetric object in 3D space. When tested with a more complex object (an L shape), however, 3D object completion was observed in 6-month-old boys, but not girls (Soska and Johnson, 2013). This is consistent with the possibility of a male advantage in imagining how an object looks from another viewpoint, a process known as mental rotation (Shepard and Metzler, 1971). Mental rotation in infants is discussed in more detail subsequently.

Development of the visual system Chapter | 16

351

FIGURE 16.11 Histograms showing oculomotor anticipations (gray bars) versus reactions (black bars) as infants view ball-and-box displays. Each eye movement is coded for latency with respect to the emergence of the ball from behind the box, time 0. Eye movements initiated prior to this time are anticipations, and eye movements initiated after this time are reactions. Top panel: 4-month-olds. Center panel: 6-month-olds. Bottom panel: 4-month-olds after “training” with a fully visible trajectory. Adapted from Johnson, S.P., Amso, D., Slemmer, J.A., 2003. Development of object concepts in infancy: Evidence for early learning in an eye tracking paradigm. Proc. Natl. Acad. Sci. U.S.A. 100, 10568e10573.

How does 3D object completion arise? One possibility is that developmental changes in infants’ motor skills might underlie the ability to perceive the unseen parts of objects. Two types of motor skills, self-sitting and coordinated visualmanual object exploration, seem particularly important, because independent sitting frees the hands for play and promotes gaze stabilization during manual actions (Rochat and Goubet, 1995). Thus, self-sitting might spur improvements in coordinating object manipulation (e.g., rotating and transferring hand-to-hand) with visual inspection, providing infants

352 PART | II Cognitive development

FIGURE 16.12 Rotating object displays from experiments on infants’ perception of three-dimensional object completion. Top panel: habituation stimulus. Center and bottom panels: test stimuli. Adapted from Soska, K.C., Johnson, S.P., 2008. Development of 3D object completion in infancy. Child Dev. 79, 1230e1236.

with multiple views of objects. We tested these hypotheses in a group of 4.5- to 7.5-month-olds by replicating the Soska and Johnson (2008) methods and evaluating the infants’ motor skills (self-sitting and manipulation of different objects) on the same day (Soska et al., 2010). We found strong and significant relations between both self-sitting and visual-manual coordination (from the motor skills assessment) and our measure of 3D object completion (from the habituation paradigm). (Other motor skills we recorded, such as holding skill and manual exploration without visual attention to the objects, did not predict 3D object completion (Fig. 16.12). These results provide evidence for a cascade of developmental events following the advent of visual-motor coordination, including learning from self-produced experiences. Evidence from spatial completion experiments reveals that newborns perceive surface segregation even under conditions in which older infants and adults see the identical surfaces as unified (Slater et al., 1990), yet under other circumstances, say when stationary surfaces are directly adjacent, their connectivity or segregation may be ambiguous (Needham, 1997). This was demonstrated by Needham and Baillargeon (1998) for 4.5-month-olds’ interpretation of stimulus displays containing two dissimilar but adjacent, stationary objects (Fig. 16.11). After viewing these objects during a familiarization trial, infants were presented with test events in which a hand pulled the cylinder; the box either remained stationary or moved with the cylinder. The authors reasoned that infants would look longer at the event that was unexpected (e.g., the “move-apart” event if the objects were perceived as connected), a result found with 8-month-olds (Needham and Baillargeon, 1997), but the 4.5-month-old infants looked about equally at the two test events, providing no evidence for either interpretation on the infants’ part. Needham and Baillargeon (1998) asked whether 4.5-month-olds would learn from a brief prior exposure to either object in isolation and subsequently perceive the two as segregated. Their hypothesis was confirmed: either a 5-s exposure to the box or a 15-s exposure to the cylinder alone supported segregation of the adjacent cylinder-and-box display into two separate units when infants were tested immediately afterward. Some effects of such training last as long as 72 h (Dueker et al., 2003). This learning effect has been extended in a number of important ways. For example, the effect generalizes from exposure to objects in different orientations (Needham, 2001) (Fig. 16.13), but not to objects with distinct features, unless infants are introduced to the different objects in a variety of settings or contexts prior to testing, prompting formation of a perceptual category for the objects (Dueker and Needham, 2005). Categorization is facilitated as well by increasing the number or variety of exemplars during the learning phase of the experiment (Needham et al., 2005).

16.5.4 Hormonal and environmental influences on object perception As noted previously, mental rotation is the ability to imagine how an object that is seen from one perspective would look if it were rotated in space and viewed from a different perspective. There is a relatively strong sex difference in performance

Development of the visual system Chapter | 16

353

Move-apart Event

Move-together Event

FIGURE 16.13 Schematic depictions of stimuli used to assess object segregation in infants. Reproduced from Needham, A., 2001. Object recognition and object segregation in 4.5-month-old infants. J. Exp. Child Psychol. 78, 3e24.

on mental rotation tasksdon average, male participants outperform female participantsdin adults (Schöning et al., 2007; Voyer et al., 1995) and children (Linn and Petersen, 1985). Studies of mental rotation in infants have revealed a possible male advantage as young as 3e5 months (Moore and Johnson, 2008, 2011; Quinn and Liben, 2008, 2014), raising questions about the developmental origins of sex differences in perception of complex objects. Exposure to the hormone testosterone early in life, including prenatal exposure, is linked to a substantial number of sex differences, including features such as height, sexual orientation, and gender identity (Hines, 2010); testosterone exposure also influences mental rotation performance in adults (Pintzka et al., 2016) and children (Grimshaw et al., 1995), as well as children’s genderrelated playmate and toy preferences (Constantinescu and Hines, 2012). To test for a possible role for testosterone exposure and mental rotation in infants, my colleagues and I used a visual habituation paradigm involving complex objects rotating in 3D space (Constantinescu et al., 2018). As seen in Fig. 16.14A, the object viewed during habituation was a chiral arrangement of cubesdthat is, the object’s configuration was nonsuperimposable on its mirror image (akin to human hands). The object rotated partially around the vertical axis. Posthabituation test displays depicted the same complex object, but now seen from a different perspective (Fig. 16.14B), alternating with a mirror-image object (Fig. 16.14C). We reasoned that recognizing the same object (i.e., habituation and test objects) across different viewpoints would yield a test display preference for the mirror-image display, and we observed this effect in 5-month-old boys, but not in girls (replicating the findings of Moore and Johnson, 2008). Saliva samples were obtained to measure testosterone when infants were 1e2.5 months of age, a time known as “mini-puberty” in infancy due to a surge in testosterone, particularly in males (Lamminmäki et al., 2012). As seen in Fig. 16.14D, there was a statistically reliable correlation between testosterone measured at 1e2.5 months and the visual novelty preference measured at 5 months, but only in boys (data for girls are not shown). Importantly, we also found that for girls, mental rotation performance at 5 months (i.e., the novelty preference) was correlated with parental attitudes concerning gender roles assessed with the Child Gender Socialization Scale (Blakemore and Hill, 2008), specifically disapproval of genderatypical characteristics. Thus both hormones and the social environment modulated infants’ mental rotation performance, but in different ways for girls and boys.

16.6 Summary and conclusions From its prenatal origins to its postnatal refinement, learning to see is a mixture of developmental mechanisms, some of which operate outside of experience, and some of which are dependent on it. Although newborn infants can see fairly well upon their first exposure to patterned visual stimulation, as best we can tell the initial inputs are not bound into a stable, predictable, coherent visual world. The visual world as we adults know it emerges across the first year after birth. Developmental mechanisms are not limited only to cortical maturation or experience or learning, but instead comprise all of these and their interactions. The theoretical views most relevant to questions of visual cognitive development described previouslydthe views of Piaget and the Gestalt theoristsdforecasted some of the research I have described. As Piaget proposed, an infant’s

354 PART | II Cognitive development

FIGURE 16.14 Stimuli and results from a study testing mental rotation in 5-month-old infants. A: habituation stimulus, rotating back and forth through 240 degrees around the vertical axis. B and C: identical or mirror image stimuli, seen from a different perspective. D: significant correlation between testosterone at 1e2.5 months and mental rotation performance at 5 months in boys (the correlation for girls was not statistically significant). E: significant correlation between parental disapproval of atypical gender norms and mental rotation performance at 5 months in girls (the correlation for boys was not statistically significant). Reproduced from Constantinescu, M., Moore, D.S., Johnson, S.P., Hines, M., 2018. Early contributions to infants’ mental rotation abilities. Dev. Sci. 21, e12613.

Development of the visual system Chapter | 16

355

experience of the visual world begins with a limited capacity to detect object boundaries, particularly under occlusion, and develops in part as a result of the infant’s interactions with the environment. And as the Gestalt theorists proposed, visual perception is organized at birth and elaborated with experience; many of the organizational principles characteristic of adult vision appear to be operational in infants (if not at birth). These theories have proven prescient and have given direction to many investigations of infant perception and cognition, yet neither theory is fully adequate to explain the foundations of vision and its development. Although our understanding of visual cognitive development continues to grow, the current state of knowledge is substantial and the outlines of a comprehensive account can now be summarized, as follows: l

l

l

l

l

l

l

l

To understand how infants come to experience a stable and predictable world of substantial, volumetric objects, overlapping and extending in depthdthe visual world that we adults experiencedwe must look to experiments that elucidate visual development. Visual development begins well before birth. The visual system begins to develop within weeks after conception, and continues to develop rapidly prior to the onset of patterned visual stimulation. Visual responses to light introduced into the womb indicate that the visual system is at least partly functional before birth. Vision is partially organized at birth. Neonates show systematic scanning patterns and visual preferences, in particular preferences for areas of high contrast and motion. These preferences are well suited for directing attention to features of the visual world relevant to learning about objects. But neonates most likely do not perceive objects as do adultsdas solid and substantial entities. Newborn’s experience of the visual world is fragmented and unstable. Visual and motor systems that yield an experience of coherent objects and the position of the observer relative to a stable environment emerge across the first postnatal year. There is a critical period for development of face and object perception. Normal visual experience during this time is essential to their development, as are patterns of eye movements, and other action systems, in binding features into wholes. Developmental mechanisms include cortical maturation, visual experience, and learning, and the interplay between these developmental events. Developments in some visual functions have been linked directly to maturation of specific cortical regions and visual pathways. Development of smooth pursuit eye movements and motion direction discrimination is thought to stem from maturation of cortical area V5 (also known as MT), form and motion perception from parvocellular and magnocellular processing streams, respectively, and visual memory from structures in the medial temporal lobe. These developments occur between birth and 6 months of age. Infants have multiple means of learning at their disposal, and learning is an indispensable part of understanding the visual world. Infants learn from their own behavior as well as by observing relevant events in the environment.

References Aguiar, A., Baillargeon, R., 1999. 2.5-month-old infants’ reasoning about when objects should and should not be occluded. Cogn. Psychol. 39, 116e157. Amso, D., Haas, S., Markant, J., 2014. An eye tracking investigation of developmental change in bottom-up attention orienting to faces in cluttered natural scenes. PLoS One 9, e85701. Amso, D., Johnson, S.P., 2005. Selection and inhibition in infancy: evidence from the spatial negative priming paradigm. Cognition 95, B27eB36. Amso, D., Johnson, S.P., 2006. Learning by selection: visual search and object perception in young infants. Dev. Psychol. 6, 1236e1245. Amso, D., Johnson, S.P., 2008. Development of visual selection in 3- to 9-month-olds: evidence from saccades to previously ignored locations. Infancy 13, 675e686. Aslin, R.N., 2007. What’s in a look? Dev. Sci. 10, 48e53. Aslin, R.N., Johnson, S.P., May 1994. Suppression of VOR by human infants. In: Poster Presented at the Association for Research in Vision and Ophthalmology Conference, Sarasota, FL. Aslin, R.N., Johnson, S.P., 1996. Suppression of the optokinetic reflex in human infants: implications for stable fixation and shifts of attention. Infant Behav. Dev. 19, 233e240. Atkinson, J., 1984. Human visual development over the first 6 months of life: a review and hypothesis. Hum. Neurobiol. 3, 61e74. Atkinson, J., 2000. The Developing Visual Brain. Oxford University Press, New York. Atkinson, J., Braddick, O., 1981. Development of optokinetic nystagmus in infants: an indicator of cortical binocularity? In: Fisher, D.E., Monty, R.A., Senders, J.W. (Eds.), Eye Movements: Cognition and Visual Perception. Erlbaum, Hillsdale, NJ, pp. 53e64. Banks, M.S., Aslin, R.N., Letson, R.D., 1975. Sensitive period for the development of human binocular vision. Science 190, 675e677. Banks, M., Salapatek, P., 1983. Infant visual perception. In: Mussen, P.H., Haith, M.M., Campos, J.J. (Eds.), Handbook of Child Psychology: vol. 2. Infancy and Developmental Psychobiology, fourth ed. Wiley, New York, pp. 435e572.

356 PART | II Cognitive development

Barr, R., Dowden, A., Hayne, H., 1996. Developmental changes in deferred imitation by 6- to 24-month-olds. Infant Behav. Dev. 19, 159e170. Bauer, P.J., 2004. Getting explicit memory off the ground: steps toward construction of a neuro-developmental account of changes in the first two years of life. Dev. Rev. 24, 347e373. Blakemore, J.E.O., Hill, C.A., 2008. The child gender socialization scale: a measure to compare traditional and feminist parents. Sex. Roles 58, 192e207. Bourgeois, J.,-P., Goldman-Rakic, P.S., Rakic, P., 2000. Formation, elimination, and stabilization of synapses in the primate cerebral cortex. In: Gazzaniga, M.S. (Ed.), The New Cognitive Neurosciences. MIT Press, Cambridge, MA, pp. 45e53. Bremner, J.G., Slater, A.M., Johnson, S.P., 2015. Perception of object persistence: the origins of object permanence in infancy. Child Dev. Perspect. 9, 7e13. Bronson, G., 1990. Changes in infants’ visual scanning across the 2- to 14-week age period. J. Exp. Child Psychol. 49, 101e125. Bronson, G., 1994. Infants’ transitions toward adult-like scanning. Child Dev. 65, 1243e1261. Carey, S., Diamond, R., 1977. From piecemeal to configurational representation of faces. Science 195, 312e314. Cashon, C.H., Cohen, L.B., 2003. Beyond U-shaped development in infants’ processing of faces: an information-processing account. J. Cogn. Dev. 5, 59e80. Clancy, B., Darlington, R.B., Finlay, B.L., 2000. The course of human events: predicting the timing of primate neural development. Dev. Sci. 3, 57e66. Clifton, R.K., Rochat, P., Litovsky, R.Y., Perris, E.E., 1991. Object representation guides infants’ reaching in the dark. J. Exp. Psychol. Hum. Percept. Perform. 17, 323e329. Cohen, L.B., 1976. Habituation of infant visual attention. In: Tighe, R.J., Leaton, R.N. (Eds.), Habituation: Perspectives from Child Development, Animal Behavior, and Neurophysiology. Erlbaum, Hillsdale, NJ, pp. 207e238. Cohen, L.B., Cashon, C.H., 2001. Do 7-month-old infants process independent features or facial configurations? Infant Child Dev. 10, 83e92. Colombo, J., 2001. The development of visual attention in infancy. Annu. Rev. Psychol. 52, 337e367. Constantinescu, M., Hines, M., 2012. Relating prenatal testosterone exposure to postnatal behavior in typically developing children: methods and findings. Child Dev. Perspect. 6, 407e413. Constantinescu, M., Moore, D.S., Johnson, S.P., Hines, M., 2018. Early contributions to infants’ mental rotation abilities. Dev. Sci. 21, e12613. Daw, N.W., 1995. Visual Development. Plenum Press, New York. Dueker, G., Modi, A., Needham, A., 2003. 4.5-month-old infants’ learning, retention, and use of object boundary information. Infant Behav. Dev. 26, 588e605. Dueker, G.L., Needham, A., 2005. Infants’ object category formation and use: real-world context effects on cateory use in object processing. Vis. Cogn. 12, 1177e1198. Dunn, K., Reissland, N., Reid, V.M., 2015. The functioning foetal brain: a systematic preview of methodological factors in reporting foetal visual and auditory capacity. Dev. Cognit. Neurosci. 13, 43e52. Fantz, R.L., 1961. The origin of form perception. Sci. Am. 204, 66e72. Fantz, R.L., 1964. Visual experience in infants: decreased attention to familiar patterns relative to novel ones. Science 146, 668e670. Felleman, D.J., Van Essen, D.C., 1991. Distributed hierarchical processing in the primate cerebral cortex. Cerebr. Cortex 1, 1e47. Fine, I., Wade, A.R., Brewer, A.A., May, M.G., Goodman, D.F., Boynton, G.M., Wandell, B.A., MacLeod, D.I.A., 2003. Long-term deprivation affects visual perception and cortex. Nat. Neurosci. 6, 915e916. Finlay, B.L., Clancy, B., Kingsbury, M.A., 2003. The developmental neurobiology of early vision. In: Hopkins, B., Johnson, S.P. (Eds.), Neurobiology of Infant Vision. Praeger Publishers, Westport, CT, pp. 1e41. Finlay, B.L., Darlington, R.B., 1995. Linked regularities in the development and evolution of mammalian brains. Science 268, 1578e1584. Frank, M.C., Vul, E., Johnson, S.P., 2009. Development of infants’ attention to faces during the first year. Cognition 110, 160e170. Frank, M.C., Amso, D., Johnson, S.P., 2014. Visual search and attention to faces in early infancy. J. Exp. Child Psychol. 118, 13e26. Gandhi, T., Kalia, A., Ganesh, S., Sinha, P., 2015. Immediate susceptibility to visual illusions after sight onset. Curr. Biol. 25, R345eR361. Gandhi, T., Singh, A.K., Swami, P., Ganesh, S., Sinha, P., 2017. Emergence of categorical face perception after extended early-onset blindness. Proc. Natl. Acad. Sci. U.S.A 114, 6139e6143. Gibson, J.J., 1950. The Perception of the Visual World. Greenwood Press, Westport, CT. Gibson, J.J., 1979. The Ecological Approach to Visual Perception. Erlbaum, Hillsdale, NJ. Gilmore, R.O., Baker, T.J., Grobman, K.H., 2004. Stability in young infants’ discrimination of optic flow. Dev. Psychol. 40, 259e270. Gilmore, R.O., Johnson, M.H., 1997. Body-centered representations for visually-guided action emerge during early infancy. Cognition 65, B1eB9. Greenough, W.T., Black, J.E., Wallace, C.S., 1987. Experience and brain development. Child Dev. 58, 539e559. Grimshaw, G.M., Sitarenios, G., Finegan, J.K., 1995. Mental rotation at 7 years: relations with prenatal testosterone levels and spatial play experiences. Brain Cogn. 29, 85e100. Haith, M.M., 1980. Rules that Babies Look by: The Organization of Newborn Visual Activity. Erlbaum, Hillsdale, NJ. Hebb, D.O., 1949. The Organization of Behavior: A Neuropsychological Theory. Erlbaum, Mahwah, NJ. Held, R., 1985. Binocular vision: behavioral and neural development. In: Mehler, J., Fox, R. (Eds.), Neonate Cognition: Beyond the Blooming Buzzing Confusion. Erlbaum, Hillsdale, NJ, pp. 37e44. Henderson, J.M., 2003. Human gaze control during real-world scene perception. Trends Cogn. Sci. 7, 498e504. Hines, M., 2010. Sex-related variation in human behavior and the brain. Trends Cogn. Sci. 14, 448e456. Hunter, M.A., Ames, E.W., 1989. A multifactor model of infant preferences for novel and familiar stimuli. In: Rovee-Collier, C., Lipsitt, L.P. (Eds.), Advances in Child Development and Behavior, vol. 5. Ablex, Norwood, NJ, pp. 69e93.

Development of the visual system Chapter | 16

357

Huttenlocher, P.R., de Courten, C., Garey, L.J., Van Der Loos, H., 1986. Synaptogenesis in human visual cortexdevidence for synapse elimination during normal development. Neurosci. Lett. 33, 247e252. Johnson, M.H., 1990. Cortical maturation and the development of visual attention in early infancy. J. Cogn. Neurosci. 2, 81e95. Johnson, M.H., 2005. Developmental Cognitive Neuroscience, second ed. Blackwell, Oxford, UK. Johnson, S.P., 2001. Neurophysiological and psychophysical approaches to visual development. In: Kalverboer, A.F., Gramsbergen, A., Hopkins, J.B. (Eds.), Handbook of Brain and Behaviour in Human Development. Elsevier, Amsterdam, pp. 653e675. Johnson, S.P., Amso, D., Slemmer, J.A., 2003a. Development of object concepts in infancy: evidence for early learning in an eye tracking paradigm. Proc. Natl. Acad. Sci. U.S.A 100, 10568e10573. Johnson, S.P., Aslin, R.N., 1995. Perception of object unity in 2-month-old infants. Dev. Psychol. 31, 739e745. Johnson, S.P., Bremner, J.G., Slater, A., Mason, U., Foster, K., Cheshire, A., 2003b. Infants’ perception of object trajectories. Child Dev. 74, 94e108. Johnson, S.P., Davidow, J., Hall-Haro, C., Frank, M.C., 2008. Development of perceptual completion originates in information acquisition. Dev. Psychol. 44, 1214e1224. Johnson, S.P., Náñez, J.E., 1995. Young infants’ perception of object unity in two-dimensional displays. Infant Behav. Dev. 18, 133e143. Johnson, S.P., Slemmer, J.A., Amso, D., 2004. Where infants look determines how they see: eye movements and object perception performance in 3month-olds. Infancy 6, 185e201. Johnson, S.P., Shuwairi, S.M., 2009. Learning and memory facilitate predictive tracking in 4-month-olds. J. Exp. Child Psychol. 102, 122e130. Kellman, P.J., Spelke, E.S., 1983. Perception of partly occluded objects in infancy. Cogn. Psychol. 15, 483e524. Kiorpes, L., Movshon, J.A., 2004. Neural limitations on visual development in primates. In: Chalupa, L.M., Werner, J.S. (Eds.), The Visual Neurosciences. MIT Press, Cambridge, MA, pp. 159e188. Köhler, W., 1947. Gestalt Psychology: An Introduction to New Concepts in Modern Psychology. Liveright, New York. Koffka, K., 1935. Principles of Gestalt Psychology. Routledge & Kegan Paul, London. Koffka, K., 1959. In: Ogden, R.M. (Ed.), The Growth of the Mind: An Introduction to Child Psychology. Littlefield, Adams, Patterson, NJ (Original work published 1928). Lamminmäki, A., Hines, M., Kuiri-Hänninen, T., Kilpeläinen, L., Dunkel, L., Sankilampi, U., 2012. Testosterone measured in infancy predicts subsequent sex-typed behavior in boys and in girls. Horm. Behav. 61, 611e616. Le Grand, R., Mondloch, C.J., Maurer, D., Brent, H.P., 2001. Early visual experience and face processing. Nature 410, 890. Linn, M.C., Petersen, A.C., 1985. Emergence and characterization of sex differences in spatial ability: a meta-analysis. Child Dev. 56, 1479e1498. Macchi Cassia, V., Simion, F., Milani, I., Umiltà, C., 2002. Dominance of global visual properties at birth. J. Exp. Psychol. Gen. 131, 398e411. Mareschal, D., Johnson, M.H., 2003. The “what” and “where” of object representations in infancy. Cognition 88, 259e276. Melcher, D., Kowler, E., 2001. Visual scene memory and the guidance of saccadic eye movements. Vis. Res. 41, 3597e3611. Mondloch, C.J., Geldart, S., Maurer, D., Le Grand, R., 2003. Developmental changes in face processing skills. J. Exp. Child Psychol. 86, 67e84. Mondloch, C.J., Le Grand, R., Maurer, D., 2002. Configural face processing develops more slowly than featural face processing. Perception 31, 553e566. Moore, D.S., Johnson, S.P., 2008. Mental rotation in human infants: a sex difference. Psychol. Sci. 19, 1063e1066. Moore, D.S., Johnson, S.P., 2011. Mental rotation of dynamic, three-dimensional stimuli by 3-month-old infants. Infancy 16, 435e445. Morton, J., Johnson, M.H., 1991. CONSPEC and CONLERN: a two-process theory of infant face recognition. Psychol. Rev. 98, 164e181. Needham, A., 1997. Factors affecting infants’ use of featural information in object segregation. Curr. Dir. Psychol. Sci. 6, 26e33. Needham, A., 2001. Object recognition and object segregation in 4.5-month-old infants. J. Exp. Child Psychol. 78, 3e24. Needham, A., Baillargeon, R., 1997. Object segregation in 8-month-old infants. Cognition 62, 121e149. Needham, A., Baillargeon, R., 1998. Effects of prior experience in 4.5-month-old infants’ object segregation. Infant Behav. Dev. 21, 1e24. Needham, A., Dueker, G.L., Lockhead, G., 2005. Infants’ formation and use of categories to segregate objects. Cognition 94, 215e240. Nelson, C.A., 1995. The ontogeny of human memory: a cognitive neuroscience perspective. Dev. Psychol. 31, 723e738. Noton, D., Stark, L., 1971. Scan paths in eye movements during pattern perception. Science 171, 308e331. Oakes, L.M., Ross-Sheehy, S., Luck, S.J., 2006. Rapid development of feature binding in visual short-term memory. Psychol. Sci. 17, 781e787. Ostrovsky, Y., Andalman, A., Sinha, P., 2006. Vision following extended congenital blindness. Psychol. Sci. 17, 1009e1014. Piaget, J., 1952. In: Cook, M. (Ed.), The Origins of Intelligence in Children. International Universities Press, New York (Original work published 1936). Piaget, J., 1954. In: Cook, M. (Ed.), The Construction of Reality in the Child. Basic Books, New York (Original work published 1937). Pintzka, C.W.S., Evensmoen, H.R., Lehn, H., Håberg, A.K., 2016. Changes in spatial cognition and brain activity after a single dose of testosterone in healthy women. Behav. Brain Res. 298, 78e90. Pitti, A., Kuniyoshi, Y., Quoy, M., Gaussier, P., 2013. Modeling the minimal newborn’s intersubjective mind: the visuotopic-somatotopic alignment hypothesis in the superior colliculus. PLoS One 8, e69474. Prechtl, H.F.R., 2001. Prenatal and early postnatal development of human motor behavior. In: Kalverboer, A.F., Gramsbergen, A. (Eds.), Handbook of Brain and Behaviour in Human Development. Kluwer Academic Publishers, Norwell, MA, pp. 415e427. Preston, K.L., Finocchio, D.V., 1983. Development of the vestibulo-ocular and optokinetic reflexes. In: Simons, K. (Ed.), Early Visual Development: Normal and Abnormal. Oxford University Press, New York, pp. 80e88. Putzar, L., Hötting, K., Rösler, F., Röder, B., 2007. The development of visual feature binding processes after visual deprivation in early infancy. Vis. Res. 47, 2616e2626. Quinn, P.C., Liben, L.S., 2008. A sex difference in mental rotation in young infants. Psychol. Sci. 19, 1067e1070. Quinn, P.C., Liben, L.S., 2014. A sex difference in mental rotation in young infants: convergent evidence. Infancy 19, 103e116.

358 PART | II Cognitive development

Reid, V.M., Dunn, K., Young, R.J., Amu, J., Donovan, T., Reeissland, N., 2017. The human fetus preferentially engages with face-like visual stimuli. Curr. Biol. 27, 1825e1828. Richards, J.E. (Ed.), 1998. Cognitive Neuroscience of Attention: A Developmental Perspective. Erlbaum, Mahwah, NJ. Rizzo, M., Hurtig, R., Damasio, A.R., 1987. The role of scanpaths in facial recognition and learning. Ann. Neurol. 22, 41e45. Robbins, R.A., Maurer, D., Mondloch, C.J., Nishimura, M., Lewis, T.L., April 2008. A house is not a face: the effects of early deprivation on the later discrimination of spacing and featural changes in a non-face object. In: Poster Presented at the Annual Meeting of the Cognitive Neuroscience Society, San Francisco. Rochat, P., Goubet, N., 1995. Development of sitting and reaching in 5- to 6-month-old infants. Infant Behav. Dev. 18, 53e68. Rodman, H.R., 2003. Development of temporal lobe circuits for object recognition: data and theoretical perspectives from nonhuman primates. In: Hopkins, B., Johnson, S.P. (Eds.), Neurobiology of Infant Vision. Praeger, Westport, CT, pp. 105e145. Rose, S.A., Feldman, J.F., Jankowski, J.J., 2004. Infant visual recognition memory. Dev. Rev. 24, 74e100. Rose, S.A., Jankowski, J.J., Feldman, J.F., 2008. The inversion effect in infancy: the role of internal and external features. Infant Behav. Dev. 31, 470e480. Ross-Sheehy, S., Oakes, L.M., Luck, S.J., 2003. The development of visual short-term memory capacity in infants. Child Dev. 74, 1807e1822. Rovee-Collier, C., 1999. The development of infant memory. Curr. Dir. Psychol. Sci. 8, 80e85. Schiller, P.H., Tehovnik, E.J., 2001. Look and see: how the brain moves your eyes about. Prog. Brain Res. 134, 127e142. Schöning, S., Engelien, A., Kugel, H., Schäfer, S., Schiffbauer, H., Zwitserlood, P., et al., 2007. Functional anatomy of visuo-spatial working memory during mental rotation is influenced by sex, menstrual cycle, and sex steroid hormones. Neuropsychologia 45, 3203e3214. Schwarzer, G., Zauner, N., Jovanovic, B., 2007. Evidence of a shift from featural to configural face processing in infancy. Dev. Sci. 10, 452e463. Shepard, R.N., Metzler, J., 1971. Mental rotation of three-dimensional objects. Science 171, 701e703. Sirois, S., Mareschal, D., 2002. Models of habituation in infancy. Trends Cogn. Sci. 6, 293e298. Slater, A., 1995. Visual perception and memory at birth. In: Rovee-Collier, C., Lipsitt, L.P. (Eds.), Advances in Infancy Research, vol. 9. Ablex, Norwood, NJ, pp. 107e162. Slater, A., Johnson, S.P., Brown, E., Badenoch, M., 1996. Newborn infants’ perception of partly occluded objects. Infant Behav. Dev. 19, 145e148. Slater, A., Johnson, S.P., Kellman, P.J., Spelke, E.S., 1994. The role of three-dimensional depth cues in infants’ perception of partly occluded objects. Early Dev. Parent. 3, 187e191. Slater, A.M., Morison, V., Rose, D.H., 1983. Perception of shape by the new-born baby. Br. J. Dev. Psychol. 1, 135e142. Slater, A., Morison, V., Somers, M., Mattock, A., Brown, E., Taylor, D., 1990. Newborn and older infants’ perception of partly occluded objects. Infant Behav. Dev. 13, 33e49. Soska, K.C., Adolph, K.A., Johnson, S.P., 2010. Systems in development: motor skill acquisition facilitates 3D object completion. Dev. Psychol. 46, 129e138. Soska, K.C., Johnson, S.P., 2008. Development of 3D object completion in infancy. Child Dev. 79, 1230e1236. Soska, K.C., Johnson, S.P., 2013. Development of three-dimensional completion of complex objects. Infancy 18, 325e344. Spelke, E.S., Breinlinger, K., Macomber, J., Jacobson, K., 1992. Origins of knowledge. Psychol. Rev. 99, 605e632. Sperry, R.W., 1963. Chemoaffinity in the orderly growth of nerve fiber patterns and their connections. Proc. Natl. Acad. Sci. U.S.A 50, 703e710. Teller, D.Y., Movhson, J.A., 1986. Visual development. Vis. Res. 26, 1483e1506. Thier, P., Ilg, U.J., 2005. The neural basis of smooth-pursuit eye movements. Curr. Opin. Neurobiol. 15, 645e652. Turati, C., Simion, F., Milani, I., Umiltà, C., 2002. Newborns’ preference for faces: what is crucial? Dev. Psychol. 38, 875e882. Valenza, E., Bulf, H., 2011. Early development of object unity: evidence for perceptual completion in newborns. Dev. Sci. 14, 799e808. Valenza, E., Simion, F., Macchi Cassia, V., Umiltà, C., 1996. Face preference at birth. J. Exp. Psychol. Hum. Percept. Perform. 22, 892e903. Valenza, E., Simion, F., Umiltà, C., 1994. Inhibition of return in newborn infants. Infant Behav. Dev. 17, 293e302. Van Essen, D.C., Anderson, C.H., Felleman, D.J., 1992. Information processing in the primate visual system: an integrated systems perspective. Science 255, 419e423. Van Essen, D.C., Maunsell, J.H.R., 1983. Hierarchical organization and the functional streams in the visual cortex. Trends Neurosci. 6, 370e375. van Renswoude, D.R., Visser, I., Raijmakers, M.E.J., Tsang, T., Johnson, S.P., 2019. Real-world scene perception in infants: what factors guide attention allocation? Infancy 24, 693e717. von Hofsten, C., 2004. An action perspective on motor development. Trends Cogn. Sci. 8, 266e272. von Hofsten, C., Kochukhova, O., Rosander, K., 2007. Predictive tracking over occlusions by 4-month-old infants. Dev. Sci. 10, 625e640. von Hofsten, C., Vishton, P., Spelke, E.S., Feng, Q., Rosander, K., 1998. Predictive action in infancy: tracking and reaching for moving objects. Cognition 67 (3), 255e285. Voyer, D., Voyer, S., Bryden, M.P., 1995. Magnitude of sex differences in spatial abilities: a meta-analysis and consideration of critical variables. Psychol. Bull. 117, 250e270. Wiesel, T.N., Hubel, D.H., 1963. Single-cell responses in striate cortex of kittens deprived of vision in one eye. J. Neurophysiol. 26, 1003e1017. Wong, R.O.L., 1999. Retinal waves and visual system development. Annu. Rev. Neurosci. 22, 29e47. Yin, R.K., 1969. Looking at upside-down faces. J. Exp. Psychol. 81, 141e145. Younger, B., 1992. Developmental change in infant categorization: the perception of correlations among facial features. Child Dev. 63, 1526e1535. Zilles, K., Werners, R., Busching, U., Schleicher, A., 1986. Ontogenesis of the laminar structure in areas 17 and 18 of the human visual cortex: a quantitative study. Anat. Embryol. 174, 339e353. Zuckerman, C.B., Rock, I., 1957. A re-appraisal of the roles of past experience and innate organizing processes in visual perception. Psychol. Bull. 54, 269e296.

Chapter 17

The development of visuospatial processing Joan Stiles1, 2, Natacha A. Akshoomoff 2, 3 and Frank Haist2, 3 1

Department of Cognitive Science, University of California, San Diego, La Jolla, CA, United States; 2Center for Human Development, University of

California, San Diego, La Jolla, CA, United States; 3Department of Psychiatry, University of California, San Diego, La Jolla, CA, United States

Chapter outline 17.1. The development of visuospatial processing 359 17.1.1. Anatomical organizations of the primary visual systems 361 17.1.2. Ventral stream processes 361 17.1.2.1. Perception of the global and local levels of visual pattern structure 362 17.1.2.2. Perception of faces 364 17.1.2.3. Spatial construction 365 17.1.3. Dorsal stream processes 366 17.1.3.1. Spatial localization 366 17.1.3.2. Spatial attention 368

17.1.3.3. Mental rotation 17.1.4. Trajectories of dorsal and ventral stream development 17.1.5. Neurodevelopmental disorders of visuospatial processing 17.1.5.1. Perinatal stroke 17.1.5.2. Spina bifida 17.1.5.3. Neurogenetic syndromes 17.1.6. Summary and conclusions References

369 371 372 373 376 376 379 380

17.1 The development of visuospatial processing Visual input is a critical source of knowledge about the organization and structure of the spatial world. It provides information about everything from the structure of objects and scenes to their location or movement in space. Visuospatial processing encompasses a wide variety of neurocognitive abilities ranging from the basic ability to analyze how parts or features of an object combine to form an organized whole to the dynamic and interactive spatial processes required to track moving objects, to visualize displacement, and to localize, attend, or reach for objects or visual targets in a spatial array. These varied processes work in concert to provide a seamless and immediate perception of the intricacies of the visual world. This perception provides an essential basis for precise and effective action in the world as well as a rich source of input for cognitive functions across many domains. A complex neural architecture involving dozens of interrelated visual areas in the posterior cortices supports visuospatial processing (Van Essen et al., 1992). Ungerleider and Mishkin (1982) first proposed a model for understanding the organization of this complex set of cortical areas and functions in the early 1980s (Ungerleider, 1995). In their model, the cortical visual system is anatomically and functionally subdivided into the ventral and dorsal processing pathways or streams (see Fig. 17.1). The ventral stream is dominant for processing information about patterns and objects, while the dorsal stream mediates spatial processing associated with attention to movement and location. Subsequent models describe dorsal stream functions as specialized for supporting visual processing related to action (e.g., Andersen et al., 1997; Goodale and Milner, 1992; Goodale and Westwood, 2004; Rizzolatti and Matelli, 2003; Milner, 2017). This chapter begins with a summary of the neuroarchitecture of the ventral and dorsal visual streams. The summary focuses on the flow of visual information beginning, for both streams, in the primary visual cortex and then extending to

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00017-2 Copyright © 2020 Elsevier Inc. All rights reserved.

359

360 PART | II Cognitive development

FIGURE 17.1 Dorsal and ventral visual processing pathways in monkey. Solid lines indicate connections arising from both central and peripheral visual field representations; dotted lines indicate connections restricted to peripheral field representations. Red boxes indicate ventral stream areas related primarily to object vision; green boxes indicate dorsal stream areas related primarily to spatial vision; and white boxes indicate areas not clearly allied with either stream. Shaded region on the lateral view of the brain represents the extent of the cortex included in the diagram. DP, dorsal prelunate area; FST, fundus of superior temporal area; HIPP, hippocampus; LIP, lateral intraparietal area; MSTc, medial superior temporal area, central visual field representation; MSTp, medial superior temporal area, peripheral visual field representation; MT, middle temporal area; MTp, middle temporal area, peripheral visual field representation; PO, parietaleoccipital area; PP, posterior parietal sulcal zone; STP, superior temporal polysensory area; V1, primary visual cortex; V2, visual area 2; V3, visual area 3; V3A, visual area 3, part A; V4, visual area 4; and VIP, ventral intraparietal area. Inferior parietal area 7a; prefrontal areas 8, 11e13, 45, and 46, perirhinal areas 35 and 36; and entorhinal area 28 are from Brodmann and Garey (2007). Inferior temporal areas TEO and TE, parahippocampal area TF, temporal pole area TG, and inferior parietal area PG are from Von Bonin and Bailey (1947). Rostral superior temporal sulcal (STS) areas are from Seltzer and Pandya (1978), and VTF is the visually responsive portion of area TF (Boussaoud et al., 1991). Reproduced from Ungerleider, L.G., 1995. Functional brain imaging studies of cortical mechanisms for memory. Science 270 (5237), 769e775, with permission.

the temporal and parietal lobes for the ventral and dorsal streams, respectively. Connections between the two major visual pathways as well as connections with the frontal lobes are also considered. The next two sections consider the development of cognitive processes associated with the two principal brain visual systems. The section on ventral stream processing examines the development of visual pattern processing from infancy through adolescence focusing on changes in the perception of visual patterns and faces, and in the ability to construct spatial arrays. The section on dorsal stream processing examines the development of spatial attention, location processing, and mental rotation. The final section of the chapter turns the focus to neurodevelopmental disorders where visuospatial processing is a primary feature. It examines both the effects of frank neural insult on the development of spatial processes and the effects of specific genetic abnormalities on the development of the neural systems that underlie the development of spatial processes. The original descriptions of ventral and dorsal stream organization came from studies of adults with injury to various subsystems within the cortical visual pathways. Data from children with neurodevelopmental disorders provide insight into the emergence of visual system organization following early pathology, and can address questions about how specific neural compromise and neural plasticity interact to affect the developmental trajectories of basic visuospatial functions and the neural systems that mediate them.

The development of visuospatial processing Chapter | 17

361

17.1.1 Anatomical organizations of the primary visual systems The organization of the primary visual pathways has been most fully described for rhesus macaque monkeys; thus, the description presented here uses the nomenclature typically used for nonhuman primates. However, the basic pathways in humans and monkeys appear to be largely homologous (Brewer et al., 2002; Orban et al., 2004). The ventral visual pathway begins at the retina and projects via the lateral geniculate nucleus (LGN) of the thalamus to the primary visual cortex, area V1. From there, the pathway proceeds to extrastriate visual areas V2 and V4, and then projects ventrally to the posterior (PIT) and anterior (AIT) regions of the inferior temporal lobe. Input to the ventral pathway is derived principally, though not exclusively, from P-type retinal ganglion cells that project to the parvocellular layers of the LGN and then to layer 4C beta of V1. Parvocellular input to V1 organizes into distinct areas called the blob and interblob regions (Livingstone and Hubel, 1984; Wong-Riley, 1979; Kaas and Collins, 2004). Cells in the blob regions are maximally sensitive to form, while cells in the interblob regions respond principally to color. The ventral stream processes information about visual properties of objects and patterns, and has been described as the what pathway. The dorsal visual pathway also begins at the retina and projects via the LGN to area V1. From there, the pathway proceeds to extrastriate areas V2 and V3, then projects dorsally to the medial (MT/V5) and medial superior (MST) regions of the temporal lobe, and then to the ventral inferior parietal (IP) lobe. Input to the dorsal pathway is derived principally, though not exclusively, from the large M-type retinal ganglion cells that project to the magnocellular layers of LGN and then to layer 4C alpha of V1. Cells in this pathway are maximally sensitive to movement and direction and are less responsive to color or form. The original functions identified for the dorsal stream involved processing of information about spatial location, optic flow, motion, and allocation and maintenance of spatial attention. It was thus described as the where pathway. More recently, work examining the dorsal stream’s role in visually guided movements suggests that the pathway is also involved in the integration of visual and motor functions. It has thus been called the how system (e.g., Andersen et al., 1997; Goodale and Milner, 1992; Rizzolatti and Matelli, 2003; Goodale and Westwood, 2004; Goodale, 2014). The dorsal and ventral pathways project rostrally to common and distinct, albeit adjacent, areas of the prefrontal cortex. Imaging studies suggest that these prefrontal networks are involved in a variety of dorsal and ventral stream functions (Farivar, 2009). For example, spatial working memory and attention rely on networks connecting the dorsolateral prefrontal cortex (DLPFC) and posterior parietal cortex (Corbetta et al., 2002; Awh and Jonides, 2001; Curtis, 2006), whereas object memory relies on systems connecting the prefrontal cortex with inferior temporal (Ranganath, 2006; Ranganath et al., 2004; Ranganath and D’Esposito, 2005; Mackey and Curtis, 2017). At least three principal projection pathways from the parietal lobe have been described: a parietal prefrontal pathway mediating eye movement and spatial working memory, a parietal premotor pathway mediating visually guided movement (eye movement, reach, and grasp), and a parietal medial temporal pathway that processes complex spatial information for navigation (Kravitz et al., 2011). There is substantial evidence that the dorsal and ventral pathways are richly interconnected and at least partially overlapping in the mature (e.g., Dobkins and Albright, 1994, 1995, 1998; Marangolo et al., 1998; Merigan and Maunsell, 1993; Sincich and Horton, 2005; Thiele et al., 2001; Rosa et al., 2009; Zanon et al., 2010; Takahashi et al., 2013) and the developing (Dobkins and Anderson, 2002; Dobkins and Teller, 1996a,b) visual system. The dissociation of function across the two pathways may be less complete than originally thought. Subregions within each system may respond to functions typically associated with the other pathway (Lehky and Sereno, 2007; Kawasaki et al., 2008; Konen and Kastner, 2008; Husain and Nachev, 2007). For example, regions in the parietal lobe may respond to color and shape features (Kawasaki et al., 2008), and area MT/V5 in extrastriate visual cortex may show object selective responses (Konen and Kastner, 2008).

17.1.2 Ventral stream processes A major function of the ventral visual stream is the analysis of pattern information. Here, findings from three specific functions within the ventral stream are summarized: globalelocal processing, face processing, and spatial construction. Behaviorally, visuospatial analysis is defined as the ability to specify the parts and the overall configuration of a visually presented pattern, and to understand how the parts are related to form an organized whole (e.g., Delis et al., 1988, 1986; Palmer, 1977, 1980; Palmer and Bucher, 1981; Robertson and Delis, 1986; Smith and Kemler, 1977; Vurpillot, 1976). Thus, it involves the ability to segment a pattern into a set of constituent parts (referred to as featural or local level processing), and integrate those parts into a coherent whole (referred to as configural or global level processing). Different approaches to the study of spatial analysis have focused on level of processing and type of input. Perceptual processing studies focus largely on issues of global versus local or configural versus featural processing. Much of the data on perceptual processing of global and local aspects of objects and patterns come from hierarchical form processing tasks (see Fig. 17.2). Perceptual processing of faces is a related but generally distinct line of study. Faces constitute a critically

362 PART | II Cognitive development

FIGURE 17.2 The model forms (top) provide examples of hierarchical form stimuli. Hierarchical form stimuli have two levels of organization: a large global/configural level comprised of appropriately arranged smaller forms constituting the local/featural level. A series of hierarchical form stimuli (the models) were presented one at a time and children were given 10 s to study the form. After a 30 s delay, they were asked to reproduce the forms from memory. The graph illustrates that systematic improvement in the accuracy of reproductions was observed among typically developing 4- to 8-year-old children, and also that there are no differences at any age in the relative accuracy of reproducing the global and local levels of the forms. At each age, children are equally accurate in their reproduction of global and local pattern information. Reproduced from Dukette, D., Stiles, J., 2001. The effects of stimulus density on children’s analysis of hierarchical patterns. Dev. Sci. 4 (2), 233e251, with permission.

important class of social stimuli for which most individuals acquire considerable processing expertise. Because of the importance of the information faces provide to typical social interaction and communication, faces may be processed differently compared to other classes of visual objects. Construction tasks provide a window to children’s conceptual organization of spatial arrays. The processes and strategies children use to recreate spatial scenes can provide insight into their understanding of their spatial world.

17.1.2.1 Perception of the global and local levels of visual pattern structure Differential laterality for global and local processing is well documented for adults with right hemisphere (RH) dominance for global processing and left hemisphere (LH) dominance for local processing (e.g., Han et al., 2002; Martin, 1979; Martinez et al., 1997b; Sergent, 1982; Volberg and Hubner, 2004; Yovel et al., 2001; Flevaris and Robertson, 2016; Flevaris et al., 2014). Sergent (1982) suggested that these differences arise from preferential processing of lower spatial frequencies in the RH and higher spatial frequencies in the LH. Several experiments have presented sinusoidal gratings containing a single spatial frequency presented to the right visual field (RVF) or left visual field (LVF) to evaluate this hypothesis with generally positive results. Low spatial frequencies elicit faster responses when presented to the LVF-RH than the RVF-LH, while high spatial frequencies elicit the opposite pattern (Kitterle et al., 1990, 1992; Kitterle and Selig, 1991). Event-related potential (ERP) and other functional imaging studies have supported these basic patterns of lateralization (e.g., Martinez et al., 1997a; Fink et al., 1997; Heinze et al., 1998; Flevaris et al., 2014; Kauffmann et al., 2014).

The development of visuospatial processing Chapter | 17

363

In addition to the evidence for the laterality of global and local level processing, there is also strong evidence of a globalelocal processing asymmetry. Specifically, inconsistent or competing information at the global level interferes with local processing, but inconsistent local information does not affect global processing. These two findings led to the postulation of a global precedence effect in visual pattern processing, which states that global level information is processed prior to local level information (Navon, 1977). Although many factors may mitigate the global precedence effect in adults, it remains a robust finding within the standard task (Ivry and Robertson, 1998; Kimchi, 1992; Navon, 2003; Robertson and Delis, 1986; Robertson et al., 1993; Robertson and Lamb, 1991; Flevaris and Robertson, 2016). The ability to analyze spatial patterns begins to emerge in the first year of life. Newborns exhibit configural preferences and rudimentary partewhole (Cassia et al., 2002; Farroni et al., 2000; Quinn et al., 1993; Slater et al., 1991; Guy et al., 2017, 2013). There are dramatic changes in the complexity of visual pattern processing reflecting a systematic improvement in the infant’s ability to process global and local level pattern information across the first year of life (Cohen and Younger, 1984). These patterns of change appear to reflect early hemispheric differences in processing. Infants as young as 4 months exhibit lateralized processing differences on global and local processing tasks similar to those observed in adult neuroimaging studies (Deruelle and de Schonen, 1998, 1991). Systematic changes are also observed in infant’s response to high and low frequency spatial patterns. At 6e7 months, infants show both a preference for low frequency stimuli and require longer presentation times to detect high frequency stimuli, as compared to 12e13 month olds (Otsuka et al., 2014). Studies using the standard hierarchical form stimuli (see Fig. 17.2) have also consistently documented a protracted period of developmental change in globalelocal processing that extends well into adolescence (Dukette and Stiles, 1996, 2001; Harrison and Stiles, 2009; Mondloch et al., 2003; Moses et al., 2002; Porporino et al., 2004; Vinter et al., 2010; Scherf et al., 2009; Krakowski et al., 2016). The classic global precedence effect emerges slowly over the course of development. Although some studies of children report a global processing bias (Cassia et al., 2002; Mondloch et al., 2003; Moses et al., 2002; Porporino et al., 2004), others report only modest effects that are modulated by altering task and stimulus demands. For example, Nayar et al. (2015) using illusory contour stimuli with 3- to 10-year-old children, report a gradual shift from local to global processing in the 4- to 7-year age range. Increased task demands (Harrison and Stiles, 2009) and selective degradation of the global level stimulus induce a shift in processing bias from the global to the local level that is much more pronounced in children than in adults (Dukette and Stiles, 1996). Scherf et al. (2009, #394) tested 8- to 30-year-old participants on both the standard hierarchical processing task and a primed matching paradigm originally introduced by Kimchi (1998), and found the global precedence effect emerges over a protracted period that extends well into adolescence. In addition, Krakowski et al. (2016, #370) have demonstrated the effects of inhibitory control on global precedence in middle childhood. Finally, studies of spatial frequency processing in childhood present a somewhat complex picture. While children across the 3- to 15-year age range can process both high and low spatial frequency stimuli, children under age 7 show delayed responses for high spatial frequency stimuli (van den Boomen et al., 2015), while the reverse is true for children over age 7 (Rokszin et al., 2018). The combined data from studies of hierarchical form processing show that children are clearly able to engage in global and local level processing from a very early age. However, stable and mature levels of visuospatial processing emerge slowly over a protracted period of development. Variations in stimulus and task demands play an important role in modulating the dominant level of processing. Thus, the functional role of a global or local processing bias or advantage may be different during development than it is later in life, and may reflect growing expertise and facility in processing complex visuospatial patterns. Imaging studies of typical children confirm the behavioral findings and suggest that the neural systems associated with spatial analytic processing undergo a protracted period of development. Moses et al. (2002) tested children between 11 and 15 years of age using a hemifield reaction time (RT) task and functional magnetic resonance imaging (fMRI) protocols identical to those used by Martinez et al. (1997a) with adult subjects. The pattern of RT data obtained from children across this age range differed from that of adults. Similar to the findings from the Mondloch et al. (2003) study, children were faster with global than with local targets and did not manifest the kinds of hemifield RT differences observed among adults. Importantly, children’s profiles of activation in the fMRI study differed from those of the adults. For the global and local tasks, children showed statistically greater activation in the RH than in the LH. Overall activation among children was greater than among adults, and children showed considerably more bilateral activation particularly on the local processing tasks than adults. Thus, at least for these perceptually demanding tasks, children showed a global processing advantage and overall RH dominance. Anatomical changes are shown to be associated with the shift from local to global processing biases in children. In one study, 6-year-old children were assigned to one of two groups depending on performance on a behavioral globalelocal processing task (Poirel et al., 2011). One group of children exhibited the mature profile of global level bias, and the other the more immature local level bias profile. The investigators used voxel-based morphology to assess group gray matter density differences in brain regions implicated in global processing, specifically the calcarine sulcus, the inferior occipital

364 PART | II Cognitive development

gyrus, the RH occipital lingual gyrus, the right parietal precuneus, and the precentral gyrus. The group of children exhibiting the behaviorally more mature “global bias” showed reduced gray matter density in all of the brain regions associated with global level processing. A subsequent study by this same group (Poirel et al., 2014) reported cortical thinning in occipital regions bilaterally, as well as increased cortical thickness in left frontoparietal regions in the more mature group. These findings suggest a link between brain maturation and performance on this important spatialprocessing task.

17.1.2.2 Perception of faces The ability to recognize a face is essential for everyday social exchange. Although such recognition depends on the discrimination of subtle differences among faces, the task of identifying a face is effortless for adults, suggesting considerable expertise with this important class of stimuli (see also Chapter 20 for a more extended discussion of the development of face processing). Face processing is thought to rely disproportionately on configural cues such as the spacing between features. For example, unlike other objects, face recognition is significantly impaired when the stimuli are turned upside down (Rossion and Gauthier, 2002). It has been suggested that face inversion selectively disrupts facial configural information processing (Yin, 1969). A recent developmental study suggests that face configuration processing matures in the early childhood period, but large individual differences are observed through early adulthood (Petrakova et al., 2018). Neuroimaging studies with adults find a strong RH bias for face activation within what has been described as the core brain network for face processing (Epstein et al., 2006; Gauthier et al., 2005; Grill-Spector et al., 2004; Kanwisher et al., 1997, 1999; Mazard et al., 2006; Rhodes et al., 2004; Wojciulik et al., 1998; Xu, 2005; Yovel and Kanwisher, 2004, 2005). The core face network is a ventral occipitaletemporal (VOT) system that includes the middle aspects of the lateral fusiform gyrus (Brodmann’s area 37), often referred to as the fusiform face area (FFA), the inferior occipital gyrus in Brodmann’s area 18, often referred to as the occipital face area (OFA), and the posterior superior temporal sulcus (Haxby et al., 2000). Recent evidence suggests that the FFA may contain multiple distinct subdivisions sensitive to face processing (for review, see Weiner and Zilles, 2016). Preference for face stimuli has been documented from the first hours of life (Johnson et al., 1991). Infants as young as 2e3 months show selective cortical responses to faces (Halit et al., 2004; Tzourio-Mazoyer et al., 2002). Some studies suggest that infants show an RH bias for faces (de Schonen and Deruelle, 1991; de Schonen and Mathivet, 1990; de Schonen et al., 1996; Reynolds and Roth, 2018). Despite these early competences, there is overwhelming evidence for developmental change in face processing that extends at least through the school-age period (Chung and Thomson, 1995; Taylor et al., 2001). Early studies suggested that changes in face processing might reflect a shift from a feature-based to a more configural or analytic strategy (Carey and Diamond, 1977; Diamond and Carey, 1986; Tanaka and Farah, 1993). However, accumulating evidence supports a pattern of slower, quantitative age-related change (Itier and Taylor, 2004; Taylor et al., 1999) and increasingly more effective use of the same types of cues used by adults (Freire and Lee, 2001; Baenninger, 1994). These kinds of change may be associated with the acquisition of greater expertise in processing faces and other visual objects (Carey, 1996; Diamond and Carey, 1986; Gauthier and Nelson, 2001). Recent developmental fMRI studies of face processing suggest that the core brain network for face processing undergoes a protracted change that extends through the school-age period into adolescence (Haist et al., 2013; Passarotti et al., 2003; Gathers et al., 2004; Golarai et al., 2007; Grill-Spector et al., 2008; Aylward et al., 2005). Most developmental studies have focused on individual components within the core VOT network, particularly the FFA. The preponderance of evidence indicates that school-age children may produce reliable FFA activation, but the patterns of activation within the fusiform gyrus region vary considerably from those observed among adults. Systematic increases in fMRI blood oxygen level dependent signal activation both in terms of the extent (Haist et al., 2013; Brambati et al., 2010; Golarai et al., 2007; Peelen et al., 2009) and intensity of activation (Brambati et al., 2010; Cohen Kadosh et al., 2011; Golarai et al., 2007; Joseph et al., 2011) have been reported from the early school-age period through adulthood. These developmental activation changes correlate with improvement in recognition memory for faces (Golarai et al., 2007, #90; Golarai et al., 2010, #241) and with task demand (Scherf et al., 2011). A few studies have looked at changes in the organization of brain networks for face processing outside the core face network, specifically in regions falling within the “extended” face network. Extended face network regions tend to be recruited in a task-specific manner; for example, amygdala and anterior cingulate cortex for emotion recognition and monitoring, the anterior temporal pole and dorsolateral prefrontal cortex for semantic retrieval of identity and contextual information (Haxby et al., 2000). Haist et al. (2013) found evidence that many regions within the extended face network show greater activity in children than adolescents and adults, suggesting that the modulation of these regions has a protracted developmental time course. With age and growing expertise, these networks become more focused and task-specific (Joseph et al., 2011).

The development of visuospatial processing Chapter | 17

365

17.1.2.3 Spatial construction Spatial construction tasks such as drawing or block assembly provide insight into an individual’s conceptualization of the organization of spatial arrays. They can reveal how the participant construes both the parts of an array and the relations among parts that combine to form the overall configuration. Studies of adults with unilateral brain injury use construction tasks extensively. These studies consistently report lateralized differences in the kinds of construction errors produced. Specifically, adults with injury to right posterior brain regions are able to identify, or segment, the parts of spatial forms but have difficulty organizing these parts into integrated spatial configurations. In contrast, adults with injury to left posterior brain regions are able to reproduce the overall pattern configuration, but fail to incorporate pattern detail and tend to simplify the spatial arrays (Akshoomoff et al., 1989; Delis et al., 1986, 1988; Piercy et al., 1960; Shorr et al., 1992). Studies of children’s spontaneous spatial construction activities suggest that before 12 months, children engage in very little systematic organization of objects (Forman, 1982; Gesell, 1925; Guanella, 1934; Langer, 1980). In block construction tasks stacking begins at about 12 months, and by 18 months, children begin to arrange blocks in lines by placing the blocks next to one another (Bayley, 1969; Forman, 1982; Gesell, 1925; Stiles-Davis, 1988). It is not until 3e4 years that children regularly build both vertical and horizontal components within a single spatial construction (Guanella, 1934; Stiles-Davis, 1988). There is also systematic change in processes used to generate block constructions (Stiles and Stern, 2001; Stiles-Davis, 1988). At 24 months, children rely upon a simple repetitive process with a single relation (e.g., stacking). By 36 months, they can use more than one relation (e.g., including a stack and a line in the same construction), but they typically generate them in sequence (e.g., completing the stack, and then the line). By 48 months, children are able to produce multicomponent constructions, with multiple spatial relations, extending in multiple directions in space, and with multiple points of contact between components (e.g., shifting between the line and the stack while building; creating multicomponent construction such as a bridge and several roads). More structured spatial construction tasks that require the child to copy models reveal other factors that can affect spatial construction performance. Verdine et al. (2014) showed that both model complexity, as measured by the number of construction elements, and task demands, such as requirements to rotation or translation construction elements, affect task performance in preschool age children. Zingrone (2014) demonstrated that the complexity of pattern symmetry in model constructions affect performance in children as old as 12 years of age. Similar changes are observed in studies of children’s drawings. Prather and Bacon (1986) showed that children can attend to either the parts or the whole of a spatial pattern, but their performance can be influenced by specific task and stimulus manipulations. Data from a large series of studies using different measures with children ranging in age from 3 to 12 years show that initially children segment out well-formed, independent parts and use simple combinatorial rules to integrate the parts into the overall configuration (Akshoomoff and Stiles, 1995a,b; Feeney and Stiles, 1996; Stiles and Stern, 2001; Tada and Stiles, 1996). Across the preschool and school-age period, change is observed in both the nature of the parts and the relations children use to organize the parts. Further, pattern complexity affects how children approach the problem of analysis. Martens et al. (2014) examined developmental change in strategies used to reproduce the Rey Osterreith Complex Figure (ROCF; see Fig. 17.3). They reported that 5-year-olds use more local strategies focused on

FIGURE 17.3 Models of the Rey-Osterrieth Complex Figure (ROCF) and the three simple geometric forms are shown on the left. Children were given unlimited time to copy each model form. On the right side are examples from 6- and 8-year-old children’s copies of the ROCF, and a representative example of a 5-year-old child’s copies of the simple geometric forms. Although the simpler forms are contained within the ROCF, children are more accurate in reproducing them in isolation than in the context of the more complex form. Reproduced from Akshoomoff, N.A., Stiles, J., 1995a. Developmental trends in visuospatial analysis and planning: I. Copying a complex figure. Neuropsychology 9 (3), 364e377, with permission.

366 PART | II Cognitive development

individual lines or elements, but by age 7 more global approaches such as beginning with the core rectangle begin to emerge. A study comparing drawing strategies for the ROCF and simplified variants (see Fig. 17.3) demonstrated that pattern complexity affects performance. While the local to global shift in strategies similar to that reported by Martens et al. (2014) was observed for the ROCF, the use of simpler models induced more advanced reproduction strategies (Akshoomoff and Stiles, 1995a,b).

17.1.3 Dorsal stream processes A variety of spatial processes have been associated with dorsal visual stream processing. We consider three of these processes in this section: spatial localization, spatial attention, and mental rotation. The characterization of these basic dorsal system processes as independent and distinct from those of the ventral stream is somewhat artificial in that, for example, localization of an object may also require a shift in spatial attention, and mental translation of an object must involve both localization and attention in space. Nonetheless, there is substantial evidence for functional and anatomical independence of key features of each process.

17.1.3.1 Spatial localization Evidence from human and animal studies shows that the dorsal stream plays a critical role in perceptual localization (Belger et al., 1998; Chiba et al., 2002). In a series of studies using positron emission tomography (PET) imaging, Haxby and colleagues examined profiles of posterior brain activation using tasks that required adults to compare the location of objects in two visually presented arrays (Haxby et al., 1991, 1994). In addition to activation in bilateral extrastriate cortex presumed to be involved in early visual processing, there was also robust activation of bilateral regions of parietal lobe, including posterior superior parietal areas extending rostrally to the intraparietal sulcus (Brodmann’s area 7). These human brain activation findings on location processing are consistent with animal studies (Colby and Duhamel, 1996; Colby and Goldberg, 1999; Rizzolatti and Matelli, 2003), and have been largely replicated in subsequent fMRI, PET, patient, and transcranial magnetic stimulation studies (Belger et al., 1998; Casey et al., 1998; Ellison and Cowey, 2006; Jonides et al., 1993; Nelson et al., 2000; Oliveri et al., 2001; Smith et al., 1996, 1995; Jerde and Curtis, 2013; Mackey et al., 2016). Moreover, subsequent studies identified the IP lobe as important in perceptual processing of location (Colby and Duhamel, 1996; Courtney et al., 1996). A large number of functional neuroimaging studies have demonstrated the importance of frontal regions in spatial working memory for locations. Two regions that appear to be particularly important for spatial working memory in humans include the superior frontal cortex (Atkinson and Braddick, 2012; Courtney et al., 1998; Curtis, 2006; Haxby et al., 2000; Sala et al., 2003) and DLPFC (Atkinson and Braddick, 2012; Curtis, 2006; Postle et al., 2000). The task of looking or reaching to a spatial location involves a complex network of neural areas within the dorsal frontoparietal system (Colby and Duhamel, 1996; Colby and Goldberg, 1999; Johnson et al., 1996; Pierrot-Deseilligny et al., 2004; Rizzolatti and Matelli, 2003; Wise et al., 1997). Prefrontal motor areas mediate planning and preparation for motor action; activation of these areas typically precedes the actual motor event. There is considerable evidence for superior parietal input to dorsal premotor and motor cortices; activation in frontal and superior parietal areas is concordant, suggesting a network of spatial-motor control (Rizzolatti and Matelli, 2003; Rizzolatti and Sinigaglia, 2010; Rizzolatti et al., 2014; Bernier et al., 2017). In addition, IP areas connect to frontal premotor areas and play an important modulatory role in spatial-motor activity (Andersen et al., 1997). Rizzolatti and Matelli (2003, 2010), Rizzolatti et al. (2014) have suggested that the dorsal system may comprise two separate but interrelated systems: an IP system dominated by visual perceptual inputs and a superior parietal system governed by somatosensory information that is used to guide action. Location processing is postulated to rely on the computation of two distinct types of relations: categorical and coordinate (Kosslyn, 1987, 2006; Kosslyn et al., 1992, 1989; Ruotolo et al., 2016) Categorical relations provide generalized abstract positional information about the relative location of two elements, such as above/below or right/left. Coordinate relations provide precise metric information about spatial relations. Neuroimaging studies have implicated posterior parietal regions for both categorical and coordinate relational processing (Kosslyn et al., 1989, #122; Kosslyn et al., 1995a, 1998; Trojano et al., 2002) but the laterality of the two processes appears to differ. Specifically, categorical processing is LH dominant, while coordinate processing is RH dominant (Kosslyn et al., 1989, 1995a; Kosslyn, 2006; van der Ham et al., 2009). One of the largest bodies of data on the early development of visuospatial processing comes from a simple, spatial hiding task, originally introduced by Piaget (1952). Infants watch as a toy is hidden under one of two screens (A or B) and are then encouraged to retrieve it. Eight-month-olds easily retrieve the object hidden under A (but also see Smith et al., 1999), but when the object is then hidden under B, they continue to search at A committing what has been termed the A-not-B error (AB error). This error has been widely conceptualized as an index of object permanence, that is, of the

The development of visuospatial processing Chapter | 17

367

infant’s knowledge that objects exist independently over space and time. A wide range of factors have been shown to affect the likelihood of making the AB error. For example, the beginning of self-locomotion reduces the likelihood of AB error (Bertenthal and Campos, 1990; Horobin and Acredolo, 1986; Kermoian and Campos, 1988). In addition, healthy preterm infants are more advanced on the AB search task compared to full-term peers matched for gestational age, suggesting that extra experience in the world offers the healthy preterm infants a developmental advantage (Matthews et al., 1996). Altering task demands affects AB task performance. When the task requires to look rather than reach for the object, search performance is enhanced. Other factors, such as the use of salient landmarks, distinctive screens, or increased distance between the screens, also improve performance (Butterworth et al., 1982; Wellman et al., 1987). By contrast, increasing task demands by increasing the delay between hiding and search negatively impacts performance. Introduction of a delay between hiding and retrieval increases error frequency among children as old as 12 months (Diamond, 1985; Spencer et al., 2001). Finally individual differences in temperament also affect performance (Johansson et al., 2014). Although neuropsychological data on AB task performance are limited, several studies implicate the DLPFC. In adult rhesus monkeys, bilateral lesions of the DLPFC disrupt AB search task performance (Diamond, 1991; Diamond et al., 1994). Studies using near-infrared spectroscopy to measure localized brain activation in infants provide converging evidence for the association between frontal lobe development and successful search performance (Baird et al., 2002). Electroencephalography (EEG) data have been used to examine potential markers of object representation. Gamma-band activity has been associated with maintenance of mental representations of objects among adults (Tallon-Baudry et al., 1998). Studies measuring gamma-band activity in the EEG of 6-month-old infants during object processing and object occlusion tasks suggest that the neural signature of object representation can be detected by the middle of the first year of life (Csibra et al., 2000; Kaufman et al., 2003, 2005). Systematic changes in EEG power and EEG coherence across widely distributed brain regions have been reported across the first year of life (Bell, 2012; Cuevas and Bell, 2011; Cuevas et al., 2012). In summary, these data suggest that a complex network of neural systems emerge across the first year of life to support performance on this seemingly simple task. The data point to changes in both frontal and parietal regions within the dorsal stream, and suggest comparable changes within temporal and frontal regions of the ventral stream. As Johnson noted, changes within these neural regions are unlikely to be unitary events; rather, neural development likely reflects a more gradual “coming online” of the different components of the complex neural system that progressively come to support the range of behaviors involved in the visual search task (Johnson et al., 2001). Although studies of location coding in toddlers suggest that they can make use of fine-grained distance information when searching for hidden objects, the tendency to subdivide space (hierarchical coding) to facilitate remembering an object’s location does not emerge until approximately age 4 (Huttenlocher et al., 1994). Further, it is not until age 10 that children show reliable, adultlike spatial coding of fine-grained, multidimensional categorical information (Sandberg et al., 1996; Bullens et al., 2011). This is consistent with other studies demonstrating improvements in location memory from the toddler period through mid- to late childhood (Bell, 2002; Luciana et al., 2005; Orsini et al., 1987; Zald and Iacono, 1998; Hund and Foster, 2008; Negen and Nardini, 2015; Ribordy et al., 2013; Yang and Merrill, 2018). Increasing task demands by requiring that multiple spatial positions be recalled in a certain order extends the period of immature performance into early adolescence (Pagulayan et al., 2006; Gathercole et al., 2004; Luciana et al., 2005). Fine-tuning of location information encoded in memory is reported to extend through late adolescence (Luna et al., 2004). Finally, there is also evidence that the strength of children’s object categories, that is, their knowledge of what things are, can enhance memory for the object’s location (Hund and Plumert, 2003, 2005; Plumert et al., 2017). Visual hemifield tasks are used to examine hemispheric specialization of categorical and coordinate image generation (Kosslyn et al., 1995b). In these studies, participants decide whether probe marks (X) presented on a blank grid (categorical task) or bracketed square (coordinate task) appeared on a previously studied letter (Fig. 17.4). Target grids or brackets are presented to either the right (RVF) or left visual hemifield (LVF). For adults, the grid task elicits an LH categorical advantage, whereas the bracket task elicits an RH coordinate advantage. This profile of lateralization appears to emerge gradually during middle childhood. Specifically, 8-year-olds show an RH advantage for both categorical and coordinate tasks, but 10-year-olds begin to show the profile of lateralized differences characteristic of adults (Reese and Stiles, 2005). It is notable that the overall performance of 8-year-olds is considerably poorer than that of 10-year-olds, and the RH advantage is evident primarily on less challenging trials when the probe appears in a salient location marked by global spatial cues (so-called early probes). The finding of a developmentally early RH advantage on these location processing tasks is consistent with the finding of an RH-mediated, global processing advantage for children on a globalelocal processing task discussed earlier (Moses and Stiles, 2002). It appears that the more detailed LH-mediated processing required for both local level and coordinate spatial processing are later emerging aspects of neural specialization.

368 PART | II Cognitive development

FIGURE 17.4 Examples of the categorical (grids, above) and coordinate (brackets, below) stimuli (Kosslyn et al., 1995a). During the test phase, subjects were asked to read the lowercase letter beneath the stimulus and decide whether the corresponding block letter would cover the X mark in the grid (categorical task) or brackets (coordinate task) if it were present. Reproduced from Kosslyn, S.M., Maljkovic, V., Hamilton, S.E, Horwitz, G., Thompson, W.L., 1995. Two types of image generation: evidence for left and right hemisphere processes. Neuropsychologia 33 (11), 1485e1510, with permission.

17.1.3.2 Spatial attention A closely related line of investigation focuses on the neural systems associated with the ability to shift attention to different spatial locations. In contrast to work examining profiles of brain activity when subjects are required to directly perceive or remember the location of an object, spatial attention tasks investigate the brain systems engaged when attention must shift to a new location. There is considerable clinical and experimental evidence that the posterior parietal lobes play a crucial role in the ability to shift attention (Heilman and Valenstein, 1993; Hillyard and Anllo-Vento, 1998; Ivry and Robertson, 1998; Posner, 1980; Posner et al., 1984; Rafal and Robertson, 1995; Robertson, 1992). Posner’s influential model of the attention system involves an interconnected network of structures that modulate and control different aspects of attention (Posner, 1980; Posner and Petersen, 1990). The posterior parietal network plays an essential role in disengaging attention from one location and allowing a shift of attention to another location. In the standard task used to test covert shifts of attention (Posner and Cohen, 1980), subjects are required to fixate on a point located centrally between two identical, flanking squares. After a fixed period, a visual cue is presented either centrally (e.g., an arrow) or peripherally (e.g., one box brightens), and soon after a target appears briefly in one box. The subject responds as soon as the target is detected. The critical variable is the validity of the cue. On most trials (75% e80%), the cue is “valid” and the target appears in the cued box. On the remaining trials, the cue is “invalid,” and the target appears in the opposite box. If cueing serves to covertly shift attention, it should take less time to detect the target when the cue is valid than when it is invalid. One additional, well-established finding concerns response differences associated with the length of the interval between the valid cue and target, or stimulus onset asynchrony (SOA). With short SOAs ( true belief/direct perception), rather than relying on reliable channel placement on the scalp across participants. Using this approach, in two samples of 7-month-olds (total n ¼ 50) Hyde et al. observed preferential increases in oxygenated blood over the RTPJ, specifically during false belief sequences, and specifically when the target person returned to retrieve the object. That is, the RTPJ response in 7-month-olds was strikingly similar to the RTPJ response in adults! Thus, aggregating the results of these very different studies, cortical regions of the ToM network appear to be preferentially engaged by similar stimuli in adults, 3-year-old children, and 7-month-old infants. These results point to a very early developmental origin of a cortical network for ToM. Although no other studies have directly measured preferential functional responses during ToM in very young children or infants, other approaches hint at the same conclusion. Rather than measuring stimulus evoked responses in the ToM network, another strategy is to simply measure the correlation in responses, over time, between the different regions in the ToM network. Interregion correlation analyses can be used to characterize functional networks in the brain, because brain regions that operate within a functional network are more highly correlated with one another than with brain regions outside of that functional network (Blank et al., 2014; Hasson, 2004). During the Partly Cloudy movie, responses in ToM brain regions were significantly more correlated with one another than with other regions, even among 3-year-old children (Richardson et al., 2018). Brain regions that are correlated during cognitive tasks are also correlated at rest (i.e., in absence of stimuli; Cole et al., 2014; Greicius et al., 2003; Miall and Robertson, 2006). Unlike preferential responses to specific stimuli, correlations in functional network properties can be measured with fMRI in toddlers and infants, because they can be measured during sleep. Studies that take this approach often use dimensional reduction techniques, like independent component analysis (ICA), to identify brain regions with similar (correlated) timecourses of activity. Using ICA, Gao et al. (2009) found similar response timecourses in MPFC and PC regions in 1-year-old infants, and MPFC, PC, and TPJ responses were correlated by age 2 years (Gao et al., 2009). An alternative method to ICA is to measure the correlation between responses in a particular “seed” region and spheres that surround every other voxel (2   2  2 mm cube) in cortex. For example, using the PC as a seed region, Gao et al. (2013) found correlated responses in MPFC and TPJ in 1- and 2-year-olds (Gao et al., 2013). Results from resting state connectivity studies of 1-month-old neonates and term infants have provided some evidence for correlated activity in MPFC and PC (Smyser et al., 2010; though see Gao et al., 2009, 2013); evidence for connectivity between these regions in preterm infants is sparse (Cao et al., 2016; Doria et al., 2010; for review, see van den Heuvel and Thomason, 2016). Overall, these results converge to support the hypothesis that the ToM network is functionally correlated and distinct from other cortical networks very early in development. Evidence from studies of macaques provides an additional indirect source of evidence about the early origins of social brain regions. Sliwa and Freiwald (2017) used fMRI to measure neural responses in the brains of macaques while they watched short videos of two monkeys interacting (grooming, fighting, chasing, and mounting). Control condition videos showed a single monkey interacting with objects (food, toys), two objects interacting (bouncing against each other, getting entangled), or a single inactive monkey (lying down or sitting quietly) or inactive object (hanging). Functional regions in the MPFC and superior temporal sulcus (STS), a region adjacent to temporoparietal junction and involved in social perception in humans, were selectively recruited when macaques watched clips that involved social interactions between two monkeys. These regions responded to social interaction clips and deactivated to all other kinds of clips. Regions that responded preferentially to social interactions were functionally correlated, and more correlated with one another than with other regions (Sliwa and Freiwald, 2017). Given the similar location and functional response profile of these regions, the STS and MPFC in macaques may very well be functional homologs of the ToM network in humans. Future comparative work is necessary to better understand the similarities and differences of the functional roles of these homologous regions across species, to characterize the ontogenetic development of these regions in nonhuman primates (Rosati et al., 2014), and to link neural measures of social brain regions to the vast behavioral literature on social understanding and ToM in nonhuman primates (Drayton and Santos, 2016; Martin and Santos, 2016). In all, by the most basic and general definition, the ToM network appears to emerge ontogenetically and possibly phylogenetically early. TPJ, MPFC, and PC are preferentially recruited by stimuli that evoke consideration of others’ thoughts and feelings. These regions are functionally correlated with one another, and thus constitute a distinct functional network.

21.3 Neural correlates of ongoing theory of mind development in childhood The recent imaging studies reviewed above have begun to fill in the gap between behavioral studies on ToM development in early childhood and infancy, and fMRI studies of ToM brain regions in older children and adults. The results of these studies are generally consistent with the behavioral literature on ToM development, which increasingly places false belief

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

471

task performance in the context of gradual, continued ToM development throughout childhood. The ability to represent others’ beliefs is preceded by the ability to represent others’ goals and knowledge access (Wellman and Liu, 2004), and followed by the ability to reason about nonliteral speech (Filippova and Astington, 2008; Peterson and Wellman, 2018; Peterson et al., 2012; Wellman et al., 2011) and evaluate moral blameworthiness in accidents (Cushman et al., 2013). A fully developed Theory of Mind incorporates each of these abilities, and as such, Theory of Mind development encompasses more than the achievement of one of them. Analogously, ToM brain regions show early signatures of functional organization prior to explicit false belief task success, but should also continue to undergo developmental change throughout childhood. How do the functions and computations of ToM brain regions change to support ongoing ToM development? Below, we review three kinds of neural measures that capture developmental change in ToM brain regions in childhood and correlate with behavioral improvements in Theory of Mind, cross-sectionally: response selectivity, response reliability, and functional network segregation.

21.3.1 Response selectivity: fine-tuning preferential responses Given the highly selective functional responses observed in ToM brain regions among adults, initial fMRI studies on ToM in children were designed specifically to test hypotheses about the development of selective responses. In other domains, the development of increasingly selective functional responses supports cognitive change in childhood. For example, the magnitude of selective responses in cortical regions specialized for faces (fusiform face area, FFA) and places (parahippocampal place area, PPA) is larger in adults, relative to children (despite similar whole-brain volume), and is correlated with behavioral recognition memory of faces and scenes (Golarai et al., 2007; Gomez et al., 2017; Grill-Spector et al., 2008). These cortical regions respond less to nonpreferred categories with age, and the reduced response to nonpreferred categories corresponds to improvements on category-relevant behavioral recognition tasks (Cantlon et al., 2010). Similarly, the volume of symbol selective cortex (visual word form area (VWFA); McCandliss et al., 2003) increases rapidly in children upon learning to read (Dehaene-Lambertz et al., 2018), and development of this region involves reduced responses to nonpreferred visual categories (Dehaene et al., 2010). Increasingly selective cortical responses could reflect or support the specialization of a region for the particular computational demands of a particular cognitive domain. Thus, in addition to measuring neural responses to mental states and nonmentalistic control conditions, two of the early studies of ToM in children (5e12 years old) measured responses to nonmentalistic social stimuli, which included information about the characters’ stable relationships (“Sarah and Lori played together on the school soccer team .”; “Jenny and Samantha were twins .”), physical appearances (“Old Mr. McFeegle is a gray wrinkled old farmer.”), and actions and abilities (“Once upon a time, a girl and her little brother went out to pick flowers .”; “. She was so good at playing the flute that when she played everyone immediately started dancing”). By measuring responses to social, but nonmentalistic, stimuli, these two studies found evidence for developmental change in ToM responses: responses in ToM brain regions, and in particular in RTPJ, became more selective for mental states between 5 and 12 years of age (Gweon et al., 2012; Saxe et al., 2009). Responses to mental state stories remained high throughout childhood, but responses to non-mentalistic social control stories decreased with age. Additionally, selectivity for mental states in the RTPJ correlated with behavioral performance on a ToM task that involved reasoning about similar/diverse desires, true/false beliefs, moral blame, and mistaken referents (Gweon et al., 2012). Richardson et al. (2018) provides converging evidence for developmental change in response selectivity throughout childhood. In addition to conducting interregion correlation analyses (reviewed above), Richardson et al. conducted reverse correlation analyses on ToM responses in adults (n ¼ 33) and in 3-year-old children (n ¼ 17). By identifying timepoints that reliably evoke positive responses in a brain region or functional network across participants, reverse correlation analyses offer a data-driven way to identify the kind of stimulus that drives responses in that region or network. In adults, reverse correlation analyses have offered satisfying replications of functional selectivity profiles initially discovered with well-controlled, experimenter-driven methods: the fusiform face area is driven by faces, and the parahippocampal place area is driven by scenes (Hasson, 2004). Similarly, Richardson et al. found that the pain network (a set of brain regions recruited to reason about physical/bodily pain) is driven by scenes involving salient physical sensations, and the ToM network is driven by scenes that evoke mental state reasoning among adults. Scenes that evoked responses in the pain network were nonoverlapping with those that evoked responses in the ToM network. A majority of the scenes identified as pain events (9/12) and all of the ToM scenes (7/7) were identified in a replication analysis of a second (independent) sample of adults (Richardson et al., 2018). As a data-driven approach, reverse correlation analyses aren’t constrained by experimenter-constructed hypotheses, and therefore could discover preferential responses for kinds of stimuli that are unexpected, especially in participants with different or developing cognitive abilities. Do the same ToM events evoke reliably high responses in ToM brain regions

472 PART | II Cognitive development

among 3-year-old children? As in adults, scenes that evoked responses in the pain network were nonoverlapping with those that evoked responses in the ToM network in 3-year-old children. However, only 8 timepoints (of 168) reliably evoked positive responses in the ToM network across 3-year-old participants. Of these 8 timepoints, half corresponded to an adult ToM event; the other half corresponded to an adult pain event. In the entire cross-sectional sample of children (n ¼ 122), ToM network responses to ToM events increased and ToM network responses to pain events decreased between 3 and 12 years of age. Thus, while ToM brain regions are functionally distinct from other networks and show some evidence for early preferential responses (i.e., by being significantly correlated with adult responses), responses become more selective for mental state content throughout childhood. Together, these studies provide evidence for slow, gradual development of increasingly selective brain regions for ToM. However, there is a small discrepancy in these results. In the initial fMRI studies (Gweon et al., 2012; Saxe et al., 2009), developmental change in response selectivity manifested as stable, high responses to mental state stimuli paired with decreasing responses to nonmentalistic social stimuli. In Richardson et al. (2018) responses to nonmentalistic pain events decreased with age, and responses to mental state events increased. We suggest that this discrepancy may be explained by another aspect of the response in ToM brain regions that changes with age in childhooddspontaneous responses in ToM brain regions to naturalistic stimuli become increasingly reliable and stereotyped. In Richardson et al. (2018) developmental increases in ToM responses to mental state events could reflect reduced variability in the timing and duration of the responses to these events. Consistent with this idea, functional responses in the ToM network become more similar to the average adult response with age. We review additional evidence for developmental change in the reliability of spontaneous ToM responses below.

21.3.2 Reliable spontaneous (uninstructed) responses to movies Moments or situations that require or would benefit from ToM reasoning aren’t always marked. Individuals may have different thresholds for spontaneously engaging in mental state reasoning, and this could contribute to individual differences in ToM. For example, children who spontaneously produce mentalistic explanations for behaviors subsequently achieve consistent successful performance on false belief tasks (Amsterlaw and Wellman, 2006). These individual differences could reflect an individual’s propensity to shift their focus to mental states (reliably and quickly), anticipate or predict future mental states, and/or interpret current events in the context of past mental states. Univariate measures of response magnitude or selectivity are limited in their sensitivity to this kind of individual difference, because they average responses across time and trials to calculate a single binary contrast. By contrast, timecourse analyses of data collected during naturalistic movie paradigms can capture individual differences in responses to a continuous, dynamic narrative that unfolds over time (Dubois and Adolphs, 2016b). Similarity of response timecourses across individual participants can be quantified via intersubject correlation (ISC) analyses (Hasson, 2004). Among adults, ISC analyses find highly similar responses in ToM brain regions during movieviewing or narrative-listening (Hasson, 2004; Jääskeläinen et al., 2008; Lerner et al., 2011; Nummenmaa et al., 2012; Wilson et al., 2007), as well as during free recall of a movie (Chen et al., 2016). Identifying response profiles that are generally common across individuals potentially enables discovering meaningful differences in individuals whose response timecourses vary. Indeed, adults with more correlated responses in ToM brain regions during movie-viewing have similar interpretations of the movie (Nguyen et al., 2017), and similar memories about the movie (Chen et al., 2016). Thus, ISC analyses are a promising approach for capturing meaningful neural individual differences in ToM responses to naturalistic viewing or listening paradigms (Dubois and Adolphs, 2016b). Can naturalistic movie-viewing paradigms be used to capture developmental change in ToM? One way to measure individual differences in ToM responses to naturalistic paradigms across children is to compare each child’s timecourse to an adult template time course. This measure is sometimes referred to as “functional maturity,” as it quantifies the similarity of the functional profiles between children and mature adults. By calculating this measure over responses to naturalistic stimuli, this measure likely reflects developmental change in both response selectivity to functionally preferred content and developmental change in spontaneous, unprompted cognitive computations. In other cognitive domains, functional maturity is correlated with cognitive abilities. For example, Cantlon and Li (2013) calculated functional maturity in 4- to 6-year-old children by comparing their response timecourses to an episode of Sesame Street to an average adult timecourse. Among children, functional maturity of the intraparietal sulcus, a cortical region that responds preferentially to process numbers, correlated with behavioral performance on a math assessment, and functional maturity of Broca’s area, a cortical region involved in language processing, correlated with children’s verbal abilities (Cantlon and Li, 2013). While some aspects of the developmental increase in functional maturity could be very general, i.e., reflective of more consistent and adultlike eye movements during naturalistic viewing tasks (Franchak et al., 2016; Frank et al., 2011), correlations between

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

473

functional maturity in functionally selective cortical regions and performance on relevant behavioral tasks suggests that this measure captures domain-specific development as well. Two studies have measured functional maturity of ToM brain regions in children. In Richardson et al. (2018), reviewed above, functional maturity of the ToM network increased with age: responses in these brain regions to Partly Cloudy became more adultlike between ages 3 and 12 years (Richardson et al., 2018). While functional maturity (of the entire ToM network) correlated with performance on a behavioral ToM task, this correlation did not remain significant when additionally controlled for age. In the second study, Moraczewski et al. (2018) measured response timecourses in 4- and 6-year-old children, as well as adults, while they watched a 6-min clip from the movie Toy Story (Guggenheim et al., 1995). Responses in ToM brain regions to Toy Story were more similar across individual adults, than across individual children (that is, the adult response profile was more stereotyped or reliable, across subjects). Additionally, responses in TPJ to Toy Story became more similar to that of adults between 4 and 6 years of age (Moraczewski et al., 2018). While behavioral ToM was not measured in this sample, Moraczewski et al. (2018) confirmed their results in subsequent analyses that used response timecourses from the frontal eye field as a nuisance regressor, in order to account for individual differences in attention. Nonetheless, additional work is necessary to better understand the relationship between functional maturity and individual behavioral differences in ToM. Like traditional fMRI experiments, naturalistic movie-viewing paradigms allow for measurements of neural responses to particular events. However, because events are embedded within the movie, naturalistic movie-viewing paradigms are additionally sensitive to individual differences in anticipation of future mental states, and (re-)interpretation of current events in the context of past mental states. Spontaneous activity in response to specific events whose meanings depend on integration over long periods of time may be particularly relevant to individual differences in ToM. In Richardson et al. (2018), multiple measures of the response in ToM brain regions correlated with behavioral ToM score (proportion correct on questions that involved reasoning about same/different desires, true/false beliefs, moral blame, sarcasm, lies, secondorder false beliefs, and mistaken referents). However, only the magnitude of response to a particular event during Partly Cloudy (Reher and Sohn, 2009) correlated with ToM score when additionally controlling for age. As reviewed above, similar response timecourses in adults reflect similar narrative interpretations of movies or stories. Thus, by looking at the content of the event that evoked ToM responses differentially by ToM behavioral score among children, we can hypothesize about the kinds of differences in narrative interpretations of that event that might be relevant to ToM development. The event in question involved one character (Peck, the stork) offering an alternative explanation for his previous behavior (flying away) to another character (Gus, the cloud), who had not only made an incorrect inference about the cause of the behavior (Peck was abandoning him in favor of working with a different cloud), but who had also been very upset about this inferred cause. Based on the content of this event, the correlation between responses during this event and behavioral ToM may reflect participant’s (re-)consideration of mental states in absence of explicit cues to do so. This kind of brain-behavior correlation has since been replicated in confirmatory analyses of an independent sample of 5- to 12-yearold children (n ¼ 186; Richardson, 2019). Thus, more reliable response timecourses in ToM brain regions with age may reflect developmental improvements in spontaneous, flexible engagement of mental state reasoning. Spontaneous, flexible mental state reasoning may be directly tied to the development of increasingly refined ToM concepts, which could enable children to more easily recognize the relevance of particular ToM concepts across different contexts and without cues (Amsterlaw and Wellman, 2006; also see Moreira et al., 2018 for a relevant study in adolescents).

21.3.3 Integration and separation of functional networks A third kind of neural change that may support ongoing ToM development is ongoing separation of functional networks. Functional network separation can be measured via interregion correlation analyses (reviewed above), which compare response timecourses within individuals across different brain regions (Blank et al., 2014; Hasson, 2004). Brain regions with highly correlated activity during movie-viewing are driven and deactivated by similar moments in the movie, i.e., their functional response profile is similar, and they support a specific set of cognitive functions. Naturalistic movie-viewing and story-listening experiments have been used to characterize functional networks for processing faces and scenes (Hasson, 2004), and for dissociating brain networks for processing language versus domain-general cognitively difficult tasks (the “multiple demand” network; Blank et al., 2014). These kinds of experiments have begun to characterize the functional network properties of ToM brain regions. Among adults, ToM brain regions are more correlated with one another than with brain regions involved in processing language (Paunov et al., 2017) and brain regions that process physical (bodily) pain (Richardson et al., 2018). Interestingly, ToM and pain brain regions are not just uncorrelated, but they are actually anticorrelated, among adults. As reviewed above, ToM

474 PART | II Cognitive development

brain regions are functionally dissociated from the pain network by age 3 years. However, the extent of this dissociation increases with age throughout childhooddToM brain regions become more correlated with one another (see also Xiao et al., 2019), and increasingly negatively correlated with brain regions in the pain network, with age (Richardson et al., 2018). These developmental trends were replicated in an independent sample of children and adolescents (Richardson, 2019). Is the developmental separation of functional networks a plausible neural correlate of improvements in ToM behavior? Increases in within-network correlations could reflect faster, more efficient, and less noisy communication between ToM brain regions. Similarly, decreases in across-network correlations could reflect less interference, or more functionally precise responses, across networks. Richardson et al. (2018) collected behavioral ToM data and tested for neural measures that (1) differed between young children (n ¼ 65, 3e5 years old) who passed and failed explicit false belief tasks, and (2) increased with overall ToM reasoning (including questions about concepts that develop prior to and after passing false belief tasks) in the full sample of children (n ¼ 122, 3e12 years old). Interestingly, ToM brain regions were significantly more correlated with one another in children who passed explicit false belief tasks, relative to those who failed, controlling for age. However, this measure also increased with age in the full sample of children. In the full sample of children, withinToM network correlations correlated with overall ToM performance, but this correlation did not survive controlling for age. Thus, one possibility is that within-ToM network correlations are particularly important for or reflective of ToM behavior in young children, but more correlated with age across development. These results are consistent with two studies that used two different neural measures of within-network integration. First, Grosse Wiesmann et al. (2017) used diffusion tensor imaging (DTI) to measure white matter connectionsdrather than functional responsesdin the brain, to study ToM development in young children (Grosse Wiesmann et al., 2017). Grosse Wiesmann et al. found increased connectivity in tracts surrounding ToM brain regions (including RTPJ, PC, and VMPFC) in 3- and 4-year-old children who passed false belief tasks, relative to those who failed. Second, Xiao et al. (2019) found that resting state connectivity between RTPJ and other ToM ROIs (LTPJ, PC) correlated with scores on a parent-questionnaire assessing ToM abilities in 4- to 8-year-old children. While both studies found neural differences by ToM behavior that remained significant while controlling for age, future work with a wider and older age range is necessary to determine if these neural measures are especially important for early ToM development, or if ongoing developmental change in these measures supports subsequent ToM advances. In any case, it is encouraging to see consistency in results across three types of neural measures (interregion correlations measured during movie-viewing and at rest, and physical connectivity measures). In sum, imaging studies of children provide evidence for ongoing development in ToM brain regions throughout childhood. While ToM brain regions show early preferential responses and early segregation from other functional networks, the responses in these brain regions become increasingly selective and fine-tuned for processing mental states. Responses in these regions also appear to converge with typical adult responses, and diverge from responses in other functional networks, with age. Each of these neural measures appears to relate to performance on ToM tasks in childhood. Of course, these measures are not independent from one anotherdthey each capture a slightly different aspect of the same functional response. Future work is necessary to understand the relationships between each measure. For example, what is the causal relationship between the development of selective responses, and the development of segregated functional networks? Does intrinsic or functional connectivity restrict, predict, or guide physical connectivity, or vice versa? Understanding how these measures relate to one another in development may clarify their relative contributions to behavioral ToM development.

21.4 Future directions: open questions and challenges 21.4.1 Neural correlates of structural changes in theory of mind As children get older, they not only acquire and refine ToM concepts, but they also achieve a more sophisticated understanding of the relationships between different mental states (beliefs, desires, emotions), and the relationship between mental states and their causes and consequences. While children show relatively early understanding of probable causes of different emotions (surprise, sadness, awe) (Lagattuta and Wellman, 2001; Skerry and Spelke, 2014; Wu et al., 2017), there is also evidence for ongoing change in the conceptual organization of emotions in childhood (Gao and Maurer, 2010; Weisman et al., 2017; Widen and Russell, 2003, 2008). One challenge for future neuroimaging studies of ToM development will be to discover neural correlates of this kind of conceptual change. Traditional measures of selectivitydwhich measure the relative response of a given brain region to different conceptual categories (e.g., mental states vs. bodily sensations)dmay not be sensitive to structural changes within the category of mental states. That is, because mental state stimuli evoke high responses in ToM brain regions, selectivity may not be a sensitive neural measure of the development conceptual distinctions between different mental states within the broader category. Indeed, among adults, different kinds of mental states (e.g., expected vs. unexpected) evoke the same magnitude of response in ToM brain regions (Young et al., 2010b).

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

475

Multivariate pattern analyses may provide a method for capturing developmental change in the structure of mental state representations. In adults, multivariate analyses have revealed information about the structure of representations within ToM brain regions by determining the features of mental state stimuli that drive response similarity. For example, Tamir et al. (2016) used multivariate analyses to suggest that almost half of the variation in responses in ToM brain regions to mental states can be accounted for collectively by three abstract dimensions: rationality, social impact, and valence (Tamir et al., 2016; see also Saxe, 2018). To discover these dimensions, Tamir and colleagues first used principal component analysis (PCA) on behavioral ratings of 166 mental states (e.g., “embarrassment,” “ecstasy,” “planning,” “dominance,” “friendliness,” “imagination,” “self-pity,” “satisfaction,” “affection,” “disgust,” “disarray”) on 16 plausible dimensions derived from the literature, collected in a large sample of adults (n ¼ 1205). They found that behavioral rating similarity across mental states was driven by four orthogonal dimensions: rationality, social impact, human mind, and valence. Next, they used fMRI to measure responses in an independent sample of adults (n ¼ 20) who considered and rated the extent to which a given scenario (e.g., “finding $5 on the sidewalk”, n ¼ 16 unique scenarios total) evoked a particular mental state (see examples above; n ¼ 60 unique mental states total). They used a “searchlight analysis” to examine the relative response surrounding every voxel of cortex to each mental state. The extent to which mental states were similar in rationality, social impact, and valence ratings predicted the extent to which mental states evoked similar neural responses in social brain regions. Thus, the same ToM brain regions implicated by univariate analyses contain multidimensional representations of mental states, and rationality, social impact, and valence may be particularly important dimensions by which responses in these regions are organized. Multivariate analyses can additionally provide evidence about the different functional roles of distinct social brain regions. Koster-Hale et al. (2017) used fMRI to measure neural responses in adults who listened to short stories that involved mental state reasoning. Across stories, characters formed beliefs that differed in justification (i.e., based on strong vs. weak evidence), source modality (i.e., based on hearing vs. seeing evidence), and valence (positive vs. negative). Koster-Hale and colleagues found that the epistemic featuresd that is, features that described the source and evidence for the beliefd organized responses in temporoparietal junction, whereas the valence of the resulting emotion (caused by the belief) organized responses in medial prefrontal cortex. Thus, multivariate methods can be used to discover the divisions of labor among ToM brain regions, in addition to the features that organize neural responses, and the underlying representations, within each region. Developmental change in the structure of ToM representations could be reflected in developmental change in the features that drive response similarity in ToM brain regions. Additionally, developmental change in the structure of ToM representations could result in divisions of labor among ToM brain regions, by which certain brain regions are increasingly fine-tuned to particular aspects of mental states. Studies that use multivariate analyses to characterize ToM responses in child populations are an important next step in the developmental cognitive neuroscience of ToM.

21.4.2 Discovering reliable neural markers of individual differences in theory of mind Perhaps the greatest challenge faced by developmental cognitive neuroscientists studying ToM is to discover robust neural markers that reliably measure individual differences in ToM (Dubois and Adolphs, 2016a). To date, there is no such evidence for a neural marker that reliably predicts individual differences in ToM longitudinally, despite multiple longitudinal behavioral studies suggesting that there are reliable individual differences in ToM behavior in development (Peterson and Wellman, 2018; Wellman et al., 2011; Wellman et al., 2008; Yamaguchi et al., 2009). Reliable neural individual difference measures would be useful for testing hypotheses about environmental and genetic factors that promote or hinder development of ToM brain regions, and are critical for designing and testing the effectiveness of clinical interventions that aim to improve social cognitive abilities. While many studies find group differences between older children and adolescents who have been diagnosed with autism or schizophrenia and neurotypical individuals, none of these neural differences are sensitive or early markers that could be used in a clinical or treatment setting (studies of clinical populations are not reviewed in this chapter; see Happé and Conway, 2016; Happé and Frith, 2014). Longitudinal evidence is necessary for testing candidate neural markers: such a marker should show reliable neural individual differences over time that relate to ToM behavior, and early measurement of the marker should predict subsequent ToM development and ability. Neuroimaging evidence from individuals with different developmental experiences is also useful for learning about drivers of individual differences in ToM, and about the impact of experience on social and brain development more broadly.

21.4.3 The role of developmental experience: language There is significant need for more research on the developmental factors that drive the specialization of brain regions for ToM. One open question concerns the role of experience on the maintenance and development of selective responses. By one extreme, evolutionary pressures and genetic makeup could drive ToM brain regions to develop increasingly

476 PART | II Cognitive development

functionally selective responses throughout childhood, regardless of developmental experience. On the other extreme, the development of functionally selective responses could be quite fragile or flexible: lack of a necessary input or experience during a critical or sensitive time could preclude the development of brain regions selective for ToM processes. By this account, individual differences in the development of specialized responses would be a direct result of developmental experience. What kinds of developmental experiences might be important for ToM development? Once again, behavioral studies of ToM development in children provide useful hypotheses. For example, language abilities are highly correlated with ToM performance in childhood (Astington, 2006; Astington and Jenkins, 1999; Milligan et al., 2007). One intriguing hypothesis is that this correlation reflects a facilitative role of linguistic experience for ToM development. That is, language might not only enable children to express their understanding of ToM conceptsdlinguistic experience might directly drive the development of ToM concepts per se. If linguistic experience is important for ToM-specific development (and not just for task performance), then ToM development should be delayed as a function of delayed access to language. Behavioral studies have provided evidence in favor of this hypothesis by measuring ToM performance in children who are d/Deaf.2 Deaf children born to hearing parents (DoH) often experience delayed access to sign language, whereas d/Deaf children born to signing parents receive access to sign language at birth. DoH children show delayed ToM development corresponding to the length of delay prior to exposure to language, even on nonlinguistic ToM tasks (Meristo et al., 2012; Peterson and Siegal, 1999; Peterson and Wellman, 2018; Schick and Hoffmeister, 2001; Schick et al., 2007; Woolfe et al., 2002). Does delayed access to language result in delayed development of brain regions specialized for ToM? A recent study by Richardson et al. (2019) used fMRI in conjunction with behavioral measures to ask this question. Native (n ¼ 21) and delayed (n ¼ 12) signing children (4e12 years old) completed linguistic and nonlinguistic behavioral tasks and fMRI experiments. Delayed signers experienced delays ranging from 0.25 to 7 years in length, but all children were proficient in American Sign Language (ASL) at the time of the study. Consistent with previous behavioral studies, children with delayed access to language showed corresponding delays on behavioral ToM tasks. These delays were most apparent in linguistic questions involving “advanced” ToM reasoning, like considering intended meaning in nonliteral speech (sarcasm), lies, and second-order false beliefs, and assignment of moral blame in accidents, as measured by the linguistic task. However, delayed signing children showed typical ToM performance on moral blame questions in the nonlinguistic behavioral task, and all children showed higher ToM performance on questions about false beliefs when tested linguistically. Behavioral differences across linguistic and nonlinguistic task formats suggest that language plays a role in the expression of ToM understanding. Does language additionally play a role in development of ToM per se? The linguistic fMRI experiment provided a way to test this question. This experiment was modeled after the paradigm used in initial studies of ToM brain regions in 5- to 12-year-old children, and therefore was designed to measure the selectivity of ToM responses for mental states relative to nonmentalistic social information (Gweon et al., 2012; Saxe et al., 2009). As reviewed above, developmental specialization of ToM brain regions typically occurs via suppression of responses to nonmentalistic social content. In delayed signers, the selectivity of the response in RTPJ was delayed as a function of the length of delay prior to language exposure: the response to the nonmentalistic social condition was higher in children who experienced longer language delays. This effect was present despite proficiency in sign language at the time of the study, and no differences in the responses of brain regions involved in processing language. Thus, developmental experience, and in particular, early and extensive exposure to rich linguistic input, is important for refining the functional response of RTPJ in childhood. Importantly, typical response profiles were observed in delayed signing adults (n ¼ 16), which suggests that delayed access to sign language delays, but does not permanently disrupt, the development of selective responses in RTPJ. Interestingly, children with delayed access to language did not show differences in ToM responses during the nonlinguistic fMRI experiment, which involved watching a silent version of Disney Pixar’s Partly Cloudy (Reher and Sohn, 2009). As described above, this movie stimulus has previously been used to measure responses in ToM brain regions in children ages 3e12 years old (Richardson et al., 2018). As in the original study, ToM brain regions responded preferentially to mental state content and deactivated during moments depicting physical pain during this experimental context, regardless of age of exposure to language. Given that the functional dissociation between ToM brain regions and the pain network is apparent by age 3 years (Richardson et al., 2018), one possibility is that preferential responses in RTPJ for minds relative to bodies develop early and are less dependent on linguistic experience. This study suggests that language is a key aspect of developmental experience that impacts ToM development. Is early linguistic input sufficient for typical development of brain regions specialized for ToM reasoning? Studies of children who 2. Note that in the literature, the capitalized form of the word “Deaf” has been used to refer to the cultural and linguistic minority group, and the lower-case “deaf” to refer to the audiological status.

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

477

are congenitally blind could help to address this question. Blind children have typical linguistic input, but reduced access to information about minds that is conveyed through vision. Vision provides a way to perceive consequences of mental states (e.g., if a person reaches for a teddy bear, she prefers it to the ball; if a person expresses sadness upon seeing a puppy, she’s remembering when the puppy stole her snack), and facilitates early interactions and social bonding (e.g., through eye contact, joint attention, and attention to facial expressions). While previous neuroimaging research with adults suggests that by adulthood, blindness has no effect on the functional responses in theory of mind brain regions (Bedny et al., 2009; Koster-Hale et al., 2014), behavioral studies find some evidence for delayed ToM development in children who are blind (Brambring and Asbrock, 2010; Brown et al., 1997; Minter et al., 1998; Peterson et al., 2000). Future work investigating the development of functionally selective responses in children who are congenitally blind could clarify whether linguistic input is particularly important for refining functional responses, and if visual input during development plays a similar role in refining the functional responses in RTPJ.

21.4.4 The role of developmental experience: culture A second way to study the impact of developmental experience on ToM development is to study children growing up in different cultures. Cultures provide social norms and instill individuals with a particular set of values. In adults, cultural differences appear to influence ToM reasoning: large-scale industrialized societies (e.g., Los Angeles, USA) place greater weight on intentions when assigning blame or punishment than small-scale traditional societies (e.g., Yasawa Island, Fiji; Barrett et al., 2016). Does culture also impact the development of ToM in children? Initial studies of the influence of culture on ToM development compared ToM performance in children from WEIRD (Western, Educated, Industrialized, Rich, Democratic; Henrich, 2011) and non-WEIRD societies. Studies of non-WEIRD societies are extremely difficult and expensive to conduct, and can be risky in that the presence of the research itself could have unintended consequences for these societies. As such, there have been relatively few behavioral studies and, to our knowledge, no fMRI studies of ToM in individuals living in smallscale societies. Possibly because of the sparse number of studies, those that do exist focus primarily on false belief task performance: a reasonable initial step for understanding ToM reasoning in a different culture. Interestingly, these studies suggest that children from non-WEIRD societies, e.g., children of the Baka (pygmies living in southeast Cameroon rainforests; Avis and Harris, 1991), children living in an impoverished mountain village of Peru (Callaghan et al., 2005), and children from a Polynesian settlement in Samoa pass false belief tasks by age 5 years (Callaghan et al., 2005), similar to children in WEIRD societies (Wellman et al., 2001). Of course, false belief tasks may not be the most sensitive test of the impact of culture on ToM development. Additionally, the differences between the societies examined may not be relevant for false belief task performance. Subsequent studies of the role of culture on ToM development suggest that by emphasizing or deemphasizing particular ToM concepts, cultures can influence the trajectory of ToM development; that is, the order in which children master particular concepts in ToM. For example, individualistic communities, like the United States of America or Australia, tend to value personal expression and independence, whereas collectivist or interdependent communities, like China, Iran, or Turkey, place more emphasis on shared knowledge. These different social norms and values highlight different ToM concepts for young children. Accordingly, American and Australian children understand that individuals can hold different beliefs earlier than Chinese, Iranian, and Turkish children, whereas Chinese, Iranian, and Turkish children understand the relationship between seeing and knowing earlier than Americans and Australians (Selcuk et al., 2018; Shahaeian et al., 2011; Wellman et al., 2006). The social norms of a culture can also shape how children seek out social information. For example, by 7 months of age, infants from Western (UK) versus Eastern (Japan) cultures fixate on different aspects of the face (mouth vs. eyes) to discriminate emotional expression (happy vs. fearful). These perceptual strategies map onto cultural differences in information value of mouths and eyes for communicating emotion (Geangu et al., 2016). Thus, culture shapes social input as well as the trajectory of ToM development in childhood. Does culture also influence the development of ToM brain regions, and if so, how? The role of culture for the development of ToM brain regions may differ from the role of language. Multiple studies suggest that a person’s native language does not affect the development of ToM brain regions. Individuals who are native English (Saxe and Kanwisher, 2003), Swedish (Happé et al., 1996), French (Brunet et al., 2000, 2003), German (Sommer et al., 2007; Vogeley et al., 2001), Japanese (Kobayashi et al., 2007), and American Sign Language (Richardson et al., 2019) speakers/signers recruit the same ToM brain regions during social reasoning tasks. Additionally, the semantic content, rather than the linguistic format, of stories drives response similarity in these brain regions, across adults (Honey et al., 2012). However, as reviewed above, children with delayed access to language show corresponding delays in the development of selective responses in the RTPJ (Richardson et al., 2019). Thus, while the differences in phonetic distinctions, grammatical structures, and modality that exist across languages do not appear to alter the location or

478 PART | II Cognitive development

functional profiles of brain regions recruited for ToM, early exposure to language is important for the development of ToM brain regions. The provision of some set of norms and values that guide and help to understand social behaviors may similarly be important for the development of ToM brain regions. However, given the behavioral literature reviewed above, the particular norms and values that a culture highlights may also alter the developmental trajectory of ToM brain regions, or the particular features of mental states that drive responses in ToM brain regions. What kinds of neural measures might be sensitive to cultural differences in ToM? Traditional univariate measures could reflect cultural differences if an experiment evokes a particular aspect of mental state reasoning that differs across the cultures studied. Additionally, naturalistic movie-viewing or narrative-listening paradigms could plausibly capture differences in narrative interpretation or perspective driven by cultural background. Among adults, similar response timecourses in ToM brain regions between speakers and listeners predict successful communication (that is, agreement about what was communicated; Stephens et al., 2010). Thus, differences in neural responses in ToM brain regions could indicate differences in culturally shaped perspectives. Multivoxel pattern analyses might also be useful for measuring differences in ToM reasoning driven by culture, because they could reveal differences in the features or dimensions of mental states that drive responses in ToM brain regions (e.g., degree of intent). Given evidence that culture shapes the trajectory of ToM development in children, studying children may be necessary to understand cultural differences in the trajectory of development of ToM brain regions (Selcuk et al., 2018; Shahaeian et al., 2011; Wellman et al., 2006). There is strikingly little evidence about the development of ToM brain regions across cultures. Lloyd-Fox et al. (2017) used fNIRS to measure cortical responses to social stimuli (e.g., movies of adults playing Peek-A-Boo), compared to nonsocial control stimuli (e.g., images of cars and helicopters) in infants (0e24 months old) growing up in rural Gambia. Preferential responses in the right posterior temporal lobe (plausibly STS/TPJ) to social stimuli in these infants did not differ from those of children growing up in the UK (Lloyd-Fox et al., 2017). In addition to showing some preservation of preferential responses to social stimuli in the temporal lobe, this experiment demonstrates the feasibility of using fNIRS, which is less expensive and easier to transport than fMRI, to measure neural responses in children living in many diverse areas of the world. Future work is necessary to understand the degree of specialization of ToM brain regions in infants across varying cultures, and to characterize the impact of culture on the developmental trajectory of ToM brain regions in childhood. Ultimately, studying the impact of culture on the development of ToM brain regions could refine theories about functional origins of these brain regions. For example, WEIRD adults recruit ToM brain regions when making moral judgmentsde.g., when deciding how much blame to assign or how severely to punish an individual who has caused harm. In particular, the consideration of an individual’s intentionsdwhether she meant to cause harmddrives responses in RTPJ (Koster-Hale et al., 2013; Young and Saxe, 2009; Young et al., 2010a,c). But, if WEIRD people tend to focus on individual beliefs much more than people growing up with other cultural norms, then it seems likely that the root cognitive function of RTPJ (if there is one) is more general than thinking about others’ thoughts: i.e., the RTPJ could plausibly be specialized for considering any latent variables that people use to explain and evaluate others’ actions. These two hypotheses about the root function of the RTPJ can’t be teased apart in WEIRD adults, because individual beliefs are the latent variables WEIRD adults use to explain and evaluate others’ actions. However, studies of individuals who prioritize other information to understand actions (e.g., outcomes, rather than intent) could inform the extent to which ToM brain regions are specialized for beliefs per se, versus the function they serve for understanding others’ actions.

21.4.5 The role of family on ToM: shared environment and shared genes Parents and siblings play a large role in shaping the smaller-scale culture of a child’s home life. Biological parents and siblings share their genes, in addition to shaping a child’s environment. As a result, it is sometimes difficult to tease apart the relative contributions of shared environment and genetic heritability when characterizing the role of family on social cognitive development (Avinun and Knafo, 2014). Indeed, some genes, like the oxytocin receptor gene, may impact ToM development in early childhood both through heritability and through their direct impact on social interactions (Wade et al., 2015; Wu and Su, 2015). One way to isolate the impact of heritable genes, relative to shared or nonshared environmental factors, on ToM development is to study monozygotic and dizygotic twin pairs. While both kinds of twin pairs have similar cultural, home, and school environments across twins, monozygotic twins are genetically identical, whereas dizygotic twins share on average 50% of their genes. A handful of studies have now taken this approach in order to determine the relative influences of genes and environment on ToM. However, the methods and results of these studies are quite variable. In a study of 119 three-year-old twin pairs, Hughes and Cutting (1999) found that genetic influences accounted for 60% of the variance in performance on false belief and deception ToM tasks (Hughes and Cutting, 1999). In a second study of

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

479

1116 five-year-old twin pairs, heritable genes accounted for only 7% of the variance in performance on first and secondorder false belief tasks, relative to 45% of the variance accounted for by nonshared environmental factors (e.g., siblings’ different relationships with parents, with each other, and with peers, as well as child-specific life events like illness), and 28% by shared environmental factors (e.g., other siblings and SES-related factors). Similarity in ToM performance did not vary by zygosity of twin pairs in this sample (Hughes et al., 2005). What could explain the discrepancy in these results? As the authors point out, one possibility is that the larger sample of the second study afforded the sensitivity to measure the effects of shared environment. A second possibility is that the relative impact of genes and environment changes over development, with genes having a larger impact on ToM in early childhood (Study 1, age 3 years), and environment playing a larger role in ToM development later on (Study 2, age 5 years). While the standard pattern in most aspects of development (McGue et al., 1993; Plomin and Kosslyn, 2001), including prosocial behavior (Knafo and Plomin, 2006), is that the role of genes increases with age, an initial drop followed by a rise in the role of genes on prosocial sharing and comforting behaviors has been observed in young children (Knafo et al., 2008). However, results from a longitudinal study of >1000 twin pairs studied at 2, 3, and 4 years of age argue against these two possibilities. This study, which was similar in sample size as the initial study, found a moderate role of genes, accounting for 25%e57% of the variance on a questionnaire-based assessment of ToM. Additionally, the role of shared environment on ToM tasks decreased with age (Ronald et al. (2005); though see Ronald et al. (2006) and Warrier and Baron-Cohen (2018), for evidence of little to no role of genes on ToM in 9- and 13-year-olds, respectively). A third possibility is that the inclusion of children from very low socioeconomic status (SES) families in the second study increased its sensitivity to the effects of shared environment. In other aspects of cognition, like IQ, shared environment is a stronger predictor than genes in impoverished families, whereas the opposite is true in affluent families (Turkheimer et al., 2003). Thus, the effects of shared environment on ToM may be stronger in samples that include more children living in disadvantageous environments. Future studies that take into account developmental changes with age, differences by ToM task, and gene  environment interactions (Knafo and Israel, 2009; Knafo-Noam et al., 2018) are necessary to clarify the relative role of genes on ToM development. Could neuroimaging measures help to clarify the heritability of ToM in children? Among adults, there is some evidence that cortical responses in functionally selective brain regions are heritable. For example, the pattern of response to faces and scenes in the ventral visual cortex is more similar in monozygotic twins than in dizygotic twins (Polk et al., 2007). Thus, it is plausible that responses in ToM brain regions could similarly reflect the role of genes in ToM. Additionally, more similar responses in monozygotic twins compared to dizygotic twins in the response of the RTPJ would provide strong evidence that the genetic impact is on ToM per se, rather than domain-general abilities that are heritable and correlated with ToM (like language and executive functions), given the functional specialization of this brain region. By comparing RTPJ responses in twin pairs of different ages, this kind of research could also inform developmental changes in the relative roles of genes and environment on ToM. Finally, studies that measure correlations between particular genes and functionally selective responses could provide specific insight into how genes shape ToM development. In sum, neuroimaging measures may be a promising way to make progress in this area.

21.5 Conclusion Throughout this chapter, we have summarized exciting progress in developmental cognitive neuroscience, as well as several avenues for future research. Recent methodological advances, like the utilization of naturalistic movie-viewing paradigms and fNIRS, have enabled us to begin to describe the early neural signatures of ToM brain regions. Simultaneously, studies of older children have begun to test different hypotheses about the kinds of neural changes that might support ongoing behavioral ToM development after age 5. Future studies on neural markers that reflect structural changes in ToM, or the impact of developmental or cultural experience, will provide exciting insight into ToM development: What does it mean to become better at ToM in mid- to late childhood, and what is the basis for individual differences in ToM in adults? This work will also lead to a better understanding of the development of functionally specialized brain regions for ToM: What experiences are necessary or sufficient for the maintenance and fine-tuning of preferential responses for ToM? Finally, a key challenge for future studies is to discover reliable neural markers of ToM via longitudinal studies. These markers will provide specific hypotheses about neural differences in clinical populations that struggle disproportionately with ToM reasoning. We are excited to see progress made toward these aims by developmental cognitive neuroscientists in the years to come.

480 PART | II Cognitive development

References Amsterlaw, J., Wellman, H.M., 2006. Theories of mind in transition: a microgenetic study of the development of false belief understanding. J. Cogn. Dev. 7 (2), 139e172. https://doi.org/10.1207/s15327647jcd0702_1. Astington, J.W., 2006. The developmental interdependence of theory of mind and language. In: The Roots of Human Sociality: Culture, Cognition, and Human Interaction, pp. 179e206. Astington, J.W., Jenkins, J.M., 1999. A longitudinal study of the relation between language and theory-of-mind development. Dev. Psychol. 35 (5), 1311. Avinun, R., Knafo, A., 2014. Parenting as a reaction evoked by children’s genotype: a meta-analysis of children-as-twins studies. Personal. Soc. Psychol. Rev. 18 (1), 87e102. Avis, J., Harris, P.L., 1991. Belief-desire reasoning among Baka children: evidence for a universal conception of mind. Child Dev. 62 (3), 460e467. Barrett, H.C., Bolyanatz, A., Crittenden, A.N., Fessler, D.M., Fitzpatrick, S., Gurven, M., et al., 2016. Small-scale societies exhibit fundamental variation in the role of intentions in moral judgment. Proc. Natl. Acad. Sci. U.S.A. 113 (17), 4688e4693. Bedny, M., Pascual-Leone, A., Saxe, R.R., 2009. Growing up blind does not change the neural bases of Theory of Mind. Proc. Natl. Acad. Sci. U.S.A. 106 (27), 11312e11317. Blank, I., Kanwisher, N., Fedorenko, E., 2014. A functional dissociation between language and multiple-demand systems revealed in patterns of BOLD signal fluctuations. J. Neurophysiol. 112 (5), 1105e1118. Bloom, P., German, T.P., 2000. Two reasons to abandon the false belief task as a test of theory of mind. Cognition 77 (1), B25eB31. Brambring, M., Asbrock, D., 2010. Validity of false belief tasks in blind children. J. Autism Dev. Disord. 40 (12), 1471e1484. https://doi.org/10.1007/ s10803-010-1002-2. Brown, R., Hobson, R.P., Lee, A., Stevenson, J., 1997. Are there “Autistic-like” features in congenitally blind children? J. Child Psychol. Psychiatry 38 (6), 693e703. Bruneau, E.G., Pluta, A., Saxe, R., 2012. Distinct roles of the “shared pain’and ‘theory of mind’networks in processing others” emotional suffering. Neuropsychologia 50 (2), 219e231. Brunet, E., Sarfati, Y., Hardy-Baylé, M.-C., Decety, J., 2000. A PET investigation of the attribution of intentions with a nonverbal task. Neuroimage 11 (2), 157e166. https://doi.org/10.1006/nimg.1999.0525. Brunet, E., Sarfati, Y., Hardy-Baylé, M.-C., Decety, J., 2003. Abnormalities of brain function during a nonverbal theory of mind task in schizophrenia. Neuropsychologia 41 (12), 1574e1582. Buckner, R.L., Andrews Hanna, J.R., Schacter, D.L., 2008. The brain’s default network. Ann. N. Y. Acad. Sci. 1124 (1), 1e38. Buttelmann, D., Carpenter, M., Tomasello, M., 2009. Eighteen-month-old infants show false belief understanding in an active helping paradigm. Cognition 112 (2), 337e342. https://doi.org/10.1016/j.cognition.2009.05.006. Callaghan, T., Rochat, P., Lillard, A., Claux, M.L., Odden, H., Itakura, S., et al., 2005. Synchrony in the onset of mental-state reasoning: evidence from five cultures. Psychol. Sci. 16 (5), 378e384. Cantlon, J.F., Li, R., 2013. Neural activity during natural viewing of Sesame Street statistically predicts test scores in early childhood. PLoS Biol. 11 (1), e1001462. Cantlon, J.F., Pinel, P., Dehaene, S., Pelphrey, K.A., 2010. Cortical representations of symbols, objects, and faces are pruned back during early childhood. Cerebr. Cortex 21 (1), 191e199. https://doi.org/10.1093/cercor/bhq078. Cao, M., He, Y., Dai, Z., Liao, X., Jeon, T., Ouyang, M., et al., 2016. Early development of functional network segregation revealed by connectomic analysis of the preterm human brain. Cerebr. Cortex 27 (3), 1949e1963. Carlson, S.M., Moses, L.J., 2001. Individual differences in inhibitory control and children’s theory of mind. Child Dev. 72 (4), 1032e1053. Carlson, S.M., Moses, L.J., Hix, H.R., 1998. The role of inhibitory processes in young children’s difficulties with deception and false belief. Child Dev. 69 (3), 672e691. Carrington, S.J., Bailey, A.J., 2009. Are there theory of mind regions in the brain? A review of the neuroimaging literature. Hum. Brain Mapp. 30 (8), 2313e2335. https://doi.org/10.1002/hbm.20671. Chen, J., Leong, Y.C., Norman, K.A., Hasson, U., 2016. Shared Experience, Shared Memory: A Common Structure For Brain Activity During Naturalistic Recall. bioRxiv, p. 035931. Cole, M.W., Bassett, D.S., Power, J.D., Braver, T.S., Petersen, S.E., 2014. Intrinsic and task-evoked network architectures of the human brain. Neuron 83 (1), 238e251. Cooper, R., Selb, J., Gagnon, L., Phillip, D., Schytz, H.W., Iversen, H.K., et al., 2012. A systematic comparison of motion artifact correction techniques for functional near-infrared spectroscopy. Front. Neurosci. 6, 147. Cushman, F., Sheketoff, R., Wharton, S., Carey, S., 2013. The development of intent-based moral judgment. Cognition 127 (1), 6e21. Deen, B., Koldewyn, K., Kanwisher, N., Saxe, R., 2015. Functional organization of social perception and cognition in the superior temporal sulcus. Cerebr. Cortex 25 (11), 4596e4609. Dehaene, S., Pegado, F., Braga, L.W., Ventura, P., Nunes Filho, G., Jobert, A., et al., 2010. How learning to read changes the cortical networks for vision and language. Science 330 (6009), 1359e1364. Dehaene-Lambertz, G., Monzalvo, K., Dehaene, S., 2018. The emergence of the visual word form: longitudinal evolution of category-specific ventral visual areas during reading acquisition. PLoS Biol. 16 (3), e2004103. Doria, V., Beckmann, C.F., Arichi, T., Merchant, N., Groppo, M., Turkheimer, F.E., et al., 2010. Emergence of resting state networks in the preterm human brain. Proc. Natl. Acad. Sci. U.S.A. 107 (46), 20015e20020.

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

481

Drayton, L.A., Santos, L.R., 2016. A decade of theory of mind research on Cayo Santiago: insights into rhesus macaque social cognition. Am. J. Primatol. 78 (1), 106e116. Dubois, J., Adolphs, R., 2016. Building a science of individual differences from fMRI. Trends Cogn. Sci. 20 (6), 425e443. https://doi.org/10.1016/ j.tics.2016.03.014. Dubois, J., Adolphs, R., 2016. How the brain represents other minds. Proc. Natl. Acad. Sci. U.S.A. 113 (1), 19e21. https://doi.org/10.1073/ pnas.1522316113. Filippova, E., Astington, J.W., 2008. Further development in social reasoning revealed in discourse irony understanding. Child Dev. 79 (1), 126e138. https://doi.org/10.1111/j.1467-8624.2007.01115.x. Foulkes, L., Blakemore, S.J., 2018. Studying Individual Differences in Human Adolescent Brain Development, vol. 1. Nature Publishing Group. Fox, M.D., Snyder, A.Z., Vincent, J.L., Corbetta, M., Van Essen, D.C., Raichle, M.E., 2005. The human brain is intrinsically organized into dynamic, anticorrelated functional networks. Proc. Natl. Acad. Sci. U.S.A. 102 (27), 9673e9678. Franchak, J.M., Heeger, D.J., Hasson, U., Adolph, K.E., 2016. Free viewing gaze behavior in infants and adults. Infancy 21 (3), 262e287. Frank, M.C., Vul, E., Saxe, R., 2011. Measuring the development of social attention using free-viewing. Infancy 17 (4), 355e375. https://doi.org/10.1111/ j.1532-7078.2011.00086.x. Gao, X., Maurer, D., 2010. A happy story: developmental changes in children’s sensitivity to facial expressions of varying intensities. J. Exp. Child Psychol. 107 (2), 67e86. Gao, W., Zhu, H., Giovanello, K.S., Smith, J.K., Shen, D., Gilmore, J.H., Lin, W., 2009. Evidence on the emergence of the brain’s default network from 2-week-old to 2-year-old healthy pediatric subjects. Proc. Natl. Acad. Sci. U.S.A. 106 (16), 6790e6795. Gao, W., Gilmore, J.H., Shen, D., Smith, J.K., Zhu, H., Lin, W., 2013. The synchronization within and interaction between the default and dorsal attention networks in early infancy. Cerebr. Cortex 23 (3), 594e603. Geangu, E., Ichikawa, H., Lao, J., Kanazawa, S., Yamaguchi, M.K., Caldara, R., Turati, C., 2016. Culture shapes 7-month-olds’ perceptual strategies in discriminating facial expressions of emotion. Curr. Biol. 26 (14), R663eR664. Golarai, G., Ghahremani, D.G., Whitfield-Gabrieli, S., Reiss, A., Eberhardt, J.L., Gabrieli, J.D., Grill-Spector, K., 2007. Differential development of highlevel visual cortex correlates with category-specific recognition memory. Nat. Neurosci. 10 (4), 512. Gomez, J., Barnett, M.A., Natu, V., Mezer, A., Palomero-Gallagher, N., Weiner, K.S., et al., 2017. Microstructural proliferation in human cortex is coupled with the development of face processing. Science 355 (6320), 68e71. Gopnik, A., Wellman, H.M., 1992. Why the child’s theory of mind really is a theory. Mind Lang. 7 (1-2), 145e171. Gopnik, A., Wellman, H.M., 1994. The theory theory. In: Mapping the Mind: Domain Specificity in Cognition and Culture, p. 257. Greicius, M.D., Krasnow, B., Reiss, A.L., Menon, V., 2003. Functional connectivity in the resting brain: a network analysis of the default mode hypothesis. Proc. Natl. Acad. Sci. U.S.A. 100 (1), 253e258. Grill-Spector, K., Golarai, G., Gabrieli, J., 2008. Developmental neuroimaging of the human ventral visual cortex. Trends Cogn. Sci. 12 (4), 152e162. Grosse Wiesmann, C., Schreiber, J., Singer, T., Steinbeis, N., Friederici, A.D., 2017. White matter maturation is associated with the emergence of Theory of Mind in early childhood. Nat. Commun. 8, 14692. Guggenheim, R. (Producer)., Arnold, B. (Producer)., Lasseter, J. (Director)., 1995. Toy Story [Motion Picture] USA: Pixar Animation Studios and Walt Disney Pictures. Gusnard, D.A., Akbudak, E., Shulman, G.L., Raichle, M.E., 2001. Medial prefrontal cortex and self-referential mental activity: relation to a default mode of brain function. Proc. Natl. Acad. Sci. U.S.A. 98 (7), 4259e4264. Gweon, H., Dodell-Feder, D., Bedny, M., Saxe, R., 2012. Theory of mind performance in children correlates with functional specialization of a brain region for thinking about thoughts. Child Dev. 83 (6), 1853e1868. https://doi.org/10.1111/j.1467-8624.2012.01829.x. Happé, F., Conway, J.R., 2016. Recent progress in understanding skills and impairments in social cognition. Curr. Opin. Pediatr. 28 (6), 736e742. Happé, F., Frith, U., 2014. Annual research review: towards a developmental neuroscience of atypical social cognition. J. Child Psychol. Psychiatry 55 (6), 553e577. Happé, F., Ehlers, S., Fletcher, P., Frith, U., Johansson, M., Gillberg, C., et al., 1996. “Theory of mind”in the brain. Evidence from a PET scan study of Asperger syndrome. Neuroreport 8 (1), 197e201. Hasson, U., 2004. Intersubject synchronization of cortical activity during natural vision. Science 303 (5664), 1634e1640. https://doi.org/10.1126/ science.1089506. Henrich, J., 2011. Thematic Series: The Emerging Science of Culture the Weirdest People in the World. Honey, C.J., Thompson, C.R., Lerner, Y., Hasson, U., 2012. Not lost in translation: neural responses shared across languages. J. Neurosci. 32 (44), 15277e15283. Hughes, C., Cutting, A.L., 1999. Nature, nurture, and individual differences in early understanding of mind. Psychol. Sci. 10 (5), 429e432. Hughes, C., Jaffee, S.R., Happé, F., Taylor, A., Caspi, A., Moffitt, T.E., 2005. Origins of individual differences in theory of mind: from nature to nurture? Child Dev. 76 (2), 356e370. Hyde, D.C., Aparicio Betancourt, M., Simon, C.E., 2015. Human temporal-parietal junction spontaneously tracks others’ beliefs: a functional nearinfrared spectroscopy study. Hum. Brain Mapp. 36 (12), 4831e4846. Hyde, D.C., Simon, C.E., Ting, F., Nikolaeva, J., 2018. Functional organization of the temporal-parietal junction for theory of mind in preverbal infants: a near-infrared spectroscopy study. J. Neurosci. 38 (18), 4264e4274. Jääskeläinen, I.P., Koskentalo, K., Balk, M.H., Autti, T., Kauramäki, J., Pomren, C., Sams, M., 2008. Inter-subject synchronization of prefrontal cortex hemodynamic activity during natural viewing. Open Neuroimaging J. 2, 14.

482 PART | II Cognitive development

Kilford, E.J., Garrett, E., Blakemore, S.-J., 2016. The development of social cognition in adolescence: an integrated perspective. Neurosci. Biobehav. Rev. 70, 106e120. Knafo, A., Israel, S., 2009. Genetic and environmental influences on prosocial behavior. In: Prosocial Motives, Emotions, and Behavior: the Better Angels of Our Nature, pp. 149e167. Knafo, A., Plomin, R., 2006. Prosocial behavior from early to middle childhood: genetic and environmental influences on stability and change. Dev. Psychol. 42 (5), 771. Knafo, A., Zahn-Waxler, C., Van Hulle, C., Robinson, J.L., Rhee, S.H., 2008. The developmental origins of a disposition toward empathy: genetic and environmental contributions. Emotion 8 (6), 737. Knafo-Noam, A., Vertsberger, D., Israel, S., 2018. Genetic and environmental contributions to children’s prosocial behavior: brief review and new evidence from a reanalysis of experimental twin data. Curr. Opin. Psychol. 20, 60e65. Knudsen, B., Liszkowski, U., 2012. Eighteen-and 24-month-old infants correct others in anticipation of action mistakes. Dev. Sci. 15 (1), 113e122. Knudsen, B., Liszkowski, U., 2012. 18-Month-Olds predict specific action mistakes through attribution of false belief, not ignorance, and intervene accordingly. Infancy 17 (6), 672e691. Kobayashi, C., Glover, G.H., Temple, E., 2007. Children’s and adults’ neural bases of verbal and nonverbal ‘theory of mind’. Neuropsychologia 45 (7), 1522e1532. Koster-Hale, J., Saxe, R., Dungan, J., Young, L.L., 2013. Decoding moral judgments from neural representations of intentions. Proc. Natl. Acad. Sci. U.S.A. 110 (14), 5648e5653. https://doi.org/10.1073/pnas.1207992110. Koster-Hale, J., Bedny, M., Saxe, R., 2014. Thinking about seeing: perceptual sources of knowledge are encoded in the theory of mind brain regions of sighted and blind adults. Cognition 133 (1), 65e78. https://doi.org/10.1016/j.cognition.2014.04.006. Koster-Hale, J., Richardson, H., Velez, N., Asaba, M., Young, L., Saxe, R., 2017. Mentalizing regions represent distributed, continuous, and abstract dimensions of others’ beliefs. Neuroimage 161, 9e18. https://doi.org/10.1016/j.neuroimage.2017.08.026. Lagattuta, K.H., Wellman, H.M., 2001. Thinking about the past: early knowledge about links between prior experience, thinking, and emotion. Child Dev. 72 (1), 82e102. Laird, A.R., Eickhoff, S.B., Rottschy, C., Bzdok, D., Ray, K.L., Fox, P.T., 2013. Networks of task co-activations. Neuroimage 80, 505e514. Lerner, Y., Honey, C.J., Silbert, L.J., Hasson, U., 2011. Topographic mapping of a hierarchy of temporal receptive windows using a narrated story. J. Neurosci. 31 (8), 2906e2915. https://doi.org/10.1523/JNEUROSCI.3684-10.2011. Lloyd-Fox, S., Blasi, A., Elwell, C.E., 2010. Illuminating the developing brain: the past, present and future of functional near infrared spectroscopy. Neurosci. Biobehav. Rev. 34 (3), 269e284. Lloyd-Fox, S., Richards, J.E., Blasi, A., Murphy, D.G., Elwell, C.E., Johnson, M.H., 2014. Coregistering functional near-infrared spectroscopy with underlying cortical areas in infants. Neurophotonics 1 (2), 025006. Lloyd-Fox, S., Begus, K., Halliday, D., Pirazzoli, L., Blasi, A., Papademetriou, M., et al., 2017. Cortical specialisation to social stimuli from the first days to the second year of life: a rural Gambian cohort. Accid. Anal. Prev. 25, 92e104. Mar, R.A., 2011. The neural bases of social cognition and story comprehension. Annu. Rev. Psychol. 62, 103e134. Martin, A., Santos, L.R., 2016. What cognitive representations support primate theory of mind? Trends Cogn. Sci. 20 (5), 375e382. McCandliss, B.D., Cohen, L., Dehaene, S., 2003. The visual word form area: expertise for reading in the fusiform gyrus. Trends Cogn. Sci. 7 (7), 293e299. https://doi.org/10.1016/S1364-6613(03)00134-7. McGue, M., Bouchard Jr., T.J., Iacono, W.G., Lykken, D.T., 1993. Behavioral Genetics of Cognitive Ability: A Life-span Perspective. Meristo, M., Morgan, G., Geraci, A., Iozzi, L., Hjelmquist, E., Surian, L., Siegal, M., 2012. Belief attribution in deaf and hearing infants. Dev. Sci. 15 (5), 633e640. Miall, R.C., Robertson, E.M., 2006. Functional imaging: is the resting brain resting? Curr. Biol. 16 (23), R998eR1000. Milligan, K., Astington, J.W., Dack, L.A., 2007. Language and theory of mind: meta-analysis of the relation between language ability and false-belief understanding. Child Dev. 78 (2), 622e646. Minter, M., Hobson, R.P., Bishop, M., 1998. Congenital visual impairment and“theory of mind.” Br. J. Dev. Psychol. 16, 183. Mitchell, J.P., Banaji, M.R., Macrae, C.N., 2005. General and specific contributions of the medial prefrontal cortex to knowledge about mental states. Neuroimage 28 (4), 757e762. Moraczewski, D., Chen, G., Redcay, E., 2018. Inter-subject synchrony as an index of functional specialization in early childhood. Sci. Rep. 8 (1), 2252. Moreira, J.F.G., McLaughlin, K.A., Silvers, J.A., 2018. Spatial and Temporal Cortical Variability Track With Age and Affective Experience During Emotion Regulation in Youth. bioRxiv, p. 291245. Nguyen, M., Vanderwal, T., Hasson, U., 2017. Shared Understanding is Correlated With Shared Neural Responses in The Default Mode Network. bioRxiv, p. 231019. Nummenmaa, L., Glerean, E., Viinikainen, M., Jääskeläinen, I.P., Hari, R., Sams, M., 2012. Emotions promote social interaction by synchronizing brain activity across individuals. Proc. Natl. Acad. Sci. U.S.A. 109 (24), 9599e9604. Paunov, A., Blank, I., Fedorenko, E., 2017. Functionally Distinct Language and Theory of Mind Networks Work Together, Especially During Language Comprehension, 27. PsyArXiv. Perner, J., Leekam, S.R., Wimmer, H., 1987. Three-year-olds’ difficulty with false belief: the case for a conceptual deficit. Br. J. Dev. Psychol. 5 (2), 125e137. Peterson, C.C., Siegal, M., 1999. Representing inner worlds: theory of mind in autistic, deaf, and normal hearing children. Psychol. Sci. 10 (2), 126e129.

Early signatures of and developmental change in brain regions for theory of mind Chapter | 21

483

Peterson, C.C., Wellman, H.M., 2018. Longitudinal theory of mind (ToM) development from preschool to adolescence with and without ToM Delay. Child Dev. 90 (6), 1917e1934. Peterson, C.C., Peterson, J.L., Webb, J., 2000. Factors influencing the development of a theory of mind in blind children. Br. J. Dev. Psychol. 18 (3), 431e447. Peterson, C.C., Wellman, H.M., Slaughter, V., 2012. The mind behind the message: advancing theory-of-mind scales for typically developing children, and those with deafness, autism, or Asperger syndrome. Child Dev. 83 (2) https://doi.org/10.1111/j.1467-8624.2011.01728.x. Plomin, R., Kosslyn, S.M., 2001. Genes, brain and cognition. Nat. Neurosci. 4 (12), 1153e1154. Polk, T.A., Park, J., Smith, M.R., Park, D.C., 2007. Nature versus nurture in ventral visual cortex: a functional magnetic resonance imaging study of twins. J. Neurosci. 27 (51), 13921e13925. Powell, L.J., Deen, B., Saxe, R., 2017. Using individual functional channels of interest to study cortical development with fNIRS. Dev. Sci. 21 (4), e12595. Premack, D., Woodruff, G., 1978. Does the chimpanzee have a theory of mind? Behav. Brain Sci. 1 (04), 515e526. Priewasser, B., Rafetseder, E., Gargitter, C., Perner, J., 2017. Helping as an early indicator of a theory of mind: mentalism or teleology? Cogn. Dev. 46, 69e78. Raichle, M.E., MacLeod, A.M., Snyder, A.Z., Powers, W.J., Gusnard, D.A., Shulman, G.L., 2001. A default mode of brain function. Proc. Natl. Acad. Sci. U.S.A. 98 (2), 676e682. Raschle, N.M., Lee, M., Buechler, R., Christodoulou, J.A., Chang, M., Vakil, M., et al., 2009. Making MR imaging child’s play-pediatric neuroimaging protocol, guidelines and procedure. J. Vis. Exp. (29). Reher, K. (Producer)., Sohn, P. (Directory)., 2009. Partly Cloudy [Motion Picture] USA: Pixar Animation Studios and Walt Disney Pictures. Richardson, H., 2019. Development of brain networks for social functions: confirmatory analyses in a large open source dataset. Dev. Cognit. Neurosci. 37, 100598. Richardson, H., Lisandrelli, G., Riobueno-Naylor, A., Saxe, R., 2018. Development of the social brain from age three to twelve years. Nat. Commun. 9 (1), 1027. Richardson, H., Koster-Hale, J., Caselli, N., Magid, R.W., Benedict, R., Olson, H., Saxe, R., 2019. Reduced Neural Selectivity For Mental States in Deaf Children With Delayed Exposure to Sign Language. PsyArxiv. https://doi.org/10.31234/osf.io/j5ghw (In revision). Ronald, A., Happé, F., Hughes, C., Plomin, R., 2005. Nice and nasty theory of mind in preschool children: nature and nurture. Soc. Dev. 14 (4), 664e684. Ronald, A., Viding, E., Happé, F., Plomin, R., 2006. Individual differences in theory of mind ability in middle childhood and links with verbal ability and autistic traits: a twin study. Soc. Neurosci. 1 (3e4), 412e425. Rosati, A.G., Wobber, V., Hughes, K., Santos, L.R., 2014. Comparative developmental psychology: how is human cognitive development unique? Evol. Psychol. 12 (2), 147470491401200211. Saxe, R., 2018. Seeing other minds in 3D. Trends Cogn. Sci. 22 (3), 193e195. Saxe, R., Kanwisher, N., 2003. People thinking about thinking people: the role of the temporo-parietal junction in “theory of mind.” Neuroimage 19 (4), 1835e1842. Saxe, R., Powell, L.J., 2006. It’s the thought that counts specific brain regions for one component of theory of mind. Psychol. Sci. 17 (8), 692e699. Saxe, R.R., Whitfield-Gabrieli, S., Scholz, J., Pelphrey, K.A., 2009. Brain regions for perceiving and reasoning about other people in school-aged children. Child Dev. 80 (4), 1197e1209. Schick, B., Hoffmeister, R., 2001. ASL skills in deaf children of deaf parents and of hearing parents. In: Presented at the Society for Research in Child Development International Conference, Minneapolis, MN. Schick, B., De Villiers, P., De Villiers, J., Hoffmeister, R., 2007. Language and theory of mind: a study of deaf children. Child Dev. 78 (2), 376e396. Schurz, M., Radua, J., Aichhorn, M., Richlan, F., Perner, J., 2014. Fractionating theory of mind: a meta-analysis of functional brain imaging studies. Neurosci. Biobehav. Rev. 42, 9e34. https://doi.org/10.1016/j.neubiorev.2014.01.009. Selcuk, B., Brink, K.A., Ekerim, M., Wellman, H.M., 2018. Sequence of theory-of-mind acquisition in Turkish children from diverse social backgrounds. Infant Child Dev. e2098. Shahaeian, A., Peterson, C.C., Slaughter, V., Wellman, H.M., 2011. Culture and the sequence of steps in theory of mind development. Dev. Psychol. 47 (5), 1239. Skerry, A.E., Spelke, E.S., 2014. Preverbal infants identify emotional reactions that are incongruent with goal outcomes. Cognition 130 (2), 204e216. Sliwa, J., Freiwald, W.A., 2017. A dedicated network for social interaction processing in the primate brain. Science 356 (6339), 745e749. Smith, S.M., Fox, P.T., Miller, K.L., Glahn, D.C., Fox, P.M., Mackay, C.E., et al., 2009. Correspondence of the brain’s functional architecture during activation and rest. Proc. Natl. Acad. Sci. U.S.A. 106 (31), 13040e13045. Smyser, C.D., Inder, T.E., Shimony, J.S., Hill, J.E., Degnan, A.J., Snyder, A.Z., Neil, J.J., 2010. Longitudinal analysis of neural network development in preterm infants. Cerebr. Cortex 20 (12), 2852e2862. Sommer, M., Döhnel, K., Sodian, B., Meinhardt, J., Thoermer, C., Hajak, G., 2007. Neural correlates of true and false belief reasoning. Neuroimage 35 (3), 1378e1384. Southgate, V., Senju, A., Csibra, G., 2007. Action anticipation through attribution of false belief by 2-year-olds. Psychol. Sci. 18 (7), 587e592. https:// doi.org/10.1111/j.1467-9280.2007.01944.x. Stephens, G.J., Silbert, L.J., Hasson, U., 2010. Speakerelistener neural coupling underlies successful communication. Proc. Natl. Acad. Sci. U.S.A. 107 (32), 14425e14430. Surian, L., Caldi, S., Sperber, D., 2007. Attribution of beliefs by 13-month-old infants. Psychol. Sci. 18 (7), 580e586.

484 PART | II Cognitive development

Tamir, D.I., Thornton, M.A., Contreras, J.M., Mitchell, J.P., 2016. Neural evidence that three dimensions organize mental state representation: rationality, social impact, and valence. Proc. Natl. Acad. Sci. U.S.A. 113 (1), 194e199. https://doi.org/10.1073/pnas.1511905112. Turkheimer, E., Haley, A., Waldron, M., d’Onofrio, B., Gottesman, I.I., 2003. Socioeconomic status modifies heritability of IQ in young children. Psychol. Sci. 14 (6), 623e628. van den Heuvel, M.I., Thomason, M.E., 2016. Functional connectivity of the human brain in utero. Trends Cogn. Sci. 20 (12), 931e939. Vanderwal, T., Kelly, C., Eilbott, J., Mayes, L.C., Castellanos, F.X., 2015. Inscapes: a movie paradigm to improve compliance in functional magnetic resonance imaging. Neuroimage 122, 222e232. Vogeley, K., Bussfeld, P., Newen, A., Herrmann, S., Happé, F., Falkai, P., et al., 2001. Mind reading: neural mechanisms of theory of mind and selfperspective. Neuroimage 14 (1), 170e181. https://doi.org/10.1006/nimg.2001.0789. Wade, M., Hoffmann, T.J., Jenkins, J.M., 2015. Geneeenvironment interaction between the oxytocin receptor (OXTR) gene and parenting behaviour on children’s theory of mind. Soc. Cogn. Affect. Neurosci. 10 (12), 1749e1757. Warrier, V., Baron-Cohen, S., 2018. Genetic contribution to “theory of mind”in adolescence. Sci. Rep. 8 (1), 3465. Weisman, K., Dweck, C.S., Markman, E.M., 2017. Children’s intuitions about the structure of mental life. In: Presented at the the Annual Meeting of the Cognitive Science Society. Wellman, H.M., Liu, D., 2004. Scaling of theory-of-mind tasks. Child Dev. 75 (2), 523e541. Wellman, H.M., Cross, D., Watson, J., 2001. Meta-analysis of theory-of-mind development: the truth about false belief. Child Dev. 72 (3), 655e684. Wellman, H.M., Fang, F., Liu, D., Zhu, L., Liu, G., 2006. Scaling of theory-of-mind understandings in Chinese children. Psychol. Sci. 17 (12), 1075e1081. Wellman, H.M., Lopez-Duran, S., LaBounty, J., Hamilton, B., 2008. Infant attention to intentional action predicts preschool theory of mind. Dev. Psychol. 44 (2), 618e623. https://doi.org/10.1037/0012-1649.44.2.618. Wellman, H.M., Fang, F., Peterson, C.C., 2011. Sequential progressions in a theory-of-mind scale: longitudinal perspectives. Child Dev. 82 (3), 780e792. Widen, S.C., Russell, J.A., 2003. A closer look at preschoolers’ freely produced labels for facial expressions. Dev. Psychol. 39 (1), 114. Widen, S.C., Russell, J.A., 2008. Children acquire emotion categories gradually. Cogn. Dev. 23 (2), 291e312. Wilson, S.M., Molnar-Szakacs, I., Iacoboni, M., 2007. Beyond superior temporal cortex: intersubject correlations in narrative speech comprehension. Cerebr. Cortex 18 (1), 230e242. Wimmer, H., Perner, J., 1983. Beliefs about beliefs: representation and constraining function of wrong beliefs in young children’s understanding of deception. Cognition 13 (1), 103e128. https://doi.org/10.1016/0010-0277(83)90004-5. Woolfe, T., Want, S.C., Siegal, M., 2002. Signposts to development: theory of mind in deaf children. Child Dev. 73 (3), 768e778. Wu, N., Su, Y., 2015. Oxytocin receptor gene relates to theory of mind and prosocial behavior in children. J. Cogn. Dev. 16 (2), 302e313. Wu, Y., Muentener, P., Schulz, L.E., 2017. One-to four-year-olds connect diverse positive emotional vocalizations to their probable causes. Proc. Natl. Acad. Sci. 114 (45), 11896e11901. Xiao, Y., Geng, F., Riggins, T., Chen, G., Redcay, E., 2019. Neural correlates of developing theory of mind competence in early childhood. Neuroimage 184, 707e716. Yamaguchi, M., Kuhlmeier, V.A., Wynn, K., vanMarle, K., 2009. Continuity in social cognition from infancy to childhood. Dev. Sci. 12 (5), 746e752. https://doi.org/10.1111/j.1467-7687.2008.00813.x. Young, L., Saxe, R., 2009. An FMRI investigation of spontaneous mental state inference for moral judgment. J. Cogn. Neurosci. 21 (7), 1396e1405. Young, L., Camprodon, J.A., Hauser, M., Pascual-Leone, A., Saxe, R., 2010. Disruption of the right temporoparietal junction with transcranial magnetic stimulation reduces the role of beliefs in moral judgments. Proc. Natl. Acad. Sci. U.S.A. 107 (15), 6753e6758. https://doi.org/10.1073/ pnas.0914826107. Young, L., Dodell-Feder, D., Saxe, R., 2010. What gets the attention of the temporo-parietal junction? An fMRI investigation of attention and theory of mind. Neuropsychologia 48 (9), 2658e2664. Young, L., Nichols, S., Saxe, R., 2010. Investigating the neural and cognitive basis of moral luck: it’s not what you do but what you know. Rev. Philos. Psychol. 1 (3), 333e349. https://doi.org/10.1007/s13164-010-0027-y.

Chapter 22

A developmental neuroscience perspective on empathy Jean Decety1, 2 and Kalina J. Michalska3 1

Department of Psychology, Department of Psychiatry and Behavioral Neuroscience, The University of Chicago, Chicago, IL, United States; 2The

Child Neurosuite, The University of Chicago, Chicago, IL, United States; 3Department of Psychology, University of California, Riverside, CA, United States

Chapter outline 22.1. Introduction 22.2. Clearing up definitional issues 22.3. The development of empathy 22.3.1. Affect sharing and physiological synchrony 22.3.2. Emotion recognition 22.3.3. Emotion understanding 22.3.4. Perspective-taking and theory of mind 22.3.4.1. Neurophysiological approaches to understanding cognitive empathy 22.3.5. Emotion regulation 22.3.6. Motivation to care

485 486 488 488 489 490 490 491 491 493

22.4. Neurodevelopmental changes in empathic responding 22.4.1. Evidence from event-related potential 22.4.2. Evidence from functional magnetic resonance imaging 22.5. Maladaptive alterations in developmental trajectories of empathy 22.5.1. Conduct problems 22.5.2. Autism spectrum disorder 22.6. Conclusions List of abbreviations References

493 494 494 496 496 498 499 499 500

22.1 Introduction Developmental social neuroscience is a growing discipline that has the potential to improve our understanding of human cognitive and social behavior by integrating theory and research across disciplines including psychology, neuroscience, and psychiatry. Among the psychological processes that are the basis for much of sociability and smooth social interaction, empathy plays a pivotal role. Empathy-related responding, including caring and sympathetic concern, motivates bonding between individuals and some forms of prosocial behavior, inhibits aggression, and overall facilitates group living and cooperation. On the other hand, certain developmental disorders are marked by empathy deficits, which influence the motivation to respond to others in distress or need, and care for them. Understanding how these motivations and behaviors are implemented in the brain and downstream peripheral physiology, both in typically developing children and children with antisocial tendencies, can help elucidate the role of empathy in prosociality. It is important to note that empathy can compete with moral judgment and justice principles, for instance, by inducing partiality for in-group members (Cikara and Van Bavel, 2014; Decety and Cowell, 2014a). Thus, empathy and morality should not be conflated (Decety and Cowell, 2014b). This chapter critically examines the current knowledge about the development of the mechanisms supporting empathy and its associated behavioral responses such as some forms of prosocial behaviors including caring. We review the affective and cognitive components that give rise to empathy, starting first with the automatic proclivity to share affect and emotions with others, and then the cognitive processes of perspective-taking and executive control, which enable individuals to intentionally adopt the subjective view point of another without confusion between self and other. The goal of

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00022-6 Copyright © 2020 Elsevier Inc. All rights reserved.

485

486 PART | II Cognitive development

the chapter is to address the underlying autonomic, affective, and cognitive mechanisms involved in empathy, as well as their neural architecture, and to examine their dysfunctions in developmental disorders marked by socialecognitive impairments. Based on conceptual and empirical evidence from developmental psychology, cognitive neuroscience, and neurology, a number of distinct and interacting components contribute to the experience of empathy: (1) affect sharing, a bottom-up process grounded in affective arousal and neural circuits connecting the brain stem, amygdala, basal ganglia, and orbitofrontal cortex, (2) understanding emotion that relies on self- and other-awareness and critically involves the medial and ventromedial prefrontal cortex and temporoparietal junction, and (3) executive functions instantiated in the prefrontal cortex that operate as a top-down mediator, allowing for perspective-taking; emotion regulation, and appraisal of social context (Decety and Jackson, 2004; Decety and Meyer, 2008). Drawing from multiple sources of data and levels of analysis provides a more complete picture of the phenomenological experience of empathy, as well as an understanding of the development and interaction between the mechanisms that drive the phenomenon. Furthermore, studying subcomponents of more complex psychological constructs such as empathy can be particularly useful from a developmental perspective, because only some of its components or precursors may be observable. Developmental studies can provide unique opportunities to see how the components of the system interact in ways that are not always possible in adultsdwhere all the components are fully mature and operational. Until quite recently, research on the development of empathy-related responding from a neurobiological level of analysis has been relatively sparse. We believe that integrating this perspective with behavioral observations can shed light into the neurobiological mechanisms underpinning the basic building blocks of empathy and sympathy and their age-related functional changes. Such integration helps us characterize the neurobiological processes that underpin interpersonal affective responding and prosocial behavior while also potentially informing interventions for individuals with atypical development, such as antisocial behavior problems.

22.2 Clearing up definitional issues The construct of empathy is applied to various phenomena, covering a broad spectrum ranging from feeling concerned for others, experiencing emotions that match another individual’s emotions, knowing what another is thinking or feeling, to blurring the line between self and other. Key concepts l

l

l

l

l

l l

l

Empathy, the natural ability to share and understand the subjective states of others, can motivate caring for the well-being of another (also known as sympathy, empathic concern, or compassion). Empathy is implemented by a network of distributed, recursively connected, interacting neural systems and regions, including the superior temporal sulcus, insula, medial and orbitofrontal cortices, amygdala, and anterior cingulate cortex, as well as autonomic and neuroendocrine systems implicated in sociality. Prosocial behavior is an umbrella term for actions that are benefiting another person. This umbrella concept includes many different types of behaviors such as helping, cooperating, sharing, comforting, rescuing, and informing. These various forms of prosocial behaviors have distinct underlying motivations such as caring, fairness, reputation management, group loyalty, reciprocity, social rewards, etc. Morality refers to prescriptive norms regarding how people should treat one another, including concepts such as justice, fairness, and rights. All definitions of morality minimally include judgment of the rightness or wrongness of acts or behaviors that knowingly cause harm to people. Autonomic synchronization involves any associative pattern in the physiologies of interacting partner like a mother and her child, such as synchrony in heart rate, respiration rhythm, pupil diameter, and hormonal level. Emotion contagion is the unconscious tendency to take on the sensory, motor, physiological, and affective states of others. Neurodevelopmental studies provide unique opportunities to explore how the components of empathic responding interact in ways that are not possible in adults. Investigating dysfunction of the components of empathy provides important clues for understanding deviations that can lead to the lack of concern for others in social decision-making and behavior.

In psychology, empathy is generally defined as an affective response stemming from the understanding of another’s emotional state or a condition similar to what the other person is feeling or would be expected to feel in a given situation (Eisenberg and Eggum, 2009). Some scholars more narrowly characterize empathy as one specific set of congruent

A developmental neuroscience perspective on empathy Chapter | 22

487

emotions, those feelings that are more other-focused than self-focused, and employ the notion of empathic concern (Batson, 2012), which is functionally linked with a motivation to care for the welfare of another. Importantly, empathybased responding is not restricted to consoling actions. Rather, empathic concern is a motivation that can underlie different forms of prosocial behaviors depending on whether or not they are based on an emotional reaction to the others’ perceived need. The experience of empathy can lead to caring (an other-oriented motivation) or personal distress (an egoistic motivation to reduce stress by withdrawing from the stressor, thereby decreasing the likelihood of prosocial behavior). Emotion regulation is considered to be a critical component of empathy, as the modulation of emotional experience allows an individual to remain aware of an emotionally evocative situation without being overwhelmed or numbed by it. This is particularly important in the case of negative arousal (Decety and Lamm, 2009). Developmental research indicates that children who are able to regulate their emotions are also more likely to experience sympathy (concern) rather than personal distress (Eisenberg and Eggum, 2009). For instance, in 5- to 6-year-old children, prosocial behavior is significantly correlated with ratings of the emotional state of the protagonist but not with own emotional state, suggesting that empathic concern rather than personal distress is the primary influence on prosocial behavior (Williams et al., 2014). Contemporary conceptual and empirical work from developmental science and social neuroscience converges to view empathy as a construct reflecting a natural capacity to share and understand the subjective states of others, and comprising emotional (sharing affect with another), cognitive (understanding the other’s subjective state from her point of view), and motivational (feeling concerned for another) facets (Decety, 2015). These facets operate by way of automatic and controlled processes and interact with one another. Yet they can be dissociated, as they rely on partially separable information processing systems and underlie different functions (Shdo et al., 2018). Moreover, these facets have different developmental trajectories and phylogenetic roots (Decety and Svetlova, 2012). In particular, there is evidence for agerelated changes in these neural circuits, which, together with behavioral observations, reflect how brain maturation influences the reaction to the distress of others (Decety and Michalska, 2010; Levy et al., 2018). Empathy can thus be deconstructed into a process model that includes bottom-up processing of affect sharing and topdown processing in which the perceiver’s motivation, intentions, and self-regulation influence the extent of an empathic experience. Emotion contagion, emotion recognition, perspective-taking, caring for other, and emotion regulation constitute the basic macrocomponents of empathy, which are mediated by specific and interacting neural circuits, including regions of the prefrontal cortex, anterior cingulate cortex (ACC), medial prefrontal cortex (mPFC), orbitofrontal cortex (OFC), ventromedial prefrontal cortex (vmPFC) insula, amygdala, brain stem, basal ganglia, and frontoparietal attention networks (Decety, 2015) (Fig. 22.1). Consequently, this model assumes and predicts that dysfunction in any of these

FIGURE 22.1 Brain regions that are involved in the experience of empathy. Functional imaging studies reveal that the anterior insula and the anterior cingulate cortices are conjointly activated during the experience of negative emotion and during the perception of negative emotion in others. The insula provides a foundation for the representation of subjective bodily feelings, which substantiates emotional awareness. The anterior cingulate cortex (ACC) can be divided anatomically based on cognitive (dorsal) and emotional (ventral) components: The dorsal part is connected with the prefrontal cortex and parietal cortex as well as the motor system, making it a central station for processing top-down and bottom-up stimuli and assigning appropriate control to other areas in the brain; by contrast, the ventral part of the ACC is connected with amygdala (a structure involved in assigning affective significance to stimuli), striatum, hypothalamus, and anterior insula and is involved in assessing the salience of emotion and motivational information. Many functions are attributed to ACC, such as error detection, anticipation of tasks, motivation, and modulation of emotional responses. The medial prefrontal cortex (mPFC) is critically associated with theory of mind processes and emotion understanding. The ventromedial prefrontal cortex (vmPFC) is involved in sensory integration, in representing the affective value of reinforcers, and in social decision-making. In particular, the vmPFC is thought to regulate planning behavior associated with sensitivity to reward and punishment and is closely connected to the anterior insula and amygdala. The vmPFC is a critical hub for empathic concern and caring behavior (Decety and Cowell, 2014ab).

488 PART | II Cognitive development

subcomponents and their underlying neural circuits may lead to an alteration of the experience of empathy and correspond with selective social cognitive disorders depending on which aspect is disrupted (Blair and Blair, 2009; Decety and Moriguchi, 2007; Decety et al., 2013a,b; Michalska et al., 2016). It is also important to keep in mind that both interpersonal and contextual factors impact a person’s subjective experience of empathy. For instance, current mood states, relationship to the person, social status, and the context in which the interaction occurs influence the manner and the extent to which the observer will react (e.g., Kraus et al., 2010; Stellar et al., 2012).

22.3 The development of empathy Empathy in its mature form typically begins to emerge when a child has developed a greater awareness of the experience of others, during the second and third years of life, and arises in the context of someone else’s emotional experience. In the following sections, each of the components of empathy (affect sharing, emotion understanding, perspective-taking, and emotion regulation) will be considered separately from a developmental neuroscience perspective. These components are dissociable as documented in patients with brain lesions (Shdo et al., 2018), yet mature empathic sensitivity and concern depend on their functional integration in the service of goal-directed socioemotional behavior. In addition, both genetic and environmental factors contribute to the development of empathy and prosociality (Knafo et al., 2008). Both developmentally and evolutionarily, advanced forms of empathy are preceded by and grow out of more elementary ones, such as the capacity to express and respond adaptively to emotional signals (Decety et al., 2016).

22.3.1 Affect sharing and physiological synchrony While there is some controversy concerning the nature of empathic responding in very young children, there is ample behavioral evidence demonstrating that the affective component of empathy develops earlier than the cognitive and emotion regulation components. Affective responsiveness is known to be present at an early age, is involuntary, and relies on somatosensorimotor resonance between other and self (Decety and Meyer, 2008). The general consensus is that infants and toddlers are sensitive and responsive to their caregiver’s emotional cues and that some of the basic building blocks of empathy, such as emotion contagion and autonomic synchrony, are present early in life. For instance, mothers and their children share a deep physiological connection. This type of physiological linkage is shared by most mammals and represents the earliest form of emotional contagion that occurs between a mother and a child even before birth (Feldman, 2016). Specifically, during mothereinfant interactions, mothers’ emotional states are reflected in their nonverbal behaviors (facial expressions, body postures, and eye gaze) and their physiological responses (heart rate, respiration, hormonal levels). Infants implicitly pick up these subtle social signals from their caregivers, and this in turn has an impact on their own physiology and cognition (Prochazkova and Kret, 2017). The experience of synchrony during the first months of age demonstrates how critical environmental input is for both the maturation of neural circuits that support social engagement (Johnson et al., 2005) and for the infant’s ultimate emotional development. Importantly, parenteinfant synchrony at that age has been shown to predict infants’ self-regulation, empathy, and moral internalization across childhood and up to adolescence (Feldman, 2012). One channel by which affective arousal is transmitted physiologically is pupillary contagion, responding to pupil size observed in other people with changes in one’s own pupil. In a cross-sectional study, 6- and 9-month-old infants were presented with schematic depictions of eyes with smaller and larger pupils while their own pupil sizes were recorded (Fawcett et al., 2016). In both age groups, infants’ pupil size was greater when they viewed large-center circles than when they viewed small-center circles, with no observed differences when viewing large- versus small-center squares. Similarly, another study demonstrated that 6- and 12-month-old infants had greater pupil dilation while viewing videos of other infants laughing and crying than while viewing videos of neutral babbling (Geangu et al., 2011). Together, these findings demonstrate that infants are sensitive and responsive to subtle emotional cues indexing other people’s internal states. This spontaneous transfer of internal states is fundamental for survival, social group cohesion and, we contend, the development of empathy. Interestingly, a recent functional MRI-pupillometry study revealed that when the pupils of interacting partners synchronously dilate, affiliation is promoted (Prochazkova et al., 2018). Pupil mimicry modulates trust decisions through the activation of the theory-of-mind network (precuneus, temporoparietal junction superior temporal sulcus, and medial prefrontal cortex). This demonstrates that pupil mimicry is regulated by the theory-of-mind network and informs decisions of trust. Contagious crying is another example of early emotional contagion. An infants’ cry is the product of interactions between neuroanatomical structures and physiological mechanisms located in the brain stem, which are linked with the

A developmental neuroscience perspective on empathy Chapter | 22

489

sympathetic and the parasympathetic branches of the autonomic nervous system. Crying is part of a regulatory system in which the interplay of behavioral and physiological processes functions to maintain homeostatic balance, regulate the duration of alertness and attention, and elicit caretaking when internal or external demands are not met (Lester and Boukydis, 1985). It has been observed that infants respond to others’ distress with distress. For instance, a study that examined infants’ reactions to audiotapes of neonatal crying showed that 1-day-old babies cry in response to other infant cries, but not to the sound of their own cries (Martin and Clark, 1987). Another study recorded infants’ (1, 3, 6, and 9 months of age) emotional reactions in response to different types of cries (Geangu et al., 2010). The results revealed that infants in all age categories mimicked crying, with distress reactions highest in response to cries of pain. Thus, it appears that infants are endowed with an early capacity for differentiating among elemental distress signals. Complementing the aforementioned behavioral findings, neurophysiological approaches document that neonates seem to possess prewired neural mechanisms for discriminating vocal affect. One study demonstrated a mismatched electroencephalographic (EEG) response over the right hemisphere in response to emotionally laden (happy and fearful vs. neutral) syllables within the first few days of life (Cheng et al., 2012), indicating early discrimination of emotions. Similarly, functional magnetic resonance imaging (fMRI) data show that in 3- to 7-month-olds, sad vocalizations are associated with a selective increase of hemodynamic activity in brain regions involved in processing affective stimuli, such as the orbitofrontal cortex (OFC) and insula (Blasi et al., 2011). Newborn infants’ differentiation of emotion and the relevance of prenatal experience in influencing responsiveness to emotion were tested by examining newborn responses to the presentation of a range of vocal expressions (Mastropieri and Turkewitz, 1999). Differential responding was observed, indexed by an increase in eye opening behavior in response to the presentation of happy speech patterns. More importantly, differential responding was observed only when the infants listened to emotional speech spoken by speakers of their maternal language. Together, these findings demonstrate that infants’ auditory perception of another’s aversive affective state elicits a similar distressful emotional state in the self and suggests a neurobiologically based predisposition for humans to be connected to others. The emotional and physiological synchrony with others (mostly caregivers) allows for dynamic transactions on which intersubjectivity and empathy develop. Infants’ expressive and responsive behavior in response to the affects and emotions signaled by others serves as an instrument for social learning, reinforcing the significance of social exchanges, which then becomes associated with the infant’s own emotional experience.

22.3.2 Emotion recognition While the neural substrate for a few basic emotional expressions (interest, joy) seems functional in the early months of life, over the course of the first 2 years (Izard, 2007), evidence for the ability of newborns to discriminate and respond to different emotional facial expressions remains controversial (De Haan and Nelson, 1999). A recent longitudinal study finds no support for neonatal imitation of facial expressions and vocalizations (Oostenbroek et al., 2016). Moreover, it is not clear whether newborns can discriminate a fearful from a neutral facial expression, despite the fact that the fearful expression contains wide eyes and a semiopen mouth, features that may have enhanced the salience of a face (Farroni et al., 2007). However, when fearful expressions are compared with happy ones, newborns look significantly longer at the latter, demonstrating that newborns are able to discriminate happy from fearful expressions. Importantly, newborn’s facial musculature and cortical and subcortical motor pathways are fully developed and capable of actions that are morphologically identical to adults’ facial expressions (Muri, 2016), supporting the view that humans have biologically endowed neurophysiological readiness to communicate changes in emotional and motivational states. Evolution has prepared infants with innate action readiness patterns, which are crucial for early infantecaregiver social interaction, and through interpersonal exchanges with others, specific facial configurations acquire functional significance, becoming associated with specific emotions (Cole and More, 2015). From an ontogenetic perspective, babies’ facial expressions are the precursors of adult expressions, with communicational and adaptive biological functions that are crucial during the child’s first years, letting adults know what the baby needs (Oster, 2005). By 7e8 months of age, babies have learned how to express a wide variety of emotions, and around their first birthday, they become aware of not only other peoples’ expressions but also their actual emotional states, especially distress (Saarni, 2008). A recent longitudinal study examined individual variations in attentional bias to faces in 7-month-old infants by using a face-distractor competition paradigm and tested whether these variations were associated with outcomes reflecting social behavior at 24 and 48 months of age (i.e., spontaneous helping, emotion understanding, mentalizing, and callouse unemotional traits) (Peltola et al., 2018). The results showed a robust and distinct attention bias to faces at 7 months, particularly when faces were displaying a fearful expression. Variations in attention to faces at 7 months were not associated with emotion understanding or mentalizing abilities at 48 months of age, but increased attention to faces at

490 PART | II Cognitive development

7 months (regardless of facial expression) was related to more frequent helping responses at 24 months and reduced callouseunemotional traits at 48 months of age. These results are consistent with a model whereby increased attention to faces in infancy is linked with the development of affective empathy and responsivity to others’ needs.

22.3.3 Emotion understanding Although the capacity for two individuals to resonate with one another emotionally prior to any cognitive understanding is the basis for developing shared emotional meanings and primary intersubjectivity, it is not sufficient for mature empathic understanding. Such an understanding requires forming a representation, a knowledge of the emotions of self and others, which necessitates additional computational mechanisms beyond shared physiology (Decety and Moriguchi, 2007). The cognitive components that give way to empathic understanding have a more protracted course of development than the affective components, even though many precursors are already in place very early in life. Emotion understanding refers to the explicit knowledge about emotional processes or beliefs about how emotions work. Such understanding includes the recognition of emotional expressions and one’s knowledge about of one’s own and others’ emotions, the detection of cues for others’ feelings, as well as ways of intentionally using emotion expression to communicate to others (or vice versa, e.g., hiding emotions). Children’s development of gradually more sophisticated understandings of emotion fosters many adaptive processes, such as social functioning and coping. Consequently, delayed or limited emotion understanding may place youth at risk for emotional disorders. As discussed in the previous section, children recognize facial expressions of emotions at an early age (Haviland and Lewica, 1987). Most children are using emotion labels for facial expressions and are talking about emotion topics by the age of 2 years. Research shows that very young children (18- to 25-month-old) are able to sympathize with a victim of harm even in the absence of overt emotional cues (Vaish et al., 2009), suggesting some early form of affective perspectivetaking that does not rely on emotion contagion or affect sharing. With respect to the causes and effects of emotion and the cues used to infer emotion, developmental research has detailed a progression from situation-bound, behavioral explanations of emotion to broader, more mentalistic understanding (Harris et al., 1981). For example, children’s early explanations of emotion are largely based in the external world (e.g., ‘‘I am sad because someone took my toy”), whereas as children develop, their explanations of emotions focus more on internal causes (e.g., “I am sad because that toy was important to me”). With age, children’s emotional inferences contain a more complex and differentiated use of several types of information, such as moral variables, relational and contextual factors, and the target child’s goals or beliefs (Nunner-Winkler and Sodian, 1988). This development appears to be somewhat slower for complex social emotions such as pride, shame, or embarrassment (Lewis, 2000). Children also develop an understanding of multiple emotions, comprehending that a person can feel more than one emotion at a time. Development of this understanding proceeds from lack of acknowledgment of multiple emotions in younger children, to acknowledgment, and to an appreciation of different variables, such as emotion valence and emotion intensity (e.g., one very strong and one very weak emotion are easier to understand than two strong ones) (Carroll and Steward, 1984). Understanding that appraisal can modulate a person’s emotional experience to a given situation develops from being desire based to being belief based. At first, 2- and 3-year-old children understand the role that desires or goals play in determining a person’s appraisal and ensuing emotion (Repacholi and Gopnik, 1997). By 18 months, infants can not only infer that another person can hold a desire that may be different from their own but also recognize how desires are related to emotions and understand something about the subjectivity of these desires. By 4 and 5 years of age, this desire-based concept of emotion develops to include beliefs and expectations. Children at this age begin to understand that an emotion is not necessarily triggered by whether or not a desire and an observed outcome match, but rather whether a desire and an expected outcome match. The shift from a desire to a beliefedesire conception of mind and emotion is well established. Children’s references to other people’s mental states indicate that they talk systematically about desires and goals throughout their third year, but that beginning at about their third birthday, children begin to make reference to beliefs (Bartsch and Wellman, 1995).

22.3.4 Perspective-taking and theory of mind The developmental mechanism that is most frequently used to explain age-related increases in empathy and prosocial behavior is the ability to intentionally adopt the point of view of the other person. This cognitive component of empathy largely overlaps with the processes involved in mental state attribution (also known as theory of mind or ToM), which requires executive functions such as cognitive flexibility, inhibitory control, and working memory (Decety and Jackson, 2004). While relatively little is known about whether children who have a strong grasp of mental states also are advanced

A developmental neuroscience perspective on empathy Chapter | 22

491

in their understanding of emotions, emotion understanding and mental state understanding seem to engage common, as well as distinct, computational processes. Specifically, when seeing another child who is upset, a child has to hold two different perspectives in mind in order to correctly identify what that child is feeling and comfort themdtheir own perspective, which may not be congruent with that of the other child, and the point of view of the other child. There is indeed some evidence for a link between understanding of mental state and emotion. Several studies have shown that by around 4 years of age, children can appreciate that the emotion a person feels about a given event depends upon that person’s perception of the event and their beliefs and desires about it. A longitudinal study of 50 children aged 47e 60 months, examining developmental changes in understanding of false belief and emotion, in the context of mental state conversation with friends, reported that individual differences in understanding of both false belief and emotion were stable over this time period and were significantly related to one another (Hughes and Dunn, 1998). There is an intricate relation between children’s ToM, emotion understanding, and prosociality (Lane et al., 2010). A recent metaanalysis of 76 developmental studies found that ToM predicts prosocial behavior including comforting, helping, and cooperating in children aged 2e12 years (Imuta et al., 2016). Such associations, which increase with age, support the proposal that acting prosocially is influenced by a developing cognitive understanding of others’ perspectives and needs (Dunfield, 2014). A cross-sectional study of generosity across five countries (Canada, China, Turkey, South African, and the United States) found that second-order ToM ability and executive functions are the primary drivers of costly sharing in middle and late childhood (Cowell et al., 2017). Between the ages of 4 and 12 years, extensive development in domain-general cognitive capacities, such executive function, as well as ToM occurs, supported by the maturation of the medial prefrontal cortex and reciprocal connections with the posterior superior temporal sulcus, precuneus, and inferior frontal regions (Schurz et al., 2014).

22.3.4.1 Neurophysiological approaches to understanding cognitive empathy Neuroimaging studies that have examined the neural systems engaged during mental state understanding in adolescents and adults consistently identified a neural network involving the medial prefrontal cortex (mPFC), the posterior temporal cortex (pSTS) at the junction of the parietal cortex (TPJ), and the temporal poles (Brunet et al., 2000; Hillis, 2014). These regions are engaged in children aged 6e11 years while they listened to sections of a story describing a character’s thoughts compared with sections of the same story that described a physical context (Saxe et al., 2009). Furthermore, change in response selectivity with age was observed in the right TPJ, which was recruited equally for mental and physical facts about people in younger children, but only for mental facts in older children. Further support for age-related changes in brain activity associated with metacognition and ToM is provided by a neuroimaging investigation of ToM in participants whose age ranged between 9 and 16 years (Moriguchi et al., 2007). Both children and adolescents demonstrated significant activation in the neural circuits associated with mentalizing tasks, including the TPJ, the temporal poles, and the mPFC. Furthermore, the authors found a positive correlation between age and the degree of activation in the dorsal part of the mPFC. A milestone of human cognitive development, critical for empathy, is reached around the age of 4 years, when children begin to understand others’ false beliefs. Tract-based spatial statistics and probabilistic tractography in 3- and 4-year-old children show that the developmental breakthrough in false belief understanding is associated with age-related changes in local white matter structure in temporoparietal regions, precuneus, and mPFC and with increased dorsal white matter connectivity between temporoparietal and inferior frontal regions (Wiesmann et al., 2017). Notably, these effects are independent of codeveloping cognitive abilities including language and executive functions.

22.3.5 Emotion regulation While it is frequently assumed that an empathic response to another’s distress will automatically motivate prosocial behavior, the association between these constructs is in fact often modest (Eisenberg and Eggum, 2009). One putative reason for these weak associations is the influence of moderating variables. Eisenberg et al. (1994) proposed that individual differences in both the emotion reactivity and regulation capacities are related to an individual’s level of prosocial responding. Specifically, they suggested that the perception of distress in another leads to emotional arousal, but emotion regulation, i.e., the monitoring, evaluation, and modifying of emotional reactions, influences the subsequent goal-directed behavior, either to improve one’s own situation or help the situation of the other. Individuals who are able to optimally regulate their vicarious arousal, so that they do not experience excessive distress in the face of another person’s emotions and thus do not become self-focused, are proposed to behave prosocially and be prone to sympathy. Conversely,

492 PART | II Cognitive development

unregulated affect sharing can lead to personal distress, which can deter one’s inclination to act in prosocial ways. Moreover, individuals who are anxious and overregulated are thought to display withdrawal, which also inhibits prosocial behavior. Emotion regulation can include both implicit emotion regulation, i.e., processes, which occur automatically and largely outside conscious awareness at early stages of the emotion regulation process, and explicit emotion regulation, which involves using conscious strategies to modify emotional responses (Ahmed et al., 2015). Functional brain imaging research provides insight into the neural underpinnings of emotion regulation in childhood, adolescence, and adulthood by examining the effects of using specific instructed strategies to change emotional experience. Most of this work has focused on expressive suppression, i.e., actively inhibiting ongoing emotional expression, or cognitive reappraisal, a strategy that involves reframing or thinking differently about a stimulus to change one’s feelings. Extant research contends that cognitive reappraisal is a successful emotion regulation strategy, decreasing negative affect and resulting in an attenuation of autonomic arousal, whereas expressive suppression is thought to be a suboptimal strategy because it creates a conflict between heightened emotional arousal and overt expression of the arousal (Ochsner and Gross, 2005). These two types of strategies also appear to lead to different consequences for interpersonal functioning. Specifically, while cognitive reappraisal is positively related to having closer relationships and fewer depressive symptoms, expressive suppression is associated with greater experience of negative emotions, disturbed interpersonal interactions, and decreased prosocial tendencies (Ochsner and Gross, 2005). Of note, behavioral data in adults indicate that cognitive reappraisal may moderate the associations between affective empathy and prosocial behavior (Lockwood et al., 2014). Specifically, affective empathy was found to be positively associated with prosocial behavior, but only in participants at low and average, but not high, levels of cognitive reappraisal. This suggests that affective and cognitive components of empathic responding may have complementary, rather than additive, associations with prosocial behavior and interact in unique ways with emotion regulation tendencies. In line with these observations, the development of emotion regulation and other executive functions has been functionally linked to the development of mental state understanding. There is substantial evidence of a specific relation between the development of ToM and improved emotion regulation at around the age of 4 years (Carlson and Moses, 2001). This improvement in inhibitory control is related to increasing metacognitive abilities, as well as with maturation of neural regions and pathways, which underlie working memory and inhibitory control (Tamm et al., 2002). Emotion regulation taps into executive function resources implemented in the prefrontal cortex (Zelazo et al., 2008), with different areas subserving distinct functions. Ventral and dorsolateral regions of the prefrontal cortex have been associated with response inhibition and self-control, both key components of emotion regulation (Ochsner and Gross, 2005). Support for this hypothesis in relation to empathy comes from a study that compared hemodynamic responses in a group of physicians and a group of matched control participants when they viewed short video clips depicting hands and feet being pricked by a needle (painful situations) or being touched by a cotton bud (nonpainful situations) (Cheng et al., 2007). Unlike control participants, physicians showed a significantly reduced hemodynamic responses in the anterior insula, anterior cingulate cortex, and no activation of the periaqueductal gray (PAG) (a mediator of the flight or fight response) when viewing body parts being pricked by a needle. Instead, cortical regions underpinning executive functions and self-regulation in the dorsolateral and mPFC, and executive attention (precentral, superior parietal, and TPJ) were activated. Functional connectivity analyses further demonstrated that activation in the medial and dorsolateral prefrontal cortices was inversely coupled with hemodynamic activity in the insula in physicians, indicating active suppression of the emotional response to the others’ pain. It is well documented that the prefrontal cortex and its functions follow a significantly protracted developmental course, and age-related changes continue well into adolescence (Casey et al., 2005). Frontal lobe maturation is associated with an increase in a child’s ability to activate circuitry involved in emotional control and exercise inhibitory control over their thoughts, attention, and action. The maturation of the prefrontal cortex also allows children to use verbalizations to achieve self-regulation of their feelings. As a child matures into adolescence, there is a shift in response to emotional events from using more limbic-related anatomic structures, such as the amygdala, to using more frontal lobe regions to control emotional responses (Killgore and Yurgelun-Todd, 2007). Thus, not only may there be less neural activity related to the regulation of cognition and emotion in younger children, but the neural pattern itself is also likely to differ (Decety et al., 2012). Functional MRI studies suggest that the ability to downregulate negative emotion, decrease amygdala activation, and increase activity in lateral prefrontal regions tracks with age (Martin and Ochsner, 2016). Although children can use and describe cognitive strategies like reappraisal from as young as 5e6 years, the ability to use them well continues to develop throughout childhood and adolescence (Silvers et al., 2015). Studies of emotion regulation strategy use have provided evidence of age-related changes in lateral prefrontal cortex (LPFC) recruitment, amygdala modulation, and, more recently, the integration of these two components in the ventrolateral

A developmental neuroscience perspective on empathy Chapter | 22

493

PFCeamygdala pathway hypothesis (Silvers et al., 2017). Future work in this area can build on these neural patterns supporting emotion regulation to test additional aspects of regulatory skill beyond single strategies (e.g., examinations of flexible strategy use across an unfolding distressing interaction).

22.3.6 Motivation to care Empathic concern corresponds to the motivation to care for another’s welfare. This caring motivation arises from a set of biological mechanisms that are located in subcortical neural systems in the brain stem, hypothalamus, and connections with the mesolimbic dopamine system, including the ventral tegmental area, ventral striatum and ventromedial prefrontal cortex (vmPFC). Importantly, the vmPFC, a region reciprocally connected with ancient affective systems in brain stem, amygdala, and hypothalamus, across mammalian species, is a critical hub for caregiving behavior, particularly parenting through reward-based and affective associations. This dedicated neurobiological system was shaped by evolution for parental care and drives individuals to promote the well-being of others. This other-related and resource-related information also outputs to other motivational centers, including those that generate reward-based and avoidance-driven impulses that can conflict with caregiving motivation (Brown et al., 2011). This system has been linked to a range of affiliative emotions and prosocial behaviors, including charitable giving, trust, and empathic concern (Ashar et al., 2017). Mature empathic concern in humans has been theorized to derive from an affective response evoked by another conspecific’s emotional state, combined with emotion understanding and motivated by the desire to alleviate distress and suffering (Batson, 2012; see Decety et al., 2016, for a neurobehavioral perspective on similar motivational mechanisms across species). Prior work has shown that other-oriented concern starts to emerge at the beginning of the second year of life in conjunction with expressions of concern, efforts to alleviate distress, and an understanding that others have subjective states independent on their own (Moore, 2007). Some studies have documented early signs of empathic concern during the first year during which infants make pain attributions, but they also respond to a variety of distress cues, and they direct their comforting behavior in ways that are appropriate to the target’s distress (Roth-Hanania et al., 2011). Of note, expressions of concern have been documented for parents and other adults who exhibit pain or distress, but not other toddlers. Naturalistic studies of toddlers (9e27 months of age) in daycare settings found that they respond to their peer’s distress with concern or helping in only 3% of the time and are more likely to become distressed themselves rather than responding prosocially (Lamb and Zakhireh, 1997). Older toddlers (16e33 months of age) reacted to 22% of distress episodes by stopping play, looking at the peer and trying to intervene (Howes and Farver, 1987). A more recent, and controlled, study examined 12-, 18-, and 24-month-old children’s affective and behavioral responses to a crying infant (Nichols et al., 2015). Results showed that the younger children were neither interested in nor concerned about the infant crying but could discriminate between negative and neutral vocal emotion information. In contrast, the 18- and 24-montholds were socially interested and displayed greater concern for the crying infant. Overall, over the second year, children evidence both greater social interest in their peers’ emotional communication and increasingly distinct and appropriate behavioral responses. Contextual appraisal also plays a role in empathic responding early in development. This is particularly poignantly demonstrated in a study of 3-year-olds, who showed reduced empathic concern and prosocial behavior toward a “crybaby.” Toddlers exhibited less concern toward an individual who was mildly inconvenienced but showed exaggerated signs of distress, than toward a person who was distressed after being more seriously harmed (Hepach et al., 2013), suggesting that young children’s prosocial behaviors are not simply automatic responses to emotional displays, but rather involve factoring in whether the displayed distress is justified. In children aged 3e6 years, empathic concern leads to prosocial resource allocation by both promoting sharing and decreasing envy (Williams et al., 2014). As children grow up and become increasingly sophisticated social actors, they learn to regulate their empathy so that it is more likely to occur toward familiar, close, or deserving individuals (Vaish and Warneken, 2012). For instance, experiencing a natural disaster, like an earthquake, significantly affects children’s altruistic giving (Li et al., 2013). Overall, infants’ fundamental motivation for connectedness is clearly manifested by their interest and genuine concern for the other’s welfare.

22.4 Neurodevelopmental changes in empathic responding Perceiving other people in distress triggers empathy and provides a crucial signal motivating caring. In the past decade, numerous studies have used fMRI and EEG with children and adult participants being presented with stimuli depicting physical pain, emotional distress, and need. Across all these studies, perceiving signals of distress and need are associated with brain activation in a circumscribed number of regions including the thalamus, ACC, somatosensory cortex, and anterior insula, which partly overlap with the firsthand experience of emotional pain (Lamm et al., 2011). Importantly, an

494 PART | II Cognitive development

overlap between regions detected in fMRI does not mean that the same neurons are activated. Multivariate voxel pattern analyses have demonstrated that somatic pain and vicarious pain are represented by dissociable neural processes (Krishnan et al., 2016). This neural network detects danger and threat and can trigger protective and defensive behaviors.

22.4.1 Evidence from event-related potential EEG provides a noninvasive measure of electrical activity in the brain that is recorded via scalp electrodes. The EEG signal is the result of the summation of synchronous firing of postsynaptic neuronal potentials. Various measures of emotional functioning can be derived from the signal acquired from electrodes at specific scalp locations. The temporal resolution of EEG is one of the most notable benefits of this techniquedresponses are characterized very precisely, on the scale of milliseconds. In young children aged 3e9 years, EEG/ERP has been used to identify the neurophysiological markers of the development of empathy in response to visual stimuli depicting another person in physical pain (Cheng et al., 2014). Across early and middle childhood, an early automatic component (N200), reflecting attention to emotionally salient stimuli, and a late-positive potential (LPP), indexing cognitive reappraisal of emotional stimuli, were detected when perceiving the pain of others. The neuroanatomical localization of these ERP responses can be inferred from functional neuroimaging studies conducted with older children and adults (e.g., Cheng et al., 2008; Hohmeister et al., 2010) that converge on sources in the anterior midcingulate cortex (aMCC). This region, through its reciprocal connections with limbic and paralimbic structures (amygdala, nucleus accumbens, OFC, PAG, and autonomic brain stem motor nuclei), plays a central role in nociceptive processing, specifically in the motivational affective dimension of pain, which is associated with the preparation of behavioral responses to aversive events. Similarly, adolescents relative to young adults exhibited an earlier automatic component in response to another’s pain but greater late potentials (LPP) when perceiving the neutral stimuli, indicating the continued development of empathy during adolescence (Mella et al., 2012). Age-related changes in LPP are proposed to reflect the development of regulatory abilities during adolescence. To our knowledge, only one neuroscience study to date has examined the electrophysiological responses (EEG and ERPs) associated with perspective-taking and empathic concern in preschool children (Decety et al., 2018). Consistent with a body of previous studies using stimuli depicting somatic pain in both children and adults, larger early (w200 ms) ERPs were identified when perceiving painful versus neutral stimuli. In the slow wave window (w800 ms), a significant interaction of empathy condition and stimulus type was driven by a greater difference between painful and neutral images in the empathic concern condition. Across early development, children exhibited enhanced N2 to pain when engaging in empathic concern. Children’s own prosocial behavior was predicted by several individual differences in neural function, including larger early LPP responses during cognitive empathy and greater differentiation in late LPP and slow wave responses to empathic concern versus affective perspective-taking. In adult participants, empathic concern motivates costly altruism, and this relationship is predicted by activity in regions critical for promoting social attachment and caregiving, including activation in the vmPFC and ventral striatum (Ashar et al., 2017; Feldman Hall et al., 2015). To examine changes from childhood to adult functioning in empathetic responses, a magnetoencephalography study monitored the brain response of children, adolescents, and adults to visual stimuli depicting somatic pain (Levy et al., 2018). Results indicate that children’s vicarious empathy for pain operates via rudimentary sensory predictions involving alpha oscillations in somatosensory cortex, whereas adults’ responses recruit advanced mechanisms of updating sensory predictions and activating affective empathy in visceromotor cortex via higher-level representations involving beta- and gamma-band activity. These findings suggest that full-blown empathy to others’ pain emerges only in adulthood and involves a shift from sensory self-based to interoceptive other-focused mechanisms that support human altruism, maintain self-other differentiation, modulate feedback to monitor other’s state, and activate a plan of action to alleviate other’s suffering.

22.4.2 Evidence from functional magnetic resonance imaging As several key regions and neural pathways to experiencing and understanding emotions undergo considerable growth from childhood to adolescence, a few studies have captured continuous functional changes in this response across age. For example, one fMRI study examined age-related changes associated with reactivity to others in distress in participants from middle childhood to adulthood (Decety and Michalska, 2010). Results indicated that the younger the participants, the greater the signal response in the posterior insula and amygdala when they viewed clips depicting people in physical distress. The degree of activation in posterior insula was inversely correlated with age, whereas a positive correlation with age was found in the anterior insula. Results also showed that attending to accidentally caused painful situations was associated with activation of the so-called pain matrix, including the aMCC, insula, PAG, and somatosensory cortex.

A developmental neuroscience perspective on empathy Chapter | 22

495

Interestingly, when watching one person intentionally hurting another, regions that are consistently engaged in mental state understanding and affective evaluation (mPFC, pSTS, TPJ, and vmPFC) were also engaged. The younger the participants, the more strongly the amygdala, posterior insula, and SMA were recruited when they perceived painful situations that were accidentally caused. While participants’ subjective ratings of the painful situations decreased with age and were significantly correlated with hemodynamic response in the mPFC, increases in pain ratings were correlated with bilateral amygdala activation (Fig. 22.2). A significant inverse correlation between age and degree of activation was found in the posterior insula. In contrast, a positive correlation was found in the anterior portion of the insula. A posterior-to-anterior progression of increasingly complex rerepresentations in the human insula is thought to provide a foundation for the sequential integration of the individual homeostatic condition with one’s sensory environment and motivational condition (Craig, 2004). The posterior insula receives inputs from the ventromedial nucleus of the thalamus that is highly specialized to convey bodily and homeostatic information such as pain, temperature, hunger, thirst, itch, and cardiorespiratory activity. It serves as a primary sensory cortex for each of these distinct interoceptive feelings from the body (Jackson et al., 2006). On the other hand, activation in the insula correlates directly with subjective feelings from the body and with emotional feelings, thus positioning it as a hub for subjective emotional experience (Craig, 2004). In other words, what develops is children’s ability to compute a higher-order cognitive representation of bodily activity related to feelings of distress and aversive affect. Results from this study also showed that activation in the OFC/vmPFC in response to empathy-eliciting stimuli shifts from the engagement of the medial portion in young participants to the lateral portion in older participants. The medial OFC appears integral in guiding visceral and motor responses, whereas lateral OFC integrates the external sensory features of a stimulus with its impact on the homeostatic state of the body (Hurliman et al., 2005). Greater signal change with increasing age was associated with prefrontal regions that are responsible for cognitive control and response inhibition, such as the dorsolateral prefrontal cortex (dlPFC) and the inferior frontal gyrus (IFG). Indeed, the older the participants, the

FIGURE 22.2 On the left: Subjective ratings to dynamic visual stimuli depicting painful situations accidentally caused by self (Pain CS) and painful situations intentionally caused by another individual (Pain CO) across age (in weeks) in 57 participants (aged 7e40 years). A gradual decrease in the subjective evaluation of pain intensity for both painful conditions was found across age, with younger participants rating them significantly more painful than older participants (for Pain CS r ¼ 0.327, P < .01, for Pain CO r ¼ 0.267, P < .05). Furthermore, while on average participants rated the pain caused intentionally (Pain CO) conditions as significantly more painful than when pain was accidentally caused by self (Pain CS) (t(56) ¼ 2.581, P < .01), this effect was not driven by age. On the right: Significant negative correlation between age and degree of activation in the amygdala when the participants are perceiving painful situations accidentally caused. Adapted from Decety, J., Michalska, K. J., 2010. Neurodevelopmental changes in the circuits underlying empathy and sympathy from childhood to adulthood. Dev. Sci. 13 (6), 886e899.

496 PART | II Cognitive development

greater the activity in the left dlPFC and IFG, which are involved in cognitive control and response inhibition (Swick et al., 2008). This is in line with evidence that regulatory mechanisms continue to develop into late adolescence and early adulthood. In another cross-sectional study, children, adolescents, and adult participants (age range from 4 to 37 years) were scanned while presented with visual scenarios depicting intentional and accidental harm to others (Decety et al., 2012). Across development, several neural regions involved in the integration of affective, mental state understanding and cognitive processes, including the OFC/vmPFC, increased in activation and their functional coupling with the amygdala and pSTS during third-party perception of harm. Ratings of empathic sadness for the victim of intentional harm were predicted by the neurohemodynamic response in the anterior insula, thalamus, and vmPFC. These findings reflect the continued neural growth underlying the integration of more complex contextual cues with aversion to interpersonal harm, empathic concern, and mental state understanding to arrive at more mature interpersonal sensitivity. In sum, behavioral evaluations of pain intensity or harm to others and evoked patterns of brain activity from childhood to adulthood reflect a gradual change from a visceral emotional response critical for the analysis of the affective significance of stimuli to a more evaluative function.

22.5 Maladaptive alterations in developmental trajectories of empathy Several developmental disorders are associated with deficits in aspects of emotional and cognitive empathy, such as conduct disorder (CD) and autism. In this section, we review extant behavioral and neuroscience research that has informed understanding of atypical emotional responding in children with CD and autism, and we point out the gaps in current knowledge that can usefully be addressed with neurophysiological methods.

22.5.1 Conduct problems Conduct problems (CPs) in children include aggression, theft, bullying, and cruelty to others (American Psychiatric Association, 2013). They are one of the most common reasons for a childhood referral to mental health services and represent a substantial public health cost (Romeo et al., 2006). Children with CP are considerably more likely to engage in antisocial behavior in adulthood than typically developing children and are at risk for developing adult psychopathy. Atypical empathic processing is thought to be a risk factor for the development of psychopathy (Blair et al., 2014). The ability of individuals with psychopathy to manipulate and hurt others without concern for their well-being suggests an atypical empathic response to others’ distress. Although it is not appropriate to diagnose children with psychopathy, a subgroup of children with CP who also present with callouseunemotional (CU) traits present with atypical empathic responses to the suffering and distress of others. Such a pattern is similar to that seen in adults with psychopathy (Moul et al., 2015). Children with CP and high levels of CU (CP/CUþ) are characterized by a lack of regard for others, an inability to feel remorse, and a deficit in affective empathy, though their capacity for deliberate perspective-taking appears intact (Jones et al., 2010). The lack of emotional empathy in these children is an important risk factor for persistent antisocial behavior compared with both typically developing peers and peers with CP and lower levels of CU traits (CP/ CU) (Decety and Cowell, 2018). There have been considerable efforts in recent years to understand the neurocognitive basis of empathy problems observed in these children (Michalska et al., 2015; Michalska et al., 2016). Reduced autonomic nervous system reactivity to distress cues has been documented in children and adolescents with CP/CUþ (de Wied et al., 2012). Similarly, EEG/ERPs data indicate that incarcerated juvenile psychopaths with CP/CUþ, compared with youth with CP/CU, show a reduced frontal N120 in response to viewing people in pain, coupled with insensitivity to experienced pain, suggesting an absence of early affective arousal (Cheng et al., 2012). As with earlier behavioral work (Jones et al., 2010), the capacity of these CUþ participants to understand mental states was not impaired. Recent work in incarcerated adult offenders expands on these data, finding that psychopathic individuals are less likely to automatically represent the visual perspective of another agent despite a preserved ability to deliberatively take their perspective (Drayton et al., 2018) and exhibit an atypical pattern of brain activation and effective connectivity between the anterior insula and amygdala with the OFC vmPFC (Decety et al., 2013a,b). These data provide support for the idea that the maladaptive behavior of individuals with psychopathic or CU traits may result from attention and valuation dysfunctions that prioritize a goal-relevant perspective. They also lay the ground work for future research into delineating developmental dissociations between controlled and automatic ToM processes in the emergence of CP. Functional MRI studies have shown that children with CP, particularly those with CP/HCU, show atypical response to others’ pain and suffering in regions implicated in empathic processing (e.g., Lockwood et al., 2013; Marsh et al., 2013; Michalska et al., 2016; Yoder et al., 2016). For instance, Marsh et al. (2013) reported reduced amygdala and ACC

A developmental neuroscience perspective on empathy Chapter | 22

497

responses to stimuli of pain-inducing injuries in children with CP, with those with HCU reporting the most marked differences compared with typically developing children. Similarly, Lockwood et al. (2013) reported reduced AI and ACC responses to pictures of hands and feet in painful situations in boys with CP, with the degree of activation correlating negatively with levels of CU traits. Michalska et al. (2016) studied youth with CP when they viewed animated images of intentional and unintentional harm, finding that both CP and CU traits were inversely associated with AI and aMCC responses to intentional harm, whereas reactive aggressive traits and CU were positively associated. Moreover, sex differences emerged in the association between CP and functional neuroanatomy, suggesting early emergence of sexual dimorphism in correlates of aggressive behavior (Michalska et al., 2016). Furthermore, another report on the same sample and tasks showed that CUþ was associated with disrupted functional connectivity between ACC and AI, and ACC and amygdala (Yoder et al., 2016). The authors concluded that CUþ seems to be characterized by the disruption to cortical networks involved in detecting and appropriately responding to salient environmental cues, such as other people’s pain and distress. Notably, in an earlier fMRI study with adolescent male youth using a similar paradigm (Decety et al., 2009), adolescents with CP exhibited significantly greater activation in the amygdala and the striatum when viewing others in pain (Fig. 22.3). The extent of amygdala and activation to viewing pain in others was positively correlated with their number of aggressive acts and their subjective ratings of daring and sadism. A limitation of this earlier work was the small sample size. Yet, the role of sadism in the emergence of antisocial behavior is largely understudied in youth, and thus, replication of these associations is needed. Together, these data imply subgroups of individuals with early-onset CP and posit that empathy deficits in children with CP are heterogenous. Anomalies in responses to distress cues from others may be context dependent and modulated by attention, which explains the intriguing observation that the same brain area may be reported as either hypo- or hyperactive. In other words, there is not yet clear evidence for a particular area being persistently hypo- or hyperactive; the functional activation data associated with psychopathy seem to depend critically on the experimenters’ selection of task and stimuli (Koenigs et al., 2011). It is also imperative to consider contextual and familial moderators of any observed activations. For instance, preliminary work shows that amygdala responses to fear stimuli interact with prior trauma to predict CU traits. In this study, fear intensityemodulated amygdala responses negatively predicted CU traits for youth with low trauma, whereas

FIGURE 22.3 When youth with aggressive conduct disorder (CD) viewed individuals intentionally hurting another (like closing a piano lid), regions of the brain that process nociceptive information were activated (anterior insula, dorsal and ventral anterior cingulate cortex [ACC], somatosensory cortex, and periaqueductal gray [PAG]), as well as the amygdala and ventral striatum that are part of the neural circuit involved in reward processing. These latter regions were not engaged in healthy control adolescents. Dispositional ratings of daring and sadism in youth with CD were predicted by the hemodynamic response in amygdala and striatum (P < .005). These adolescents may enjoy seeing people in pain, and this may be rewarding and lead to repeated aggression. Adapted from Decety, J., Michalska, K. J., Akitsuki, Y., Lahey, B., 2009. Atypical empathic responses in adolescents with aggressive conduct disorder: a functional MRI investigation. Biol. Psychol., 80, 203e211.

498 PART | II Cognitive development

fear intensityemodulated amygdala responses positively predicted CU traits for youth with high trauma (Meffert et al., 2018). Adolescents with CD/CU, compared with controls, showed reduced response in the vmPFC during an emotion recognition task that included angry and fearful faces and a diminished response in the amygdala during an emotional resonance task (Klapwijk et al., 2016). As with functional neuroimaging studies, studies of brain structural variations that may be associated with CP in children are largely, though not always, consistent. Several studies report reduced amygdala and insula gray matter volume in adolescents with CD (Sterzer and Stadler, 2009). Other work does not replicate findings in these two regions but does replicate findings of reduced gray matter volumes in temporal regions, particularly in girls (Michalska et al., 2015). Studies that examined antisocial traits in children and adolescents have also consistently reported alterations in the uncinate fasciculus (Olson et al., 2015). This anatomical pathway serves as a critical link between structures that are implicated in several components of empathydparticularly between OFC/vmPFC, anterior insula, temporal pole, and amygdala. It is one of the last white matter tracts to reach its maturational peak with a developmental time course extending throughout adolescence and into young adulthood (Olson et al., 2015). Overall, youth with CP and CU traits show atypical patterns of hemodynamic response and effective connectivity in regions that are typically associated with processing distress cues from others as well as emotion recognition. Ensuring that scientific understanding of neurodevelopment in CP generalizes to both sexes and to children residing in different sociocultural strata is the crucial next step for work in this area.

22.5.2 Autism spectrum disorder Individuals with autism have an impaired understanding of mental states. They also appear to show problems with some facets of empathic responding (Jones et al., 2010). Autism spectrum disorder (ASD) refers to a neurodevelopmental disorder characterized by social interaction and communication problems and the presence of restrictive, repetitive behavioral patterns (American Psychiatric Association, 2013). Face emotion recognition precedes emotional understanding, and studies using pictures of facial emotional expressions suggest that abnormalities in emotion recognition may underlie some of the social difficulties associated with ASD. For instance, younger individual with ASD (12e16 years) demonstrated a tendency to require more time recognizing facial emotions than controls (Høyland et al., 2017). A metaanalysis of 13 fMRI studies including 226 individuals with ASD and 251 typically developing controls found ASD-related hyperactivation in subcortical structures, including bilateral thalamus, bilateral caudate, and right precuneus, and ASD-related hypoactivation in the hypothalamus during emotional face processing (Aoki et al., 2015). Subanalyses with more homogeneous contrasts confirmed the findings of the main analysis in particular abnormalities in the subcortical structures, such as amygdala, hypothalamus, and basal ganglia, which seems to be associated with atypical emotional face processing in individuals with ASD. Children and adults with ASD often show lower scores on cognitive perspective-taking and ToM using self-report measures, but they are no different from controls on affective empathy scale and frequently score higher than controls on personal distress (Rogers et al., 2007). Adolescents with ASD perceive themselves as having empathic concern but are not necessarily able to use it to support decision-making during challenging sociomoral situations, perhaps due to higher personal distress (Senland and Higgins-D’Alessandro, 2013). Initial work has examined affect sharing and social understanding in adolescents and young adults with ASD by combining pressure pain thresholds (PPTs) with fMRI, as well as EEG/ERP combined with eye tracking in response to empathy-eliciting stimuli depicting physical bodily injuries (Fan et al., 2014). Interestingly, results indicate that participants with ASD had lower PPT than controls. When viewing body parts being accidentally injured, increased hemodynamic response in the somatosensory cortex (SI/SII) but decreased response in the anterior midcingulate and anterior insula as well as heightened N2 but preserved LPPs were detected in ASD participants compared with controls. When viewing a person intentionally hurting another, decreased hemodynamic response in the medial prefrontal cortex and reduced LPP were observed in the ASD group. PPT was a mediator for the SI/ SII response in predicting subjective unpleasantness ratings to others’ pain. Both ASD and control groups had comparable mu suppression, indicative of typical sensorimotor resonance. Supporting earlier behavioral work (Jones et al., 2010), these findings demonstrate that, in addition to reduced pain thresholds, individuals with ASD exhibit heightened affect sharing but impair social understanding when perceiving others in distressful situations. Only a few fMRI studies to date have attempted to directly investigate whether activation in the brain areas responding to the perception of pain in others is altered in people with ASD. One such study used fMRI with short video clips of facial expressions of people experiencing pain to examine the neural response associated with the spontaneous empathic response in autism (Hadjikhani et al., 2014). No significant differences in brain activation between ASD individuals and controls were found during the perception of pain experienced by others. Both groups showed similar levels of activation in areas

A developmental neuroscience perspective on empathy Chapter | 22

499

associated with affect sharing such as the anterior insula and ACC. Differences between groups could be observed at a more liberal statistical threshold and revealed increased activations in areas involved in cognitive reappraisal in ASD participants compared with controls. Scores of emotional empathy were positively correlated with brain activation in areas involved in emotion of pain in the ASD group only. The fMRI study by Fan et al. (2014) found a hypersensitivity, the pain in individual with ASD, as well as greater activation in the somatosensory cortex in response to perceiving others in pain. These findings show that neural mechanisms involved in affect sharing are preserved in high-functioning individuals with autism and suggest that increased reappraisal may have a role in their apparent lack of caring behavior. Lastly, one study compared children with ASD and different subtypes of CP to describe disorder-specific empathy profiles in clinical samples (Schwenck et al., 2012). Empathy profiles showed differential impairment in children with ASD and CP/CUþ. Boys with ASD were impaired in cognitive empathy, whereas participants with CP/CUþ were impaired in emotional empathy. Children with CP/CU did not differ from controls. However, boys with CD/CU were less emotionally reactive in response to film stimuli than children with ASD. Together, these results suggest that affective empathy is intact, or even heightened in ASD, although as with CP, comparisons among subgroups (i.e., with and without anxiety) are needed to further elucidate the biobehavioral bases of empathic responding.

22.6 Conclusions Empathy is a natural ability that develops as a result of biological processes involving affect sharing, emotional regulation, mental state understanding, and cognitive abilities that are continuously interactive between an individual and her social environment. Empathy can thus be viewed both as an intrapersonal and an interpersonal process. Breaking down empathy into components and examining their neurodevelopment can contribute to a more complete model of interpersonal sensitivity (Decety and Jackson, 2004). Likewise, drawing from multiple sources of data and methods can improve our understanding of the nature and causes of empathy deficits in individuals with antisocial behavior disorders or atypical social cognitive abilities. Recent advances in developmental social neuroscience indicate that distinct but interacting brain circuits underpin the different components of empathy, each having their own developmental trajectory. Given the importance of empathy for healthy social interaction, it is clear that a developmental approach using functional neuroimaging to elucidate the computational mechanisms underlying affective reactivity, regulation, and behavioral outcomes is essential to complement traditional behavioral methods and gain a better understanding of how deficits may arise in the context of development. The link between empathy and prosocial behaviors is complex and is intertwined with social and motivational contingencies (Decety and Cowell, 2014b; Konrath and Grynberg, 2013). For instance, perceiving people in distress triggers a neural response associated with aversion. This response, in turn, may initiate helping or soothing behaviors motivated to both reduce one’s own discomfort, to feel good about oneself, and lessen another’s distress. Such behaviors may be reinforced by both endogenous reward (dopamine system) and positive social feedback from others. Empathy is strongly influenced by social categorization and group dynamics, and this is reflected at both the neurobiological level (changes in brain activation and neurohormone levels such as oxytocin) and behavioral outcomes (such as helping and caring) (Cikara and Van Bavel, 2014; De Dreu and Kret, 2016). Understanding how these biases in empathy develop in young children is an important endeavor for future research.

List of abbreviations ANS autonomic nervous system dACC dorsal anterior cingulate cortex aMCC anterior medial cingulate cortex CADS child and adolescent dispositions scale CD conduct disorder dlPFC dorsolateral prefrontal cortex EEG electroencephalography ERPs event-related potentials fMRI functional magnetic resonance imaging IFG inferior frontal gyrus mPFC medial prefrontal cortex OFC orbitofrontal cortex PAG periaqueductal gray SMA supplementary motor area

500 PART | II Cognitive development

SPL superior parietal lobule TPJ temporoparietal junction vmPFC ventromedial prefrontal cortex

References Ahmed, S.P., Bittencourt-Hewitt, A., Sebastian, C.L., 2015. Neurocognitive bases of emotion regulation development in adolescence. Dev. Cogn. Neurosci. 15, 11e25. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders, fifth ed. American Psychiatric Publishing, Arlington, VA. Aoki, Y., Cortese, S., Tansella, M., 2015. Neural bases of atypical emotional face processing in autism: a meta-analysis of fMRI studies. World J. Biol. Psychiatry 16 (5), 291e300. Ashar, Y.K., Andrews-Hanna, J.R., Dimidjian, S., Wager, T.D., 2017. Empathic care and distress: predictive brain markers and dissociable brain systems. Neuron 94 (6), 1263e1273. Bartsch, K., Wellman, H.M., 1995. Children Talk about the Mind. Oxford University Press, New York. Batson, C.D., 2012. The empathy-altruism hypothesis: issues and implications. In: Decety, J. (Ed.), Empathy e from Bench to Bedside. MIT Press, Cambridge, pp. 41e53. Blair, R.J.R., Blair, K.S., 2009. Empathy, morality, and social convention: evidence from the study of psychopathy and other psychiatric disorders. In: Decety, J., Ickes, W. (Eds.), The Social Neuroscience of Empathy. MIT press, Cambridge, pp. 139e152. Blair, R.J.R., Leibenluft, E., Pine, D.S., 2014. Conduct disorder and callouseunemotional traits in youth. N. Engl. J. Med. 371 (23), 2207e2216. Blasi, A., Mercure, E., Lloyd-Fox, S., Thomson, A., Brammer, M., Sauter, D., Deeley, Q., Barker, G.J., Renvall, V., Deoni, S., Gasston, D., Williams, S.C.R., Murphy, D.G.M., 2011. Early specialization for voice and emotion processing in the infant brain. Curr. Biol. 21 (14), 1220e1224. Brown, S.L., Brown, R.M., Preston, S.D., 2011. The human caregiving system: a neuroscience model of compassionate motivation and behavior. In: Brown, S.L., Brown, R.M., Penner, L.A. (Eds.), Moving beyond Self-Interest: Perspectives from Evolutionary Biology, Neuroscience, and the Social Sciences. Oxford University Press, New York, NY, pp. 75e88. Brunet, E., Sarfati, Y., Hardy-Bayle, M.C., Decety, J., 2000. A PET investigation of attribution of intentions to others with a non-verbal task. Neuroimage 11, 157e166. Carlson, S.M., Moses, L.J., 2001. Individual differences in inhibitory control and children’s theory of mind. Child Dev. 72, 1032e1053. Carroll, J.J., Steward, M.S., 1984. The role of cognitive development in children’s understandings of their own feelings. Dev. Psychol. 55, 1486e1492. Casey, B.J., Tottenham, N., Liston, C., Durston, S., 2005. Imaging the developing brain: what have we learned about cognitive development? Trends Cogn. Sci. 9, 104e110. Cheng, Y., Lin, C., Liu, H.L., Hsu, Y., Lim, K., Hung, D., Decety, J., 2007. Expertise modulates the perception of pain in others. Curr. Biol. 17, 1708e1713. Cheng, Y., Yang, C.Y., Ching-Po, L., Lee, P.L., Decety, J., 2008. The perception of pain in others suppresses somatosensory oscillations: a magnetoencephalography study. Neuroimage 40, 1833e1840. Cheng, Y., Lee, S., Chen, H.Y., Wang, P.Y., Decety, J., 2012. Voice and emotion processing in the human neonatal brain. J. Cogn. Neurosci. 24 (6), 1411e1419. Cheng, Y., Hung, A., Hung, D., Decety, J., 2012. Dissociation between affective sharing and emotion understanding in juvenile psychopaths. Dev. Psychopathol. 24 (2), 623e636. Cheng, Y., Chen, C., Decety, J., 2014. An EEG/ERP investigation of the development of empathy during early childhood. Developmental Cognitive Neuroscience 10, 160e169. Cikara, M., Van Bavel, J.J., 2014. The neuroscience of intergroup relations: an integrative review. Perspect. Psychol. Sci. 9 (3), 245e274. Cole, P.M., Moore, G.A., 2015. About face! Infant facial expression of emotion. Emotion Rev. 7 (2), 116e120. Cowell, J.M., Lee, K., Malcolm-Smith, S., Selcuk, B., Zhou, X., Decety, J., 2017. The development of generosity and moral cognition across five cultures. Dev. Sci. 20 (4), e12403. Craig, A.D., 2004. Human feelings: why are some more aware than others? Trends Cogn. Sci. 8, 239e241. De Dreu, C.K., Kret, M.E., 2016. Oxytocin conditions intergroup relations through upregulated in-group empathy, cooperation, conformity, and defense. Biol. Psychiatry 79 (3), 165e173. De Haan, M., Nelson, C.A., 1999. Brain activity differentiates face and object processing in 6-month-old infants. Dev. Psychol. 35 (4), 1113e1121. de Wied, M., van Boxtel, A., Matthys, W., Meeus, W., 2012. Verbal, facial and autonomic responses to empathy-eliciting film clips by disruptive male adolescents with high versus low callous-unemotional traits. J. Abnorm. Child Psychol. 40 (2), 211e223. Decety, J., 2015. The neural pathways, development and functions of empathy. Curr. Opin. Behav. Sci. 3, 1e6. Decety, J., Cowell, J.M., 2014. The complex relation between morality and empathy. Trends Cogn. Sci. 18 (7), 337e339. Decety, J., Cowell, J.M., 2014. Friends or foes: is empathy necessary for moral behavior? Perspect. Psychol. Sci. 9 (5), 525e537. Decety, J., Cowell, J.M., 2018. Interpersonal harm aversion as a necessary foundation for morality: a developmental neuroscience perspective. Dev. Psychopathol. 30, 153e164. Decety, J., Jackson, P.L., 2004. The functional architecture of human empathy. Behav. Cogn. Neurosci. Rev. 3 (2), 71e100. Decety, J., Lamm, C., 2009. Empathy versus personal distress - recent evidence from social neuroscience. In: Decety, J., Ickes, W. (Eds.), The Social Neuroscience of Empathy. MIT press, Cambridge, pp. 199e213.

A developmental neuroscience perspective on empathy Chapter | 22

501

Decety, J., Meyer, M., 2008. From emotion resonance to empathic understanding: a social developmental neuroscience account. Dev. Psychopathol. 20, 1053e1080. Decety, J., Michalska, K.J., 2010. Neurodevelopmental changes in the circuits underlying empathy and sympathy from childhood to adulthood. Dev. Sci. 13 (6), 886e899. Decety, J., Moriguchi, Y., 2007. The empathic brain and its dysfunction in psychiatric populations: implications for intervention across different clinical conditions. Biopsychosoc. Med. 1, 22e65. Decety, J., Svetlova, M., 2012. Putting together phylogenetic and ontogenetic perspectives on empathy. Dev. Cogn. Neurosci. 2, 1e24. Decety, J., Michalska, K.J., Akitsuki, Y., Lahey, B., 2009. Atypical empathic responses in adolescents with aggressive conduct disorder: a functional MRI investigation. Biol. Psychol. 80, 203e211. Decety, J., Michalska, K.J., Kinzler, K.D., 2012. The contribution of emotion and cognition to moral sensitivity: a neurodevelopmental study. Cerebr. Cortex 22, 209e220. Decety, J., Chen, C., Harenski, C., Kiehl, K.A., 2013. An fMRI study of affective perspective taking in individuals with psychopathy: imagining another in pain does not evoke empathy. Front. Hum. Neurosci. 7, 489. Decety, J., Skelly, L.R., Kiehl, K.A., 2013. Brain response to empathy-eliciting scenarios involving pain in incarcerated individuals with psychopathy. JAMA Psychiatry 70 (6), 638e645. Decety, J., Ben-Ami Bartal, I., Uzefovsky, F., Knafo-Noam, A., 2016. Empathy as a driver of prosocial behavior: highly conserved neurobehavioral mechanisms across species. Proc. Royal Soc. London Biol. 371, 20150077. Decety, J., Meidenbauer, K.L., Cowell, J.M., 2018. The development of cognitive empathy and concern in preschool children: a behavioral neuroscience investigation. Dev. Sci. 21 (3), e12570. Drayton, L.A., Santos, L.R., Baskin-Sommers, A., 2018. Psychopaths fail to automatically take the perspective of others. Proc. Natl. Acad. Sci. 115 (13), 3302e3307. Dunfield, K.A., 2014. A construct divided: prosocial behavior as helping, sharing, and comforting subtypes. Front. Psychol. 5, 958. Eisenberg, N., Eggum, N., 2009. Empathic responding: sympathy and personal distress. In: Decety, J., Ickes, W. (Eds.), The Social Neuroscience of Empathy. MIT Press, Cambridge, pp. 71e83. Eisenberg, N., Fabes, R.A., Murphy, B., Karbon, M., Maszk, P., Smith, M., O’Boyle, C., Suh, K., 1994. The relations of emotionality and regulation to dispositional and situational empathy-related responding. J. Personal. Soc. Psychol. 66, 776e797. Fan, Y.T., Chen, C., Chen, S.C., Decety, J., Cheng, Y., 2014. Empathic arousal and social understanding in individuals with autism: evidence from fMRI and ERP measurements. Soc. Cogn. Affect. Neurosci. 9 (8), 1203e1213. Farroni, T., Menon, E., Rigato, S., Johnson, M.H., 2007. The perception of facial expressions in newborns. Eur. J. Dev. Psychol. 4 (1), 2e13. Fawcett, C., Wesevich, V., Gredebäck, G., 2016. Pupillary contagion in infancy: evidence for spontaneous transfer of arousal. Psychol. Sci. 27 (7), 997e1003. Feldman, R., 2012. Parenteinfant synchrony: a biobehavioral model of mutual influences in the formation of affiliative bonds. Monogr. Soc. Res. Child Dev. 77 (2), 42e51. Feldman, R., 2016. The neurobiology of mammalian parenting and the biosocial context of human caregiving. Horm. Behav. 77, 3e17. Feldman-Hall, O., Dalgleish, T., Evans, D., Mobbs, D., 2015. Empathic concern drives costly altruism. Neuroimage 105, 347e356. Geangu, E., Benga, O., Stahl, D., Striano, T., 2010. Contagious crying beyond the first days of life. Infant Behav. Dev. 33 (3), 279e288. Geangu, E., Hauf, P., Bhardwaj, R., Bentz, W., 2011. Infant pupil diameter changes in response to others’ positive and negative emotions. PLoS One 6 (11), e27132. Hadjikhani, N., Zürcher, N.R., Rogier, O., Hippolyte, L., Lemonnier, E., Ruest, T., et al., 2014. Emotional contagion for pain is intact in autism spectrum disorders. Transl. Psychiatry 4 (1), e343. Harris, P.L., Olthof, T., Meerum-Terwogt, M., 1981. Children’s knowledge of emotion. J. Child Psychol. Psychiatry 22, 247e261. Haviland, J.M., Lewica, M., 1987. The induced affect response: ten-week old infants’ responses to three emotion expressions. Dev. Psychol. 23, 97e104. Hepach, R., Vaish, A., Tomasello, M., 2013. Young children sympathize less in response to unjustified emotional distress. Dev. Psychol. 49 (6), 1132e1138. Hillis, A.E., 2014. Inability to empathize: brain lesions that disrupt sharing and understanding another’s emotions. Brain 137 (4), 981e997. Hohmeister, J., Kroll, A., Wollgarten-Hadamek, I., Zohsel, K., Demirakca, S., Flor, H., Hermann, C., 2010. Cerebral processing of pain in school-aged children with neonatal nociceptive input: an exploratory fMRI study. Pain 150, 257e267. Howes, C., Farver, J., 1987. Toddler’s response to the distress of their peers. J. Appl. Dev. Psychol. 8, 441e452. Høyland, A.L., Nærland, T., Engstrøm, M., Lydersen, S., Andreassen, O.A., 2017. The relation between face-emotion recognition and social function in adolescents with autism spectrum disorders: a case control study. PLoS One 12 (10), e0186124. Hughes, C., Dunn, J., 1998. Understanding mind and emotion: longitudinal associations with mental state talk between young friends. Dev. Psychol. 34, 1026e1037. Hurliman, E., Nagode, J.C., Pardo, J.V., 2005. Double dissociation of exteroceptive and interoceptive feedback systems in the orbital and ventromedial prefrontal cortex of humans. J. Neurosci. 25, 4641e4648. Imuta, K., Henry, J.D., Slaughter, V., Selcuk, B., Ruffman, T., 2016. Theory of mind and prosocial behavior in childhood: a meta-analytic review. Dev. Psychol. 52 (8), 1192e1205. Izard, C.E., 2007. Basic emotions, natural kinds, emotion schemas, and a new paradigm. Perspect. Psychol. Sci. 2 (3), 260e280. Jackson, P.L., Rainville, P., Decety, J., 2006. To what extent do we share the pain of others? Insight from the neural bases of pain empathy. Pain 125, 5e9.

502 PART | II Cognitive development

Johnson, M.H., Griffin, R., Csibra, G., Halit, H., Farroni, T., De Haan, M., Tucker, L.A., Baron-Cohen, S., Richards, J., 2005. The emergence of the social brain network: evidence from typical and atypical development. Dev. Psychopathol. 17 (3), 599e619. Jones, A.P., Happé, F.G., Gilbert, F., Burnett, S., Viding, E., 2010. Feeling, caring, knowing: different types of empathy deficit in boys with psychopathic tendencies and autism spectrum disorder. J. Child Psychol. Psychiatry 51 (11), 1188e1197. Killgore, W.D.S., Yurgelun-Todd, D.A., 2007. Unconscious processing of facial affect in children and adolescents. Soc. Neurosci. 2, 28e47. Klapwijk, E.T., Aghajani, M., Colins, O.F., Marijnissen, G.M., Popma, A., van Lang, N.D., et al., 2016. Different brain responses during empathy in autism spectrum disorders versus conduct disorder and callous-unemotional traits. J. Child Psychol. Psychiatry 57 (6), 737e747. Knafo, A., Zahn-Waxler, C., Van Hulle, C., Robinson, J.L., Rhee, S.H., 2008. The developmental origins of a disposition toward empathy: genetic and environmental contributions. Emotion 8, 737e752. Koenigs, M., Baskin-Sommers, A., Zeier, J., Newman, J.P., 2011. Investigating the neural correlates of psychopathy: a critical review. Mol. Psychiatry 16 (8), 792. Konrath, S., Grynberg, D., 2013. The positive (and negative) psychology of empathy. In: Watt, D., Panksepp, J. (Eds.), The Neurobiology and Psychology of Empathy. Nova Science Publishers, Hauppauge, NY, pp. 63e107. Kraus, M.W., Côté, S., Keltner, D., 2010. Social class, contextualism, and empathic accuracy. Psychol. Sci. 21 (11), 1716e1723. Krishnan, A., Woo, C.W., Chang, L.J., Ruzic, L., Gu, X., López-Solà, M.,., Wager, T.D., 2016. Somatic and vicarious pain are represented by dissociable multivariate brain patterns. Elife 5, e15166. Lamb, S., Zakhireh, B., 1997. Toddlers’ attention to the distress of peers in a daycare setting. Early Educ. Dev. 8, 105e118. Lamm, C., Decety, J., Singer, T., 2011. Common and distinct neural networks associated with empathy for pain elicited by two experimental paradigms: an image-based meta-analysis. Neuroimage 54, 2492e2502. Lane, J.D., Wellman, H.M., Olson, S.L., LaBounty, J., Kerr, D.C., 2010. Theory of mind and emotion understanding predict moral development in early childhood. Br. J. Dev. Psychol. 28 (4), 871e889. Lester, B.M., Boukydis, C.F.Z., 1985. Infant Crying: Theoretical and Research Perspectives. Plenum Press, New York. Levy, J., Goldstein, A., Pratt, M., Feldman, R., 2018. Maturation of pain empathy from child to adult shifts from single to multiple neural rhythms to support interoceptive representations. Sci. Rep. 8 (1), 1810. Lewis, M., 2000. Self-conscious emotions: embarrassment, pride, shame, and guilt. In: Lewis, M., Haviland, J.M. (Eds.), Handbook of Emotions. Guilford Press, New York, pp. 623e636. Li, Y., Li, H., Decety, J., Lee, K., 2013. Experiencing a natural disaster alters children’s altruistic giving. Psychol. Sci. 24 (9), 1686e1695. Lockwood, P.L., Sebastian, C.L., McCrory, E.J., Hyde, Z.H., Gu, X., De Brito, S.A., Viding, E., 2013. Association of callous traits with reduced neural response to others’ pain in children with conduct problems. Curr. Biol. 23 (10), 901e905. Lockwood, P.L., Seara-Cardoso, A., Viding, E., 2014. Emotion regulation moderates the association between empathy and prosocial behavior. PLoS One 9 (5), e96555. Marsh, A.A., Finger, E.C., Fowler, K.A., Adalio, C.J., Jurkowitz, I.T., Schechter, J.C., et al., 2013. Empathic responsiveness in amygdala and anterior cingulate cortex in youths with psychopathic traits. J. Child Psychol. Psychiatry 54 (8), 900e910. Martin, G.B., Clark, R.D., 1987. Distress crying in neonates: species and peer specificity. Dev. Psychol. 18, 3e9. Martin, R.E., Ochsner, K.N., 2016. The neuroscience of emotion regulation development: implications for education. Curr. Opin. Behav. Sci. 10, 142e148. Mastropieri, D., Turkewitz, G., 1999. Prenatal experience and neonatal responsiveness to vocal expressions of emotion. Dev. Psychobiol. 35 (3), 204e214. Meffert, H., Thornton, L.C., Tyler, P.M., Botkin, M.L., Erway, A.K., Kolli, V., et al., 2018. Moderation of prior exposure to trauma on the inverse relationship between callous-unemotional traits and amygdala responses to fearful expressions: an exploratory study. Psychol. Med. 1e9. Mella, N., Studer, J., Gilet, A.L., Labouvie-Vief, G., 2012. Empathy for pain from adolescence through adulthood: an event-related brain potential study. Front. Psychol. 3, 501. Michalska, K.J., Decety, J., Zeffiro, T.A., Lahey, B.B., 2015. Association of regional gray matter volumes in the brain with disruptive behavior disorders in male and female children. Neuroimage: Clinical 7, 252e257. Michalska, K.J., Zeffiro, T.A., Decety, J., 2016. Brain response to viewing others being harmed in children with conduct disorder symptoms. J. Child Psychol. Psychiatry 57 (4), 510e519. Moore, C., 2007. Understanding self and others in the second year. In: Brownell, C.A., Kopp, C.B. (Eds.), Socioemotional Development in the Toddler Years: Transitions and Transformations. Guilford Press, New York, pp. 43e65. Moriguchi, Y., Ohnishi, T., Mori, T., Matsuda, H., Komaki, G., 2007. Changes of brain activity in the neural substrates for theory of mind in childhood and adolescence. Psychiatry Clin. Neurosci. 61, 355e363. Moul, C., Hawes, D., Dadds, M., 2015. The moral brain: Psychopathology. In: Decety, J., Wheatley, T. (Eds.), The Moral Brain e A Multidisciplinary Perspective. MIT Press, Cambridge MA, pp. 253e264. Müri, R.M., 2016. Cortical control of facial expression. J. Comp. Neurol. 524 (8), 1578e1585. Nichols, S.R., Svetlova, M., Brownell, C.A., 2015. Toddlers’ responses to infants’ negative emotions. Infancy 20 (1), 70e97. Nunner-Winkler, G., Sodian, B., 1988. Children’s understanding of moral emotions. Child Dev. 59, 1323e1338. Ochsner, K.N., Gross, J.J., 2005. The cognitive control of emotion. Trends Cogn. Sci. 9, 242e249. Olson, I.R., Von Der Heide, R.J., Alm, K.H., Vyas, G., 2015. Development of the uncinate fasciculus: implications for theory and developmental disorders. Dev. Cogn. Neurosci. 14, 50e61.

A developmental neuroscience perspective on empathy Chapter | 22

503

Oostenbroek, J., Suddendorf, T., Nielsen, M., Redshaw, J., Kennedy-Costantini, S., Davis, J., et al., 2016. Comprehensive longitudinal study challenges the existence of neonatal imitation in humans. Curr. Biol. 26 (10), 1334e1338. Oster, H., 2005. The repertoire of infant facial expressions: an ontogenetic perspective. In: Nadel, J., Muir, D. (Eds.), Emotional Development: Recent Research Advances. Oxford University Press, New York, NY, pp. 261e292. Peltola, M.J., Yrttiaho, S., Leppänen, J.M., 2018. Infants’ attention bias to faces as an early marker of social development. Dev. Sci. e12687. Prochazkova, E., Kret, M.E., 2017. Connecting minds and sharing emotions through mimicry: a neurocognitive model of emotional contagion. Neurosci. Biobehav. Rev. 80, 99e114. Prochazkova, E., Prochazkova, L., Giffin, M.R., Scholte, H.S., De Dreu, C.K., Kret, M.E., 2018. Pupil mimicry promotes trust through the theory-of-mind network. Proc. Natl. Acad. Sci. 115 (31), E7265eE7274, 201803916. Repacholi, B.M., Gopnik, A., 1997. Early reasoning about desires: evidence from 14- and 18-month-olds. Dev. Psychol. 33, 12e21. Rogers, K., Dziobek, I., Hassenstab, J., Wolf, O.T., Convit, A., 2007. Who cares? Revisiting empathy in Asperger syndrome. J. Autism Dev. Disord. 37 (4), 709e715. Romeo, R., Knapp, M., Scott, S., 2006. Economic cost of severe antisocial behaviour in childreneand who pays it. Br. J. Psychiatry 188 (6), 547e553. Roth-Hanania, R., Davidov, M., Zahn-Waxler, C., 2011. Empathy development from 8 to 16 months: early signs of concern for others. Infant Behav. Dev. 34 (3), 447e458. Saarni, C., 2008. The interface of emotional development with social context. In: Lewis, M., Haviland-Jones, J., Feldman Barrett, L. (Eds.), The Handbook of Emotions, third ed. Guilford Press, New York, pp. 332e347. Saxe, R.R., Whitfield-Gabrieli, S., Pelphrey, K.A., Scholz, J., 2009. Brain regions for perceiving and reasoning about other people in school-aged children. Child Dev. 80, 1197e1209. Schurz, M., Radua, J., Aichhorn, M., Richlan, F., Perner, J., 2014. Fractionating theory of mind: a meta-analysis of functional brain imaging studies. Neurosci. Biobehav. Rev. 42, 9e34. Schwenck, C., Mergenthaler, J., Keller, K., Zech, J., Salehi, S., Taurines, R.,., Freitag, C.M., 2012. Empathy in children with autism and conduct disorder: group-specific profiles and developmental aspects. J. Child Psychol. Psychiatry 53 (6), 651e659. Senland, A.K., Higgins-D’Alessandro, A., 2013. Moral reasoning and empathy in adolescents with autism spectrum disorder: implications for moral education. J. Moral Educ. 42 (2), 209e223. Shdo, S.M., Ranasinghe, K.G., Gola, K.A., Mielke, C.J., Sukhanov, P.V., Miller, B.L., Rankin, K.P., 2018. Deconstructing empathy: neuroanatomical dissociations between affect sharing and prosocial motivation using a patient lesion model. Neuropsychologia 116, 126e135. Silvers, J.A., Shu, J., Hubbard, A.D., Weber, J., Ochsner, K.N., 2015. Concurrent and lasting effects of emotion regulation on amygdala response in adolescence and young adulthood. Dev. Sci. 18 (5), 771e784. Silvers, J.A., Insel, C., Powers, A., Franz, P., Helion, C., Martin, R.E., et al., 2017. vlPFCevmPFCeamygdala interactions underlie age-related differences in cognitive regulation of emotion. Cerebr. Cortex 27 (7), 3502e3514. Stellar, J.E., Manzo, V.M., Kraus, M.W., Keltner, D., 2012. Class and compassion: socioeconomic factors predict responses to suffering. Emotion 12 (3), 449. Sterzer, P., Stadler, C., 2009. Neuroimaging of aggressive and violent behaviour in children and adolescents. Front. Behav. Neurosci. 3, 35. Swick, D., Ashley, V., Turken, A.U., 2008. Left inferior frontal gyrus is critical for response inhibition. BMC Neurosci. 9, 102e. Tamm, L., Menon, V., Reiss, A.L., 2002. Maturation of brain function associated with response inhibition. J. Am. Child. Adoles. Psychiatry 41, 1231e1238. Vaish, A., Warneken, F., 2012. Social-cognitive contributors to young children’s empathic and prosocial behavior. In: Decety, J. (Ed.), Empathy: From Bench to Bedside. MIT Press, Cambridge, pp. 131e146. Vaish, A., Carpenter, M., Tomasello, M., 2009. Sympathy through affective perspective-taking, and its relation to prosocial behavior in toddlers. Dev. Psychol. 45, 534e543. Wiesmann, C.G., Schreiber, J., Singer, T., Steinbeis, N., Friederici, A.D., 2017. White matter maturation is associated with the emergence of Theory of Mind in early childhood. Nat. Commun. 8, 14692. Williams, A., O’Driscoll, K., Moore, C., 2014. The influence of empathic concern on prosocial behavior in children. Front. Psychol. 5, 425. Yoder, K.J., Lahey, B.B., Decety, J., 2016. Callous traits in children with and without conduct problems predict reduced connectivity when viewing harm to others. Sci. Rep. 6, 20216. Zelazo, P., Carlson, S., Kesek, A., 2008. The development of executive function in childhood. In: Nelson, C.A., Luciana, M. (Eds.), Handbook of Developmental Cognitive Neuroscience. MIT Press, Cambridge, pp. 553e574.

Chapter 23

Developing attention and self-regulation in infancy and childhood M.R. Rueda1, 2 and A. Conejero2 1

Department of Experimental Psychology, University of Granada, Granada, Spain; 2Mind, Brain and Behavior Research Center (CIMCYC),

University of Granada, Granada, Spain

Chapter outline 23.1. Introduction 23.2. Facets of attention 23.2.1. Attention and self-regulation 23.3. Brain networks 23.3.1. Alerting 23.3.2. Orienting 23.3.3. Executive attention 23.4. Development of brain and behavior 23.4.1. Infancy 23.4.2. Toddlerhood

505 506 506 507 507 507 508 509 509 510

23.4.3. Childhood 23.5. Individual differences 23.5.1. Temperament 23.5.2. Genes 23.5.3. Environment 23.6. Plasticity of attention networks 23.7. Summary and integration Acknowledgments References

511 513 513 514 515 516 517 517 517

23.1 Introduction Attention is an aspect of cognition involved in most of our daily life activities. Mostly from the second half of the 20th century on, research in the field of cognitive psychology has provided precise experimental methods to measure the different processes involved in this multidimensional construct. Aspects of alertness, orienting and executive control have been differentiated at the cognitive and neural levels. The study of the neural mechanisms of attention has greatly benefited from the impressive technological developments that happened in the past decades, which allow the examination of a wide range of brain processes in living individuals. In this chapter, we discuss information available on the particular brain circuitry associated with attention and self-regulation, as well as neurochemicals modulating this anatomy. Information about the neural mechanisms of this central aspect of cognition is key to an understanding of individual differences in attentional efficiency. An important source of variation in efficiency is the level of maturation of the system. Developmental studies provide valuable information on the cognitive and neural changes that occur with age. In turn, this evidence informs the possible mechanisms underlying differences in competence across individuals over and above age, which can be related to genetic as well as environmental factors or the interactions of both. Finally, in recent years, an increasing number of studies have examined the impact of cognitive training programs of different nature on behavioral and neural measures, providing compelling evidence on the plastic nature of the brain. Although much research is needed before we fully understand processes of cognitive and brain plasticity following intervention, undoubtedly this research will inform the best possible strategies to promote children’s ability to regulate thoughts and behavior.

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00023-8 Copyright © 2020 Elsevier Inc. All rights reserved.

505

506 PART | II Cognitive development

23.2 Facets of attention Just as the structure of our body imposes limits to the number of things we can interact with at a given time, our mind is limited in the amount of information it can consciously process at a time. However, our environment is complex and provides an enormous amount of stimulation at any given time. In this complex world, attention serves as the interface between all the stimulation reaching our senses and the more limited set of information we are able to consciously process at a time. In this sense, attention is a selection mechanism that serves to choose a particular source of stimulation for priority processing and is closely connected to consciousness. On the other hand, attention has been largely linked to the voluntary and effortful control of action, as opposed to well-learned automatic behavior. Most of the actions we do on a daily basis are effortless and unfold in an automatic nonsupervised way. For instance, we can perform a quite complex motoric action such as walking upstairs, whereas our attention is focused on a conversation we are having with a friend. Automatic actions do not require attention control. However, in certain situations, attention is necessary to supervise goaldirected action. These are situations that involve overcoming an automatic course of action and detecting the need to do so. Also, attention is necessary for detecting errors and controlling behavior in dangerous and novel or unpracticed conditions (Norman and Shallice, 1986). Thus, attention mechanisms are also central to the generation of voluntary behavior, which often involves inhibition of automatic response tendencies. Finally, attention also demands an optimal level of activation. Efficiency of attention is greatly affected by conditions in which our level of activation is compromised, such as fatigue or drowsiness. These three facets of attention are represented in Fig. 23.1. Hence, attention can be defined as a multidimensional construct that refers to a state in which we have an optimal level of activation that allows selecting the information we want to prioritize to control the course of our actions. In fact, aspects of activation, selection, and control are represented in the construct of attention in Posner’s theoretical model, which are, respectively, referred to as alerting, orienting, and executive attention (Petersen and Posner, 2012; Posner and Petersen, 1990). Moreover, the attentive state can be primarily driven from external stimulation or be under the voluntary control of the individual. Novel, distinctive, and relevant stimulation capture attention automatically. This type of attention is referred to as bottom-up, stimulus-driven, or exogenous attention. On the other hand, attention can be voluntarily controlled and deployed according to internally represented goals and intentions. This second type of attention control is referred to as top-down, goal-directed, or endogenous attention (Corbetta and Shulman, 2002).

23.2.1 Attention and self-regulation The content of the mind, at least the part of the mind involved in making decisions and planning behavior according to internal goals, is shaped by the information we experience at any given time. Attention is the mechanism that regulates the flow of information within this mental working space. In William James’ words, “My experience is what I agree to attend to. Only those items that I notice shape my mind e without selective interest [.] the consciousness of every creature would be a gray chaotic indiscriminateness, impossible for us even to conceive” (James, 1890, pp. 402e403). Relevant classical models of attention have emphasized its role in filtering out irrelevant information (Broadbent, 1958) and administering cognitive resources among relevant tasks (Kahneman, 1973), to maximize the efficient processing of information in and out of the mental working space of individuals (see Fig. 23.1). The attention system is thus in charge of supervising that the flow of information in the conscious working space is tuned in to the current (or future) intentions of the individual. Therefore, attention is the primary mechanism for self-regulating thoughts and actions.

FIGURE 23.1 Representation of the different facets of attention.

Developing attention and self-regulation in infancy and childhood Chapter | 23

507

The study of attention and self-regulation arose from different research traditions. Attention has been a prolific area in experimental psychology since the second half of the 20th century and is currently being studied at many levels in cognitive neuroscience. Self-regulation, most commonly with respect to emotional control, has also been a central topic in developmental psychology and in child and family studies (Vohs and Baumeister, 2011). However, in the past decades, efforts have been made to integrate these two constructs and to investigate the role individual differences in attentional efficiency play in the development of self-regulation (Posner and Rothbart, 2000; Rueda et al., 2011). Studying attention within the framework of cognitive neuroscience provides information about the brain underpinnings supporting individual differences in attention and self-regulation. In the next sections of this chapter, we first delineate the attentional networks of the human brain, derived mostly from adult studies. Then, we describe brain changes that occur between infancy and adulthood and seek to link them to developmental differences in behavior.

23.3 Brain networks Imaging studies have suggested that at least three somewhat independent networks are involved in different aspects of attention, carrying out the functions of alerting, orienting, and executive attention (Petersen and Posner, 2012).

23.3.1 Alerting Arousal of the central nervous system involves input from brain stem systems that modulate activation of the cortex. Primary among these is the locus coeruleus (LC), which is the source of the brain’s norepinephrine (NE). We know that drugs that block NE prevent the changes in the alert state that lead to improved performance after a warning signal is provided (Coull et al., 2001). It has been demonstrated that the influence of warning signals operates via this LC-NE system, which exhibits phasic and tonic modes of activity (Aston-Jones and Cohen, 2005). In the phasic mode, the system reacts to warning signals in a short timescale by facilitating decision processes that optimize performance in a particular task. In the tonic mode of activation, the system optimizes performance across tasks promoting a more exploratory mode of alertness. To monitor for task-related utility, the LC is prominently connected to the anterior cingulate and orbitofrontal cortices, structures that are involved in representing action goals and intentions. Event-related potentials (ERPs) provide precise information about the time in which the brain responds differently to presence/absence of alerting cues. The presence of alerting cues produces dramatic changes in brain activation early on, followed by a sustained negative ERP component called the contingent-negative variation (CNV). Several studies have shown that the CNV is generated by activation in the frontal lobe, with specific regions depending on the type of task being used (Cui et al., 2000). When using fixed cue-target intervals, warning signals are informative of when a target will occur, thus producing a preparation in the time domain. Under these conditions, warning cues appear to activate frontoparietal structures in the left hemisphere, instead of the right (Coull et al., 2000).

23.3.2 Orienting The attention system achieves selection by modulating the functioning of sensory systems. Studies using fMRI and cellular recording have demonstrated that brain areas that are activated by attention cues, such as the superior parietal lobe and temporal parietal junction, play a key role in modulating activity within primary and extrastriate visual systems when attentional orienting occurs (Corbetta and Shulman, 2002; Desimone and Duncan, 1995). Studies that combine neuroimaging techniques with attention orienting cues have led to the identification of two different brain networks involved in endogenous and exogenous selective attention. The two networks are distinctively activated (1) when focusing attention voluntarily using top-down control mechanisms or (2) when exogenous and relevant stimuli appear in the environment inducing reorienting of attention according to task demands. In the first case, activation is registered in a bilateral dorsalefrontoparietal network that involves the intraparietal sulcus (IPS), the superior parietal lobule (SPL), and the frontal eye fields (FEF). In the second case, detection of infrequent or miscued targets has been related to increased activation in a right-lateralized network of ventral frontoparietal structures including the temporoparietal junction (TPJ) and inferior frontal cortex (Corbetta et al., 2008; see Fig. 23.2). Additionally, analyses of spontaneous fluctuations in blood oxygenation (BOLD signal) at rest provide a measure of functional connectivity, because regions that are functionally connected show correlated fluctuations in the bold signal over time. Studies using this method have revealed that the two attention systems are clearly segregated and exhibit only a small overlapping region in the prefrontal cortex (Fox et al., 2006). At the functional level, temporal differences in the neural

508 PART | II Cognitive development

FIGURE 23.2 Main regions of the three attention networks. Regions of the alerting network (red): thalamus, superior parietal and frontal cortex; regions of the orienting network (green): intraparietal sulcus (IPS), inferior parietal lobe (IPL), temporoparietal junction (TPJ), frontal eye fields (FEFs), and areas of the thalamus. Regions of the executive attention network (blue): dorsal division of the anterior cingulate cortex (dACC), precuneus, and dorsolateral and anterior regions of the prefrontal cortex (PFC). Combined striped colors indicate regions thar are common to different networks.

activation of the two networks have also been reported. Electrode recordings in monkeys have demonstrated that activation of structures within the dorsal network reaches significance levels before ventral structures when monkeys perform a visual search task (endogenous orienting), whereas activation of structures in the right-lateralized ventral network reaches significance levels earlier under pop-out (exogenous orienting) conditions (Buschman and Miller, 2007).

23.3.3 Executive attention A frequently used method to study executive control in the lab consists of inducing conflict between responses by instructing people to execute a subdominant response while suppressing a dominant tendency. Conflict-inducing tasks vary in the stage of information processing when interference is induced (see Fig. 23.3). For instance, in the flanker task,

FIGURE 23.3

Marker tasks of executive attention.

Developing attention and self-regulation in infancy and childhood Chapter | 23

509

interference is produced by presenting distracting stimuli that compete with the target at the perceptual level. In the Stroop task, the dominant (word reading) and required (color naming) responses are brought about by the very same stimulus; thus, interference is induced at the level of response selection. Finally, in Go/NoGo tasks, the rapid responses to frequent Go stimuli interfere with the need to hold the response to infrequent NoGo stimuli, in which cases control is mostly operating at the level of response execution. In the different types of conflict-inducing tasks, inhibition is necessary to withhold the dominant incorrect response and develop the appropriate one. We know from numerous neuroimaging studies that diverse conflict tasks show a common node of activation in the anterior cingulate cortex (ACC) together with other regions of the lateral prefrontal cortex (Botvinick et al., 1999; Fan et al., 2003). Different parts of the ACC are well connected to a variety of other brain regions, including limbic structures as well as parietal and frontal areas. Studies have examined the connectivity of the executive network at rest and have shown that two functionally different, but complementary circuits are engaged when implementing attentional control: the frontoparietal and the cinguloopercular networks (Dosenbach et al., 2007). The frontoparietal network is related to processing of cognitive control signals that potentially initiate response adjustments on a trial-by-trial basis. This network includes the dorsolateral prefrontal cortex (dlPFC), inferior parietal lobe (IPL), dorsal frontal cortex (dFC), intraparietal sulcus (IPS), precuneous, and middle cingulate cortex (mCC). On the other hand, the cinguloopercular network is involved in maintaining a stable task set during performance, that is, representing the goal of the individual in the context of the task and the corresponding stimulus-to-response mapping over many trials. This network includes the anterior prefrontal cortex (aPFC), anterior insula/frontal operculum (aI/fO), dorsal anterior cingulate cortex/medial superior frontal cortex (dACC/ msFC), and the thalamus (Dosenbach et al., 2008) (see Fig. 23.2). Studies using electroencephalography also inform about the temporal dynamics of conflict processing. Attention to a target, particularly in the presence of conflict, modulates the N2 and P3a potentials, two midline frontoparietal eventrelated potentials that peak around 200e400 and 300e500 ms, respectively, after the presentation of the target stimulus (Kopp et al., 1996; Polich, 2007). The amplitude of these potentials increases in conflict trials, a signal of the greater attention effort needed. In particular, the N2 is thought to be associated with suppressing the processing of the dominant but incorrect response. This effect has been linked to activation originating in the ACC (Veen and Carter, 2002).

23.4 Development of brain and behavior The three attention networks described earlier develop rapidly during infancy, childhood, and, to a lesser extent, also during adolescence. In the following sections, we present the main changes in attention that occur during infancy, toddlerhood, and childhood.

23.4.1 Infancy Infants are ready to react to the environment from the very moment they are born. The alert system (involving subcortical structures such as the brain stem) is functional enough at birth, so that the newborn infant is sensitive to physical changes in the environment (Arditi et al., 2006; Trapanotto et al., 2004). Likewise, the ability to maintain optimal arousal levels develops rapidly during infancy. Sleep dominates at birth, and the waking state is relatively rare in the first weeks after birth. In fact, the newborn infant dedicates most of the time to sleep, sleeping about three quarters of the day (Colombo and Horowitz, 1987). By 3 months of age, infants’ awake time increases from the initial 4e6 h a day (Figueiredo et al., 2016). Sleepeawake patterns also become increasingly regular. At the beginning, infants’ alert state is mainly modulated by external stimulation, much of it provided by caregivers. Changes in sustained attention during infancy differ depending on the complexity of stimulation. As young infants grow, they lose interest in simple stimuli such as geometric figures or static images of human faces. Conversely, the time infants spend looking at more complex stimuli, such as dynamic video clips, increases exponentially (Courage et al., 2006). Recent studies further suggest that phasic changes in arousal levels influence individual patterns of sustained attention in the first months of life. Increases in arousal level relate to infants’ shorter look times to objects, whereas low arousal levels relate to longer looking times (de Barbaro et al., 2017). Thus, alert mechanisms may serve as a primitive way of modulating attention focus before executive attention mechanisms are sufficiently developed. High arousal levels may act by enhancing sensitivity to incoming inputs, whereas low arousal may help to focus attention longer in a certain stimulus by reducing reactivity to potential distractors. The ability of infants to orient attention is also evident in the first months of life. Eye-tracking technology allows us to study the development of this function at these early stages of development. Newborns direct their head and gaze toward new stimuli. They show a preference to certain kind of stimuli, such as faces or moving objects (Courage et al., 2006). However, disengaging attention is a hard task before the third month of life. Younger infants find very difficult to move

510 PART | II Cognitive development

attention away from the stimulus they are currently focused on (even after stimulus has become boring or uninteresting) and redirect attention to a new one (Hood, 1995). The superior colliculus (a subcortical thalamic structure) has been proposed to control the shifting of visual attention at first (Johnson, 1990). As cortical structures of the orienting system develop (e.g., parietal cortex and frontal eye fields), infants become increasingly able to disengage attention from a central stimulus, which occurs at around 3 months of age (Atkinson et al., 1992; Johnson et al., 1991), significantly diminishing the time needed to disengage between 4 and 6 months of age (Hunnius et al., 2006). Also, infants respond to orienting cues from about 4 months of age. At this stage, infants orienting of attention to the location of a target stimulus speed up after presenting a valid cue, same as adults do (Johnson and Tucker, 1996; Johnson, 1994). In the first months of life, orienting of attention mainly relies on the characteristics of external stimulation (e.g., saliency) and is controlled by external events with a predominant role of caregivers. Thus, infants automatically shift their attention toward objects with attractive colors that make funny sounds or what caregivers encourage them to look at. First signs of endogenous orienting can be observed at about 4 months of age. After presenting a repeated sequence of events, 4-month-olds are able to anticipate gaze toward the expected location in which stimuli will appear (Canfield and Haith, 1991; Clohessy et al., 2001). This endogenous control of orienting antecedes the more flexible voluntary control achieved by the executive attention network. Young infants are only able to learn simple unambiguous sequences of events (e.g., stimuli appearing at location 1, then location 2, then 1, 2, 1, 2, etc.). Learning more complex sequences of events including additional locations (e.g., 1, 2, 1, 3, 1, 2, 1, 3, etc.) is not possible until toddlerhood as it requires monitoring the context, and therefore, the maturation of lateral prefrontal cortex (Clohessy et al., 2001). Remarkably, a certain degree of functional connectivity within the structures of the executive attention network is already observable in infants (Doria et al., 2010). However, the earliest behavioral signs of executive attention are not evident until the end of the first year of life. At about 9 months of age, infants are able to inhibit attention to irrelevant distractors and remain focused on an interesting stimulus (Holmboe et al., 2008). At this age, babies are also able to pass the easier versions of the A-not B task, a task that requires the ability to inhibit searching for a toy in the location that was initially hidden to successfully retrieve the toy from the location where it was newly hidden (Diamond, 1990). In fact, infants’ ability to ignore distractors and performance in the A-not B task has been found to be intercorrelated (Holmboe et al., 2018). This finding points to a common underlying neural system, presumably the executive attention network, which activates for inhibitory control. Studies including electrophysiological measures of brain activity give further support to the idea that executive attention mechanisms come into play at this early age. Detecting errors and reacting to the violation of expectations is a function attributed to the ACC, a principal node in the executive attention network. By 9 months of age, infants’ brains react to errors and unexpected events in a similar way as adults and older children (Berger et al., 2006; Reid et al., 2009). However, during infancy, executive attention is still quite rudimentary, and the regulation of behavior develops over the following years thanks to the further maturation of the executive attention network.

23.4.2 Toddlerhood During the second and third year of life, substantial development of attention neural networks takes place. The myelination process experiences a maximum peak about the second year of life (Dean et al., 2014; Deoni et al., 2015). These structural changes in white matter bring about the enhancement of neural connections, which become more efficient and reorganized into more modular networks (Hagmann et al., 2010). Changes in toddlers’ brains translate into behaviorally observable changes in toddlers’ attention abilities. Time toddlers can spend focused on a certain cognitive task increases between 17 and 24 months of age and is related to better performance on such a cognitive task (Choudhury and Gorman, 2000). However, 2-year-olds’ performance in vigilance tasks is still poor, missing the target stimuli about 60% of the times (Silverman and Gaines, 1996). The advances of toddlers in sustaining attention are also observable in more natural situations such as during free play. Compared with toddlers, infants get bored of toys relatively quickly. Whereas 2-year-olds are engaged with the toys the whole time, infants’ attention to toys progressively declines during play time (Ruff and Lawson, 1990). Evidence of the relative separation of alerting, orienting, and executive aspects of attention emerges at about 18 months of age (de Jong et al., 2016). Among all attention functions, executive attention undergoes the greatest development during this period. Projections connecting thalamus, cingulate cortex, and prefrontal structures (mostly overlapping with executive attention network) grow and consolidate as a differentiated neural network, predicting cognitive achievements during toddlerhood (Alcauter et al., 2014). There is considerable progress in executive attention skills throughout the second year of life. Whereas 14-month-olds barely pass inhibitory control tasks (e.g., a more complex version of the A-not B task or a task in which toddlers have to resist touching a toy they have been told not to touch), toddlers are able to perform adequately such inhibitory control tasks at 18 months of age (Miller and Marcovitch, 2015). Cohesion and consistency

Developing attention and self-regulation in infancy and childhood Chapter | 23

511

among these executive attention measures also improve with age during this period. Between 18 months and 3 years of age, performance on simple inhibitory control tasks in which toddlers are only required to withhold a response also experience a major development (Garon et al., 2008). The same applies to simplified versions of the Dimension Carde Sorting Task. Children can be ready to engage with the original Dimension CardeSorting Task from the age of 3 years. Even so, this task is too complex for children that young who tend to make too many perseverative errors or show an inconsistent sorting strategy (Blakey et al., 2016; Zelazo et al., 1996). Interestingly, 3-year-olds who perform better activate prefrontal areas during the task to a greater extent than those who perform worse (Moriguchi and Hiraki, 2009). Likewise, between the second and the third year of life, toddlers become more efficient in resolving cognitive conflict. Computerized tasks can be used with older toddlers so that it is possible to register both reaction time and accuracy of motor responses. An example is the spatial conflict task designed by Gerardi-Caulton (Gerardi-Caulton, 2000). In this task, a target stimulus (e.g., an animal cartoon) appears on the screen either on the left or on the right. The two possible response options are positioned at the bottom of the screen one on the right and one on the left. Toddlers have to touch the option matching target in identity, suppressing the tendency to touch the option on the same side of target on spatial incompatible trials (that is, when target and matching response are located in opposite sites). Performance on this task notably changes from 2 to 3 years of age becoming more accurate and faster, considerably reducing the conflict effect (Gerardi-Caulton, 2000; Rothbart et al., 2003). At the same time, self-regulation flourishes during toddlerhood in close relationship with the development of inhibitory control. For instance, between 22 and 33 months of age, toddlers gain control over their behavior and manage to refrain from eating a tempting snack placed in front of them, waiting longer before taking the snack as they get older (Kochanska et al., 2000). However, toddlers still need some external aid, such as some support from caregivers, to comply with simple rules (Kochanska et al., 2001; Kopp, 1982). They become more and more independent with age, and control over behavior is more likely to be self-initiated. There are also some differences in the developmental pace of self-regulation skills depending on the context in which toddlers have to regulate their behavior. Self-regulating to prevent or stop doing something is easier to attain for toddlers compared with self-regulation in situations that require sustaining an unpleasant or demanding activity such as cleaning up toys after playing (Kochanska et al., 2001).

23.4.3 Childhood As previously mentioned, the alerting and orienting networks are functional soon after birth, and infants and toddlers rapidly make a huge progress in these two aspects of attention. However, there is still a long way until children’s competence reaches comparable adult levels. The alerting and orienting networks continue developing at a slower pace during childhood. In the case of the alerting network, vigilance tasks are commonly used to measure sustained attention. In these tasks, children should produce a particular response (e.g., pressing a key) every time a certain event occurs (e.g., letter A appears on the screen). Low probability of target events together with the simplicity of the design makes the task a monotonous activity in which maintaining optimal activation levels over time turns into a more challenging task. The ability of children to remain alert during a vigilance task improves considerably during the preschool years (Danis et al., 2008). Unlike toddlers, 3-year-olds reengage the task after involuntary periods of inattention, and by the age of 4 years, children maintain attention in the task the whole time with no interruption. Children’s sustained attention increases progressively after preschool years, performing similarly to adults at about 13 years of age (Lin et al., 1999). Regarding the orienting network, children become faster, more accurate, and efficient in redirecting attention as they grow older (Plude et al., 1994; Schul et al., 2003; Wainwright and Bryson, 2002). There is also a decrease in the orienting cost when invalid cues are presented. From 6 years of age, children benefit from having longer time intervals between an orienting cue and a target (Wainwright and Bryson, 2005). Younger children need some extra time to activate endogenous selection mechanism that allows the voluntary orienting of attention. However, for both covert (i.e., without head and/or eye movements) and overt (i.e., accompanied by eyes/head orientation) orienting, there are no significant age-related changes from preschool age on to adulthood in the facilitation effect produced by valid cues (Wainwright and Bryson, 2002). Furthermore, some evidence suggests that preschool children can already use central orienting cues (e.g., arrows) to anticipate the location of target stimuli (Ristic et al., 2002; Wainwright and Bryson, 2005). Central cues are thought to elicit the endogenous orienting of attention as the symbolic meaning activates the shift of attention in a top-down manner. However, perceptual features of central cues may have dominance over symbolic meaning for children below the age of 4 years so that the orienting effect in very young children would be better explained by automatic orienting responses triggered by central cues (Jakobsen et al., 2013). During the preschool years, children internalize the meaning of symbolic cues, and thus, there is a transition from exogenous to endogenous orienting of attention in reaction to central cues.

512 PART | II Cognitive development

An experimental task that has been widely used in the study of the development of attention networks is the Child Attention Network Task (C-ANT; Rueda et al., 2004). This is a child-friendly version of the original ANT (Fan et al., 2002), which incorporates different procedures (warning signals, location cues, and flanker stimuli) to assess the efficiency of alerting, orienting, and executive attention networks in the same task. As for adults, children are asked to respond according to the direction in which a central fish is pointing by pressing a button. Fish take the place of arrows as target stimuli in the C-ANT (see Fig. 23.4). The target fish is flanked by distractors: fish pointing either in the same or the opposite direction (congruent and incongruent conditions respectively). To engage children with the task, it is contextualized as a “catch the fish” or “feed the fish” game. The task is simple enough for children from 3 to 4 years of age, providing information about reaction time and accuracy of children responses. Measures of attention functioning are obtained by contrasting performance between presence versus absence of warning cues (alert index), valid versus invalid orienting cues (orienting index), and congruent versus incongruent flankers (executive attention index, also called the conflict index). Using the task with children revealed that between 6 and 9 years of age, attention networks develop at different developmental trajectories (Rueda et al., 2004; Pozuelos et al., 2014). Alerting and orienting scores remain stable during this period, indicating an earlier development of these two attention networks in consonance with studies using others experimental paradigms. In contrast, conflict scores noticeably diminish between the 6 and 7 years of age with no important changes after this age, revealing that the early years of childhood seem to be particularly relevant for the development of executive attention. Longitudinal data using this task confirm this developmental pattern for attention networks during childhood (Lewis et al., 2018). During preschool years, children experience substantial improvements in inhibitory control. A longitudinal study on behavior inhibition development during childhood showed that inhibitory control skills improve with age between the 3 and 7 years (Kochanska et al., 1997). Inhibitory control as measured with simple inhibitory control tasks rapidly grows at the beginning of preschool years, developing more steadily during the rest of early childhood (Garon et al., 2014). Conversely, inhibitory control in more demanding tasks develops at a slower pace during the early preschool years, increasing substantially after that. Children also experience marked changes in the ability to resolve cognitive conflict. The proportion of children who pass conflict tasks (such as a child version of a Stroop task, Snow/Grass task, consisting in touching the opposite color of either the snow or the grass) rises considerably from the 3 to the 5 years of age (Carlson, 2005). However, the development of executive attention skills extends to adolescence and up to adulthood (Waszak et al., 2010). Data from neuroimaging studies suggest that brain mechanisms related to executive attention are not as efficient in children as in adults. Adults but not children recruit the right ventrolateral prefrontal cortex in cognitive tasks involving inhibition and the suppression of interference (Bunge et al., 2002). Indeed, the executive attention network appears to be the one that undergoes foremost changes during childhood compared with the other attention networks. Connectivity

FIGURE 23.4 Child version of the Attention Network Task (Pozuelos et al., 2014).

Developing attention and self-regulation in infancy and childhood Chapter | 23

513

between areas within the executive network (e.g., anterior cingulate cortex and insula) increases from midchildhood to adulthood, but no meaningful changes are observed in the connections among areas within the orienting network (Menon, 2013). The number of short-range connections progressively begins to decrease in childhood as a consequence of the growing specialization of neural networks during this period, whereas long-range connections between distal areas increase favoring the coordination among the different attention networks (Fair et al., 2007). Using the C-ANT, we can investigate not only how attention networks independently develop but also the interplay among the different networks during the developmental process. In adults, the presence of warning cues appears to reduce the efficacy of the executive attention network in conflict tasks (Callejas et al., 2004). An interaction between the alerting and executive networks has been also observed in children (Mullane et al., 2016; Pozuelos et al., 2014). However, this interaction is in turn modulated by age. Young children benefit from the presence of alerting cues when facing targets involving conflict, whereas older children show the same pattern of results as adults (Pozuelos et al., 2014). This result suggests that maximizing the efficacy of executive attention requires a moderate level of alertness. Too much alerting, as in the case of older children and adults immediately after receiving a warning cue, or too little, as seems to be the case of young children without the help of alerting cues, is associated with poorer implementation of executive attention mechanisms. There is also some evidence for developmental changes in brain activity associated with performance on the C-ANT. In one study (Abundis-Gutiérrez et al., 2014), children’s EEG activity was registered while performing the C-ANT. Agerelated changes are more prominent in early-evoked potentials corresponding with the first stages of stimulus processing. Regarding the alerting network, warning signals seem to be poorly processed by children younger than 10 years of age. Older children and adults typically show an early brain component (N1) in response to warning signals. The N1 component is a negative deflection over frontocentral sites on the scalp that has been related to the preparation for processing subsequent incoming stimuli. Children do not exhibit this ERP component. Concerning orienting of attention, main differences between ages can be seen in N1 and P3 components following the orienting cue, which suggest that reorienting attention after invalid cues requires young children to activate the orienting network to a greater extent than older children and adults. Finally, regarding the executive attention network, children under 13 years of age present an immature pattern of the N2 component. This component is modulated by the congruency of stimuli in adults but not in children. In addition, the N2 shows longer latencies for children than adults. These findings further indicate that the executive attention network has a protracted development extending over adolescence.

23.5 Individual differences No two children are alike. Individual developmental trajectories of attention networks are susceptible to the effects of both external and internal factors. Constitutional characteristics of the child (e.g., temperament or genes) as well as environmental and educational factors shape the development of attention.

23.5.1 Temperament Temperament refers to individual differences in motor, emotional, and attentional reactivity together with the mechanisms involved in regulating such reactivity (Rothbart and Bates, 2006). Temperament is considered the core of personality traits, mainly stable during life span with an important biological basis (Casalin et al., 2012; Stifter et al., 2008). Individual differences in behavioral, emotional, and attention reactivity can be observed from very early on development. Parents distinguish whether their children are, for example, more or less irritable or active from infancy (Rothbart et al., 2011). According to Rothbart and Bates model, three main temperament factors can be distinguished: surgency (SUR), negative affectivity (NA), and regulation/effortful control (EC) (Rothbart and Bates, 2006). The first one, SUR, refers to a tendency toward positive affect, high activity levels, and impulsivity. The second one, NA, is a temperamental disposition to experiment negative emotions such as fear, frustration, distress, or anger. Finally, EC applies to the voluntary control of attention, emotion, or behavior. For instance, EC mechanisms may serve to downregulate high reactivity levels, either positive or negative, to accommodate to the cultural context or to achieve long-term goals. Different temperamental profiles have been linked to differences in children’s executive attention. Brain areas associated with EC mostly overlap with brain structures within the executive attention network. Prefrontal structures associated with the control of attention (e.g., ACC, DLPFC, or VLPC) modulate the activation of limbic structures (essentially amygdala, hippocampus, and nucleus accumbens), so that the interaction among these neural circuits is proposed to explain individual variability in temperament (Rohr et al., 2015; Whittle et al., 2006). On the one hand, higher levels of temperamental reactivity, either SUR or NA, are associated with poorer executive attention skills. Already from infancy,

514 PART | II Cognitive development

higher NA is related to difficulties in disengaging attention (Johnson et al., 1991; McConnell and Bryson, 2005) and less attention flexibility as measured with an attention shifting task (Conejero and Rueda, 2018). Infants with poorer attention skills also show higher SUR levels (Papageorgiou et al., 2015). Preschool children showing higher temperamental NA also commit more perseverative errors during a spatial conflict task and poorer inhibitory control in a delay task (GerardiCaulton, 2000). Higher SUR levels also relate to poorer inhibitory control of children (Davis et al., 2002; Wolfe and Bell, 2007). On the other hand, greater executive attention skills come along with greater EC. The ability of infants to disengage attention has been related to greater soothability (Johnson et al., 1991). Indeed, infants’ control of attention predicts EC at childhood (Papageorgiou et al., 2015). In particular, inhibitory control is a key component of EC. Parental ratings in EC positively correlate with the proportion of correct choices made by 2-year-olds in the A not-B task, which is considered as an indicator of inhibitory control (Morasch and Bell, 2011). Developmental changes in executive attention (specifically in conflict resolution and inhibitory control) during the third year of life also correlate with parental reports of temperamental EC (Gerardi-Caulton, 2000; Kochanska et al., 2000). Likewise, EC is associated with performance of children in inhibitory control tasks such as the go/no-go task and cognitive conflict tasks such as the flanker task (Simonds et al., 2007; Wolfe and Bell, 2007). Overall, these studies suggest that self-regulation and EC rest on the maturation of the executive attention network. Temperament can in fact be a predictor of later executive attention skills (Ursache et al., 2013). Higher levels of emotional reactivity need greater regulation, so that more reactive children might be required to apply control more often and to a greater extent than less reactive children.

23.5.2 Genes One approach to study the influence of genes in the development of attention is to examine heritability with twin designs; one such study compared the performance of pairs of monozygotic and dizygotic twins in the ANT. Researchers found a greater concordance among monozygotic than dizygotic twins in alerting and executive attention scores, although not for orienting scores (Fan et al., 2001). These results suggest that the functions of the alerting and executive attention networks are influenced by genetic endowment to a greater extent than that of the orienting network. Another research approach focuses on identifying candidate genes that might have an influence in the development of attention networks. Specific polymorphic variations of genes can be associated with higher or lower levels of a certain neurotransmitter in particular brain regions. A number of candidate genes have been selected for their implication in the codification of neurotransmitters modulating the three attention networks (see Table 23.1 for a summary). Variations of MAOA gene, involved in the codification of enzymes degrading noradrenaline, are associated with the function of the alerting network at both behavioral and brain levels (Fan et al., 2003; Fossella et al., 2002). In the case of the orienting

TABLE 23.1 Summary of the main genes involved in brain functions and circuits related to attention networks. Attention network

Genes

Variations

Function

Behavioral correlates

Alerting

MAOA

MAOA 2, 3 or 5 repeats

Encode enzymes degrading noradrenaline: low activity

Related to alerting scores in the Attention Network Task (Fan et al., 2003; Fossella et al., 2002)

Orienting

CHRNA4

T

Codify cholinergic receptors, less availability of acetylcholine

More distractibility in selective attention tasks (Greenwood et al., 2012)

Executive attention

DAT1

10 repeats

Increase the density of dopamine transporters

Poorer inhibitory control (Co´mbita et al., 2017)

DRD4

7 repeats

Regulate the amount of dopamine receptors D4 in mesolimbic pathways:

Sensation seeking, less efficiency in resolving cognitive conflict, and more susceptibility to intervention (Congdon et al., 2008; Langley et al., 2004; Fan et al., 2003)

COMT

Val polymorphism

Encode enzymes degrading dopamine

Poorer attention control and less efficiency of executive attention network (Blasi et al., 2005)

Developing attention and self-regulation in infancy and childhood Chapter | 23

515

network, cholinergic neurotransmitters, such as acetylcholine, seem to play a key role. Variations of the CHRNA4, a gene regulating the function of cholinergic receptors, have been related to performance in selective attention tasks (Greenwood et al., 2012). A larger number of studies provide evidence for the heritability of executive attention. Genes modulating dopamine levels in the brain seem to play a predominant role in executive attention network. Some polymorphic variations of genes such as the DAT1, DRD4, or COMT have an influence on the amount of dopamine available in prefrontal cortex contributing to individual differences in executive attention. The SLC6A3/DAT1 10-repeat allele is associated to less availability of dopamine in frontostriatal neural pathways, in contrast to the 9-repeat allele, which is associated to poorer inhibitory control in children (Cómbita et al., 2017). Variations of the DRD4 gene codifying dopamine receptors have been also associated to individual differences in executive attention. People carrying 7-repeat allele of the DRD4 find it more difficult to inhibit the response in a stop-signal task (Congdon et al., 2008). Children diagnosed with ADHD with the 7-repeat DRD4 perform more impulsively on a set of neuropsychological tasks measuring attention, independent of the severity of the symptoms (Langley et al., 2004). Moreover, DRD4 variations are related to the degree to which the anterior cingulate is activated while performing the ANT (Fan et al., 2003). Finally, the COMT gene has been also studied in relation to the executive attention network for its involvement in the codification of enzymes degrading dopamine. When comparing two variants of the COMT (Val vs. Met), the COMT Val genotype is associated with less efficiency of the executive attention network at both behavioral and neural levels in contrast to the Met variant (Blasi et al., 2005). The contribution of genes to individual differences in attention is noticeable from very early in development. Toddlers with the COMT Met polymorphic variation anticipate targets more frequently during an attention-shifting task compared with toddlers with other COMT variations (Voelker et al., 2009). A recent study assessing the genetic influence on attention development from a longitudinal perspective suggests that the importance of different genes may vary according to age (Lundwall et al., 2015). Whereas SLC6A3/DAT1 is associated with attention skills in infancy, COMT and DRD4 explain individual differences in attention when infants grow up as children. Hence, relevant genes in the development of attention should be explored along the life span.

23.5.3 Environment Although genetic heritability has a significant effect in the development of attention, environmental variables also influence its development. One important environmental factor is the family socioeconomic status (SES), which includes parental education, parental occupation, and family income. There is a growing literature evidencing the detrimental effect of experiencing poor socioeconomic conditions on attention development. In general, low SES is associated with worse performance in cognitive tasks (Greg et al., 1998; Noble et al., 2007). More specifically, data from a study measuring attention in 6- to 7-year-old children with the ANT revealed that children with lower SES have poorer scores in alerting and executive attention networks compared with children from families with higher SES, with no differences in the orienting network (Mezzacappa, 2004). In the same line, another study found differences in selective attention of children as a function of caregivers’ education attainment (Stevens et al., 2009). Children performed a dichotic listening task in which they were asked to attend to the input from one ear while ignoring input from the other. Unlike children with highly educated parents, children from families with a lower educational level show no differences in brain activation patterns when processing ignored and attended information. These results suggest that children of lower-educated parents find it harder to suppress the information coming from irrelevant channels, an important function of the executive attention network. The effects of environmental factors on the development of executive attention can be observed from infancy. Infants about 12 months of age living in deprived environments make more perseveration responses in the A-not B task (Lipina et al., 2005). Infants from low SES contexts are more likely to look for the hidden toy in the previous location where the toy was hidden, which indicates lack of attentional inhibition and flexibility. There are some data indicating that during the first years of life, the influence of family SES on the development of cognitive abilities seems to be particularly relevant. The SESecognitive outcomes relationship strengthens during this time and appears to weaken after the age of 4 years (Mollborn et al., 2014). As children grow up, social contexts expand beyond the nuclear family to school, providing children with new and different experiences that are also affecting the development of attention. SES also exerts an impact on the development of different brain structures related to attention. Lower SES is associated with decreased volume of prefrontal structures (Clearfield and Jedd, 2013; Lawson et al., 2013). Children from low-SES backgrounds show diminished gray matter volume in frontal and parietal regions during the first years of life (Hanson et al., 2013). Reduced gray matter volume in structures within the executive attention network may contribute to the poorer functional efficiency of this network. Likewise, children raised in low-SES families show reduced sensitivity to detect

516 PART | II Cognitive development

errors at neural level. Evoked theta oscillatory activity in response to errors is diminished in toddlers from low-SES families compared with those from high-SES families (Conejero et al., 2016). Growing up in an enriched environment seems to promote the maturity of brain circuits involved in error detection in children, as evidenced by the relationship between high SES and development of functional brain activity in response to committed errors during preschool years (Brooker, 2018). Another environmental aspect that has been related to the efficiency of executive attention is the quality of parentechild interactions. Low-quality parenting may have a negative impact on children’s executive attention from very early on (Fay-Stammbach et al., 2014). Inconsistent parenting strategies, low sensitivity to their children’s needs, and a coercive patenting style have been related to poorer executive attention and later behavioral problems (Bindman et al., 2013; Morrell and Murray, 2003; Rioux et al., 2015). In contrast, high-quality parenting appears to promote the development of executive attention skills. It has been observed that when caregivers support toddlers by promoting their autonomy (for instance, teaching them strategies appropriate to their competence and giving them the opportunity to use them) children perform better on attention flexibility tasks (Bernier et al., 2010). Similarly, children whose parents make use of scaffolding when involved in a common activity (that is, encouraging children to be independent and, at the same time, providing them support and feedback about their performance) regulate attention more efficiently (Robinson et al., 2009). Positive parenting strategies may promote the development of attention skills important for self-regulation as long as parents encourage children to actively control impulsive responses (Fay-Stammbach et al., 2014). Conversely, directive parents who exert excessive control over their children’s behavior may hinder the development of their executive attention skills, as they have less opportunities to apply inhibitory control on their own. Conway and Stifter (2012) found that the ability of toddlers to inhibit in a delay task was reduced when mothers have a general tendency to redirect children’s attention. Environmental factors also seem to interact with intrinsic characteristics of children such as temperament, genes, or brain structures. Some authors have proposed that individual differences may cause a differential susceptibility to the environmental variables (Belsky and Pluess, 2009). For instance, carriers of particular genetic variations, as the presence of the DRD4 7-repeat, appear to be more susceptible to external influence (van Ijzendoorn and Bakermans-Kranenburg, 2015). Given these findings, this type of genetic variation might not be considered so much as a risk factor. Instead, it could be seen as more of an indicator for greater potential of change regarding the effects that a prospective intervention may have in stimulating the development of attention skills.

23.6 Plasticity of attention networks Interest in studying whether the efficacy of attention at both the cognitive and brain function level can be improved by means of intervention has been growing in the past decades. A variety of intervention methods have been tested (see Strobach and Karbach, 2016). Training programs often consist of computerized exercises that engage the skills they aim to train at increased levels of difficulty. Several studies using these so-called process-based training interventions have shown efficacy gains in selective attention (Stevens et al., 2008), attentional flexibility (Karbach and Kray, 2009), working memory (Jaeggi et al., 2011), and inhibitory control tasks (Thorell et al., 2009) following training. Because of the high cost of sustained interventions, very often, training studies are limited to just a few sessions of training, rarely going beyond 10 or 12 sessions. This is an important limitation of many studies. However, many training studies have tested whether posttraining improvements are related to training-induced brain plasticity using neuroimaging techniques. Using brain measures may provide a more sensitive test of training effects for short-term interventions because observable effects at the behavior level necessarily reflect changes in underlying brain processes. Reported findings show that cognitive training influences brain plasticity at different levels. Using EEG, we studied training-induced changes in the efficiency of the executive attention networks in a sample of preschool-age children (Rueda et al., 2005; 2012). Results revealed that attention training produced a reduction of latency and a shift of topography of conflict-related activations, suggesting a more advanced pattern of activation after training. In a more recent study, we showed that training executive attention accompanied by metacognitive scaffolding provided by an adult boosts transfer of training to fluid intelligence in 5-year-old children and that the fluid IQ gain following training was predicted by changes in conflict-related brain activation in the frontal midline (Pozuelos et al., 2019). The transfer of training benefits to nontrained tasks (i.e., far-transfer) has been a matter of intense debate. Although research yields mix results, a considerable number of attention training studies have been shown to produce gains in fluid intelligence (fIQ) in children (Neville et al., 2013; Rueda et al., 2012) and adults (Jausovec and Jausovec, 2012; Karbach and Kray, 2009). Nonetheless, other studies have failed to find significant transfer effects of executive processes to fluid

Developing attention and self-regulation in infancy and childhood Chapter | 23

517

intelligence (Colom et al., 2013; Sarzy
nska et al., 2017), and these researchers argue that transfer to intelligence may need more sustained training or that training might not impact intelligence at the construct level but benefit more basic processes taxed by training activities. Training studies have been important to assess and understand how much we can impact the efficiency of brain networks with theoretically grounded interventions. The evidence shows the potential of modifying brain systems by means of education and practice to improve self-regulatory processes. Although more studies are needed to replicate and validate these findings, evidence to date has shed light onto the beneficial impact of training intervention on different processes that promotes mental capital. However, additional research is needed before we determine the nature of the most effective interventions.

23.7 Summary and integration In this chapter, we discussed many different aspects related to attention, including cognitive function, brain networks, development, genes, experience, and intervention. Each of these areas is involved as we try to understand attention, and an integral model of this central cognitive aspect of human behavior is useful for connecting these different levels of analysis. The extraordinary technological developments that have taken place in the past decades allow a much deeper understanding of the mindebrain relationship. Combined with grounded theoretical accounts of cognition, brain imaging methods are being applied to studies of the circuitry, plasticity, and development of neural networks underlying cognitive skills. The integrated model of attention presented in this paper is likely to inform of possible pathophysiological mechanisms of developmental disorders involving attention. In turn, the development of efficient interventions aimed at stimulating attention in both typical and atypical development will be greatly facilitated by knowing the biological factors underlying pathological mechanisms as well as the environmental factors that influence them.

Acknowledgments Work of the authors was supported by grant PSI2017-82670-P of the Spanish Agency of Research awarded to MRR.

References Abundis-Gutiérrez, A., Checa, P., Castellanos, C., Rueda, M.R., 2014. Electrophysiological correlates of attention networks in childhood and early adulthood. Neuropsychologia 57, 78e92. https://doi.org/10.1016/j.neuropsychologia.2014.02.013. Alcauter, S., Lin, W., Smith, J.K., Short, S.J., Goldman, B.D., Reznick, J.S., et al., 2014. Development of thalamocortical connectivity during infancy and its cognitive correlations. J. Neurosci. 34 (27), 9067e9075. https://doi.org/10.1523/JNEUROSCI.0796-14.2014. Arditi, H., Feldman, R., Eidelman, A.I., 2006. Effects of human contact and vagal regulation on pain reactivity and visual attention in newborns. Dev. Psychobiol. 48 (7), 561e573. https://doi.org/10.1002/dev.20150. Aston-Jones, G., Cohen, J.D., 2005. AN integrative theory OF locus coeruleus-norepinephrine function: adaptive gain and optimal performance. Annu. Rev. Neurosci. 28 (1), 403e450. https://doi.org/10.1146/annurev.neuro.28.061604.135709. Atkinson, J., Hood, B., Wattam-Bell, J., Braddick, O., 1992. Changes in infants’ ability to switch visual attention in the first three months of life. Perception 21 (5), 643e653. https://doi.org/10.1068/p210643. Belsky, J., Pluess, M., 2009. Beyond diathesis stress: differential susceptibility to environmental influences. Psychol. Bull. 135 (6), 885e908. https:// doi.org/10.1037/a0017376. Berger, A., Tzur, G., Posner, M.I., 2006. Infant brains detect arithmetic errors. Proc. Natl. Acad. Sci. U.S.A 103 (33), 12649e12653. https://doi.org/ 10.1073/pnas.0605350103. Bernier, A., Carlson, S.M., Whipple, N., 2010. From external regulation to self-regulation: early parenting precursors of young children’s executive functioning. Child Dev. 81 (1), 326e339. https://doi.org/10.1111/j.1467-8624.2009.01397.x. Bindman, S.W., Hindman, A.H., Bowles, R.P., Morrison, F.J., 2013. The contributions of parental management language to executive function in preschool children. Early Child. Res. Q. 28 (3), 529e539. https://doi.org/10.1016/j.ecresq.2013.03.003. Blakey, E., Visser, I., Carroll, D.J., 2016. Different executive functions support different kinds of cognitive flexibility: evidence from 2-, 3-, and 4-yearolds. Child Dev. 87 (2), 513e526. https://doi.org/10.1111/cdev.12468. Blasi, G., Mattay, V.S., Bertolino, A., Elvevåg, B., Callicott, J.H., Das, S., et al., 2005. Effect of catechol-O-methyltransferase val158met genotype on attentional control. J. Neurosci. 25 (20), 5038e5045. https://doi.org/10.1523/JNEUROSCI.0476-05.2005. Botvinick, M., Nystrom, L.E., Fissell, K., Carter, C.S., Cohen, J.D., 1999. Conflict monitoring versus selection-for-action in anterior cingulate cortex. Nature 402 (6758), 179e181. https://doi.org/10.1038/46035. Broadbent, D.E., 1958. The effects of noise on behaviour. In: Perception and Communication. Pergamon Press, Inc, Elmsford, NY, US, pp. 81e107. https://doi.org/10.1037/10037-005.

518 PART | II Cognitive development

Brooker, R.J., 2018. Maternal behavior and socioeconomic status predict longitudinal changes in error-related negativity in preschoolers. Child Dev. 89 (3), 725e733. https://doi.org/10.1111/cdev.13066. Bunge, S.A., Dudukovic, N.M., Thomason, M.E., Vaidya, C.J., Gabrieli, J.D.E., 2002. Immature frontal lobe contributions to cognitive control in children: evidence from fMRI. Neuron 33 (2), 301e311. Buschman, T.J., Miller, E.K., 2007. Top-down versus bottom-up control of attention in the prefrontal and posterior parietal cortices. Science 315 (5820), 1860e1864. https://doi.org/10.1126/science.1138071. Callejas, A., Lupiáñez, J., Tudela, P., 2004. The three attentional networks: on their independence and interactions. Brain Cogn. 54 (3), 225e227. https:// doi.org/10.1016/j.bandc.2004.02.012. Canfield, R.L., Haith, M.M., 1991. Young infants’ visual expectations for symmetric and asymmetric stimulus sequences. Dev. Psychol. 27 (2), 198e208. https://doi.org/10.1037/0012-1649.27.2.198. Carlson, S.M., 2005. Developmentally sensitive measures of executive function in preschool children. Dev. Neuropsychol. 28 (2), 595e616. https:// doi.org/10.1207/s15326942dn2802. Casalin, S., Luyten, P., Vliegen, N., Meurs, P., 2012. The structure and stability of temperament from infancy to toddlerhood: a one-year prospective study. Infant Behav. Dev. 35 (1), 94e108. https://doi.org/10.1016/j.infbeh.2011.08.004. Choudhury, N., Gorman, K.S., 2000. The relationship between sustained attention and cognitive performance in 17e 24-month old toddlers. Infant Child Dev. 9 (December 1999), 127e146. https://doi.org/10.1002/1522-7219(200009)9:33.0.CO;2-5. Clearfield, M.W., Jedd, K.E., 2013. The effects of socio-economic status on infant attention. Infant Child Dev. 22 (1), 53e67. https://doi.org/10.1002/ icd.1770. Clohessy, A.B., Posner, M.I., Rothbart, M.K., 2001. Development of the functional visual field. Acta Psychol. 106 (1e2), 51e68. https://doi.org/10.1016/ S0001-6918(00)00026-3. Colom, R., Román, F.J., Abad, F.J., Shih, P.C., Privado, J., Froufe, M., et al., 2013. Adaptive n-back training does not improve fluid intelligence at the construct level: gains on individual tests suggest that training may enhance visuospatial processing. Intelligence 41 (5), 712e727. https://doi.org/ 10.1016/j.intell.2013.09.002. Colombo, J., Horowitz, F.D., 1987. Behavioral state as a lead variable in neonatal research. Merrill-Palmer Quaterly 33 (4), 423e437. Cómbita, L.M., Voelker, P., Abundis-Gutiérrez, A., Pozuelos, J.P., Rueda, M.R., 2017. Influence of the SLC6A3-DAT1 gene on multifaceted measures of self-regulation in preschool children. Front. Psychol. 8 https://doi.org/10.3389/fpsyg.2017.00026. Conejero, Á., Guerra, S., Abundis-Gutiérrez, A., Rueda, M.R., 2016. Frontal theta activation associated with error detection in toddlers: influence of familial socioeconomic status. Dev. Sci. 21 (1), 1e10. https://doi.org/10.1111/desc.12494. Conejero, Á., Rueda, M.R., 2018. Infant temperament and family socio-economic status in relation to the emergence of attention regulation. Sci. Rep. 8 (1), 11232. https://doi.org/10.1038/s41598-018-28831-x. Congdon, E., Lesch, K.P., Canli, T., 2008. Analysis of DRD4 and DAT polymorphisms and behavioral inhibition in healthy adults: implications for impulsivity. Am. J. Med. Genet. Part B, Neuropsychiatr. Genet. 147B (1), 27e32. https://doi.org/10.1002/ajmg.b.30557. Conway, A., Stifter, C.A., 2012. Longitudinal antecedents of executive function in preschoolers. Child Dev. 83 (3), 1022e1036. https://doi.org/10.1111/ j.1467-8624.2012.01756.x. Corbetta, M., Patel, G., Shulman, G.L., 2008. The reorienting system of the human brain: from environment to theory of mind. Neuron 58 (3), 306e324. https://doi.org/10.1016/j.neuron.2008.04.017. Corbetta, M., Shulman, G.L., 2002. Control of goal-directed and stimulus-driven attention in the brain. Nat. Rev. Neurosci. 3 (3), 201e215. https:// doi.org/10.1038/nrn755. Coull, J.T., Frith, C.D., Büchel, C., Nobre, A.C., 2000. Orienting attention in time: behavioral and neuroanatomical distinction between exogenous and endogenous shifts. Neuropsychologia 38 (6), 808e819. https://doi.org/10.1016/s0028-3932(99)00132-3. Coull, J.T., Nobre, A.C., Frith, C.D., 2001. The noradrenergic a2 agonist clonidine modulates behavioural and neuroanatomical correlates of human attentional orienting and alerting. Cerebr. Cortex 11 (1), 73e84. https://doi.org/10.1093/cercor/11.1.73. Courage, M.L., Reynolds, G.D., Richards, J.E., 2006. Infants’ attention to patterned stimuli: developmental change from 3 to 12 months of age. Child Dev. 77 (3), 680e695. https://doi.org/10.1111/j.1467-8624.2006.00897.x. Cui, R.Q., Egkher, A., Huter, D., Lang, W., Lindinger, G., Deecke, L., 2000. High resolution spatiotemporal analysis of the contingent negative variation in simple or complex motor tasks and a non-motor task. Clin. Neurophysiol. 111 (10), 1847e1859. https://doi.org/10.1016/S1388-2457(00)00388-6. Danis, A., Pêcheux, M.-G., Lefèvre, C., Bourdais, C., Serres-Ruel, J., 2008. A continuous performance task in preschool children: relations between attention and performance. Eur. J. Dev. Psychol. 5 (4), 401e418. https://doi.org/10.1080/17405620600866800. Davis, E.P., Bruce, J., Gunnar, M.R., 2002. The anterior attention network: associations with temperament and neuroendocrine activity in 6-year-old children. Dev. Psychobiol. 40 (1), 43e56. https://doi.org/10.1002/dev.10012. de Barbaro, K., Clackson, K., Wass, S.V., 2017. Infant attention is dynamically modulated with changing arousal levels. Child Dev. 88 (2), 629e639. https://doi.org/10.1111/cdev.12689. de Jong, M., Verhoeven, M., Hooge, I.T.C., van Baar, A.L., 2016. Factor structure of attention capacities measured with eye-tracking tasks in 18-monthold toddlers. J. Atten. Disord. 20 (3), 230e239. https://doi.org/10.1177/1087054713516002. Dean, D.C.I.I.I., O’Muircheartaigh, J., Dirks, H., Waskiewicz, N., Lehman, K., Walker, L., et al., 2014. Modeling healthy male white matter and myelin development: 3 through 60months of age. Neuroimage 84, 742e752. https://doi.org/10.1016/j.neuroimage.2013.09.058. Deoni, S.C.L., Dean, D.C., Remer, J., Dirks, H., O’Muircheartaigh, J., 2015. Cortical maturation and myelination in healthy toddlers and young children. Neuroimage 115, 147e161. https://doi.org/10.1016/j.neuroimage.2015.04.058.

Developing attention and self-regulation in infancy and childhood Chapter | 23

519

Desimone, R., Duncan, J., 1995. Neural mechanisms of selective visual attention. Annu. Rev. Neurosci. 18 (1), 193e222. https://doi.org/10.1146/ annurev.ne.18.030195.001205. Diamond, A., 1990. Developmental time course in human infants and infant monkeys, and the neural bases of, inhibitory control in reaching. Ann. N. Y. Acad. Sci. 608 (1), 637e676. https://doi.org/10.1111/j.1749-6632.1990.tb48913.x. Doria, V., Beckmann, C.F., Arichi, T., Merchant, N., Groppo, M., Turkheimer, F.E., et al., 2010. Emergence of resting state networks in the preterm human brain. Proc. Natl. Acad. Sci. U.S.A 107 (46), 20015e20020. https://doi.org/10.1073/pnas.1007921107. Dosenbach, N.U.F., Fair, D.A., Cohen, A.L., Schlaggar, B.L., Petersen, S.E., 2008. A dual-networks architecture of top-down control. Trends Cogn. Sci. 12 (3), 99e105. Dosenbach, N.U.F., Fair, D.A., Miezin, F.M., Cohen, A.L., Wenger, K.K., Dosenbach, R.A.T., et al., 2007. Distinct brain networks for adaptive and stable task control in humans. Proc. Natl. Acad. Sci. U.S.A 104 (26), 11073e11078. https://doi.org/10.1073/pnas.0704320104. Fair, D.A., Dosenbach, N.U.F., Church, J.A., Cohen, A.L., Brahmbhatt, S., Miezin, F.M., et al., 2007. Development of distinct control networks through segregation and integration. Proc. Natl. Acad. Sci. U.S.A. 104 (33), 13507e13512. https://doi.org/10.1073/pnas.0705843104. Fan, J., Flombaum, J.I., McCandliss, B.D., Thomas, K.M., Posner, M.I., 2003. Cognitive and brain consequences of conflict. Neuroimage 18 (1), 42e57. https://doi.org/10.1006/nimg.2002.1319. Fan, J., Fossella, J., Sommer, T., Wu, Y., Posner, M.I., 2003. Mapping the genetic variation of executive attention onto brain activity. Proc. Natl. Acad. Sci. U.S.A. 100 (12), 7406e7411. https://doi.org/10.1073/pnas.0732088100. Fan, J., McCandliss, B.D., Sommer, T., Raz, A., Posner, M.I., 2002. Testing the efficiency and independence of attentional networks. J. Cogn. Neurosci. 14 (3), 340e347. https://doi.org/10.1162/089892902317361886. Fan, J., Wu, Y., Fossella, J.A., Posner, M.I., 2001. Assessing the heritability of attentional networks. BMC Neurosci. 2 (1), 14. https://doi.org/10.1186/ 1471-2202-2-14. Fay-Stammbach, T., Hawes, D.J., Meredith, P., 2014. Parenting influences on executive function in early childhood: a review. Child Dev. Perspect. 8 (4), 258e264. https://doi.org/10.1111/cdep.12095. Figueiredo, B., Castro, C., Miguel, T., Field, T., 2016. Infant sleep-wake behaviors at two weeks , three and six months. Infant Behav. Dev. 44, 169e178. https://doi.org/10.1016/j.infbeh.2016.06.011. Fossella, J., Sommer, T., Fan, J., Wu, Y., Swanson, J.M., Pfaff, D.W., et al., 2002. Assessing the molecular genetics of attention networks. BMC Neurosci. 3 (1), 14. https://doi.org/10.1186/1471-2202-3-14. Fox, M.D., Corbetta, M., Snyder, A.Z., Vincent, J.L., Raichle, M.E., 2006. Spontaneous neuronal activity distinguishes human dorsal and ventral attention systems. Proc. Natl. Acad. Sci. U.S.A. 103 (26), 10046e10051. https://doi.org/10.1073/pnas.0604187103. Garon, N., Bryson, S.E., Smith, I.M., 2008. Executive function in preschoolers: a review using an integrative framework. Psychol. Bull. 134 (1), 31e60. https://doi.org/10.1037/0033-2909.134.1.31. Garon, N., Smith, I.M., Bryson, S.E., 2014. A novel executive function battery for preschoolers: sensitivity to age differences. Child Neuropsychol. 20 (6), 713e736. https://doi.org/10.1080/09297049.2013.857650. Gerardi-Caulton, G., 2000. Sensitivity to spatial conflict and the development of self-regulation in children 24-36 months of age. Dev. Sci. 3 (4), 397e404. https://doi.org/10.1111/1467-7687.00134. Greenwood, P.M., Parasuraman, R., Espeseth, T., 2012. A cognitive phenotype for a polymorphism in the nicotinic receptor gene CHRNA4. Neurosci. Biobehav. Rev. 36 (4), 1331e1341. https://doi.org/10.1016/j.neubiorev.2012.02.010. Greg, J., Jean, W., Judith, R., Duncan, G.J., Yeung, W.J., Brooks-Gunn, J., Smith, J.R., 1998. How much does childhood poverty affect the life chances of children ? Am. Sociol. Rev. 63 (3), 406. https://doi.org/10.2307/2657556. Hagmann, P., Sporns, O., Madan, N., Cammoun, L., Pienaar, R., Wedeen, V.J., et al., 2010. White matter maturation reshapes structural connectivity in the late developing human brain. Proc. Natl. Acad. Sci. U.S.A. 107 (44), 19067e19072. https://doi.org/10.1073/pnas.1009073107. Hanson, J.L., Hair, N., Shen, D.G., Shi, F., Gilmore, J.H., Wolfe, B.L., Pollak, S.D., 2013. Family poverty affects the rate of human infant brain growth. PLoS One 8 (12). https://doi.org/10.1371/journal.pone.0080954. Holmboe, K., Bonneville-Roussy, A., Csibra, G., Johnson, M.H., 2018. Longitudinal development of attention and inhibitory control during the first year of life. Dev. Sci. e12690. https://doi.org/10.1111/desc.12690. Holmboe, K., Pasco Fearon, R.M., Csibra, G., Tucker, L.A., Johnson, M.H., 2008. Freeze-Frame: a new infant inhibition task and its relation to frontal cortex tasks during infancy and early childhood. J. Exp. Child Psychol. 100 (2), 89e114. https://doi.org/10.1016/j.jecp.2007.09.004. Hood, B.M., 1995. Shifts of visual attention in the human infant: a neuroscientific approach. Adv. Infancy Res. 9, 163e216. Hunnius, S., Geuze, R.H., van Geert, P., 2006. Associations between the developmental trajectories of visual scanning and disengagement of attention in infants. Infant Behav. Dev. 29 (1), 108e125. https://doi.org/10.1016/j.infbeh.2005.08.007. Jaeggi, S.M., Buschkuehl, M., Jonides, J., Shah, P., 2011. Short- and long-term benefits of cognitive training. Proc. Natl. Acad. Sci. U.S.A 108 (25), 10081e10086. https://doi.org/10.1073/pnas.1103228108. Jakobsen, K.V., Frick, J.E., Simpson, E.A., 2013. Look here! The development of attentional orienting to symbolic cues. J. Cogn. Dev. 14 (2), 229e249. https://doi.org/10.1080/15248372.2012.666772. James, W., 1890. Principles of Psychology. Henry Holt, Hardpress Limited, New York. Jausovec, N., Jausovec, K., 2012. Working memory training: improving intelligence–changing brain activity. Brain Cogn. 79 (2), 96e106. https://doi.org/ 10.1016/j.bandc.2012.02.007. Johnson, M.H., 1990. Cortical maturation and the development of visual attention in early infancy. J. Cogn. Neurosci. 2 (2), 81e95. https://doi.org/ 10.1162/jocn.1990.2.2.81.

520 PART | II Cognitive development

Johnson, M.H., 1994. Visual attention and the control of eye movements in early infancy. In: Attention and Performance: XV. Conscious and Nonconscious Information Processing, pp. 291e310. Johnson, M.H., Posner, M.I., Rothbart, M.K., 1991. Components of visual orienting in early infancy: contingency learning, anticipatory looking, and disengaging. J. Cogn. Neurosci. 3 (4), 335e344. https://doi.org/10.1162/jocn.1991.3.4.335. Johnson, M.H., Tucker, L.a., 1996. The development and temporal dynamics of spatial orienting in infants. J. Exp. Child Psychol. 63 (1), 171e188. https://doi.org/10.1006/jecp.1996.0046. Kahneman, D., 1973. Attention and Effort. Citeseer. Karbach, J., Kray, J., 2009. How useful is executive control training? Age differences in near and far transfer of task-switching training. Dev. Sci. 12 (6), 978e990. https://doi.org/10.1111/j.1467-7687.2009.00846.x. Kochanska, G., Coy, K.C., Murray, K.T., 2001. The development of self-regulation in the first four years of life. Child Dev. 72 (4), 1091e1111. https:// doi.org/10.1111/1467-8624.00336. Kochanska, G., Murray, K., Coy, K.C., 1997. Inhibitory control as a contributor to conscience in childhood: from toddler to early school age. Child Dev. 68, 263e277. https://doi.org/10.1111/j.1467-8624.1997.tb01939.x. Kochanska, G., Murray, K.T., Harlan, E.T., 2000. Effortful control in early childhood: continuity and change, antecedents, and implications for social development. Dev. Psychol. 36 (2), 220e232. https://doi.org/10.1037/0012-1649.36.2.220. Kopp, B., Rist, F., Mattler, U., 1996. N200 in the flanker task as a neurobehavioral tool for investigating executive control. Psychophysiology 33, 282e294. https://doi.org/10.1111/j.1469-8986.1996.tb00425.x. Kopp, C.B., 1982. Antecedents of self-regulation: a developmental perspective. Dev. Psychol. 18 (2), 199e214. https://doi.org/10.1037/00121649.18.2.199. Langley, K., Marshall, L., van den Bree, M., Thomas, H., Owen, M., O’Donovan, M., Thapar, A., 2004. Association of the dopamine D4 receptor gene 7repeat allele with neuropsychological test performance of children with ADHD. Am. J. Psychiatry 161 (1), 133e138. https://doi.org/10.1176/ appi.ajp.161.1.133. Lawson, G.M., Duda, J.T., Avants, B.B., Wu, J., Farah, M.J., 2013. Associations between children’s socioeconomic status and prefrontal cortical thickness. Dev. Sci. 16 (5), 641e652. https://doi.org/10.1111/desc.12096. Lewis, F., Reeve, R., Johnson, K., 2018. A longitudinal analysis of the attention networks in 6- to 11-year-old children. Child Neuropsychol. 24 (2), 145e165. https://doi.org/10.1080/09297049.2016.1235145. Lin, C.C.H., Hsiao, C.K., Chen, W.J., 1999. Development of sustained attention assessed using the Continuous Performance Test among children 6-15 years of age. J. Abnorm. Child Psychol. 27 (5), 403e412. https://doi.org/10.1023/a:1021932119311. Lipina, S.J., Martelli, M.I., Vuelta, B., Colombo, J.A., 2005. Performance on the A-not-B task of argentinean infants from unsatisfied and satisfied basic needs homes. Interam. J. Psychol. 39 (1), 49e60. Retrieved from: http://www.redalyc.org/articulo.oa?id¼28439106. Lundwall, R.A., Dannemiller, J.L., Goldsmith, H.H., 2015. Genetic associations with reflexive visual attention in infancy and childhood. Dev. Sci. 20 (3), e12371. https://doi.org/10.1111/desc.12371. McConnell, B.A., Bryson, S.E., 2005. Visual attention and temperament: developmental data from the first 6 months of life. Infant Behav. Dev. 28 (4), 537e544. https://doi.org/10.1016/j.infbeh.2005.09.002. Menon, V., 2013. Developmental pathways to functional brain networks: emerging principles. Trends Cogn. Sci. 17 (12), 627e640. https://doi.org/ 10.1016/j.tics.2013.09.015. Mezzacappa, E., 2004. Alerting, orienting, and executive attention: developmental properties and sociodemographic correlates in an epidemiological sample of young, urban children. Child Dev. 75 (5), 1373e1386. https://doi.org/10.1111/j.1467-8624.2004.00746.x. Miller, S.E., Marcovitch, S., 2015. Examining executive function in the second year of life : coherence , stability, and relations to joint attention and language. Dev Psychol. 51 (1), 101e114. https://doi.org/10.1037/a0038359. Mollborn, S., Lawrence, E., James-hawkins, L., Fomby, P., 2014. Advances in Life Course Research when do socioeconomic resources matter most in early childhood ? Adv. Life Course Res. 20, 56e69. https://doi.org/10.1016/j.alcr.2014.03.001. Morasch, K.C., Bell, M.A., 2011. Journal of Experimental Child the role of inhibitory control in behavioral and physiological expressions of toddler executive function. J. Exp. Child Psychol. 108 (3), 593e606. https://doi.org/10.1016/j.jecp.2010.07.003. Moriguchi, Y., Hiraki, K., 2009. Neural origin of cognitive shifting in young children. Proc. Natl. Acad. Sci. U.S.A. 106 (14), 6017e6021. https://doi.org/ 10.1073/pnas.0809747106. Morrell, J., Murray, L., 2003. Parenting and the development of conduct disorder and hyperactive symptoms in childhood: a prospective longitudinal study from 2 months to 8 years. J. Child Psychol. Psychiatry 44 (4), 489e508. https://doi.org/10.1111/1469-7610.t01-1-00139. Mullane, J.C., Lawrence, M.A., Corkum, P.V., Klein, R.M., McLaughlin, E.N., 2016. The development of and interaction among alerting, orienting, and executive attention in children. Child Neuropsychol. 22 (2), 155e176. https://doi.org/10.1080/09297049.2014.981252. Neville, H.J., Stevens, C., Pakulak, E., Bell, T.A., Fanning, J., Klein, S., Isbell, E., 2013. Family-based training program improves brain function, cognition, and behavior in lower socioeconomic status preschoolers. Proc. Natl. Acad. Sci. U.S.A. 110 (29), 12138e12143. https://doi.org/10.1073/ pnas.1304437110. Noble, K.G., McCandliss, B.D., Farah, M.J., 2007. Socioeconomic gradients predict individual differences in neurocognitive abilities. Dev. Sci. 10 (4), 464e480. https://doi.org/10.1111/j.1467-7687.2007.00600.x. Norman, D.A., Shallice, T., 1986. Attention to action: willed and automatic control of behavior. In: Davison, R.J., Schwartz, G.E., Shapiro, D. (Eds.), Consciousness and Self-Regulation. Plenum Press, New York, pp. 1e18.

Developing attention and self-regulation in infancy and childhood Chapter | 23

521

Papageorgiou, K.A., Farroni, T., Johnson, M.H., Smith, T.J., Ronald, A., 2015. Individual differences in newborn visual attention associate with temperament and behavioral difficulties in later childhood. Sci. Rep. 5 (January), 1e8. https://doi.org/10.1038/srep11264. Petersen, S.E., Posner, M.I., 2012. The attention system of the human brain: 20 Years after. Annu. Rev. Neurosci. 35 (1), 73e89. https://doi.org/10.1146/ annurev-neuro-062111-150525. Plude, D.J., Enns, J.T., Brodeur, D., 1994. The development of selective attention: a life-span overview. Acta Psychol. 86 (2e3), 227e272. https:// doi.org/10.1016/0001-6918(94)90004-3. Polich, J., 2007. Updating P300: an integrative theory of P3a and P3b. Clin. Neurophysiol. 118, 2128e2148. https://doi.org/10.1016/j.clinph.2007.04.019. Posner, M.I., Petersen, S.E., 1990. The attention system of the human brain. Annu. Rev. Neurosci. 13 (1), 25e42. https://doi.org/10.1146/ annurev.ne.13.030190.000325. Posner, M.I., Rothbart, M.K., 2000. Developing mechanisms of self-regulation. Dev. Psychopathol. 12 (3), 427e441. https://doi.org/10.1017/ S0954579400003096. Pozuelos, J.P., Paz-Alonso, P.M., Castillo, A., Fuentes, L.J., Rueda, M.R., 2014. Development of attention networks and their interactions in childhood. Dev. Psychol. 50 (10), 2405e2415. https://doi.org/10.1037/a0037469. Pozuelos, J.P., Cómbita, L.M., Abundis-Gutiérrez, A., Paz-Alonso, P.M., Conejero, A., Guerra, S., Rueda, M.R., 2019. Metacognitive scaffolding boosts cognitive and neural benefits following executive attention training in children. Dev. Sci. 22 (2), e12756. https://doi.org/10.1111/desc.12756. Reid, V.M., Hoehl, S., Grigutsch, M., Groendahl, A., Parise, E., Striano, T., 2009. The neural correlates of infant and adult goal prediction: evidence for semantic processing systems. Dev. Psychol. 45 (3), 620e629. https://doi.org/10.1037/a0015209. Rioux, C., Castellanos-Ryan, N., Parent, S., Vitaro, F., Tremblay, R.E., Séguin, J.R., 2015. Differential susceptibility to environmental influences: interactions between child temperament and parenting in adolescent alcohol use. Dev. Psychopathol. 28, 1e11. https://doi.org/10.1017/ S0954579415000437. Ristic, J., Friesen, C.K., Kingstone, A., 2002. Are eyes special? It depends on how you look at it. Psychon. Bull. Rev. 9 (3), 507e513. https://doi.org/ 10.3758/BF03196306. Robinson, J.B., Burns, B.M., Davis, D.W., 2009. Maternal scaffolding and attention regulation in children living in poverty. J. Appl. Dev. Psychol. 30 (2), 82e91. https://doi.org/10.1016/j.appdev.2008.10.013. Rohr, C.S., Dreyer, F.R., Aderka, I.M., Margulies, D.S., Frisch, S., Villringer, A., Okon-Singer, H., 2015. Individual differences in common factors of emotional traits and executive functions predict functional connectivity of the amygdala. Neuroimage 120, 154e163. https://doi.org/10.1016/ j.neuroimage.2015.06.049. Rothbart, M.K., Bates, J.E., 2006. Temperament. In: Handbook of Child Psychology Social, Emotional, and Personality Development, sixth ed., vol. 3. John Wiley & Sons Inc, Hoboken, NJ, pp. 99e166. Rothbart, M.K., Ellis, L.K., Rueda, M.R., Posner, M.I., 2003. Developing mechanisms of temperamental effortful control. J. Personal. 71 (6), 1113e1143. https://doi.org/10.1111/1467-6494.7106009. Rothbart, M.K., Sheese, B.E., Rueda, M.R., Posner, M.I., 2011. Developing mechanisms of self-regulation in early life. Emotion Rev. 3 (2), 207e213. https://doi.org/10.1177/1754073910387943. Rueda, M.R., Fan, J., McCandliss, B.D., Halparin, J.D., Gruber, D.B., Lercari, L.P., Posner, M.I., 2004. Development of attentional networks in childhood. Neuropsychologia 42 (8), 1029e1040. https://doi.org/10.1016/j.neuropsychologia.2003.12.012. Rueda, M.R., Checa, P., Cómbita, L.M., 2012. Enhanced efficiency of the executive attention network after training in preschool children: immediate changes and effects after two months. Dev. Cogn. Neurosci. 2S, S192eS204. https://doi.org/10.1016/j.dcn.2011.09.004. Rueda, M.R., Posner, M.I., Rothbart, M.K., 2011. Attentional control and self-regulation. In: Vohs, K.D., Baumeister, R.F. (Eds.), Handbook of SelfRegulation: Research, Theory, and Applications, second ed., vol. 2. The Guldford Press, New York, pp. 284e299. Rueda, M.R., Rothbart, M.K., McCandliss, B.D., Saccomanno, L., Posner, M.I., 2005. Training, maturation, and genetic influences on the development of executive attention. Proc. Natl. Acad. Sci. U.S.A 102 (41), 14931e14936. http://www.pnas.org/cgi/content/abstract/102/41/14931. Ruff, H.A., Lawson, K.R., 1990. Development of sustained, focused attention in young children during free play. Dev. Psychol. 26 (1), 85e93. https:// doi.org/10.1037/0012-1649.26.1.85. _ Sarzy
nska, J., Zelechowska, D., Falkiewicz, M., Necka, E., 2017. Attention training in schoolchildren improves attention but fails to enhance fluid intelligence. Stud. Psychol. 59 (1), 50e65. https://doi.org/10.21909/sp.2017.01.730. Schul, R., Townsend, J., Stiles, J., 2003. The development of attentional orienting during the school-age years. Dev. Sci. 6 (3), 262e272. https://doi.org/ 10.1111/1467-7687.00282. Silverman, I.W., Gaines, M., 1996. Using standard situations to measure attention span and persistence in toddler-aged children: some cautions. J. Genet. Psychol. 157 (4), 397e410. https://doi.org/10.1080/00221325.1996.9914874. Simonds, J., Kieras, J.E., Rueda, M.R., Rothbart, M.K., 2007. Effortful control, executive attention, and emotional regulation in 7-10-year-old children. Cogn. Dev. 22 (4), 474e488. https://doi.org/10.1016/j.cogdev.2007.08.009. Stevens, C., Fanning, J., Coch, D., Sanders, L., Neville, H., 2008. Neural mechanisms of selective auditory attention are enhanced by computerized training: electrophysiological evidence from language-impaired and typically developing children. Brain Res. 1205, 55e69. https://doi.org/10.1016/ j.brainres.2007.10.108. Stevens, C., Lauinger, B., Neville, H., 2009. Differences in the neural mechanisms of selective attention in children from different socioeconomic backgrounds: an event-related brain potential study. Dev. Sci. 12 (4), 634e646. https://doi.org/10.1111/j.1467-7687.2009.00807.x. Stifter, C.A., Putnam, S., Jahromi, L., 2008. Exuberant and inhibited toddlers: stability of temperament and risk for problem behavior. Dev. Psychopathol. 20 (2), 401e421. https://doi.org/10.1017/S0954579408000199.

522 PART | II Cognitive development

Strobach, T., Karbach, J., 2016. Cognitive Training: An Overview of Features and Applications. https://doi.org/10.1007/978-3-319-42662-4. Thorell, L.B., Thorell, L.B., Lindqvist, S., Lindqvist, S., Bergman Nutley, S., Bergman Nutley, S., et al., 2009. Training and transfer effects of executive functions in preschool children. Dev. Sci. 12, 106e113. https://doi.org/10.1111/j.1467-7687.2008.00745.x. Trapanotto, M., Benini, F., Farina, M., Gobber, D., Magnavita, V., Zacchello, F., 2004. Behavioural and physiological reactivity to noise in the newborn. J. Paediatr. Child Health 40 (5e6), 275e281. https://doi.org/10.1111/j.1440-1754.2004.00363.x. Ursache, A., Blair, C., Stifter, C.A., Voegtline, K., 2013. Emotional reactivity and regulation in infancy interact to predict executive functioning in early childhood. Dev. Psychol. 49 (1), 127e137. https://doi.org/10.1037/a0027728. van Ijzendoorn, M.H., Bakermans-Kranenburg, M.J., 2015. Genetic differential susceptibility on trial: meta-analytic support from randomized controlled experiments. Dev. Psychopathol. 27 (01), 151e162. https://doi.org/10.1017/S0954579414001369. Veen, V. van, Carter, C.S., 2002. The timing of action-monitoring processes in the anterior cingulate cortex. J. Cogn. Neurosci. 14 (4), 593e602. https:// doi.org/10.1162/08989290260045837. Voelker, P., Sheese, B.E., Rothbart, M.K., Posner, M.I., 2009. Variations in Catechol-O-Methyltransferase Gene interact with parenting to influence attention in early development. Neuroscience 164 (1), 121e130. https://doi.org/10.1016/j.neuroscience.2009.05.059. Vohs, K.D., Baumeister, R.F., 2011. Handbook of Self-Regulation. Research, Theory and Applications, second ed. Guilford Press, New York. Wainwright, A., Bryson, S.E., 2002. The development of exogenous orienting: mechanisms of control. J. Exp. Child Psychol. 82 (2), 141e155. https:// doi.org/10.1016/S0022-0965(02)00002-4. Wainwright, A., Bryson, S.E., 2005. The development of endogenous orienting: control over the scope of attention and lateral asymmetries. Dev. Neuropsychol. 27 (2), 237e255. https://doi.org/10.1207/s15326942dn2702_3. Waszak, F., Li, S.-C., Hommel, B., 2010. The development of attentional networks: cross-sectional findings from a life span sample. Dev. Psychol. 46 (2), 337e349. https://doi.org/10.1037/a0018541. Whittle, S., Allen, N.B., Lubman, D.I., Yucel, M., 2006. The neurobiological basis of temperament: towards a better understanding of psychopathology. Neurosci. Biobehav. Rev. 30 (4), 511e525. https://doi.org/10.1016/j.neubiorev.2005.09.003. Wolfe, C.D., Bell, M.A., 2007. Sources of variability in working memory in early childhood: a consideration of age, temperament, language, and brain electrical activity. Cogn. Dev. 22 (4), 431e455. https://doi.org/10.1016/j.cogdev.2007.08.007. Zelazo, P.D., Frye, D., Rapus, T., 1996. An age-related dissociation between knowing rules and using them. Cogn. Dev. 11 (1), 37e63. https://doi.org/ 10.1016/S0885-2014(96)90027-1.

Chapter 24

The neural correlates of cognitive control and the development of social behavior G.A. Buzzell1, A. Lahat2 and N.A. Fox1 1

University of Maryland, College Park, MD, United States; 2University of Toronto, Toronto, ON, Canada

Chapter outline 24.1. The development of cognitive control and its neural basis523 24.1.1. Error monitoring 524 24.1.2. Control instantiation 526 24.2. The role of cognitive control in decision-making, motivation, and social behavior 528 24.2.1. Motivation, decision-making, and cognitive control 528 24.2.2. Cognitive control and social behavior 530

24.3. Individual differences in cognitive control 24.3.1. Temperament, cognitive control, and psychopathology 24.3.2. Cross-cultural differences in the development of cognitive control 24.4. Chapter summary and future directions References

531 531 532 533 534

This chapter reviews the extant literature on the neural correlates of cognitive control across development, with an emphasis on relations to social behavior, motivation, and individual differences. Increasingly, researchers have sought to identify the subprocesses that make up cognitive control, with “monitoring” and “control instantiation” being the primary divisions of this construct, along with partitions based on the domains within which monitoring and control operate. In this chapter, we adopt this more nuanced approach to describing and understanding how the cognitive control system develops. Broadly, cognitive control has been found to develop nonlinearly, with rapid development across childhood and approaching maturity in adolescence. At the neural level, improvements in these abilities are generally associated with maturation of a frontoparietal network. Alongside such neurocognitive developments in cognitive control, changes in limbic-related motivation processes also exhibit stark developmental changes, particularly during the periadolescence period. In this chapter, we review the neurocognitive development of cognitive control and relations to motivation and social behavior. Additionally, how these functions differ across individuals, as a function of temperament and psychopathology, as well as cross-cultural variation, is reviewed.

24.1 The development of cognitive control and its neural basis With development, children and adolescents improve in their ability to control their thoughts, behavior, and impulses, working toward goals while ignoring distracting information (Casey et al., 2005; Diamond, 2002; Zelazo, 2004). This ability has been termed cognitive control (Bunge and Crone, 2009; Nigg, 2017). Broadly, cognitive control can be divided into two distinct, but highly related, processes, including (1) monitoring, which detects situations where control might be needed, and (2) control instantiation, which biases neurocognitive activity to achieve task goals. Despite this distinction, monitoring and control instantiation are highly related and, in healthy adults, typically work together in tandem to allow for cognitive control. That is, monitoring is only useful if it is followed by control instantiation, and similarly, control instantiation would be aimless in the absence of monitoring to signal when control is required. Nonetheless, there are situations in adults where monitoring and control do not always closely track one another (Buzzell et al., 2017a; Van Der Borght et al., 2016), and from a developmental perspective, there are some indications that these two processes exhibit

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00024-X Copyright © 2020 Elsevier Inc. All rights reserved.

523

524 PART | II Cognitive Development

distinct, albeit highly contingent developmental trajectories. Monitoring and control instantiation can further be partitioned based on what is being monitored or how neural activity is being controlled. With this in mind, this chapter focuses primarily on error monitoring and inhibitory control, as specific examples of monitoring and control instantiation, respectively. The chapter begins with a description of monitoring, leveraging more recent conceptualizations inspired by work in adults, and then provides a review of what is known about the development of monitoring at the behavioral and neural levels. Next, we expand on current conceptualizations of control instantiation, again leveraging recent adult work, before charting its developmental course. With a picture of the two core subdomains of cognitive control and their development in place, the chapter then turns to the highly related concept of motivation and motivation-control relations. Given that social rewards and social contexts represent one of the most common forms of human motivation, we then review what is known about the developmental relations between social behavior and cognitive control. Finally, the last portion of this chapter expands the review to explore how the development of cognitive control is related to individual differences in temperament and psychopathology, as well as cross-cultural variation. This chapter focuses mostly on childhood and adolescence given that majority of research into the neural correlates of cognitive control has been conducted for this age range. Moreover, cognitive control is known to develop markedly during this period.

24.1.1 Error monitoring Monitoring, also referred to as “performance monitoring,” refers to the human brain’s ability to evaluate its own responses, or external stimuli, to determine whether adaptations are required to maintain goal-directed behavior (Ullsperger et al., 2014b). Critically, monitoring is not a unitary construct, nor mapped directly onto a single neural structure. Rather, monitoring is better characterized as a set of processes that all share the common function of detecting when control is needed, and a common anatomical hub in the medialefrontal cortex (MFC), to include the dorsal anterior cingulate cortex (dACC) (Shenhav et al., 2013). The three most common domains for which monitoring has been studied are (1) conflict monitoring, referring to interference and interactions between different information processing pathways in the brain at the level of stimulus or response-related neural activity (Braver et al., 2001); (2) error monitoring, referring to the detection of self-directed actions that deviate from task goals (Gehring et al., 1993); (3) feedback monitoring, referring to the processing of feedback related to task performance or the presence/absence of rewards and punishments (Holroyd and Coles, 2002). This chapter focuses on error monitoring. Methodological considerations and adult studies. In the following, we focus our review of monitoring on the domain of error monitoring. However, before doing so, it is important to note inherent difficulties in the study of monitoring and cognitive control more generally. To study monitoring in the laboratory, computer-based tasks that require children or adults to make relatively quick decisions and responses are typically employed. One approach that has been proposed and widely used to studying error monitoring in children is to assess how quickly participants respond on the trials immediately following errors. A common finding is that individuals slow down on trials that follow errors, and this has traditionally been interpreted as reflecting the monitoring system first detecting the mistake, followed by subsequent control instantiation to respond more cautiously on the following trial (Botvinick et al., 2001; Danielmeier and Ullsperger, 2011). Thus, posterror slowing (PES) had been used as an index of monitoring. However, there are two issues with such an approach. First, it remains debated as to whether PES actually reflects more cautious responding or not (Danielmeier and Ullsperger, 2011; Jentzsch and Dudschig, 2009; Notebaert et al., 2009), with the possibility that what PES reflects differs across tasks (Danielmeier and Ullsperger, 2011), and perhaps, even across development. Second, even if slowing always reflects increased response caution as the result of error monitoring detecting a mistake, this presents a measure of monitoring that is confounded with control instantiation. Given the confounded nature of behavioral metrics, a complementary approach to assessing monitoring is to leverage neuroimaging techniques, which can sometimes provide a more direct assessment of monitoring itself. However, such neural measures can also sometimes be confounded by overlapping neural processes or a lack of certainty as to what they index. Therefore, in studying the development of monitoring, it is important to consider both behavioral and neural metrics, while keeping in mind the limitations of each approach. As previously mentioned, adult work finds that the monitoring system includes a network of brain regions, with a hub in MFC (Shenhav et al., 2013; Ullsperger et al., 2014a). Critically, activation of the monitoring system can be measured using the two most commonly employed noninvasive measures of neural activity: functional magnetic resonance imaging (fMRI) and electroencephalography (EEG). For example, when errors occur, conflict is detected, or negative feedback processed, increased activation of the MFC is typically observed via fMRI (Ridderinkhof et al., 2004). Moreover, a similar pattern of event-related potentials (ERPs) in the EEG is observed following conflict, errors, and feedback, consisting first of a rapid negativity over frontocentral scalp regions (i.e., the N2, error-related negativity [ERN], and feedback negativity,

The neural correlates of cognitive control and the development of social behavior Chapter | 24

525

respectively) and a slower positivity over more centroparietal scalp regions (i.e., the P3, Pe, and feedback P3, respectively; Ullsperger et al., 2014b). Source analysis studies suggest that the frontocentral negativities tend to share a common neural hub within MFC (Debener, 2005; Hauser et al., 2014; van Veen and Carter, 2002). Additionally, timeefrequency analyses of EEG demonstrate that monitoring is associated with oscillatory changes within the theta band over frontocentral regions (Cavanagh et al., 2012). To summarize, adult work suggests that monitoring reflects a class of processes that operate across domains, with a common neural hub within the MFC. Although monitoring activity can be indirectly inferred via behavioral metrics, neuroimaging approaches can also be used and may sometimes provide a more direct assessment of monitoring. With a basic understanding of adult monitoring in place, we now turn to a review of how monitoring develops across childhood and adolescence, with an emphasis on error monitoring. Error monitoring development. Behavioral studies generally suggest that error monitoring emerges early in childhood. Using PES as a way to indirectly assess monitoring in children, Backen Jones et al. (2003) found that as 3- to 4-year-old children’s performance on a Simon Says-type task improved, they tended to show PES. However, using a more fast-paced task-switching paradigm, Davidson et al. (2006) failed to find evidence of PES in 4- to 5-year-old children. Typically, younger children make more errors than adults and may be less aware of them; some studies have shown that children are able to detect errors but are unable to correct them. For example, Bullock and Lutkenhaus (1984) found that 18- and 24-month-old children could distinguish between correctly and incorrectly built towers, even when they themselves failed to build the towers correctly. In a study designed to examine this question with preschool children, Jacques et al. (1999) presented 3-year-old children with the dimensional change card sort, in which children are shown two target cards (e.g., a blue rabbit and a red boat) and are asked to sort a series of test cards (e.g., red rabbits and blue boats) first according to one dimension (e.g., color) and then according to the other (e.g., shape). Most 3-year-olds perseverate during the postswitch phase, continuing to sort test cards by the first dimension (e.g., Zelazo et al., 2003). To assess error detection, Jacques et al. (1999) asked children to evaluate the sorting of a puppet. When 3-year-olds watched the puppet perseverate, they judged the puppet to be correct, whereas when they saw the puppet sort correctly, they judged the puppet to be wrong. This pattern of results suggests that 3-year-olds’ perseverating performance and error detection are closely linked in this task. When studying PES in slightly older, school-aged children, mixed findings have been reported. On the one hand, a study of 6- to 10-year-olds found that PES increased from ages 6 to 7 years but then steadily decreased from ages 7 to 10 years (Gupta et al., 2009). On the other hand, across a wide age range (7e25 years), Davies et al. (2004) identified the presence of PES for all age groups, albeit no significant age-related changes emerged. Bearing the known limitations of using PES as an index of monitoring, the literature appears to support the relatively early emergence of error monitoring within the first few years of life, some evidence for an initial increase in monitoring into the early school years, but less development during adolescence. Electrophysiological studies in young children have provided additional support for the early emergence of error monitoring. For example, EEG studies have demonstrated the presence of an ERN, thought to reflect activation of the monitoring system, as young as 4 years of age (Brooker et al., 2011). Moreover, this same study provides some evidence that the ERN increases in magnitude from 4 to 8 years, particularly for girls (Brooker et al., 2011). Other work has demonstrated the continued development of the ERN from ages 8 to 13 years (Meyer et al., 2012). Additionally, in one of the most extensive investigations of error monitoring development, a study by Davies at al. (2004) found that error monitoring generally increases throughout adolescence. However, the authors also found that such age-related changes were nonlinear and qualified by sex interactions, with females tending to show earlier changes than males. A more detailed review of developmental changes in the ERN specifically can be found elsewhere (Tamnes et al., 2013). As previously mentioned, prior work in adults has consistently demonstrated a primarily cingulate-based source for the ERN, providing a direct link between this electrophysiological measure and the monitoring system (Debener, 2005; Hauser et al., 2014; van Veen and Carter, 2002). Similarly, developmental studies demonstrate that the ERN also includes a primarily cingulate-based source (Mathewson et al., 2005; Ladouceur et al., 2007; Santesso and Segalowitz, 2008), providing a direct link between developmental changes in the ERN and monitoring activity within the cingulate (MFC). Indeed, fMRI investigations have also found increased error-related MFC activation in adults, compared with children (Fitzgerald et al., 2010; Rubia et al., 2007). However, it is worth noting that the majority of source analysis studies have leveraged an approach that does not readily allow for identifying multiple sources of ERN activity across the brain and have employed approaches with limited spatial resolution. A more recent ERN source localization study in a sample of 9to 35-year-olds employed individual MRI constraints and a distributed source model approach to build on this prior work in two important ways (Buzzell et al., 2017b). First, it was confirmed that across the age range studied, a primarily cingulate-based source was again identified, and critically, no significant changes across adolescence were found within

526 PART | II Cognitive Development

this core region of the monitoring system. Second, age-related changes in the ERN were localized to ventralefrontal neural regions that make up the extended error monitoring network: the insula, orbitofrontal cortex, and more caudal portions of the inferior frontal gyrus (frontal operculum; Buzzell et al., 2017b). Collectively, these data provide corroborating evidence for developmental increases in error monitoring through adolescence and even into early adulthood; however, at least for increases in error monitoring at these later stages of development, neural regions more closely tied to salience and emotional processing within ventralefrontal brain regions seem to be the focus of developmental change. These findings are consistent with structural (MRI) and functional (fMRI) studies that point to continued development and reorganization of widespread areas of the brain throughout adolescence and into adulthood, particularly those areas linked to higher cognitive functions like cognitive control (Sowell et al., 1999; Gogtay et al., 2004; Tamm et al., 2002). Moreover, these more extended changes within the monitoring system are consistent with other investigations of executive function development reporting marked changes in network dynamics with increasing age (Fair et al., 2007, 2009). Studies of the ERN generally support the notion that error monitoring emerges early and exhibits continued development across childhood and adolescence, findings that are corroborated by source analyses, behavioral assessments of PES, and fMRI-based investigations. However, such studies leave open the question as to whether such development reflects increases in the underlying brain activity driving the ERN, i.e., signal strength, or whether developmental changes are better characterized by more efficient and synchronous neuronal firing. Leveraging timeefrequency analyses of EEG data, DuPuis et al. (2015) demonstrated that developmental increases in the ERN from ages 5 to 7 years were best described by increases in signal consistency, as opposed to signal strength. Specifically, these authors demonstrated that theta band power (signal strength) and theta phase synchrony (signal consistency) both predicted the ERN but that longitudinal increases in the ERN were predicted by theta phase synchrony increases, whereas theta power actually exhibited decreases across this time period (DuPuis et al., 2015). These data suggest that error monitoring development, at least in the early school years, is best characterized by improved efficiency of the monitoring system. Future work will be needed to further characterize how theta band dynamics change across childhood and adolescence. To summarize, error monitoring seems to emerge early, as supported by behavioral and electrophysiological data, and demonstrates rapid development during childhood but also continues to change throughout adolescence and even into adulthood. A general principle of such error monitoring development appears to be increased efficiency of this system, characterized by increased fMRI activations within error-specific brain regions, along with increased synchrony of such activations, as opposed to an overall increase in signal strength. Finally, at least for later development of this system throughout adolescence and into adulthood, development appears to be driven most prominently by the extended monitoring network in more ventralefrontal brain regions, as opposed to continued changes within the central MFC hub.

24.1.2 Control instantiation Detecting when something goes wrong, monitoring, is only advantageous if the brain can do something about it and engage control instantiation. Specifically, control instantiation refers to top-down control over the brain and behavior to achieve task goals (Miller and Cohen, 2001). However, much like monitoring, control instantiation is not a unitary construct, nor mapped to a single neural structure. Instead, control instantiation refers to a broad class of neurocognitive processing that allows for maintaining goal-directed behavior. For example, inhibiting unwanted responses (Aron, 2007), increasing attention (Desimone and Duncan, 1995), and switching from one task to another (Braver et al., 2003) reflect a few examples of control instantiation. Thus, control instantiation can be partitioned as a function of the domain within which control operates; in this review, we focused primarily on inhibitory control, with some discussion of attentional control and task switching as well. We first describe adult work briefly, as an introduction to the neural basis of control instantiation and its relations with monitoring. We then turn to a review of how control instantiation develops across childhood and adolescence. Control instantiation in adults. Control instantiation can be readily assessed using behavioral metrics. Simply put, if participants are asked to perform a task that is believed to require a particular form of control instantiation for successful completion, then accuracy and real-time performance can be taken as a measure of control instantiation. However, such behavioral approaches to the study of control instantiation are not without limitation. First, control instantiation rarely occurs in the absence of monitoring first detecting that control is required. Thus, behavioral measures of control instantiation are necessarily confounded by monitoring. Second, when attempting to study a specific domain of control instantiation, such as inhibitory control, it is essentially impossible to design a task that does not also involve other forms of control to be performed properly. For example, a “go/no-go” task involves making repeated responses to frequently presented stimuli, and then infrequently withholding responses to others, believed to require inhibitory control (Bokura et al., 2001). This task is known to rely heavily on inhibitory control for successful completion; nonetheless, this task also

The neural correlates of cognitive control and the development of social behavior Chapter | 24

527

requires some degree of attentional control for successful performance (Schröger, 1993), yielding an imperfect metric of control instantiation. To summarize, similar to the limits of behavioral assessments of monitoring, behavioral approaches for studying control instantiation are also imperfect. However, here again, neural measures can be leveraged to provide a complementary method for studying control instantiation. Work in adults generally supports the notion that whereas monitoring relies primarily on midline frontal structures, like the cingulate and MFC more generally, control instantiation is more closely linked to lateral prefrontal structures, along with superior parietal regions (Dosenbach et al., 2008; Miller and Cohen, 2001; Petersen and Posner, 2012). Moreover, given that monitoring and control typically work together in concert to allow for goal-directed behavior, a common observation is the dynamic interaction between these two cognitive control subprocesses. For example, following errors or other events that signal a need for control instantiation, initial activation of the MFC, associated with monitoring, is typically followed by activation of lateraleprefrontal cortical regions associated with control instantiation (MacDonald et al., 2000). This dynamic interplay between medial and lateral frontal structures has been observed using fMRI (MacDonald et al., 2000), as well as EEG (Cavanagh et al., 2009) by observing increased coupling within the theta band between these two regions. Despite these general themes of monitoring and control instantiation relations, the neural correlates of control instantiation also exhibit a high degree of domain specificity. For example, inhibitory control has been linked to a relatively specified frontostriatal network, including the inferior frontal cortex (Aron et al., 2004). In contrast, attentional control is more closely associated with a frontoparietal network that includes dorsolateral prefrontal cortex (PFC), the frontal eye fields, and superior parietal cortex (Petersen and Posner, 2012). Activation of these regions, linked to the sources of inhibitory and attentional control, can thus be taken as evidence for domain-specific control instantiation. To summarize, control instantiation often emerges shortly after monitoring processes reach their peak, and direct interactions between monitoring and control instantiation emerge through medialelateral communication within the frontal cortex. Moreover, control instantiation exhibits a high degree of domain specificity, which can be leveraged to more closely track changes in particular forms of control. With this basic understanding of control instantiation in place, we now turn to a review of how these processes develop. In particular, we place an emphasis on the study of inhibitory control, but studies relating to attentional control and task switching are also discussed. Development of inhibitory control. Inhibitory control, as defined earlier, is the ability to inhibit a dominant response in favor of a more appropriate response or no response at all (Aron, 2007; Rothbart et al., 2003). At the behavioral level, inhibitory control has a well-characterized developmental course, with improvements in children’s ability to inhibit prepotent responses detected during toddlerhood (Gerardi-Caulton, 2000), along with improvements in inhibitory control occurring during the preschool years (e.g., Gerstadt et al., 1994; Rueda et al., 2004; Zelazo, 2006). Moreover, marked improvements in inhibitory control continue during middle childhood (e.g., Simonds et al., 2007) and adolescence (e.g., Anderson, 2002; Huizinga et al., 2006). Regarding changes in neural activation during inhibitory control tasks, research has pointed to several changes that occur with development. As noted earlier, a core region involved in inhibitory control, particularly for inhibition of motor responses, is the inferior frontal cortex (IFC; Aron et al., 2004). Using a go/no-go task to study longitudinal development of inhibitory control across ages 9e11 years, Durston et al. (2006) found activation of the right IFC to increase across the two time points studied. More importantly, the IFC was found to be associated with improved inhibitory performance and that this region in particular exhibited increased activation with increasing age, whereas developmental decreases in activation were found in prefrontal regions that were not relevant for task performance. These findings are in line with another study that employed a modified go/no-go task, observing increased right IFC activation in adults, compared with children (age 8e12 years). Specifically, Bunge et al. (2002) had participants perform a go/no-go task that included an additional manipulation so that some trials required only motor inhibition of infrequent responses (similar to a classic go/no-go task), whereas other trials required participants to ignore distracting information. As an aside, tasks that require participants to ignore distracting information are said to involve “distractor suppression,” which not only likely involves attentional control to minimize processing of the distracting information but also likely involves inhibitory control of motor responses associated with the distracting information. Critically, the authors found that across both conditions, children failed to recruit the right IFC. Given the close link between IFC activity and inhibitory control, this suggests that the feature shared by each condition of this modified taskdinhibitory control of motor responsesdwas less developed in children. It is also worth noting that children appeared to adopt different neural strategies, compared with adults, for each condition of this task. Specifically, children recruited the opposite prefrontal hemisphere (left) for successful performance of the condition involving distractor suppression, whereas they recruited more posterior brain regions, but not frontal, for successful performance in the condition similar to the classic go/no-go task (Bunge et al., 2002). Beyond developmental changes within the IFC, studies have generally indicated that activation in the PFC becomes more focal with age (e.g., Bunge et al., 2002; Durston et al., 2006; Luna et al., 2001). In a study by Rubia et al. (2006),

528 PART | II Cognitive Development

functional brain activation was compared between adolescents and adults during three different executive tasks measuring inhibitory control (go/no-go), cognitive interference inhibition (Simon task, for which the spatial location of stimuli conflict with the proper response location), and attentional set shifting (switch task, during which interference from a previous stimuluseresponse association has to be inhibited). In all three tasks, adults recruited portions of the PFC, the anterior cingulate cortex (ACC), and the striatum more strongly than adolescents. Additionally, adults engaged the inferior parietal cortex more strongly than adolescents on the Simon and switch tasks, a finding similar to that reported for the go/no-go task by Bunge et al. (2002). Additional evidence for focalization comes from a study by Luna et al. (2001), who administered an antisaccade task to children (7e12 years), adolescents (13e17 years), and young adults (18e22 years). This task requires participants to suppress looking at a visual target that appears suddenly in the peripheral visual field and instead look away from the target in the opposite direction. These authors found that a network of regions thought to be involved in inhibitory control and spatial attention was increased for adults, relative to children and adolescents. In particular, adults exhibited increased activity within the frontal eye fields (FEFs), intraparietal sulcus, thalamus, superior colliculus, and regions of the cerebellum. However, adolescents exhibited a marked increase in striatal activity compared with children, as well as increased dorsolateral prefrontal cortex (DLPFC) activation relative to either children or adults (Luna et al., 2001). As will be discussed further in the section on motivation, these findings regarding the striatum and DLPFC might point to the unique developmental changes in motivation and cognitive control that occurs during adolescence. In sum, research examining the neural correlates of inhibitory control development suggests that, with age, improvements in this ability are associated with increased focalization and efficiency in recruitment of the PFC. Thus, with development, brain regions that are not relevant for task performance decrease in their activation, whereas regions relevant for the task increase in their activation. This focalization of the PFC has been consistently found with various tasks that tap inhibitory control, as well as for tasks that require other forms of control instantiation. These changes in functional activation of inhibitory control that occur with development may reflect disengagement from immature neural circuits used by children and adolescents and recruitment of more mature, alternative networks. However, the differences in activation patterns between children and adults may also reflect children’s reliance on different cognitive strategies, ongoing strategy-learning processes, or differ from adults in their efforts to complete the tasks (Ernst and Mueller, 2008).

24.2 The role of cognitive control in decision-making, motivation, and social behavior 24.2.1 Motivation, decision-making, and cognitive control Cognitive control would be aimless in the absence of motivation. That is, while cognitive control is what allows humans to sustain goal-directed behavior through monitoring and control instantiation, motivation is what provides a direction to behavior. Therefore, to better understand cognitive control and its developmental trajectory, it is critical to also consider the development of motivation. Existing work supports the idea of a frontostriatal network linking cognitive control and motivation during decision-making (see Somerville and Casey, 2010). However, few studies have examined the developmental neuroscience of motivationecontrol relations. Here, behavioral studies have more frequently been employed, as will be discussed further. Nevertheless, recent years have shown an increase in the number of studies exploring the neural basis of motivationecontrol relations, and critically, how such processes develop; these studies will be reviewed as well. A simple paradigm that has been used to assess relations between motivation and cognitive control with young children is the delay-of-gratification task, which measures children’s ability to give up an immediate reward in favor of a larger reward later (Mischel et al., 1972). For example, participants are seated in front of a piece of candy while an experimenter leaves the room. If they wait for the experimenter to return, they get two treats; otherwise they get only one (Mischel et al., 1972). Variations of the standard delay-of-gratification paradigm have shown developmental differences throughout the preschool years. For example, in one study (Kochanska et al., 1996), children were asked to hold candy in their mouth without eating it until they were told to do so. In a second task, children were asked not to peek while they could hear that the experimenter was wrapping a gift for them. Children’s ability to delay gratification increased significantly from 3 to 4 years of age. In a study linking delay of gratification and decision-making, Prencipe and Zelazo (2005) examined children’s delay of gratification for self and other (the experimenter). Three-year-olds typically chose an immediate reward for themselves and a delayed reward for others (the experimenter). According to Prencipe and Zelazo (2005), these findings suggest that 3-year-olds are capable of adaptive decision-making but still have difficulty regulating their own approach behavior in motivationally salient situations. According to these authors, it is possible that their behavior is driven by the relatively automatic processes rather than by more deliberate prefrontal networks. The findings can be explained in light of Barresi and Moore’s (1996) model of the development of perspective-taking. It is possible that 3-year-olds may have made

The neural correlates of cognitive control and the development of social behavior Chapter | 24

529

impulsive choices in the self-condition because they took an exclusively first-person perspective on their own behavior and had difficulty adopting a more objective, third-person perspective according to which delay would be preferred. In contrast, 3-year-olds may have chosen delayed rewards in the other condition because they took an exclusively third-person perspective and had difficulty appreciating the experimenter’s subjective perspective (i.e., his or her own desire for immediate gratification). In more complex situations, individuals are often required to make approach-avoidance decisions in the face of uncertainty. One common measure of complex decision-making in adults is the Iowa gambling task (IGT; Bechara et al., 1994). In this task, participants are asked to choose cards from four decks that contain a different number of cards that could lead to their winning or losing money. During the task, the participants learn that some decks are more advantageous than others. Kerr and Zelazo (2004) modified the IGT to create a version for children that included two decks of cards, one advantageous and one disadvantageous. Feedback on participants’ decisions was provided in the form of happy (reward) and sad (loss) faces. Choosing cards from the disadvantageous deck resulted in more rewards on every trial but also with occasional (unpredictable) large losses. Three-year-olds failed to develop a preference for the advantageous deck. However, 4- and 5-year-olds were able to make advantageous decisions (Kerr and Zelazo, 2004). Although these studies suggest a steady improvement in delay of gratification during early childhood, motivatione control relations do not appear to follow a linear pattern of development throughout development. For example, Figner et al. (2009) used a gambling task in which reward feedback was given either during or after decision-making. The findings revealed that adolescents made riskier choices compared with adults, but only in the condition in which the reward was given during the decision. Another study examined participants between the ages of 10 and 30 years and used a delay discounting task (Steinberg et al., 2009) in which participants were asked to choose between an immediate reward of less value (e.g., $400 today) and a variety of delayed rewards of more value (e.g., $700 1 month from now or $800 6 months from now). The findings indicate that, before age 16 years, children exhibited a greater willingness to accept a smaller reward immediately than a large reward that was delayed. In a third study, Cauffman et al. (2010) obtained similar results examining the same age range with a modified version of the IGT. Results indicate that approach behaviors (a tendency to choose from the advantageous decks) display an inverted U-shape relation to age, peaking in mid- to late adolescence. In contrast, avoidance behaviors (a tendency to refrain from choosing from the disadvantageous decks) increased linearly with age, with adults avoiding disadvantageous decks at higher rates than both preadolescents and adolescents. Taken together, these studies show that risky choices tend to peak between 14 and 16 years of age, followed by a decline in risky behavior. Critically, this pattern of behavior on laboratory tasks matches reports that adolescence is characterized by heightened rates of risky behavior, such as drug use and risky sexual conduct (Casey et al., 2005; Steinberg, 2008). The behavioral findings described earlier appear to suggest that while cognitive control might improve across childhood and adolescence, selective impairments on cognitive control tasks might occur during the adolescence period. These behavioral patterns have led several researchers to propose that such apparent deficits in adolescent cognitive control arise from a mismatch between cognitive control and motivation processes during this period (Luciana and Collins, 2012; Luna and Wright, 2016; Somerville and Casey, 2010; Steinberg, 2010). These theories predict that due to different developmental trajectories of cognitive control and motivational processes, the adolescent period reflects a developmental window within which motivational processes are more salient relative to cognitive control ability. Thus, researchers have suggested that although adolescents show improvements in cognitive control relative to children, their goal-oriented behavior can be diminished in light of motivational cues of potential reward (Cauffman et al., 2010; Figner et al., 2009; Steinberg et al., 2009). Indeed, fMRI studies examining the role of the striatum in salient and motivational contexts support the idea that adolescents show enhanced sensitivity to incentives relative to children and adults (Ernst et al., 2005; Galvan et al., 2006; Geier et al., 2010; May et al., 2004). For example, Ernst et al. (2005) used a monetary reward task and found stronger activation among adolescents than adults in the left nucleus accumbens, a structure in the striatum thought to be involved in reward processing. In addition, a reduction in amygdala response to reward omission was larger for adults than for adolescents. Moreover, in line with the importance of motivationecontrol interactions for effective cognitive control behavior, research has shown that connection strength between frontal and striatal regions is associated with cognitive control ability in typically and atypically developing individuals. In addition to the developmental mismatch between motivation and control during the adolescent period, which can lead to apparent control deficits, motivation is also known to have a facilitative effect on cognitive control. For example, when participants were promised a financial reward for accurate performance on certain trials of an antisaccade task, cognitive control was improved for adolescents more than adults (Jazbec et al., 2006). Geier et al. (2010) studied the neural underpinnings of reward processing and its influence on cognitive control in adolescence using a modified version of an antisaccade task. The results indicated that faster correct inhibitory responses were made on reward trials than on neutral trials by both adolescents and adults. Additionally, fewer inhibitory errors were committed by adolescents. For reward

530 PART | II Cognitive Development

trials, activation was attenuated within the ventral striatum in adolescents during cue assessment. This was followed by overactivation in adolescents during response preparation (i.e., during fixation after the reward cue) in the ventral striatum, as well as the precentral sulcus, which is important for controlling eye movements. These findings suggest enhanced activation in adolescents in control regions as a result of reward anticipation (Geier et al., 2010). In sum, studies examining the development of cognitive control in decision-making have pointed to an inverted U-shaped trend in the development of motivation (see Somerville and Casey, 2010). Studies with young children show steady improvements in the ability to delay gratification (e.g., Prencipe and Zelazo, 2005) as well as more effective decision-making (e.g., Kerr and Zelazo, 2004). However, studies have shown a greater propensity for risky decisions during adolescence in light of motivationally salient situations (e.g., Cauffman et al., 2010). Additionally, fMRI studies examining the link between cognitive control processes and motivation have found enhanced activation during adolescence in brain regions associated with cognitive control during anticipation of reward (Geier et al., 2010).

24.2.2 Cognitive control and social behavior Research has increasingly focused on the relations between cognitive control and social behavior. Following our review of motivation development more generally, we first overview how social stimuli can serve as motivating factors that directly influence cognitive control. Next, we discuss how social stimuli can be the focus of cognitive control, such as when social rejection is processed by the monitoring system. Furthermore, we review research across each of these domains to illustrate the emerging neuroscience on the relations between cognitive control and social behavior. It has long been known that, at least at the behavioral level, social stimuli or social contexts can serve to motivate behavior (Triplett, 1898; Zajonc and Paulus, 1980; Zajonc, 1965). In line with this notion, similar to the finding that cognitive control performance is improved in the presence of monetary rewards (Jazbec et al., 2006), similar findings have been reported for social rewards, such as happy faces (Kohls et al., 2009). Social observation can also serve as a form of motivation, and researchers have shown that the presence of others leads to an increase in error monitoring, as measured by the ERN (Hajcak et al., 2005). Critically, work has recently demonstrated that the effects of social observation on error monitoring differ across development. For example, Barker et al. (2018) had a group of participants between the ages of 8e17 years perform a cognitive control task under two conditions, once while being watched by peers, and once alone. The authors found that this social observation manipulation only led to an increase in the ERN for the younger adolescents, with no effect present for the older adolescents. Moreover, age-related changes in error monitoring were diminished when this effect of social context was considered (Barker et al., 2018). Consistent with these findings, another study by Barker et al. (2015) did not find evidence for social-related changes in the ERN for a sample of adults (unless social anxiety was also considered), whereas a third study in children aged 12 years did identify such an effect (Buzzell et al., 2017c). Earlier, we reviewed how the error monitoring system develops across childhood and adolescence; however, such work did not consider the role of social context; these more recent data suggest that age-related changes in error monitoring may be more complex, with social context playing a critical role. Beyond the influence that social context can have on cognitive control, research has also investigated how social feedback itself is processed by the brain, investigating direct relations between the monitoring system and social behavior. Using a feedback processing paradigm similar to other investigations of the monitoring system, Somerville et al. (2006) found that while expectancy violations and social rejection are both processed by the MFC, social rejection was exclusively associated with more ventral activation of the MFC. Whereas this initial study was performed in adults, subsequent work has studied how the processing of social feedback differs across age. For example, a similar study by Gunther Moor et al. (2010) found that across ages 8e25 years, increasing activation of not only ventral regions of the MFC, but also the orbitofrontal cortex, lateral regions of the frontal cortex, and the striatum, was found when processing rejection feedback. Similarly, other work has shown that when participants participate in a virtual game, which is manipulated so that the participants feel left out, social distress is similarly associated with increased activity within the ventral MFC (Gunther Moor et al., 2010). Critically, younger participants, aged 10e12 years, showed increased activation within the ventral MFC, specifically the subgenual anterior cingulate cortex, compared with older adolescents and adults (Gunther Moor et al., 2010). To summarize, research has demonstrated that monitoring and control can be upregulated by the presence of others, effects that change over the course of development. In particular, increases in monitoring that are driven by the presence of others appear to be strongest during young adolescence. Similarly, younger adolescents appear to uniquely demonstrate activation of the subgenual ACC during social exclusion, suggesting that this reflects a critical developmental period in terms of the monitoring of social contexts. More generally, the processing of social rejection appears to associate with increased ventral MFC activity more broadly. Having reviewed the relations between cognitive control and social behavior, we now turn to a discussion of how cognitive control differs across individuals. Here, we continue a social focus, reviewing relations with other socioemotional factors such as temperament, social anxiety, and cross-cultural differences.

The neural correlates of cognitive control and the development of social behavior Chapter | 24

531

24.3 Individual differences in cognitive control 24.3.1 Temperament, cognitive control, and psychopathology This chapter has so far reviewed the typical development of cognitive control that is observed, on average, across children. However, a substantial and growing literature has characterized individual differences in the development of cognitive control (e.g., Gehring et al., 2000; Henderson, 2010; Lewis et al., 2006; McDermott et al., 2009; Stieben et al., 2007; White et al., 2011). Critically, such individual differences have been linked to early differences in a child’s temperament, as well as normative variation in later social outcomes and even psychopathology. For example, behavioral and physiological measures of cognitive control have been linked to negative mood induction conditions (Lewis et al., 2006, 2007) and to heightened levels of trait anxiety and internalizing symptoms (e.g., Gehring et al., 2000; Stieben et al., 2007). Here, we review literature on such relations, before turning to much broader differences across cultures. A child’s temperament refers to persistent individual differences that are thought to have a basis in biology. Relations between temperament and cognitive control are natural, as one influential model of temperament by Rothbart and colleagues (Rothbart, 1981; Rothbart and Bates, 2006) explicitly includes “regulation,” referring to a child’s ability to self-sooth and control their own behavior (Rothbart, 1981; Rothbart and Bates, 2006), as one of the dimensions of temperament. Critically, a growing corpus of studies has investigated how behavioral inhibition (BI) either relates to cognitive control directly or interacts with cognitive control to predict later outcomes in socioemotional development and psychopathology. We now turn to a review of relations between BI and cognitive control, as well as relations with other forms of temperament. A number of studies from different laboratories have found that specific cognitive control functions moderate the developmental trajectory of BI in children (Henderson, 2010; Thorell et al., 2004; White et al., 2011). For example, Thorell et al. (2004) examined how BI and inhibitory control assessed at 5 years of age were associated with socioemotional functioning at 9 years of age. The results indicated that behaviorally inhibited children with high levels of inhibitory control were reported as being more socially anxious than behaviorally inhibited children with low levels of inhibitory control. Similarly, Fox and Henderson (2000) found that behaviorally inhibited 4-year-olds with high inhibitory control were less socially competent and more socially withdrawn than behaviorally inhibited children with low inhibitory control. Other neurophysiological evidence for the moderating role cognitive control has been found for the association between shyness and socialeemotional maladjustment (Henderson, 2010). In 9- to 13-year-old children, Henderson (2010) found that shyness was associated with poor social outcomes primarily among children with greater neural activity associated with cognitive control, as measured by the N2 ERP component. Lahat et al., (2014b) found that among children with a large N2, BI was positively related to withdrawal and negatively related to assertiveness during social exclusion. Taken together, these results point to the role of conflict in shy children’s social adjustment (Henderson, 2010; Lahat et al., 2014a,b). In a study using dense-array ERPs and source analyses, Lamm et al. (2012) showed that, during an inhibitory control task, children high in temperamental fearfulness showed modeled source activation in areas suggestive of ventrolateral PFC across both emotional and nonemotional conditions of the task. However, children low in temperamental fearfulness only showed this pattern of activation during the emotional condition. Results from this study suggest that while children low in temperamental fearfulness recruited increased inhibitory control only during the emotional conditions, or those conditions in which more cognitive control recourses were likely needed, children with high fearful temperaments sustain this increased level of inhibitory control across both neutral and emotional contexts (Lamm et al., 2012). These findings suggest that temperamentally fearful individuals show increased vigilance not only in emotional situations but also in nonemotional ones. Thus, these fearful individuals may refrain even from social situations that do not induce negative emotions. Findings regarding the role of inhibitory control for temperamental BI or shyness are the opposite of results regarding another form of control instantiation: attention shifting (Eisenberg et al., 1998). For example, Eisenberg et al. (1998) found that children low on attention shifting and high on parental reports of negative emotions, such as fear, sadness, and anxiety, were rated by their parents and teachers as shyer 2 years later. Investigating this difference further, White et al. (2011) examined how attention shifting and inhibitory control, which were tested at 48 months of age, moderated the association between BI assessed at 24 months of age and anxiety problems in the preschool years. The results indicated that high levels of inhibitory control increased the risk for anxiety disorders among behaviorally inhibited children, whereas high levels of attention shifting decreased the risk for anxiety problems in these children. In a different study, Lahat et al. (2012) examined how these two forms of control instantiationdattention shifting and inhibitory controldmoderated the relation between exuberant temperament in infancy and propensity for risk taking in childhood. Temperamental exuberance has been defined by positive reactivity to novelty, approach behavior, and sociability (Putnam and Stifter, 2005). Children with an exuberant temperament are also characterized by impulsivity, sensitivity to reward, fearlessness, and risk taking (Fox

532 PART | II Cognitive Development

et al., 2001; Polak-Toste and Gunnar, 2006; Rothbart and Bates, 2006). Control instantiation was assessed at 48 months of age. Risk taking propensity was measured when children were 60 months of age. The results indicated that exuberance was positively associated with risk taking propensity among children relatively low in attention shifting but unrelated for children high in attention shifting. Inhibitory control did not significantly moderate the link between exuberance and risk taking. Taken together, the findings from these sets of studies on temperament and different types of control instantiation demonstrate that attention shifting and inhibitory control have differential influences on levels of risk or adaptation. Furthermore, these two studies (Lahat et al., 2012; White et al., 2011) suggest that high levels of attention shifting may serve as a protective factor in the link between temperament and negative outcomes. This conclusion may have important implications for prevention and intervention efforts in the form of training to improve attention-shifting skills. While the aforementioned findings provide evidence for the potential negative risks of excessive inhibitory control ability for children high in BI, it is important to note that the majority of such findings are primarily based on behavioral measures, which are confounded with monitoring. Similarly, neurophysiological research that leverages the N2 as an index of “inhibitory control” is not without limitation, as it remains debated as to whether the N2 ERP component more closely maps onto monitoring or control, with evidence suggesting a closer relation to monitoring (Donkers and van Boxtel, 2004). Thus, it is possible that many of the purported links between overactive inhibitory controls might more closely reflect increased monitoring. Indeed, as we will now review, individual differences in monitoring consistently relate to temperament and moderate risk for negative social outcomes or psychopathology. The first study to provide evidence that BI directly relates to monitoring came from a study by McDermott et al. (2009). These authors studied adolescents with a mean age of 15 years who were part of a larger longitudinal study and were assessed during infancy and early childhood for BI. Using the ERN as an index of monitoring, it was found that adolescents with high childhood BI exhibited increased error monitoring relative to those low in childhood BI. Moreover, magnitude of the ERN moderated relations between BI and clinical anxiety, such that a higher risk for anxiety disorders was present for children assessed as high in BI and also exhibiting a larger ERN (McDermott et al., 2009). In a separate cohort of children, this basic set of relations was replicated at an earlier age, assessing BI in toddlerhood, the ERN at age 7 years and anxiety outcomes at age 9 years (Lahat et al., 2014a). Also in this second cohort of children, the ERN was found to relate to BI a third time when assessed at age 12 years, particularly when social context was taken into account (Buzzell et al., 2017c). In this third study, adolescents were required to perform a cognitive control task twice, once alone, and once while under social observation. The authors found that greater increases for the ERN, while under social observation, were predicted by early BI. Critically, longitudinal relations between early BI and social anxiety in adolescence were mediated by a combination of this increased ERN and also slowed behavioral responses after errors (Buzzell et al., 2017c). Thus, these more recent results point to the importance of social context when considering relations between temperament, cognitive control, and psychopathology. The reviewed studies illustrate a consistent relation between the BI temperament and later individual differences in error monitoring. This pattern is consistent with similar relations that have been found for children characterized as high in fearfulness during early toddlerhood (2 years of age), who also exhibit larger ERNs later at 4.5 years of age (Brooker and Buss, 2014). However, one study has also shown that children with a temperament of negative emotionality that also have mothers with high anxiety actually demonstrate a smaller ERN (Torpey et al., 2013). Beyond error monitoring, children high in BI also demonstrate increased activation of the performance monitoring system in response to a variety of events. For example, children high in BI exhibit increased cingulate activation in response to stimulus conflict, likely reflecting conflict monitoring (Jarcho et al., 2013). Similarly, children high in BI demonstrate an enhanced N2 ERP, which can be interpreted as an index of conflict monitoring as well (Lahat et al., 2014b). Indeed, following source localization of the N2 in a go/no-go task, it was found that children high in BI exhibited greater cingulate activity for high conflict trials, consistent with increased conflict monitoring for these children (Lamm et al., 2014). To summarize, substantial variation in cognitive control can be observed during childhood and adolescence, and such variation relates to temperament, social outcomes, and psychopathology. In general, it appears that certain temperaments, such as BI, are associated with increased monitoring for errors or conflict later in development. Moreover, at least for children high in BI, increased monitoring is actually a risk factor for anxiety. Similarly, it appears that enhanced inhibitory control is also a risk factor for children high in the BI temperament; however, more work using neural measures is needed to confirm this relation. Having reviewed individual variation in cognitive control, we now turn to more global variation by reviewing cross-cultural differences.

24.3.2 Cross-cultural differences in the development of cognitive control Research on the development of cognitive control has mostly been conducted in Western cultures. However, an emerging body of cross-cultural studies suggests that Asian children may outperform Western children on measures of cognitive control (e.g., Lahat et al., 2010; Sabbagh et al., 2006). Research comparing children from Western and Asian cultures has

The neural correlates of cognitive control and the development of social behavior Chapter | 24

533

shown that Chinese children perform better on behavioral and neurophysiological measures of EFs than North American children (e.g., Chen et al., 1998; Ho, 1994; Lahat et al., 2010; Sabbagh et al., 2006; Wu, 1996). For example, Sabbagh et al. (2006) administered a battery of EF and theory of mind (ToM) tasks to preschoolers from China and the United States. ToM is the ability to attribute mental statesdbeliefs, intents, and desiresdto oneself and others and to recognize that mental representations can differ across individuals (Premack and Woodruff, 1978). The Chinese preschoolers showed better performance than the US preschoolers on all measures of EF, but not on measures of ToM. However, individual differences in EF predicted ToM performance in both groups. Chinese children’s advanced performance on EF tasks may stem from the opportunities to exercise and practice these abilities that they encounter within their culture. For example, Chinese parents expect children to master impulse control at a much younger age than North American parents (Chen et al., 1998; Ho, 1994; Wu, 1996). Compared with Western parents, Chinese parents are more controlling and protective in child rearing. For example, they often encourage their young children to stay close to and to be dependent on them. Indeed, most Chinese infants and toddlers sleep in the same bed or in the same room as their parents (Chen et al., 1998). In addition, impulse control is more highly valued in Chinese daycare settings than in North American daycare settings (Tobin et al., 1989). Another possibility for Chinese children’s superior cognitive control was suggested by Lahat et al. (2010) in a study that focused on the N2 component of ERP. The study compared 5-year-old children from a Chinese-Canadian ethnic background with children from a European-Canadian background on a go/no-go task. No behavioral differences between the two cultural groups were observed, but robust N2 amplitude differences were found. Chinese-Canadian children showed larger (i.e., more negative) N2 amplitudes than European-Canadian children on the right side of the scalp on no-go trials as well as on the left side of the scalp on go trials. Source analyses of the N2 showed greater modeled source activation for Chinese-Canadian children in dorsomedial, ventromedial, and (bilateral) ventrolateral PFC. These findings reveal that Chinese-Canadian children show greater hemispheric differentiation than European-Canadian children, perhaps relating to more advanced cognitive control. In sum, individual differences in cognitive control have also been observed in cross-cultural variations. Specifically, an emerging body of research comparing children from Asian and Western cultures has shown behavioral (e.g., Sabbagh et al., 2006) as well as neurophysiological (Lahat et al., 2010) evidence, suggesting advanced cognitive control abilities among children from a Chinese cultural background compared with children from a Western cultural background. Although the reasons for these differences are not clear yet, it is possible that differences in socialization between the two cultures play a major role.

24.4 Chapter summary and future directions In this chapter, cognitive control, its development, and relations to social behavior were reviewed. The chapter described how cognitive control is composed of two primary subdivisions, monitoring, and control instantiation. Toward this end, a focus was placed on reviewing error monitoring and inhibitory control in particular. As described earlier, these functions have been found to develop with the maturation of the PFC, and patterns of activation generally increase with age (Zelazo et al., 2008). Based on behavioral and neural measures, error monitoring is shown to emerge early in childhood and exhibit sustained development through adolescence (e.g., Backen Jones et al., 2003; Bullock and Lutkenhaus, 1984; Davies et al., 2004). Error monitoring is shown to rely primarily on a central hub within the MFC that exhibits increasing activation across childhood and early adolescence (Ladouceur et al., 2007). However, continued development through adolescence and into adulthood appears to derive from more nuanced development within ventralefrontal structures linked to salience and emotional processing (Buzzell et al., 2017b). Studies examining the neural correlates of inhibitory control development have shown that PFC activation not only increases with age but also becomes more focalized (e.g., Bunge et al., 2002; Durston et al., 2006; Luna et al., 2001). As discussed, with development, brain regions not associated with task performance decrease in activation, whereas regions relevant to task performance increase in activation. In this chapter, the role of cognitive control in the development of decision-making and motivation was also reviewed. Studies have shown that the development of decision-making takes an inverted U-shaped form, with steady improvements during childhood (e.g., Kerr and Zelazo, 2004) and a propensity for risk taking during adolescence due to a mismatch between the developmental trajectories of motivation and cognitive control (Somerville and Casey, 2010). fMRI studies with adolescents (e.g., Geier et al., 2010) have shown increased activation in brain regions associated with cognitive control during anticipation of reward. Moreover, early adolescence appears to be marked by heightened sensitivity to social influences on cognitive control, with social observation yielding increases in error monitoring sensitivity (Barker et al., 2015; Buzzell et al., 2017c), as well as heightened sensitivity to social rejection during this developmental window (Gunther Moor et al., 2010).

534 PART | II Cognitive Development

Inhibitory control and error monitoring have been found to show associations with individual differences and their socialeemotional outcomes (e.g., Fox and Henderson, 2000; Lamm et al., 2012; McDermott et al., 2009; White et al., 2011). This body of research has shown that children with a fearful temperament who are also high in inhibitory control or error monitoring are at risk of developing negative socialeemotional outcomes, such as social withdrawal and anxiety disorders. However, evidence for individual differences associated with cognitive control comes mostly from behavioral and electrophysiological studies, and more work is needed. Individual differences relating to cultural variation have also been found, particularly when comparing Chinese children with children from Western cultures. These studies provide behavioral (Sabbagh et al., 2006) as well as neurophysiological (Lahat et al., 2010) evidence for advanced performance on cognitive control tasks among Chinese children. It is possible that socialization processes that differ across the two cultures contribute to these differences in cognitive control. However, given that the cross-cultural neurophysiological work has been conducted with Asian children who grew up in North America, future research should compare Western children with Asian children who are raised in Asia.

References Aron, A.R., 2007. The neural basis of inhibition in cognitive control. Neuroscientist 13 (3), 214e228. https://doi.org/10.1177/1073858407299288. Aron, A.R., Robbins, T.W., Poldrack, R.A., 2004. Inhibition and the right inferior frontal cortex. Trends Cognit. Sci. 8 (4), 170e177. https://doi.org/ 10.1016/j.tics.2004.02.010. Anderson, P., 2002. Assessment and development of executive function (EF) during childhood. Child Neuropsychol. 8, 71e82. Backen Jones, L., Rothbart, M., Posner, M., 2003. Development of executive attention in preschool children. Dev. Sci. 6, 498e504. Barker, T.V., Troller-Renfree, S.V., Bowman, L.C., Pine, D.S., Fox, N.A., 2018. Social influences of error monitoring in adolescent girls. Psychophysiology 55 (9), e13089. https://doi.org/10.1111/psyp.13089. Barker, T.V., Troller-Renfree, S., Pine, D.S., Fox, N.A., 2015. Individual differences in social anxiety affect the salience of errors in social contexts. Cognit. Affect Behav. Neurosci. 15 (4), 723e735. Barresi, J., Moore, C., 1996. Intentional relations and social understanding. Behav. Brain Sci. 19, 107e154. Bechara, A., Damasio, A.R., Damasio, H., Anderson, S.W., 1994. Insensitivity to future consequences following damage to human prefrontal cortex. Cognition 50, 7e15. Bokura, H., Yamaguchi, S., Kobayashi, S., 2001. Electrophysiological correlates for response inhibition in a Go/NoGo task. Clin. Neurophysiol. 112 (12), 2224e2232. Botvinick, M.M., Braver, T.S., Barch, D.M., Carter, C.S., Cohen, J.D., 2001. Conflict monitoring and cognitive control. Psychol. Rev. 108 (3), 624. Braver, T.S., Barch, D.M., Gray, J.R., Molfese, D.L., Snyder, A., 2001. Anterior cingulate cortex and response conflict: effects of frequency, inhibition and errors. Cerebr. Cortex 11, 825e836. Braver, T.S., Reynolds, J.R., Donaldson, D.I., 2003. Neural mechanisms of transient and sustained cognitive control during task switching. Neuron 39 (4), 713e726. https://doi.org/10.1016/S0896-6273(03)00466-5. Brooker, R.J., Buss, K.A., 2014. Toddler fearfulness is linked to individual differences in error-related negativity during preschool. Dev. Neuropsychol. 39 (1), 1e8. https://doi.org/10.1080/87565641.2013.826661. Brooker, R.J., Buss, K.A., Dennis, T.A., 2011. Error-monitoring brain activity is associated with affective behaviors in young children. Dev. Cognit. Neurosci. 1 (2), 141e152. https://doi.org/10.1016/j.dcn.2010.12.002. Bullock, M., Lutkenhaus, P., 1984. The development of volitional behavior in the toddler years. Child Dev. 59, 664e674. Bunge, S.A., Crone, E.A., 2009. Neural correlates of the development of cognitive control. In: Rumsey, J., Ernst, M. (Eds.), Neuroimaging in Developmental Clinical Neuroscience. Cambridge University Press, pp. 22e37. Bunge, S.A., Dudukovic, N.M., Thomason, M.E., Vaidya, C.J., Gabrieli, J.D.E., 2002. Development of frontal lobe contributions to cognitive control in children: evidence from fMRI. Neuron 33, 301e311. Buzzell, G.A., Beatty, P.J., Paquette, N.A., Roberts, D.M., McDonald, C.G., 2017. Error induced blindness: error detection leads to impaired sensory processing and lower accuracy at short responseestimulus intervals. J. Neurosci. 37 (11), 2895e2903. Buzzell, G.A., Richards, J.E., White, L.K., Barker, T.V., Pine, D.S., Fox, N.A., 2017. Development of the error-monitoring system from ages 9e35: unique insight provided by MRI constrained source localization of EEG. NeuroImage 157, 13e26. Buzzell, G.A., Troller-Renfree, S.V., Barker, T.V., Bowman, L.C., Chronis-Tuscano, A., Henderson, H.A., et al., 2017. A neurobehavioral mechanism linking behaviorally inhibited temperament and later adolescent social anxiety. J. Am. Acad. Child Adolesc. Psychiatry 56 (12), 1097e1105. https:// doi.org/10.1016/j.jaac.2017.10.007. Casey, B.J., Galvan, A., Hare, T.A., 2005. Changes in cerebral functional organization during cognitive development. Curr. Opin. Neurobiol. 15, 239e244. Cauffman, E., Shulman, E.P., Steinberg, L., et al., 2010. Age differences in affective decision making as indexed by performance on the Iowa gambling task. Dev. Psychol. 46, 193e207. Cavanagh, J.F., Cohen, M.X., Allen, J.J., 2009. Prelude to and resolution of an error: EEG phase synchrony reveals cognitive control dynamics during action monitoring. J. Neurosci. 29 (1), 98e105.

The neural correlates of cognitive control and the development of social behavior Chapter | 24

535

Cavanagh, J.F., Zambrano-Vazquez, L., Allen, J.J., 2012. Theta lingua franca: a common mid-frontal substrate for action monitoring processes. Psychophysiology 49 (2), 220e238. Chen, X., Hastings, P.D., Rubin, K.H., Chen, H., Cen, G., Stewart, S.L., 1998. Child rearing attitudes and behavioral inhibition in Chinese and Canadian toddlers: a cross cultural study. Dev. Psychol. 34, 677e686. Chronis-Tuscano, A., Degnan, K.A., Pine, D.S., et al., 2009. Stable early maternal report of behavioral inhibition predicts lifetime social anxiety disorder in adolescence. J. Am. Acad. Child Adolesc. Psychiatry 48, 928e935. Danielmeier, C., Ullsperger, M., 2011. Post-error adjustments. Front. Psychol. 2, 233. Davidson, M.C., Amso, D., Cruess-Anderson, L., Diamond, A., 2006. Development of cognitive control and executive functions from 4e13 years: evidence from manipulations of memory, inhibition and task switching. Neuropsychologia 44, 2037e2078. Davies, P.L., Segalowitz, S.J., Gavin, W.J., 2004. Development of response-monitoring ERPs in 7- to 25-year-olds. Dev. Neuropsychol. 25, 355e376. Debener, S., 2005. Trial-by-trial coupling of concurrent electroencephalogram and functional magnetic resonance imaging identifies the dynamics of performance monitoring. J. Neurosci. 25 (50), 11730e11737. https://doi.org/10.1523/JNEUROSCI.3286-05.2005. Desimone, R., Duncan, J., 1995. Neural mechanisms of selective visual attention. Annu. Rev. Neurosci. 18 (1), 193e222. Diamond, A., 2002. Normal development of prefrontal cortex from birth to young adulthood: cognitive functions, anatomy and biochemistry. In: Stuss, D.T., Knight, R.T. (Eds.), Principles of Frontal Lobe Function. Oxford University Press, London, pp. 466e503. Donkers, F.C.L., van Boxtel, G.J.M., 2004. The N2 in go/no-go tasks reflects conflict monitoring not response inhibition. Brain Cognit. 56 (2), 165e176. https://doi.org/10.1016/j.bandc.2004.04.005. Dosenbach, N.U., Fair, D.A., Cohen, A.L., Schlaggar, B.L., Petersen, S.E., 2008. A dual networks architecture of top-down control. Trends Cognit. Sci. 12 (3), 99e105. DuPuis, D., Ram, N., Willner, C.J., Karalunas, S., Segalowitz, S.J., Gatzke-Kopp, L.M., 2015. Implications of ongoing neural development for the measurement of the error-related negativity in childhood. Dev. Sci. 18 (3), 452e468. https://doi.org/10.1111/desc.12229. Durston, S., Davidson, M.C., Tottenham, N., et al., 2006. A shift from diffuse to focal cortical activity with development. Dev. Sci. 9, 1e20. Eisenberg, N., Shepard, S.A., Fabes, R.A., Murphy, B.C., Guthrie, I.K., 1998. Shyness and children’s emotionality, regulation, and coping: contemporaneous, longitudinal, and across-context relations. Child Dev. 69, 767e790. Ernst, M., Mueller, S.C., 2008. The adolescent brain: insights from functional neuroimaging research. Dev. Neurobiol. 68, 729e743. Ernst, M., Nelson, E.E., Jazbec, S., et al., 2005. Amygdala and nucleus accumbens in responses to receipt and omission of gains in adults and adolescents. NeuroImage 25, 1279e1291. Fair, D.A., Cohen, A.L., Power, J.D., Dosenbach, N.U.F., Church, J.A., Miezin, F.M., Schlaggar, B.L., Petersen, S.E., 2009. Functional Brain Networks Develop from a “Local to Distributed” Organization. PLoS Comput. Biol. 5 (5) https://doi.org/10.1371/journal.pcbi.1000381. Fair, D.A., Dosenbach, N.U., Church, J.A., Cohen, A.L., Brahmbhatt, S., Miezin, F.M., Barch, D.M., Raichle, M.E., Petersen, S.E., Schlaggar, B.L., 2007. Development of distinct control networks through segregation and integration. Proc. Natl. Acad. Sci. 104 (33), 13507e13512. Figner, B., Mackinlay, R.J., Wilkening, F., Weber, E.U., 2009. Affective and deliberative processes in risky choice: age differences in risk taking in the Columbia Card Task. J. Exp. Psychol. Learn. Mem. Cognit. 35, 709e730. Fitzgerald, K.D., Perkins, S.C., Angstadt, M., Johnson, T., Stern, E.R., Welsh, R.C., Taylor, S.F., 2010. The development of performance-monitoring function in the posterior medial frontal cortex. NeuroImage 49 (4), 3463e3473. https://doi.org/10.1016/j.neuroimage.2009.11.004. Fox, N.A., Henderson, H.A., 2000. Temperament, emotion, and executive function: influences on the development of self-regulation. In: Paper Presented at the Annual Meeting of the Cognitive Neuroscience Society. San Francisco. Fox, N.A., Henderson, H.A., Rubin, K.H., Calkins, S.D., Schmidt, L.A., 2001. Continuity and discontinuity of behavioral inhibition and exuberance: psychophysiological and behavioral influences across the first four years of life. Child Dev. 72, 1e21. Fox, N.A., Henderson, H.A., Marshall, P.J., Nichols, K.E., Ghera, M.M., 2005. Behavioral inhibition: linking biology and behavior within a developmental framework. Annu. Rev. Psychol. 56, 235e262. Galvan, A., Hare, T.A., Parra, C.E., et al., 2006. Earlier development of the accumbens relative to orbitofrontal cortex might underlie risk-taking behavior in adolescents. J. Neurosci. 26, 6885e6892. Gehring, W.J., Goss, B., Coles, M.G.H., Meyer, D.E., Donchin, E., 1993. A neural system for error detection and compensation. Psychol. Sci. 4, 385e390. Gehring, W.J., Himle, J., Nisenson, L.G., 2000. Action-monitoring dysfunction in obsessive-compulsive disorder. Psychol. Sci. 11, 1e6. Geier, C.F., Terwilliger, R., Teslovich, T., Velanova, K., Luna, B., 2010. Immaturities in reward processing and its influence on inhibitory control in adolescence. Cerebr. Cortex 20, 1613e1629. Gerardi-Caulton, G., 2000. Sensitivity to spatial conflict and the development of self-regulation in children 24e26 months of age. Dev. Sci. 3, 397e404. Gerstadt, C.L., Hong, Y.J., Diamond, A., 1994. The relationship between cognition and action: performance of children 3 1/2-7 years old on a Stroop-like day-night test. Cognition 53, 129e153. Gogtay, N., Giedd, J.N., Lusk, L., Hayashi, K.M., Greenstein, D., Vaituzis, A.C., Nugent, T.F., Herman, D.H., Clasen, L.S., Toga, A.W., Rapoport, J.L., Thompson, P.M., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. 101 (21), 8174e8179. https://doi.org/10.1073/pnas.0402680101. Gunther Moor, B., van Leijenhorst, L., Rombouts, S.A.R.B., Crone, E.A., Van der Molen, M.W., 2010. Do you like me? Neural correlates of social evaluation and developmental trajectories. Soc. Neurosci. 5 (5e6), 461e482. https://doi.org/10.1080/17470910903526155. Gupta, R., Kar, B.R., Srinivasan, N., 2009. Development of task switching and post-error slowing in children. Behav. Brain Funct. 5 (1), 38.

536 PART | II Cognitive Development

Hajcak, G., Moser, J.S., Yeung, N., Simons, R.F., 2005. On the ERN and the significance of errors. Psychophysiology 42 (2), 151e160. https://doi.org/ 10.1111/j.1469-8986.2005.00270.x. Hauser, T.U., Iannaccone, R., Stämpfli, P., Drechsler, R., Brandeis, D., Walitza, S., Brem, S., 2014. The feedback-related negativity (FRN) revisited: new insights into the localization, meaning and network organization. NeuroImage 84, 159e168. Henderson, H.A., 2010. Neural correlates of cognitive control and the regulation of shyness in children. Dev. Neuropsychol. 35, 177e193. Ho, D.Y.F., 1994. Cognitive socialization in Confucian heritage cultures. In: Greenfield, P.M., Cocking, R.R. (Eds.), Cross-Cultural Roots of Minority Development. Erlbaum, Hillsdale, NJ, pp. 285e313. Holroyd, C.B., Coles, M.G., 2002. The neural basis of human error processing: Reinforcement learning, dopamine, and the error-related negativity. Psychol. Rev. 109 (4), 679. Huizinga, M., Dolan, C.V., van der Molen, M.W., 2006. Age-related change in executive function: developmental trends and a latent variable analysis. Neuropsychologia 44, 2017e2036. Jacques, S., Zelazo, P.D., Kirkham, N.Z., Semcesen, T.K., 1999. Rule selection versus rule execution in preschoolers: an error-detection approach. Dev. Psychol. 35, 770e780. Jarcho, J.M., Fox, N.A., Pine, D.S., Etkin, A., Leibenluft, E., Shechner, T., Ernst, M., 2013. The neural correlates of emotion-based cognitive control in adults with early childhood behavioral inhibition. Biol. Psychol. 92 (2), 306e314. Jazbec, S., Hardin, M.G., Schroth, E., McClure, E., Pine, D.S., Ernst, M., 2006. Age-related influence of contingencies on a saccade task. Exp. Brain Res. 174, 754e762. Jentzsch, I., Dudschig, C., 2009. Short Article: why do we slow down after an error? Mechanisms underlying the effects of posterror slowing. Q. J. Exp. Psychol. 62 (2), 209e218. https://doi.org/10.1080/17470210802240655. Kagan, J., Reznick, J.S., Clarke, C., Snidman, N., Garcia-Coll, C., 1984. Behavioral inhibition to the unfamiliar. Child Dev. 55 (6), 2212e2225. Kagan, J., Reznick, S., Snidman, N., 1988. Biological bases of childhood shyness. Science 240 (4849), 167. Kerr, A., Zelazo, P.D., 2004. Development of ‘hot’ executive function: the Children’s Gambling Task. Brain Cognit. 55, 148e157. Kochanska, G., Murray, K., Jacques, T.Y., Koenig, A., Vandegeest, K.A., 1996. Inhibitory control in young children and its role in emerging internalization. Child Dev. 67, 490e507. Kohls, G., Petzler, J., Hepertz-Dahlmann, B., Konrad, K., 2009. Differential effects of social and nonsocial reward on response inhibition in children and adolescents. Dev. Sci. 12, 614e625. Ladouceur, C.D., Dahl, R.E., Carter, C.S., 2007. Development of action monitoring through adolescence into adulthood: ERP and source localization. Dev. Sci. 10 (6), 874e891. https://doi.org/10.1111/j.1467-7687.2007.00639.x. Lahat, A., Todd, R.M., Mahy, C.E.V., Lau, K., Zelazo, P.D., 2010. Neurophysiological correlates of executive function: a comparison of EuropeanCanadian and Chinese-Canadian 5-year-old children. Front. Hum. Neurosci. 3, 72. https://doi.org/10.3389/neuro.09.072.2009. Lahat, A., Degnan, K.A., White, L.K., et al., 2012. Temperamental exuberance and executive function predict propensity for risk-taking in childhood. Dev. Psychopathol. 24 (3), 847e856. Lahat, A., Lamm, C., Chronis-Tuscano, A., Pine, D.S., Henderson, H.A., Fox, N.A., 2014. Early behavioral inhibition and increased error monitoring predict later social phobia symptoms in childhood. J. Am. Acad. Child Adolesc. Psychiatry 53 (4), 447e455. https://doi.org/10.1016/ j.jaac.2013.12.019. Lahat, A., Walker, O.L., Lamm, C., Degnan, K.A., Henderson, H.A., Fox, N.A., 2014. Cognitive conflict links behavioral inhibition and social problem solving during social exclusion in childhood. Infant Child Dev. 23 (3), 273e282. https://doi.org/10.1002/icd.1845. Lamm, C., Walker, O.L., Degnan, K.A., Henderson, H.A., Pine, D.S., McDermott, J.M., Fox, N.A., 2014. Cognitive control moderates early childhood temperament in predicting social behavior in seven year old children: an ERP study. Dev. Sci. 17 (5), 667e681. Lamm, C., White, L.K., McDermett, J.M., Fox, N.A., 2012. Neural activation underlying cognitive control in the context of neutral and affectively charged pictures in children. Brain Cognit. 79, 181e187. Lewis, M.D., Lamm, C., Segalowitz, S.J., Steiben, J., Zelazo, P.D., 2006. Neurophysiological correlates of emotion regulation in children and adolescents. J. Cognit. Neurosci. 18, 430e443. Lewis, M.D., Todd, R.M., Honsberger, M.J.M., 2007. Event-related potential measures of emotion regulation in early childhood. NeuroReport 18, 61e65. Luciana, M., Collins, P.F., 2012. Incentive motivation, cognitive control, and the adolescent brain: is it time for a paradigm shift? Child Dev. Perspect. 6 (4), 392e399. https://doi.org/10.1111/j.1750-8606.2012.00252.x. Luna, B., Thulborn, K.R., Munoz, D.P., et al., 2001. Maturation of widely distributed brain function subserves cognitive development. NeuroImage 13, 786e793. Luna, B., Wright, C., 2016. Adolescent Brain Development: Implications for the Juvenile Criminal Justice System. MacDonald, A.W., Cohen, J.D., Stenger, V.A., Carter, C.S., 2000. Dissociating the role of the dorsolateral prefrontal and anterior cingulate cortex in cognitive control. Science 288 (5472), 1835e1838. Mathewson, K.J., Dywan, J., Segalowitz, S.J., 2005. Brain bases of error-related ERPs as influenced by age and task, Bio. Psychol 70 (2), 88e104. https:// doi.org/10.1016/j.biopsycho.2004.12.005. May, J.C., Delgado, M.R., Dahl, R.E., et al., 2004. Event-related functional magnetic resonance imaging of reward-related brain circuitry in children and adolescents. Biol. Psychiatry 55, 359e366. McDermott, J.M., Perez-Edgar, K., Henderson, H.A., Chronis-Tuscano, A., Pine, D.S., Fox, N.A., 2009. A history of childhood behavioral inhibition and enhanced response monitoring in adolescence are linked to clinical anxiety. Biol. Psychiatry 65, 445e448.

The neural correlates of cognitive control and the development of social behavior Chapter | 24

537

Meyer, A., Weinberg, A., Klein, D.N., Hajcak, G., 2012. The development of the error related negativity (ERN) and its relationship with anxiety: evidence from 8 to 13 year-olds. Dev. Cognit. Neurosci. 2 (1), 152e161. https://doi.org/10.1016/j.dcn.2011.09.005. Miller, E.K., Cohen, J.D., 2001. An integrative theory of prefrontal cortex function. Annu. Rev. Neurosci. 24 (1), 167e202. https://doi.org/10.1146/ annurev.neuro.24.1.167. Mischel, W., Ebbesen, E.B., Zeiss, A.R., 1972. Cognitive and attentional mechanisms in delay of gratification. J. Pers. Soc. Psychol. 21, 204e218. Nigg, J.T., 2017. Annual Research Review: On the relations among self-regulation, self-control, executive functioning, effortful control, cognitive control, impulsivity, risk-taking, and inhibition for developmental psychopathology. J Child Psychol. Psychiatry 58 (4), 361e383. Notebaert, W., Houtman, F., Van Opstal, F., Gevers, W., Fias, W., Verguts, T., 2009. Post error slowing: an orienting account. Cognition 111 (2), 275e279. Petersen, S.E., Posner, M.I., 2012. The attention system of the human brain: 20 years after. Annu. Rev. Neurosci. 35, 73e89. Polak-Toste, C.P., Gunnar, M.R., 2006. Temperamental exuberance: correlates and consequences. In: Marshall, P.J., Fox, N.A. (Eds.), TheDevelopment of Social Engagement: Neurobiological Perspectives. Series in Affective Science. Oxford University Press, New York, NY, pp. 19e45. Premack, D., Woodruff, G., 1978. Does the chimpanzee have a theory of mind? Behav. Brain Sci. 1, 516e526. Prencipe, A., Zelazo, P.D., 2005. Development of affective decision-making for self and other: evidence for the integration of first- and third-person perspectives. Psychol. Sci. 16, 501e505. Putnam, S.P., Stifter, C.A., 2005. Behavioral approach-inhibition in toddlers: prediction from infancy, positive and negative affective components, and relations with behavior problems. Child Dev. 76, 212e226. Ridderinkhof, K.R., Ullsperger, M., Crone, E.A., Nieuwenhuis, S., 2004. The role of the medial frontal cortex in cognitive control. Science 306, 443e447. Rothbart, M.K., 1981. Measurement of temperament in infancy. Child Dev. 569e578. Rothbart, M.K., Bates, J.E., 2006. Temperament. In: Damon, W., Eisenberg, N. (Eds.), Handbook of Child Psychology, Social, Emotional, and Personality Development, sixth ed., vol. 3. Wiley, New York, pp. 99e166. Rothbart, M.K., Ellis, L.K., Rueda, M.R., Posner, M.I., 2003. Developing mechanisms of temperamental effortful control. J. Pers. 71, 1113e1143. Rubia, K., Smith, A.B., Woolley, J., et al., 2006. Progressive increase of frontostriatal brain activation from childhood to adulthood during event-related tasks of cognitive control. Hum. Brain Mapp. 27, 973e993. Rubia, K., Smith, A.B., Taylor, E., Brammer, M., 2007. Linear age-correlated functional development of right inferior fronto-striato-cerebellar networks during response inhibition and anterior cingulate during error-related processes. Hum. Brain Mapp. 28, 1163e1177. Rueda, M.R., Fan, J., McCandliss, B.D., et al., 2004. Development of attentional networks in childhood. Neuropsychologia 42, 1029e1040. Sabbagh, M.A., Xu, F., Carlson, S.M., Moses, L.J., Lee, K., 2006. The development of executive functioning and theory of mind: a comparison of Chinese and U.S. preschoolers. Psychol. Sci. 17, 74e81. Santesso, D.L., Segalowitz, S.J., 2008. Developmental differences in error-related ERPs in middle- to late-adolescent males. Dev. Psychol. 44 (1), 205e217. https://doi.org/10.1037/0012-1649.44.1.205. Schröger, E., 1993. Event-related potentials to auditory stimuli following transient shifts of spatial attention in a Go/Nogo task. Biol. Psychol. 36 (3), 183e207. Shenhav, A., Botvinick, M.M., Cohen, J.D., 2013. The expected value of control: an integrative theory of anterior cingulate cortex function. Neuron 79 (2), 217e240. Simonds, J., Kieras, J.E., Rueda, M.R., Rothbart, M.K., 2007. Effortful control, executive attention, and emotional regulation in 7e10-year-old children. Cognit. Dev. 22, 474e488. Somerville, L.H., Casey, B., 2010. Developmental neurobiology of cognitive control and motivational systems. Curr. Opin. Neurobiol. 20 (2), 236e241. Somerville, L.H., Heatherton, T.F., Kelley, W.M., 2006. Anterior cingulate cortex responds differentially to expectancy violation and social rejection. Nat. Neurosci. 9 (8), 1007e1008. https://doi.org/10.1038/nn1728. Sowell, E.R., Thompson, P.M., Holmes, C.J., Jernigan, T.L., Toga, A.W., 1999. In vivo evidence for post-adolescent brain maturation in frontal and striatal regions. Nat. Neurosci. 2 (10), 859e861. https://doi.org/10.1038/13154. Steinberg, L., 2008. A social neuroscience perspective on adolescent risk taking. Dev. Rev. 28, 78e106. Steinberg, L., 2010. A dual systems model of adolescent risk-taking. Dev. Psychobiol. 52 (3), 216e224. Steinberg, L., Graham, S.J., O’Brien, L., Woolard, J., Cauffman, E., Banich, M., 2009. Age differences in future orientation and delay discounting. Child Dev. 80, 28e44. Stieben, J., Lewis, M.D., Granic, I., Zelazo, P.D., Segalowitz, S., Pepler, D., 2007. Neurophysiological mechanisms of emotion regulation for subtypes of externalizing children. Dev. Psychopathol. 19, 455e480. Tamm, L., Menon, V., Reiss, A.L., 2002. Maturation of brain function associated with response inhibition. J. Am. Acad. Child Adolesc. Psychiatry 41, 1231e1238. Tamnes, C.K., Walhovd, K.B., Torstveit, M., Sells, V.T., Fjell, A.M., 2013. Performance monitoring in children and adolescents: a review of developmental changes in the error-related negativity and brain maturation. Dev. Cognit. Neurosci. 6, 1e13. Thorell, L., Bohlin, G., Rydell, A., 2004. Two types of inhibitory control: predictive relations to social functioning. Int. J. Behav. Dev. 28, 193e203. Tobin, J.J., Wu, D.Y.H., Davidson, D.H., 1989. Preschool in Three Cultures: Japan China and the United States. Yale University Press, London. Torpey, D.C., Hajcak, G., Kim, J., Kujawa, A.J., Dyson, M.W., Olino, T.M., Klein, D.N., 2013. Error-related brain activity in young children: associations with parental anxiety and child temperamental negative emotionality. J. Child Psychol. Psychiatry 54 (8), 854e862. https://doi.org/10.1111/ jcpp.12041. Triplett, N., 1898. The dynamogenic factors in pacemaking and competition. Am. J. Psychol. 9 (4), 507e533.

538 PART | II Cognitive Development

Ullsperger, M., Danielmeier, C., Jocham, G., 2014. Neurophysiology of performance monitoring and adaptive behavior. Physiol. Rev. 94 (1), 35e79. Ullsperger, M., Fischer, A.G., Nigbur, R., Endrass, T., 2014. Neural mechanisms and temporal dynamics of performance monitoring. Trends Cognit. Sci. 18 (5), 259e267. Van Der Borght, L., Schevernels, H., Burle, B., Notebaert, W., 2016. Errors disrupt subsequent early attentional processes. PloS One 11 (4), e0151843. Van Veen, V., Carter, C.S., 2002. The timing of action-monitoring processes in the anterior cingulate cortex. J. Cognit. Neurosci. 14, 593e602. White, L.K., McDermott, J.M., Degnan, K.A., Henderson, H.A., Fox, N.A., 2011. Behavioral inhibition and anxiety: the moderating roles of inhibitory control and attention shifting. J. Abnorm. Child Psychol. 39, 735e747. Wu, D.Y.H., 1996. Chinese childhood socialization. In: Bond, M.H. (Ed.), The Handbook of Chinese Psychology. Oxford University Press, New York, pp. 143e151. Zajonc, R.B., Paulus, P.B., 1980. Psychology of Group Influence. Erlbaum, Hillsdale, NJ. Zajonc, R.B., 1965. Social facilitation. Science 149 (3681), 269e274. Zelazo, P.D., 2004. The development of conscious control in childhood. Trends Cognit. Sci. 8, 12e17. Zelazo, P.D., 2006. The dimensional change card sort (DCCS): a method of assessing executive function in children. Nat. Protoc. 1, 297e301. Zelazo, P.D., Müller, U., Frye, D., Marcovitch, S., 2003. The development of executive function in early childhood. Monogr. Soc. Res. Child Dev. 68, viie137. Zelazo, P.D., Carlson, S.M., Kesek, A., 2008. Development of executive function in childhood. In: Nelson, C.A., Luciana, M. (Eds.), Handbook of Developmental Cognitive Neuroscience, second ed. MIT Press, Cambridge, MA, pp. 553e574.

Chapter 25

Executive function: development, individual differences and clinical insights Hughes Claire1, 2 1

Newnham College, Cambridge University, Cambridge, United Kingdom; 2Centre for Family Research, Cambridge University, Cambridge, United

Kingdom

Chapter outline 25.1. Introduction 539 25.1.1. Normative developmental trajectories for executive function from infancy to adolescence 540 25.2. Clinical insights, from infancy to adolescence 543 25.3. From biological to environmental predictors of individual differences in executive function 545

25.3.1. Early executive function predicts academic, sociocognitive and social success at school 548 25.4. Conclusions 551 References 552

25.1 Introduction Executive function (EF) is an umbrella term that encompasses the set of higher-order processes (such as inhibitory control, working memory, attentional flexibility) that govern goal-directed action and adaptive responses to novel, complex, or ambiguous situations (Hughes et al., 2005). Research into the neural substrate for EF (Golden, 1981) has long focused on the prefrontal cortex (PFC), but this traditional view is open to at least two distinct challenges. First, positron emission tomography (PET) studies demonstrate that EF tasks activate parietal areas involved in basic attentional processes more strongly than the PFC (for a review, see Jurado and Rosselli, 2007); similar findings have also been reported in an MRI study of EF in children and adolescents (Tamnes et al., 2010). Second, clinical studies show that early pathology in any brain region leads to executive deficits, such that, for children at least, intact EF depends upon the integrity of the entire brain, not just frontal regions (Anderson and Catroppa, 2005), with evidence from more recent studies suggesting that timing (rather than location) of traumatic brain injury predicts the severity of EF impairment (Resch et al., 2019). Research into EF in children has grown exponentially over the past few decades: A recent Scopus search using the terms “executive functions” and “children” showed just 5 studies prior to 1980; 26 studies between 1980 and 1990; 216 studies between 1990 and 2000; 1092 studies between 2000 and 2010; and 7807 studies between 2010 and 2020. This massive expansion of research reflects the growth of interest in childhood clinical groups, as EF deficits are apparent in a number of different developmental disordersdespecially children with attention-deficit hyperactivity disorder (ADHD) and children with autism spectrum disorder (ASD) (Craig et al., 2016; Friedman and Sterling, 2019; Pineda-Alhucema et al., 2018). However, the rapid growth of neuroimaging studies of normative age-related changes in EF has led to a valuable cross-over between research on typical and atypical development (Fiske and Holmboe, 2019). In particular, in comparison with traditional imaging methods such as EEG and fMRI, the advent of fNIRS has greatly facilitated the investigation of neural substrates of EF in infants and young children. As Tau and Peterson (2010) put it, documenting normative pathways of brain development “provides the Archimedean point from which to interpret and understand the aberrant pathways of brain development that produce disease,” while studying atypical groups “informs our knowledge of normal brain development by throwing into relief the developmental pathways that are most sensitive to perturbation.”

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00025-1 Copyright © 2020 Elsevier Inc. All rights reserved.

539

540 PART | II Cognitive development

A key insight to emerge within the past decade of research on EF is that the neural systems supporting EF differ by affective context, such that motivationally salient situations activate a “hot” EF system that has its neural base in the orbitofrontal cortex, whereas neutral situations evoke “cool” EF that depends primarily on the dorsolateral prefrontal cortex. In a review of this field, Zelazo and Carlson (2012) showed that both hot and cool EF appear surprisingly malleable, such that there are promising possibilities for both intervention and prevention. From a developmental perspective, it is worth noting that a recent analysis of pooled data from four separate studies in Southern USA, involving a sample of 1900 children, supported a two-factor structure (i.e., hot vs. cool EF) but revealed only partial measurement invariance across age groups, indicating that how certain tests represent EF changes with age (Montroy et al., 2019). Note also that the protracted nature of EF development helps explain both the pervasiveness of EF deficits in childhood disorders and the salience of EF for studies of normal brain development. Thus, EF skills begin to emerge in infancy (Devine et al., 2019; Diamond, 1988) and show marked improvements across toddlerhood and the preschool period (Carlson et al., 2004; Hughes and Ensor, 2007; Hughes et al., 2010; Moriguchi, 2014; Pauen and Bechtel-Kuehne, 2016) and continue to improve across the school years (e.g., Huizinga et al., 2006), with some aspects of EF continuing to develop throughout adolescence (Luciana et al., 2005; Luna et al., 2004). Interestingly, although several reviews of EF development are available (e.g., Anderson, 2002; Best et al., 2009; Blair et al., 2005; Blakemore and Choudhury, 2006; Garon et al., 2008; Hughes and Graham, 2002; Moriguchi, 2014; Moriguchi and Hiraki, 2013), very few span the full period of development (for an exception, see Diamond, 2002). To address this gap, the first section of this chapter provides a synopsis of research findings on developmental trajectories from infancy to adolescence, in both typical and atypical populations. Having a protracted developmental course also makes EF a focus of interest for researchers interested in exploring environmental influences. As noted elsewhere (Hackman et al., 2015), this topic of environmental influence on EF is a promising new direction for research, because the slow maturation of the frontal cortex and its networks make it heavily dependent on the environment (e.g., Noble et al., 2005). In the past decade, there has been substantial progress in our understanding of the processes underpinning environmental influences on EF. To illustrate this work, the second section of this chapter provides an overview of three strands of research that investigate (1) effects of training or intervention programs on children’s emerging EF skills; (2) influences of parentechild interactions on EF; and (3) studies of environmental influence on EF in clinical groups. Interwoven with perspectives on EF that emphasize typical or atypical development, the expansion of research in this field also mirrors the intensity of research into children’s understanding of mind, as numerous studies have reported a robust association between EF and children’s “theory of mind” skills that is evident across a broad age range, independent of covarying effects of language ability and IQ, and applies to both typical and atypical populations (for a metaanalytic review, see Devine and Hughes, 2014). In addition, recent research has demonstrated that early emerging skills in both EF and theory of mind are strong predictors of children’s school readiness and their performance in academic subjects. The third section of this chapter aims to bring together the findings from research tracing links between EF and children’s academic, sociocognitive, and social success at school.

25.1.1 Normative developmental trajectories for executive function from infancy to adolescence The past two decades have seen a massive increase in the availability of child-friendly EF tasks (for a useful summary, see Blair, 2016), leading to dramatic improvements in our understanding of the development of EF. For instance, it is now known that EF is a unitary construct with partially dissociable components (Friedman and Miyake, 2017) that begins to emerge in the first few years of life (e.g., Devine et al., 2019; Diamond, 1991) and continues to develop through to adulthood (Huizinga et al., 2006). However, most studies in this field have been cross-sectional in design and have not controlled for peripheral task demands. As a result, both cohort effects and developmental changes in how children cope with peripheral task demands may contribute to age-related contrasts in EF performance. In addition, simplifying adult tasks to make them age-appropriate for young children carries the danger of losing the critical EF component (Garon et al., 2008). These notes of caution need to be borne in mind when considering the summary review of findings from infants to adolescents given in the following. Infancy. The first evidence that EF emerges in the first year of life (much earlier than previously believed) came from studies using Piaget’s object permanence task, in which babies are repeatedly allowed to retrieve an attractive object from one location (A) before seeing it hidden at a new location (B). Early studies indicated that while eight-month-olds and older babies typically search correctly at location B, 5-month olds persist in searching for the object at location A (e.g., Harris, 1975). Although Piaget interpreted this “A-not B” error as a failure to recognize that objects have an independent existence in the world (i.e., a lack of “object permanence”), later studies showed that when looking times rather than physical reaches

Executive function: development, individual differences and clinical insights Chapter | 25

541

are used to assess perceptual understanding, even 4-months-olds do well (e.g., Baillargeon et al., 1985). Thus, young babies who make the A-not B error in their reaching responses may know that the object has been moved but fail to inhibit their previously successful (and therefore prepotent) reach to A. Thus, success on this task can be seen as reflecting infants’ growing cognitive flexibility and volitional control (Diamond and Goldman-Rakic, 1989). Other tasks that also demonstrate early executive skills in infancy include a detour-reaching task, in which infants are invited to retrieve an object that is visible behind a Perspex screen; success on this task depends on making a “detour reach” around the side of the screen (Diamond et al., 1989). In short, a rudimentary ability to inhibit prepotent responses is clearly evident by 7e12 months (Diamond, 2002). More recent studies of infants also indicate an early emergence of other aspects of EF. In particular, building on Rothbart et al., (2003) finding that 2-year-olds’ anticipatory looking (i.e., looking to the location of a target prior to its appearance) is associated with parental ratings of self-regulation, Sheese et al. (2008) have reported that 6- and 7-montholds who display high levels of anticipatory looking also show more signs of self-regulation in their approach toward novel toys. Together, these two studies indicate that executive attention (indexed by anticipatory looking) is also evident very early in life. Interestingly, Sheese et al. (2008) found that anticipatory looking was also associated with looking away from disturbing stimuli (face masks), supporting proposals (e.g., Aksan and Kochanska, 2004; Rothbart et al., 2000) for a link between early systems of emotional control (e.g., fear and caution) and later systems of cognitive control. Most recently, Devine et al. (2019) have extended the developmental scope of this research still further in a study involving more than 400 infants, by demonstrating that attention at 4 months predicts working memory performance 10 months later. Preschoolers. Research into the preschool years accounts for the lion’s share of studies of EF in childhood. As will be discussed in the third section of this chapter, this focus on preschoolers partly reflects the intensity of research into children’s understanding of mind, which shows dramatic improvements in the preschool years (Wellman, 2018). Another reason for a research focus on preschoolers is that a wide variety of age-appropriate tasks have been developed in response to the previous dearth of tasks suitable for young children (e.g., Devine et al., 2019; Mulder et al., 2017; Nieto et al., 2016); however, it is worth noting that different methods of creating composite scores yield contrasting testeretest reliabilities (Allan et al., 2015; Morra et al., 2018; Mulder et al., 2017; Willoughby et al., 2017). However, even when these conceptual and methodological factors are taken into account, the growth of research on preschool EF remains remarkable. Although it is difficult to summarize briefly, three points deserve particular mention. First, while key components of EF (working memory, response inhibition, and set shifting) in adults are all evident before the age of 3 years (e.g., Garon et al. (2008), studies provide growing evidence for an age-related shift in the nature of neural substrate that can be characterized by increased focalization, such that diffuse neural structures underpin EF in young children, but more localized structures provide the substrate for EF in older children (for a recent fNIRS study, see Chevalier et al., 2019). Second, predictions from theoretical work highlighting the overlap between attention and EF (e.g., Rothbart and Posner, 2005) or the pivotal role of attention as a driver of developmental improvements in children’s ability to overcome prepotent thoughts/acts (Diamond, 2002) or to integrate conflicting rules (Zelazo et al., 2003) are supported by both longitudinal findings and EEG findings. Specifically, investigations that straddle early infancy to toddlerhood (Cuevas and Bell, 2014; Devine et al., 2019) and toddlerhood to early preschool (Veer et al., 2017) all highlight developmental associations between early attention and later EF. Likewise, a recent EEG study of young school-aged children provides further empirical evidence for associations between attention and EF (Weiss et al., 2018). Third, the increased accessibility of latent variable analysis has greatly enhanced the methodological rigor of EF research. For example, by minimizing effects of measurement error, latent variable analyses enable more effective testing of different models of the structure of EF. As demonstrated in a metaanalysis involving data from 46 separate studies (N ¼ 9756) (Karr et al., 2018), the structure of EF shows significant age-related contrasts in the best-fitting model (i.e., preschool ¼ one/two-factor; school-age ¼ three-factor; adolescent/adult ¼ three/nested-factor; older adult ¼ two/threefactor). Thus, despite the rising availability of child-friendly computerized tasks that enable EF assessments to include both reaction times and accuracy scores (Willoughby et al., 2018), the comparability of EF performance across different age groups remains open to question. From this perspective, it is therefore reassuring that measurement invariance has been demonstrated for latent factors for EF in preschool and early school age (Hughes et al., 2010). Latent variable analyses have also been used to conduct person-centered analyses (e.g., latent growth analysis, latent class analysis) to examine the origins and consequences of variability in two aspects of EF: stable individual differences and contrasts in age-related gains in EF. Importantly, these two approaches help explain the contrasting conclusions regarding the relative salience of genetic and environmental influences on EF. Specifically, as noted by Asbury and Plomin (2013), genetic influences are typically diffuse rather than specific and underpin stability rather than change over time. For example, Friedman et al., (2011) assessed EF in 950 twins at 14, 20, 24, and 36 months and then again at age 17 years,

542 PART | II Cognitive development

using a latent variable approach to distinguish a “common EF” latent factor from two more specific “updating” and “shifting” factors. Latent class analysis showed that persistent problems of self-regulation in toddlerhood predicted poor EF at age 17 years, with genetic factors accounting for the association with common EF. Conversely, latent variable analyses of EF trajectories typically highlight the importance of environmental factors (e.g., Friedman et al., 2016; Hughes and Devine, 2019; Hughes et al., 2010, 2013; Narad et al., 2017). Consistent with this emphasis on environmental influences on early EF development, metaanalytic findings demonstrate significant associations between EF and measures of socioeconomic status (SES) that vary from small to medium effect size, depending on the sample variation in SES and the number of EF tasks included (Lawson et al., 2018). School age. As may be inferred from the age-related shifts in the structure of EF noted earlier, different aspects of EF improve rapidly at different points in development. Specifically, while (as noted earlier) the preschool years are characterized by dramatic improvements in inhibitory control, evidence from young school-aged children and preadolescents highlight, respectively, improvements in cognitive flexibility and in working memory/planning ability. For example, using the intradimensional/extradimensional (IDED) shift task from the CANTAB battery, Luciana (2003) reported findings that showed a marked improvement in children’s ability to shift mental set around the age of 6 or 7 years. The IDED task is a multistage task that is based on the Wisconsin Card Sorting Task, which requires participants to work out a rule for sorting cards (e.g., sort by color, or number or shape of stimuli) and then, when they receive negative feedback indicating that the rule has changed, to shift strategy to sort by a new rule. Luciana (2003) also showed that clear improvements on the Tower of London planning task or on a self-ordered search test of working memory were often not evident before the age of 11 or 12 years. Other studies using different tasks have reported a similar contrast in the developmental trajectories of different aspects of EF. For example, several studies report improvements in mental flexibility around age 8 years (e.g., Anderson, 2002; Anderson et al., 2001b), whereas planning, organizing, and strategic thinking are typically reported to emerge later and to show age-related improvements throughout adolescence (Anderson et al., 1996, 2001a; De Luca et al., 2003; Krikorian et al., 1994; Welsh et al., 1991). At odds with this general pattern, however, are findings from another study that suggest a long developmental progression for cognitive flexibility, with 13-year-olds still not at adult levels (Davidson et al., 2006). A closer look at the specific tasks suggests that the contrast in these findings may be explained by Diamond’s (2009) “all or none” theory. According to this theory, the brain and mind work effortlessly (or under difficult conditions) at a gross level, but require effort (or more optimal conditions), to work in a more selective manner. Thus, it is easier to inhibit a dominant response all the time than only some of the time. As a result, even older children are likely to show frequent errors on task-switching paradigms, such as that used in Davidson et al.’s (2006) study. This study also revealed several other interesting developmental contrasts. For example, adults slowed down on difficult trials to preserve accuracy, but children (and especially young children) were impulsive and so made errors on difficult trials. These contrasting speed accuracy trade-offs highlight the value of using computerized tasks to assess EF. At the same time, these advantages may be offset by a reduction in sensitivity. For example, children with ASD have been found to perform significantly better on a computerized set-shifting task than on a manual version (Ozonoff, 1995). In addition, computerized tasks are likely to have lower ecological validity, such that there is still a need for manual tasks that mimic everyday EF demands. Here, one task battery that deserves mention is the behavioral assessment of dysexecutive syndrome (Roy et al., 2015) that includes tasks that tap into abilities for multitasking, problem-solving, and strategic thinking. At this point, it is worth noting that the widespread use of computerized tasks within studies of EF in school-aged children enables researchers not only to standardize administration and include large numbers of trials but also to collect information on reaction times as well as accuracy. Importantly, these two indicators can then be combined to examine age-related improvements in the efficiency of EF performance. In a cross-cultural study to adopt this approach with a sample of 1427 participants (children aged 9e16 and their parents), Ellefson et al. (2017) demonstrated that children from Hong Kong were not only more accurate but also substantially quicker at completing EF tasks than their British counterparts, with a between-country contrast that was equivalent to a lag of 2 years for children living in the United Kingdom. While mirroring results from other EF cross-cultural studies involve either school-aged children (Wang et al., 2016) or preschoolers (Sabbagh et al., 2006), this contrast provides an interesting counterpoint to results from cross-cultural studies of theory of mind (Hughes et al., 2017; Nawaz et al., 2015; Wang et al., 2016). Together, these divergent findings highlight the need for more nuanced studies of family and cultural influences on children’s EF development. Adolescence: Consistent with the findings from younger samples described earlier, different aspects of EF have been reported to show distinct trajectories across the adolescent years. In particular, while age-related improvements in “cool” EF are typically linear, improvements in “hot”’ EF are nonlinear, which may help explain why risk-taking problems peak in middle adolescence (Poon, 2018). Thus an emerging theme from the adolescent literature concerns the need to consider

Executive function: development, individual differences and clinical insights Chapter | 25

543

EF alongside other key brain systems. For example, Tau and Peterson (2010) note that adolescents and adults differ not just in the maturity of their EF functions but also in the extent to which they avoid risk and respond to reward/peer influence (adolescents are less risk averse, more driven by reward, and more easily influenced by peers). As a result, accounts of developmental change in everyday behavior should consider not only top-down EF systems but also bottom-up motivational and emotional responses to situations of risk and reward. Similar conclusions emerge from a community-based study of 10- to 12-year-olds in which self-reported impulsivity was inversely associated with both working memory and reversal learning (Romer et al., 2009). Noting that interventions to improve children’s working memory have led to reductions in impulsive behaviors (e.g., Klingberg et al., 2005), the authors of this study concluded that young people who have difficulties in considering multiple (and potentially conflicting) goals will be less likely either to “look before they leap” or to temper their interest in novel or exciting experiences. More recently, however, the same research group (Romer et al., 2017) have challenged this simple “dual systems” account, noting that while some forms of risk taking (e.g., sensation seeking) peak in adolescence, other forms (e.g., impulsive choices) show a linear decline. Arguing that sensation seeking is motivated by a desire to explore the environment in ambiguous risk contexts, these authors call for theoretical models that recognize adaptive roles of both cognition and experience during adolescence.

25.2 Clinical insights, from infancy to adolescence In parallel with, and perhaps fueling the growth of research on normative EF development, studies of EF in atypical child populations have grown dramatically and yielded several interesting findings that together emphasize the extent to which EF can be impaired by brain abnormalities or insults, especially if these occur early in life. Infancy. The first clinical study of EF in infancy involved children treated early and continuously for phenylketonuria (PKU), a metabolic disorder characterized by a failure to convert phenylalanine to tyrosine (the precursor of dopamine) that, if untreated, is the most common biochemical cause of intellectual disability. Highlighting the importance of dopamine for EF, this 4-year longitudinal study showed that children with PKU with high plasma levels of phenylalanine performed more poorly than all other control groups (including siblings and PKU children with lower levels of phenylalanine) on tests of working memory and inhibitory control (Diamond et al., 1997). Moreover, this impairment was specific to performance on EF tasks, directly related to levels of phenylalanine and evident across the age groups, from the youngest (6e12 months) to the oldest (3½e7 years). Long-term deficits in EF are also found in infants born prematurely and infants exposed prenatally to high levels of alcohol. In a metaanalysis of longitudinal studies involving children born very prematurely (i.e., before 32 weeks) (Brydges et al., 2018), impairments in EF were reported to be medium to large in effect size, regardless of age of assessment and, for younger children only, varied in magnitude with gestational age (i.e., extremely premature infants show greater deficits). A corresponding metaanalysis of EF in children with prenatal alcohol exposure has demonstrated an association with global EF impairments but most consistent deficits for higher-order aspects of EF (i.e., planning, fluency, and set shifting) (Kingdon et al., 2016). Moreover, a friendship training study involving primary school-aged children with prenatal alcohol exposure showed that parental ratings of EF predicted gains in social skills, even when effects of IQ were taken into account (Schonfeld et al., 2009). In short, prenatal alcohol exposure has long-term adverse effects of children’s emerging EF skills, which in turn not only underpin academic achievement (see later in this chapter) but also play a key role in these children’s social competencies. Preschoolers. From a clinical perspective, preschool milestones in EF have attracted considerable attention from researchers studying the cognitive profiles associated with ASD. For example, Hughes and Russell (1993) demonstrated that children with ASD (with a verbal mental age of at least 4 years) displayed significant difficulties on two tasks that most 4-year-olds passed with ease. The first of these was the “Windows” task in which, for each of 20 trials, children could win a treat (visible through a window in a box) by choosing a visibly empty box. Children with ASD (and many 3-year-olds) chose the baited box and often persisted in this incorrect response across all 20 trials. Moreover, administering the task in four different conditions (verbal/nonverbal; competitor present/absent) led to no change in results. In the second task, to retrieve a large and attractive marble from inside a metal box, children needed to perform a simple but arbitrary meansdend action (flicking a switch at the side of the box that turned off a simple circuit with an infrared beam that, if tripped, activated a small trapdoor on which the marble was resting). Once again, older children with ASD and typically developing 3-year-olds (but not 4-year-olds) persisted in making a direct and unsuccessful reach into the box. These findings added weight to reports of impaired performance by children with ASD on more traditional (and complex) EF tasks (e.g., Hughes et al., 1994; Ozonoff et al., 1991; Prior, 1979) and sparked a sustained program of research on EF deficits in ASD (for reviews, see Hill, 2004; Pennington and Ozonoff, 1996). The breadth and

544 PART | II Cognitive development

sophistication of research in this field are nicely illustrated by (1) evidence for associations between impaired inhibitory control and high-level repetitive behaviors (e.g., compulsions, preoccupations) in children with ASD spectrum disorders (Schmitt et al., 2018); (2) imaging results that indicate reduced functional connectivity and network integration between frontal, parietal, and occipital regions among individuals with ASD completing EF tasks (Solomon et al., 2009), with subsequent research demonstrating that contrasts in the coherence of the frontoparietal network help explain the greater magnitude of EF deficits in low-functioning as compared with high-functioning children with autism (Han and Chan, 2017); (3) longitudinal evidence for the importance of early EF in shaping the developmental trajectory of theory-of-mind skills in children with ASD (Pellicano, 2010), with more recent work confirming this view for both “hot” and “cool” measures of EF (Kouklari et al., 2019); and (4) evidence for age-related improvements in executive function from childhood to adolescence in ASD, indicating the presence of plasticity and suggesting a prolonged window for effective treatment (for metaanalyses, see Demetriou et al., 2018; Habib et al., 2019). School age: Earlier, in the section on EF in typically developing school-aged children, concern was raised about the ecological validity of the computerized tasks that are often used with this age group. This issue of ecological validity is also salient from a clinical perspective, as EF deficits in children with ADHD (who, by definition, show marked problems of impulsivity/inattention/disorganization in their everyday lives) are often less pervasive and severe than the EF deficits shown by children with ASD (e.g., Geurts et al., 2004; Goldberg et al., 2005; but see also Happé et al. (2006)), such that causal accounts of ADHD also include additional deficits in the signaling of delayed rewards (for a recent metaanalysis, see Doidge et al., 2018; e.g., Sonuga-Barke, 2005). In his review, Sonuga-Barke (2005) also argued for the need to consider EF deficits alongside environment influences, a view that is supported by the peak in diagnosis for ADHD at age 12 years (Mandell et al., 2005); that is, just following the transition to secondary (or “middle”) school, which brings a significant increase in demands for self-controlled and planful behavior. Indeed, a longitudinal study that modeled trajectories for parent-rated ADHD symptoms confirms that the age-related decline in symptoms is at least transiently reversed following the transition to secondary school (Langberg et al., 2008). Likewise, in a study that involved carefully matched samples, Happé et al. (2006) found that children with ASD, but not children with ADHD, showed age-related improvements in EF performance. Together, these findings highlight the importance of adopting a developmental perspective when examining EF deficits in atypical groups. Another area of recent interest within EF research stems from the growth in transdiagnostic models of psychopathology (e.g., Goschke, 2014). That is, while the lack of specificity in the effects of EF impairment was viewed as problematic within traditional diagnostic systems that adopted a categorical approach, the fifth edition of the Diagnostic and Statistical Manual to mental disorders, published by the American Psychiatric Association (2013) heralded dimensional accounts that in turn led to an interest in mechanisms (such as EF impairment) likely to contribute to a range of different disorders. That said, various researchers have presented an array of different accounts of how EF may have transdiagnostic significance. For example, findings from one study (Huang-Pollock et al., 2017) indicate that working memory deficits represent both a common cognitive liability for mental health disorders and a specific liability for externalizing disorders; another view, however, is that different aspects of EF contribute to different types of symptoms (e.g., poor inhibitory control underpins externalizing problems, while poor working memory contributes to internalizing symptoms) (Quistberg and Mueller, 2019). In addition, common processes may be disturbed in different ways within distinct disorders. For example, EF clearly underpins decision-making, which can be impaired in distinct ways within different childhood disorders (e.g., inefficient and inconsistent in ADHD, reckless in conduct disorder [CD] disengaged and pessimistic in depression and risk aversive/self-deprecating in anxiety) (Sonuga-Barke et al., 2016). Future research may illuminate how these alternative accounts have contrasting salience for different age groups or in the context of different risk factors. Adolescence. Much of the clinical work on EF in adolescence has concerned risk-taking and substance abuse. In particular, several researchers have reported a robust predictive association between poor EF and impulsivity in preadolescence and high levels of drug use in late adolescence (e.g., Aytaclar et al., 1999), with more recent findings from a study of 850 twins followed from ages 17 to23 years (Gustavson et al., 2017), demonstrating that poor EF helps explain genetic influences upon polysubstance use in late adolescence, but non-EF factors play a key role in the progression to substance abuse. Similarly, a study of 300 young heavy drinkers showed that impaired self-control plays a mediating role in the pathway from response impulsivity (difficulties inhibiting thoughts and behaviors, especially in the context of reinforcement) to alcohol problems but did not mediate the link between reflection impulsivity (i.e., the tendency to make decisions without sufficiently gathering or evaluating relevant information) and alcohol problems) to alcohol problems (Wardell et al., 2016). Note that this model adopts Tau and Peterson’s (2010) recommended dual focus on top-down and bottom-up processes. In another dual-focus study, adolescent boys with early- versus late-onset CD (or controls) were compared on tests of EF (Wisconsin Card Sort Test) and decision-making (Risky Choice task) under conditions of varying motivation and stress

Executive function: development, individual differences and clinical insights Chapter | 25

545

(Fairchild et al., 2009). With effects of IQ controlled, group differences in EF were nonsignificant, but both CD groups made more risky choices than controls, and adolescent boys with early-onset CD made risky choices even when the gains were relatively small. These findings suggest a shift in the balance between sensitivity to reward and punishment among boys with CD (particularly the early-onset form) that is similar to Wiers et al.’s (2007) proposed imbalance in top-down and bottom-up processes in addictive adolescents. Perhaps unsurprisingly then, CD is, as Fairchild et al. (2009) note, associated with a significantly increased susceptibility to substance use disorders. Likewise, a study by Perry et al. (2018) highlights the value of behavioral and emotional control in differentiating between children with declining trajectories for externalizing problems and those who show persistently high levels of conduct problems through adolescence.

25.3 From biological to environmental predictors of individual differences in executive function Biological studies have been hugely influential in research on EF in childhood. For example, the impact of Diamond and colleagues’ EF account of age-related improvements in infants’ performance on the A-not B task was greatly increased by the convergent findings from parallel studies of (1) rhesus monkeys (Diamond and Goldman-Rakic, 1989), which highlighted the dorsolateral PFC as pivotal to success on this task, and (2) children with PKU, which demonstrated the importance of dopamine for success on a wide battery of simple EF tasks (Diamond et al., 1997). More recently, in their review of studies using functional magnetic resonance imaging (fMRI), Tau and Peterson (2010) concluded that agerelated improvements in EF in childhood are associated with increased activation of (dopamine-rich) frontal and striatal circuits. Other seminal findings include Golden’ (1981) demonstration that myelination of the prefrontal cortex is associated with age-related improvements in EF in children. Technological advances have led to significant progress in documenting parallels between milestones in EF development and changes in brain myelination (Kharitonova et al., 2013). Gains in white matter have clear functional consequences, which include faster and more efficient sharing of information within the frontostriatal circuits and smoother communication between the frontal cortex and other brain regions (Paus, 2010). Peak periods of reduction in gray matter occur just after puberty and at the transition from adolescence to adulthood; although typically attributed to synaptic pruning, this “loss” of gray matter may simply reflect gains in white matter (Paus, 2010). As noted in a highly cited review (Blakemore and Choudhury, 2006), structural changes in the adolescent brain are particularly evident in the frontal cortex and are linked to age-related improvements in inhibitory control (Constantinidis and Luna, 2019; Hwang et al., 2016), working memory (Breukelaar et al., 2017), and decision-making (Churchwell and Yurgelun-Todd, 2013). The studies of EF in atypical groups summarized in the previous section provide a third strand of research with a clear biological perspective. In particular, impairments in EF are most pronounced among children with ADHD or ASD (Pennington and Ozonoff, 1996), two disorders that show substantial genetic influence (Azeredo et al., 2018; Richards et al., 2015). Indeed, EF has been implicated in the genetic basis for ADHD (e.g., Bidwell et al., 2007; Gau and Shang, 2010). More direct evidence for strong genetic influences on early EF comes from studies that include genotyping (e.g., Fossella et al., 2002; Rueda et al., 2005), which demonstrate that children with the homozygous long allele for the DAT1 gene (associated with high effortful control and low extroversion) outperform those with the heterozygous (long/ short) allele on EF tests of conflict resolution (for a recent review of molecular genetic studies of EF in children, see Brocki et al., 2009). However, none of the above findings precludes environmental factors also contributing to either developmental change or individual differences in EF. Indeed, genetic factors often show substantial interactions with environmental influences, such that genetic vulnerability is only expressed among individuals exposed to environmental stressors, such as harsh parenting or family chaos (Asbury et al., 2003, 2005); conversely, recent evidence indicates a buffering effect of positive parenting on children at cognitive risk of developing ADHD or CD (Frick, 2019). Moreover, as noted earlier, the development of EF has a very protracted course, which makes it particularly sensitive to environmental influence (Farah et al., 2006; Noble et al., 2005). In addition, cognitive models of EF (e.g., Duncan, 2001) highlight the fluidity of relations between the prefrontal cortex and EF performance. Specifically, neurons within the prefrontal cortex show rapid adaptation to changing task demands (Freedman et al., 2001), making it difficult to map between behavioral and neuronal functions. In comparison with the well-established research on biological influences, studies of environmental influences on EF are much more recent, although there is growing recognition that early experiences contribute to children’s neurocognitive development. Specifically, unfavorable early environmental experiences adversely affect both brain structure and function (Belsky and De Haan, 2011); conversely, positive experiences (especially with caregivers) appear to have a positive impact

546 PART | II Cognitive development

on brain development (Silva et al., 2018). As Bierman et al. (2008), put it: “EF development depends, in part, upon sensitive responsive caregiving and opportunities for guided exploration of the social and physical environment, fostering sustained joint attention, emotional understanding, planning, and problem-solving.” Direct evidence for environmental effects on EF comes from intervention studies, which can be considered in two sets. The first set of studies involves direct training on task analogs. In one of two seminal studies, Kloo and Perner (2003) gave a simplified version of the Wisconsin Card Sort task to typically developing preschoolers and found that both card-sorting training and false-belief training (each delivered in two sessions on separate days) led to positive effects on EF at posttest. In the second early training study, Rueda et al. (2005) gave 5 days of attention training to groups of 4- and 6-year-olds and reported that children who performed poorly at pretest showed the greatest gains from this training program, with overall improvements in executive attention and IQ performance being equivalent to 1 year’s difference in age. Reviewing this field, Klingberg (2010) reported that, across a wide variety of age groups, training leads to significant improvements in working memory. However, not all aspects of EF appear so malleable to training. For example, a 5-week preschool training program produced significant improvements in working memory but not in inhibitory control (Thorell et al., 2009), raising interesting questions about whether the distinct processes that underpin different aspects of EF lead to contrasts in the extent to which performance can be improved by training. In another training study, Karbach and Kray (2009) administered task-switching training to school-aged children (aged 8e10 year) and both young adults (aged 18e26 year) and older adults (aged 62e76 year). All three groups showed positive effects that transferred to other EF tasks and to tests of fluid intelligence (e.g., tests of abstract thinking and reasoning); however, when the training tasks were variable, improvements were reduced for children but increased for adults. This interaction effect highlights the importance of adopting a developmentally sensitive approach to the development of interventions. The second set of intervention studies adopts a broader and more naturalistic approach and is thus perhaps of greater relevance for theories of how everyday social environments might impinge on children’s EF development. At least three such interventions have been assessed using randomized controlled trial (RCT) designs. The first of these, Head Start REDI (Research-based, Developmentally Informed), is integrated into the Head Start prekindergarten program for disadvantaged children and involves brief lessons, “hands-on” extension activities, and specific teaching strategies linked empirically with the promotion of socialeemotional competencies, language development, and emergent literacy skills. In their RCT, Bierman et al. (2008) reported that the REDI intervention led to significant improvements in children’s abilities to stay on-task, coupled with marginally significant gains in set-shifting performance. Another wellrecognized intervention is the Vygotskian “Tools in the Mind” preschool curriculum (Bodrova and Leong, 1996; Diamond et al., 2007), which includes a variety of specially designed activities (e.g., sociodramatic play, shared reading) that enable children to progress from external- to shared- to self-regulation; teachers are also trained to foster early skills in literacy and mathematics by encouraging reflective thinking and metacognition. Interestingly, although language plays a pivotal role in Vygotskian accounts of cognitive development, the Tools curriculum appears to have positive effects on EF (as indexed by low problem behavior scores), but no significant impact on language development (Barnett et al., 2008). The third RCT focused on an 8-week school-based intervention for older children (7- to 9-year-olds), which aims to promote mindful awareness practices (MAPs) through twice-weekly half-hour sessions. Comparing teachers’ and parents’ pre- and postprogram ratings of EF skills, Flook et al. (2010) reported that less well-regulated children showed particularly clear treatment-related gains (evident both at school and at home); note that this finding echoes earlier results from preschoolers (e.g., Rueda et al., 2005). Together, the findings from these three RCTs support two of the three dimensions of adultechild interactions that Carlson (2003) has proposed as likely to favor child EF: scaffolding (which provides children with successful experiences of problem-based learning) and mind-mindedness (which provides children with verbal tools for progressing from external to internal forms of self-regulation). Positive effects on EF of the third dimension of sensitivity (which provides infants with successful experiences of impacting on the environment) have been reported by Bernier et al. (2010), in their longitudinal study of 80 infants, in which maternal sensitivity, mind-mindedness, and scaffolding (or “autonomy support”) were evaluated at 12e15 months of age, and EF was assessed at 18 and 26 months. All three measures of parenting predicted child EF, but once effects of maternal education and general cognitive ability were taken into account, scaffolding was the strongest predictor of EF at each age. Building on this work, Hughes and Devine (2019) applied a latent variable approach to examine the independence of four distinct measures of parenting (e.g., scaffolding, negative control, mean length of utterance (MLU), quality of home learning environment) as predictors of gains in EF across the preschool years. Importantly, this study also included verbal ability as an outcome measure, providing a rare opportunity to also assess the specificity of parental influences. This study

Executive function: development, individual differences and clinical insights Chapter | 25

547

showed that while all four parenting measures displayed independent associations with gains in EF, only scaffolding and negative parenting showed specific associations (in the expected directions) with EF gains; in contrast, home learning environment and MLU were also associated with improvements in children’s verbal skills. In short, there is now good evidence that parents’ deliberate efforts to scaffold children’s goal-directed activities do indeed foster the development of early EF skills. But, as highlighted by research on children’s developing social understanding, family influences are often incidental rather than deliberate. For example, young children are extremely acute observers of family life, paying particularly close attention to injustices such as parents’ differential treatment of siblings (Dunn, 1993). This point can be applied to young children’s observational skills more generally; most parents will be able to recall episodes in which their actions have been mimicked with uncanny accuracy. This attention to detail favors children’s rapid mastery of complex action plans: While simple mimicry of adults’ planful behavior does not constitute executive control, acquiring a repertoire of “goal-directed” acts is likely to promote EF skills. Support for this view comes from a longitudinal observational study by Hughes and Ensor (2009) involving a socially diverse sample of 125 children. In this study, both maternal scaffolding (in structured play with jigsaws) and opportunities for observational learning (indexed by maternal strategic behavior in a multitasking paradigm and in a shared tidy-up task) predicted improvements between the ages of 2 and 4 years in children’s EF scores, even when effects of verbal ability were controlled. Note that including EF assessments at both time points enabled the temporal stability of individual differences in EF to be taken into account and so minimize the confounding effects of genetic factors (Kovas et al., 2007). Reinforcing the importance of incidental effects, it is worth recalling that Hughes and Devine (2019) showed an independent and inverse association between negative parental control and EF gains from ages 4 to 5 years, echoing the negative correlation between EF development from age 2 to age 4 years and parental ratings of disorganized and unpredictable family life reported by Hughes and Ensor (2009). The take-home message here is that families can hinder as well as help young children’s emerging EF skills. Using an expanded sample from the same longitudinal study (i.e., the target children, plus friends recruited at age 4 years; N ¼ 191), Hughes and colleagues (Roman et al., 2016) also applied latent-variable analyses to demonstrate that individual differences in EF mediate the adverse effect of exposure to maternal depression on children’s behavioral adjustment. These findings are strengthened by converging results from a metaanalysis of associations between parenting and early EF (Valcan et al., 2018). In addition, these findings extend the results from studies assessing the impact on EF of exposure to more severe environmental risk factors, such as those experienced by children from homeless families (Monn et al., 2017). Expanding on the aforementioned literature on environmental influences on children’s EF development, findings from a large-scale longitudinal study of 1292 children from low-income families participating in the Family Life Project have shown that, above quality of the postnatal environment, a composite of prenatal risk factors (i.e., low birth weight, prematurity, maternal emotional problems, maternal prepregnancy obesity, and obstetric complications) predicted variation in EF and IQ at age 3 years (Camerota and Willoughby, 2019). Similarly, another prospective longitudinal study (involving a community subsample of 310 Canadian families) reported a significant interaction between birth weight and the major allele of a specific oxytocin receptor gene, rs2254298 (associated with brain functioning in regions implicated in EF) as predictors of EF at age 4.5 years (Wade et al., 2018). Another direction for emerging literature on environmental influences on children’s EF development concerns potentially bidirectional associations between EF and obesity. As noted in a systematic review of this field (Reinert et al., 2013), establishing causality between EF and obesity requires both longitudinal designs and greater uniformity in assessing EF. Responding to this challenge, Groppe and Elsner (2015) conducted a 1-year longitudinal study of 1657 children attending German elementary schools and reported a bidirectional association between BMI (body mass index) and both hot and cool aspects of EF, providing valuable guidance for intervention studies, summarized in a Cochrane review (Martin et al., 2018). Clinical studies have also highlighted the importance of environmental influences. For example, studies of children who have experienced maltreatment or severe neglect highlight the impact of such extreme adverse environments on neuroendocrine and autonomic stress reactivity, which in turn leads to increased demands on EF systems of regulatory control (e.g., Martin et al., 2017). Moreover, findings from children with traumatic brain injury indicate that higher-order brain functions (such as EF) are particularly vulnerable while they are still emerging. Specifically, in their review of the effects of early brain injury on EF, Anderson et al. (2010) examined findings from children with focal brain pathology evident on MRI scans and compared EF performance in late childhood/adolescence (assessed across a variety of domains) for children who sustained early brain injury at each of six developmental periods (congenital/perinatal/infancy/ preschool/midchildhood/late childhood), with these six groups being matched for gender, SES, lesion size, location, or laterality. Their findings clearly supported theoretical perspectives that emphasize vulnerability rather than plasticity in brain function. Specifically, children who experienced brain injury very early in life displayed markedly more severe

548 PART | II Cognitive development

deficits in EF (and IQ). In other words, while the development of EF can be disrupted, with either transient or more permanent consequences, once established, EF skills are relatively robust (Johnson, 2005; Thomas and Johnson, 2008), with more recent findings highlighting the utility of EF as an index of lesion severity (Ringdahl et al., 2019).

25.3.1 Early executive function predicts academic, socio-cognitive and social success at school This third section aims to bring together the findings from research tracing links between EF and children’s academic, sociocognitive, and social success at school. Interestingly, each of these research areas highlights different aspects of EF: working memory appears central in accounts of EF and academic performance; cognitive flexibility is highlighted by more than one account of EF and social cognition; and inhibitory control is central to accounts of EF and social success. At the same time, overlapping associations are likely, as there is such close interplay between children’s academic ability, social understanding, and social behavior at school. EF and academic performance. Over the past three decades, research into the neurocognitive underpinnings of children’s competence in core academic domains such as literacy and numeracy has expanded rapidly; interestingly, several accounts give EF (especially working memory) a prominent role (e.g., Blair and Razza, 2007; Gathercole and Pickering, 2000; Geary, 1990). Empirical research confirms the importance of working memory for academic achievement. For example, in a longitudinal study, Alloway and Alloway (2010) found that working memory at age 5 years (i.e., at the start of formal education) eclipsed IQ as a predictor of academic success 6 years later. Likewise, accounts of academic failure among children and adolescents with ADHD also highlight the importance of deficits in EF (Alloway et al., 2010; Biederman et al., 2004; Clark et al., 2000), whereas Dahlin (2010) found that primary-school children with special needs who completed a cognitive training program designed to enhance working memory also displayed accelerated reading development. That said, the role of working memory in mathematical competence is more complex than previously thought (Geary, 2010). In particular, echoing the finding (discussed in the first section of this chapter) that developmental timing is a stronger predictor of EF impairment than location of brain injury (Anderson et al., 2010), studies have shown significant changes in brainemathematics relations as children develop and mature (Ansari, 2010; Meyer et al., 2010). Moreover, RCT findings indicate that intervention strategies that rely on repeatedly playing computer games provide only transient improvements in working memory (Roberts et al., 2016), although effects may be more robust among specific groups of children, such as those with language disorders (Acosta et al., 2019). Working with younger children, and adopting a rather different theoretical perspective (in which early EF rather than academic performance is center stage), Blair and colleagues have shown that individual differences in EF in preschool predict both school readiness and children’s success in numeracy and literacy (Blair and Diamond, 2008; Blair and Peters, 2003; Blair and Razza, 2007; Razza and Blair, 2009). In their review of this field, Blair and Diamond (2008) argue that early self-regulation reflects an emerging balance between emotional arousal and cognitive regulation such that selfregulation (and hence children’s school readiness) is likely to be enhanced by school interventions that link emotional/ motivational arousal with activities designed to promote EF. Consistent with this view, findings from the Family Life Project highlight EF as a key mediator of effects of poverty on both academic success and social competence (R. Perry et al., 2018). Likewise, Hughes and Ensor (Hughes and Ensor, 2011; Hughes et al., 2010) have extended this research field in two ways: adopting a developmentally dynamic approach to examine growth of EF across the transition to school, and considering individual differences in EF trajectories in relation to children’s own perceptions of their (academic and social) success at school. Research on children’s self-perceived academic abilities has burgeoned over recent years, fueled by the finding that individual differences in IQ fail to account for up to 50% of the variance in academic performance (e.g., Chamorro-Premuzic and Furnham, 2005; Rhode and Thompson, 2007). Indeed, a metaanalytic review has shown that, even controlling for previous achievement, self-perceived abilities exert small but consistent effects on later achievement (Valentine et al., 2004). Conversely, poor self-perceptions in early childhood are associated with loneliness, withdrawal, and peer exclusion (Coplan et al., 2004). Together, these findings suggest that self-perceptions may be a key aspect of children’s psychological (as opposed to practical) school readiness. In their study (in which 191 children were followed from ages 4 to 6 years), Hughes and Ensor (2011) found that, even with effects of concurrent verbal ability and EF controlled, children who had made rapid gains in EF across the transition to school reported higher levels of academic competence at age 6 years. Given that this is the first study to report an association between preschool EF and schoolchildren’s self-perceived abilities, it is worth noting that similar findings have been obtained from adult samples. In particular, Tangney et al. (2004) have reported that, among adults, individual differences in self-control (a construct that is

Executive function: development, individual differences and clinical insights Chapter | 25

549

closely related to EF) show robust associations with individual differences in self-esteem. Complementing this work, however, findings from other studies highlight the need to include effects of classmates’ achievements, as there is a concern regarding “big-fish-little-pond” effects on academic achievement (e.g., Stäbler et al., 2017). EF and social cognition. The finding that individual differences in EF trajectories predict children’s self-perceived academic competence brings us to another hot topic, namely, the robust association between variation in preschoolers’ EF skills and in their understanding of mind. Interestingly, while most often investigated in studies of preschoolers (for a metaanalytic review, see Devine and Hughes, 2014), this link between EF and understanding of mind is also evident in adolescence (for a review, seeDerksen et al., 2018). Equally remarkable, the association between EF and theory of mind has been reported for a variety of clinical groups, including children with ASD (e.g., Pellicano, 2007), hyperactivity or conduct problems (Hughes et al., 1998), fetal alcohol syndrome (Rasmussen et al., 2009; but see also Lindinger et al., 2016), and traumatic brain injuries (Bosco et al., 2017). Thus the link between EF and theory of mind appears pervasive. How then should it be explained? In an early review of the evidence for associations between EF and theory of mind, Perner and Lang (1999) offered five possible explanations: (1) theory-of-mind skills are necessary for children to pass EF tasks (e.g., to inhibit a particular response, children have to be able to represent it as maladaptive); (2) EF is needed for children to develop their understanding of mental states (e.g., experience of goal-directed action improves children’s conceptual understanding of intentional states; (3) the relevant theory-of-mind tasks require EF (e.g., standard false-belief tasks place heavy demands on children’s inhibitory control and/or working memory); (4) tests of both EF and theory of mind require the same kind of embedded conditional reasoning; and (5) the two systems are not functionally related but have overlapping or neighboring neural substrates. Having reviewed the evidence, Perner and Lang (1999) concluded that only the third of these proposals could be ruled out with confidence; a subsequent study (see also Perner et al., 2002) ruled out the fourth account. Since then, evidence that similar brain regions are implicated in EF and theory of mind has grown (for reviews, see Perner and Aichhorn, 2008; Perner et al., 2006); thus, the fifth account (neuroanatomical proximity may well contribute to the association between the two domains) remains plausible. Indeed, dopamine, long recognized as pivotal to EF (e.g., see studies of PKU described in the first section of this chapter), is now recognized as playing a key role in children’s growing understanding of mind (Sabbagh, 2016). For example, findings from an EEG study indicate that the dorsal medial prefrontal cortex (which is rich in dopamine receptors and lies at the end of the mesocortical dopamine pathway) is a specific neurodevelopmental correlate of preschoolers’ theory-of-mind development (Sabbagh et al., 2009). Likewise, using an archive of preschoolers’ EEG recordings, Lackner et al. (2010) have reported that individual differences in eye-blink rate (an indirect but reliable measure of dopamine function) predicted theory-of-mind performance even controlling for several other related factors, including age, verbal ability, gender, and performance on a stroop task (which taps the ability to inhibit a maladaptive response). Interestingly, as Lackner et al. (2010) note, dopamine provides a mechanism that may explain both neurobiological and experiential influences on theory-of-mind development. Specifically, dopamine promotes the neural plasticity needed to respond flexibly to environmental feedback by changing goals and expectations (e.g., Montague et al., 2004) and so may mediate the impact of family factors known to predict theory-of-mind development (e.g., frequencies of family conversations about mental states (Ensor and Hughes, 2008), or of interactions with siblings (Devine and Hughes, 2017) that depend on children’s ability to reflect on (and revise) their own concepts of mind to accommodate new information from the environment. Another area of progress in this research field has been the growth of longitudinal studies, including microgenetic studies (e.g., Flynn, 2006) and studies of toddlers (Carlson et al., 2004; Hughes and Ensor, 2007). One consistent finding to emerge in a metaanalytic review (Devine and Hughes, 2014) is that early EF predicts later mental-state awareness more strongly than early mental-state awareness predicts later EF. This asymmetry in predictive relationships challenges Perner and Lang’s first account (namely that theory of mind provides a foundation for EF). Instead, without going as far as stipulating that EF is, in some sense, necessary for the emergence of mental-state awareness, it seems reasonable to argue that EF improvements in the preschool years help explain how children make use of their early intuitive understanding of mind. It is also worth noting that the relationship between EF and theory of mind may well be developmentally dynamic. For example, in a critique of the original theory-of-mind account of ASD (which is often diagnosed long before children are expected to pass false-belief tasks), Tager-Flusberg (2001) proposed that early-onset “socioperceptual” skills (or intuitive mentalizing) depend on modular cognitive processes, whereas later-onset “sociocognitive” skills (or off-line mental-state reasoning) depends on other aspects of cognition, such as language and EF. This model of dual processes (e.g., Apperly et al., 2009; de Vignemont, 2009) also goes some way to explaining why typically developing young children can show quite sophisticated mentalizing skills in their everyday interactions and yet fail experimental false-belief tasks.

550 PART | II Cognitive development

Other models of theory of mind development also suggest a developmentally dynamic relationship with EF. For example, H. Wellman and Liu (2004) proposed that improvements in children’s understanding of mind involve a series of distinct achievements, such that different aspects of EF may be particularly important for specific milestones (for a more recent overview, see Wellman, 2018). EF and social competence. For adults, there is robust evidence that deficits in EF are associated with problems of antisocial behavior: in a metaanalytic review of the adult literature, Morgan and Lilienfeld (2000) reported that the average EF performance of antisocial groups fell 0.62 standard deviations below that of control groups. Building on this work, Raine (2002) reviewed evidence from neuropsychological, neurological, and brain imaging studies and concluded that prefrontal structural and functional deficits are implicated in antisocial or aggressive behavior throughout the life span. Beauchamp and Anderson (2010) conducted a theoretical review of the cognitive underpinnings of children’s developing social skills and noted that, within EF, attentional control (i.e., self-monitoring, response inhibition and self-regulation) is especially critical. In one of the earliest studies to link EF to young children’s social competence (and the first to use direct observational methods), Hughes et al. (2000) found that, in a socially diverse sample of preschoolers (half of whom had been rated by parents as “hard to manage”) poor performance on a battery of EF tasks (but not on theory of mind tasks) was associated with higher frequencies of angry and antisocial behavior toward friends. In other words, the interpersonal problems of these hard-to-manage preschoolers appear not to reflect difficulties in social understanding per se, but rather failure of behavioral regulation. In a follow-up study with the same sample, Hughes et al. (2001) showed that poor EF at age 4 years predicted negative behavior at age 5 and that this group of hard-to-manage children continued to show rule violations and perseverative errors at age 7 years (Brophy et al., 2002). In a further longitudinal observational study (for more details, see Hughes, 2011), Hughes and colleagues followed a socially diverse sample of 140 children from toddlerhood through to school age, recruiting best friends for each target child at age 4 years, such that their findings are best presented in two parts. In the first, Hughes and Ensor (2008) examined children’s EF, verbal ability, and theory of mind scores at ages 2, 3, and 4 years in relation to aggregate (multiinformant, multisetting) measures of problem behaviors at each time point and found that (1) poor EF at age 2 years predicted worsening problem behaviors from ages 2 to 4 years; (2) individual differences in EF at age 3 years fully mediated the influence of age 2 language deficits upon age 4 problem behaviors; and (3) by age 4 years, individual differences in problem behaviors showed specific associations with individual differences in EF (but not theory of mind or verbal ability). Capitalizing on the expanded sample (N ¼ 191), Hughes and Ensor applied latent growth models, which showed that high EF gains across the transition to school predicted low levels of teacher-rated emotional symptoms, hyperactivity, conduct problems, and peer problems at age 6 years (as well as higher self-reported academic competence, as noted earlier). Together, these findings highlight the importance of preschool EF for early social adjustment and demonstrate the value of tracking developmental change in EF and in social competence in tandem (e.g., Devine et al., 2016). Links between EF and social competence may also be indirect. For example, Razza and Blair (2009) have reported that false-belief understanding mediates the association between early individual differences in children’s EF and later teacher ratings of children’s social competence. On a related note, Maszk et al. (1999) found that 4- to 6-year-olds rated by peers and teachers as high in behavioral and emotional self-control became increasingly popular over the school year and so argued that individual differences in self-control may be meaningful for how children are viewed by others, and hence for how they view themselves. This raises the interesting possibility that children’s awareness of their own gains in EF across the transition to school may shape both social behavior and self-concepts. If so, interventions to improve children’s social adjustment might aim beyond increasing children’s EF to ensuring that children are aware of their own progress in regulating and organizing their thoughts and behaviors. At this point, it’s worth noting that the findings reviewed in this chapter and elsewhere suggest that EF can act as moderator, mediator, and outcome of interventions. For example, the effects of interventions are often particularly impressive for children with poor EF (e.g., Sasser et al., 2017). Likewise, improvements in children’s inhibitory control have been shown to at least partially mediate the positive effects of the PATHS curriculum on children’s behavior (Riggs et al., 2006). Similarly, all the three RCTs reviewed earlier in this chapter demonstrated that positive effects on EF can be expected from enriched and structured curricula that promote scaffolding (and hence lead to experiences of successful problem-based learning) and adult mind-mindedness (enabling a progression from external to internal forms of selfregulation). These multiple roles for EF in interventions to promote social behavior highlight the importance of adopting a broad and contextualized approach to identifying underlying mechanisms. Support for this conclusion comes from a recent metaanalytic review of EF interventions (Takacs and Kassai, 2019) that included 90 studies (N ¼ 8925) and demonstrated that atypically developing children benefit less from explicit training (e.g., via computerized games) than

Executive function: development, individual differences and clinical insights Chapter | 25

551

from implicit training aimed at enabling children to acquire new strategies for self-regulation. Within a developmental perspective, the next step is to identify whether specific age groups require different intervention strategies to facilitate their EF development, to maximize all children’s long-term positive outcomes.

25.4 Conclusions This chapter on EF in childhood has covered considerable ground, including development from infancy to adolescence in typically developing and atypical groups; positive and negative effects of environmental influences (from training on specific tasks to exposure to enriched and predictable environments); and academic, sociocognitive, and social outcomes associated with individual differences in early EF (or EF trajectories). Perhaps the main pair of conclusions to emerge from this review is that both continuities and contrasts in EF in children of different ages are striking. One striking continuity across studies of typically developing children of different ages is a close interplay between top-down systems of EF and bottom-up reward-oriented systems, such that, from infancy through to adolescence, poor EF appears associated with risk taking and sensation seeking. Indeed, research on several atypical groups, including children with ADHD, CD, or problems of substance abuse, also highlights this interplay between top-down and bottom-up processes. Although not yet evident in research on ASD, it’s worth noting that the amygdala, which is a key substrate involved in reward processing, is central to at least one prominent account of ASD (Baron-Cohen et al., 2000). Thus, extending this dual focus on EF and reward processing to children with ASD would appear a fruitful direction for future research. A second notable developmental continuity is that, across a wide age range, typically developing individuals with good EF are more likely than their peers to do well on tests of theory of mind and show positive self-concepts, and less likely to display antisocial behaviors. Perhaps related to these stable correlates of EF, longitudinal studies support EF as a predictor of later academic achievement in both young children and adolescents. Finally, across a wide variety of ages, at least some aspects of EF (e.g., working memory) appear malleable to training effects. Examples of age-related contrasts include differences in the nature of EF: Improvements in some aspects of EF (such as inhibitory control) can be seen from a very early age, whereas other aspects (e.g., planning) do not show marked improvements until much later on in development. Another important contrast concerns the extent to which EF can be associated with a localized neural base: Age-related improvements in EF appear hand in hand with an increase in frontostriatal activation, such that development is characterized by a progression from diffuse to specific neural substrate. Several age-related functional changes in children’s performance on EF tasks suggest that this progressive localization of neural substrate may reflect increases in how strategic children and adolescents are when completing EF tasks. For example, adults and children differ markedly in how they respond to more challenging situations; while adults can reduce their speed of response to remain accurate, young children typically show a drop in accuracy. Similarly, young children are particularly likely to show an “all or none” effect, in which they can inhibit a response if this is consistently required of them, but find it much harder to cope with situations that place varying demands on this system of inhibitory control. Finally, related to these contrasts in strategy use, training studies indicate an age-related contrast in the optimal format of the training tasks, with task variability increasing training benefits in adults, but reducing training benefits in children. Together, the above age-related contrasts lead to a third key conclusion; namely the need to take developmental issues seriously when examining a construct such as EF that shows such a protracted developmental course. For example, if the differences noted above do indeed reflect an age-related contrast in strategy use on EF tasks, then the validity of cross-age comparisons is in question, as different sets of skills may well underpin performance on the same task for children of different ages. An important first step in addressing this issue is to establish measurement invariance before comparing EF skills across different age groups (Hughes et al., 2010). Developmental issues are also raised by findings from studies of atypical groups. Thus, studies comparing different clinical groups (e.g., children with ASD and children with ADHD) should be designed, so that contrasts in developmental trajectories can be elucidated. The few existing studies that adopt a developmental perspective indicate that children with ASD may show greater progress than children with ADHD, but the reasons for this are not yet known. The final conclusions to emerge from this review concern the interplay between EF and children’s environments. First, although individual differences in EF have been viewed as almost entirely genetic in origin (e.g., Friedman et al., 2008), there is growing evidence that, for young children at least, environmental influences can be substantial. Thus, detailed longitudinal studies highlight the importance of family factors (e.g., maternal well-being, sensitivity, and consistency of parenting). In addition, at least three randomized controlled trials show positive effects on EF of early educational interventions. Second, these positive effects may be (1) strongest for children with low levels of EF (i.e., EF moderates the impact of interventions); (2) pivotal to explaining the substantial improvement in children’s behavior as a result of such

552 PART | II Cognitive development

interventions (i.e., EF is a mediator of intervention effects); and (3) achieved indirectly, via improvements in children’s theory of mind skills, or in how children are viewed by others or view themselves. Clearly then, tracing out the mechanisms that underpin associations between family environments and children’s growing EF skills, and between interventions and children’s social and cognitive achievements, is the important challenge for future research.

References Acosta, V., Hernandez, S., Ramirez, G., 2019. Effectiveness of a working memory intervention program in children with language disorders. Appl. Neuropsychol. 8, 15e23. Aksan, N., Kochanska, G., 2004. Links between systems of inhibition from infancy to preschool years. Child Dev. 75, 1477e1490. Allan, D., Allan, N., Lerner, M., Farrington, A., Lonigan, C., 2015. Identifying unique components of preschool children’s self-regulatory skills using executive function tasks and continuous performance tests. Early Child. Res. Q. 32, 40e50. https://doi.org/10.1016/j.ecresq.2015.02.001. Alloway, T., Alloway, R., 2010. Investigating the predictive roles of working memory and IQ in academic attainment. J. Exp. Child Psychol. 106, 20e29. Alloway, T., Gathercole, S., Elliott, J., 2010. Examining the link between working memory behaviour and academic attainment in children with ADHD. Dev. Med. Child Neurol. 52, 632e636. Anderson, P., 2002. Assessment and development of executive function (EF) during childhood. Child Neuropsychol. 8, 71e82. Anderson, P., Anderson, V., Garth, J., 2001. Assessment and development of organisational ability: the Rey Complex Figure organisational strategy score (RCF-OSS). Clin. Neuropsychol. 15, 81e94. Anderson, P., Anderson, V., Lajoie, G., 1996. The tower of London test: validation and standardization for pediatric populations. Clin. Neuropsychol. 10, 54e65. Anderson, V., Catroppa, C., 2005. Recovery of executive skills following pediatric traumatic brain injury (TBI): a two year follow-up. Brain Inj. 19, 459e470. Anderson, V., Spencer-Smith, M., Coleman, L., Anderson, P., Williams, J., Greenham, M., et al., 2010. Children’s executive functions: are they poorer after very early brain insult. Neuropsychologia 48, 2041e2050. Anderson, W.W., Damasio, H., Tranel, D., Damasio, A.R., 2001. Long-term sequelae of prefrontal cortex damage acquired in early childhood. Dev. Neuropsychol. 18, 281e296. Ansari, D., 2010. Neurocognitive approaches to developmental disorders of numerical and mathematical cognition: the perils of neglecting the role of development. Learn. Individ. Differ. 20, 123e129. Apperly, I., Samson, D., Humphreys, G., 2009. Studies of adults can inform accounts of theory of mind development. Dev. Psychol. 45, 190e201. Asbury, K., Dunn, J., Pike, A., Plomin, R., 2003. Nonshared environmental influences on individual differences in early behavioral development: a monozygotic twin differences study. Child Dev. 74, 933e943. Asbury, K., Plomin, R., 2013. G Is for Genes: The Impact of Genetics on Education and Achievement. Wiley Blackwell, Oxford. Asbury, K., Wachs, T., Plomin, R., 2005. Environmental moderators of genetic influence on verbal and nonverbal abilities in early childhood. Intelligence 33, 643e661. Association, A.P., 2013. Diagnostic and Statistical Manual of Mental Disorders, fifth ed. American Psychiatric Publishing, Arlington, VA. Aytaclar, S., Tarter, R., Kirisci, L., Sandy, L., 1999. Association between hyperactivity and executive cognitive functioning in childhood and substance use in early adolescence. J. Am. Acad. Child Adolesc. Psychiatry 38, 172e178. Azeredo, A., Moreira, D., Barbosa, F., 2018. ADHD, CD, and ODD: systematic review of genetic and environmental risk factors. Res. Dev. Disabil. 82, 10e19. https://doi.org/10.1016/j.ridd.2017.12.010. Baillargeon, R., Spelke, E., Wasserman, S., 1985. Object permanence in five-month-old infants. Cognition 20, 191e208. Barnett, W., Jung, K., Yarosz, D., Thomas, J., Hornbeck, A., Stechuk, R., Burns, S., 2008. Educational effects of the tools of the mind curriculum: a randomized trial. Early Child. Res. Q. 23, 299e313. Baron-Cohen, S., Ring, H., Bullmore, E., Wheelwright, S., Ashwin, C., Williams, S., 2000. The amygdala theory of autism. Neurosci. Biobehav. Rev. 24, 355e364. Beauchamp, M., Anderson, V., 2010. SOCIAL: an integrative framework for the development of social skills. Psychol. Bull. 136, 39e64. Belsky, J., De Haan, M., 2011. Annual research review: parenting and children’s brain development: the end of the beginning. J. Child Psychol. Psychiatry Allied Discip. 52, 409e428. Bernier, A., Carlson, S., Whipple, N., 2010. From external regulation to self-regulation: early parenting precursors of young children’s executive functioning. Child Dev. 81, 326e339. https://doi.org/10.1111/j.1467-8624.2009.01397.x. Best, J., Miller, P., Jones, L., 2009. Executive functions after age 5: changes and correlates. Dev. Rev. 29, 180e200. Bidwell, L., Willcutt, E., DeFries, J., Pennington, B., 2007. Testing for neuropsychological endophenotypes in siblings discordant for attention-deficit/ hyperactivity disorder. Biol. Psychiatry 62, 991e998. Biederman, J., Monuteaux, M.C., Doyle, A.E., Seidman, L.J., Wilens, T.E., Ferrero, F., Morgan, C.L., Faraone, S.V., 2004. Impact of Executive Function Deficits and Attention-Deficit/Hyperactivity Disorder (ADHD) on Academic Outcomes in Children. J. Consult. Clin. Psychol 72, 757e766. Bierman, K., Nix, R., Greenberg, M., Blair, C., Domitrovich, C., 2008. Executive functions and school readiness intervention: impact, moderation, and mediation in the Head Start REDI program. Dev. Psychopathol. 20, 821e843. Blair, C., 2016. Executive function and early childhood education. Curr. Opin. Behav. Sci. 10, 102e107. https://doi.org/10.1016/j.cobeha.2016.05.009.

Executive function: development, individual differences and clinical insights Chapter | 25

553

Blair, C., Diamond, A., 2008. Biological processes in prevention and intervention: the promotion of self-regulation as a means of preventing school failure. Dev. Psychopathol. 20, 899e911. Blair, C., Peters, R., 2003. Physiological and neurocognitive correlates of adaptive behavior in preschool among children in Head Start. Dev. Neuropsychol. 24, 479e497. Blair, C., Razza, R., 2007. Relating effortful control, executive function, and false belief understanding to emerging math and literacy ability in kindergarten. Child Dev. 78, 647e663. Blair, C., Zelazo, P., Greenberg, M., 2005. The measurement of executive function in early childhood. Dev. Neuropsychol. 28, 561e571. Blakemore, S.-J., Choudhury, S., 2006. Development of the adolescent brain: implications for executive function and social cognition. J. Child Psychol. Psychiatry Allied Discip. 47, 296e312. Bodrova, E., Leong, D., 1996. Tools of the Mind: The Vygotskian Approach to Early Childhood Education. Merrill/Prentice Hall, Englewood Cliffs, New Jersey. Bosco, F., Parola, A., Sacco, K., Zettin, M., Angeleri, R., 2017. Communicative-pragmatic disorders in traumatic brain injury: the role of theory of mind and executive functions. Brain Lang. 168, 73e83. https://doi.org/10.1016/j.bandl.2017.01.007. Breukelaar, I., Antees, C., Grieve, S., Foster, S., Gomes, L., Williams, L., Korgaonkar, M., 2017. Cognitive control network anatomy correlates with neurocognitive behavior: a longitudinal study. Hum. Brain Mapp. 38, 631e643. https://doi.org/10.1002/hbm.23401. Brocki, K., Clerkin, S., Guise, K., Fan, J., Fossella, J., 2009. Assessing the molecular genetics of the development of executive attention in children: focus on genetic pathways related to the anterior cingulate cortex and dopamine. Neuroscience 164, 241e246. Brophy, M., Taylor, E., Hughes, C., 2002. To go or not to go: inhibitory control in ’hard to manage’ children. Infant Child Dev. 11, 125e140. Brydges, C., Landes, J., Reid, C., Campbell, C., French, N., Anderson, M., 2018. Cognitive outcomes in children and adolescents born very preterm: a meta-analysis. Dev. Med. Child Neurol. 60, 452e468. https://doi.org/10.1111/dmcn.13685. Camerota, M., Willoughby, M., 2019. Prenatal risk predicts preschooler executive function: a cascade model. Child Dev. https://doi.org/10.1111/ cdev.13271. Carlson, S., 2003. Executive function in context: development, measurement, theory, and experience. Monogr. Soc. Res. Child Dev. 68, 138e151. Carlson, S., Mandell, D., Williams, L., 2004. Executive function and theory of mind: stability and prediction from ages 2 to 3. Dev. Psychol. 40, 1105e1122. Chamorro-Premuzic, T., Furnham, A., 2005. Personality and Intellectual Competence. Erlbaum, Hillsdale, NJ. Chevalier, N., Jackson, J., Revueltas Roux, A., Moriguchi, Y., Auyeung, B., 2019. Differentiation in prefrontal cortex recruitment during childhood: evidence from cognitive control demands and social contexts. Dev. Cognit. Neurosci. 36 https://doi.org/10.1016/j.dcn.2019.100629 art. no. 100629. Churchwell, J., Yurgelun-Todd, D., 2013. Age-related changes in insula cortical thickness and impulsivity: significance for emotional development and decision-making. Dev. Cognit. Neurosci. 6, 80e86. https://doi.org/10.1016/j.dcn.2013.07.001. Clark, C., Prior, M., Kinsella, G., 2000. Do executive function deficits differentiate between adolescents with AD/HD and Oppositional Defiant/Conduct Disorder? A neuropsychological study using the six elements test and Hayling Sentence Completion Test. J. Abnorm. Child Psychol. 28, 403e414. Constantinidis, C., Luna, B., 2019. Neural substrates of inhibitory control maturation in adolescence. Trends Neurosci. 42, 604e616. https://doi.org/ 10.1016/j.tins.2019.07.004. Coplan, R., Findlay, L., Nelson, L., 2004. Characteristics of preschoolers with lower perceived competence. J. Abnorm. Child Psychol. 32, 399e408. Craig, F., Margari, F., Legrottaglie, A., Palumbi, R., de Giambattista, C., Margari, L., 2016. A review of executive function deficits in autism spectrum disorder and attention-deficit/hyperactivity disorder. Neuropsychiatric Dis. Treat. 12, 1191e1202. https://doi.org/10.2147/NDT.S104620. Cuevas, K., Bell, M., 2014. Infant attention and early childhood executive function. Child Dev. 85, 397e404. Dahlin, K., 2010. Effects of working memory training on reading in children with special needs. Read. Writ. 23, 1e13. Davidson, M., Amso, D., Anderson, L., Diamond, A., 2006. Development of cognitive control and executive functions from 4 to 13 years: evidence from manipulations of memory, inhibition, and task switching. Neuropsychologia 44, 2037e2078. De Luca, C., Wood, S., Anderson, V., Buchanan, J., Profitt, T., Mahoney, K., Pantelis, C., 2003. Normative data from the CANTAB: development of executive function over the lifespan. J. Clin. Exp. Neuropsychol. 25, 242e254. de Vignemont, F., 2009. Drawing the boundary between low-level and high-level mindreading. Philos. Stud. 144, 457e466. Demetriou, E., Lampit, A., Quintana, D., Naismith, S., S, Y.J.C., Pye, J., Hickie, I., Guastella, A.J., 2018. Autism spectrum disorders: a meta-analysis of executive function. Mol. Psychiatry 23, 1198e1204. https://doi.org/10.1038/mp.2017.75. Derksen, D., Hunsche, M., Giroux, M., Connolly, D., Bernstein, D., 2018. A systematic review of theory of mind’s precursors and functions. Z. Psychol./ J. Psychol. 226, 87e97. https://doi.org/10.1027/2151-2604/a000325. Devine, R., Hughes, C., 2014. Relations between false-belief understanding and executive function in early childhood: a meta-analysis. Child Dev. Devine, R., Hughes, C., 2017. Family correlates of false belief understanding in early childhood: a meta-analysis. Child Dev. https://doi.org/10.1111/ cdev.12682. Early Online View. Devine, R., Ribner, A., Hughes, C., 2019. Measuring and predicting individual differences in executive function at 14 months. Child Dev. 85 (5), 1777e1794. Devine, R., White, N., Ensor, R., Hughes, C., 2016. Theory of mind, executive function and social adjustment across middle childhood: a longitudinal study. Dev. Psychol. 52, 758e771. Diamond, A., 1988. Abilities and neural mechanisms underlying A not B performance. Child Dev. 59, 523e527. Diamond, A., 1991. Developmental time course in human infants and infant monkeys, and the neural bases of inhibitory control in reaching. Ann. N.Y. Acad. Sci. 637e704.

554 PART | II Cognitive development

Diamond, A., 2002. Normal development of prefrontal cortex from birth to young adulthood: cognitive functions, anatomy, and biochemistry. In: Stuss, D., Knight, R. (Eds.), Principles of Frontal Lobe Function. Oxford University Press, London, UK, pp. 466e503. Diamond, A., 2009. All or none hypothesis: a global-default mode that characterizes the brain and mind. Dev. Psychol. 45, 130e138. Diamond, A., Barnett, W., Thomas, J., Munro, S., 2007. Preschool program improves cognitive control. Science 318, 1387e1388. Diamond, A., Goldman-Rakic, P., 1989. Comparison of human infants and rhesus monkeys on Piaget’s A-not-B task: evidence for dependence on dorsolateral prefrontal cortex. Exp. Brain Res. 74, 24e40. Diamond, A., Prevor, M., Callender, G., Druin, D., 1997. Prefrontal cortex cognitive deficits in children treated early and continuously for PKU. Monogr. Soc. Res. Child Dev. 62. Diamond, A., Zola-Morgan, S., Squire, L., 1989. Successful performance by monkeys with lesions of the hippocampal formation on AB and object retrieval, two tasks that mark developmental changes in human infants. Behav. Neurosci. 103, 526e537. Doidge, J., Flora, D., Toplak, M., 2018. A meta-analytic review of sex differences on delay of gratification and temporal discounting tasks in ADHD and typically developing samples. J. Atten. Disord. https://doi.org/10.1177/1087054718815588. Duncan, J., 2001. An adaptive coding model of neural function in prefrontal cortex. Nat. Rev. Neurosci. 2, 820e829. Dunn, J., 1993. Young Children’s Close Relationships: Beyond Attachment, vol. 4. Sage Publications, Newbury Park, CA. Ellefson, M., Ng, F., Wang, Q., Hughes, C., 2017. Efficiency of executive function: a two-generation cross-cultural comparison of samples from Hong Kong and the United Kingdom. Psychol. Sci. 28, 555e566. https://doi.org/10.1177/0956797616687812. Ensor, R., Hughes, C., 2008. Content or connectedness? Early family talk and theory of mind in the toddler and preschool years. Child Dev. 79, 201e216. Fairchild, G., van Goozen, S., Stollery, S., Aitken, M., Savage, J., Moore, S., Goodyer, I., 2009. Decision making and executive function in male adolescents with early-onset or adolescence-onset conduct disorder and control subjects. Biol. Psychiatry 66, 162e168. Farah, M., Shera, D., Savage, J., Betancourt, L., Giannetta, J., Brodsky, N., et al., 2006. Childhood poverty: specific associations with neurocognitive development. Brain Res. 1110, 166e174. Fiske, A., Holmboe, K., 2019. Neural substrates of early executive function development. Dev. Rev. 52, 42e62. https://doi.org/10.1016/j.dr.2019.100866. Flook, L., Smalley, S., Kitil, M., Galla, B., Kaiser-Greenland, S., Locke, J., et al., 2010. Effects of mindful awareness practices on executive functions in elementary school children. J. Appl. Sch. Psychol. 26, 70e95. Flynn, E., 2006. A microgenetic investigation of stability and continuity in theory of mind development. Br. J. Dev. Psychol. 24, 631e654. Fossella, J., Sommer, T., Fan, J., Wu, Y., Swanson, J., Pfaff, D., Posner, M., 2002. Assessing the molecular genetics of attention networks. BMC Neurosci. 3, 14. Freedman, D.J., Riesenhuber, M., Poggio, T., Miller, E.K., 2001. Categorical representation of visual stimuli in the primate prefrontal cortex. Science 291, 312e316. Frick, M., Brocki, K.C., 2019. A multi-factorial perspective on ADHD and ODD in school-aged children: what is the role of cognitive regulation, temperament, and parental support? J. Clin. Exp. Neuropsychol. https://doi.org/10.1080/13803395.2019.1641185. Friedman, L., Sterling, A., 2019. A review of language, executive function, and intervention in autism spectrum disorder. Semin. Speech ̄ Lang. 40, 291e304. https://doi.org/10.1055/s-0039-1692964. Friedman, N., Miyake, A., 2017. Unity and diversity of executive functions: individual differences as a window on cognitive structure. Cortex 186e204. Friedman, N., Miyake, A., Altamirano, L., Corley, R., Young, S., Rhea, S., Hewitt, J., 2016. Stability and change in executive function abilities from late adolescence to early adulthood: a longitudinal twin study. Dev. Psychol. 52, 326e340. https://doi.org/10.1037/dev0000075. Friedman, N., Miyake, A., Robinson, J., Hewitt, J., 2011. Developmental trajectories in toddlers’ self-restraint predict individual differences in executive functions 14 years later: a behavioral genetic analysis. Dev. Psychol. 47, 1410e1430. Friedman, N., Miyake, A., Young, S., DeFries, J., Corley, R., Hewitt, J., 2008. Individual differences in executive functions are almost entirely genetic in origin. J. Exp. Psychol. Gen. 137, 201e225. Garon, N., Bryson, S., Smith, I., 2008. Executive function in preschoolers: a review using an integrative framework. Psychol. Bull. 134, 31e60. Gathercole, S., Pickering, S., 2000. Working memory deficits in children with low achievements in the national curriculum at 7 years of age. Br. J. Educ. Psychol. 70, 177e194. Gau, S.-F., Shang, C.-Y., 2010. Executive functions as endophenotypes in ADHD: evidence from the Cambridge neuropsychological test battery (CANTAB). J. Child Psychol. Psychiatry Allied Discip. 51, 838e849. Geary, D., 1990. A componential analysis of an early learning deficit in mathematics. J. Exp. Child Psychol. 49, 363e383. Geary, D., 2010. Mathematical disabilities: reflections on cognitive, neuropsychological, and genetic components. Learn. Individ. Differ. 20, 130e133. Geurts, H., Verté, S., Oosterlaan, J., Roeyers, H., Sergeant, J., 2004. How specific are executive functioning deficits in attention deficit hyperactivity disorders and autism? J. Child Psychol. Psychiatry Allied Discip. 45, 836e854. Goldberg, M., Mostofsky, S., Cutting, L., Mahone, E., Astor, B., Denckla, M., Landa, R., 2005. Subtle executive impairment in children with autism and children with ADHD. J. Autism Dev. Disord. 35, 279e293. Golden, C., 1981. Diagnosis and Rehabilitation in Clinical Neuropsychology. Charles C. Thomas, Springfield. Goschke, T., 2014. Dysfunctions of decision-making and cognitive control as transdiagnostic mechanisms of mental disorders: advances, gaps, and needs in current research. Int. J. Methods Psychiatr. Res. 23, 41e57. https://doi.org/10.1002/mpr.1410. Groppe, K., Elsner, B., 2015. The influence of hot and cool executive function on the development of eating styles related to overweight in children. Appetite 87, 127e136. https://doi.org/10.1016/j.appet.2014.12.203. Gustavson, D., Stallings, M.C., Corley, R.P., Miyake, A., Hewitt, J., F, N.P., 2017. Executive functions and substance use: relations in late adolescence and early adulthood. J. Abnorm. Psychol. 126, 257e270. https://doi.org/10.1037/abn0000250.

Executive function: development, individual differences and clinical insights Chapter | 25

555

Habib, A., Harris, L., Pollick, F., Melville, C., 2019. A meta-analysis of working memory in individuals with autism spectrum disorders. PLoS One 14, e0216198. https://doi.org/10.1371/journal.pone.0216198 art. no. Hackman, D., Gallop, R., Evans, G., Farah, M., 2015. Socioeconomic status and executive function: developmental trajectories and mediation. Dev. Sci. 18, 686e702. https://doi.org/10.1111/desc.12246. Han, Y., Chan, A., 2017. Disordered cortical connectivity underlies the executive function deficits in children with autism spectrum disorders. Res. Dev. Disabil. 61, 19e31. https://doi.org/10.1016/j.ridd.2016.12.010. Happé, F., Hughes, C., Booth, R., Charlton, R., 2006. Executive dysfunction in autism spectrum disorders and attention deficit/hyperactivity disorder: developmental profiles. Brain Cogn. 61, 25e39. Harris, P., 1975. Development of search and object permanence during infancy. Psychol. Bull. 82, 332e344. Hill, E., 2004. Executive function in autism. Trends Cogn. Sci. 8, 26e32. Huang-Pollock, C., Shapiro, Z., Galloway-Long, H., Weigard, A., 2017. Is poor working memory a transdiagnostic risk factor for psychopathology? J. Abnorm. Child Psychol. 45, 1477e1490. https://doi.org/10.1007/s10802-016-0219-8. Hughes, C., 2011. Social Understanding, Social Lives: From Toddlerhood through to the Transition to School. Psychology Press, London. Hughes, C., Cutting, A., Dunn, J., 2001. Acting nasty in the face of failure? Longitudinal observations of ’hard to manage’ children playing a rigged competitive game with a friend. J. Abnorm. Child Psychol. 29, 403e416. Hughes, C., Devine, R., 2019. For better or for worse? Positive and negative parental influences on young children’s executive function. Child Dev. 90, 593e609. https://doi.org/10.1111/cdev.12915. Hughes, C., Devine, R., Wang, Z., 2017. Does parental mind-mindedness account for cross-cultural differences in preschoolers’ theory of mind? Child Dev. https://doi.org/10.1111/cdev.12746. Early Online View. Hughes, C., Dunn, J., White, A., 1998. Trick or treat?: uneven understanding of mind and emotion and executive function among “hard to manage” preschoolers. J. Child Psychol. Psychiatry Allied Discip. 39, 981e994. Hughes, C., Ensor, R., 2007. Executive function and theory of mind: predictive relations from ages 2- to 4-years. Dev. Psychol. 43, 1447e1459. Hughes, C., Ensor, R., 2008. Does executive function matter for preschoolers’ problem behaviors? J. Abnorm. Child Psychol. 36, 1e14. Hughes, C., Ensor, R., 2009. How do families help or hinder the development of executive function? New Dir. Child Adolesc. Psychiatry 123, 35e50. https://doi.org/10.1002/cd.234. Hughes, C., Ensor, R., 2011. Executive function trajectories across the transition to school predict externalizing and internalizing behaviors and children’s self-perceived academic success at age 6. J. Exp. Child Psychol. 108, 663e676. Hughes, C., Ensor, R., Wilson, A., Graham, A., 2010. Tracking executive function across the transition to school: A latent variable approach. Dev. Neuropsychol. 35, 20e36. Hughes, C., Graham, A., 2002. Measuring executive functions in childhood: problems & solutions? Child Adolesc. Ment. Health 7, 131e142. Hughes, C., Graham, A., Grayson, A., 2005. Executive function in childhood: development and disorder. In: Oates, J. (Ed.), Cognitive Development. Open University Press, pp. 205e230. Hughes, C., Roman, G., Hart, M., Ensor, R., 2013. Does maternal depression predict young children’s executive function? A 4-year longitudinal study. J. Child Psychol. Psychiatry 54, 169e177. https://doi.org/10.1111/jcpp.12014. Hughes, C., Russell, J., 1993. Autistic children’s difficulty with mental disengagement from an object: its implications for theories of autism. Dev. Psychol. 29, 498e510. Hughes, C., Russell, J., Robbins, T., 1994. Evidence for central executive dysfunction in autism. Neuropsychologia 32, 477e492. Hughes, C., White, A., Sharpen, J., Dunn, J., 2000. Antisocial, angry and unsympathetic: ‘Hard to manage’ preschoolers’ peer problems, and possible social and cognitive influences. J. Child Psychol. Psychiatry 41, 169e179. Huizinga, M., Dolan, C., van der Molen, M., 2006. Age-related change in executive function: developmental trends and a latent variable analysis. Neuropsychologia 44, 2017e2036. Hwang, K., Ghuman, A., Manoach, D., Jones, S., Luna, B., 2016. Frontal preparatory neural oscillations associated with cognitive control: a developmental study comparing young adults and adolescents. Neuroimage 136, 139e148. https://doi.org/10.1016/j.neuroimage.2016.05.017. Johnson, M., 2005. Sensitive periods in functional brain development: problems and prospects. Dev. Psychobiol. 46, 287e292. Jurado, M., Rosselli, M., 2007. The elusive nature of executive functions: a review of our current understanding. Neuropsychol. Rev. 17, 213e233. Karbach, J., Kray, J., 2009. How useful is executive control training? Age differences in near and far transfer of task-switching training. Dev. Sci. 12, 978e990. Karr, J., Areshenkoff, C., Rast, P., Hofer, S., Iverson, G., Garcia-Barrera, M., 2018. The unity and diversity of executive functions: a systematic review and re-analysis of latent variable studies. Psychol. Bull. 144, 1147e1185. https://doi.org/10.1037/bul0000160. Kharitonova, M., Martin, R., Gabrieli, J., Sheridan, M., 2013. Cortical gray-matter thinning is associated with age-related improvements on executive function tasks. Dev. Cognit. Neurosci. 6, 61e71. https://doi.org/10.1016/j.dcn.2013.07.002. Kingdon, D., Cardoso, C., McGrath, J., 2016. Research Review: executive function deficits in fetal alcohol spectrum disorders and attention-deficit/ hyperactivity disorder e a meta-analysis. J. Child Psychol. Psychiatry Allied Discip. 57, 116e131. https://doi.org/10.1111/jcpp.12451. Klingberg, T., 2010. Training and plasticity of working memory. Trends Cogn. Sci. 14, 317e324. Klingberg, T., Olesen, P., Johnson, M., Gustafsson, P., Dahlström, K., Gillberg, C., et al., 2005. Computerized training of working memory in children with ADHD - a randomized, controlled trial. J. Am. Acad. Child Adolesc. Psychiatry 44, 177e186. Kloo, D., Perner, J., 2003. Training transfer between card sorting and false belief understanding: helping children apply conflicting descriptions. Child Dev. 74, 1823e1839.

556 PART | II Cognitive development

Kouklari, E.-C., Tsermentseli, S., Monks, C., 2019. Developmental trends of hot and cool executive function in school-aged children with and without autism spectrum disorder: links with theory of mind. Dev. Psychopathol. 31, 541e556. https://doi.org/10.1017/S0954579418000081. Kovas, Y., Haworth, C., Dale, P., Plomin, R., 2007. The genetic and environmental origins of learning abilities and disabilities in the early school years. Monogr. Soc. Res. Child Dev. 72 (Boston: Blackwell). Krikorian, R., Bartok, J., Gay, N., 1994. Tower of London procedure: a standard method and developmental data. J. Clin. Exp. Neuropsychol. 16, 840e850. Lackner, C., Bowman, L., Sabbagh, M., 2010. Dopaminergic functioning and preschoolers’ theory of mind. Neuropsychologia 48, 1767e1774. Langberg, J., Epstein, J., Altaye, M., Molina, B., Arnold, L., Vitiello, B., 2008. The transition to middle school is associated with changes in the developmental trajectory of ADHD symptomatology in young adolescents with ADHD. J. Clin. Child Adolesc. Psychol. 37, 651e663. Lawson, G., Hook, C.J., Farah, M., 2018. A meta-analysis of the relationship between socioeconomic status and executive function performance among children. Dev. Sci. 21, e12529. https://doi.org/10.1111/desc.12529 art. no. Lindinger, N., Malcolm-Smith, S., Dodge, N., Molteno, C., Thomas, K., Meintjes, E., et al., 2016. Theory of mind in children with fetal alcohol spectrum disorders. Alcohol Clin. Exp. Res. 40, 367e376. https://doi.org/10.1111/acer.12961. Luciana, M., 2003. Practitioner review: computerized assessment of neuropsychological function in children: clinical and research applications of the Cambridge neuropsychological testing automated battery (CANTAB). J. Child Psychol. Psychiatry 45, 649e663. Luciana, M., Conklin, H., Cooper, C., Yarger, R., 2005. The development of nonverbal working memory and executive control processes in adolescents. Child Dev. 76, 697e712. Luna, B., Garver, K., Urban, T., Lazar, N., Sweeney, J., 2004. Maturation of cognitive processes from late childhood to adulthood. Child Dev. 75, 1357e1372. Mandell, D., Thompson, W., Weintraub, E., DeStefano, F., Blank, M., 2005. Trends in diagnosis rates for autism and ADHD at hospital discharge in the context of other psychiatric diagnoses. Psychiatr. Serv. 56, 56e62. Martin, A., Booth, J., Laird, Y., Sproule, J., Reilly, J., Saunders, D., 2018. Physical activity, diet and other behavioural interventions for improving cognition and school achievement in children and adolescents with obesity or overweight. Cochrane Database Syst. Rev. 2018, CD009728. https:// doi.org/10.1002/14651858.CD009728.pub3 art. no. Martin, M., Davies, P., Cummings, E.M., Cicchetti, D., 2017. The mediating roles of cortisol reactivity and executive functioning difficulties in the pathways between childhood histories of emotional insecurity and adolescent school problems. Dev. Psychopathol. 29, 1483e1498. https://doi.org/ 10.1017/S0954579417000402. Maszk, P., Eisenberg, N., Guthrie, I., 1999. Relations of children’s social status to their emotionality and regulation: a short-term longitudinal study. Merrill-Palmer Q. 45, 468e492. Meyer, M., Salimpoor, V., Wu, S., Geary, D., Menon, V., 2010. Differential contribution of specific working memory components to mathematics achievement in 2nd and 3rd graders. Learn. Individ. Differ. 20, 101e109. Monn, A., Narayan, A., Kalstabakken, A.W., Schubert, E., M, A.S., 2017. Executive function and parenting in the context of homelessness. J. Fam. Psychol. 31, 61e70. https://doi.org/10.1037/fam0000275. Montague, P., Hyman, S., Cohen, J., 2004. Computational roles for dopamine in behavioural control. Nature 431, 760e767. Montroy, J., Merz, E., Williams, J., Landry, S., Johnson, U., Zucker, T., et al., 2019. Hot and cool dimensionality of executive function: model invariance across age and maternal education in preschool children. Early Child. Res. Q. 49, 188e201. https://doi.org/10.1016/j.ecresq.2019.06.011. Morgan, A., Lilienfeld, S., 2000. A meta-analytic review of the relation between antisocial behavior and neuropsychological measures of executive function. Clin. Psychol. Rev. 20, 113e156. Moriguchi, Y., 2014. The early development of executive function and its relation to social interaction: a brief review. Front. Psychol. 5 https://doi.org/ 10.3389/fpsyg.2014.00388 art. no. Article 388. Moriguchi, Y., Hiraki, K., 2013. Prefrontal cortex and executive function in young children: a review of NIRS studies. Front. Hum. Neurosci. 7 https:// doi.org/10.3389/fnhum.2013.00867. Morra, S., Panesi, S., Traverso, L., Usai, M., 2018. Which tasks measure what? Reflections on executive function development and a commentary on Podjarny, Kamawar, and Andrews (2017). J. Exp. Child Psychol. 167, 246e258. https://doi.org/10.1016/j.jecp.2017.11.004. Mulder, H., Verhagen, J., Van der Ven, S., Slot, P., Leseman, P., 2017. Early executive function at age two predicts emergent mathematics and literacy at age five. Front. Psychol. 8, 1706. https://doi.org/10.3389/fpsyg.2017.01706 art. no. Narad, M., Treble-Barna, A., Peugh, J., Yeates, K., Taylor, H., Stancin, T., Wade, S., 2017. Recovery trajectories of executive functioning after pediatric TBI: a latent class growth modeling analysis. J. Head Trauma Rehabil. 32, 98e106. https://doi.org/10.1097/HTR.0000000000000247. Nawaz, S., Hanif, R., Lewis, C., 2015. ‘Theory of mind’ development of Pakistani children: do preschoolers acquire an understanding of desire, pretence and belief in a universal sequence? Eur. J. Dev. Psychol. 12, 177e188. https://doi.org/10.1080/17405629.2014.973843. Nieto, M., Ros, L., Medina, G., Ricarte, J., Latorre, J., 2016. Assessing executive functions in preschoolers using shape school task. Front. Psychol. 7, 1489. https://doi.org/10.3389/fpsyg.2016.01489 art. no. Noble, K., Norman, M.F., Farah, M., 2005. Neurocognitive correlates of socioeconomic status in kindergarten children. Dev. Sci. 8, 74e87. Ozonoff, S., 1995. Reliability and validity of the Wisconsin card sorting test in studies of autism. Neuropsychology 9, 491e500. Ozonoff, S., Pennington, B.F., Rogers, S.J., 1991. Executive function deficits in high functioning autistic children: relationship to theory of mind. J. Child Psychol. Psychiatry 32, 1081e1105. Pauen, S., Bechtel-Kuehne, S., 2016. How toddlers acquire and transfer tool knowledge: developmental changes and the role of executive functions. Child Dev. 87, 1233e1249. https://doi.org/10.1111/cdev.12532.

Executive function: development, individual differences and clinical insights Chapter | 25

557

Paus, T., 2010. Growth of white matter in the adolescent brain: myelin or axon? Brain Cogn. 72, 26e35. Pellicano, E., 2007. Links between theory of mind and executive function in young children with autism: clues to developmental primacy. Dev. Psychol. 43, 974e990. Pellicano, E., 2010. Individual differences in executive function and central coherence predict developmental changes in theory of mind in autism. Dev. Psychol. 46, 530e544. Pennington, B., Ozonoff, S., 1996. Executive function and developmental psychopathology. J. Child Psychol. Psychiatry 37, 51e87. Perner, J., Aichhorn, M., 2008. Theory of mind, language and the temporoparietal junction mystery. Trends Cogn. Sci. 12, 123e126. Perner, J., Aichhorn, M., Kronbichler, M., Staffen, W., Ladurner, G., 2006. Thinking of mental and other representations: the roles of left and right temporo-parietal junction. Soc. Neurosci. 1, 245e258. Perner, J., Lang, B., 1999. Development of theory of mind and executive control. Trends Cogn. Sci. 3, 337e344. Perner, J., Lang, B., Kloo, D., 2002. Theory of mind and self-control: more than a common problem of inhibition. Child Dev. 73 (3), 752e767. Perry, N., Calkins, S., Dollar, J., Keane, S., Shanahan, L., 2018. Self-regulation as a predictor of patterns of change in externalizing behaviors from infancy to adolescence. Dev. Psychopathol. 497e510. https://doi.org/10.1017/S0954579417000992. Perry, R., Braren, S., Blair, C., Vernon-Feagans, L., Cox, M., Blair, C., et al., 2018. Socioeconomic risk and school readiness: longitudinal mediation through children’s social competence and executive function. Front. Psychol. 9, 1544. https://doi.org/10.3389/fpsyg.2018.01544 art. no. Pineda-Alhucema, W., Aristizabal, E., Escudero-Cabarcas, J., Acosta-López, J., Vélez, J., 2018. Executive function and theory of mind in children with ADHD: a systematic review. Neuropsychol. Rev. 28, 341e358. https://doi.org/10.1007/s11065-018-9381-9. Poon, K., 2018. Hot and cool executive functions in adolescence: development and contributions to important developmental outcomes. Front. Psychol. 8, 2311. https://doi.org/10.3389/fpsyg.2017.02311 art. no. Prior, M.R., 1979. Cognitive abilities and disabilities in infantile autism: a review. J. Abnorm. Child Psychol. 7, 357e380. Quistberg, K., Mueller, U., 2019. Prospective relations between kindergarteners’ executive function skills and their externalizing and internalizing behaviors. Clin. Neuropsychol. https://doi.org/10.1080/13854046.2019.1591510. Raine, A., 2002. Annotation: the role of prefrontal deficits, low autonomic arousal, and early health factors in the development of antisocial and aggresive behavior in children. J. Child Psychol. Psychiatry Allied Discip. 43, 417e434. Rasmussen, C., Wyper, K., Talwar, V., 2009. The relation between theory of mind and executive functions in children with fetal alcohol spectrum disorders. Can. J. Clin. Pharmacol. 16, 370e380. Razza, R., Blair, C., 2009. Associations among false-belief understanding, executive function, and social competence: a longitudinal analysis. J. Appl. Dev. Psychol. 30, 332e343. Reinert, K., Po’e, E., Barkin, S.L., 2013. The relationship between executive function and obesity in children and adolescents: a systematic literature review. J. Obesity 2013, 820956. https://doi.org/10.1155/2013/820956/. Resch, C., Anderson, V., Beauchamp, M., Crossley, L., Hearps, S., van Heugten, C., et al., 2019. Age-dependent differences in the impact of paediatric traumatic brain injury on executive functions: a prospective study using susceptibility-weighted imaging. Neuropsychologia 24, 236e245. https:// doi.org/10.1016/j.neuropsychologia.2018.12.004. Rhode, T., Thompson, L., 2007. Predicting academic achievement with cognitive ability. Intelligence 35, 83e92. Richards, C., Jones, C., Groves, L., Moss, J., Oliver, C., 2015. Prevalence of autism spectrum disorder phenomenology in genetic disorders: a systematic review and meta-analysis. Lancet Psychiatry 2, 909e916. https://doi.org/10.1016/S2215-0366(15)00376-4. Riggs, N., Greenberg, M., Kusché, C., Pentz, M., 2006. The mediational role of neurocognition in the behavioral outcomes of a social-emotional prevention program in elementary school students: effects of the PATHS Curriculum. Prev. Sci. 7, 91e102. Ringdahl, E., Becker, M., Hussey, J., Thaler, N., Vogel, S., Cross, C., et al., 2019. Executive function profiles in pediatric traumatic brain injury. Dev. Neuropsychol. 44, 172e188. https://doi.org/10.1080/87565641.2018.1557190. Roberts, G., Quach, J., Spencer-Smith, M., Anderson, P., Gathercole, S., Gold, L., et al., 2016. Academic outcomes 2 years after working memory training for children with lowworking memory: a randomized clinical trial. JAMA Pediatrics 170, e154568. https://doi.org/10.1001/jamapediatrics.2015.4568 art. no. Roman, G., Ensor, R., Hughes, C., 2016. Does executive function mediate the path from mothers’ depressive symptoms to young children’s problem behaviors? J. Exp. Child Psychol. 142, 158e170. Romer, D., Betancourt, L., Giannetta, J., Brodsky, N., Farah, M., Hurt, H., 2009. Executive cognitive functions and impulsivity as correlates of risk taking and problem behavior in preadolescents. Neuropsychologia 47, 2916e2926. Romer, D., Reyna, V., Satterthwaite, T., 2017. Beyond stereotypes of adolescent risk taking: placing the adolescent brain in developmental context. Dev. Cognit. Neurosci. 27, 19e34. https://doi.org/10.1016/j.dcn.2017.07.007. Rothbart, M., Derryberry, D., Hershey, K., 2000. Stability of temperament in childhood: laboratory infant assessment to parent report at seven years. In: Molfese, V., Molfese, D. (Eds.), Temperament and Personality Development across the Life Span, p. 85. Rothbart, M., Ellis, L., Rueda, M., Posner, M., 2003. Developing mechanisms of temperamental effortful control. J. Personal. 71, 1113e1144. Rothbart, M., Posner, M., 2005. Genes and experience in the development of executive attention and effortful control. New Dir. Child Adolesc. Dev. 109, 101e108. Roy, A., Allain, P., Roulin, J.L., Fournet, N., Le Gall, D., 2015. Ecological approach of executive functions using the behavioural assessment of the dysexecutive syndrome for children (BADS-C): developmental and validity study. J. Clin. Exp. Neuropsychol. 37, 956e971. https://doi.org/10.1080/ 13803395.2015.1072138.

558 PART | II Cognitive development

Rueda, M., Rothbart, M., McCandliss, B., Saccomanno, L., Posner, M., 2005. Training, maturation, and genetic influences on the development of executive attention. Proc. Natl. Acad. Sci. U.S.A 102, 14931e14936. Sabbagh, M., 2016. Dopamine and theory of mind in preschoolers. In: Sommerville, J., Decety, J. (Eds.), Social Cognition: Development Across the Life Span. Routledge, pp. 187e209. Sabbagh, M., Bowman, L., Evraire, L., Ito, J., 2009. Neurodevelopmental correlates of theory of mind in preschool children. Child Dev. 80, 1147e1162. Sabbagh, M., Xu, F., Carlson, S., Moses, L., Lee, K., 2006. The development of executive functioning and theory-of-mind: a comparison of Chinese and U.S. preschoolers. Psychol. Sci. 17, 74e81. Sasser, T., Bierman, K., Heinrichs, B., Nix, R., 2017. Preschool intervention can promote sustained growth in the executive-function skills of children exhibiting early deficits. Psychol. Sci. 28, 1719e1730. https://doi.org/10.1177/0956797617711640. Schmitt, L., White, S., Cook, E., Sweeney, J., Mosconi, M., 2018. Cognitive mechanisms of inhibitory control deficits in autism spectrum disorder. J. Child Psychol. Psychiatry Allied Discip. 59, 586e595. https://doi.org/10.1111/jcpp.12837. Schonfeld, A., Paley, B., Frankel, F., O’Connor, M., 2009. Behavioral regulation as a predictor of response to children’s friendship training in children with fetal alcohol spectrum disorders. Clin. Neuropsychol. 23, 428e445. Sheese, B., Rothbart, M., Posner, M., White, L., Fraundorf, S., 2008. Executive attention and self-regulation in infancy. Infant Behav. Dev. 31, 501e510. Silva, N., Linhares, M., Gaspardo, C., 2018. Developmental care approaches for mitigating stress in preterm neonates in the neonatal intensive care unit: a systematic review. Psychol. Neurosci. 11, 117e131. https://doi.org/10.1037/pne0000119. Solomon, M., Ozonoff, S., Ursu, S., Ravizza, S., Cummings, N., Ly, S., Carter, C., 2009. The neural substrates of cognitive control deficits in autism spectrum disorders. Neuropsychologia 47, 2515e2526. Sonuga-Barke, E., 2005. Causal models of attention-deficit/hyperactivity disorder: from common simple deficits to multiple developmental pathways. Biol. Psychiatry 57, 1231e1238. Sonuga-Barke, E., Cortese, S., Fairchild, G., Stringaris, A., 2016. Annual research review: transdiagnostic neuroscience of child and adolescent mental disorders - differentiating decision making in attention-deficit/hyperactivity disorder, conduct disorder, depression, and anxiety. J. Child Psychol. Psychiatry Allied Discip. 57, 321e349. https://doi.org/10.1111/jcpp.12496. Stäbler, F., Dumont, Becker, M., Baumert, J., 2017. What happens to the fish’s achievement in a little pond? A simultaneous analysis of class-average achievement effects on achievement and academic self-concept. J. Educ. Psychol. 109, 191e207. https://doi.org/10.1037/edu0000135. Tager-Flusberg, H., 2001. A re-examination of the Theory of Mind hypothesis of autism. In: Burack, J., Charman, T., Yirmiya, N., Zelazo, P. (Eds.), The Development of Autism: Perspectives from Theory and Research. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 173e194. Takacs, Z., Kassai, R., 2019. The efficacy of different interventions to foster children’s executive function skills: a series of meta-analyses. Psychol. Bull. 145, 653e697. https://doi.org/10.1037/bul0000195. Tamnes, C., Østby, Y., Walhovd, K., Westlye, L., Due-Tønnessen, P., Fjell, A., 2010. Neuroanatomical correlates of executive functions in children and adolescents: a magnetic resonance imaging (MRI) study of cortical thickness. Neuropsychologia 48, 2496e2508. Tangney, J., Baumeister, R., Boone, A., 2004. High self-control predicts good adjustment, less pathology, better grades, and interpersonal success. J. Personal. Soc. Psychol. 72, 271e324. Tau, G., Peterson, B., 2010. Normal development of brain circuits. Neuropsychopharmacology 35, 147e168. Thomas, M., Johnson, M., 2008. New advances in understanding sensitive periods in brain development. Curr. Dir. Psychol. Sci. 17, 1e5. Thorell, L., Lindqvist, S., Nutley, S., Bohlin, G., Klingberg, T., 2009. Training and transfer effects of executive functions in preschool children. Dev. Sci. 12, 106e113. Valcan, D., Davis, H., Pino-Pasternak, D., 2018. Parental behaviours predicting early childhood executive functions: a meta-analysis. Educ. Psychol. Rev. 30, 607e649. https://doi.org/10.1007/s10648-017-9411-9. Valentine, J., DuBois, D., Cooper, H., 2004. The relation between self-beliefs and academic achievement: a meta-analytic review. Educ. Psychol. 39, 111e133. Veer, I., Luyten, H., Mulder, H., van Tuijl, C., Sleegers, P., 2017. Selective attention relates to the development of executive functions in 2,5- to 3-yearolds: a longitudinal study. Early Child. Res. Q. 41, 84e94. https://doi.org/10.1016/j.ecresq.2017.06.005. Wade, M., Prime, H., Hoffmann, T., Schmidt, L., O’Connor, T., Jenkins, J., 2018. Birth weight interacts with a functional variant of the oxytocin receptor gene (OXTR) to predict executive functioning in children. Dev. Psychopathol. 30, 203e211. https://doi.org/10.1017/S0954579417000578. Wang, Z., Devine, R., Wong, K., Hughes, C., 2016. Theory of mind and executive function during middle childhood across cultures. J. Exp. Child Psychol. 149, 6e22. https://doi.org/10.1016/j.jecp.2015.09.028. Wardell, J., Quilty, L., Hendershot, C., 2016. Impulsivity, working memory, and impaired control over alcohol: a latent variable analysis. Psychol. Addict. Behav. 30, 544e554. https://doi.org/10.1037/adb0000186. Weiss, S., Meltzoff, A.N., Marshall, P., 2018. Neural measures of anticipatory bodily attention in children: relations with executive function. Dev. Cognit. Neurosci. 34, 148e158. https://doi.org/10.1016/j.dcn.2018.08.002. Wellman, H., 2018. Theory of mind: the state of the art. Eur. J. Dev. Psychol. 15, 728e755. https://doi.org/10.1080/17405629.2018.1435413. Wellman, H., Liu, D., 2004. Scaling of theory-of-mind tasks. Child Dev. 75, 523e541. Welsh, M.C., Pennington, B.F., Groisser, D.B., 1991. A normative-developmental study of executive function: a window on prefrontal function in children. Dev. Neuropsychol. 7, 131e149. Wiers, R., Bartholow, B., van den Wildenberg, E., Thush, C., Engels, R., Sher, K., et al., 2007. Automatic and controlled processes and the development of addictive behaviors in adolescents: a review and a model. Pharmacol. Biochem. Behav. 86, 263e283.

Executive function: development, individual differences and clinical insights Chapter | 25

559

Willoughby, M., Blair, C., Kuhn, L., Magnus, B., 2018. The benefits of adding a brief measure of simple reaction time to the assessment of executive function skills in early childhood. J. Exp. Child Psychol. 170, 30e44. Willoughby, M., Kuhn, L., Blair, C., Samek, A., List, J., 2017. The testeretest reliability of the latent construct of executive function depends on whether tasks are represented as formative or reflective indicators. Child Neuropsychol. 23 https://doi.org/10.1080/09297049.2016.1205009. Zelazo, P., Carlson, S., 2012. Hot and cool executive function in childhood and adolescence: development and plasticity. Child Dev. Perspect. 6, 354e360. https://doi.org/10.1111/j.1750-8606.2012.00246.x. Zelazo, P., Muller, U., Frye, D., Marcovitch, S., 2003. The development of executive function: cognitive complexity and controldrevised. Monogr. Soc. Res. Child Dev. 68, 93e119.

Chapter 26

The effects of stress on early brain and behavioral development Amanda N. Noron˜a1, 2, Jenalee R. Doom1, 3, Elysia Poggi Davis1, 4 and Megan R. Gunnar5 1

Department of Psychology, University of Denver, Denver, CO, United States; 2Department of Psychiatry, University of Colorado Anschutz Medical

Campus, Aurora, CO, United States; 3Center for Human Growth and Development, University of Michigan, Ann Arbor, MI, United States; 4

Department of Psychiatry and Human Behavior, University of California Irvine, Irvine, CA, United States; 5Institute of Child Development, University of Minnesota, Minneapolis, MN, United States

Chapter outline 26.1. Introduction 561 26.2. The anatomy and physiology of stress 562 26.3. Prenatal stress and neurobehavioral development 565 26.3.1. Fetal programming 565 26.3.2. Stress regulation and pregnancy 566 26.3.2.1. Changes in the maternal hypothalamice pituitaryeadrenocortical and placental axes over the course of pregnancy 566 26.3.2.2. Fetal adrenal development 567 26.3.2.3. Fetal brain development and susceptibility to stress and stress hormones 567 26.3.3. Gestational stress influences the human fetus 567 26.3.4. Prenatal maternal psychosocial stress and infant and child development 568 26.3.4.1. Socioemotional development 568 26.3.4.2. Hypothalamicepituitaryeadrenocortical axis functioning 569 26.3.4.3. Cognitive development 569 26.3.5. Prenatal maternal biological stress signals and infant and child development 569

26.3.5.1. Social/emotional development 569 26.3.5.2. Hypothalamicepituitaryeadrenocortical axis functioning 570 26.3.5.3. Cognitive development 570 26.3.6. Sex differences 570 26.3.7. Epigenetics 570 26.3.8. Interactions with the postnatal environment 571 26.3.9. Is this fetal programming? 571 26.3.9.1. Summary 571 26.4. Postnatal stress and neurobehavioral development 572 26.4.1. Social regulation of the hypothalamicepituitary eadrenocortical axis and the role of caregivers 573 26.4.2. Early adversity 574 26.4.2.1. Diurnal cortisol following postnatal stress 574 26.4.2.2. Effects of early care on cortisol set points and reactivity 574 26.4.3. Individual differences in sensitivity to experience 575 26.4.4. Summary 577 26.5. Future directions 577 References 578

26.1 Introduction Over a half century of research using animal models has documented the impact of early-life stress on neurobehavioral development (Sanchez et al., 2001). Both stress to the mother or more directly to the fetus during prenatal development and stressors that affect mother and infant during postnatal development impact circuits that are developing during the period of stressor exposure, including the development of stress-mediating systems. Alterations in stress-mediating systems, in turn, influence responses to stressors throughout development, producing cascading effects that can produce significant physical and mental health problems later in life. Research on the neurobiological sequelae of stress during human pre- and postnatal development has a much shorter history. However, inroads are being made in understanding how exposure to stress early in life influences neurobehavioral development and lifelong health (Koss and Gunnar, 2018).

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00026-3 Copyright © 2020 Elsevier Inc. All rights reserved.

561

562 PART | II Cognitive development

Activity of the hypothalamicepituitaryeadrenocortical (HPA) axis, a stress-sensitive neuroendocrine system, has figured prominently in animal studies of early-life stress since the 1950s when it was noted that early experiences permanently altered HPA reactivity and regulation (Levine and Wiener, 1988). Because the HPA axis produces hormones that function as gene transcription factors in numerous organs and tissues and because experience alters its activity as well as the activity of its receptors, research on early-life stress has continued to include a focus on this neuroendocrine system. Attention to activity of this system in studies of human development has been promoted by the availability of assays that allow noninvasive measurement of cortisol, its end hormone, in small samples of saliva (Kirschbaum and Hellhammer, 1989). Consistent with the history of research in this area, we will maintain a focus, though not exclusively, on the HPA axis in reviewing the research on early-life stress and human neurobehavioral development. This chapter consists of four parts: (1) a brief review of the anatomy and physiology of the mature HPA axis and related stress-mediating system; (2) a discussion of prenatal stress and fetal programming; (3) a discussion of the postnatal development of the HPA system, the importance of social regulation of HPA axis in early human development, and what is currently known about long-term impacts of early-life stress on later physical and mental health; and (4) issues that need to be addressed as this field moves forward.

26.2 The anatomy and physiology of stress Stressors are real or perceived threats to psychological or physical viability that are responded to by stressor-specific release of molecules termed stress mediators. These molecules bind to their receptor targets and orchestrate integrated responses that have evolved to increase survival in the immediate face of threat (Joëls and Baram, 2009). Glucocorticoids (cortisol in humans) are steroid hormones that serve as a major mediator of the mammalian stress response. Glucocorticoids are produced by the cortex of the adrenal glands; the medulla of the adrenals produces adrenaline, a hormone that is central to the fight/flight response. Glucocorticoids serve multiple roles in defensive responding (Sapolsky et al., 2000). At basal levels, they are permissive, in the sense that they maintain organs and tissues in a state that permits rapid and sustained mobilization by other neurotransmitters or hormones. At elevated levels, they suppress the actions of other stressmediating systems and, through negative feedback, return the HPA system to basal levels of activity. Via effects on gene transcription, glucocorticoids also can have long-term effects on neural systems mediating perception and response to threat, both up- and downregulating reactions to subsequent stressors. Critically, the effects of acute activations of the HPA system and those of chronic activation are markedly different, with chronic activation, resulting in progressive changes in the expression of stress-mediating genes, alteration in neuronal systems that process signals of threat, and changes in neuronal firing patterns throughout the brain. The cascade of events that produce changes in cortisol release by the adrenals begins with the release of corticotropinreleasing hormone (CRH) and arginine vasopressin (AVP) by cells in the paraventricular nuclei of the hypothalamus (see Fig. 26.1, reviewed in (Gunnar and Vazquez, 2006)). CRH and AVP are released through small blood vessels to the anterior pituitary where they stimulate the release of adrenocorticotropic (ACTH) hormone into the blood stream. Cells on the cortex of the adrenal glands respond to ACTH and start a cascade of enzymatic actions that convert cholesterol to cortisol (corticosterone in rodents). Activation of the adrenal cortex by ACTH also results in production of dehydroepiandrosterone (DHEA), an adrenal androgen, that because of its anabolic effects has antistress properties. Once released into circulation, because of its lipid solubility, cortisol enters the cytoplasm of cells throughout the body and brain where it interacts with its receptors if they are present. Cortisol has affinity to two receptors: mineralocorticoid receptors (MRs) and glucocorticoid receptors (GRs). Its affinity with MRs is many times greater than to GRs; hence, if both are present, MRs will be occupied and activated first, followed by GRs. In most areas of the body, however, cortisol cannot access MRs because an enzyme is present (11-beta hydroxysteroid dehydrogenase or 11b-HSD) that converts cortisol to a form with low MR affinity. As will be discussed, this enzyme is also present in the placenta where it serves to regulate impacts of maternal cortisol on the placenta and fetus. In the brain, however, the enzyme is not present, allowing the levels of cortisol in circulation to determine the balance between MR and GR activation. Under basal levels, MRs tend to be almost wholly occupied; while when cortisol rises to stress levels and also at the peak of the diurnal rhythm, GRs become occupied as well. MRs tend to mediate many of the permissive effects of cortisol; whereas GRs mediate many of the more catabolic stress effects. GRs are also involved in negative feedback of the axis, functioning at the level of the pituitary, hypothalamus, hippocampus, and likely also the medial frontal cortex, to contain the HPA response and help return the axis to basal levels of activity. While there is increasing evidence that, under conditions of stress, rapid cell membrane-mediated effects of cortisol occur, most effects of cortisol involve translocation of the cortisol-receptor complex from the cytoplasm to the cell nucleus where cortisol interacts with glucocorticoid response elements (GREs) in the promotor region of many genes. Activation of

The effects of stress on early brain and behavioral development Chapter | 26

563

FIGURE 26.1 The HPA system. Panel A depicts the anatomy of the HPA system and structures important in its regulation. PFCtx ¼ prefrontal cortex, AMY ¼ amygdala, HYP ¼ hypothalamus, HC ¼ hippocampus, NTS ¼ nucleus of the tractus solitarius. Panel B depicts the activation (þ) and negative feedback inhibition () pathways of the HPA system. Increases in GCs are initiated by the release of CRH/AVP from the medial parvocellular region of the paraventricular nucleus (mpPVN) in the hypothalamus. Negative feedback inhibition operates through GCs acting at the level of the pituitary, hypothalamus and hippocampus. GABA, gamma animobutyric acid; CRH, corticotropin releasing hormone; AVP, arginine vasopressin; ACTH, adrenocorticotropic hormone. Used with permission from Gunnar and Vazquez (2006).

GREs increases or decreases gene transcription in interaction with other gene transcription factors. When cortisol operates as a gene transcription factor, its effects on organs and tissues take minutes to hours to be produced. This means that while acute threat may stimulate increases in cortisol production, cortisol itself is not a major factor in fight/flight responses that proceed on the basis of seconds to minutes. Activation of the HPA axis is regulated by complex signals derived across a number of pathways that carry information about the state of the internal and external milieu (see Fig. 26.2). Activation of the system in response to threats to internal homeostasis (e.g., blood volume loss) travels to the CRHproducing cells in the hypothalamus through brainstem pathways. Psychological threats that require integration of information about external events are mediated through pathways involving the amygdala and bed nucleus of the stria terminalis (BNST). Notably, neural systems involved in activating the axis in response to psychosocial threats either produce CRH or have receptors for CRH. The central nucleus of the amygdala is one region rich in CRH-producing neurons, activation of which plays a role not only on activating the HPA axis but also in stimulating increases in central norepinephrine and peripheral activation of the sympathetic nervous system. The extrahypothalamic CRH system is part of the fight/flight system and a key orchestrator of fear behavior (McCall et al., 2015). The expression of CRH and its

564 PART | II Cognitive development

FIGURE 26.2 Schematic representation of the activating (right side) and inhibiting (left side) circuits that contribute to regulation of the HPA system. Catecholamines, norepinephrine (NE) and epinephrine (E) arising from medullary nuclei of the brainstem are the primary neurotransmitters providing activation of CRH synthesis and release from the mpPVN. Serotonin originating from dorsal raphe is weakly activating; it act both directly on mpPVN and indirectly through excitatory glutamate neurons or inhibitory gamma animobutyric acid (GABA) inputs. Paradoxically, the inhibitory GABA neurotransmitter activates the mpPVN to secrete CRH since two GABA neurons activated in series leads to excitation and not inhibition. Extra-hypothalamic CRH also acts as a neurotransmitter to initiate autonomic and behavioral responses to stress. The activation of the extra-hypothalamic CRH system is initiated by rising glucocorticoids levels that operate on the amygdala to secrete CRH that, in turn, impacts on the locus cerouleus (LC). Through the activation of catecholaminergic brainstem nuclei there is also stimulation of descending pathways leading to NE/E release form the adrenal medulla that facilitates cardiovascular autonomic responses to stress. Inhibition of the LHPA axis seen in the left side is provided by glucocorticoids acting on glucocorticoids receptors (GR) in the hypothalamus and pituitary where CRH and ACTH release is halted. The hippocampus serves to inhibit the stress response via multiple circuits, some of which are direct inhibitory GABA inputs; others are indirect through glutamate excitatory inputs to GABA neurons converging in the mpPVN. GABA neurons located in each of the structures further modify the stress reactivity and inhibition from other brain regions such as the thalamus, association cortex, cortical and limbic afferents. * Glucocorticoids provide positive stimulation to the amygdala for the synthesis and release of CRH, but negative to the pituitary, hypothalamus and hippocampus. y interaction is through glutamate outflow from these regions that synapse on local GABAergic neurons, producing inhibition of mpPVN. Used by permission from Gunnar and Vazquez (2006).

receptors in the various brain regions involved in emotion and cognition is age dependent and regulated by stress throughout the life span. Recent evidence indicates that effects of stress on neurodevelopment are mediated by CRH, as well as cortisol (Korosi and Baram, 2008). The activity of the HPA axis is closely related to the sympathetic and parasympathetic nervous systems. The sympathetic nervous system governs the fight-flight response, which is a critical component of our response to threat.

The effects of stress on early brain and behavioral development Chapter | 26

565

Immediately following the perception of threat, preganglionic sympathetic nerves are activated in the spinal cord, which then leads directly to organs involved in responding to stress. The most important of these organs is the adrenal medulla, which produces epinephrine that travels throughout the body to influence cardiovascular function and energy metabolism (Ulrich-Lai and Herman, 2009). In concert with norepinephrine release from nerve terminals, these neurotransmitters coordinate responses to stress in the periphery, although norepinephrine is also produced in the central nervous system. The sympatheticeadrenal medullary (SAM) system is a component of the sympathetic nervous system involving hormones, and it supports arousal and attention in addition to stress responses. Thus, the SAM system will respond to effortful tasks that do not involve threats to the physical or social self, even though the sympathetic nervous system and HPA axis do not respond. The parasympathetic nervous system (PNS) helps to restore the body to baseline following a stressor, assisting with rest and repair (Porges, 2009). Activation of the PNS is more targeted, with projection between the body and brain to relay information about the body’s state. The nucleus tractus solitarius relays these inputs to the amygdala and other neural systems that detect and respond to threat. One of the main effectors of the PNS is the vagus nerve, which provides tonic inhibitory input to the heart. In children performing negative emotional tasks, primarily, changes in vagal tone are responsible for changes in heart rate. Notably, the PNS provides important input not only to the amygdala, as noted, but also to the HPA axis as this is the primary mode through which systemic (physical) as opposed to psychological stressors impact the regulation of CRH, ACTH, and thus cortisol (Herman, 2017).

26.3 Prenatal stress and neurobehavioral development 26.3.1 Fetal programming Fetal development proceeds at a more rapid pace than any later developmental stage (Barker, 1998). For this reason, the human fetus is particularly vulnerable to both organizing and disorganizing influences, which have been described as programming. Programming is the process by which a stimulus or insult during a vulnerable developmental period has a long-lasting or permanent effect (Barker, 1998; Kuzawa, 2005). The trajectory of fetal development adjusts, in response to cues in utero, to optimize adaptability to the environment and ultimately survival of the organism. Such adaptations, however, can have long-term negative consequences postnatally, particularly if there is significant mismatch between prenatal and postnatal environments (Sandman et al., 2012). The effects of programming are dependent on the timing (i.e., the developmental stage of organ systems and the changes in maternal and placental physiology) and the duration of exposure (Davis and Sandman, 2010; Nathanielsz, 1999), as well as the dosage of prenatal stress (DiPietro et al., 2006). Programming also has differential effects by fetal sex (Glover and Hill, 2012) and genotype (Buss et al., 2012a,b). There is convincing support for fetal programming of adult health outcomes, including heart disease, diabetes, and obesity; however, the evidence comes primarily from studies that rely on birth phenotype (e.g., small size at birth or preterm delivery) as an index of fetal development (e.g., Barker, 1998). It is unlikely that birth phenotype alone is the cause of subsequent health outcomes. Birth phenotype, instead, likely is a marker of fetal adaptations that shape the structure and function of those physiological systems that are part of the causal pathway to health outcomes (Gluckman and Hanson, 2004). Prenatal exposure to maternal stress signals is one of the primary pathways for programming of later health and development. The HPA axis participates in a surveillance and response system for stress that is present in many species, from the desert-dwelling Western Spadefoot tadpole to the human fetus, and allows for the detection of threat so that development can be adjusted accordingly. For instance, rapidly evaporating pools of desert water result in elevation of CRH in the tadpole, accelerating metamorphosis and increasing the likelihood of survival. If the CRH response is blocked during environmental desiccation, then development is not accelerated and the tadpole’s survival is compromised. There are, however, long-term consequences for the tadpole that survives this early-life stress because its growth is stunted, and it is at a disadvantage in the competition for food and reproduction (Denver, 1997). It has been argued that a similar signaling pathway participates in the regulation of human fetal development. Detection by the fetal/placental unit of stress signals from the maternal environment (e.g., cortisol) informs the fetus that there may be a threat to survival. This information may prime or advance the placental clock (McLean et al., 1995) by activating the promoter region of the CRH gene and increasing the placental synthesis of CRH (Sandman et al., 2006). Trajectories of CRH levels affect fetal growth and development; early increases or rapidly rising levels of placental CRH lead to shorter gestation and preterm birth (McGrath et al., 2002). Early departure from an inhospitable host environment may be essential for survival, but it also may have long-term consequences for the human fetus just as it does for the tadpole. The developmental trajectory of the fetus, whether born early or at term, is influenced by the maternal environment, and

566 PART | II Cognitive development

adaptation of the developmental program to maternal stress signals may prepare the fetus for postnatal survival. The goal of the second section is to discuss the various pathways through which prenatal maternal stress signals may prepare the fetus for adaptation to the postnatal world.

26.3.2 Stress regulation and pregnancy 26.3.2.1 Changes in the maternal hypothalamicepituitaryeadrenocortical and placental axes over the course of pregnancy Regulation of the HPA axis is altered dramatically during pregnancy. The maternal pituitary gland doubles in size, and the production of maternal HPA axis hormones increases several-fold. The growth and development of the placenta is primarily responsible for the profound changes in the HPA axis throughout gestation. The placenta expresses the genes for CRH (hCRHmRNA) and the precursor for ACTH and b-endorphin (proopiomelanocortin) (see Fig. 26.3). All of these stress-responsive hormones increase as pregnancy advances, but the exponential increase in placental CRH (pCRH) in maternal plasma is especially dramatic, reaching levels observed only in the hypothalamic portal system during physiological stress. Placental CRH (pCRH) is identical to hypothalamic CRH in structure, immunoreactivity, and bioactivity. There is, however, one crucial difference in its regulation. In contrast to the negative feedback regulation of hypothalamic CRH, cortisol stimulates the expression of hCRHmRNA in the placenta, establishing a positive feedback loop that allows for the simultaneous increase of pCRH, ACTH, and cortisol over the course of gestation (see for review (Sandman and Davis, 2010)). The normative increase in stress-responsive hormones such as cortisol and pCRH plays an important role in the regulation of pregnancy as well as facilitating maturation of the fetus. However, because of the positive feedback between cortisol and pCRH, the effects of maternal stress on the fetus may be amplified with potentially negative consequences for the developing fetus. The effects of CRH and cortisol are modulated by the activities of binding proteins and enzymes, including CRHbinding protein, cortisol-binding globulin (CBG), and 11b-HSD2. Levels of these proteins and enzymes increase progressively with advancing gestation, to protect the fetus from overexposure to cortisol. Prior to term, levels decline to ensure maturation of the fetal lungs, CNS, and other organ systems in full-term births (Murphy and Clifton, 2003). Levels of binding protein have been associated with birth outcome (Hobel et al., 1999), and variations in CBG may contribute to individual differences in developmental outcomes because levels have been shown to be lower in women with growthrestricted fetuses (Ho et al., 2007). In addition, despite the presence of 11b-HSD2 early in gestation, maternal cortisol

FIGURE 26.3 The regulation of the maternal HPA axis changes dramatically over the course of gestation with profound implications for the mother and the fetus. One of the most significant changes during pregnancy is the development of the placenta, a fetal organ with significant endocrine properties. In non-pregnant women, exposure to stress activates a cascade of events including the release of CRH, ACTH and cortisol. This stress system is regulated by a negative feedback loop in which cortisol “turns off” the HPA axis. During pregnancy CRH is released from the placenta into both the maternal and fetal compartments. In contrast to the negative feedback regulation of hypothalamic CRH, cortisol increases the production of CRH from theplacenta. Placental CRH (pCRH) concentrations rise exponentially over the course of gestation. Because of the positive feedback between cortisol and pCRH the effects of maternal stress on the fetus may be amplified representing one pathway by which stress may exert influences on the fetus. In addition to its effects on pCRH, maternal cortisol passes through the placenta. However, the effects of maternal cortisol on the fetus are modulated by the presence of a placental enzyme 11bHSD2 which oxidizes it into an inactive form, cortisone. Activity of this enzyme increases as pregnancy advances, and then drops precipitously so that maternal cortisol is available to promote maturation of the fetal lungs, central nervous system as well as other organ systems. Structures of the HPA axis begin their development early in gestation and become increasingly functional with the progression towards term. See text for description.

The effects of stress on early brain and behavioral development Chapter | 26

567

does reach the fetus, and the amount varies with circulating maternal levels (Gitau et al., 1998). Maternal stress downregulates 11b-HSD2 activity in the placenta allowing a greater proportion of maternal cortisol to cross the placenta to reach the fetus (Mairesse et al., 2007), impacting fetal growth and development (Baibazarova et al., 2013; Reinisch et al., 1978). This is another mechanism through which the consequences of maternal stress for the developing fetus may be amplified. Because of the timetable of fetal development and the changes in maternal and placental physiology, the consequences of stress exposures will vary based on the gestational period of exposure. Although findings regarding the nature of prenatal influence on postnatal functioning are well established in animal models, it is critical to acknowledge that the differences in reproductive and stress physiology, even in very closely related species such as humans and nonhuman primates, limit the validity of generalizing from animal models (Smith and Nicholson, 2007). For these reasons, this chapter primarily focuses on studies of gestational stress in humans.

26.3.2.2 Fetal adrenal development The fetal adrenals make unique contributions to both the regulation of fetal development and the timing of parturition. Cortisol is thought to play critical roles in the promotion of fetal maturation in preparation for extrauterine life. Furthermore, dehydroepiandrosterone sulfate (DHEAS) produced by the fetal adrenal is an obligate precursor for placental estrogen and is also thought to contribute to the initiation of parturition. Morphologically, the fetal adrenal gland comprises two zones: the outer, definitive zone, and the large, inner fetal zone. Between these two zones is the transitional zone. The fetal and definitive zones can be recognized after the eighth gestational week. The fetal adrenals grow rapidly until the third trimester so that at term the fetal adrenals are significantly larger, relative to body weight, than the adult adrenals. At the end of human pregnancy, the fetal zone begins to atrophy. The human fetal adrenal has steroidogenic enzymes as early as the seventh gestational week, and cortisol secretion can be detected as early as eighth week. Cortisol production from the fetal adrenal is regulated by ACTH and ACTH-containing cells can be seen in the pituitary by eight gestational weeks (Jaffe et al., 1998; Kempná and Flück, 2008). There is evidence that the fetus responds to pain with an increase in cortisol during the latter half of gestation (Gitau et al., 2001).

26.3.2.3 Fetal brain development and susceptibility to stress and stress hormones The rapid changes in the developing fetal brain render it particularly susceptible to influences of stress-responsive hormones such as cortisol and pCRH. Neurodevelopmental processes including neurogenesis, migration, neuronal differentiation, dendritic arborization, axonal elongation, synapse formation and collateralization, and myelination proceed at an exceptionally rapid pace throughout the fetal period (Bourgeois, 1997; Sidman and Rakic, 1973). We specifically discuss regions of the brain that are both integral to the regulation of stress responses and vulnerable to exposure to stress hormones, including the hippocampus and amygdala. Both are identifiable between 6 and 8 gestational weeks, and by term, the basic neuroanatomical architecture of these regions is present. Limited information exists regarding the time course of prenatal development of cortisol receptors in humans. There is evidence that both types of cortisol receptors are present in the human hippocampus by 24 gestational weeks (Noorlander et al., 2006). Exposure to prenatal maternal biological and psychosocial stress influences the developing fetal brain and endocrine systems producing long-term effects on cognition, emotion, and physiology in the offspring (Kapoor et al., 2008). Evidence for persistent organizational changes or programming influences on the nervous system has been growing and may include changes in neurotransmitter levels, cell growth and survival, and adult neurogenesis. For instance, high concentrations of glucocorticoids (e.g., cortisol) and CRH may inhibit growth and differentiation of the developing nervous system. In a human study of over 400 pregnant women, maternal gestational cortisol secretion was found to be inversely related to ultrasound measures of fetal brain size in early, middle, and late pregnancy (Li et al., 2012). Prenatally, elevated prenatal placental CRH predicted cortical thinning in childhood (Sandman et al., 2018). Experimental animal studies similarly show that high levels of CRH cause dendritic atrophy in cortical neurons (Curran et al., 2017). Lastly, glucocorticoids may be especially neurotoxic to hippocampal CA3 pyramidal cells; fetal exposure to high levels of glucocorticoids produces irreversible damage to the hippocampus (Liston and Gan, 2011). These data from human and animal models suggest endocrine and brain mechanisms by which early-life stress may provoke long-term effects on stress, emotional regulation, and cognition (see Joëls and Baram, 2009; Seckl, 2007 for reviews).

26.3.3 Gestational stress influences the human fetus In humans, a compelling body of work has documented that both maternal report of elevated maternal stress or anxiety and exposure to traumatic events during pregnancy are associated with increased risk for preterm birth. Sociodemographic

568 PART | II Cognitive development

stressors, including low socioeconomic status, experiences of discrimination, and residing in unsafe neighborhoods, increase risk for preterm birth (Giurgescu et al., 2012). However, studies that have controlled for these key sociodemographic factors as well as obstetric risk factors continue to find a correlation between maternal stress and preterm birth, suggesting that psychological functioning over and above contextual stress affects birth outcomes. There are several pathways by which maternal stress may lead to preterm birth including accelerated production of pCRH and altered vascular and immune functioning (Dunkel-Schetter and Glynn, 2011). Preterm birth is associated with pervasive developmental delays (Aarnoudse-Moens et al., 2009); however, intrauterine exposures, including stress, likely contribute to these impairments independently from birth outcomes. Furthermore, fetal brain development differs between fetuses who go on to be delivered preterm and their counterparts who go on to be delivered at term, suggesting that the neurodevelopmental correlates of preterm birth begin in utero (Thomason et al., 2017). The study of human fetal development is important because it provides a direct test of the fetal programming hypothesis with the opportunity to assess the effects of gestational stress on development before the effects of external forces, such as birth outcome, parenting, and socialization, are exerted. More direct tests of the effect of maternal stress on the fetus come from studies manipulating maternal stress and evaluating the consequences for fetal behavior and studies measuring fetal responses to external stimulation. Fetuses display a consistent response profile (e.g., suppression of motor activity) during maternal exposures to moderate laboratory challenges such as the Stroop color-word test or viewing videos of labor and delivery (DiPietro, 2004). The nature of these responses appears to be moderated by maternal psychological state (Kinsella and Monk, 2009). Direct measures of fetal responses to external stimulation provide an index of fetal nervous system development and have been used to assess the developmental consequences of exposure to physical or maternal psychological stress. The response to a vibroacoustic stimulus (VAS) is an indication of fetal maturity reflecting maturation and integrity of neural pathways through the cerebral cortex, midbrain, brainstem, vagus nerve, and the cardiac conduction system. Using the fetal response to VAS, it has been shown that stress signals, most clearly pCRH trajectories, influence the developing fetal nervous system. Low pCRH is associated with more mature or earlier development of the fetus’ ability to mount a response to the VAS and with a more mature profile to a classic habituation/dishabituation paradigm (Sandman et al., 1999). Other maternal stress signals including overexpression of b-endorphin and underexpression of ACTH have additionally been linked to the fetal response to VAS (Sandman et al., 2003). Fetal responses may be linked to brain development, as one study found that fetal heart rate variability was correlated with connectivity between the dACC and mPFC (Spann et al., 2018). These studies raise the possibility that repeated exposures over the course of gestation may influence the developing fetal nervous system. They also provide evidence that signals of maternal stress during gestation exert programming influences on the nervous system that cannot be explained by postnatal experiences. Continuity between the fetal and infant periods in assessments of movement and heart rate indicates that maternal influences that shape developmental trajectories during the prenatal period will continue to influence functioning postnatally (DiPietro, 2004).

26.3.4 Prenatal maternal psychosocial stress and infant and child development 26.3.4.1 Socioemotional development Prenatal exposure to elevated levels of maternal psychosocial stress is associated with behavioral and emotional disturbances during infancy and childhood among healthy full-term infants who are independent of postpartum maternal psychosocial stress (Bergman et al., 2007; Davis et al., 2004, 2007). Both maternal report of psychosocial stress and stressful life events are associated with more fearful and reactive behaviors during infancy and toddlerhood. Effects on social and emotional development continue to be observed during childhood and adolescence. Maternal antenatal stress, anxiety, and depression predict childhood behavioral and emotional problems, including attention-deficit/hyperactivity disorder (ADHD) and both internalizing and externalizing problems, after controlling for birth outcomes and postnatal maternal psychological state (Barker, 1998; Bergman et al., 2007; Van den Bergh and Marcoen, 2004). Recent evidence suggests that lack of predictability in maternal mood additionally predicts internalizing problems in childhood beyond the effects of level of maternal distress (Glynn et al., 2018). Maternal distress may impact the development of limbic and prefrontal brain regions, and changes in these regions may be a mechanism underlying the observed behavioral findings. Recent studies have shown that elevated prenatal maternal distress predicts microstructure of the right and left amygdala (lower fractional anisotropy and axial diffusivity; Rifkin-Graboi et al., 2013) as well as connectivity between prefrontal and limbic regions in the newborn (Posner et al., 2016). Prenatal maternal distress further predicts elevated amygdala response to negative emotional stimuli and cortical thinning, primarily in the right frontal cortex during childhood (Lebel et al., 2016; Sandman et al., 2015). Thus, maternal distress shapes the trajectory of brain development, and developing prefrontal and limbic regions may be particularly susceptible.

The effects of stress on early brain and behavioral development Chapter | 26

569

26.3.4.2 Hypothalamicepituitaryeadrenocortical axis functioning Alterations to the fetal HPA axis are frequently proposed as the primary biological pathway underlying fetal programming of later health and development. Animal studies suggest that the fetal HPA axis may be particularly vulnerable to prenatal exposure to maternal stress (Kapoor et al., 2006); however, relatively little is known about the consequences of prenatal maternal psychosocial stress for HPA axis functioning among humans. One study found that infants of women with higher economic strain during pregnancy had elevated cortisol response to vaccination (Thayer and Kuzawa, 2014). Prenatal psychosocial stress is also associated with altered circadian regulation during childhood and adolescence on typical days (O’Connor et al., 2005; Van den Bergh et al., 2008). Prenatal exposure to maternal psychosocial stress may influence the developing fetal HPA axis in ways that implicate the regulation of cortisol production during infancy, childhood, and adolescence.

26.3.4.3 Cognitive development The influence of gestational exposure to maternal psychosocial stress on cognitive and motor development is less clear. There is evidence that maternal self-report of elevated stress and anxiety, as well as exposure to traumatic life events, such as severe ice storms, during pregnancy is associated with delayed infant and child cognitive, language, and neuromotor development and that these deficits may persist into adolescence. However, a recent metaanalysis found only a small relation (r ¼ 0.05) between maternal stress and child cognitive development as assessed using the Bayley Scales or general measures of IQ in children under 5 years (Tarabulsy et al., 2014). The impact of prenatal stress on cognitive outcome may vary based on the aspect of functioning considering. For example, it may be the case that executive functions are particularly susceptible to the impact of gestational stress (Buss et al., 2011). It is also plausible, however, that generalized self-report measures of psychological distress do not adequately characterize stress that is unique during pregnancy. As reviewed in Davis and Sandman (2010), measures of pregnancy-specific stress (e.g., “I am fearful regarding the health of my baby,” “I am concerned or worried about losing my baby”) are better than measures of generalized psychological distress for predicting neurodevelopmental outcomes. It is important to note that these associations are not explained by actual medical risk associated with pregnancy and birth outcome. Furthermore, pregnancy-specific stress is linked to trajectories of gestational stress hormones (Kane et al., 2014). Support for the importance of pregnancy-specific stress for developmental outcomes comes from a recent study documenting associations between elevated pregnancy-specific anxiety and decreased gray matter density in the prefrontal cortex, the premotor cortex, the medial temporal lobe, the lateral temporal cortex, the postcentral gyrus as well as the cerebellum extending to the middle occipital gyrus and the fusiform gyrus at 6e10 years of age (Davis and Sandman, 2010). These brain regions are associated with a variety of cognitive processes including reasoning, planning, attention, memory, and language and raise the possibility that developmental alterations to these regions may underlie associations between elevated pregnancyspecific anxiety and cognitive performance observed in prior studies (Buss et al., 2010).

26.3.5 Prenatal maternal biological stress signals and infant and child development Given the accumulating evidence that prenatal psychosocial stress affects the developing offspring, evaluation of the underlying biological mechanisms will elucidate critical mediators of this relationship. Alterations to the maternal HPA and placental axis are most frequently cited as the mechanism that underlies fetal programming of later health and developmental outcomes.

26.3.5.1 Social/emotional development Prenatal exposure to elevated maternal cortisol and placental CRH predicts increased fussiness, negative behavior, and fearfulness during infancy (Davis et al., 2005, 2007; de Weerth et al., 2003) and toddlerhood (Bergman et al., 2007). Furthermore, both maternal cortisol and placental CRH predict childhood anxiety symptoms (Davis and Sandman, 2012; Howland et al., 2016). Prenatal stress hormones may impact development of internalizing problems via alterations to development of limbic regions. Prenatal maternal cortisol predicts greater right amygdala volume among girls (Buss et al., 2012a,b), which mediated the relation between prenatal maternal cortisol and child anxiety. Recent work indicates that prenatal stress hormones may also shape the connectivity of the human brain. Fetal exposure to elevated maternal cortisol predicts network segregation in girls. Specifically, girls generated more connections than boys to maintain topologically capable and efficient neural circuits, and this increase in neural cost was associated with higher levels of internalizing problems (Kim et al., 2016).

570 PART | II Cognitive development

26.3.5.2 Hypothalamicepituitaryeadrenocortical axis functioning Prenatal elevations in maternal cortisol production may impact the development of the fetal HPA axis with consequences for postnatal HPA axis regulation. Higher levels of prenatal cortisol have been associated with elevated cortisol reactivity following a blood draw in infants (Davis et al., 2011a,b) and with higher levels of cortisol on the day of inoculation and the first day of school in preschoolers (Gutteling et al., 2005). Amniotic fluid cortisol also predicted infant cortisol response patterns to separationereunion stress at age 17 months, after controlling for key covariates, including prenatal and obstetric factors and parentechild attachment (O’Connor et al., 2013). These studies provide evidence for effects of prenatal maternal cortisol on HPA axis functioning although interpretation is limited by modest sample sizes and varied methodologies (e.g., type of stressor, mode and timing of maternal cortisol measurement). Complementary lines of evidence from animal research (Henry et al., 1994; Mustoe et al., 2014) and evaluation of prenatal synthetic glucocorticoid administration (Davis et al., 2011a,b) substantiate the likely role of prenatal cortisol in HPA axis development.

26.3.5.3 Cognitive development Evidence suggests that the trajectory of maternal cortisol across gestation is a strong predictor of child neurodevelopment, and the effects likely vary by timing of the exposure. Elevated maternal cortisol during early and midgestation has been associated with decreased neuromuscular maturity in the newborn (Ellman et al., 2008) and delayed cognitive development during toddlerhood (Bergman et al., 2010). Conversely, elevated maternal cortisol late in gestation has been associated with significantly higher scores on measures of mental development at 1 year (Davis and Sandman, 2010) and during middle childhood (Davis et al., 2017), as well as greater cortical thickness during childhood (Davis et al., 2017). These findings linking cortisol exposure in late gestation to neurodevelopment are remarkably consistent with its function in the maturation of the human fetus. As pregnancy advances toward term, exposure to cortisol is necessary and beneficial for fetal maturation, and exposure to increased cortisol is facilitated by the sharp drop in 11b-HSD2 activity allowing a greater proportion of maternal cortisol to cross the placental barrier (Murphy and Clifton, 2003). The beneficial effects of modestly elevated cortisol during late gestation are consistent with animal models demonstrating that modest cortisol increases during the early postnatal period are associated with persisting beneficial effects for the developing brain (Catalani et al., 2000).

26.3.6 Sex differences There is clear evidence that the male and female fetuses respond differentially to prenatal stress. Clifton and colleagues have shown that male fetuses respond to an adverse maternal environment with minimal placental adaptation, to ensure increased or continued growth. Female fetuses, on the other hand, respond with multiple changes involving placental genes and proteins to ensure survival in the case of additional adverse events in utero (Clifton, 2010). These differences in fetal adaptation may explain why male fetuses are more vulnerable to threats to morbidity and mortality (DiPietro and Voegtline, 2017). In contrast, the more subtle adaptations made by female fetuses in response to stress may underlie sexspecific vulnerabilities to subsequent psychopathology (Sandman et al., 2013). For example, females exposed to prenatal stress are more likely to show elevations in symptoms of anxiety and depression and have heightened stress responses (Davis and Pfaff, 2014; Sandman et al., 2013), whereas males are more likely to show deficits in memory and learning and exhibit increased aggression (Glover and Hill, 2012).

26.3.7 Epigenetics Researchers are in the early stages of examining specific biological pathways through which maternal stress produces changes in offspring development. Because the placenta regulates the prenatal environment, including cortisol levels, an important area of study involves effects of maternal stress on the expression of genes in the placental glucocorticoid pathway. Prenatal anxiety and depression predict the downregulation of placental 11b-HSD2 (O’Donnell et al., 2012). Experimental work with rats has supported the hypothesis that maternal stress altered HSD11B2 gene expression and fetal brain development (Peña et al., 2012). Human studies have similarly found that prenatal stress was associated with the methylation of glucocorticoid pathway genes in the placenta (i.e., HSD11B2, NR3C1, and FKBP5). Maternal stress and methylation in the placenta have been correlated with fetal CNS development (Monk et al., 2016) and infant neurodevelopment and behavior (Conradt et al., 2013). Methylation of genes related to HPA axis functioning is also impacted by prenatal stress. In a sample of 24 mothereinfant dyads residing in the conflict-ridden region of the Democratic Republic of Congo, Kertes et al. (2016)

The effects of stress on early brain and behavioral development Chapter | 26

571

found that chronic stress and war trauma significantly affected the methylation of HPA axis genes, particularly at transcription factor binding sites. Furthermore, gene methylation levels were related to birth weight outcomes, collectively explaining over half of birth weight’s variance. Maternal prenatal stress (i.e., material deprivation, daily psychosocial stressors, war stress) was also related to newborn methylation of the glucocorticoid receptor NR3C1 (Mulligan et al., 2012). These findings provide promising preliminary support for possible biological mechanisms underlying the link between maternal stress and infant HPA functioning and highlight the need for further research to understand the nature of this relation.

26.3.8 Interactions with the postnatal environment The next section of this chapter will extensively cover postnatal influences on brain and behavioral development; however, it is important to note that prenatal and postnatal influences likely interact with and/or build upon each other. It is unlikely that the programming window closes at birth (O’Connor et al., 2014). Interestingly, breastmilk may serve as one pathway through which mothers continue to shape the development of their children, as biologically active components in breastmilk, including maternal cortisol, have been associated with infant behavior (Gray et al., 2013). Adverse childhood experiences, when paired with a history of prenatal stress exposure, may exacerbate negative developmental outcomes. Pawlby et al. (2011) found that children exposed to prenatal stress and childhood maltreatment were at almost 12 times greater risk of developing psychopathology compared with children exposed to either prenatal stress or childhood maltreatment. On the other hand, positive postnatal experiences may buffer the negative impact of prenatal stress. Rat pups exposed to prenatal stress have improved outcomes when postnatal rearing is enriched, such as when they are cross-fostered to unstressed mothers (Del Cerro et al., 2010; Maccari et al., 1995) or when subjected to “neonatal handling” (removed from mother and placed in an individual compartment for a brief period of time; Wakshlak and Marta, 1990). In humans, high levels of positive parenting act as a buffer against the negative effects of prenatal stress on cognitive functioning in early childhood (Schechter et al., 2017). Similarly, maternal sensitivity quells infant distress, and this association is strengthened for infants exposed to maternal prenatal anxiety (Grant et al., 2010).

26.3.9 Is this fetal programming? One concern that challenges research with humans is whether associations between maternal stress and fetal outcomes should be interpreted as fetal programming or alternatively as a reflection of shared genetic factors or as continuity between the prenatal and postnatal environment. In the studies of naturally occurring variations in maternal cortisol or maternal selfreported stress, it is difficult to differentiate between these alternative explanations. The programming findings reported here, however, are consistent with animal models where random assignment is possible (Kapoor et al., 2006), with human studies of randomly occurring traumatic events, such as natural disasters (Laplante et al., 2004; Yehuda et al., 2005), and with prenatal exposure to synthetic glucocorticoids (Davis et al., 2006; French et al., 1999). Furthermore, in studies involving measurement of prenatal and postnatal stress, prenatal effects generally hold when postnatal effects are statistically controlled (O’Connor et al., 2002; Van den Bergh and Marcoen, 2004). More convincingly, similar effects of prenatal stress on child outcomes were documented among children conceived by in vitro fertilization, and thus in a model where mother and fetus were genetically unrelated (Rice et al., 2010). Thus, while in most human studies of prenatal stress, genetic and postnatal mechanisms cannot be ruled out as a possible explanation, there is reasonable evidence to warrant the conclusion that maternal stress has effects on the neurodevelopment of her fetus. Future research can examine the impact of experimental manipulations of prenatal stress using a randomized controlled design (Davis et al., 2018).

26.3.9.1 Summary Both psychosocial and biological maternal stress signals are associated with developmental consequences for the fetus. Furthermore, these effects cannot be accounted for by birth outcome or postnatal maternal psychological distress. However, it is important to acknowledge that although there is some evidence that psychosocial and biological stress signals converge (Hoffman et al., 2016), others have suggested that these signals may not be correlated during pregnancy and may exert independent influences on developmental outcomes (Sandman and Davis, 2010). Future research will have to examine vascular or immune pathways that could be mechanisms by which increases in maternal psychosocial stress might also affect the fetus.

572 PART | II Cognitive development

The studies discussed in this section of the chapter emphasize the importance of performing prospective longitudinal studies to evaluate the trajectory of maternal stress signals across gestation and its association with infant and child developmental outcomes. Data indicate that the trajectory or profile of biological and psychosocial stress signals may be more critical for determining developmental outcomes as compared with level at a given gestational interval (Davis and Sandman, 2010; Glynn et al., 2008). Both severity and timing of exposure, as well as the sex of the fetus, must be considered to evaluate associations between prenatal stress measures and infant outcomes. The adaptive significance of these associations is yet to be determined and requires long-term follow-up evaluating the interaction between the prenatal and the postnatal environment.

26.4 Postnatal stress and neurobehavioral development The HPA axis is not yet mature at birth (Gunnar and Herrera, 2013). The fetal zone of the adrenal cortex involutes over the first 6 postnatal months. As it involutes, the structure of the mature adrenal cortex becomes more distinct. Cortisol-binding globulin (CBG) is low in the newborn but gradually increases during the first postnatal months. Thus, even small increases in plasma cortisol in response to stress may result in large increases of biologically active unbound cortisol as CBGbinding sites become filled. Adrenal sensitivity to ACTH also appears to decrease over these first postnatal months as sensitivity is higher in the first 4 months than later in development. For the first few months of postnatal life, there is higher HPA reactivity to stimuli than in later development, which is consistent with evidence than mild stimuli such as being undressed, or undergoing a physical exam provokes elevations in cortisol in the first 3e4 postnatal months but not later. Diurnal cortisol patterning changes dramatically in the first few years of life. Basal cortisol levels are not related to time-of-day at birth; rather cortisol is associated with behavioral arousal. As early as 6 weeks postnatal, a morning cortisol peak and a nadir in the evening can be observed. However, there is a great deal of variability in the diurnal rhythm at this age. The morning peak and evening nadir become more distinct across the next few months, but the more mature adultlike pattern with consistent cortisol decreases from midmorning to late afternoon is not observed until around age 4 years, when children give up their daily naps. Napping in young children has been associated with decreases in cortisol during sleep and increases in cortisol to prenap levels around 30e45 min after awakening. The cortisol awakening response (CAR), or the increase in cortisol levels soon after awakening that precedes decreases in cortisol throughout the day, can be detected already in young infants and is reliably detected through adulthood (Bäumler et al., 2013). These developmental considerations make measurement of the HPA axis more difficult in infants and young children. There are few developmental changes in cortisol reactivity and regulation from age 5 until puberty, when basal cortisol levels and cortisol reactivity both show increases (Netherton et al., 2004; Stroud et al., 2009). This increase in HPA reactivity and the greater reactivity of neural systems underlying emotion have been hypothesized to increase risk for psychiatric disorders in adolescence (Spear, 2000). Shifts in neurodevelopment and stress system functioning across puberty may make this developmental period a time during which the impacts of early experiencesdboth positive and negativedmay be fully realized. Alternatively, the peripubertal period may be a time of heightened plasticity of the HPA axis and the neuroaffective circuits that stimulate its activation, thus allowing the system to calibrate to current life conditions, for better or for worse (DePasquale et al., 2019; Romeo, 2010). The HPA axis is regulated by multiple factors (Herman et al., 2016), so understanding developmental changes in multiple systems with inputs to the HPA axis are important to consider. For example, neural systems including the amygdala and BNST are involved with activation of the HPA axis, whereas the medial prefrontal cortex (mPFC) and hippocampus are involved in inhibition of the axis. Interestingly, the human hippocampus does not show developmental increases in GR, which suggests that the hippocampus may have a relatively mature ability to terminate the HPA response to stress at birth (Pryce, 2008). However, GR mRNA expression levels in the prefrontal cortex do increase into adolescence, and GR expression appears to be as high or even higher in the human neocortex as in the hippocampus (Pryce, 2008). Thus, the protracted development of the prefrontal cortex continues to impactdand is impacted bydthe development of the HPA axis well beyond infancy. The complex stress-mediating circuitry that unfolds over time in response to environmental signals, such as early caregiving, has been termed the “neurosymphony of stress” (Joëls and Baram, 2009). Receptors for CRH outside the HPA axis have been detected in the basolateral and medial nuclei of the amygdala, prefrontal cortex, hippocampus, reticular formation, and cerebellum, suggesting broad effects in many parts of the brain following the release of CRH. Interestingly, cortisol in the PVN downregulates CRH, whereas in the amygdala, chronic elevations of cortisol upregulate CRH and increase fear behavior. For the HPA axis to react to psychological stressors, the central nucleus of the amygdala activates a pathway that crosses multiple synapses and involves the bed nucleus of the stria

The effects of stress on early brain and behavioral development Chapter | 26

573

terminalis, which then converges on the PVN to release CRH (Ulrich-Lai and Herman, 2009). Threats to the social self, especially when uncontrollable or unpredictable, are one of the most reliable and relatively common triggers of HPA activation (Dickerson and Kemeny, 2004). Cortisol regulates its own release through negative feedback mechanisms in the hypothalamus, pituitary, medial prefrontal cortex (mPFC), and hippocampus (Tasker and Herman, 2011). GABA-producing cells surrounding the PVN have tonic inhibition on the HPA axis, and downregulation of GABA input due to chronic stress reduces the tonic inhibition on the system. CRH and ACTH can also regulate their own production. Negative feedback of cortisol can act quickly through the endocannabinoid system in the PVN or slowly through genomic mechanisms in the mPFC and hippocampus (Joëls and Baram, 2009). Chronic stress may also reduce GR expression in the mPFC and hippocampus as a result of negative feedback mechanisms to protect the brain from chronic HPA activation (Joëls and Baram, 2009). These adaptations to stress occurring early in life may have implications for neurodevelopment years later. There is evidence that early adversity speeds up certain aspects of neurodevelopment, including connectivity of the amygdala and mPFC, which is important for emotion regulation (Gee et al., 2013). In both human and animal models, these changes are mediated by altered HPA activity (Gee et al., 2013), which suggests that the HPA axis may facilitate faster neurodevelopment in the context of adversity to promote adaptation to the current environment. Overall, the HPA axis is a complex system involving multiple inputs from various physiological systems that is sensitive to the physical and social environment to coordinate basal regulation and reactivity to and recovery from stress.

26.4.1 Social regulation of the hypothalamicepituitaryeadrenocortical axis and the role of caregivers Across rodents, nonhuman primates, and humans, there is a wealth of evidence that both caregiver proximity and contact are critical to regulation of the HPA axis in early postnatal development (Gunnar et al., 2015; Sanchez et al., 2001). Sensitive and responsive caregiving in the first few months of life is associated with a more rapid recovery of the HPA axis following activation (Blair et al., 2006), whereas insensitive care is associated with increases in cortisol during infante parent play (Spanglar et al., 1994). These findings extend to nonparent caregivers including childcare providers (Vermeer and van IJzendoorn, 2006), as young children in out-of-home care with less sensitive and/or more intrusive care providers show larger increases in cortisol across the childcare day than do those with more sensitive caregivers. Sensitive and responsive caregiving is an important component in the development of secure attachment, which characterizes the relationship between the child and their caregiver and not a trait of the child. Children may be securely attached to some caregivers while being insecurely attached to others. Nearly all studies on attachment security have centered on the motherechild relationship, and these studies suggest that the presence and availability of the mother in secure attachment relationships provides a powerful buffer for HPA axis activity, whereas the presence of the mother in insecure relationships is less capable of buffering HPA responses when infants are frightened and distressed (Gunnar and Donzella, 2002). While the exact mechanism behind social buffering in humans is unknown, neurobiological and developmental work suggests that neural priming in areas of the brain associated with fear and emotion and the use of attachment figures as safety signals likely contribute (Hostinar et al., 2014). Oxytocin, a neuropeptide associated with social behaviors, is thought to be one of the proposed mediators of social buffering (Smith and Wang, 2014). Intranasal oxytocin infusions have been shown to enhance social support’s stress-buffering effects (Heinrichs et al., 2003), but negative early-care experiences may influence oxytocin’s role in social behavior (Fries et al., 2005). The actual presence of the attachment figure is needed to buffer HPA axis activity early in life, although secure attachment relationships are associated with better regulated anxiety and stress responses across development. When mothers and toddlers have a secure attachment relationships, toddlers show lower cortisol levels on days when the mother is present at a new child care arrangement than when she is absent (Ahnert et al., 2004). These elevations in cortisol decreased over several weeks but did not dissipate after 5 months in child care, suggesting the importance of parental presence early in life. Parents are capable of buffering increases in cortisol in response to stressors throughout childhood but less so during adolescence (Hostinar et al., 2015). The decrease in effectiveness of parents as a social buffer appears to track pubertal development more than age (Doom et al., 2015). In addition, friends do not take over the buffering role in adolescence at least when the stressor involves social evaluation; indeed, for these stressors, their presence actually increases cortisol responses relative to responses when parents serve as stress buffers (Doom et al., 2017). This loss of social buffering effectiveness may be one potential contributor to the increase in psychopathology during early adolescence. At this point, it is unclear whether the reduction in social buffering effectiveness extends beyond threats to the social self, as only these stressors have been studied. Nor is it clear how long a period of reduced social buffering effectiveness youth experience.

574 PART | II Cognitive development

It is clear that by adulthood, the presence of one’s spouse is a powerful social buffer, particularly if the relationship is supportive and secure (Beckes and Coan, 2011). In addition, social support figures in adulthood appear to serve as prepared safety stimuli (Hornstein et al., 2016). Finally, the HPA axis is powerfully regulated by spousal attachment figures in adulthood (Heinrichs et al., 2003). There are variations in buffering ability by sex, physical touch, and relationship to the buffering figure (e.g., friend vs. romantic partner), though emotional intimacy does appear to be a key element of social buffering effects as strangers will generally not buffer the HPA axis. Whether parents again become powerful stress buffers for their adult children is not known. We are also unsure what qualities of friend and parent relationships beyond infancy and toddlerhood lead to the most effective social buffering and if more effective social buffering of the HPA axis prevents the onset of psychopathology in the face of stress.

26.4.2 Early adversity Children’s early-care histories shape the development of the HPA axis, with both normal variations (e.g., sensitive vs. harsh parenting) and extreme variations (e.g., abuse or neglect) in care associated with HPA regulation across development. Individual differences in genetics, as well as risk and protective factors, play a large role in the effects of these earlycare experiences on development (Koss and Gunnar, 2018). Stressors that are particularly potent involve either threat, such as physical or sexual abuse, interparental conflict, and community violence, or deprivation, such as neglect, institutional care, or parental loss. Many children experience stressors that involve both threat and deprivation, such as extreme poverty, and there is evidence that the effects of these early stressors are cumulative (Doan et al., 2014). These conditions of early stress are typically chronic, with the rare exceptions (e.g., children adopted out of institutions), which makes it challenging to understand the effects of stressor timing on development.

26.4.2.1 Diurnal cortisol following postnatal stress The diurnal HPA rhythm that is maturing during the first postnatal years appears to be sensitive to immediate contexts of care. In a study of 46 2- and 3-year-olds living in an institution in Romania, no child exhibited the typical diurnal pattern that would be expected at this age. In addition, the institutionalized children showed lower cortisol than home-reared Romanian children in the morning and slightly higher cortisol in the later afternoon hours (Carlson and Earls, 1997). However, institutions are quite heterogeneous, and the flatter diurnal rhythm is not always found (Dobrova-Krol et al., 2008). Nonetheless, similar findings have been reported for children recently placed in foster care after being removed from maltreating homes (Bruce et al., 2009). It is still not clear what factors are associated with the flatter diurnal rhythm in young children when it is found; however, there are several findings that point to neglect rather than abuse as an important factor. Bruce et al. (2009) found that it was the severity of neglect that predicted low early morning and a flatter diurnal slope in children entering a new foster placement. Similarly, in toddlers and preschoolers adopted from institutions, those who experienced less supportive social care (i.e., fewer caregivers to children, less interaction) produced the lower and flatter cortisol diurnal activity (Koss et al., 2014; see also Kertes et al., 2008). Adults adopted as children from maltreating homes also show a lower morning cortisol than adoptees without this history, and the main predictor of a flatter diurnal slope appeared again to be the severity of neglect (van der Vegt et al., 2009). Notably, the altered pattern of diurnal activity was persistent for several years postadoption in one study (Koss et al., 2014) or from childhood to adulthood in another (van der Vegt et al., 2009). Thus at least in some cases, removing the child from conditions of neglect does not appear to normalize the diurnal rhythm. In other cases, no differences in rhythm have been noted between those with abuse and neglect histories and individuals without such histories, particularly for children with later-onset abuse and low levels of internalizing symptoms (Cicchetti et al., 2010). While it is as yet uncertain whether these differences reflect the type or severity of adversity, there is some evidence that the age of removal from adversity may affect the HPA axis.

26.4.2.2 Effects of early care on cortisol set points and reactivity In rodent models, poor quality postnatal care shapes development in a unidirectional manner, increasing both behavioral and physiological reactivity to stress. However, in primates, including humans, adverse early life care does not always result in hyperactivity of the HPA axis. For example, rhesus infants reared on nonanimate surrogates exhibit hypoactivity of the axis in terms of basal levels, in response to psychosocial stressors and in response to pharmacological challenge (Capitanio et al., 2005). Similarly, squirrel monkeys separated repeatedly from their mothers exhibit a more regulated HPA axis (Lyons et al., 2010). Thus, several researchers have proposed that the relationship between early adversity and stress system functioning may be a J-shaped curve, with modest amounts of adversity, leading to a better regulated system,

The effects of stress on early brain and behavioral development Chapter | 26

575

whereas high amounts result in hyperactivity of stress systems (Boyce and Ellis, 2005; Del Giudice et al., 2011). The problem with these models is that, as noted for infant rhesus, complete removal of the mother and other potential caregivers, which must be viewed as extreme adversity, results in hypocortisolism. This also appears to be true for children reared in institutions. Not only, as noted, are flatter diurnal rhythms observed, but for those who remain in institutional care for their first 2 years of life or more, the HPA and SNS response to stressors also appears to be blunted (McLaughlin et al., 2015). In addition, we have noted a blunted cortisol awakening response for children adopted internationally from institutional care if they were over 16 months at removal, but not if they were 16 months or younger (Leneman et al., 2018). Likewise, adults adopted as young children from Romanian orphanages still exhibit a blunted CAR in adulthood (Kumsta et al., 2017). Evidence that low care or neglect is associated with low cortisol reactivity has also been found in family-reared children. Thus both low and very high maternal care has been associated with low cortisol reactivity (Engert et al., 2010). Note that the association of high maternal care with low reactivity argues against the J-shaped model described earlier. Similarly, again in contrast to that model, children who have been bullied were found to have blunted cortisol responses compared with their nonbullied twin and that these differences were not attributable to genetics or shared environments (Ouellet-Morin et al., 2011). Likewise, a study of teenage girls who had been maltreated showed blunted cortisol but not cardiac responses to a social evaluative stressor, suggesting a block on the HPA axis (MacMillan et al., 2009). While these studies would seem to argue that extreme early adversity shapes a hyporesponsive HPA axis, other studies do tend to support the more J-shaped model, at least with regard to higher levels of adversity predicting hyperactivity of the axis. Many of these studies, however, have involved children of depressed mothers, raising the possibility that the genetics of depression may also play a role. Thus, children whose mothers reported high depressive symptoms during infancy showed higher afternoon cortisol levels at age 4 (Essex et al., 2002). A cash transfer program for extremely poor families in Mexico resulted in lower cortisol levels for children aged 2e6 years, but this effect was only noted for the children of mothers with high depressive symptoms (Fernald and Gunnar, 2009). This finding, then, indicates that poverty was associated with elevated cortisol for children of depressed, but not nondepressed mothers. Children of women who experienced postnatal depression showed higher and more variable morning cortisol concentrations at 13 years (Halligan et al., 2007). Other studies also support the idea that psychiatric disorders moderate the association between childhood adversity and HPA axis functioning. For example, survivors of child maltreatment have been found to exhibit higher cortisol levels and great reactivity, but primarily if they have also developed major depression (e.g., Heim et al., 2008). For adults without a psychiatric diagnosis, those with histories of adversity often show a lower HPA axis set point and more blunted cortisol responses to psychosocial stress (Carpenter et al., 2007). In children, those who were physically or sexually abused before age 5 showed altered diurnal cortisol production only if they were also high in internalizing problems (Cicchetti et al., 2010). One of the challenges of determining traitlike HPA activity is that both child extremes in temperament (Gunnar et al., 1997) and life conditions (Doom et al., 2014) have been shown to be associated with day-to-day variability in basal cortisol activity, and likely also in reactivity. Importantly, maltreatment that leads to posttraumatic stress disorder (PTSD) is not associated with hyperactivity of the HPA axis, even though it is often comorbid with depression. Instead, in adults, it is associated with normal to low levels of cortisol activity (Yehuda, 2001). PTSD before puberty appears to be related to elevated cortisol production (Carrion et al., 2002). Longitudinal work on sexually abused girls showed that elevated cortisol levels decreased beginning in adolescence with lower cortisol levels by adulthood, with at least this study supporting time-since-trauma as the critical factor over pubertal onset (Trickett et al., 2010).

26.4.3 Individual differences in sensitivity to experience Not every child exposed to chronic or severe stress goes on to develop emotional or behavioral problems, and developmental science has been attempting to understand the wide variation in outcomes following early stress. For example, two children who have experienced maltreatment from a caregiver may have vastly different emotional and behavioral outcomes if there are variations in social support, the environment, genetics, and previous early experiences that influence risk and resilience processes. One theory about variation in responses to early experiences is the biological sensitivity to context theory (Boyce and Ellis, 2005). This theory argues that individuals whose biology is more reactive to context are more likely to have outcomes that vary by the contexts they encountered during development. Similarly, the differential susceptibility framework argues that there are genetic polymorphisms that result in greater plasticity and reactions to contextual variations (Belsky and Pluess, 2009). Both of these theories are distinct from the diathesisestress framework

576 PART | II Cognitive development

that frequently guides work on early stress, which only frames individual variation as vulnerability to adversity. Both biological sensitivity to context and differential susceptibility theories argue that the same individual difference fact that may make one vulnerable to adverse conditions may allow superlative developmental outcomes under supportive conditions. Thus, with regard to genes, those referred to as “risk alleles” might be better named “plasticity genes” in reference to their vulnerability to risky environments and enhanced outcomes in enriched environments. There are a number of factors that influence susceptibility to effects on both the fear- and stress-responsive systems, including temperament, sex differences, and genetics. One dimension of temperament closely tied to the fear and stress systems is behavioral inhibition, which includes greater behavioral reactivity and less approaching of novel or unpredictable stimuli. Infants who are behaviorally inhibited often go on to be shy and more likely to have social anxiety (Henderson et al., 2015). These infants do not always have higher cortisol levels and reactivity, as stress system functioning in behaviorally inhibited infants likely depends on factors such as support from the attachment figure or whether avoidance is an effective way to regulate stress. In the absence of supportive care, higher cortisol reactivity in inhibited children could mediate their greater vulnerability to adverse early caregiving (Phillips et al., 2011). Their sensitivity to the environment could help them to excel in high-quality care (biological sensitivity to context, discussed above; Boyce and Ellis, 2005). Animal models of early stress document sex-specific changes as a result of postnatal stress and stress hormone exposure (Bale and Epperson, 2015). Increasing evidence points to sex differences in neurological development that begin prenatally, with developmental hormone exposure largely guiding the organization of the sexually dimorphic brain (Bale and Epperson, 2015). Hormonal effects on cell migration early in life may program stress-regulatory brain regions, including the hypothalamus and limbic circuitry, which contribute to sex differences in stress responses throughout the life span (Bale and Epperson, 2015). Sex-specific developmental trajectories observed postnatally are likely a continuation of prenatal changes in neurodevelopment, which may lead to sex differences in psychiatric disorders. There is evidence that females may be at greater risk for early postnatal and peripubertal stressors, leading to affective disorders later in life. Sex differences postnatally may be due to biological differences, including interactions between sex chromosome genes and hormonal changes, as well as differences in the effects of parents on developing fear- and stress-responsive systems between males and females. Sex is a major factor in guiding prenatal and early postnatal brain development, and it also contributes to brain development across puberty. Limbic brain regions, such as the amygdala and hippocampus, express many androgen and estrogen steroid hormone receptors and increase in volume in rodents and humans (Bale and Epperson, 2015). Sex-specific changes in brain development across puberty influence functioning of fear- and stress-responsive systems across the life span. Genetics play a role in one’s sensitivity to the environment and the likelihood of significant alterations in cognitive and socioemotional development following early stress. The CRHR1 gene, for example, has been shown to mediate both behavioral and physiological stress responses, and variants in this gene have been shown to interact with early abuse to increase depression risk (Gillespie et al., 2009). In the context of childhood stress, genes regulating GR and the HPA axis are of particular interest, as these have been associated with risk and resilience in the context of early stress (DeRijk and de Kloet, 2008; Gillespie et al., 2009). The same genetic polymorphism associated with risk in the context of early harsh caregiving could be associated with enhanced outcomes in the context of sensitive, responsive caregiving (Belsky and Pluess, 2009). More recently, genetic work has shifted to understanding how early caregiving may “get under the skin” to influence children’s health and development through the study of epigenetics. Epigenetics involves modifications to the genome that alter the accessibility of DNA and may alter gene expression but do not change the base-pair sequence. Epigenetic changes are thought to be one of the primary pathways by which early experiencesdboth positive and negativedmay be embedded to influence later cognitions, emotions, and behavior. A growing body of literature has begun to document associations between early stress and epigenetic modifications, including DNA methylation. A study of adolescents adopted during early childhood from institutions in Eastern Europe showed altered DNA methylation patterns in genes associated with neurodevelopment compared to adolescents raised in their birth families (Esposito et al., 2016). Even less extreme variation in early-life stressors has been shown to be associated with adolescent DNA methylation (Essex et al., 2013). Evidence in twins demonstrates that the twin who was bullied showed higher serotonin transporter methylation at 10 years compared with their nonbullied cotwin, which could not be attributed to genetic makeup or shared environments, and this higher methylation was associated with blunted cortisol reactivity to stress (Ouellet-Morin et al., 2013). A recent systematic review identified that 89% of publications in humans on early parental stress and methylation of the promoter of the GR gene (NR3C1) showed greater methylation with increased stress at this site (Turecki and Meaney, 2016). Epigenetics could be one way by which experiences are transmitted intergenerationally (Champagne, 2008). Further work must be done on

The effects of stress on early brain and behavioral development Chapter | 26

577

investigating the type and timing of early experiences that impact epigenetics, understanding who are the most sensitive to epigenetic alterations in response to the environment, and delineating the pathway by which epigenetic changes could lead to downstream effects on health and behavior.

26.4.4 Summary Accumulating research is illuminating the impact of early postnatal experiences on neurodevelopment and the development of stress-mediating systems. In humans, the HPA axis develops significantly over the first 6 months of postnatal life. The fear- and stress-responsive systems are strongly regulated by caregivers, and sensitive, responsive caregiving has been associated with better regulation of these systems and the ability to return to baseline quickly following challenges. Adverse early care, including maltreatment and institutional care, has been associated with disruptions in the HPA axis that can persist for years and may partially mediate cognitive, behavioral, and emotional outcomes following stress. More normative variations in early care are also associated with HPA profiles. The timing and type of stressor and the age at assessment are important moderators, as developmental changes like puberty may play an important role in the functioning of the HPA axis poststressor. In addition, current psychiatric status, including experiencing depression or PTSD, is associated with unique HPA profiles. Factors such as genetics and temperament are important moderators of HPA functioning and behavior, and specific genetic profiles or temperamental characteristics may act as a vulnerability factor in certain contexts and a resilience factor in others. Thus, developmental timing and individual differences are vital to consider when understanding the development of the HPA axis and the brain, particularly in the context of postnatal stress.

26.5 Future directions The study of pre- and postnatal stress and its impact on neurodevelopment and health has thus far proceeded largely independently. It is time for this work to become integrated, both empirically and theoretically. While researchers studying prenatal stress have sometimes obtained measures of outcomes later in life, it is rare that postnatal experiences are examined as anything other than potential confounds to be controlled statistically. Yet, postnatal experiences have the potential to either ameliorate or exacerbate prenatal effects. In addition, alterations in infant functioning related to prenatal experiences may result in differential sensitivity to variations in early-care experiences and/or behaviors, which elicit different responses from caregivers. As noted, some researchers are beginning to address how postnatal care interacts with prenatal stress exposure to influence cognitive, behavioral, and health outcomes. More of this work is needed. From a fetal programming perspective, it is especially critical that conceptual models are examined via longitudinal studies that track postnatal development. If, as some models posit, fetal programming via stress mediators prepares the fetus to survive in a harsh postnatal world (Gluckman and Hanson, 2004), then we need to design studies to directly test such hypotheses. With regard to nutrition, there is evidence that concordance in pre- and postnatal nutrition leads to more functional health outcomes than discordance (Cleal et al., 2007). Within the socioemotional domain, infant developmental outcomes appear to be improved when mothers experience similar levels of depression during and after pregnancy (Sandman et al., 2012). Replication and extension of these findings is essential to identify the prenatal stressors and postnatal outcomes for which adaptive advantage exists. Consistent with the early literature on outcomes for premature infants, it is also possible that a supportive postnatal environment may ameliorate and a harsh environment may exacerbate the neurobehavioral sequelae of prenatal stress. In this case, discordance in the harshness of pre- and postnatal experience may predict better outcomes as long as the later environment is supportive. Recent human studies have provided support for this possibility, demonstrating that highquality maternal care can compensate for the negative effects of prenatal stress exposures (Bergman et al., 2010; Schechter et al., 2017). It remains likely that the effects of stress during the prenatal and postnatal period will differ by developmental outcome, and there is a clear need for prospective studies with multiple prenatal and postnatal assessments. A more compelling reason to study prenatal stress in the context of postnatal stress is if prenatal experiences alter the individual’s susceptibility to postnatal experiences. This would be the case if, as it has been argued, prenatal stress programs postnatal plasticity (Pluess and Belsky, 2011) or if prenatal stress increases suitability to certain types of postnatal environments. If true, this might mean that postnatal experiences might not only exacerbate or reduce the negative impacts of prenatal stress, which is more a diathesisestress model, prenatal stress might result in more positive outcomes if postnatal conditions are favorable, and more negative if they are not, as has been shown in at least one study of voles (Hartman et al., 2018). This argument, of course, completely contradicts the Barker hypothesis, which argues that adaptations to harsh conditions during fetal development increase survival advantage if the postnatal environment is harsh

578 PART | II Cognitive development

and poor in resources, but impairs outcomes if it is a mismatch (supportive, well-resourced) with the prenatal environment. Strong support for this survival advantage of a match between the prenatal and postnatal environment comes from both animal and human studies looking at nutrition (Cleal et al., 2007; Gluckman and Hanson, 2004; Gluckman et al., 2005). Few studies, however, have looked at the impact of a mismatch between the prenatal and postnatal environments in the psychosocial domain (Sandman et al., 2012). Which predictions turn out to be accurate requires the longitudinal studies, beginning before birth that we have been calling for in this future directions section. They also likely require a range of outcome to be assessed, as the predictive utility of the model may be a function of the outcomes we study. That is, psychological functioning might follow a sensitivity to context/susceptibility model as prenatally stress and emotional sensitivity and physiological reactive children blossom under supportive parenting, whereas body mass and cardiovascular health suffer when a harsh and low nutrient fetal environment is followed by a nutrient-rich and high-calorie postnatal context. Thus far, we have only focused these future direction comments on the value of incorporating research and theory on postnatal stress into models and research on prenatal stress. Perhaps a more critical need is for those studying postnatal stress to consider the role that prenatal stress may be playing in their findings. Certainly, it seems likely that children who are abused, neglected, or abandoned to orphanage care may be the product of stressed pregnancies. It also is likely that they are the products of pregnancies complicated by poor nutrition and exposure to alcohol and drugs. Unfortunately, in many studies of postnatal stress, there is meager information about prenatal conditions or even birth outcomes. Retrospective reports obtained from parents in studies, for example, of child maltreatment must be suspect. For children abandoned to orphanage care, even the child’s age at abandonment is often unknown, let alone their gestational age, health at birth, and prenatal conditions. Nonetheless, despite the challenge of obtaining accurate information on prenatal conditions for children identified because of their poor postnatal care, we need studies where such information can be or has been obtained. This is particularly important in studies examining interventions to improve outcomes as prenatal conditions may moderate how the child responds. Finally, one area that would seem ripe for integration into research on pre- and postnatal stress is the role that the dramatic hormonal changes that accompany pregnancy contribute to the quality of postnatal caregiving. Stress and reproductive hormones during pregnancy are associated with maternal cognitive functioning (Glynn, 2010), the development of postpartum depression (Yim et al., 2009), and the quality of maternal care (Feldman and Eidelman, 2007; Glynn et al., 2016). The maternal hormonal milieu also impacts the developing fetus, as was discussed. Thus, to close the loop on our understanding of the ways in which stress impacts the developing fetus and young child, we need studies that incorporate the impacts of stress and maternal hormonal changes on the mother, her caregiving, and her response to her infant. This research will directly inform prenatal interventions through considering the intervention’s effects in the context of the anticipated postnatal environment. Likewise, postnatal interventions may be adapted based on knowledge of the infant’s prenatal environment.

References Aarnoudse-Moens, C.S.H., Weisglas-Kuperus, N., van Goudoever, J.B., Oosterlaan, J., 2009. Meta-analysis of neurobehavioral outcomes in very preterm and/or very low birth weight children. Pediatrics 124 (2), 717e728. Ahnert, L., Gunnar, M.R., Lamb, M.E., Barthel, M., 2004. Transition to child care: associations with infantemother attachment, infant negative emotion, and cortisol elevations. Child Dev. 75 (3), 639e650. Baibazarova, E., van de Beek, C., Cohen-Kettenis, P.T., Buitelaar, J., Shelton, K.H., van Goozen, S.H., 2013. Influence of prenatal maternal stress, maternal plasma cortisol and cortisol in the amniotic fluid on birth outcomes and child temperament at 3 months. Psychoneuroendocrinology 38 (6), 907e915. Bale, T.L., Epperson, C.N., 2015. Sex differences and stress across the lifespan. Nat. Neurosci. 18 (10), 1413. Barker, D.J., 1998. In utero programming of chronic disease. Clin. Sci. 95 (2), 115e128. Bäumler, D., Kirschbaum, C., Kliegel, M., Alexander, N., Stalder, T., 2013. The cortisol awakening response in toddlers and young children. Psychoneuroendocrinology 38 (11), 2485e2492. Beckes, L., Coan, J.A., 2011. Social baseline theory: the role of social proximity in emotion and economy of action. Social and Personality Psychology Compass 5 (12), 976e988. Belsky, J., Pluess, M., 2009. Beyond diathesis stress: differential susceptibility to environmental influences. Psychol Bull. 135 https://doi.org/10.1037/ a0017376 (American Psychological Association, US). Bergman, K., Sarkar, P., Glover, V., O’Connor, T.G., 2010. Maternal prenatal cortisol and infant cognitive development: moderation by infantemother attachment. Biol. Psychiatry 67 (11), 1026e1032. Bergman, K., Sarkar, P., O’connor, T.G., Modi, N., Glover, V., 2007. Maternal stress during pregnancy predicts cognitive ability and fearfulness in infancy. J. Am. Acad. Child Adolesc. Psychiatry 46 (11), 1454e1463.

The effects of stress on early brain and behavioral development Chapter | 26

579

Blair, C., Granger, D., Willoughby, M., Kivlighan, K., 2006. Maternal sensitivity is related to hypothalamic-pituitary-adrenal Axis stress reactivity and regulation in response to emotion challenge in 6-month-old infants. Ann. N. Y. Acad. Sci. 1094 (1), 263e267. https://doi.org/10.1196/ annals.1376.031. Bourgeois, J., 1997. Synaptogenesis, heterochrony and epigenesis in the mammalian neocortex. Acta Paediatr. 86 (S422), 27e33. Boyce, W.T., Ellis, B.J., 2005. Biological sensitivity to context: I. An evolutionaryedevelopmental theory of the origins and functions of stress reactivity. Dev. Psychopathol. 17 (02), 271e301. Bruce, J., Fisher, P.A., Pears, K.C., Levine, S., 2009. Morning cortisol levels in preschool-aged foster children: differential effects of maltreatment type. Dev. Psychobiol. 51 (1), 14e23. Buss, C., Davis, E., Hobel, C., Sandman, C., 2011. Maternal pregnancy-specific anxiety is associated with child executive function at 6e9 years age. Stress 14 (6), 665e676. Buss, C., Davis, E.P., Muftuler, L.T., Head, K., Sandman, C.A., 2010. High pregnancy anxiety during mid-gestation is associated with decreased gray matter density in 6e9-year-old children. Psychoneuroendocrinology 35 (1), 141e153. Buss, C., Davis, E.P., Shahbaba, B., Pruessner, J.C., Head, K., Sandman, C.A., 2012. Maternal cortisol over the course of pregnancy and subsequent child amygdala and hippocampus volumes and affective problems. Proc. Natl. Acad. Sci. 109 (20). E1312eE1319. Buss, C., Entringer, S., Wadhwa, P.D., 2012. Fetal programming of brain development: intrauterine stress and susceptibility to psychopathology. Sci. Signal. 5 (245) pt.7-pt7. Capitanio, J.P., Mendoza, S.P., Mason, W.A., Maninger, N., 2005. Rearing environment and hypothalamic-pituitary-adrenal regulation in young rhesus monkeys (Macaca mulatta). Dev. Psychobiol. 46 (4), 318e330. Carlson, M., Earls, F., 1997. Psychological and neuroendocrinological sequelae of early social deprivation in institutionalized children in Romania. Ann. N. Y. Acad. Sci. 807 (1), 419e428. Carpenter, L.L., Carvalho, J.P., Tyrka, A.R., Wier, L.M., Mello, A.F., Mello, M.F., et al., 2007. Decreased adrenocorticotropic hormone and cortisol responses to stress in healthy adults reporting significant childhood maltreatment. Biol. Psychiatry 62 (10), 1080e1087. Carrion, V.G., Weems, C.F., Ray, R.D., Glaser, B., Hessl, D., Reiss, A.L., 2002. Diurnal salivary cortisol in pediatric posttraumatic stress disorder. Biol. Psychiatry 51 (7), 575e582. Catalani, A., Casolini, P., Scaccianoce, S., Patacchioli, F., Spinozzi, P., Angelucci, L., 2000. Maternal corticosterone during lactation permanently affects brain corticosteroid receptors, stress response and behaviour in rat progeny. Neuroscience 100 (2), 319e325. Champagne, F.A., 2008. Epigenetic mechanisms and the transgenerational effects of maternal care. Front. Neuroendocrinol. 29 (3), 386e397. Cicchetti, D., Rogosch, F.A., Gunnar, M.R., Toth, S.L., 2010. The differential impacts of early physical and sexual abuse and internalizing problems on daytime cortisol rhythm in school-aged children. Child Dev. 81 (1), 252e269. Cleal, J.K., Poore, K.R., Boullin, J.P., Khan, O., Chau, R., Hambidge, O., et al., 2007. Mismatched pre-and postnatal nutrition leads to cardiovascular dysfunction and altered renal function in adulthood. Proc. Natl. Acad. Sci. 104 (22), 9529e9533. Clifton, V., 2010. Sex and the human placenta: mediating differential strategies of fetal growth and survival. Placenta 31, S33eS39. Conradt, E., Lester, B.M., Appleton, A.A., Armstrong, D.A., Marsit, C.J., 2013. The roles of DNA methylation of NR3C1 and 11b-HSD2 and exposure to maternal mood disorder in utero on newborn neurobehavior. Epigenetics 8 (12), 1321e1329. Curran, M.M., Sandman, C.A., Davis, E.P., Glynn, L.M., Baram, T.Z., 2017. Abnormal dendritic maturation of developing cortical neurons exposed to corticotropin releasing hormone (CRH): insights into effects of prenatal adversity? PLoS One 12 (6), e0180311. Davis, E., Glynn, L.M., Waffarn, F., Sandman, C.A., 2011. Prenatal maternal stress programs infant stress regulation. J. Child Psychol. Psychiatry 52 (2), 119e129. Davis, E., Hankin, B.L., Swales, D.A., Hoffman, M.C., 2018. An experimental test of the fetal programming hypothesis: can we reduce child ontogenetic vulnerability to psychopathology by decreasing maternal depression? Dev. Psychopathol. 30 (3), 787e806. Davis, E., Head, K., Buss, C., Sandman, C.A., 2017. Prenatal maternal cortisol concentrations predict neurodevelopment in middle childhood. Psychoneuroendocrinology 75, 56e63. Davis, E., Sandman, C.A., 2010. The timing of prenatal exposure to maternal cortisol and psychosocial stress is associated with human infant cognitive development. Child Dev. 81 (1), 131e148. Davis, E., Townsend, E., Gunnar, M., Guiang, S., Lussky, R., Cifuentes, R., Georgieff, M., 2006. Antenatal betamethasone treatment has a persisting influence on infant HPA axis regulation. J. Perinatol. 26 (3), 147. Davis, E., Waffarn, F., Sandman, C.A., 2011. Prenatal treatment with glucocorticoids sensitizes the hpa axis response to stress among full-term infants. Dev. Psychobiol. 53 (2), 175e183. Davis, E.P., Glynn, L.M., Schetter, C.D., Hobel, C., Chicz-Demet, A., Sandman, C.A., 2005. Corticotropin-releasing hormone during pregnancy is associated with infant temperament. Dev. Neurosci. 27 (5), 299e305. Davis, E.P., Glynn, L.M., Schetter, C.D., Hobel, C., Chicz-Demet, A., Sandman, C.A., 2007. Prenatal exposure to maternal depression and cortisol influences infant temperament. J. Am. Acad. Child Adolesc. Psychiatry 46 (6), 737e746. Davis, E.P., Pfaff, D., 2014. Sexually dimorphic responses to early adversity: implications for affective problems and autism spectrum disorder. Psychoneuroendocrinology 49, 11e25. Davis, E.P., Sandman, C.A., 2012. Prenatal psychobiological predictors of anxiety risk in preadolescent children. Psychoneuroendocrinology 37 (8), 1224e1233. Davis, E.P., Snidman, N., Wadhwa, P.D., Glynn, L.M., Schetter, C.D., Sandman, C.A., 2004. Prenatal maternal anxiety and depression predict negative behavioral reactivity in infancy. Infancy 6 (3), 319e331.

580 PART | II Cognitive development

de Weerth, C., van Hees, Y., Buitelaar, J.K., 2003. Prenatal maternal cortisol levels and infant behavior during the first 5 months. Early Hum. Dev. 74 (2), 139e151. Del Cerro, M., Perez-Laso, C., Ortega, E., Martín, J., Gomez, F., Perez-Izquierdo, M., Segovia, S., 2010. Maternal care counteracts behavioral effects of prenatal environmental stress in female rats. Behav. Brain Res. 208 (2), 593e602. Del Giudice, M., Ellis, B.J., Shirtcliff, E.A., 2011. The adaptive calibration model of stress responsivity. Neurosci. Biobehav. Rev. 35 (7), 1562e1592. Denver, R.J., 1997. Environmental stress as a developmental cue: corticotropin-releasing hormone is a proximate mediator of adaptive phenotypic plasticity in amphibian metamorphosis. Horm. Behav. 31 (2), 169e179. DePasquale, C., Donzella, B., Gunnar, M., 2019. Pubertal recalibration of cortisol reactivity following early life stress: a cross-sectional analysis. J. Child Psychol. Psychiatry. DeRijk, R.H., de Kloet, E.R., 2008. Corticosteroid receptor polymorphisms: determinants of vulnerability and resilience. Eur. J. Pharmacol. 583 (2e3), 303e311. Dickerson, S.S., Kemeny, M.E., 2004. Acute stressors and cortisol responses: a theoretical integration and synthesis of laboratory research. Psychol. Bull. 130 (3), 355e391. https://doi.org/10.1037/0033-2909.130.3.355. DiPietro, J.A., 2004. The role of prenatal maternal stress in child development. Curr. Dir. Psychol. Sci. 13 (2), 71e74. DiPietro, J.A., Novak, M.F., Costigan, K.A., Atella, L.D., Reusing, S.P., 2006. Maternal psychological distress during pregnancy in relation to child development at age two. Child Dev. 77 (3), 573e587. DiPietro, J.A., Voegtline, K.M., 2017. The gestational foundation of sex differences in development and vulnerability. Neuroscience 342, 4e20. https:// doi.org/10.1016/j.neuroscience.2015.07.068. Doan, S.N., Dich, N., Evans, G.W., 2014. Childhood cumulative risk and later allostatic load: mediating role of substance use. Health Psychol. 33 (11), 1402. Dobrova-Krol, N.A., van IJzendoorn, M.H., Bakermans-Kranenburg, M.J., Cyr, C., Juffer, F., 2008. Physical growth delays and stress dysregulation in stunted and non-stunted Ukrainian institution-reared children. Infant Behav. Dev. 31 (3), 539e553. Doom, J.R., Cicchetti, D., Rogosch, F.A., 2014. Longitudinal patterns of cortisol regulation differ in maltreated and nonmaltreated children. J. Am. Acad. Child Adolesc. Psychiatry 53 (11), 1206e1215. https://doi.org/10.1016/j.jaac.2014.08.006. Doom, J.R., Doyle, C.M., Gunnar, M.R., 2017. Social stress buffering by friends in childhood and adolescence: effects on HPA and oxytocin activity. Soc. Neurosci. 12 (1), 8e21. https://doi.org/10.1080/17470919.2016.1149095. Doom, J.R., Hostinar, C.E., VanZomeren-Dohm, A.A., Gunnar, M.R., 2015. The roles of puberty and age in explaining the diminished effectiveness of parental buffering of HPA reactivity and recovery in adolescence. Psychoneuroendocrinology 59, 102e111. https://doi.org/10.1016/ j.psyneuen.2015.04.024. Dunkel-Schetter, C., Glynn, L.M., 2011. Stress in pregnancy: empirical evidence and theoretical issues to guide interdisciplinary research. In: The Handbook of Stress Science Biology, Psychology and Health. Springer Publishing Company, New York. Ellman, L.M., Schetter, C.D., Hobel, C.J., Chicz-DeMet, A., Glynn, L.M., Sandman, C.A., 2008. Timing of fetal exposure to stress hormones: effects on newborn physical and neuromuscular maturation. Dev. Psychobiol. 50 (3), 232e241. Engert, V., Efanov, S.I., Dedovic, K., Duchesne, A., Dagher, A., Pruessner, J.C., 2010. Perceived early-life maternal care and the cortisol response to repeated psychosocial stress. J. Psychiatry Neurosci. 35 (6), 370. Esposito, E.A., Jones, M.J., Doom, J.R., MacIsaac, J.L., Gunnar, M.R., Kobor, M.S., 2016. Differential DNA methylation in peripheral blood mononuclear cells in adolescents exposed to significant early but not later childhood adversity. Dev. Psychopathol. 28 (4pt.2), 1385e1399. https://doi.org/ 10.1017/S0954579416000055. Essex, M.J., Klein, M.H., Cho, E., Kalin, N.H., 2002. Maternal stress beginning in infancy may sensitize children to later stress exposure: effects on cortisol and behavior. Biol. Psychiatry 52 (8), 776e784. Essex, M.J., Thomas Boyce, W., Hertzman, C., Lam, L.L., Armstrong, J.M., Neumann, S.M., Kobor, M.S., 2013. Epigenetic vestiges of early developmental adversity: childhood stress exposure and DNA methylation in adolescence. Child Dev. 84 (1), 58e75. Feldman, R., Eidelman, A.I., 2007. Maternal postpartum behavior and the emergence of infantemother and infantefather synchrony in preterm and fullterm infants: the role of neonatal vagal tone. Dev. Psychobiol. 49 (3), 290e302. Fernald, L., Gunnar, M.R., 2009. Effects of a poverty-alleviation intervention on salivary cortisol in very low-income children. Soc. Sci. Med. 68 (12), 2180e2189. https://doi.org/10.1016/j.socscimed.2009.03.032. French, N.P., Hagan, R., Evans, S.F., Godfrey, M., Newnham, J.P., 1999. Repeated antenatal corticosteroids: size at birth and subsequent development. Am. J. Obstet. Gynecol. 180 (1), 114e121. Fries, A.B.W., Ziegler, T.E., Kurian, J.R., Jacoris, S., Pollak, S.D., 2005. Early experience in humans is associated with changes in neuropeptides critical for regulating social behavior. Proc. Natl. Acad. Sci. 102 (47), 17237e17240. Gee, D.G., Gabard-Durnam, L.J., Flannery, J., Goff, B., Humphreys, K.L., Telzer, E.H., et al., 2013. Early developmental emergence of human amygdalaeprefrontal connectivity after maternal deprivation. Proc. Natl. Acad. Sci. 10 (39), 15638e15643. Gillespie, C.F., Phifer, J., Bradley, B., Ressler, K.J., 2009. Risk and resilience: genetic and environmental influences on development of the stress response. Depress. Anxiety 26 (11), 984e992. https://doi.org/10.1002/da.20605. Gitau, R., Cameron, A., Fisk, N.M., Glover, V., 1998. Fetal exposure to maternal cortisol. The Lancet 352 (9129), 707e708. Gitau, R., Fisk, N.M., Teixeira, J.M., Cameron, A., Glover, V., 2001. Fetal hypothalamic-pituitary-adrenal stress responses to invasive procedures are independent of maternal responses. J. Clin. Endocrinol. Metab. 86 (1), 104e109.

The effects of stress on early brain and behavioral development Chapter | 26

581

Giurgescu, C., Zenk, S.N., Dancy, B.L., Park, C.G., Dieber, W., Block, R., 2012. Relationships among neighborhood environment, racial discrimination, psychological distress, and preterm birth in African American women. J. Obstet. Gynecol. Neonatal Nurs. 41 (6), E51eE61. Glover, V., Hill, J., 2012. Sex differences in the programming effects of prenatal stress on psychopathology and stress responses: an evolutionary perspective. Physiol. Behav. 106 (5), 736e740. Gluckman, P.D., Hanson, M.A., 2004. Living with the past: evolution, development, and patterns of disease. Science 305 (5691), 1733e1736. Gluckman, P.D., Hanson, M.A., Spencer, H.G., 2005. Predictive adaptive responses and human evolution. Trends Ecol. Evol. 20 (10), 527e533. Glynn, L.M., 2010. Giving birth to a new brain: hormone exposures of pregnancy influence human memory. Psychoneuroendocrinology 35 (8), 1148e1155. Glynn, L.M., Davis, E.P., Sandman, C.A., Goldberg, W.A., 2016. Gestational hormone profiles predict human maternal behavior at 1-year postpartum. Horm. Behav. 85, 19e25. Glynn, L.M., Howland, M.A., Sandman, C.A., Davis, E.P., Phelan, M., Baram, T.Z., et al., 2018. Prenatal maternal mood patterns predict child temperament and adolescent mental health. J. Affect. Disord. 228, 83e90. Glynn, L.M., Schetter, C.D., Hobel, C.J., Sandman, C.A., 2008. Pattern of perceived stress and anxiety in pregnancy predicts preterm birth. Health Psychol. 27 (1), 43. Grant, K.-A., McMahon, C., Reilly, N., Austin, M.-P., 2010. Maternal sensitivity moderates the impact of prenatal anxiety disorder on infant mental development. Early Hum. Dev. 86 (9), 551e556. Grey, K.R., Davis, E.P., Sandman, C.A., Glynn, L.M., 2013. Human milk cortisol is associated with infant temperament. Psychoneuroendocrinology 38 (7), 1178e1185. Gunnar, M.R., Donzella, B., 2002. Social regulation of the cortisol levels in early human development. Psychoneuroendocrinology 27 (1), 199e220. Gunnar, M.R., Doom, J.R., Esposito, E.A., 2015. Psychoneuroendocrinology of stress. In: Handbook of Child Psychology and Developmental Science. Gunnar, M.R., Herrera, A.M., 2013. The development of stress reactivity. In: The Oxford Handbook of Developmental Psychology, vol. 2. Gunnar, M.R., Tout, K., de Haan, M., Pierce, S., Stanbury, K., 1997. Temperament, social competence, and adrenocortical activity in preschoolers. Dev. Psychobiol. 31 (1), 65e85. Gunnar, M.R., Vazquez, D., 2006. Stress neurobiology and developmental psychopathology. In: Cicchetti, D., Cohen, D. (Eds.), Developmental Psychopathology, vol. 2. Wiley, Hoboken, NJ. Gutteling, B.M., de Weerth, C., Buitelaar, J.K., 2005. Prenatal stress and children’s cortisol reaction to the first day of school. Psychoneuroendocrinology 30 (6), 541e549. Halligan, S.L., Herbert, J., Goodyer, I., Murray, L., 2007. Disturbances in morning cortisol secretion in association with maternal postnatal depression predict subsequent depressive symptomatology in adolescents. Biol. Psychiatry 62 (1), 40e46. Hartman, S., Freeman, S.M., Bales, K.L., Belsky, J., 2018. Prenatal stress as a riskdand an opportunitydfactor. Psychol. Sci. 29 (4), 572e580. Heim, C., Newport, D.J., Mletzko, T., Miller, A.H., Nemeroff, C.B., 2008. The link between childhood trauma and depression: insights from HPA axis studies in humans. Psychoneuroendocrinology 33 (6), 693e710. Heinrichs, M., Baumgartner, T., Kirschbaum, C., Ehlert, U., 2003. Social support and oxytocin interact to suppress cortisol and subjective responses to psychosocial stress. Biol. Psychiatry 54 (12), 1389e1398. Henderson, H.A., Pine, D.S., Fox, N.A., 2015. Behavioral inhibition and developmental risk: a dual-processing perspective. Neuropsychopharmacology 40 (1), 207. Henry, C., Kabbaj, M., Simon, H., Le Moal, M., Maccari, S., 1994. Prenatal stress increases the hypothalamo-pituitary-adrenal axis response in young and adult rats. J. Neuroendocrinol. 6 (3), 341e345. Herman, J.P., 2017. Regulation of hypothalamo-pituitary-adrenocortical responses to stressors by the nucleus of the solitary tract/dorsal vagal complex. Cell. Mol. Neurobiol. 1e11. Herman, J.P., McKlveen, J.M., Ghosal, S., Kopp, B., Wulsin, A., Makinson, R., et al., 2016. Regulation of the hypothalamic-pituitary-adrenocortical stress response. Comprehensive Physiology 6 (2), 603. Ho, J.T., Lewis, J.G., O’Loughlin, P., Bagley, C.J., Romero, R., Dekker, G.A., Torpy, D.J., 2007. Reduced maternal corticosteroid-binding globulin and cortisol levels in pre-eclampsia and gamete recipient pregnancies. Clin. Endocrinol. 66 (6), 869e877. Hobel, C.J., Arora, C.P., Korst, L.M., 1999. Corticotrophin-releasing hormone and CRH-binding protein: differences between patients at risk for preterm birth and hypertension. Ann. N. Y. Acad. Sci. 897 (1), 54e65. Hoffman, M.C., Mazzoni, S.E., Wagner, B.D., Laudenslager, M.L., 2016. Measures of maternal stress and mood in relation to preterm birth. Obstet. Gynecol. 127 (3), 545. Hornstein, E.A., Fanselow, M.S., Eisenberger, N.I., 2016. A safe haven: investigating social-support figures as prepared safety stimuli. Psychol. Sci. 27 (8), 1051e1060. Hostinar, C.E., Johnson, A.E., Gunnar, M.R., 2015. Parent support is less effective in buffering cortisol stress reactivity for adolescents compared to children. Dev. Sci. 18 (2), 281e297. Hostinar, C.E., Sullivan, R.M., Gunnar, M.R., 2014. Psychobiological mechanisms underlying the social buffering of the hypothalamicepituitarye adrenocortical axis: a review of animal models and human studies across development. Psychol. Bull. 140 (1), 256. Howland, M.A., Sandman, C.A., Glynn, L.M., Crippen, C., Davis, E.P., 2016. Fetal exposure to placental corticotropin-releasing hormone is associated with child self-reported internalizing symptoms. Psychoneuroendocrinology 67, 10e17. Jaffe, R.B., Mesiano, S., Smith, R., Coulter, C.L., Spencer, S.J., Chakravorty, A., 1998. The regulation and role of fetal adrenal development in human pregnancy. Endocr. Res. 24 (3e4), 919e926.

582 PART | II Cognitive development

Joëls, M., Baram, T.Z., 2009. The neuro-symphony of stress. Nat. Rev. Neurosci. 10 (6), 459e466. https://doi.org/10.1038/nrn2632. Kane, H.S., Schetter, C.D., Glynn, L.M., Hobel, C.J., Sandman, C.A., 2014. Pregnancy anxiety and prenatal cortisol trajectories. Biol. Psychol. 100, 13e19. Kapoor, A., Dunn, E., Kostaki, A., Andrews, M.H., Matthews, S.G., 2006. Fetal programming of hypothalamo-pituitary-adrenal function: prenatal stress and glucocorticoids. J. Physiol. 572 (1), 31e44. Kapoor, A., Petropoulos, S., Matthews, S.G., 2008. Fetal programming of hypothalamicepituitaryeadrenal (HPA) axis function and behavior by synthetic glucocorticoids. Brain Res. Rev. 57 (2), 586e595. Kempná, P., Flück, C.E., 2008. Adrenal gland development and defects. Best Pract Res Clin Endocrinol Metab. 22 (1), 77e93. Kertes, D.A., Gunnar, M.R., Madsen, N.J., Long, J.D., 2008. Early deprivation and home basal cortisol levels: a study of internationally adopted children. Dev. Psychopathol. 20 (2), 473e491. Kertes, D.A., Kamin, H.S., Hughes, D.A., Rodney, N.C., Bhatt, S., Mulligan, C.J., 2016. Prenatal maternal stress predicts methylation of genes regulating the hypothalamicepituitaryeadrenocortical system in mothers and newborns in the Democratic Republic of Congo. Child Dev. 87 (1), 61e72. Kim, D.-J., Davis, E.P., Sandman, C.A., Sporns, O., O’Donnell, B.F., Buss, C., Hetrick, W.P., 2016. Prenatal maternal cortisol has sex-specific associations with child brain network properties. Cerebr. Cortex 27 (11), 5230e5241. Kinsella, M.T., Monk, C., 2009. Impact of maternal stress, depression & anxiety on fetal neurobehavioral development. Clin. Obstet. Gynecol. 52 (3), 425. Kirschbaum, C., Hellhammer, D.H., 1989. Salivary cortisol in psychobiological research: an overview. Neuropsychobiology 22 (3), 150e169. Korosi, A., Baram, T.Z., 2008. The central corticotropin releasing factor system during development and adulthood. Eur. J. Pharmacol. 583 (2e3), 204e214. Koss, K.J., Gunnar, M.R., 2018. Annual Research Review: early adversity, the hypothalamicepituitaryeadrenocortical axis, and child psychopathology. J. Child Psychol. Psychiatry 59 (4), 327e346. Koss, K.J., Hostinar, C.E., Donzella, B., Gunnar, M.R., 2014. Social deprivation and the HPA axis in early development. Psychoneuroendocrinology 50, 1e13. Kumsta, R., Schlotz, W., Golm, D., Moser, D., Kennedy, M., Knights, N., et al., 2017. HPA axis dysregulation in adult adoptees twenty years after severe institutional deprivation in childhood. Psychoneuroendocrinology 86, 196e202. Kuzawa, C., 2005. The origins of the developmental origins hypothesis and the role of postnatal environments: response to Koletzko. Am. J. Hum. Biol. 17 (5), 662e664. Laplante, D.P., Barr, R.G., Brunet, A., Du Fort, G.G., Meaney, M.L., Saucier, J.-F., et al., 2004. Stress during pregnancy affects general intellectual and language functioning in human toddlers. Pediatr. Res. 56 (3), 400. Lebel, C., Walton, M., Letourneau, N., Giesbrecht, G.F., Kaplan, B.J., Dewey, D., 2016. Prepartum and postpartum maternal depressive symptoms are related to children’s brain structure in preschool. Biol. Psychiatry 80 (11), 859e868. https://doi.org/10.1016/j.biopsych.2015.12.004. Leneman, K.B., Donzella, B., Desjardins, C.D., Miller, B.S., Gunnar, M.R., 2018. The slope of cortisol from awakening to 30 min post-wake in postinstitutionalized children and early adolescents. Psychoneuroendocrinology 96, 93e99. Levine, S., Wiener, S.G., 1988. Psychoendocrine aspects of mother-infant relationships in nonhuman primates. Psychoneuroendocrinology 13 (1e2), 143e154. Li, J., Wang, Z.-N., Chen, Y.-P., Dong, Y.-P., Shuai, H.-L., Xiao, X.-M., et al., 2012. Late gestational maternal serum cortisol is inversely associated with fetal brain growth. Neurosci. Biobehav. Rev. 36 (3), 1085e1092. Liston, C., Gan, W.-B., 2011. Glucocorticoids are critical regulators of dendritic spine development and plasticity in vivo. Proc. Natl. Acad. Sci. 108 (38), 16074e16079. Lyons, D.M., Parker, K.J., Schatzberg, A.F., 2010. Animal models of early life stress: implications for understanding resilience. Dev. Psychobiol. 52 (7), 616e624. Maccari, S., Piazza, P.V., Kabbaj, M., Barbazanges, A., Simon, H., Le Moal, M., 1995. Adoption reverses the long-term impairment in glucocorticoid feedback induced by prenatal stress. J. Neurosci. 15 (1), 110e116. MacMillan, H.L., Georgiades, K., Duku, E.K., Shea, A., Steiner, M., Niec, A., et al., 2009. Cortisol response to stress in female youths exposed to childhood maltreatment: results of the youth mood project. Biol. Psychiatry 66 (1), 62e68. Mairesse, J., Lesage, J., Breton, C., Bréant, B., Hahn, T., Darnaudéry, M., et al., 2007. Maternal stress alters endocrine function of the feto-placental unit in rats. Am. J. Physiol. Endocrinol. Metabol. 292 (6), E1526eE1533. McCall, J.G., Al-Hasani, R., Siuda, E.R., Hong, D.Y., Norris, A.J., Ford, C.P., Bruchas, M.R., 2015. CRH engagement of the locus coeruleus noradrenergic system mediates stress-induced anxiety. Neuron 87 (3), 605e620. McGrath, S., McLean, M., Smith, D., Bisits, A., Giles, W., Smith, R., 2002. Maternal plasma corticotropin-releasing hormone trajectories vary depending on the cause of preterm delivery. Am. J. Obstet. Gynecol. 186, 257e260. McLaughlin, K.A., Sheridan, M.A., Tibu, F., Fox, N.A., Zeanah, C.H., Nelson, C.A., 2015. Causal effects of the early caregiving environment on development of stress response systems in children. Proc. Natl. Acad. Sci. 112 (18), 5637e5642. McLean, M., Bisits, A., Davies, J., Woods, R., Lowry, P., Smith, R., 1995. A placental clock controlling the length of human pregnancy. Nat. Med. 1, 460e463. Monk, C., Feng, T., Lee, S., Krupska, I., Champagne, F.A., Tycko, B., 2016. Distress during pregnancy: epigenetic regulation of placenta glucocorticoidrelated genes and fetal neurobehavior. Am. J. Psychiatry 173 (7), 705e713. Mulligan, C., D’Errico, N., Stees, J., Hughes, D., 2012. Methylation changes at NR3C1 in newborns associate with maternal prenatal stress exposure and newborn birth weight. Epigenetics 7 (8), 853e857.

The effects of stress on early brain and behavioral development Chapter | 26

583

Murphy, V., Clifton, V., 2003. Alterations in human placental 11b-hydroxysteroid dehydrogenase type 1 and 2 with gestational age and labour. Placenta 24 (7), 739e744. Mustoe, A.C., Taylor, J.H., Birnie, A.K., Huffman, M.C., French, J.A., 2014. Gestational cortisol and social play shape development of marmosets’ HPA functioning and behavioral responses to stressors. Dev. Psychobiol. 56 (6), 1229e1243. Nathanielsz, P.W., 1999. Life in the Womb: The Origin of Health and Disease, first ed. Promethian Press, Ithaca, NY. Netherton, C., Goodyer, I., Tamplin, A., Herbert, J., 2004. Salivary cortisol and dehydroepiandrosterone in relation to puberty and gender. Psychoneuroendocrinology 29 (2), 125e140. Noorlander, C., De Graan, P., Middeldorp, J., Van Beers, J., Visser, G., 2006. Ontogeny of hippocampal corticosteroid receptors: effects of antenatal glucocorticoids in human and mouse. J. Comp. Neurol. 499 (6), 924e932. O’Connor, T., Heron, J., Glover, V., Team, A.S., 2002. Antenatal anxiety predicts child behavioral/emotional problems independently of postnatal depression. J. Am. Acad. Child Adolesc. Psychiatry 41 (12), 1470e1477. O’Connor, T.G., Ben-Shlomo, Y., Heron, J., Golding, J., Adams, D., Glover, V., 2005. Prenatal anxiety predicts individual differences in cortisol in preadolescent children. Biol. Psychiatry 58 (3), 211e217. O’Connor, T.G., Bergman, K., Sarkar, P., Glover, V., 2013. Prenatal cortisol exposure predicts infant cortisol response to acute stress. Dev. Psychobiol. 55 (2), 145e155. O’Connor, T.G., Monk, C., Fitelson, E.M., 2014. Practitioner review: maternal mood in pregnancy and child developmenteimplications for child psychology and psychiatry. J. Child Psychol. Psychiatry 55 (2), 99e111. O’Donnell, K.J., Jensen, A.B., Freeman, L., Khalife, N., O’Connor, T.G., Glover, V., 2012. Maternal prenatal anxiety and downregulation of placental 11b-HSD2. Psychoneuroendocrinology 37 (6), 818e826. Ouellet-Morin, I., Danese, A., Bowes, L., Shakoor, S., Ambler, A., Pariante, C.M., et al., 2011. A discordant monozygotic twin design shows blunted cortisol reactivity among bullied children. J. Am. Acad. Child Adolesc. Psychiatry 50 (6), 574e582 e573. Ouellet-Morin, I., Wong, C., Danese, A., Pariante, C., Papadopoulos, A., Mill, J., Arseneault, L., 2013. Increased serotonin transporter gene (SERT) DNA methylation is associated with bullying victimization and blunted cortisol response to stress in childhood: a longitudinal study of discordant monozygotic twins. Psychol. Med. 43 (9), 1813e1823. Pawlby, S., Hay, D., Sharp, D., Waters, C.S., Pariante, C.M., 2011. Antenatal depression and offspring psychopathology: the influence of childhood maltreatment. Br. J. Psychiatry 199 (2), 106e112. Peña, C.J., Monk, C., Champagne, F.A., 2012. Epigenetic effects of prenatal stress on 11b-hydroxysteroid dehydrogenase-2 in the placenta and fetal brain. PLoS One 7 (6), e39791. Phillips, D.A., Fox, N.A., Gunnar, M.R., 2011. Same place, different experiences: bringing individual differences to research in child care. Child Development Perspectives 5 (1), 44e49. https://doi.org/10.1111/j.1750-8606.2010.00155.x. Pluess, M., Belsky, J., 2011. Prenatal programming of postnatal plasticity? Dev. Psychopathol. 23 (1), 29e38. Porges, S.W., 2009. The polyvagal theory: new insights into adaptive reactions of the autonomic nervous system. Clevel. Clin. J. Med. 76 (Suppl. 2), S86. Posner, J., Cha, J., Roy, A., Peterson, B., Bansal, R., Gustafsson, H., . . Monk, C., 2016. Alterations in amygdalaeprefrontal circuits in infants exposed to prenatal maternal depression. Transl. Psychiatry 6 (11), e935. Pryce, C.R., 2008. Postnatal ontogeny of expression of the corticosteroid receptor genes in mammalian brains: inter-species and intra-species differences. Brain Res. Rev. 57 (2), 596e605. Reinisch, J.M., Simon, N.G., Karow, W.G., Gandelman, R., 1978. Prenatal exposure to prednisone in humans and animals retards intrauterine growth. Science 202 (4366), 436e438. Rice, F., Harold, G., Boivin, J., Van den Bree, M., Hay, D., Thapar, A., 2010. The links between prenatal stress and offspring development and psychopathology: disentangling environmental and inherited influences. Psychol. Med. 40 (2), 335e345. Rifkin-Graboi, A., Bai, J., Chen, H., Hameed, W.B., Sim, L.W., Tint, M.T., et al., 2013. Prenatal maternal depression associates with microstructure of right amygdala in neonates at birth. Biol. Psychiatry 74 (11), 837e844. https://doi.org/10.1016/j.biopsych.2013.06.019. Romeo, R.D., 2010. Pubertal maturation and programming of hypothalamicepituitaryeadrenal reactivity. Front. Neuroendocrinol. 31 (2), 232e240. Sanchez, M.M., Ladd, C.O., Plotsky, P.M., 2001. Early adverse experience as a developmental risk factor for later psychopathology: evidence from rodent and primate models. Dev. Psychopathol. 13 (3), 419e449. Sandman, C., Davis, E., 2010. Gestational stress influences cognition and behavior. Future Neurol. 5 (5), 675e690. Sandman, C.A., Buss, C., Head, K., Davis, E.P., 2015. Fetal exposure to maternal depressive symptoms is associated with cortical thickness in late childhood. Biol. Psychiatry 77 (4), 324e334. https://doi.org/10.1016/j.biopsych.2014.06.025. Sandman, C.A., Curran, M.M., Davis, E.P., Glynn, L.M., Head, K., Baram, T.Z., 2018. Cortical thinning and neuropsychiatric outcomes in children exposed to prenatal adversity: a role for placental CRH? Am. J. Psychiatry 175 (5), 471e479. Sandman, C.A., Davis, E.P., Glynn, L.M., 2012. Prescient human fetuses thrive. Psychol. Sci. 23 (1), 93e100. Sandman, C.A., Glynn, L., Wadhwa, P.D., Chicz-DeMet, A., Porto, M., Garite, T., 2003. Maternal hypothalamic-pituitary-adrenal disregulation during the third trimester influences human fetal responses. Dev. Neurosci. 25 (1), 41e49. Sandman, C.A., Glynn, L.M., Davis, E.P., 2013. Is there a viabilityevulnerability tradeoff? Sex differences in fetal programming. J. Psychosom. Res. 75 (4), 327e335. Sandman, C.A., Glynn, L.M., Schetter, C.D., Wadwha, P.D., Garite, T., Hobel, C., 2006. Elevated maternal cortisol early in pregnancy predicts third trimester levels of placental corticotropin releasing hormone (CRH): priming the placental clock. Peptides 27, 1457e1463.

584 PART | II Cognitive development

Sandman, C.A., Wadhwa, P.D., Chicz-DeMet, A., Porto, M., Garite, T.J., 1999. Maternal corticotropin-releasing hormone and habituation in the human fetus. Dev. Psychobiol. 34 (3), 163e173. Sapolsky, R.M., Romero, L.M., Munck, A.U., 2000. How do glucocorticoids influence stress responses? Integrating permissive, suppressive, stimulatory, and preparative actions. Endocr. Rev. 21 (1), 55e89. Schechter, J.C., Brennan, P.A., Smith, A.K., Stowe, Z.N., Newport, D.J., Johnson, K.C., 2017. Maternal prenatal psychological distress and preschool cognitive functioning: the protective role of positive parental engagement. J. Abnorm. Child Psychol. 45 (2), 249e260. Seckl, J.R., 2007. Glucocorticoids, developmental ‘programming’and the risk of affective dysfunction. Prog. Brain Res. 167, 17e34. Sidman, R.L., Rakic, P., 1973. Neuronal migration, with special reference to developing human brain: a review. Brain Research 62 (1), 1e35. Smith, A.S., Wang, Z., 2014. Hypothalamic oxytocin mediates social buffering of the stress response. Biol. Psychiatry 76 (4), 281e288. Smith, R., Nicholson, R.C., 2007. Corticotrophin releasing hormone and the timing of birth. Front. Biosci. 12, 912e918. Spanglar, G., Schieche, M., Ilg, U., Maier, U., Ackermann, C., 1994. Maternal sensitivity as an external organizer for biobehavioral regulation in infancy. Dev. Psychobiol. 27 (7), 425e437. Spann, M.N., Monk, C., Scheinost, D., Peterson, B.S., 2018. Maternal immune activation during the third trimester is associated with neonatal functional connectivity of the salience network and fetal to toddler behavior. J. Neurosci. 38 (11), 2877e2886. Spear, L.P., 2000. Neurobehavioral changes in adolescence. Curr. Dir. Psychol. Sci. 9 (4), 111e114. Stroud, L.R., Foster, E., Papandonatos, G.D., Handwerger, K., Granger, D.A., Kivlighan, K.T., Niaura, R., 2009. Stress response and the adolescent transition: performance versus peer rejection stressors. Dev. Psychopathol. 21 (1), 47e68. Tarabulsy, G.M., Pearson, J., Vaillancourt-Morel, M.-P., Bussières, E.-L., Madigan, S., Lemelin, J.-P., et al., 2014. Meta-analytic findings of the relation between maternal prenatal stress and anxiety and child cognitive outcome. J. Dev. Behav. Pediatr. 35 (1), 38e43. Tasker, J.G., Herman, J.P., 2011. Mechanisms of rapid glucocorticoid feedback inhibition of the hypothalamicepituitaryeadrenal axis. Stress 14 (4), 398e406. Thayer, Z.M., Kuzawa, C.W., 2014. Early origins of health disparities: material deprivation predicts maternal evening cortisol in pregnancy and offspring cortisol reactivity in the first few weeks of life. Am. J. Hum. Biol. 26 (6), 723e730. Thomason, M.E., Scheinost, D., Manning, J.H., Grove, L.E., Hect, J., Marshall, N., et al., 2017. Weak functional connectivity in the human fetal brain prior to preterm birth. Sci. Rep. 7, 39286. Trickett, P.K., Noll, J.G., Susman, E.J., Shenk, C.E., Putnam, F.W., 2010. Attenuation of cortisol across development for victims of sexual abuse. Dev. Psychopathol. 22 (1), 165e175. Turecki, G., Meaney, M.J., 2016. Effects of the social environment and stress on glucocorticoid receptor gene methylation: a systematic review. Biol. Psychiatry 79 (2), 87e96. Ulrich-Lai, Y.M., Herman, J.P., 2009. Neural regulation of endocrine and autonomic stress responses. Nat. Rev. Neurosci. 10 (6), 397e409. https:// doi.org/10.1038/nrn2647. Van den Bergh, B.R., Marcoen, A., 2004. High antenatal maternal anxiety is related to ADHD symptoms, externalizing problems, and anxiety in 8-and 9year-olds. Child Dev. 75 (4), 1085e1097. Van den Bergh, B.R., Van Calster, B., Smits, T., Van Huffel, S., Lagae, L., 2008. Antenatal maternal anxiety is related to HPA-axis dysregulation and self-reported depressive symptoms in adolescence: a prospective study on the fetal origins of depressed mood. Neuropsychopharmacology 33 (3), 536. van der Vegt, E.J., Van Der Ende, J., Kirschbaum, C., Verhulst, F.C., Tiemeier, H., 2009. Early neglect and abuse predict diurnal cortisol patterns in adults: a study of international adoptees. Psychoneuroendocrinology 34 (5), 660e669. Vermeer, H.J., van IJzendoorn, M.H., 2006. Children’s elevated cortisol levels at daycare: a review and meta-analysis. Early Child. Res. Q. 21 (3), 390e401. Wakshlak, A., Marta, W., 1990. Neonatal handling reverses behavioral abnormalities induced in rats by prenatal stress. Physiol. Behav. 48 (2), 289e292. Yehuda, R., 2001. Biology of posttraumatic stress disorder. J. Clin. Psychiatry 61 (Suppl. 7), 15e21. Yehuda, R., Engel, S.M., Brand, S.R., Seckl, J., Marcus, S.M., Berkowitz, G.S., 2005. Transgenerational effects of posttraumatic stress disorder in babies of mothers exposed to the World Trade Center attacks during pregnancy. J. Clin. Endocrinol. Metab. 90 (7), 4115e4118. Yim, I.S., Glynn, L.M., Schetter, C.D., Hobel, C.J., Chicz-DeMet, A., Sandman, C.A., 2009. Risk of postpartum depressive symptoms with elevated corticotropin-releasing hormone in human pregnancy. Arch. Gen. Psychiatr. 66 (2), 162e169.

Chapter 27

Sex differences in brain and behavioral development Adriene M. Beltz1, Dominic P. Kelly1 and Sheri A. Berenbaum2 1

University of Michigan, Ann Arbor, MI, United States; 2The Pennsylvania State University, University Park, PA, United States

Chapter outline 27.1. Introduction 27.1.1. Issues in studying sex differences 27.1.2. Interpreting sex differences 27.2. Psychological sex differences: nature and development 27.2.1. Cognitive skills 27.2.1.1. Spatial skills 27.2.1.2. Mathematical skills 27.2.1.3. Verbal skills 27.2.1.4. Memory 27.2.1.5. Perceptual speed 27.2.2. Noncognitive sex differences 27.2.2.1. Gender identity 27.2.2.2. Sexual orientation 27.2.2.3. Physical and motor skills 27.2.2.4. Activity interests 27.2.2.5. Temperament and personality 27.2.2.6. Social behaviors 27.2.2.7. Psychological disorders 27.3. Explanations for psychological sex differences 27.3.1. Socialization perspectives 27.3.1.1. Socialization of cognitive sex differences 27.3.1.2. Socialization of noncognitive sex differences 27.3.2. Genetic perspectives 27.3.3. Hormone perspectives 27.3.3.1. Evidence for hormone influences on nonhuman sex-typed behavior 27.3.3.2. Early hormone influences on human behavior 27.3.3.3. Adolescent hormone influences on human behavior 27.3.3.4. Circulating hormone influences on human behavior 27.3.3.5. Exogenous hormone influences on human behavior 27.3.4. Integrated perspectives

586 586 587 587 587 588 588 589 589 590 590 590 591 591 591 592 592 593 593 593 593 594 595 595 595 596 598 599 599 599

Neural Circuit and Cognitive Development. https://doi.org/10.1016/B978-0-12-814411-4.00027-5 Copyright © 2020 Elsevier Inc. All rights reserved.

27.4. Brain sex differences: nature, development, and consequences 600 27.4.1. Issues in studying the brain 600 27.4.2. Sex differences in brain structure and their development 601 27.4.2.1. Brain volume 601 27.4.2.2. Regional structure volume 602 27.4.2.3. Gray matter 604 27.4.2.4. White matter 605 27.4.2.5. Implications of sex differences in brain structure 605 27.4.3. Sex differences in brain function (activation) 606 27.4.3.1. Lateralization 606 27.4.3.2. Spatial skills 607 27.4.3.3. Language 608 27.4.3.4. Emotion-related processing 609 27.4.3.5. Functional connectivity 610 27.4.3.6. Development of sex differences in brain function 610 27.4.3.7. Implications of sex differences in brain function 611 27.5. Explanations for brain sex differences 613 27.5.1. Socialization perspectives 613 27.5.2. Genetic perspectives 614 27.5.3. Hormone perspectives 615 27.5.3.1. Prenatal hormone influences on brain sex differences 615 27.5.3.2. Adolescent hormone influences on brain sex differences 615 27.5.3.3. Circulating hormone influences on brain sex differences 616 27.5.3.4. Exogenous hormone influences on brain sex differences 617 27.6. Conclusions and future directions 618 Acknowledgments 620 References 620

585

586 PART | II Cognitive development

27.1 Introduction Psychological and neural sex differences are a topic of great interest for their relevance to both applied and basic science questions. From an applied perspective, they figure into discussions ranging from women’s underrepresentation in science and mathematics careers (Ceci and Williams, 2010; Halpern et al., 2007) to the malleability (or lack thereof) of gender identity and sexual orientation (Aitken et al., 2015; Callens et al., 2016) to variability in incidences and forms of psychopathology (Hartung and Lefler, 2019; Zahn-Waxler et al., 2008). From a basic science perspective, sex represents an important dimension of individual difference that marks social, genetic, and hormonal processes; thus, factors that lead to differences between the sexes can help to understand mechanisms underlying between-person variations generally. The focus of this chapter is sex differences in the human brain and behavior, with an emphasis on cognitive skills; the goal is to describe what is known and hypothesized about the differences, their development, and their etiology and to begin to link brain differences to cognitive differences. The chapter is organized into several sections. First, evidence concerning psychological sex differences in human beings and the ways in which they develop across age is summarized, focusing on changes from infancy through young adulthood. Second, the main theories that have been offered to explain the differences and the evidence that supports those perspectives are considered. Third, evidence concerning structural and functional sex differences in the human brain and their development is presented and synthesized. Fourth, the growing body of research about the mechanisms underlying brain sex differences is reviewed. Fifth, knowledge gaps are highlighted, and suggestions for further research are provided.

27.1.1 Issues in studying sex differences There is some controversy over the study of sex differences. While the National Institutes of Health have recently mandated that sex be considered as a biological variable in basic research (McCarthy et al., 2017; Miller et al., 2017), others have argued that sex differences do not exist or are even unethical to study, presumably because they endorse sexism (Jordan-Young and Rumiati, 2012). When sex differences are studied, there is also debate over the size of the differences and whether they have declined over time (Hyde, 2005; Miller and Halpern, 2014). The controversy relates to several factors, including assumptions made about innateness and malleability, study methodology, and inferences made from sex differences research. Regarding assumptions, there is an inaccurate, but surprisingly widespread notion that gender differences in behavior reflect universal and immutable differences between boys and men and girls and women, especially if the differences are linked to biology, such as genes, sex hormones, or neural mediation (e.g., Eliot and Richardson, 2016; Fine, 2017; Jordan-Young and Rumiati, 2012). Conversely, there is an assumption that gender-related variation is the sole result of flexible social or cultural influences (Hyde et al., 2019). Neither assumption is logically true nor accurately represents the perspective of gender scientists and sex differences research (Berenbaum, 2018; Eagly, 2018; McCarthy, 2016). Consider two examples. Hair color is genetically determined and easily changeable through environmental influences, whereas racism is a cultural phenomenon and remarkably resistant to remediation (Bonilla-Silva, 2017). Regarding methodology, there are critiques that sex differences research is not replicable and ignores within-sex variability (Fausto-Sterling, 2008; Fine, 2010; Joel and Fausto-Sterling, 2016). While publication bias likely contributes to an over-estimation of the number and size of brain and behavioral sex differences (Button et al., 2013; David et al., 2018), this (unfortunately) is not unique to this area of research; for instance, similar arguments characterize medical research (Ioannidis, 2016). Moreover, this criticism can be significantly offset by focusing on properly conducted metaanalyses and replicated results from well-powered studiesdas will be done in this review. Moreover, the detection of average differences between the sexes does not mean the sexes are dimorphic or represent two (and only two), nonoverlapping patterns of behavior, function, or morphology. In fact, the description of sex differences used here and throughout the literature is effect size, d (Cohen, 1988), which is expressed in standard deviation units and reflects the percent of overlap between the distributions of girls and boys or men and women: small (d w 0.2, 85% overlap in distributions of the sexes), moderate (d w 0.5, 67% overlap, probably noticeable), and large (d w 0.8, 53% overlap, very noticeable). Regarding inferences, some researchers claim that sex differencesdeven if they do exist and reflect overlap between the sexesdare too small to be meaningful (Halpern et al., 2011; Hyde, 2005). But sex differences can be among the largest effects in psychology. The effects described throughout this chapter are compared with typical effects in psychology and medicine described elsewhere (e.g., Meyer et al., 2001), which is illustrated in Fig. 27.1. Most sex differences are not small, and even small effects, particularly when consistently found, are important for scientific understanding and innovation.

Sex differences in brain and behavioral development Chapter | 27

587

FIGURE 27.1 Psychological sex differences in behavior compared with other health and psychological effects, described in standard deviation units, d. Data are from Hyde, 2005.

27.1.2 Interpreting sex differences A sex difference reflects a statistically significant, average difference between girls and boys or women and men in a specific behavioral or biological domain. There is variability both within and between the sexes in that domain, and gender is multidimensional (Ruble et al., 2006), so findings do not necessarily generalize across domains. In other words, a given sex difference may not describe every individual (i.e., an individual girl can outperform the “average” boy), and just because an individual is gender-typical (i.e., well-characterized by the average for their sex) in one domain does not mean that they will be equally gender-typical in another domain. This complexity is partly why the study of sex differences is interesting and scientifically valuable. Sex reflects the integration of social (e.g., culture, parents, peers, and aspects of the environment) and biological (e.g., genes, hormones) influences on an individual systemdall of which are complexly mediated by brain structure and function. Thus, the detection of sex differences fuels discovery (Klein et al., 2015), providing insight into the factors that lead to individual differences generally and ultimately painting a detailed picture about the uniqueness of each individualdfacilitating precision healthcare and education.

27.2 Psychological sex differences: nature and development Sex differences are found in many psychological domains, from sensory thresholds to parenting practices. The focus here is on cognitive sex differences, but other significant psychological domains in which the sexes differ in behavior are noted. The goal is not necessarily to provide an exhaustive review of the extensive literature, but rather to provide a summary of the findings, highlighting the key differences and, in later sections, explanations of the differences and the links between cognitive and brain sex differences. Detailed reviews, with supporting references, are available elsewhere, as cited throughout the chapter.

27.2.1 Cognitive skills The sexes do not differ in overall intelligence, but as detailed in the following, they do differ in their pattern of cognitive skills: On average, boys and men are better than girls and women in spatial and some mathematical tasks, whereas girls and women are better than boys and men in many verbal tasks, some aspects of memory, and processing speed (discussed in Camarata and Woodcock, 2006; Halpern, 2012; Roivainen, 2011). Fig. 27.2 provides a summary of the cognitive sex differences expressed in standard deviation units, d. In fact, an analysis of the structure of intelligence shows that there are three sex-typed dimensions (Johnson and Bouchard, 2007): a rotationeverbal dimension (with spatial skills at one end and verbal tasks at the other), a focuse diffusion dimension (with focused attention on a single stimulus on one end and diffuse attention on multiple stimuli on the other), and memory. Men were found to be nearer the rotation and focus ends, and women to be nearer the verbal and diffusion ends; women also had better memory.

588 PART | II Cognitive development

FIGURE 27.2 Sample sex differences in cognition: Differences are expressed in standard deviation units, d, with red bars reflecting characteristics with higher scores for females than males and blue bars reflecting characteristics with higher scores for males than females. Asterisks (*) indicate results from meta-analyses. Data are from Beltz and Berenbaum, 2013; Blakemore et al., 2009; Maeda and Yoon, 2013, 2016; Maylor et al., 2007; Reilly et al., 2015; Voyer, 2011; Voyer and Voyer, 2014.

27.2.1.1 Spatial skills One widely studied cognitive sex difference is spatial ability. There are several ways to parse the domain, but boys and men outperform girls and women in most aspects, with the size of the difference depending upon the aspect (Halpern, 2012; Lawton, 2010). The largest sex difference is in mental rotation, especially rotation of objects in three dimensions (Maeda and Yoon, 2013; Miller and Halpern, 2014) when under strict time constraints (Voyer, 2011). This sex difference is apparent beginning in infancy (Moore and Johnson, 2008; Quinn and Liben, 2008), with boys having better skills than girls and with the difference increasing slightly from childhood to adulthood (Geiser et al., 2008). There are moderate-sized sex differences in spatial perception and the ability to identify spatial relations with respect to one’s body in reference to external space or to identify the true vertical or horizontal (Halpern, 2012). This is measured by tasks such as the “rod and frame task” (Voyer and Bryden, 1993) and the “water level task” (Vasta and Liben, 1996). There are also sex differences in skills related to navigating in the real world (Lawton, 2010), typically referred to as “wayfinding.” Boys and men are better than girls and women at remembering and navigating to distant locations in large spaces; some of the difference results from sex differences in strategy, with men primarily relying on cardinal (north, south, east, and west) directions and women on landmarks. Considering larger issues of the spatial environment, there is a huge sex disparity among National Geography Bee winners (despite equal participation from boys and girls); the sex ratio increases at each level of competition, so that in many years, all 10 finalists are boys (Liben, 1995). There is one spatial skill for which the sex difference is reversed: memory for spatial location. Girls and women are better than boys and men in remembering the location of objects (Voyer et al., 2007).

27.2.1.2 Mathematical skills The sexes differ in quantitative skills, with the differences again varying by aspect and age (Halpern, 2012). The differences also tend to be smaller than those in spatial skills. Thus, it is not surprising that meta-analyses that combine across aspects of mathematics and ages disagree on the direction of the effect (e.g., Lindberg et al., 2010; Voyer and Voyer, 2014).

Sex differences in brain and behavioral development Chapter | 27

589

In school, girls get better grades than do boys in arithmetic and mathematics classes (as they do in all classes); teachers’ evaluations especially indicate that girls perform better than boys (Robinson and Lubienski, 2011). On standardized tests of mathematics knowledge and skills (e.g., the Standardized Achievement Test and National Assessment of Educational Progress), however, boys outperform girls (Reilly et al., 2015). Historically, the main sex difference was thought to be in problem-solving tasks, with boys outperforming girls, especially at older ages, but recent data show the sex difference to be small and to vary across nations. There is some indication that national differences are related gender equality across countries (Else-Quest et al., 2010; Guiso et al., 2008), although this has been disputed (Stoet and Geary, 2013). The sexes also differ in their variability in mathematical (and spatial) skills (Hedges and Nowell, 1995). For example, for the past 20 years, about four times as many seventh-grade boys as girls have scored in the high ranges of the SATMathematics (Wai et al., 2010); before that time, there were even higher ratios of boys to girls at the top end. Given the different patterns of sex differences in mathematics grades and standardized test scores, it is unclear the extent to which sex differences in mathematics may reflect other factors, such as anxiety. Indeed, mathematics anxiety has been shown to be greater for girls than boys and is a sex-specific predictor of performance (Devine et al., 2012).

27.2.1.3 Verbal skills There are many kinds of verbal and language-related abilities, such as vocabulary size, use of correct grammar, reading, doing anagrams, and following verbal instructions. The sexes are similar on some skills. When there are differences, they are generally in the direction of girls and women having better skills than boys and men, although the size of the difference varies with age (Halpern, 2012). Language learning occurs earlier in girls than in boys, with boys catching up by age 6 years (Bornstein et al., 2004; Wallentin, 2009). Language disorders are more common in boys than in girls (Wallentin, 2009). In terms of specific verbal skills, females have a small to moderate advantage over males in several skills, most prominently in reading comprehension, verbal fluency, and phonological processing, but males have a small edge in analogies (Halpern, 2012). In terms of verbal skills that are particularly important for formal education, teachers’ evaluations throughout elementary and middle school consistently rate girls’ verbal skills as superior to those of boys (Robinson and Lubienski, 2011), and meta-analytic results show that girls outperform boys (with a moderate effect size) in language courses (Petersen, 2018; Voyer and Voyer, 2014). There is also great concern about boys’ lag in reading performance (Chudowsky and Chudowsky, 2010). For instance, in the Programme for International Student Assessment, which includes data from 1.5 million 15-year-olds, girls outperform boys in reading in every one of the 75 countries assessed (Stoet and Geary, 2013). Boys write less well than girls, too, and this may be especially notable at the highest skill levels (Hedges and Nowell, 1995; Pargulski and Reynolds, 2017; Petersen, 2018; Wai et al., 2010). Girls and women also process verbal materials more rapidly than boys and men, including intelligence scale subtests, verbal fluency tests asking for many words to be generated quickly, letters of the alphabet, and digits (Camarata and Woodcock, 2006; Roivainen, 2011). Indeed, faster processing speed (combined with an association-based strategy; e.g., Scheuringer and Pletzer, 2017) may be a key contributor to the superior reading and writing skills of girls and women.

27.2.1.4 Memory As with other cognitive domains, there are several aspects of memory. In some, the sexes do not differ. But in several, girls and women are better than boys and men, with differences that are small to moderate in size: Compared with boys and men, girls and women show more accurate recall for learning facts or material that they read, more readily learn lists of words, have better recall for lists of common objects such as animals, food, furniture, and appliances, and have better recognition memory (Gur et al., 2012; Halpern, 2012; Johnson and Bouchard, 2007). The differences may be due in part to sex differences in strategy, with girls and women more likely to use clustering than boys and men. Sex differences in memory are especially consistent for episodic memory, that is, memory of specific events and episodes (Asperholm et al., 2019a). For example, women are better than men at recognizing faces that they have seen before. Also, the types of events remembered seem to differ by sex, with women remembering more negative and less positive autobiographical experiences compared with men (Young et al., 2013). Many episodic memories are verbally based, and there is emerging evidence of heightened female superiority of this skill in cultures with high levels of female education and employment (Asperholm et al., 2019b), but women have better episodic memories even when the tasks are not verbal (e.g., face recognition; Herlitz and Loven, 2013). Women also do better on episodic tasks, which require both verbal and visuospatial memory, but not when the tasks are clearly spatial, in which case, men do better (Herlitz et al., 2010; Herlitz and Rehnman, 2008). Importantly, these sex differences seem to be consistent across age (Herlitz and Rehnman, 2008).

590 PART | II Cognitive development

Results vary when considering other aspects of memory. For example, women recall the identities and locations of objects better than men, although this might be due to women’s better memories overall (Voyer et al., 2007). Moreover, research on working memory does not point to clear sex differences, perhaps because findings depend upon the stimuli to be remembered and task difficulty (e.g., Gur et al., 2012; Reed et al., 2017), or more likely because working memory is a distinct kind of memory (e.g., part of executive function that should be separately studied; Diamond, 2013).

27.2.1.5 Perceptual speed In addition to faster processing of verbally based cognitive tasks, girls and women are faster than boys and men in perceptual speed tasks (Burns and Nettelbeck, 2005; Daseking et al., 2017), which are included in most intelligence tests. In general, these tasks concern the ability to perceive details and shift attention quickly, often while using fine motor skills such as finger movements (Halpern, 2012). The difference ranges from small to large, depending on the particular measure.

27.2.2 Noncognitive sex differences There are many other psychological domains in which the sexes differ, but the focus here is on those that have received the most attention, that are relatively large in size, or that illustrate important points about studying sex differences (e.g., the importance of developmental status, measurement, and social context). Fig. 27.3 provides a summary of the noncognitive sex differences expressed in standard deviation units, d.

27.2.2.1 Gender identity The biggest sex differencedamong all cognitive and noncognitive psychological characteristicsdis in gender identity. Boys and men typically identify as male, whereas girls and women typically identify as female (Blakemore et al., 2009).

FIGURE 27.3 Sample sex differences in noncognitive characteristics: Differences are expressed in standard deviation units, d, with red bars reflecting characteristics with higher scores for females than males and blue bars reflecting characteristics with higher scores for males than females. Asterisks (*) indicate results from meta-analyses. Data are from Berenbaum, 1999; Berenbaum and Snyder, 1995; Blakemore et al., 2009; Bleidorn et al., 2016; Butterfield et al., 2012; Costa et al., 2001; Cross et al., 2011, 2013; Else-Quest et al., 2006; Fabes et al., 2003; Hoff et al., 2018; Su et al., 2009; Weisberg et al., 2011.

Sex differences in brain and behavioral development Chapter | 27

591

As are most gendered phenomena, gender identity is multidimensional, with different aspects having different psychological antecedents, correlates, and consequences (Egan and Perry, 2001; Lippa, 2008), and there is recent work highlighting the value of considering degree of identification with both sexes (Martin et al., 2017). There are members of the population whose gender identity does not match their biological sex, including gender queer and nonconforming youth. There is considerable question and controversy over the nature and causes of transgender identity (see Korpaisarn and Safer, 2019; Polderman et al., 2018; Turban and Ehrensaft, 2018), and representative, rigorous research is needed. For instance, gender dysphoria (distress) is difficult to determine, partly because it is highly dependent upon sampling and measurement (e.g., clinical referrals vs. surveys). Based solely on diagnosis rates and referrals for hormone treatment in adults, 0.005%e0.014% of biological males and 0.002%e0.003% of biological females experience gender dysphoria (American Psychiatric Association, 2013), but this is likely a significant underestimate. Moreover, rates of clinical referrals for gender dysphoria have historically been higher for males than females (Crissman et al., 2017; Eklund et al., 1988; Zucker, 2017; Zucker et al., 1997), but there is indication that this is no longer the case, particularly in late childhood and adolescence, and that girls may even be more likely to experience gender dysphoria (Kaltiala-Heino et al., 2015; Steensma et al., 2018).

27.2.2.2 Sexual orientation The second largest sex differencedamong all cognitive and noncognitive psychological characteristicsdis in sexual orientation. Most men and women have heterosexual orientation, with men being primarily attracted to women and women being primarily attracted to men (Blakemore et al., 2009). Not all individuals have exclusively heterosexual orientations or sexual experiences. By some counts, 8.2% of the adult population in the United States has had a same-sex sexual experience, and 11% have been attracted to a member of the same sex (Gates, 2011). Moreover, an estimated 1.7% of the adult population in the United States identifies as gay or lesbian, and an additional 1.8% identifies as bisexual (Gates, 2011). There are sex differences in these rates, however, with women being more likely than men to consider themselves bisexual and to report changes in their sexual attraction and identities (Bailey et al., 2016; Diamond, 2008; Diamond et al., 2017; Gates, 2011; Kinnish et al., 2005). There is also evidence of sex differences in sexual arousal and acts regardless of sexual orientation. For instance, men show greater arousal to erotic images of their preferred sex (i.e., heterosexual men to women and homosexual men to men), but womendboth heterosexual and homosexualdare similarly aroused by male and female erotic images (Chivers, 2017), consistent with other data suggesting greater sexual fluidity in women than men (Diamond, 2008). Moreover, men are more likely to engage in casual sex, masturbation, and pornography use (Petersen and Hyde, 2010). Given these sex differences in sexual behavior, some scholars have argued that sexual orientation has, to some degree, different determinants in women and men (Diamond, 2014).

27.2.2.3 Physical and motor skills Sex differences in physical and motor skills vary considerably by aspect and age. There are few sex differences in early milestones of reaching, sitting, crawling, and walking, but differences in motor skills begin to appear in the second year and continue through preschool (Kokstejn et al., 2017; Mondschein et al., 2000). The earlier neurological development of girls compared with boys results in early development of some skills, such as eyeehand coordination. Overall, girls develop sooner and are initially superior to boys with respect to fine motor skills, whereas boys tend to have greater strength and better gross motor skills (Flatters et al., 2014; Largo et al., 2001a, 2001b), which combine with boys’ advantage in spatial skills to result in large sex differences in targeting, that is, hitting a target with a ball and other object control skills (Butterfield et al., 2012; Flatters et al., 2014; Kimura, 1999; Yang et al., 2015). There are, however, some exceptions to this, such as catching, in which there is evidence for a female advantage (Butterfield et al., 2012). Boys are also more physically active than girls in childhood and adolescence (Breslin et al., 2012; Else-Quest et al., 2006; Mota and Esculcas, 2002).

27.2.2.4 Activity interests Beyond gender identity and sexual orientation, one of the largest and most important sex differences concerns interests (reviewed in Blakemore et al., 2009; Ruble et al., 2006). In childhood, boys and girls prefer and engage with different toys (e.g., trucks and dolls) and participate in different activities (e.g., playing sports and playing dress-up). In adolescence, boys and girls continue to prefer and participate in different leisure activities (e.g., building things and dance), household chores (e.g., taking out the garbage and preparing food), and academic pursuits (e.g., math and language arts).

592 PART | II Cognitive development

This continues into adulthood and extends to career outcomes, as women and men are differentially represented in different occupations; for example, men prefer occupations that involve working with things, and women prefer occupations that involve working with people (Su et al., 2009). Meta-analytic results indicate that sex differences in career interests peak in early adolescence, owing to boys’ low interest in socially oriented careers, and then differences decline (Hoff et al., 2018). Although the sexes overlap in their interests, the average differences are large.

27.2.2.5 Temperament and personality Overall, most aspects of temperament show small or no sex differences (Else-Quest et al., 2006). The main exception is effortful control, which shows a moderate advantage for girls. In addition, boys have higher levels of surgency than girls, perhaps reflecting their greater activity level. Research on “the big five” personality factors (neuroticism, extroversion, openness, agreeableness, and conscientiousness) shows that sex differences emerge in adolescence, leading to small, but relatively consistent, differences between adult men and women (De Bolle et al., 2015). It has been reported that women score higher than men in neuroticism (especially the anxiety and depression subscales), agreeableness, and extroversion (Costa et al., 2001; Soto et al., 2011; Weisberg et al., 2011). Cross-cultural similarities in the patterns of sex differences have also been found, but the differences may be largest in European and North American countries and smallest in African and Asian countries (Costa et al., 2001). There are also small sex differences favoring boys and men in global self-esteem, especially after childhood (Kling et al., 1999), although there appear to be some cross-cultural differences in the size of the effect, in which differences were pronounced in wealthy, individualistic, and egalitarian cultures (Bleidorn et al., 2016). Sex differences vary across aspects of self-esteem, with males having consistently higher appearance-related, personal self, self-satisfaction, and athletic selfesteem and females having higher ethical and behavioral conduct self-esteem (Gentile et al., 2009).

27.2.2.6 Social behaviors Stereotypes about sex differences in social behaviordthat girls and women are emotionally expressive and perceptive, and more attuned to babies, and that boys and men take risks and are aggressivedhave some basis in evidence, but the differences are not large and depend on context (Blakemore et al., 2009). Regarding emotional expression, there is some suggestion that women have more nonverbal emotion expression than men (Fischer and LaFrance, 2015) and that boys and men express more anger and less fear and sadness than girls and women; more convincing, however, is evidence that boys come to hide emotions such as sadness and fear purposefully, especially in the presence of peers (Chaplin, 2015; Kyratzis, 2001). Regarding perceiving emotion in others, girls and women are somewhat more sympathetic, empathic, and accurate at decoding emotions than are boys and men (ChristovMoore et al., 2014; Thompson and Voyer, 2014), but the differences are small and depend on how they are measured; for instance, they are larger on self-report than in physiological responses (Eisenberg et al., 2006; McClure, 2000). Girls also show more interest in and have more nurturant interactions with babies than boys do, although the size of the difference varies across situations (summarized in Blakemore et al., 2009). Regarding risk taking, boys are more likely than girls to take risks and to be injured (Byrnes et al., 1999; Morrongiello and Matheis, 2007). Risk-taking behavior is relatively normative in adolescence (Steinberg, 2010), but the sex difference is generally seen in adulthood, as men report greater sensation seeking, disinhibition, boredom susceptibility, and adventure seeking than do women (Cross et al., 2013). Aggression is one social behavior in which sex differences are well known. Aggression is defined as behavior intended to harm others and includes physical, verbal, and indirect aggression. Males are more physically and verbally aggressive than females from childhood through adulthood (Björkqvist, 2018; Card et al., 2008; Dodge et al., 2006). Although the social context affects aggression, and the size of the difference varies across age and culture, there are no instances in which females are more directly aggressive than males. Sometimes, physical aggression moves into the domain of seriously antisocial or criminal behavior. Although few people show the type of high levels of physical aggression that would be called violent or antisocial, or at the very extreme, commit murder, the majority of those who do are male, in childhood, adolescence, and adulthood (Archer, 2004; Dodge et al., 2006; Moffitt et al., 2001; Stone, 2015). Another form of aggression is called indirect, social, or relational aggression (e.g., Crick, 1995). This form of aggression includes manipulating social relationships or purposefully excluding others. It has been seen as the “feminine” form of aggression and is often reported to be more common in girls (see discussion in Blakemore et al., 2009). However, a meta-analysis (Card et al., 2008) found no significant differences between boys and girls in indirect aggression from childhood through adolescence, and some subsequent reports are consistent with this (e.g., Sánchez-Martín et al., 2011).

Sex differences in brain and behavioral development Chapter | 27

593

27.2.2.7 Psychological disorders Many forms of serious behavioral problems and mental illness occur at different rates in the two sexes. There are several reviews on this topic (Hartung and Lefler, 2019; Zahn-Waxler, 1993; Zahn-Waxler et al., 2008), so only key differences are summarized here. Psychological problems are often described as being internalizing (e.g., anxiety and depression) or externalizing (e.g., conduct disorder, antisocial behavior, including, among other things, criminal acts and excessive aggression and attention-deficit/hyperactivity disorder). The incidence of internalizing problems is higher in girls and women, whereas the incidence of externalizing problems is higher in boys and men (Eaton et al., 2012; Zahn-Waxler et al., 2008). Disorders that are predominant in males tend to have their onset in childhood, whereas the female preponderance of depression, anxiety, and eating disorders begins at puberty (Hartung and Widiger, 1998; Martel, 2013; Nolen-Hoeksema and Hilt, 2009). Regarding internalizing problems, females tend to be diagnosed with depressive disorders at earlier ages than males, consistent with the adolescent emergence of the sex difference (Faravelli et al., 2013; Schuch et al., 2014). There is also evidence that the risk factors for depression show sex differences, with low self-esteem pathways in boys and men, but interpersonal pathways in girls and women (Kendler and Gardner, 2014). Finally, internalizing problems, especially anxiety, seem to have a greater burden on women than men, with affected women reporting more work absences and medical visits than affected men (McLean et al., 2011). Regarding externalizing problems, boys and men have higher incidences throughout life, but adolescence is a particularly important developmental period. For example, conduct disorder is diagnosed more often in adolescent boys than girls (Merikangas et al., 2010), perhaps due to sex differences in impulsivity and its antecedents; in boys more than in girls, impulsivity is thought to result from high sensation-seeking tendencies coupled with weak motivational control (Cross et al., 2011). Not all adolescent impulsivity is indicative of conduct disorder or later criminal activity, however. Some adolescent risk taking is normative (Steinberg, 2010), and some antisocial behavior is limited to adolescence (Moffitt, 1993; Moffitt et al., 2002). Boys are also more likely than girls to have learning and reading disabilities (Blakemore et al., 2009) and to be diagnosed with autism spectrum disorder (Werling and Geschwind, 2013) and attention-deficit/hyperactivity disorder (Ramtekkar et al., 2010; Willcutt, 2012). These conditions tend to be diagnosed earlier in life for boys than girls (Begeer et al., 2013). There is some speculation, however, that sex differences in diagnosis frequency and age may be linked to sex differences in disorder presentation, so future work is needed (Mandy et al., 2012; Skogli et al., 2013; Werling and Geschwind, 2013).

27.3 Explanations for psychological sex differences Hypothesized causes of psychological sex differences tend to focus on either genetic and biological factors or on social and cultural factors. Key theoretical perspectives focusing on proximal processes are considered here, with a particular emphasis on early hormonal contributions to sex differences. Readers are referred elsewhere for discussion and critiques of evolutionary explanations of cognitive and neural sex differences that focus on distal processes (e.g., sexual selection; Eagly and Wood, 1999; Geary, 1998; Hannagan, 2008; Wood and Eagly, 2002).

27.3.1 Socialization perspectives Most work on psychological sex differences comes from a socialization perspective, that is, that sex differences develop as an individual navigates, observes others, is socialized, and internalizes information about the social world. There are several types of socialization theories, differing in the extent to which they emphasize the role of basic social learning mechanisms, subtle socialization practices, social identity (as male or female), cognitive schemas derived from gender identity that guide behavior, and the role of gendered social roles and resulting stereotypes and expectancies.

27.3.1.1 Socialization of cognitive sex differences Socialization perspectives have been applied less to cognition than to other sex-typed characteristics. Nevertheless, there is evidence for the importance of social influences on sex-typed skills. Sex differences in spatial skills have been seen to depend on socioeconomic status (SES), with differences apparent in children from middle and high SES backgrounds, but not in children from low SES backgrounds (Levine et al., 2005). Such SES effects were suggested to result in part from access to experiences that facilitate spatial skills. It is important to note, however, that findings have not been widely replicated, and cross-national data show that the difference in mental rotations skill is present in 40 countries, questioning social effects (Silverman et al., 2007).

594 PART | II Cognitive development

Spatial skills have been suggested to develop from childhood sex-typed activities (e.g., Connor and Serbin, 1977; Lawton, 2010). In particular, boy-typed toys and activities (such as building with Legos) are seen to encourage manipulation and exploration of the environment, and some have claimed that sex differences in spatial skills would be reduced or even eliminated if girls were encouraged to play more with boys’ toys (Halpern, 1986; Tracy, 1990). Evidence does show a weak-to-moderate link between spatial skill and aspects of sex-typed activities (e.g., Newcombe et al., 1983), but there is some variability and inconsistency that likely reflects methodological and conceptual issues (Baenninger and Newcombe, 1989; Voyer et al., 2000). It is important to note that these associations are not evidence of causation: Engagement in boy-typed activities might enhance spatial skills or instead reflect that ability, that is, children with high spatial skills might be attracted to toys that allow manipulation and exploration. Some longitudinal data suggest that the causal path is from skills to activities rather than the reverse (Newcombe and Dubas, 1992). It is, therefore, important to note direct experimental evidence that spatial skills can be enhanced by experience. In particular, spatial ability can be improved through practice and training, with generalization beyond training stimuli. For example, playing an action video game was seen to improve both spatial attention and mental rotation skills (Feng et al., 2007). Training benefits both sexes, with women sometimes benefiting more than men, so that training may eliminate a sex difference in this domain (Lawton, 2010; Uttal et al., 2013). Sex-typed academic skills, such as mathematics and language, are influenced by family socialization. For example, although parents think that academic achievement is equally important for sons and daughters, they provide support for extracurricular involvement in mathematics and science more for sons than for daughters (Simpkins et al., 2005), and parents’ subtle beliefs about the inherent superiority of boys in such domains appear to undermine girls’ subsequent academic performance, especially in mathematics (Eccles et al., 2000). Finally, stereotypes that emphasize women’s cognitive inferiority appear to impair their performance (Steele, 1997). This has been demonstrated in studies that involve experimental manipulations, as illustrated for both mathematics and spatial skills. Women who were told that sex differences in mathematics have genetic causes performed worse on tests than those who were told that the differences have experiential causes (Dar-Nimrod and Heine, 2006). Women who were told that men outperform women on spatial tasks performed worse on a mental rotations test than women who received neutral information, and the poorer performance of the group given negative stereotypes appeared to reflect increased emotional load (Wraga et al., 2007).

27.3.1.2 Socialization of noncognitive sex differences Most studies of gender socialization have focused on social behaviors. This literature has been reviewed elsewhere (Blakemore et al., 2009; Martin and Ruble, 2010; Ruble et al., 2006), so a brief summary is provided here. In essence, boys and girls are socialized differently in ways that affect a variety of psychological outcomes, with this gendered socialization coming from a variety of sources. Much of the focus has been on socialization by peers and parents, but there are powerful influences from other social forces, including other adults such as teachers, coaches, and clergy, and information received via the many forms of media. Peers are a key enforcer of sex typing. Children have strong preferences for interaction with members of their own sex, with these preferences maintained by children themselves and resistant to change by adults (Maccoby, 1998; Ruble et al., 2006). The more children play with others of the same sex, the more they engage in gendered activities and play styles (Martin and Fabes, 2001). Parents also shape sex typing, as seen in two examples. First, parents influence the career choices of their offspring in several ways: by differential encouragement of sex-appropriate activities, by the attitudes they espouse regarding what is appropriate for boys versus girls, and by the resources they provide to their children (e.g., paying for and providing computer-related material and encouraging extracurricular involvement in mathematics and science for sons, but not for daughters; Simpkins et al., 2005). Second, parents socialize emotion differently in their sons and their daughters, through their conversations (e.g., Fivush, 1998; Fivush and Buckner, 2000): With daughters as compared to with sons, parents use more emotion-related words, elaborate their discussion of emotion more extensively, and focus conversations on the emotional aspects of interpersonal relationships; over time, girls’ own discussions include more extensive focus on emotion and emotional issues than do boys’ own discussions (Rose et al., 2016). Children also help to socialize themselves through their use of gender schemas. Children are motivated to be like others of their own sex and form cognitive constructions or networks of associations about the sexes that influence their behavior and thinking (Martin and Ruble, 2004; Martin et al., 2002). These gender schemas direct children’s attention, influence how information is interpreted, organized, and remembered, and guide behavior with objects and people; specifically, children selectively attend to and remember sex-typed information and show biases toward members of their own group

Sex differences in brain and behavioral development Chapter | 27

595

(reviewed in Martin and Ruble, 2004; Martin et al., 2002). Many studies confirm the power of gender schemas to influence many aspects of behavior and thinking, including children’s toy play, play partners, the ability to learn about activities that are traditionally for the other sex, and impressions of others (reviewed in Blakemore et al., 2009; Martin et al., 2002; Ruble et al., 2006).

27.3.2 Genetic perspectives It is reasonable to expect that sex-related characteristics would be influenced by genes on the sex chromosomes, reflecting sex differences in sex chromosome composition (XX for typical females, XY for typical males)dand there is a lot of interest in this issue (e.g., Arnold, 2017), but not a lot of definitive evidence in human beings. Evidence in rodents suggests that genes on the Y chromosome influence sex-typed behavior, including spatial ability and parenting (Arnold, 2009; Arnold and Chen, 2009). Studies of individuals with sex chromosome abnormalities show that they have altered patterns of cognition compared with individuals without those abnormalities (reviewed in Hong and Reiss, 2014), but it is difficult to disentangle effects specific to genes on the sex chromosomes from downstream consequences of having a missing or extra sex chromosome, particularly effects due to atypical hormone exposure. Some intriguing evidence using sophisticated approaches suggests that genes on the X chromosome may influence aspects of cognition, including some spatial skills (Ross et al., 2006), but this requires further study. Perhaps the best test of direct effects of genes on the Y chromosome is provided by women with complete androgen insensitivity syndrome (CAIS): They have a Y chromosome but no effective androgen exposure, so any male-typical behavior likely reflects direct effects of genes on the Y chromosome. The condition is very rare, so there are not many studies in this population, but the extant data generally show that women with CAIS have behavior that is female typical (consistent with their low effective androgen exposure and female-typical rearing). Thus, compared with typical women, women with CAIS report similar childhood and adulthood gender role behavior (Hines et al., 2003a) and have similar performance on measures of cognition (Mueller et al., 2016; Strandqvist et al., 2018); there are conflicting findings about whether women with CAIS are worse at recognizing others’ emotions (Khorashad et al., 2018a; Strandqvist et al., 2018). Unfortunately, most studies are too small for meaningful inferences.

27.3.3 Hormone perspectives Most of the research on biological mechanisms underlying gendered characteristics has focused on sex hormones, primarily androgens (including testosterone) and estrogens. This research is rooted in the work of Phoenix, Goy, and colleagues (Gibber and Goy, 1985; Phoenix et al., 1959, 1973), showing the long-lasting effects of early sex hormones on sex differences in behavior in rodents and primates, and in the subsequent work of Money and Ehrhardt (1972) in human beings. Hormones affect behavior in two ways (Becker et al., 2008; Goy and McEwen, 1980). First, sex hormones produce permanent changes to brain structures and the behaviors they subserve (“organizational” effects). Such effects usually occur early in life (in human beings, during prenatal development and perhaps in the early postnatal period), although adolescence may also be an important organizational period (Schulz et al., 2009; Sisk and Zehr, 2005) along with other periods of hormone transition experienced across the life span (Beltz and Moser, 2019). Second, sex hormones produce temporary alterations to the brain and behavior (through ongoing changes to neural circuitry) as the hormones circulate in the body throughout adolescence and adulthood (“activational” effects). The main distinctions between organizational and activational effects concern timing and permanence, although these distinctions are not absolute (Arnold and Breedlove, 1985).

27.3.3.1 Evidence for hormone influences on nonhuman sex-typed behavior Studies in many nonhuman animal species show that sex hormones are crucial for behavioral sex differences. Much work has confirmed and extended the early work of Phoenix, Goy, and colleagues, showing that hormones present during early life organize the brain so that they have long-lasting effects (Becker et al., 2002, 2008; Ryan and Vandenbergh, 2002; Wallen, 2005, 2009). These studies generally involve experimental manipulations of hormones (e.g., injecting females with testosterone, castrating males), but behavior is also influenced by naturally occurring variations in hormones, such as those that result from an animal’s position in the uterus, particularly the sex of its littermates (reviewed in Clark and Galef, 1998; Ryan and Vandenbergh, 2002).

596 PART | II Cognitive development

Even in nonhuman animals, however, the effects are complex. Studies in monkeys highlight two aspects of this complexity. First, there are several sensitive periods for androgen effects on behavior, even during the prenatal period, with some behaviors masculinized by exposure early (but not late) in gestation, and other behaviors masculinized by exposure late (but not early) in gestation (Goy et al., 1988). Second, environmental context modifies behavioral effects of hormones (Wallen, 1996). Furthermore, sex hormones may continue to exert organizational effects well beyond early life. Evidence from rodents indicates that puberty is another organizational period, with sex hormones at puberty producing permanent changes to the brain and behavior (Schulz et al., 2009; Sisk and Zehr, 2005). There is also an extensive literature showing that sex hormones are necessary for the expression of sex-typed behaviors in adulthood (activational effects). Much work has focused on the importance for sexual behavior of testosterone in males and estradiol in females, but these hormones also play a role in the expression of nonsexual behaviors in adult animals, such as maternal behavior and aggression (Becker et al., 2002, 2008).

27.3.3.2 Early hormone influences on human behavior Studies in people cannot involve experimental manipulations of hormones but have taken considerable advantage of natural experiments by studying individuals whose hormone levels were atypical for their sex during early development as a result of a disruption in the processes of sexual differentiation; these individuals are generally considered to have a disorder/difference of sex development (DSD; Blakemore et al., 2009). Although CAIS (discussed in Section 3.2) is a DSD, the most extensively studied DSD is congenital adrenal hyperplasia (CAH), which is a genetic disease resulting in exposure to high levels of androgens beginning early in gestation because of an enzyme defect affecting cortisol production. If human psychological sex differences are affected by androgens present during critical periods of development (as occurs in nonhuman animals), then females with CAH should be behaviorally more male typed and less female typed than a comparison group of females without CAH. Because CAH is not a perfect experiment (e.g., high levels of prenatal androgen also lead to masculinized genitalia, which might affect socialization, and CAH requires lifelong treatment with cortisone), it is important to seek converging evidence from other sources. Such evidence has been obtained for a number of behavioral domains from individuals with other DSD, as noted throughout this chapter. It has been more challenging to explore effects in individuals with typical variations in hormones, and three general approaches have been used to approximate fetal exposure to hormones. The first approach involves examining behavior in individuals whose hormones have been directly measured in amniotic fluid at one point in time, but this reflects a selected group (mothers who receive amniocentesis), and it is unlikely that a single sample of hormones in amniotic fluid reflects a fetus’ ongoing hormonal exposure. The second approach involves studying behavior in individuals whose hormone levels have been inferred by virtue of sharing a uterine environment with an opposite-sex twin; the rationale stems from animal studies showing that behavior is affected by uterine position, but the strongest effects are associated with gestating between two opposite-sex littermates, rather than residing near one. The third approach involves relating behavior to a presumed marker of early hormone exposure. Various markers have been proposed over the years, but the one that has received the most attention is digit ratio, specifically the ratio of the second to the fourth digit (2D:4D). But evidence makes clear that 2D:4D does not mark variations in prenatal androgen exposure (discussed in Berenbaum and Beltz, 2016; Berenbaum et al., 2009). The best evidence for the role of androgens in digit ratio comes from women with CAIS who have no effective androgen exposure (Berenbaum et al., 2009; van Hemmen et al., 2017): They have been shown to have moderately feminized digit ratio compared with men but considerable variability in digit ratio despite minimal variability in androgen exposure; furthermore, digit ratios do not even provide high discrimination between typical men and women, despite the marked sex difference in prenatal androgen exposure. Recent work in typical individuals has focused attention on the first few months of postnatal life as another early sensitive period for the behavioral effects of androgens, consistent with the testosterone surge in boys during the first to fifth postnatal months (sometimes called minipuberty; Hines et al., 2016). The importance of this period has been explored in studies linking hormones and genital markers during minipuberty to behavior and cognition at the time and later in childhood. Below is a brief summary of the evidence that early hormones (during prenatal and early postnatal life)dparticularly androgensdinfluence human psychological sex differences; key findings and issues are highlighted. This evidence has been discussed in detail elsewhere (Berenbaum, 2018; Berenbaum and Beltz, 2011, 2016; Blakemore et al., 2009; Hines, 2010; Hines et al., 2015).

Sex differences in brain and behavioral development Chapter | 27

597

27.3.3.2.1 Early hormone influences on human behavior: cognitive sex differences Evidence from multiple sources provides moderate support for the notion that prenatal androgens influence later spatial and related skills. The most compelling evidence comes from natural experiments, although there is some intriguing support from typical samples. Females with CAH have been found to have better spatial skills than their sisters in childhood, adolescence, and adulthood (Berenbaum et al., 2012; Hampson et al., 1998; Mueller et al., 2008; Resnick et al., 1986), with a meta-analysis suggesting that the effect is small to moderate in size (Puts et al., 2008). The effect is not always seen, however (e.g., Hines et al., 2003b; Malouf et al., 2006), and inconsistencies likely reflect both low power with small samples and consequences of CAH that may have subtle adverse effects on the brain that counter the facilitative effects of androgens on spatial skills (Hampson and Rovet, 2015). Confirming evidence for the effects of androgen on spatial skills comes from individuals at the other end of androgen levels: Males with low early androgen levels due to idiopathic hypogonadotropic hypogonadism (IHH) have poorer spatial skills than controls (Hier and Crowley, 1982). Importantly, the external genitals of males with IHH appear typical, suggesting that the enhanced spatial skills of females with CAH are not due to social responses to their genitals. Some converging evidence of early androgen effects on spatial skills has been found in typical samples, but, again, there are inconsistencies. On the one hand, positive findings come from two sources. First, in infants aged 1e2.5 months, salivary testosterone was positively correlated with mental rotation performance in boys, but not in girls (Constantinescu et al., 2018). Second, females with a male co-twin (who are thought to have above-average prenatal exposure to testosterone) had better spatial skills than females with a female co-twin (Cole-Harding et al., 1988; Heil et al., 2011; Vuoksimaa et al., 2010); effects of postnatal socialization (rearing with a male sibling of similar age) were ruled out in one study by finding that those with a brother did not have better spatial skills than those with a sister (Heil et al., 2011). On the other hand, studies linking spatial skills to amniotic hormones are generally negative (discussed in Berenbaum and Beltz, 2016; Hines et al., 2015). Given the limitations of these types of studies, it is necessary to be cautious in drawing inferences about androgen effects on spatial skills in typical samples. There is insufficient evidencedeither in individuals with DSD or in typical samplesdto know how early androgens affect other sex-typed skills, including stereotypically male-superior abilities such as mathematical skills, and femalesuperior abilities, such as verbal fluency, verbal memory, emotion recognition, and perceptual speed (discussed in Berenbaum and Beltz, 2016; Hines et al., 2016). These skills generally show only modest-sized sex differences, and most studies are insufficiently powered or have other methodological limitations that prevent inferences. Recent intriguing evidence suggests that an aspect of social cognition may be influenced by prenatal androgens: On the test “Reading the Mind in the Eyes” (variously considered to measure emotion recognition, ability to understand others’ emotions, theory of mind, and empathy), women with CAH had lower scores than women with CAIS or typical women who were not significantly different from each other (Khorashad et al., 2018a). 27.3.3.2.2 Early hormone influences on human behavior: noncognitive sex differences Data from multiple groups and countries with a variety of sound methods (including observations, tests, self-reports, and parent-reports) make clear that girls and women with CAH are more male typed and less female typed in some, but not all, aspects of their feelings, preferences, and behavior than are girls and women without CAH (in most studies, their unaffected sisters; reviewed in Berenbaum, 2018; Berenbaum and Beltz, 2011, 2016; Blakemore et al., 2009; Hines, 2010; Hines et al., 2015). The largest difference between females with and without CAH is in sex-typed activity interests and engagement: Girls and women with CAH prefer and are more likely to participate in male-typed activities from childhood through adulthood. They also have male-typed occupational interests; for example, females with CAH reported more interest in occupations that involve working with things (vs. people) than unaffected female siblings (Beltz et al., 2011). These male-typical interests translate into increased participation in occupations typically dominated by men and into higher salary (consistent with gender disparities in pay; Frisen et al., 2009). The male-typed activity preferences and engagement in girls with CAH have been shown to be directly associated with prenatal androgen and not to result from parents’ behavior (MeyerBahlburg et al., 2006; Nordenström et al., 2002; Pasterski et al., 2005). Females with CAH are sex atypical in other domains, but not in all aspects of behavior (reviewed in Berenbaum, 2018; Berenbaum and Beltz, 2011, 2016; Blakemore et al., 2009; Hines, 2010; Hines et al., 2015). Compared with typical females, females with CAH are more aggressive and less interested in babies and are more likely to have bisexual or homosexual orientation (although most are exclusively heterosexual). These masculinized characteristics stand in contrast to female-typical identity in the overwhelming majority of females with CAH (Berenbaum et al., 2018; Dessens et al., 2005).

598 PART | II Cognitive development

Data from other natural experiments also show that male-typical prenatal androgen levels are associated with maletyped activity interests and nonheterosexual orientation but have smaller effects on gender identity (Berenbaum, 2018; Berenbaum and Beltz, 2011, 2016; Blakemore et al., 2009; Hines et al., 2015; Meyer-Bahlburg, 2005). Recent studies emphasize that early androgens have differential effects on different aspects of behavior, with large effects on activity interests and participation, moderate effects on social behaviors, and minimal effects on gender identity (e.g., Berenbaum et al., 2018; Endendijk et al., 2016; Khorashad et al., 2018b; Kreukels et al., 2018; Spencer et al., 2017). It is noteworthy that converging evidence is seen across multiple methods in samples from different countries. There is limited converging evidence for the behavioral role of prenatal androgens from typical samples. Studies examining links between amniotic testosterone and behavior in childhood are inconsistent (see discussion in Hines et al., 2015), and it is unclear why: whether they reflect limitations related to small sample sizes, varying sensitivity of measures, insufficient sampling of hormones, or the failure of amniotic hormones to index fetal exposure. Recent evidence from typical samples provides some evidence for a role for early postnatal hormones on sex-typed play. Urinary testosterone in the first 6 months of life correlated with aspects of sex-typed play at age 14 months, specifically with parent-reported play in boys, and with observed toy choice in both boys and girls (Lamminmaki et al., 2012). Penile growth from birth to 3 months correlated with parent-reported gender-typed play at 3e4 years of age (Pasterski et al., 2015). These results are intriguing but require further extension to girls and replication, especially in light of difficulties replicating findings regarding amniotic hormones in typical samples, and potential for confounding with prenatal hormones. There has also been interest in effects of early hormones on sex differences in psychopathology. Early organizational androgens have been hypothesized to contribute to the male vulnerability for childhood disorders (e.g., Baron-Cohen et al., 2004; Martel, 2013; Martel et al., 2009); however, there has been little consideration of the ways that different disorders would be affected by hormones. For example, autism has been claimed to reflect “the extreme male brain” and therefore result from exposure to high prenatal androgens (Baron-Cohen et al., 2004), but many other disorders show similar male predominance and have been suggested by others to also result from high prenatal androgens (e.g., Martel et al., 2009). In any event, there is limited evidence that prenatal androgens affect sex-related psychopathology (see Berenbaum and Beltz, 2016; Hines et al., 2015). Given the difficulties studying androgen effects on rare conditions, and concerns about categorical diagnoses, it might be profitable to study androgen influences on continuous dimensions of behavior thought to contribute to psychopathology, including positive and negative valence systems (e.g., reward valuation and sustained threat, respectively), cognitive systems (e.g., attention), and social processes (e.g., understanding mental states; Patrick and Hajcak, 2016; Sanislow et al., 2010).

27.3.3.3 Adolescent hormone influences on human behavior In light of the recent animal evidence on the behavioral importance of both activational and organizational hormones at puberty, there has been increased attention to these effects in human beings. Studies of pubertal status concern how changes at adolescence are triggered by the surge in sex hormones at that time (activational effects). There is limited research on the relationship between pubertal hormones and cognition in humans, and the available data suggest minimal effects (e.g., Herlitz et al., 2013), although there may be an inverse link between pubertal testosterone and mental rotations in boys (Vuoksimaa et al., 2012). Links between pubertal hormones and psychological problems, however, seem to be present, including girls’ increased vulnerability to depression and eating disorders (Crick and Zahn-Waxler, 2003; Martel et al., 2009) and boys’ increased risk taking and substance use (Forbes et al., 2010; Steinberg, 2008). For instance, estrogen levels have been shown to fully account for the association between pubertal development and depression diagnoses in adolescent girls (Angold et al., 1999), and increased testosterone has been associated with increased sensation seeking in adolescents (Harden et al., 2018). Studies of pubertal timing concern how early, on time, or late development compared with same-sex peers affects later behavior (organizational effects). Findings are again limited for cognition (despite this being a longstanding topic of investigation; e.g., Waber, 1976). There is some evidence that the timing of puberty matters for spatial ability in boys, with early maturers having better spatial skills in young adulthood than later maturers, consistent with an organizational hypothesis (Beltz and Berenbaum, 2013), but this is not always found (Herlitz et al., 2013), perhaps due to the confounding of pubertal status and timing in many studies (e.g., adolescents with late timing may still be in the midst of puberty at age 17 years). There are more consistent findings for pubertal timing links to psychological problems, as reported in several reviews. For instance, early (and sometimes late)-maturing girls and late-maturing boys have greater depressive symptomatology in young adulthood compared with their peers with on-time development (Beltz, 2018b; Copeland et al., 2019; Graber, 2013; Negriff and Susman, 2011), and both early-maturing girls and boys are at risk for a variety of externalizing behaviors, such as alcohol use, conduct disorder, and delinquency (reviewed in Negriff and Susman, 2011).

Sex differences in brain and behavioral development Chapter | 27

599

27.3.3.4 Circulating hormone influences on human behavior There is an extensive literature (primarily in adolescents and adults) linking circulating levels of sex hormones (especially estradiol and testosterone through saliva and blood assays) to sex-typed characteristics, primarily aggression, mood, and cognition (reviewed in Buchanan et al., 1992; Hampson, 2007; Maki and Sundermann, 2009; Puts et al., 2010). This work focuses on activational effects of hormones. Findings are complex, with much of the complexity reflecting small effects, reliance on cross-sectional and observational studies in adults, and bidirectional effects of behavior and hormones (e.g., aggressive behavior can increase testosterone; reviewed in Duke et al., 2014). Hormones do not produce simple changes in behavior, and the most valuable studies are those that examine the ways in which hormones act indirectly and interact with social factors to influence sex-related characteristics (see discussion in Berenbaum and Beltz, 2018; Berenbaum et al., 2011). There is some consistency in findings regarding cognitive effects of circulating sex hormones, however, with data coming from studies of natural variations in hormones within individuals, particularly in association with the menstrual cycle (see Beltz and Moser, 2019). Results generally suggest that verbal skills and memory are enhanced during the luteal phase (characterized by moderate levels of estrogen and high levels of progesterone) compared with the menstrual and early follicular phases (characterized by low estrogen and progesterone); however, the opposite pattern is observed for spatial skills, with worse performance in the high-hormone versus low-hormone phases (Hampson, 1990; Maki et al., 2002; McCormick and Teillon, 2001). Nevertheless, links between cognition and hormones are not always found, probably due to factors that modify the effects of both hormones (e.g., diet) and cognition (e.g., experiences) and due to study methodology, as most samples are small and studies vary in their definition of cycle phases. Another important area of circulating hormone research concerns cognitive changes with the menopausal transition, which is marked by decreases in endogenous estrogen and progesterone. Although many women complain of deficits in episodic memory and executive functioning around this time (Gold et al., 2000), cognitive performance is not strongly related to estrogen levels in older women (for a review, see Henderson, 2008; Henderson and Popat, 2011). But studies are often small, cross-sectional, and fail to consider progesterone, so future work is needed.

27.3.3.5 Exogenous hormone influences on human behavior There is an emerging literature on exogenous hormone influences on human behavior. These synthetic hormone influences primarily include hormone replacement therapy (around menopause or associated with surgical removal of the ovaries) and hormonal contraceptive use, although there are others (e.g., androgen therapy for clinically low testosterone; considered in Beltz and Moser, 2019). For hormone replacement therapy, findings across longitudinal and cross-sectional studies converge in indicating that verbal memory is facilitated by estrogen replacement therapy, especially in younger perimenopausal women (Maki and Resnick, 2000; Maki and Sundermann, 2009). Although it has been long suggested that hormone replacement therapy helps prevent decline in memory for all older women, including postmenopausal women (Resnick et al., 1997), more recent evidence shows that effects are specific to estrogen therapy in younger women; moreover, estrogen when combined with progesterone or when administered long after menopause can actually have detrimental effects on cognition and health (Craig and Murphy, 2010; Maki and Henderson, 2012; Sherwin, 2012). Future work is needed to determine the exact timing and nature of these links between exogenous ovarian hormones and cognition. For oral contraceptive (OC) use (which is the most common form of hormonal contraceptive and contains synthetic estrogens and progestins that vary in androgenicity), findings do not always converge due to small sample sizes, heterogeneity among pills types, or variability across study designs (reviewed in Beltz and Moser, 2019; Gogos et al., 2014; Pletzer et al., 2014; Warren et al., 2014). Nonetheless, some consistent effects of OC use on cognition and affect are emerging. Compared with naturally cycling women, OC users have enhanced verbal memory (Gogos, 2013), consistent with exogenous estradiol effects around menopause, but poorer verbal fluency (Griksiene and Ruksenas, 2011). There is also evidence that users of pills with androgenic progestins outperform naturally cycling women in mental rotations tests, but that users of pills with antiandrogenic progestins perform worse than naturally cycling women (Beltz et al., 2015b; Griksiene et al., 2018), consistent with evidence that prenatal and pubertal androgens facilitate spatial skills. Furthermore, OC users have differential patterns of affect and risk for depression compared with naturally cycling women (Keyes et al., 2013; Oinonen and Mazmanian, 2002; Skovlund et al., 2016), but links seem to depend upon age and length of OC use and do not extend to other personal characteristics, such as personality (Beltz et al., 2019).

27.3.4 Integrated perspectives Evidence for differential effects of sex hormones on behavior combines with evidence for the importance of genes and socialization to make clear that sex-typed characteristics are influenced by multiple factors, and it is unfortunate that most

600 PART | II Cognitive development

studies focus on a single set of factors, rather than examining them in concert. Sex differences can best be understood by integrating different perspectives; focusing on only one set of causes (either social or biological) can lead to a distorted or even misleading understanding of sex-typed processes. Consider spatial ability as an example of the ways in which an understanding of sex differences could be enhanced by attention to both biology and social experiences (for detailed discussion and other examples, see Berenbaum et al., 2011). As discussed earlier, evidence is clear that spatial skills depend on social experiences, genes on the X chromosome, and sex hormones during prenatal development and again in adolescent and adult life. But spatial skills almost certainly develop from joint influences of biology and social experiences, as illustrated in two ways. First, biological influences on spatial skills are likely mediated through experience. As noted earlier, girls who are exposed to high levels of androgen during prenatal life (because of normal variation or CAH) are more likely than girls with low levels to play with boys’ toys, and those toys may facilitate the development of spatial skills (e.g., Newcombe et al., 1983). Preliminary evidence from females with CAH does indeed indicate that their enhanced spatial skills are in part mediated by their masculinized activity interests (Berenbaum et al., 2012). Second, biological predispositions likely facilitate learning. Although existing training studies are not compelling in this regard (e.g., men, who have high androgen levels, do not appear to be more likely to benefit from practice than women, who have low androgen levels), the situation might differ in childhood, when skills are developing, and on tests that do not show ceiling effects. This question can be studied in girls with CAH and in typical children, by examining the effects of practice at varying ages and with varying socialization experiences (e.g., presence of a same-sex sibling).

27.4 Brain sex differences: nature, development, and consequences The brains of men and women, and of boys and girls, are similar in many ways, but there are nonetheless some systematic, average differences in morphology, physiological functioning, and development. An understanding of these differencesdand what causes themdprovides insight into the mechanisms underlying sex differences in human health, disease, and behavior, including the behaviors discussed earlier (Cosgrove et al., 2007; McCarthy and Arnold, 2011; Miller et al., 2017). In this section, work on human brain sex differences is reviewed. The focus is on topics in which there is converging evidence, but null results and some exciting new findings that will likely spark future research are also presented. Most of the reviewed work comes from studies using magnetic resonance imaging (MRI), as it is the dominant research tool in human developmental neuroscience (Casey et al., 2005; Giedd et al., 2015; Luna et al., 2010), but work using other techniques, such as positron emission tomography (PET), perceptual asymmetries, and postmortem examinations, is also considered.

27.4.1 Issues in studying the brain MRI is widely used as a measure of brain anatomy and physiology. There are several scan sequences that can be used to produce a variety of image types that are then transformed using different analytic techniques. Although details of all scan sequences, images, and analyses will not be covered here, some general distinctions are helpful. Structural MRI (sMRI) provides measures of brain morphology and architecture (e.g., volumes of gray matter, white matter, and subcortical structures). Diffusion tensor imaging (DTI) is also a measure of brain morphology; it is the most frequently used assessment of water diffusion in the brain, thought to reflect white matter microstructure, and thus, information transmission among brain regions. Blood oxygen leveledependent (BOLD) functional MRI (fMRI) is a measure of change in blood oxygenation thought to reflect neural activity during rest or while performing a psychological task. All MRI measures are indirect and involve multiple assumptions and inferences, and each reflects multiple dynamic processes occurring at cellular and subcelluar levels (for discussion relevant to developmental science, see Casey et al., 2005; Paus, 2010). The analysis of MRI data is complex. Not only do preprocessing steps vary by the MRI measure, planned inferential analyses, and software version and workstation type (e.g., Gronenschild et al., 2012), but the data are also highly sensitive to motion artifactsda particular problem in children (Dosenbach et al., 2017; Satterthwaite et al., 2012). Moreover, for most studies of functional localization (in which regional activation during a particular behavior or task is identified), the brain is partitioned into volumetric pixels, or voxels, and statistics are conducted on each voxel as if it were independent of all others. This “mass univariate approach” can result in a large number of false positives unless appropriate corrections are made for multiple comparisons (Eklund et al., 2016; Friston, 2004). In studies linking behavior to regional brain activation, findings may be spurious if only voxels exceeding some preset threshold are examined (Vul et al., 2009). This is essentially

Sex differences in brain and behavioral development Chapter | 27

601

a selection bias: Only data from voxels reflecting significant brain activation are examined in relation to behavior. This problem can be avoided through a priori selection of brain regions of interest (ROIs) for analyses (Poldrack, 2000; Vul et al., 2009). Finally, for connectivity studies of distributed function (in which the synchronicity of activity among brain regions is considered), effects of motion differentially affect long-range versus short-range connections (Power et al., 2012; Satterthwaite et al., 2012), and debate surrounds how best to select ROIs (Marrelec and Fransson, 2011; Sohn et al., 2015) and calculate synchronicity among them (Beltz, 2018a; Smith et al., 2011). Beyond analytic challenges, human neuroimaging data are often inappropriately interpreted. For example, people may overinterpret brain images; research containing irrelevant neuroscience explanations and brain images is rated by nonexperts as more satisfying or reasonable than accurate work without such information in part because reductionist explanations are considered compelling (Beck, 2010; Hopkins et al., 2016; McCabe and Castel, 2008; Weisberg et al., 2008). Moreover, neural measures, especially when they show sex differences, are often interpreted in simplistic ways. For example, a bigger brain is not necessarily a better brain (see Section 4.2.1), and brain sex differences are not indicators of predetermined inequalities (see also McCarthy and Arnold, 2011; Miller et al., 2017; Poldrack, 2000). Finally, data from neuroimaging measures, including evidence for brain sex differences, must be considered in conjunction with other brain measures and behavioral data (see Sections 4.2.5 and 4.3.7). Interpreting only one piece of an intricate puzzle is likely to provide a distorted picture (Nesselroade, 2011). As reviewed in the following, there is much careful and important work demonstrating brain sex differences, and the openness, rigor, and reproducibility of psychological and neuroscience is only increasing (Kidwell et al., 2016; Poldrack et al., 2017). Research questions developed from a clear conceptual framework are often investigated using sound methodology and a careful approach to examining brainebehavior relations. Inferences are often balanced and appropriate, and there is some convergence of evidence. Of course, as in any area of science, not all studies are perfect, and results are not always straightforward, but limitations in methodology and inference, which are noted in the following, are no more prevalent or problematic in this field than elsewhere (Fiedler, 2011).

27.4.2 Sex differences in brain structure and their development An understanding of brain sex differences requires consideration of development. Some of the most compelling work on brainebehavior relations considers developmental trajectories (i.e., pattern of changes over time), not just at a single measurement. For example, general intelligence has been shown to relate to changes in cortical thickness across childhood and adolescence, but not to absolute measures of cortical thickness at a given age (Shaw et al., 2006). Thus, sex differences in brain development are best examined in longitudinal neuroimaging studies, which are becoming increasingly prevalent (Crone and Elzinga, 2015; Giedd et al., 2015). There are challenges to conducting longitudinal research, though: These include reliability and repeated use of stimuli, sample representativeness and attrition, and analytic complications (Herting et al., 2018; Telzer et al., 2018). For these reasons, the following sections include both developmental longitudinal research on sex differences and the many studies of sex differences at a single point in time (typically done in adults). Findings of sex differences in brain structure are first reviewed, followed by the ways in which structural differences relate to behavioral sex differences. This is followed by a discussion on sex differences in functional brain activity (Section 4.3).

27.4.2.1 Brain volume There are sex differences in intracranial, cerebral, cerebellar, and total brain volumes, with each approximately 10% larger in males than in females. These differences are seen in measurements made on postmortem brains (Holloway, 1980), in live ones using MRI (De Bellis et al., 2001; Giedd et al., 1997; Giedd and Rapoport, 2010; Goldstein et al., 2001; Lenroot and Giedd, 2010; Lenroot et al., 2007; Nopoulos et al., 2000; Sowell et al., 2002; Tiemeier et al., 2010), and summarized in a large meta-analysis (Ruigrok et al., 2014); they are also seen in several nonhuman primate species (Falk et al., 1999; Holloway, 1980). The human differences may primarily be due to the larger occipital and frontal poles of men compared with women (Sowell et al., 2007), although there is evidence that all lobes are bigger in men than women (Koolschijn and Crone, 2013). The differences in brain size are due, but only in part, to sex differences in body size, reflecting overall growth differences (Halpern, 2012; Holloway, 1980; Peters, 1991; Peters et al., 1998). The brain reaches about 95% of its adult size by age 6 years but appears to continue to grow into late adolescence (Giedd et al., 2015; Mills et al., 2016). The peak cerebral volume occurs about 4 years earlier in girls than in boys (a similar sex difference is seen in cerebellum development; Tiemeier et al., 2010), even though boys have larger absolute cerebrum and intracranial volumes than girls throughout development (Giedd and Rapoport, 2010; Lenroot et al., 2007;

602 PART | II Cognitive development

Ruigrok et al., 2014). A similar pattern is observed in gray and white matter, with relative proportions of each continuing to change throughout childhood and adolescence, with peak growth consistently occurring earlier in girls than in boys, who have larger absolute volumes than girls (Lenroot et al., 2007; Mills et al., 2016). The implications of brain size sex differences for brain function or behavior, however, are not clear (see also Section 4.2.5). Sex differences in the size of other body structures (e.g., men’s larger hearts and noses) do not translate into sex differences in function. Furthermore, a larger brain does not necessarily mean a smarter brain. Although there is a moderate association between brain size and intelligence (Flashman et al., 1997; Luders et al., 2009b), there are no sex differences in general intelligence, and neural connectivity (rather than size) is key to understanding the neural substrates of intelligence (Song et al., 2008; van den Heuvel et al., 2009). Interpretations of structural size differences are complicated for several reasons: Normal brain maturation involves cell death and synaptic pruning, brain size may be larger in individuals with developmental disorders including autism (Shaw et al., 2010; Sowell et al., 2001; Sowell et al., 2002), and there are dynamic relations between experiences and brain anatomy (e.g., taxi driving experience is associated with hippocampal shape; Maguire et al., 2000; Maguire et al., 2003). The sex difference in body size creates difficulties in studying sex differences in brain size, which, in turn, creates difficulties in studying other aspects of brain sex differences, leading to controversies over whether to correct for body size and overall brain size. Some have argued that sex differences in brain size are best corrected by body weight or height (Halpern, 2012; Holloway, 1980). Others have claimed that the relation between body height or weight and brain size is weak (Peters, 1991), particularly because there is great variation in human body size and type (Peters et al., 1998). Both arguments are dated. Current concerns are that the relation between brain and body size changes across development (Giedd and Rapoport, 2010; Lenroot and Giedd, 2010); for instance, in early adolescence, girls tend to be taller than boys (due to an earlier puberty-related growth spurt) despite having smaller brains (Giedd et al., 2015). This perhaps suggests that brainebody size corrections should only be made in adult samples. There are several ways to correct for sex differences in intracranial or cerebral volume to examine sex differences in specific brain structures (see Bishop and Wahlsten, 1997). First, brain or intracranial volume can be covaried. This approach is most appropriate if there is a linear relationship between brain size and the size of the brain structure or region being investigated, but there is little systematic investigation of this (for an exception, see Dennison et al., 2013); in fact, there is evidence for nonlinearities (Zhang and Sejnowski, 2000). This approach can also misrepresent sex differences if the relation is present in only one sex. Second, a ratio can be computed, reflecting the volume of the structure or region being investigated as a proportion of brain size; this approach is easy to understand, but its anatomical interpretation is unclear. A study comparing these approaches (i.e., correcting for vs. covarying whole brain or intracranial volume vs. creating proportional volumes) in four longitudinal samples showed that patterns of sex differences in gray and white matter volumes unsurprisingly varied by approach (Mills et al., 2016). Thus, recent innovative approaches have been suggested. For example, only men and women in samples matched on brain size have been compared (e.g., Luders et al., 2014), but men with particularly small brains and women with particularly large brains are unlikely to be representative of their sexes. Also, different numerical scaling factors for different characteristics of brain anatomy have been used because larger brains are not uniform expansions of smaller brains (e.g., there is greater relative increase in cortical surface area than cortical thickness with brain size increases; Im et al., 2008). Some have attempted to side-step the correction problem through transparency, that is, by providing both uncorrected results and results that correct for brain volume using one or more methods (e.g., Dennison et al., 2013; Filippi et al., 2013; Koolschijn and Crone, 2013). Longitudinal studies also side-step this problem by allowing examination of within-individual change across time and subsequent comparison of rates of change between the sexes (e.g., Thambisetty et al., 2010), but this is rarely done even when the data are available (e.g., Mills et al., 2016).

27.4.2.2 Regional structure volume There are several consistently replicated sex differences in the size of specific regions in the human brain, but these effects are not always detected, likely because they are small and approaches to correcting for sex differences in brain size vary across studies. Generally, as detailed in the following, some key regions of the brain implicated in interhemispheric communication and memory (e.g., the hippocampus) are larger in women than men, and other subcortical regions implicated in affective behaviors and sensory processing (e.g., hypothalamus, thalamus, and amygdala) are larger in men than women.

Sex differences in brain and behavioral development Chapter | 27

603

27.4.2.2.1 Sex differences in interhemispheric commissures Several studies have investigated sex differences in interhemispheric commissures (fiber bundles spanning the two halves of the brain). Commissures are important for facilitating information flow across hemispheres and, thus, for promoting cognitive function (Bryden, 1982; Kimura, 1999). There has been particular focus on sex differences in the corpus callosum (CC), especially its most posterior portion, the splenium, following an early histological report of a sex difference in this region (De Lacoste-Utamsing and Holloway, 1982). A meta-analysis of 49 studies concluded that there are no systematic sex differences in this structure when appropriate corrections are made for brain size (Bishop and Wahlsten, 1997), and MRI studies support this finding for both the splenium (Leonard et al., 2008; Luders et al., 2006b) and the entirety of the CC (Luders et al., 2014). Nonetheless, research with large, adult samples using sophisticated analysis techniques, including multimodal assessments (Bjornholm et al., 2017) and methods in which derivations of an individual’s CC from a template CC are captured at many points (Davatzikos and Resnick, 1998; Dubb et al., 2003), indicates that women do, in fact, have larger splenia than men. Other research has positively linked the size of the splenium in women (but not men) to cognition, such as verbal fluency, mental rotations, and memory (Davatzikos and Resnick, 1998; Hines et al., 1992); however, these links are not consistently seen in children and adolescents (e.g., Luders et al., 2011). The CC develops linearly in a rostral-to-caudal direction throughout childhood and early adolescence, and the rate of growth may be greater in girls than in boys (Giedd et al., 2015; Luders et al., 2010; Thompson et al., 2000). Even from late adolescence through late adulthood, the splenium appears to increase in size in women more than in men (Dubb et al., 2003). Effects are likely small, though, because they are not readily detected in all samples (e.g., Berman et al., 2018). There are also sex differences in the anterior commissure (connecting the right and left temporal lobes) and massa intermedia (connecting the right and left thalamus), but less is known about their functional significance than that of the CC. Both structures are generally found to be larger in women than men (Allen and Gorski, 1991, 1992; Kimura, 1999). The massa intermedia is also more frequently absent in men than in women (Rabl, 1958), and absence of this structure is linked to greater performance intelligence scores for men, but not for women (Lansdell and Davie, 1972). 27.4.2.2.2 Sex differences in structures involved in learning and memory Brain structures implicated in learning and memory also tend to be larger in females than males, consistent with the female advantage in aspects of memory (Section 2.1.4). Specifically, the hippocampus, which is important for memory formation, retention, and recall, is larger in females than males (Cahill, 2005; Filipek et al., 1994; Goldstein et al., 2001; Halpern, 2012; Lenroot and Giedd, 2010; Persson et al., 2014). It is also generally found to grow at a faster rate, and thus, to reach peak volumes earlier in girls than boys (Dennison et al., 2013; Giedd et al., 1997; Lenroot and Giedd, 2010), consistent with evidence that the sex difference does not emerge until after puberty (Satterthwaite et al., 2014). Results from recent meta-analyses, however, contradict this body of work. One reported that the hippocampi of women are not proportionately larger than those of men (Tan et al., 2016), but results are challenging to interpret because studies using different procedures and metrics for adjusting for overall brain volume were combined (introducing method variance), and analyses separating studies by correction method were likely underpowered; even so, most effect sizes suggested larger hippocampi in women, just not significantly so. Moreover, the same meta-analysis did not find evidence of age differences in hippocampi volume, which contradicts the developmental and aging literatures (e.g., Lenroot and Giedd, 2010; Voineskos et al., 2015). Another meta-analysis reported that men actually have larger hippocampi than women (Ruigrok et al., 2014), but results are again difficult to interpret due to age: In this analysis (including newborns through adults over age 60), age effects and sex by age interactions were not examined. The plasticity of the brain is important to consider. For example, the sex difference in verbal memory remained in a sample of men and women who had left anterior temporal lobectomy, suggesting that other brain regions and sex differences in strategy are also important for sex differences in memory (Berenbaum et al., 1997), consistent with a connectivity perspective of distributed brain function. 27.4.2.2.3 Sex differences in subcortical structures involved in affective behaviors and sensory processing Brain structures consistently found to be larger in men than women include the regions of the hypothalamus as well as the thalamus and amygdala. Sex differences in the human hypothalamus parallel early studies showing a very large sex difference in the rodent preoptic hypothalamus (Gorski et al., 1978). In particular, one of the four interstitial nuclei of the anterior hypothalamus (INAH-3) is larger in men than women (reviewed in Halpern, 2012; Kimura, 1999; Levay, 1991). There are also reports that some regions of the bed nucleus of the stria terminalis (BNST), the central portion of the

604 PART | II Cognitive development

connection between the hypothalamus and amygdala, are larger in men than women (Allen and Gorski, 1990; Zhou et al., 1995). Although the functional implications of these sex differences are unclear, there is some suggestion that the INAH and BNST play a role in gender identity and sexual orientation (for discussion, see Hines, 2011; Savic et al., 2010). There is also mounting evidence showing that males have larger thalami than females (Koolschijn and Crone, 2013; Raznahan et al., 2014), although this is not always found likely due to methodological reasons (e.g., Ruigrok et al., 2014; Tomasi and Volkow, 2012). The thalami are a pair of subcortical structures superior to the right and left hypothalami and adjacent to the third ventricle; they play a pivotal role in sensory processing and, thus, are broadly implicated in cognitive processes (Fama and Sullivan, 2015; Wang et al., 2012). They develop in an inverted U-shaped trajectory with volumes in girls peaking in mid-adolescence and boys peaking in late adolescence (Koolschijn and Crone, 2013; Raznahan et al., 2014). Finally, the amygdalae, a pair of bilateral structures in the medial temporal lobe, are larger in boys and men than in girls and women (Cahill, 2005; Goldstein et al., 2001; Halpern, 2012; Koolschijn and Crone, 2013; Lenroot and Giedd, 2010; Ruigrok et al., 2014), and they show some evidence of faster development in boys than in girls (Giedd et al., 1997; Lenroot and Giedd, 2010), but this is not always found (e.g., Koolschijn and Crone, 2013; Wierenga et al., 2014). They are thought to be important for the detection of and behavioral response to affective visual cues, perhaps because of their connectivity to many cortical and subcortical regions (Pessoa and Adolphs, 2010). Although a recent meta-analysis questioned this sex difference, concluding that “{amygdala volume} is not significantly larger in men than women when individual brain size is taken into account” (Marwha et al., 2017, p. 288), this statement arguably misrepresents the results of the meta-analysis, which actually showed a notable effect (Hedge’s g ¼ 0.23, which has a similar interpretation to Cohen’s d). The sex difference was not statistically significant but still reflects a small and meaningful effect, especially because studies using different procedures and metrics for adjusting for overall brain volume were combined (introducing method variance), and the meta-analysis may not have been adequately powered (containing only 12 samples).

27.4.2.3 Gray matter There are sex differences in gray matter, with girls and women having relatively more than boys and men on several different, albeit related, measures. The sex difference is seen in overall percentage or proportion of gray matter volume across the life span (Gur et al., 1999; Lenroot et al., 2007; Leonard et al., 2008; Luders et al., 2002; Tomasi and Volkow, 2012), but results can depend upon age and the method of correction for the sex difference in brain volume (see Mills et al., 2016; Ruigrok et al., 2014). For instance, in one study, greater gray matter-to-white matter ratios were reported in females, even though intracranial volumeecorrected gray matter volumes were larger in men in all four lobes (Koolschijn and Crone, 2013). Also, recent work suggests that the presumed sex difference in proportional gray matter volume is actually due to a sex difference in gray matter density (Gennatas et al., 2017; Tomasi and Volkow, 2012). Women also appear to have greater cortical thickness than do men in some areas of the brain, particularly in the frontal and parietal lobes. Supporting evidence comes from several different samples and research methodologies, including MRI combined with surface morphometry (Im et al., 2006; Lv et al., 2010), pattern algorithms (Luders et al., 2006a; Sowell et al., 2007), and tissue segmentation (Koscik et al., 2009). This is not always found, though (e.g., Koolschijn and Crone, 2013), perhaps because findings depend on age, as there are sex differences in adolescent patterns of cortical thinning (Mutlu et al., 2013), and on method (e.g., interpretation of histological work is complicated by small sample sizes; Rabinowicz et al., 1999). There are parallel differences in cortical complexity, or patterning of cerebral convolutions, with women showing greater complexity than men, particularly in the frontal and parietal lobes (Luders et al., 2004; Luders et al., 2006c). Findings of sex differences appear to vary with method: Differences favoring females have been found using a mesh-based approach (in which the convolutions at several thousand surface points are estimated from three-dimensional brain scans), but not when using the gyrification index (ratio of cortical surface to visible gyral surface, calculated from two-dimensional postmortem brain slices; Zilles et al., 1988). Cortical and subcortical gray matter generally develop in an inverted U-shaped trajectory across childhood and adolescence, increasing until puberty and then decreasing through early adulthood (Giedd et al., 1999; Giedd and Rapoport, 2010; Koolschijn and Crone, 2013; Lenroot and Giedd, 2010; Lenroot et al., 2007; Vijayakumar et al., 2016; Wierenga et al., 2014). The shape and timing of the peak in gray matter trajectories differ across brain regions: Parietal lobe volumes peak first, followed by frontal and temporal lobe volumes. Volumes in subcortical regions appear to peak after cortical volumes (Raznahan et al., 2014). There is some counterevidence that gray matter development actually declines linearly (not quadratically, or in an inverted U-shape) from late childhood onward (Mills et al., 2016). Discrepancies in the shape of developmental trajectories of gray matter likely reflect differences across studies in modeling, measurement

Sex differences in brain and behavioral development Chapter | 27

605

(e.g., number of observations and intervals between them), and sample characteristics (e.g., ages of participants). Regardless, the sexes have similarly shaped trajectories of gray matter development across all brain regions, but girls have earlier average peaks than boys, consistent with their earlier pubertal maturation. Generally, girls show a peak in regional gray matter between 8 and 10 years of age, whereas boys show a peak between 9 and 11 years of age (Lenroot et al., 2007). Relatedly, cortical thinning, marked by decreases in cortical density thought to reflect maturation, occurs in a specified pattern: The medial sensorimotor cortex thins first, with development proceeding in rostral and lateral directions, such that the prefrontal and lateral temporal cortices are the last regions to thin (Gogtay et al., 2004). This reduction in gray matter density appears to be linear (although nonlinearities have been reported; Vijayakumar et al., 2016) and likely reflects the elimination of irrelevant brain connections and strengthening of relevant ones (i.e., pruning). Sex differences are not well established, but there are reports for quicker thinning in both males and females, depending upon the region (Ducharme et al., 2016; Vijayakumar et al., 2016).

27.4.2.4 White matter There is not consistency in MRI studies regarding sex differences in white matter, but some evidence indicates that men have a greater proportional or corrected volume than do women (Allen et al., 2003; Gur et al., 1999; Koolschijn and Crone, 2013; Ruigrok et al., 2014; Tomasi and Volkow, 2012). The difference is not always found, though, and inconsistencies are not explained by differential corrections for the sex difference in brain size (Luders et al., 2002; Nopoulos et al., 2000). Women are rarely found to have greater white matter volume than men, suggesting that small samples and differential corrections for the sex difference in total brain volume make it difficult to reliably detect the increased white matter in men compared with women. Inconsistencies may also reflect age differences; sex differences in the proportion of brain white matter generally emerge in adolescence and persist through adulthood (Paus, 2010). Converging evidence from MRI also suggests no sex differences in fractional anisotropy (FA), which is thought to reflect directed information transmission along white matter paths. This evidence comes from large, cross-sectional DTI studies (Eluvathingal et al., 2007; Giorgio et al., 2008; Lebel et al., 2008). White matter volume increases linearly through childhood and into adulthood, and this pattern is generally consistent across brain regions (Giedd et al., 1999, 2015; Koolschijn and Crone, 2013; Lenroot et al., 2007; Mills et al., 2016; Simmonds et al., 2014). Girls and boys both experience linear increases in white matter volume, but both sexes have been reported in different studies to have more rapid rates of increase. The discrepancy may be due to age range restriction or differential periods of white matter growth in boys and girls (e.g., Simmonds et al., 2014); when trajectories are stopped in adolescence, boys appear to show steeper growth (Lenroot et al., 2007), but when trajectories extend to age 30 years, women show steeper growth (Koolschijn and Crone, 2013). FA similarly increases across early childhood and into adulthood in a general inferior-to-superior and posterior-to-anterior fashion, with emerging evidence of sex differences in late adolescence such that boys have greater FA (Colby et al., 2011; Krogsrud et al., 2016; Simmonds et al., 2014). White matter connections around subcortical structures and in the CC seem to undergo the greatest change across development, and frontaletemporal and corticolimbic tracts are the last to mature (Eluvathingal et al., 2007; Giorgio et al., 2008; Lebel et al., 2008; Simmonds et al., 2014). Increases in FA may mark a transition from functional localization to distributed neural network functioning in the developing brain. For example, adolescents show a decrease in frontal gray matter density as FA increases in pathways connecting frontal regions to other brain areas (Giorgio et al., 2008). Indeed, there is increasing investigation into the structural connectome (or pattern of brain-wide fiber connections) by sex and across development. Compared with women, men have connectomes characterized by greater connectivity within each hemisphere and greater modularity (number of specialized subnetworks); conversely, women have connectomes characterized by greater connectivity between hemispheres, consistent with sex differences in interhemispheric commissures (Ingalhalikar et al., 2014). Although these sex differences seem to emerge in adolescence, the general features of the structural connectome (e.g., in terms of modularity and connections between modules) are constant across development, except for an increasing loss of pathways between connected regions with age; this loss parallels the notion of adolescent pruning and, like pruning, begins earlier in girls than in boys (Lim et al., 2015).

27.4.2.5 Implications of sex differences in brain structure Several considerations are important when interpreting sex differences in brain structure. For instance, sex differences may not be consistent across contralateral brain regions (i.e., corresponding regions in opposite hemispheres of the brain). Sometimes, differences are found in one hemisphere, but not in the other (see, e.g., Dennison et al., 2013), which is not

606 PART | II Cognitive development

surprising given sex differences in intra- and interhemispheric connectivity (Ingalhalikar et al., 2014) and in functional lateralization (Section 4.3.1). Relatedly, there may be sex differences in only specific subsections of a brain region (e.g., the splenium of the corpus callosum and INAH-3 of the hypothalamus; see also Bjornholm et al., 2017; Persson et al., 2014; Raznahan et al., 2014), challenging interpretations when only the region in its entirety is analyzed. Another important consideration regarding sex differences in brain development is the definition of the “mature” or “developed” brain (Somerville, 2016). Often, “maturity” is operationalized as the state or age when the overall size of a brain structure reaches its adult size. This, however, is overly simplistic, failing to consider subregions of the structure or its cellular composition, the functional properties of the structure, and the structure’s connectivity with other regions. With these considerations in mind, it is important to note that somedbut not all (e.g., Luders et al., 2009a)dof the sex differences in brain morphology might result from the different shaping of larger versus smaller brains (Allen et al., 2003; Im et al., 2008; Luders et al., 2002; Luders and Toga, 2010; Seldon, 2005; Zhang and Sejnowski, 2000). In larger as compared with smaller brains, the cortex tends to be flatter and thinner because it fills a larger intracranial space; this is consistent with reports of greater cortical density, thickness, and complexity in women than men. There is also more white matter in larger brains because longer axonal connections are made to cortical regions that are farther apart than in smaller brains. This is consistent with evidence that men seem to have greater white matter and axonal volumes than women (e.g., Kodiweera et al., 2016). What do sex differences in brain structure mean for cognitive sex differences? There is increasing investigation into links between white matter development and cognitive development, with white matter serving as a marker of neural connectivity (see Giedd et al., 2015). For instance, there is emerging evidence that the timing of white matter growth likely matters for cognition, with earlier adolescent growth linked to enhanced inhibition (Simmonds et al., 2014), and that individual differences in brain-wide metrics of connectivity are related to individual differences in cognition, including in spatial and language processing (Ponsoda et al., 2017). There are also studies linking spatial and language processing to gray matter structure. Compared with women, men appear to have larger regions of the parietal lobe, which is the primary brain area implicated in spatial ability (Brun et al., 2009). Parietal lobe surface area is, in fact, positively related to mental rotations performance in men, whereas parietal lobe gray matter volume is negatively related to mental rotations performance in women (Koscik et al., 2009). This suggests that different parietal lobe morphology subserves mental rotations performance in men and women, but it does not indicate whether men and women equally engage parietal regions during mental rotations tasks (this requires data on brain activation). Compared with men, women appear to have some larger brain regions that are implicated in language, especially within the temporal lobe (Brun et al., 2009; Harasty et al., 1997). But, studies on this topic are difficult to interpret because verbal skills are not always assessed. Sex differences in brain activation during spatial and language tasks are reviewed in the following (Sections 4.3.2 and 4.3.3, respectively). It is probable that not all structural sex differences have behavioral significance, as there are likely multiple paths to the same outcome, and individuals do not uniformly display sex-typical (or atypical) morphological patterns (De Vries, 2004; Joel et al., 2015; McCarthy and Arnold, 2011). Despite large variations in structure, brain function is remarkably similar and stable across people (Lim et al., 2015; Sporns, 2011). If this “functional homeostasis” is maintained, then direct links between brain structure and behavior will not always exist.

27.4.3 Sex differences in brain function (activation) There is a considerable literature examining sex differences in brain function measured by fMRI and PET, and inferences about sex differences in brain function have also been made using behavioral assessments, such as tests of perceptual asymmetries (e.g., dichotic listening). Task-related studies examine brain function during a particular cognitive task, with the most meaningful work relating brain activation to task performance. An important issue concerns the sex difference on the behavioral outcome of interest. If men and women perform at different levels, then brain activation differences may reveal the neural substrates of the performance difference. If men and women perform at similar levels, then brain activation patterns may reveal the (potentially different) processes or strategies they use to arrive at the same outcome. Studies of resting state functional connectivity are also becoming increasingly prevalent and providing novel insights into sex differences in cognitive development. They reveal the nature of brain networks (thought to reflect the physiological and psychological baseline of the brain) in the absence of a goal-directed task, making them especially valuable in developmental research (Shen, 2015). Again, this work is most meaningful when network features are linked to behavior.

27.4.3.1 Lateralization Lateralization, also referred to as hemispheric specialization or functional asymmetry, has been a historically popular explanation for cognitive sex differences, particularly for spatial and language skills. Typically, the left hemisphere, which

Sex differences in brain and behavioral development Chapter | 27

607

houses both Broca’s and Wernicke’s areas, is dominant for sequential processing, including language, and the right hemisphere is dominant for simultaneous processing, including spatial skills (for reviews, see Bryden, 1982; Hall et al., 2008; Kansaku and Kitazawa, 2001; Kimura, 1999). Data going back several decades, derived from patients with brain damage and behavioral tasks in typical individuals (e.g., split-field visual tasks, dichotic listening, and electroencephalogram asymmetry), suggest that women are less lateralized than men, especially for language (Bryden, 1982; Gur and Gur, 2017; McGlone, 1980). The topic has also been studied using fMRI, but results are often task and method dependent and subject to misinterpretation. For instance, lateralization must be explicitly tested by comparing activation in the two hemispheres; it is not sufficient to show that the activity of only one hemisphere is above baseline or that there are significant sex differences in specific regions of a single hemisphere. When the correct analyses are conducted, women generally show less lateralization than men, but the effect is small. Women’s reduced lateralization has been suggested to be linked to greater interhemispheric communication facilitated by their larger interhemispheric commissures and is consistent with data showing their greater interhemispheric structural connectivity (Halpern, 2012; Hines et al., 1992; Ingalhalikar et al., 2014). Functional imaging studies linking lateralization to cognitive performance are reviewed below for spatial skills, language, and emotion processing. Briefly, results regarding sex differences are somewhat inconsistent, partly due to study variations in task designs and the poor temporal resolution of fMRI and PETda general limitation that is particularly critical in examinations of the speed and timing of interhemispheric processing (for discussion, see Hall et al., 2008; Kansaku and Kitazawa, 2001; Kitazawa and Kansaku, 2005; Ortigue et al., 2005). It is important to note that rarely are women reported to be more lateralized than men, so inconsistencies likely reflect differential measurement sensitivity and small sex differences. Nonetheless, combining the relatively small sex differences in brain lateralization and the relatively large sex differences in some skills, it is very unlikely that lateralization differences completely account for cognitive sex differences (consistent with the conclusions of others; Hirnstein et al., 2019).

27.4.3.2 Spatial skills There has been interest in finding the neural correlates of the male advantage in spatial skills. Generally, in both sexes, spatial processing is associated with activation of the parietal lobes as well as temporal, frontal (especially premotor), and extrastriate areas (Butler et al., 2006; Christova et al., 2008; Halari et al., 2006; Hugdahl et al., 2006; Thomsen et al., 2000; Weiss et al., 2003a; Yu et al., 2009). More recent work, however, focuses not on regional differences, but on sex differences in patterns of connectivity underlying spatial processing (Section 4.3.5). There is also some evidence that right hemisphere regions are more engaged in spatial tasks than left hemisphere regions, particularly for men, but findings from neuroimaging (e.g., PET and fMRI) are not always consistent with those from perceptual asymmetry assessments (e.g., split visual field and dichotic listening tasks), suggesting that the methods are measuring different aspects of lateralization. A meta-analysis of sex differences in lateralization of spatial tasks measured by a variety of methods (e.g., fMRI, PET, split visual field, brain damage) found that men primarily engage the right hemisphere and women engage both hemispheres when solving spatial tasks (Vogel et al., 2003). However, subsequent findings from fMRI studies in which lateralization was explicitly examined are mixed: Some reported greater right than left hemisphere activity for both men and women (Halari et al., 2006; Hugdahl et al., 2006), others reported more right-lateralized activity in men than women (e.g., Christova et al., 2008) or in women than men (e.g., Clements et al., 2006), and still others reported no lateralization effects (e.g., Weiss et al., 2003a). There are few studies directly examining the link between sex differences in lateralized activation and sex differences in cognitive performance. For spatial skills, greater (usually right) lateralization is thought to be linked to better performance because the processing of spatial information is unhindered by language circuitry. Beyond lateralization, the brain regions activated by men and women during mental rotations tasks are largely overlapping when there are no sex differences in behavioral performance, but some brain differences have been reported. Women engage portions of the frontal lobe, in particular the right inferior frontal gyrus, that men generally do not (Hugdahl et al., 2006; Jordan et al., 2002; Weiss et al., 2003a). For women, it is the activation of these frontal regions (as well as some temporal and parietal regions) that positively predicts task performance accuracy, whereas activation of parietal regions (e.g., postcentral gyrus and precuneus) positively predicts accuracy for men (Butler et al., 2006). The different patterns of task-related brain activation for women and men might reflect women’s use of top-down processing and men’s use of bottom-up processing to complete mental rotations tasks (Butler et al., 2006). Functional brain sex differences during mental rotations tasks when the sexes differ in performance may reveal the neural substrates of the cognitive difference, although the performance differences may complicate interpretation. Most behavioral tests of mental rotations involve comparisons among multiple figures, whereas most neuroimaging studies of

608 PART | II Cognitive development

mental rotations involve a pairwise comparison because of scanner space constraints; the sex difference in mental rotations performance is likely reduced in the latter compared with the former (Peters and Battista, 2008). In one fMRI study in which men outperformed women, there were no sex differences in brain regions activated by the task, but women had significantly greater activation of those regions than men (Halari et al., 2006); this is consistent with the notion that increased activation reflects the need to work harder (e.g., Gur et al., 2000). In a study using event-related potentials (with better temporal, but worse spatial resolution than fMRI), there were no sex differences in parietal responses during mental rotations, but there was a sex difference in the frontal lobe: Compared with men, women had greater negative amplitudes (reflecting a stronger neural response) in the right hemisphere early in the task, suggesting that sex differences may be driven by differences in early mental processes, such as perception, instead of later processes, such as rotation strategy (Yu et al., 2009). There are also sex differences in brain activation during other visuospatial tasks. When matching the orientation of angled lines, men had greater activation than women of left occipital and cingulate regions, and this effect increased with age (Clements-Stephens et al., 2009; Clements et al., 2006). In contrast, no regional brain sex differences were found in a similar task in which a joystick was used to move a cursor in a specified angle away from a central stimulus (Christova et al., 2008). During spatial navigation, men engaged more left temporal regions than women, and women engaged more right frontal, parietal, and hippocampal regions than men (Grön et al., 2000), although sex differences in related tasks have not always been found (Ohnishi et al., 2006). During spatial attention tasks, men appear to activate left parietal regions more than women, whereas women engage more right frontal regions than men (Rubia et al., 2010), although this difference is not always found either (Bell et al., 2006). These studies converge to indicate that visuospatial tasks seem to engage different brain systems in the two sexes: Men recruit left hemisphere regions to a greater degree than women, and women recruit frontal lobe regions that men generally do not. Unfortunately, it is difficult to review these findings in relation to behavioral performance because of the limited number of studies available and differences across studies in the tasks used.

27.4.3.3 Language Several studies have focused on finding the neural substrates of the female advantage in verbal skills. Generally, in both sexes, activity in regions of the left hemisphere, including the temporal lobe, prefrontal cortex, inferior frontal gyrus, cingulate, and regions of the parietal lobe, is associated with the performance of most language tasks, including verbal fluency, rhyming, and comprehension (Allendorfer et al., 2012; Buckner et al., 1995; Burman et al., 2008; Chang et al., 2018; Clements et al., 2006; Frank et al., 2015; Frost et al., 1999; Gauthier et al., 2009; Halari et al., 2006; Plante et al., 2006; Shaywitz et al., 1995; Weiss et al., 2003b). Because most individuals have dominant left hemispheric language processing, sex differences in the neural lateralization of language have been widely investigated. Women appear to have somewhat greater bilateral representation for language than men. Two meta-analyses of fMRI and PET studies on language lateralization reported weak support for left lateralization in men but not women in several domains (Sommer et al., 2004, 2008), and meta-analyses using other behavioral methods (e.g., dichotic listening) revealed similarly small differences (Bryden, 1982; Sommer et al., 2008; Voyer, 2011). Evidence from a large sample of nearly 1800 participants and a study using an advanced imaging technique with high temporal and spatial resolution (i.e., magnetoencephalography, or MEG) suggests that sex differences may depend on age as well as the aspect of language lateralization being measured; for instance, some differences (e.g., in right ear asymmetry) seem to emerge in adolescence (Hirnstein et al., 2013), whereas others (e.g., in event-related neural frequencies) dissipate by adolescence (Yu et al., 2014). Yet, the lateralization-performance link is unclear: Greater lateralization has been associated with poorer language skills in males (compared with females), and incomplete lateralization has been hypothesized to be a risk factor for language disorders (Hall et al., 2008). As with spatial skills, findings of sex differences in brain activation for language tasks using fMRI extend beyond lateralization and depend upon whether the sexes are matched on behavioral performance. When they are matched on task performance, most studies do not find differences in brain activation (Allendorfer et al., 2012; Chang et al., 2018; Clements et al., 2006; Donnelly et al., 2011; Frost et al., 1999; Weiss et al., 2003b). Nonetheless, differences favoring both sexes have also been reported: Men have been seen to have greater activation than women in language-related regions (Buckner et al., 1995; Gauthier et al., 2009); the reverse has also been found, with women having greater activation than men, particularly in temporal and right hemisphere regions (Frank et al., 2015; Plante et al., 2006; Shaywitz et al., 1995). When the sexes are not matched on task performance, women behaviorally outperform and display greater brain activation than men, especially in right hemisphere regions (Burman et al., 2008; Halari et al., 2006; Plante et al., 2006), consistent with data showing greater bilateral representation of language in women than men. (This argument might appear inconsistent

Sex differences in brain and behavioral development Chapter | 27

609

with that presented in Section 4.3.2, that is, that the greater frontal lobe activity of women during spatial tasks reflected their need to work harder than men. But, issues pertaining to the interpretation of sex differences in brain function are complicated and depend upon sex differences in behavior, as discussed in Section 4.3.7.) Furthermore, the link between performance accuracy in language tasks and brain activation differs for men and women: In women, accuracy is strongly linked to primary language areas in the bilateral frontal and left temporal lobes; in men, accuracy is weakly linked to primary left hemisphere language regions and also to secondary language areas, such as the left caudate and cingulate and right parietal lobe (Allendorfer et al., 2012; Burman et al., 2008; Donnelly et al., 2011).

27.4.3.4 Emotion-related processing Much imaging work has concerned the neural substrates of sex differences in the processing of emotions, enough for several reviews and meta-analyses, with increasingly nuanced perspectives. There is a particular focus on amygdala activation during the viewing of emotional human faces, which generally elicits activation in several brain regions other than the amygdala, including frontal and prefrontal cortex, anterior cingulate cortex, insula, as well as areas in the temporal, parietal, and occipital lobes (Fusar-Poli et al., 2009; Sergerie et al., 2008). The amygdala is typically engaged in the viewing of all emotions, in particular, happy, sad, and fearful faces (Fusar-Poli et al., 2009); this likely reflects its central connectivity to visual, subcortical, and cortical regions of the brain, all of which are engaged in the detection of emotional cues and subsequent planning of behavioral responses (Pessoa and Adolphs, 2010). Just as there are sex differences in amygdala volume (larger in men than in women), there appear to be sex differences in amygdala activation during the viewing of emotional faces (men have greater activation than do women), but results can be challenging to interpret, especially since women tend to have greater emotion expression and to be better at emotion recognition than men (see Section 2.2.6). Two meta-analyses of over 100 empirical studies each provide evidence for the sex difference: Men have greater bilateral activity than women in limbic areas, including the amygdala, and in prefrontal regions, particularly the medial prefrontal cortex (Fusar-Poli et al., 2009; Sergerie et al., 2008). This bilateral pattern of activation may depend on development, however, as right-lateralized amygdala activity during facial emotion recognition has been found for adolescent boys, but not girls (Schneider et al., 2011). Also, task performance and behavioral sex differences are not consistently reported in emotion recognition studies, and few studies have, in fact, investigated sex differences in the link between amygdala activation and performance on emotion recognition tasks: Many studies utilize passive viewing paradigms because the frontal lobe inhibits amygdala activation during explicit identification of emotions versus the passive viewing of emotions (Critchley et al., 2000; Hariri et al., 2000). The limited available data, however, suggest that bilateral amygdala activity is positively linked to emotion recognition accuracy for both men and women (Derntl et al., 2009; Habel et al., 2007). There are also sex differences in amygdala activation during other emotion-related tasks. In a manner consistent with findings on facial emotion recognition, men show more amygdala activity than women during the viewing of sexual stimuli (reviewed in Hamann, 2005). But greater amygdala activation in men than in women may not be a general phenomenon linked to emotion-related processing. A recent meta-analysis examining sex differences in neural responses to a wide range of emotion tasks found that women had greater activation of the amygdala (as well as other subcortical structures) than men and that men had greater activation in medial prefrontal regions than women (Filkowski et al., 2017). Other studies examining neural responses to negatively valenced emotional stimuli provide insight that may explain the discrepancies: Not only do task differences matter, but so do stimuli novelty, development, and connectivity. In males more than females, amygdala activation decreases with repeated presentations of the same negative stimuli and from childhood to adolescence, perhaps due to greater connectivity between the amygdala and medial frontal regions (Andreano et al., 2014; Hardee et al., 2017; Lungu et al., 2015). Meanwhile, females do not show amygdala activation dampening with repeated negative stimuli presentation and actually show heightened activity across adolescence (Andreano et al., 2014; Hardee et al., 2017). A more complex pattern of sex differences in amygdala activation occurs during emotional memory tasks. The amygdala is activated in both sexes, but activation is greater in the left amygdala in women and in the right amygdala in men. This finding is not easy to interpret, but it has been replicated in other emotion-processing regions (such as the ventromedial prefrontal cortex; Reber and Tranel, 2017). It has been suggested that the sex differences are linked to internalizing psychopathology through differences in memory for the gist of (right hemisphere) versus the details of (left hemisphere) negative emotional episodes (reviewed in Cahill, 2010; Hamann, 2005). It is important to note that these memory tasksdand many emotion processing tasksdemphasize emotions with a negative valence and that tasks that emphasize emotions with a positive valence may reveal different patterns of brain sex differences, for example, with women showing more frontal and temporal activity than men, and men showing more left amygdala activity than women (Stevens and Hamann, 2012).

610 PART | II Cognitive development

Partly due to variability across tasks and development, it is unclear whether there are sex differences in amygdala lateralization during emotion-related processing. Some studies provide no support for sex differences (Derntl et al., 2009; Habel et al., 2007), but men appear to have right-lateralized and women to have left-lateralized amygdala activity for negatively valenced emotional memories (Cahill, 2010; Hamann, 2005), consistent with a meta-analysis of fMRI and PET studies on a combination of emotion-related tasks (Wager et al., 2003). Other meta-analyses, however, suggest that sex differences disappear with improved meta-analytic methodology (e.g., incorporating effect sizes of original findings into the analysis and use of activation likelihood estimation; Filkowski et al., 2017; Sergerie et al., 2008).

27.4.3.5 Functional connectivity Research on sex differences in resting brain function, and its links to cognition, is emerging, following mounting interest in resting state functional connectivity in neuroscience more broadly (Shen, 2015). This is a relatively new area of research, so findings can be difficult to interpret due to the huge variability across studies in data processing, network definition, and statistical modeling. Yet, some reports of sex differences in full-brain and regional metrics of connectivity converge with what is known about sex differences in brain structure and task-related function. First, local functional connectivity density is greater in women than men, consistent with sex differences in brain volume. Second, there is suggestion that men have more right-lateralized resting state activity than women (Liu et al., 2009; Tian et al., 2011), consistent with data showing greater interhemispheric structural connectivity in women. Third, there are indications that the hippocampus is a functional hub (a highly connected region important for brain efficiency) for the resting state brain function of women, but not men (Lopez-Larson et al., 2011; Zuo et al., 2012), consistent with data showing that women have larger hippocampi than men. There are also sex differences in particular brain networks or sets of brain regions with synchronized activity during resting state. For instance, large studies (e.g., with samples sizes over 1400) and meta-analyses show that women have greater connectivity than men within the default mode network (DMN), which consists of the precuneus, lateral parietal regions, and the medial prefrontal cortex; this network is the hallmark of the resting state because it is more active during rest than tasks (Beltz et al. 2015a; Biswal et al., 2010; Mak et al., 2017; Zuo et al., 2010). There also seems to be a sex difference in sensory and motor networks, including visual and auditory networks, during rest, with men having greater connectivity than women (Filippi et al., 2013; Smith et al., 2014). The implications of this are unclear, but the differences may be related to sex differences in brain volume or body size. Finally, there is considerable interest in sex differences in the frontoparietal network, which is thought to play a key role in cognitive development, especially during adolescence (Vendetti and Bunge, 2014). Findings are mixed, with some studies reporting greater connectivity in women than men (Hjelmervik et al., 2014) and others reporting the opposite effect (Smith et al., 2014). Discrepancies may be due to study methodology, including the regions used to define different networks, as well as sample characteristics, such as size and age (e.g., Gao et al., 2015). Other resting state networks and amygdala-linked connectivity patterns have also been considered, with similarly mixed findings (Alarcon et al., 2015; Engman et al., 2016; Filippi et al., 2013; Lopez-Larson et al., 2011; Zuo et al., 2010). Importantly, sex differences in resting state brain function are beginning to be associated with cognition. Most findings are preliminary, but there is agreement that functional connectivity reflects something meaningful about cognition. For instance, patterns in resting state brain function align with patterns of brain activity occurring during cognitive tasks, indicating that resting state and task-based function mark overlapping neural processes (Cole et al., 2016). Also, links between the DMN, which shows a sex difference, and the frontoparietal network, which is implicated in cognitive processing, are expectedly associated with working memory in young adults (Keller et al., 2015). Finally, explicit links to sex differences have been reported, with one study showing a relation between the gendered nature of cognitive profiles (consisting of motor, spatial, emotion, and reasoning tasks) and the gendered nature of functional connectivity profiles; gendered profiling was based on machine learning algorithms in a large sample of nearly 700 children, adolescence, and young adults (Satterthwaite et al., 2015). Findings are mixed, however, in work concentrating on particular networks thought to underlie cognition, and discrepancies are likely due to limited sample sizes and methodological differences across studies (Douw et al., 2011; Sole-Padulles et al., 2016). Although future work with converging methods is needed, it is clear that studies of functional connectivity will provide important future insights into the development of sex differences in the brain and behavior, including cognition.

27.4.3.6 Development of sex differences in brain function There are fewer studies of functional brain development than structural brain development, although structure and function are certainly interdependent. Nonetheless, a consistent finding regarding functional network development is that brain

Sex differences in brain and behavioral development Chapter | 27

611

function in children is driven by short-range connections among anatomically close regions, whereas in adults, many local connections are replaced by long-range ones with overlapping functionality, reflecting greater neural integration with development (Fair et al., 2007, 2009; Power et al., 2010). There are many interesting observations about this anatomicallylocal to functionally-similar developmental connectivity pattern, despite the lack of investigations into sex differences in children and adolescents (for an example related to adult aging, see Scheinost et al., 2015). First, the pattern was detected for the DMN before other networks, as the DMN was the early focus of resting state research (Fair et al., 2008; Kelly et al., 2009; Power et al., 2010). Second, the pattern ontogenetically unfolds for networks involved in sensory and motor processing before networks involved in associative processing (Grayson and Fair, 2017). Third, the pattern is so pervasive that machine learning algorithms can use it to predict whether a given brain belongs to a child (aged 7e11 years) or to an adult (aged 24e30 years) with 91% accuracy (Dosenbach et al., 2010). This finding has been recently questioned, however, with emerging evidence that motion (which is greater in children than adults) could produce similar local-to-diffuse connectivity patterns across development (Power et al., 2012); in corrected studies addressing motion, the pattern holds, but it is attenuated (Satterthwaite et al., 2012). Beyond investigations of resting state functional brain development, studies of task-related brain development suggest that brain activity underlying cognitive task performance becomes less diffuse and more fine-tuned with age (Casey et al., 2005; Giedd and Denker, 2015; Giorgio et al., 2008), with adolescence being a particularly important period of developmental change. Brain function in adolescence is likely affected by the different trajectories of gray and white matter development, and these brain changes may underlie the increased risk-taking behavior characteristic of this developmental period. Development of limbic regions, which are implicated in affective processing, peaks in early-to-mid adolescence, whereas the prefrontal cortex, which is thought to subserve cognitive control functions, is among the last brain regions to undergo cortical thinning and myelination in early-to-mid adulthood. The discordant maturational timing between these affective and cognitive neural processing networks is thought to result in limited top-down control of responses to appetitive stimuli such as peers and addictive substances (Casey et al., 2011; Galvan, 2013; Shulman et al., 2016; Steinberg, 2008). Because girls have faster brain development compared to boys, they are thought to experience a shorter period of discordance between limbic and prefrontal regions, and therefore, to be less likely to engage in the type of risktaking behavior that emerges in adolescence (e.g., risky sex and substance use). This dual systems model has been critiqued, however, for being overly simplistic (e.g., a similar model applies to depression), devoid of context effects (including other aspects of brain development in adolescence), and failing to consider individual differences (Crone and Dahl, 2012; Pfeifer and Allen, 2012; Somerville, 2016). Data explicitly examining the mismatch are surprisingly rare, and results are mixed in the few relevant studies. For instance, a study of structural brain development showed that individual differences in the neural mismatch between development of prefrontal and subcortical regions were not related to risk taking (Mills et al., 2014), whereas a functional connectivity study showed that connectivity between the dorsolateral prefrontal cortex and thalamus was associated with reward-related learning and actually mediated the link between learning and age (van Duijvenvoorde et al., 2016). Methods in the two studies obviously differed, and there is reason to expect greater correspondence between functional brain activity and behavior than between structural brain development and behavior (Section 4.2.5). Future work would likely benefit from utilizing advanced connectivity methods that accurately model individual differences in the direction of connections between prefrontal and subcortical regions (as described in Beltz, 2018a).

27.4.3.7 Implications of sex differences in brain function There are several conclusions that emerge from studies on the development of sex differences in cognition-related brain activation. Regarding lateralization, the evidence is mixed, but it appears that women are more likely than men to process information in both hemispheres. The lateralized activity of men appears to contribute to some cognitive advantages (e.g., in spatial skills) and disadvantages (e.g., in language). Regarding resting state functional connectivity, women tend to have greater connectivity within the DMN, whereas men may have greater connectivity within sensorimotor regions. Although the cognitive implications of these sex differences are not yet clear, it is clear that functional brain networks reflect cognition, and it is likely that sex-related processes contribute. Regarding sex differences in neural activation to spatial, language, and emotion-related tasks, a simplified summary is provided in Fig. 27.4. With respect to spatial tasks, particularly mental rotations, both men and women engage parietal regions, but women engage frontal regions that men typically do not, and men engage some left hemisphere regions that women typically do not (Fig. 27.4A). With respect to language tasks, women recruit more bilateral frontal and temporal regions than do men (Fig. 27.4B). Thus, for spatial and language brain activationetask performance links, performance accuracy seems to be associated with activation of different brain regions for men and women, perhaps reflecting the use of

612 PART | II Cognitive development

FIGURE 27.4 Summary of sex differences in neural activation to sex-typed tasks. Colors indicate general areas of brain activation, with blue for males, red for females, and purple for both sexes. (A) Spatial tasks; right: right hemisphere, left: left hemisphere. (B) Language tasks; right: right hemisphere, left: left hemisphere. (C) Amygdala during emotion-related tasks depicted on coronal slice; right: facial recognition, left: emotional memories (see text for details).

sex-typed strategies. With respect to the processing of emotional faces, men show greater bilateral amygdala activity than do women (Fig. 27.4C, right), perhaps reflecting their need to work harder than women to decode emotions. But, with respect to negative emotional memory, women show greater left amygdala activity than men, and men show greater right amygdala activity than women (Fig. 27.4C, left). It is important to note that evidence regarding sex differences in brain function during performance of spatial, language, and emotion-related tasks is not completely consistent, for several reasons. It is obvious that task performance, sample size, and participant characteristics, especially age, matter. But, tasks and stimulus presentation methods vary across studies, too, making brain activation patterns difficult to compare. The scanning environment also constrains behavioral assessments (e.g., space and stimulus presentation), so the behavior measured in the scanner likely differs from the behavior that shows a sex difference outside the scanner (e.g., Peters and Battista, 2008).

Sex differences in brain and behavioral development Chapter | 27

613

Nonetheless, the sexes generally engage similar brain regions to solve cognitive tasks, even though they appear to diverge in some notable ways. The sex differences primarily reflect two brain processes: efficiency and integration. Greater brain activity in the sex that typically performs worse on a behavioral task than the sex that performs better can be understood in terms of neural efficiency. When performance is poor, the task is seen to be difficult, and thus, more neural resources are recruited to solve it. This interpretation is consistent with findings from studies in which task difficulty was manipulated: The neural substrates of a task become more widely distributed as task difficulty increases (e.g., Gur et al., 2000). Greater brain activity in the sex that typically performs better on a behavioral task than the sex that performs worse can be understood in terms of neural integration. When performance is good, the task elicits a diverse set of associations, subserved by greater neural engagement. This interpretation is consistent with findings on the neural systems underlying intelligence, as functional connectivity among brain regions is positively associated with intelligence (Song et al., 2008). Whether efficiency or integration is invoked as an explanation for functional brain sex differences is likely dependent upon the nature of the task (e.g., complexity) and methodology (e.g., sample characteristics, task design, and data analysis procedures); they are also not mutually exclusive explanations.

27.5 Explanations for brain sex differences The brain is plastic, and its development is influenced by a combination of genetic, epigenetic, hormonal, and social experiential effects throughout the life span. Most aspects of brain development reflect a dynamic interplay among these factors and are therefore difficult to isolate and investigate (Cosgrove et al., 2007; McCarthy, 2016; McCarthy and Arnold, 2011; McEwen, 2016; Poldrack, 2000). Animal models are important for informing examinations of human neural processes. Many brain regions that show sex differences in human beings correspond to sexually dimorphic brain regions in nonhuman animals (Goldstein et al., 2001), and animal studies have helped to guide many human studies on sex differences (McCarthy, 2016); however, the degree to which animal findings generalize to human beings is not always clear, particularly with respect to cognitive domains that are species specific, such as language. Research on the explanations for human brain sex differences is relatively new but quickly escalating; thus, the studies reviewed below must be interpreted critically and require replication. Parallel to the discussion of the explanations for psychological sex differences, influences on brain sex differences of socialization, genes, and particularly hormones (prenatal, adolescent, circulating, and exogenous) are considered in the following.

27.5.1 Socialization perspectives Brain sex differences are usually considered to be purely innate, but this is a gross oversimplification. The brain changes in response to experiences, and social experiences are gendered or have gendered consequences across development (reviewed in Blakemore et al., 2009; Ruble et al., 2006), making socialization influences on brain sex differences important to consider (Cosgrove et al., 2007). Although sex-differential experiences are not measured in most studies of sex differences in brain structure and function, socialization effects could account for variability in findings and inconsistencies across reports. For example, there is increasing evidence that childhood stress (e.g., reflected by low SES and cortisol) influences brain function for girls and boys in different ways (e.g., Burghy et al., 2012), including in ways linked to cognition: During a response inhibition task, adolescent girls from low SES backgrounds performed worse than boys from low SES backgrounds and had less connectivity between the dorsolateral prefrontal cortex and anterior cingulate cortex; this reduced connectivity was presumably (but only partially) compensated for by greater anterior cingulate cortex activity in girls than in boys (Spielberg et al., 2015b). There is much work to be done regarding socialization influences on brain sex differences, particularly examining the ways in which training on cognitive tasks might differentially mediate (or be mediated by) brain sex differences. For example, sex differences in brain activation during the performance of spatial tasks might reflect sex-typed strategy use during task completion or be reduced (or even eliminated) through training. Thus, the sex difference in brain activation during spatial task performance might be partially explained by sex-differential experiences with spatial stimuli, and consequently, learned strategies for manipulation of those stimuli. Evidence from a recent fMRI training study suggests that this is an important area for future work: Children were scanned during a mental rotations task before and after either spatially related play with blocks or nonspatially related play with a board game; groups did not differ prior to the play intervention, but after the intervention, the block play group had improved mental rotations performance and greater activity in frontal and motor regions compared with the other group (Newman et al., 2016). The sample was small, though, and primarily male, so investigations of cognitive training effects on brain function that are statistically powered to compare males and females are warranted.

614 PART | II Cognitive development

27.5.2 Genetic perspectives There is little conclusive evidence regarding sex chromosome contributions to human brain development and the ways in which genes and the environment interact to influence the brain in sex-differential ways. Nonetheless, it is clear that this is an important area of investigation. Findings from a postmortem study led to the conclusion that “sex-biased gene expression in the adult human brain is widespread both in terms of the number of genes and range of brain regions involved” (Trabzuni et al., 2013, p. 5). Moreover, twin data show greater genetic influences on white matter development in boys than in girls (Chiang et al., 2011) and suggest that sex differences in heritability change across development, with boys having greater genetic influences on cortical thickness in the temporal lobe (near language regions) in childhood and in the frontal lobe in adolescence compared to girls (Schmitt et al., 2014). Although effects are small in all studies, other findings are consistent in showing greater genetic influences for boys than girls. There is evidence for sex-differential gene expression in the cingulate cortex, with greater expression in males than females (Gershoni and Pietrokovski, 2017). Intriguing preliminary evidence also indicates that several genes on the Y chromosome are expressed in the human male prenatal brain at midgestation (Reinius and Jazin, 2009) and the adult male brain (Vawter et al., 2004). It is important to note, however, that development of the external genitalia is underway by midgestation, hormone levels change with puberty in adolescence, and complete maturation has occurred by adulthood, so genetic findings may reflect hormones (e.g., endogenous testosterone) or experiential effects on the brain. Some insight into the role of the X chromosome in brain sex differences is provided by studies of Turner syndrome (TS), a DSD in which girls and women have a single (X) sex chromosome. Compared with unaffected controls, girls and women with TS have decreased parietal lobe volume and surface area as well as decreased activation in parietal lobe regions during a variety of visuospatial tasks (e.g., Green et al., 2014); they also have increased amygdala volume and increased amygdala activation during the recognition of fearful faces. These findings are consistent with their cognitive profile, which is characterized, in part, by visuospatial and emotion recognition deficits (reviewed in Knickmeyer and Davenport, 2011; Zhao and Gong, 2017). There is also emerging evidence of functional connectivity differences between women with TS and controls, with some differences linked to cognition (Xie et al., 2017), but replications are needed. The implications of these findings for brain sex differences in typical samples are unclear for several reasons. First, brain differences are confounded by cognitive performance differences: TS and unaffected control groups generally differ on task performance in functional imaging studies. Second, sex hormone production is low in individuals with TS; thus, differential patterns of brain structure and function might reflect sex hormone influences instead of X-chromosome influences on the brain. This is consistent with evidence showing that estradiol replacement in adolescents with TS reduces differences between girls with TS and control girls (Lepage et al., 2013). Third, it is unclear whether findings in individuals with TS reflect effects of the X chromosome, having a single (be it X or Y) sex chromosome, or X-chromosome dosage. Effects of X-chromosome dosage have been investigated by comparing brain structure in men with Klinefelter syndrome (XXY), which is another DSD, to unaffected men and women. Unfortunately, results from these studies are also difficult to interpret because (as in studies in women with TS) X-dosage effects are confounded with effects of sex hormones (for discussion, see Bryant et al., 2011; Lentini et al., 2013). Finally, there is recent neuroimaging work in women with CAIS; because they have XY karyotypes but no effective androgen exposure, the ways in which they differ from female controls likely reflect Y chromosome effects, whereas the ways in which they differ from male controls likely reflect androgen effects (prenatal, pubertal, and circulating) or genderrelated socialization (women with CAIS are generally reared and identify as female). In a study of brain structure and functional connectivity, women with CAIS were similar to control women, but different from control men, in measures of parietal and occipital cortical thickness, FA in several white matter tracts including in the corpus callosum, and amygdalarelated and DMN functional connectivity; they were similar to men, though, in motor cortex thickness and caudate volume (Savic et al., 2017). In a study of brain activation during mental rotations performance, women with CAIS showed brain activity that did not differ from control women, with men outperforming and showing greater parietal activation than both groups of women (van Hemmen et al., 2016). Similarly, in a study of brain activation during the viewing of sexual images, men showed greater amygdala activity than control women and women with CAIS, who did not differ from each other (Hamann et al., 2014). These findings generally converge to suggest that androgens and/or socializationdand not genes on the Y chromosomedinfluence (at least some specific) aspects of sex differential brain structure and activity. But samples are small, and results could also be affected by ovarian hormones (women with CAIS are often treated with exogenous hormones). Thus, despite this novel work with rare samples, there is still much opportunity for future studies on sex chromosome influences on brain sex differences.

Sex differences in brain and behavioral development Chapter | 27

615

27.5.3 Hormone perspectives Sex hormones are the most investigated influence on sex differences in the human brain. Most work concerns androgens, but there is emerging consideration of ovarian hormone influences (Beltz and Moser, 2019). There is evidence from clinical and typical samples for prenatal, adolescent, circulating, and exogenous sex hormone influences on the brain, but data and their interpretations are complex, so conclusions must be tentative.

27.5.3.1 Prenatal hormone influences on brain sex differences The influence of androgens present during early development on structural brain sex differences has been examined in individuals with CAH and in typical children whose amniotic testosterone levels were assessed. As discussed earlier, androgens masculinize a variety of characteristics in girls and women with CAH (including activity interests, some social behaviors, and spatial skills). Structural imaging studies reveal effects that are generally more consistent with the disease process than with prenatal hormone effects on the brain. For instance, smaller amygdala volumes were reported in both males and females with CAH than in unaffected males and females (Merke et al., 2003), likely reflecting effects of postnatal cortisone treatment in individuals with CAH. Moreover, women with CAH were recently reported to have reduced white matter integrity as well as smaller hippocampal volumes than female controls (Webb et al., 2018); this coincided with poorer performance on working memory and processing speed assessments and is consistent with early reports of white matter abnormalities and hippocampal atrophy in individuals with CAH (Bergamaschi et al., 2006; Nass et al., 1997). Findings regarding prenatal androgen influences on brain function suggest altered amygdala and hippocampal activity in individuals with CAH, but interpretation is not straightforward. Consider findings from one set of studies in which images of emotional faces were viewed, rated, and recalled (Ernst et al., 2007; Mazzone et al., 2011). Brain activation findings showed that girls with CAH had greater amygdala activation while viewing negative facial expressions and less hippocampal activation while recalling emotional faces compared with unaffected girls; however, there were also group differences in behavioral ratings and task performance, complicating interpretation of brain activation differences. In a separate PET study, women with CAH did not differ from unaffected women in their neural response to olfactory stimuli: Both groups displayed increased amygdala activation to a masculine pheromone and increased hypothalamus activation to a feminine pheromone, compared with men who showed the reciprocal pattern of brain activity (Ciumas et al., 2009). Information about prenatal androgen effects on brain function also comes from typical samples, in which testosterone levels were measured in amniotic fluid and linked to performance on dichotic listening tasks in childhood, brain structure, and functional connectivity. Amniotic testosterone levels were positively associated with left-lateralized language processing in one sample of 6-year-old girls and boys (Lust et al., 2010) and another sample of 10-year-old girls (Grimshaw et al., 1995). This is consistent with the sex difference in language lateralization reported in children and adults. Amniotic testosterone levels have also been associated with right-lateralized emotion processing (measured with an emotioneword dichotic listening task) in 10-year-old boys (Grimshaw et al., 1995), consistent with meta-analytic findings on rightlateralized amygdala activation in men (but not women) during emotion-related tasks (Wager et al., 2003). In typical boys aged 8e11 years, testosterone levels from amniotic fluid have also been linked to gray matter volumes in brain regions that show sex differences. Specifically, testosterone was positively related to gray matter volumes in temporal regions but negatively related to volumes in occipital and frontal regions (Lombardo et al., 2012). Finally, functional connectivity within the DMN decreased linearly with increases in amniotic testosterone for adolescent boys, but there was no relation for adolescent girls (Lombardo et al., 2019); although circulating testosterone was also negatively linked to DMN connectivity in boys, the association is difficult to interpret because it did not account for the amniotic testosterone result. These findings are consistent with the sex difference in DMN connectivity and suggest that both prenatal testosterone and adolescent testosterone are related to reduced DMN connectivity. The associations between amniotic testosterone and sex differences in brain structure and function are intriguing, but the method has significant limitations (discussed in Section 3.3.2.2), and more research with other methods is necessary to determine why results are often seen in one sex, but not the other.

27.5.3.2 Adolescent hormone influences on brain sex differences There is considerable interest in the ways that brain changes in adolescence reflect the direct effects of pubertal sex hormones. Most work has not been able to differentiate organizational (permanent) effects of hormones from activational (transient) effects, but an increasing number of longitudinal studies facilitates inferences about development (Herting and Sowell, 2017). Generally, pubertal increases in sex-specific gonadal hormones (estrogen in girls and testosterone in boys) are associated with decreases in cortical gray matter volume (reviewed in Goddings et al., 2019; Herting and Sowell, 2017;

616 PART | II Cognitive development

Peper et al., 2011). There are early indications from twin studies that these hormoneebrain structure links are driven by shared environmental influences in girls, but samples are small and require replication (Brouwer et al., 2015). When boys and girls are matched on pubertal development, however, sex differences in links between testosterone and regional gray matter volumes are more difficult to interpret, with boys showing fewer significant associations than girls (Bramen et al., 2012). This apparent discrepancy may be partially explained by androgen sensitivity. The efficiency of the androgen receptor gene has been suggested to moderate the influence of sex hormones on cortical thinning, as fewer repeats of a functional polymorphism within the gene predicted more “masculinized” patterns of maturation (Paus, 2010; Raznahan et al., 2010). There is also consistent evidence that sex-specific gonadal hormones are associated with increases in white matter volume across adolescence, but the link may be stronger in boys than in girls (reviewed in Goddings et al., 2019; Herting and Sowell, 2017; Peper et al., 2011), perhaps because the efficiency of the androgen receptor gene moderates the association (Paus et al., 2010; Perrin et al., 2008). Testosterone levelsdbut not pubertal stagedhave also been linked to FA increases in boys, with estradiol levels negatively associated with FA in girls (Bava et al., 2011; Herting et al., 2012). Differences in brain links with hormone levels versus pubertal stage may indicate that the measures reflect different constructs or have unique limitations (discussed in Dorn et al., 2006). There is also evidence for pubertal hormone influences on specific aspects of brain structure, with a focus on regional volumes and on cortical thickness. Regarding regional structures, pubertal development is generally positively linked to hippocampal and amygdala volumes (Blanton et al., 2012; Bramen et al., 2011; Goddings et al., 2014; Herting et al., 2014; Satterthwaite et al., 2014), but sometimes links are not found or indicate that volumes decrease with pubertal development (e.g., Bramen et al., 2011; Satterthwaite et al., 2014). Regarding cortical thickness, pubertal influences have been detected, but converging patterns have not yet emerged (e.g., Herting et al., 2015; Nguyen et al., 2013). Across thesedand all studies of pubertal hormone influences on brain structuredindividual sets of findings may diverge from the generally convergent results presented here: Discrepancies are likely due to sex, pubertal development measure (e.g., self-report, physician rating, hormone assay), and age, including how age is statistically modeled due to its colinearity with pubertal development (Berenbaum et al., 2015; Dorn and Biro, 2011). Thus, these features are important to consider when designing future studies. Other important considerations for future work concern sex-related brain changes that vary with pubertal status (i.e., current pubertal stage) versus with pubertal timing (i.e., early vs. late pubertal development compared with peers regardless of current status). Pubertal hormone effects on brain function have been studied less than effects on brain structure, but the literature is quickly growing (reviewed in Goddings et al., 2019). Some insight is provided by studies on the neural substrates of reward processing (Forbes and Dahl, 2010). Results generally show that advanced pubertal development, particularly circulating testosterone, is related to increased striatum and nucleus accumbens activation to reward; the relation has been found for both sexes, although not in all studies, potentially due to small sample sizes and restricted ranges of pubertal development (Braams et al., 2015; Forbes and Dahl, 2010; Op de Macks et al., 2011, 2016). Other studies have investigated the link between pubertal development and the viewing of emotional faces, but results are inconsistent regarding amygdala activation, with some studies reporting increased activity with advancing puberty and others reporting decreased activity with advancing puberty (e.g., Ferri et al., 2014; Moore et al., 2012); discrepancies are likely due to differential foci on negative (e.g., vs. neutral) emotions across studies. There is also evidence of testosterone links to decreased amygdalaorbitofrontal cortex activity in both boys and girls (Spielberg et al., 2015a). Finally, some studies have begun to investigate links between pubertal development and the neural substrates of social emotional processing known to change in adolescence; current findings vary, but this is a promising area for future work (Goddings et al., 2019; Guyer et al., 2016). The influence of pubertal hormones on brain sex differences has also been investigated in boys with extremely early (disordered) puberty (Mueller et al., 2009, 2011a, 2011b). Results are difficult to interpret, however, because of age and performance differences between clinical and control groups. Nonetheless, this research illustrates a valuable approach (i.e., neuroimaging individuals with atypical pubertal timing) to understanding hormonal influences on brain development.

27.5.3.3 Circulating hormone influences on brain sex differences Examinations of structural and functional brain changes during different phases of the menstrual cycle and at menopause provide indirect evidence for activational effects of ovarian hormones on the brain, with most work focusing on estrogen. Generally, the hippocampus and frontal lobes are implicated as the primary sites for estrogen effects on brain structure in humans; thus, estrogen effects on brain function are most evident in verbal memory tasks, which show a sex difference that favors females, and are subserved by the hippocampus and frontal lobes (reviewed in Maki and Resnick, 2001).

Sex differences in brain and behavioral development Chapter | 27

617

In high-estrogen phases of the menstrual cycle as compared with low-estrogen phases, women generally have greater hippocampal volume, less amygdala volume, and greater frontal and temporal activation during the completion of verbal and mental rotation tasks (Dietrich et al., 2001; Fernández et al., 2003; Konrad et al., 2008; Lisofsky et al., 2015; Ossewaarde et al., 2013; Pletzer et al., 2010; Protopopescu et al., 2008; Schöning et al., 2007); across cycle phases, estradiol has also been negatively linked to volume of the anterior cingulate cortex (De Bondt et al., 2013). The implications of this research are not clear because there is currently little evidence for a relation between estrogen-influenced neural processes and task performance (Craig et al., 2008; Dietrich et al., 2001; Fernández et al., 2003; Konrad et al., 2008), but it highlights that the brain is plastic. There is also increasing investigation into menstrual cycle effects on functional connectivity. Early findings suggest that high-estrogen phases are linked to increased hippocampal connectivity and DMN connectivity compared with low-estrogen phases (Lisofsky et al., 2015; Petersen et al., 2014), providing evidence for estrogen influences on sex differences in connectivity, but effects are not always found (e.g., Weis et al., 2019). Menstrual cycle effects have also been investigated, but not detected, in the frontoparietal and auditory networks (Hjelmervik et al., 2014; Weis et al., 2019); this is not surprising given the weak evidence for sex differences or differences favoring men, respectively, in these networks (Section 4.3.5). Research on menstrual cycle influences on brain structure and function is in its relative infancy, with small samples and variable methodology across studies, so results require replication. Also, much research has focused on estrogen, but there are large fluctuations in progesterone across the menstrual cycle that are beginning to be related to brain function (e.g., Arelin et al., 2015), so much future work is needed. At menopause, there is evidence for decreases in whole brain, frontal lobe, and hippocampal volumes (Goto et al., 2011; Robertson et al., 2009). There is also evidence that prefrontal and hippocampal activity differs according to menopausal status during verbal memory tasks. Compared with premenopausal women, postmenopausal women show defeminized neural patterns, including increased prefrontal activation, reduced hippocampal deactivation, and increased hippocampal connectivity; these patterns are related to low estradiol (Jacobs et al., 2016, 2017). Because there are no differences in behavior, findings suggest that brain function mediates the link between estradiol and verbal task performance, reflecting compensation or strategy use. Menopausal status is linked to brain activity during other tasks as well, but this is novel work and samples are often small, so replication is required. For instance, postmenopausal (but not pre- or perimenopausal) women had prefrontal cortex activation during an emotion task, with estradiol negatively related to prefrontal activity (Berent-Spillson et al., 2017), and they also had reduced activation compared with premenopausal women in the amygdala, thalamus, and anterior cingulate cortex while viewing erotic videos, with regional activity positively linked to estradiol (Kim and Jeong, 2017).

27.5.3.4 Exogenous hormone influences on brain sex differences There are also opportunities to study hormone influences on brain sex differences by leveraging natural experiments of exogenous hormones, that is, individuals taking synthetic androgens, estrogens, and progestins. The mechanisms through which exogenous hormones influence the human brain are unclear, as exogenous hormones likely bind to endogenous hormone receptors in the brain, interact with endogenous hormones in complex ways, and are difficult to measure (because there are not assays available for all of them). Nonetheless, findings from both clinical and typical samples are intriguing and provide an underutilized resource for future work. Changes in overall brain volume and subcortical structures were observed in samples of transsexuals after several months of hormone treatment (i.e., antiandrogens þ estrogens for male-to-female, and androgens for female-to-male; Hulshoff Pol et al., 2006; Zubiaurre-Elorza et al., 2014). Male-to-female transsexuals showed decreases in overall brain volume and subcortical structures, and female-to-male transsexuals showed increases in overall brain and hypothalamus volumes. These findings are consistent with structural sex differences reported above and suggest that circulating androgens increase overall and some regional brain volumes. Results of (cross-sectional) studies also show that white matter diffusivity was positively linked to testosterone levels, but that testosterone alone could not account for differences among male-to-female, female-to-male, control female, and control male subgroups (Kranz et al., 2014) and that male-to-female transsexuals had larger amygdala and putamen volumes than control women (Mueller et al., 2016). Thus, results converge to suggest that brain sex differences have gendered influences beyond exogenous hormones and point to prenatal androgens or interactions between prenatal and circulating hormones as key contributors. Exogenous hormone influences on brain structure and function can also be examined via studies of women using hormonal contraceptives, with combined OC pills being the most common form. Combined pill packets contain a synthetic estrogen and progestin (that varies in androgenicity) for about 3 weeks a month (active phase) and a placebo pill for about 1 week. Data on OC influences on brain structure suggest that users have larger frontal and temporal regions than naturally

618 PART | II Cognitive development

cycling women (De Bondt et al., 2013; Pletzer et al., 2010), but effects are complex, depending upon the androgenicity of the progestin in the pill and the length of OC use (Pletzer et al., 2015). Cognitive taskerelated activation differences between OC users and naturally cycling women have begun to be investigated (Pletzer et al., 2014, 2015), but unfortunately, traditionally sex-typed spatial and language skills have yet to be assessed. While viewing negatively valenced emotional images, though, OC users show less amygdala reactivity compared with naturally cycling women (Petersen and Cahill, 2015). When compared with data from typical males (Section 4.3.4), this result suggests that OC users have a more feminized neural response than naturally cycling women. There is also early indication that OC use affects functional connectivity, with pill users having less DMN connectivity than naturally cycling women (Petersen et al., 2014). When compared with data from typical males (Section 4.3.5), this result suggests that OC users have a masculinized neural response. Thus, studies about OC effects on brain structure and function are somewhat inconsistent: They highlight a valuable way to examine exogenous estrogen, progestin, and androgen effects on the brain, but current data are difficult to interpret. Effects depend upon sample size (which is usually small), active or placebo pill phase (De Bondt et al., 2013; Petersen et al., 2014), menstrual cycle phase of the naturally cycling comparison group (Pletzer et al., 2010, 2014), and androgenicity of the pills (which is not consistently reported; e.g., Petersen and Cahill, 2015; Petersen et al., 2014). Given these complexities, it is unlikely that a simple story of neural masculinization or feminization due to OC use will be revealed, so future methodologically rigorous research is needed. Finally, studying estrogen therapy around the time of menopause provides insight into exogenous hormone influences on brain structure and function. Although estrogen therapy, which often cooccurs with progesterone prescription, of a relatively short duration (about 5 years) appears to offset hippocampal volume reductions that characterize menopause (Erickson et al., 2010; Lord et al., 2008), longer-term therapy appears to have a negative effect on hippocampal volumes as well as volumes in frontal regions (Lord et al., 2008; Resnick et al., 2009; Zhang et al., 2016), consistent with data on exogenous hormone influences in male-to-female transsexuals and indications that length of OC use alters neural effects. Postmenopausal women undergoing hormone therapy also have greater activation of frontal regions during verbal and working memory tasks in comparison with their pretherapy performance or to postmenopausal women who never used estrogen therapy (Berent-Spillson et al., 2010, 2015; Dumas et al., 2010; Li et al., 2015; Persad et al., 2009; Shaywitz et al., 1999); frontal regions tend to be smaller, but more active, in therapy users. Interestingly, increased activation with hormone therapy generally does not predict better memory in neuroimaging studies, perhaps because statistical power is low. Estrogen therapy users, however, had greater cerebral blood flow and better memory task performance than nonusers in research using PET, suggesting that cerebral blow flood may be one mechanism through which estrogen influences cognition (and is best measured by PET; Maki and Resnick, 2000; Resnick et al., 1998). Hormone therapy around menopause has also been related to other neural processes; for example, users have been shown to have increased striatal activity to reward in a placebo-controlled study (Thomas et al., 2014), consistent with studies of adolescent hormone influences on reward processing. This is a promising area for future research. In summary, there is little conclusive evidence regarding social, genetic, or hormonal contributions to sex differences in human brain structure and function, but research in this area is increasing rapidly. Most of the available evidence concerns hormones. Convergent findings suggest sex hormone contributions to the size and function of the hippocampus and amygdala, with a particular role for prenatal androgens. There is also indication that sex hormones are linked to gray matter volume, especially in adolescence. Finally, there is evidence for endogenous and exogenous effects of sex hormones on the adult brain, particularly for estrogen influences on frontal lobe structure and function. Although there are few data linking sex hormone influences on the brain with behavior, research on brain function (both task-related and resting state connectivity) is on the rise. There is also limited evidence regarding sex chromosome expression in the human brain, and virtually no data regarding gendered socialization influences on brain sex differences. Thus, there is much opportunity for future research.

27.6 Conclusions and future directions As reviewed earlier, the sexes differ in significant ways in both behavior, including cognition, and the brain (including structure and function), although there is still much to be learned about the development and causes of the differences. It is important to be aware of the ways in which study methodology can affect conclusions regarding the existencedor notdof sex differences. Many differences are small, but some are moderate-to-large, and change with age (and over chronological time; Miller and Halpern, 2014), so sample size, participant selection and characteristics, and statistical modeling all influence results. There is good evidence that both sex hormones, especially androgens, and social factors influence the development of human sex-related psychological characteristics, although little is currently known about the pathways by which

Sex differences in brain and behavioral development Chapter | 27

619

socialization acts on and modifies biological predispositions (see discussion in Berenbaum and Beltz, 2018; Berenbaum et al., 2011). Determinants of brain sex differences are still largely unknown, but sex hormones seem to be a key contributor. It is also important to remember that the brain is plastic and that sex differences emerge from the interplay of genes, sex hormones, and social experiences, so brain sex differences cannot be assumed to be innate. There are several opportunities to study the etiology of sex differences in brain and behavior. Sex differences in behavior and brain development can be studied in natural experiments in which prenatal hormone exposure is sex atypical, such as DSD, particularly CAH. Increasingly, the topic is also being studied in typical individuals, with prenatal hormones measured in amniotic fluid and neonatal hormones measured in blood or urine. Such studies help to separate the relative influences of early (prenatal and neonatal) hormones and postnatal socialization, and, more importantly, allow the study of their interplay (for reviews, see Berenbaum and Beltz, 2011, 2016; Blakemore et al., 2009). Moreover, sex differences in adolescent brain development can be studied in individuals with disordered pubertal development. This provides an opportunity to test hypotheses about permanent changes to the brain induced by sex hormones (Giedd et al., 2006; Sisk and Zehr, 2005). Sex differences in adolescent brain development can also be studied in typical individuals, by linking variations in pubertal development to changes in behavior and in the brain; this has been a particularly active area of research in recent years (for reviews, see Goddings et al., 2019; Herting and Sowell, 2017)dalthough such studies often do not distinguish between organizational and activational effects. Finally, circulating and exogenous hormone influences on sex differences in the brain and behavior can be studied in women during the menstrual cycle and menopause and women who use the synthetic hormone treatments related to them. This is a relatively new research area, but it holds unique opportunities to examine biopsychosocial mechanisms underlying gendered phenomena (discussed in Beltz and Moser, 2019). The links between neural and behavioral sex differences are beginning to be understood, but they must be directly tested. It is not sufficient to find sex differences in regions of the brain known to subserve specific sex-related behavioral characteristics; sex differences in those regions of the brain must be shown to explain the behavioral sex differences. Given the complexity of the brain, including the interplay among cellular, regional, and global structure as well as localized function and functional connectivity, it is unlikely that such explanations will be simple; thus, revelations from the study of gendered neural processes will likely provide insights into the mechanisms underlying brainebehavior relations more broadly. There are several opportunities to understand direct links between sex differences in the brain and behavior. First, the meaning of sex differences in brain activationdwhen they reflect strategy use (i.e., when the sexes perform similarly) or behavioral differences (i.e., when the sexes differ in performance)dcan be understood by manipulating task difficulty during functional brain scanning. Such designs are rare in examinations of sex differences, but they are typical in other research areas (e.g., Poldrack, 2000). Second, causal inferences about links between brain and behavior can be strengthened by using directed connectivity mapping, which takes advantage of the time series nature of functional neuroimaging data or intensive longitudinal behavioral and cognitive data (e.g., Beltz and Gates, 2017). Work on functional and structural brain connectivity reviewed in this chapter has identified key networks, but directed connectivity mapping would go beyond network identification, allowing prediction of regional brain activity from activity in other brain regions or task performance (see Beltz, 2018a; Beltz and Molenaar, 2016). For example, directed functional connectivity mapping enables inferences regarding the direction of a connection between two brain regions: The connection might go from region A to region B for women but from region B to region A for men. Research on all forms of brain connectivity, however, is increasing quickly and will be important in the future. Third, neural effects of sex-differential experiences can be studied using longitudinal neuroimaging designs. Longitudinal studies are optimal for learning how genes, hormones, and socialization influence brain changes and for examining the permanence of the changes; they also reveal the plasticity of the brain and indicate how sex differences arise and how they might be modified (and the within-subject design avoids the confounding influence of sex differences in brain size). Notable large-scale efforts are underway. For example, the Adolescent Brain Cognitive Development (ABCD) study is following over 10,000 10-year-old boys and girls for 10 years, providing survey data on social experiences along with neural (via MRI), cognitive, genetic, and hormonal measures (Jernigan et al., 2018). In sum, the sexes are similar in many ways, but there are notable differences in their brain and behavior, including in cognition. An understanding of sex differences has implications for discussions about women’s underrepresentation in science and mathematics careers, for policies surrounding gender identity and sexual orientation, and for sex differences in the etiology, appearance, and treatment of psychopathology. Studying sex differences can also reveal individual differences generally. By identifying the differences, delineating their development, and investigating their causes, the mechanisms underlying variation in human health, disease, and behavior can be uncovered.

620 PART | II Cognitive development

Acknowledgments We dedicate this chapter to the memory of Dr. J. Elaine O. Blakemore, an eminent scholar of gender development, and our wonderful collaborator and friend. Elaine contributed to the previous version of this work, and her influence can still be seen in the comprehensive descriptions and thoughtful critiques contained in the current pages. We are lucky to have known and worked with her.

References Aitken, M., Steensma, T.D., Blanchard, R., VanderLaan, D.P., Wood, H., Fuentes, A., et al., 2015. Evidence for an altered sex ratio in clinic-referred adolescents with gender dysphoria. J. Sex. Med. 12 (3), 756e763. Alarcon, G., Cservenka, A., Rudolph, M.D., Fair, D.A., Nagel, B.J., 2015. Developmental sex differences in resting state functional connectivity of amygdala sub-regions. Neuroimage 115, 235e244. Allen, J.S., Damasio, H., Grabowski, T.J., Bruss, J., Zhang, W., 2003. Sexual dimorphism and asymmetries in the gray-white composition of the human cerebrum. Neuroimage 18 (4), 880e894. Allen, L.S., Gorski, R.A., 1990. Sex difference in the bed nucleus of the stria terminalis of the human brain. J. Comp. Neurol. 302 (4), 697e706. Allen, L.S., Gorski, R.A., 1991. Sexual dimorphism of the anterior commisure and massa intermedia of the human brain. J. Comp. Neurol. 312 (1), 97e104. Allen, L.S., Gorski, R.A., 1992. Sexual orientation and the size of the anterior commissure in the human brain. Proc. Natl. Acad. Sci. U.S.A. 89 (15), 7199e7202. Allendorfer, J.B., Lindsell, C.J., Siegel, M., Banks, C.L., Vannest, J., Holland, S.K., Szaflarski, J.P., 2012. Females and males are highly similar in language performance and cortical activation patterns during verb generation. Cortex 48 (9), 1218e1233. American Psychiatric Association, 2013. Diagnostic and Statistical Manual of Mental Disorders, fifth ed. American Psychiatric Association Publishing, Washington, D.C. Andreano, J.M., Dickerson, B.C., Barrett, L.F., 2014. Sex differences in the persistence of the amygdala response to negative material. Soc. Cogn. Affect. Neurosci. 9 (9), 1388e1394. Angold, A., Costello, E.J., Erkanli, A., Worthman, C.M., 1999. Pubertal changes in hormone levels and depression in girls. Psychol. Med. 29 (5), 1043e1053. Archer, J., 2004. Sex differences in aggression in real-world settings: a meta-analytic review. Rev. Gen. Psychol. 8 (4), 291e322. Arelin, K., Mueller, K., Barth, C., Rekkae, P.V., Kratzsch, J., Burmann, I., et al., 2015. Progesterone mediates brain functional connectivity changes during the menstrual cycle-a pilot resting state MRI study. Front. Neurosci. 9, 11. Arnold, A.P., 2009. The organizationaleactivational hypothesis as the foundation for a unified theory of sexual differentiation of all mammalian tissues. Horm. Behav. 55 (5), 570e578. Arnold, A.P., 2017. A general theory of sexual differentiation. J. Neurosci. Res. 95 (1e2), 291e300. Arnold, A.P., Breedlove, S.M., 1985. Organizational and activational effects of sex steroids on brain and behavior: a reanalysis. Horm. Behav. 19, 469e498. Arnold, A.P., Chen, X., 2009. What does the "four core genotypes" mouse model tell us about sex differences in the brain and other tissues? Front. Neuroendocrinol. 30 (1), 1e9. Asperholm, M., Högman, N., Rafi, J., Herlitz, A., 2019. What did you do yesterday? A meta-analysis of sex differences in episodic memory. Psychol. Bull. 145 (8), 785e821. Asperholm, M., Nagar, S., Dekhtyar, S., Herlitz, A., 2019. The magnitude of sex differences in verbal episodic memory increases with social progress: data from 54 countries across 40 years. PLoS One 14 (4), e0214945. Baenninger, M., Newcombe, N., 1989. The role of experience in spatial test performance: a meta-anaylsis. Sex Roles 20, 327e344. Bailey, J.M., Vasey, P.L., Diamond, L.M., Breedlove, S.M., Vilain, E., Epprecht, M., 2016. Sexual orientation, controversy, and science. Psychol. Sci. Public Interest 17 (2), 45e101. Baron-Cohen, S., Lutchmaya, S., Knickmeyer, R., 2004. Prenatal Testosterone in Mind. MIT Press, Cambridge, MA. Bava, S., Boucquey, V., Goldenberg, D., Thayer, R.E., Ward, M., Jacobus, J., Tapert, S.F., 2011. Sex differences in adolescent white matter architecture. Brain Res. 1375, 41e48. Beck, D.M., 2010. The appeal of the brain in the popular press. Perspect. Psychol. Sci. 5 (6), 762e766. Becker, J.B., Berkley, K.J., Geary, N., Hampson, E., Herman, J., Young, E. (Eds.), 2008. Sex Differences in the Brain: From Genes to Behavior. Oxford University Press, New York. Becker, J.B., Breedlove, S.M., Crews, D., McCarthy, M.M. (Eds.), 2002. Behavioral Endocrinology, second ed. MIT Press, Cambridge, MA. Begeer, S., Mandell, D., Wijnker-Holmes, B., Venderbosch, S., Rem, D., Stekelenburg, F., Koot, H.M., 2013. Sex differences in the timing of identification among children and adults with autism spectrum disorders. J. Autism Dev. Disord. 43 (5), 1151e1156. Bell, E.C., Willson, M.C., Wilman, A.H., Dave, S., Silverstone, P.H., 2006. Males and females differ in brain activation during cognitive tasks. Neuroimage 30 (2), 529e538. Beltz, A.M., 2018. Connecting theory and methods in adolescent brain research. J. Res. Adolesc. 28 (1), 10e25. Beltz, A.M., 2018. Gendered mechanisms underlie the relation between pubertal timing and adult depressive symptoms. J. Adolesc. Health 62 (6), 722e728.

Sex differences in brain and behavioral development Chapter | 27

621

Beltz, A.M., Berenbaum, S.A., 2013. Cognitive effects of variations in pubertal timing: is puberty a period of brain organization for human sex-typed cognition? Horm. Behav. 63 (5), 823e828. Beltz, A.M., Berenbaum, S.A., Wilson, S.J., 2015. Sex differences in resting state brain function of cigarette smokers and links to nicotine dependence. Exp. Clin. Psychopharmacol 23 (4), 247e254. Beltz, A.M., Gates, K.M., 2017. Network mapping with GIMME. Multivar. Behav. Res. 52 (6), 789e804. Beltz, A.M., Hampson, E., Berenbaum, S.A., 2015. Oral contraceptives and cognition: a role for ethinyl estradiol. Horm. Behav. 74, 209e217. Beltz, A.M., Loviska, A.M., Kelly, D., 2019. No personality differences between oral contraceptive users and naturally cycling women: implications for research on sex hormones. Psychoneuroendocrinology 100, 127e130. Beltz, A.M., Molenaar, P.C.M., 2016. Dealing with multiple solutions in structural vector autoregressive models. Multivar. Behav. Res. 51 (2e3), 357e373. Beltz, A.M., Moser, J.S., 2019. Ovarian hormones: a long overlooked but critical contributor to cognitive brain structures and function. Ann. N. Y. Acad. Sci. Beltz, A.M., Swanson, J.L., Berenbaum, S.A., 2011. Gendered occupational interests: prenatal androgen effects on psychological orientation to Things versus People. Horm. Behav. 60, 313e317. Berenbaum, S.A., 1999. Effects of early androgens on sex-typed activities and interests in adolescents with congenital adrenal hyperplasia. Horm. Behav. 35 (1), 102e110. Berenbaum, S.A., 2018. Beyond pink and blue: the complexity of early androgen effects on gender development. Child Dev. Persp. 12 (1), 58e64. Berenbaum, S.A., Baxter, L., Seidenberg, M., Hermann, B., 1997. Role of the hippocampus in sex differences in verbal memory: memory outcome following left anterior temporal lobectomy. Neuropsychology 11 (4), 585e591. Berenbaum, S.A., Beltz, A.M., 2011. Sexual differentiation of human behavior: effects of prenatal and pubertal organizational hormones. Front. Neuroendocrinol. 32 (2), 183e200. Berenbaum, S.A., Beltz, A.M., 2016. How early hormones shape gender development. Curr. Opin. Behav. Sci. 7, 53e60. Berenbaum, S.A., Beltz, A.M., 2018. From genes to behavior through sex hormones and socialization: the example of gender development. Twin Res. Hum. Genet. 21 (4), 289e294. Berenbaum, S.A., Beltz, A.M., Bryk, K., McHale, S., 2018. Gendered peer involvement in girls with congenital adrenal hyperplasia: effects of prenatal androgens, gendered activities, and gender cognitions. Arch. Sex. Behav. 47 (4), 915e929. Berenbaum, S.A., Beltz, A.M., Corley, R., 2015. The importance of puberty for adolescent development: conceptualization and measurement. Adv. Child Dev. Behav. vol. 48, 53e92 (San Diego: Elsevier Academic Press Inc). Berenbaum, S.A., Blakemore, J.E.O., Beltz, A.M., 2011. A role for biology in gender-related behavior. Sex Roles 64 (11e12), 804e825. Berenbaum, S.A., Bryk, K.K., Nowak, N., Quigley, C.A., Moffat, S., 2009. Fingers as a marker of prenatal androgen exposure. Endocrinology 150 (11), 5119e5124. Berenbaum, S.A., Bryk, K.L.K., Beltz, A.M., 2012. Early androgen effects on spatial and mechanical abilities: evidence from congenital adrenal hyperplasia. Behav. Neurosci. 126 (1), 86e96. Berenbaum, S.A., Snyder, E., 1995. Early hormonal influences on childhood sex-typed activity and playmate preferences: implications for the development of sexual orientation. Dev. Psychol. 31 (1), 31e42. Berent-Spillson, A., Briceno, E., Pinsky, A., Simmen, A., Persad, C.C., Zubieta, J.K., Smith, Y.R., 2015. Distinct cognitive effects of estrogen and progesterone in menopausal women. Psychoneuroendocrinology 59, 25e36. Berent-Spillson, A., Marsh, C., Persad, C., Randolph, J., Zubieta, J.K., Smith, Y., 2017. Metabolic and hormone influences on emotion processing during menopause. Psychoneuroendocrinology 76, 218e225. Berent-Spillson, A., Persad, C.C., Love, T., Tkaczyk, A., Wang, H., Reame, N.K., et al., 2010. Early menopausal hormone use influences brain regions used for visual working memory. Menopause 17 (4), 692e699. Bergamaschi, R., Livieri, C., Uggetti, C., Candeloro, E., Egitto, M.G., Pichiecchio, A., et al., 2006. Brain white matter impairment in congenital adrenal hyperplasia. Arch. Neurol. 63 (3), 413e416. Berman, S., West, K.L., Does, M.D., Yeatman, J.D., Mezer, A.A., 2018. Evaluating g-ratio weighted changes in the corpus callosum as a function of age and sex. Neuroimage 182, 304e313. Bishop, K.M., Wahlsten, D., 1997. Sex differences in the human corpus callosum: myth or reality? Neurosci. Biobehav. Rev. 21 (5), 581e601. Biswal, B.B., Mennes, M., Zuo, X.N., Gohel, S., Kelly, C., Smith, S.M., et al., 2010. Toward discovery science of human brain function. Proc. Natl. Acad. Sci. U.S.A. 107 (10), 4734e4739. Björkqvist, K., 2018. Gender differences in aggression. Curr. Opin. Psychol. 19, 39e42. Bjornholm, L., Nikkinen, J., Kiviniemi, V., Nordstrom, T., Niemela, S., Drakesmith, M., et al., 2017. Structural properties of the human corpus callosum: multimodal assessment and sex differences. Neuroimage 152, 108e118. Blakemore, J.E.O., Berenbaum, S.A., Liben, L.S., 2009. Gender Development. Psychology Press/Taylor & Francis, New York. Blanton, R.E., Cooney, R.E., Joormann, J., Eugene, F., Glover, G.H., Gotlib, I.H., 2012. Pubertal stage and brain anatomy in girls. Neuroscience 217, 105e112. Bleidorn, W., Arslan, R.C., Denissen, J.J.A., Rentfrow, P.J., Gebauer, J.E., Potter, J., Gosling, S.D., 2016. Age and gender differences in self-esteem: a cross-cultural window. J. Personal. Soc. Psychol. 111 (3), 396e410. Bonilla-Silva, E., 2017. Racism without Racists: Color-Blind Racism and the Persistence of Racial Inequality in America, fifth ed. Rowman & Littlefield, Lanham, MD.

622 PART | II Cognitive development

Bornstein, M.H., Hahn, C.-S., Haynes, O.M., 2004. Specific and general language performance across early childhood: stability and gender considerations. First Lang. 24 (3), 267e304. Braams, B.R., van Duijvenvoorde, A.C.K., Peper, J.S., Crone, E.A., 2015. Longitudinal changes in adolescent risk-taking: a comprehensive study of neural responses to rewards, pubertal development, and risk-taking behavior. J. Neurosci. 35 (18), 7226e7238. Bramen, J.E., Hranilovich, J.A., Dahl, R.E., Chen, J., Rosso, C., Forbes, E.E., et al., 2012. Sex matters during adolescence: testosterone-related cortical thickness maturation differs between boys and girls. PLoS One 7 (3). Bramen, J.E., Hranilovich, J.A., Dahl, R.E., Forbes, E.E., Chen, J., Toga, A.W., et al., 2011. Puberty influences medial temporal lobe and cortical gray matter maturation differently in boys than girls matched for sexual maturity. Cerebr. Cortex 21 (3), 636e646. Breslin, G., Gossrau-Breen, D., McCay, N., Gilmore, G., MacDonald, L., Hanna, D., 2012. Physical activity, gender, weight status, and wellbeing in 9-to 11-year-old children: a cross-sectional survey. J. Phys. Act. Health 9 (3), 394e401. Brouwer, R.M., Koenis, M.M.G., Schnack, H.G., van Baal, G.C., van Soelen, I.L.C., Boomsma, D.I., Pol, H.E.H., 2015. Longitudinal development of hormone levels and grey matter density in 9 and 12-year-old twins. Behav. Genet. 45 (3), 313e323. Brun, C.C., Leporé, N., Luders, E., Chou, Y.Y., Madsen, S.K., Toga, A.W., Thompson, P.M., 2009. Sex differences in brain structure in auditory and cingulate regions. Neuroreport 20 (10), 930e935. Bryant, D.M., Hoeft, F., Lai, S., Lackey, J., Roeltgen, D., Ross, J., Reiss, A.L., 2011. Neuroanatomical phenotype of Klinefelter Syndrome in childhood: a voxel-based morphometry study. J. Neurosci. 31 (18), 6654e6660. Bryden, M.P., 1982. Laterality: Functional Asymmetry in the Intact Brain. Academic Press, New York. Buchanan, C.M., Eccles, J.S., Becker, J.B., 1992. Are adolescents the victims of raging hormones: evidence for activational effects of hormones on moods and behavior at adolescence. Psychol. Bull. 111, 62e107. Buckner, R.L., Raichle, M.E., Petersen, S.E., 1995. Dissociation of human prefrontal cortical areas across different speech production tasks and gender groups. J. Neurophysiol. 74 (5), 2163e2173. Burghy, C.A., Stodola, D.E., Ruttle, P.L., Molloy, E.K., Armstrong, J.M., Oler, J.A., et al., 2012. Developmental pathways to amygdala-prefrontal function and internalizing symptoms in adolescence. Nat. Neurosci. 15 (12), 1736. Burman, D.D., Bitan, T., Booth, J.R., 2008. Sex differences in neural processing of language among children. Neuropsychologia 46 (5), 1349e1362. Burns, N.R., Nettelbeck, T., 2005. Inspection time and speed of processing: sex differences on perceptual speed but not IT. Personal. Individ. Differ. 39 (2), 439e446. Butler, T., Imperato-McGinley, J., Pan, H., Voyer, D., Cordero, J., Zhu, Y.S., et al., 2006. Sex differences in mental rotation: top-down versus bottom-up processing. Neuroimage 32 (1), 445e456. Butterfield, S.A., Angell, R.M., Mason, C.A., 2012. Age and sex differences in object control skills by children ages 5 to 14. Percept. Mot. Skills 114 (1), 261e274. Button, K.S., Ioannidis, J.P.A., Mokrysz, C., Nosek, B.A., Flint, J., Robinson, E.S.J., Munafo, M.R., 2013. Power failure: why small sample size undermines the reliability of neuroscience. Nat. Rev. Neurosci. 14 (5), 365e376. Byrnes, J.P., Miller, D.C., Schafer, W.D., 1999. Gender differences in risk taking: a meta-analysis. Psychol. Bull. 125 (3), 367e383. Cahill, L., 2005. His brain, her brain. Sci. Am. 292 (5), 40e47. Cahill, L., 2010. Sex influences on brain and emotional memory: the burden of proof has shifted. Prog. Brain Res. 186, 29e40. Callens, N., Van Kuyk, M., van Kuppenveld, J.H., Drop, S.L.S., Cohen-Kettenis, P.T., Dessens, A.B., Dutch Study Grp, D. S. D., 2016. Recalled and current gender role behavior, gender identity and sexual orientation in adults with disorders/differences of sex development. Horm. Behav. 86, 8e20. Camarata, S., Woodcock, R.W., 2006. Sex differences in processing speed: developmental effects in males and females. Intelligence 34 (3), 231e252. Card, N.A., Stucky, B.D., Sawalani, G.M., Little, T.D., 2008. Direct and indirect aggression during childhood and adolescence: a meta-analytic review of gender differences, intercorrelations, and relations to maladjustment. Child Dev. 79 (5), 1185e1229. Casey, B.J., Jones, R.M., Somerville, L.H., 2011. Braking and accelerating of the adolescent brain. J. Res. Adolesc. 21 (1), 21e33. Casey, B.J., Tottenham, N., Liston, C., Durston, S., 2005. Imaging the developing brain: what have we learned about cognitive development? Trends Cogn. Sci. 9 (3), 104e110. Ceci, S.J., Williams, W.M., 2010. Sex differences in math-intensive fields. Curr. Dir. Psychol. Sci. 19 (5), 275e279. Chang, Y.A., Javadi, S.S., Bahrami, N., Uttarwar, V.S., Reyes, A., McDonald, C.R., 2018. Mapping lexical-semantic networks and determining hemispheric language dominance: do task design, sex, age, and language performance make a difference? Brain Lang. 179, 42e50. Chaplin, T.M., 2015. Gender and emotion expression: a developmental contextual perspective. Emot. Rev. 7 (1), 14e21. Chiang, M.C., McMahon, K.L., de Zubicaray, G.I., Martin, N.G., Hickie, I., Toga, A.W., et al., 2011. Genetics of white matter development: a DTI study of 705 twins and their siblings aged 12 to 29. Neuroimage 54 (3), 2308e2317. Chivers, M.L., 2017. The specificity of women’s sexual response and its relationship with sexual orientations: a review and ten hypotheses. Arch. Sex. Behav. 46 (5), 1161e1179. Christov-Moore, L., Simpson, E.A., Coude, G., Grigaityte, K., Iacoboni, M., Ferrari, P.F., 2014. Empathy: gender effects in brain and behavior. Neurosci. Biobehav. Rev. 46, 604e627. Christova, P.S., Lewis, S.M., Tagaris, G.A., Ugurbil, K., Georgopoulos, A.P., 2008. A voxel-by-voxel parametric fMRI study of motor mental rotation: hemispheric specialization and gender differences in neural processing efficiency. Exp. Brain Res. 189 (1), 79e90. Chudowsky, N., Chudowsky, V., 2010. State Test Score Trends through 2007-08, Part 5: Are There Differences in Achievement between Boys and Girls? Retrieved from: https://files.eric.ed.gov/fulltext/ED509023.pdf.

Sex differences in brain and behavioral development Chapter | 27

623

Ciumas, C., Hirschberg, A.L., Savic, I., 2009. High fetal testosterone and sexually dimorphic cerebral networks in females. Cerebr. Cortex 19 (5), 1167e1174. Clark, M.M., Galef, B.G., 1998. Effects of intrauterine position on the behavior and genital morphology of litter-bearing rodents. Dev. Neuropsychol. 14, 197e211. Clements-Stephens, A.M., Rimrodt, S.L., Cutting, L.E., 2009. Developmental sex differences in basic visuospatial processing: differences in strategy use? Neurosci. Lett. 449 (3), 155e160. Clements, A.M., Rimrodt, S.L., Abel, J.R., Blankner, J.G., Mostofsky, S.H., Pekar, J.J., et al., 2006. Sex differences in cerebral laterality of language and visuospatial processing. Brain Lang. 98 (2), 150e158. Cohen, J., 1988. Statistical Power Analysis for the Behavioral Sciences, second ed. Academic Press, New York. Colby, J.B., Van Horn, J.D., Sowell, E.R., 2011. Quantitative in vivo evidence for broad regional gradients in the timing of white matter maturation during adolescence. Neuroimage 54 (1), 25e31. Cole-Harding, S., Morstad, A.L., Wilson, J.R., 1988. Spatial ability in members of opposite-sex twin pairs (Abstract). Behav. Genet. 18, 710. Cole, M.W., Ito, T., Bassett, D.S., Schultz, D.H., 2016. Activity flow over resting-state networks shapes cognitive task activations. Nat. Neurosci. 19 (12), 1718e1726. Connor, J.M., Serbin, L.A., 1977. Behaviorally based masculine and feminine activity-preference scales for preschoolers: correlates with other classroom behaviors and cognitive tests. Child Dev. 48, 1411e1416. Constantinescu, M., Moore, D.S., Johnson, S.P., Hines, M., 2018. Early contributions to infants’ mental rotation abilities. Dev. Sci. 21 (4), 15. Copeland, W.E., Worthman, C., Shanahan, L., Costello, E.J., Angold, A., 2019. Early pubertal timing and testosterone associated with higher levels of adolescent depression in girls. J. Am. Acad. Child Adolesc. Psychiatry 58 (12), 1197e1206. Cosgrove, K.P., Mazure, C.M., Staley, J.K., 2007. Evolving knowledge of sex differences in brain structure, function, and chemistry. Biol. Psychiatry 62 (8), 847e855. Costa Jr., P., Terracciano, A., McCrae, R.R., 2001. Gender differences in personality traits across cultures: robust and surprising findings. J. Personal. Soc. Psychol. 81 (2), 322e331. Craig, M.C., Fletcher, P.C., Daly, E.M., Rymer, J., Brammer, M., Giampietro, V., Murphy, D.G.M., 2008. Physiological variation in estradiol and brain function: a functional magnetic resonance imaging study of verbal memory across the follicular phase of the menstrual cycle. Horm. Behav. 53 (4), 503e508. Craig, M.C., Murphy, D.G.M., 2010. Estrogen therapy and Alzheimer’s dementia. In: Creatsas, G., Mastorakos, G. (Eds.), Women’s Health and Disease, vol. 1205. Blackwell Publishing, Oxford, pp. 245e253. Crick, N.R., 1995. Relational aggression: the role of intent attributions, feelings of distress, and provocation type. Dev. Psychopathol. 7 (2), 313e322. Crick, N.R., Zahn-Waxler, C., 2003. The development of psychopathology in females and males: current progress and future challenges. Dev. Psychopathol. 15, 719e742. Crissman, H.P., Berger, M.B., Graham, L.F., Dalton, V.K., 2017. Transgender demographics: a household probability sample of US adults, 2014. Am. J. Public Health 107 (2), 213e215. Critchley, H., Daly, E., Phillips, M., Brammer, M., Bullmore, E., Williams, S., et al., 2000. Explicit and implicit neural mechanisms for processing of social information from facial expressions: a functional magnetic resonance imaging study. Hum. Brain Mapp. 9 (2), 93e105. Crone, E.A., Dahl, R.E., 2012. Understanding adolescence as a period of social-affective engagement and goal flexibility. Nat. Rev. Neurosci. 13 (9), 636e650. Crone, E.A., Elzinga, B.M., 2015. Changing brains: how longitudinal functional magnetic resonance imaging studies can inform us about cognitive and social-affective growth trajectories. Wiley Interdisc. Rev.-Cogn. Sci. 6 (1), 53e63. Cross, C.P., Copping, L.T., Campbell, A., 2011. Sex differences in impulsivity: a meta-analysis. Psychol. Bull. 137 (1), 97e130. Cross, C.P., Cyrenne, D.M., Brown, G.R., 2013. Sex differences in sensation-seeking: a meta-analysis. Sci. Rep. 3, 5. Dar-Nimrod, I., Heine, S.J., 2006. Exposure to scientific theories affects women’s math performance. Science 314, 435. Daseking, M., Petermann, F., Waldmann, H.C., 2017. Sex differences in cognitive abilities: analyses for the German WAIS-IV. Personal. Individ. Differ. 114, 145e150. Davatzikos, C., Resnick, S.M., 1998. Sex differences in anatomic measures of interhemispheric connectivity: correlations with cognition in women but not men. Cerebr. Cortex 8 (7), 635e640. David, S.P., Naudet, F., Laude, J., Radua, J., Fusar-Poli, P., Chu, I., et al., 2018. Potential reporting bias in neuroimaging studies of sex differences. Sci. Rep. 8, 8. De Bellis, M.D., Keshavan, M.S., Beers, S.R., Hall, J., Frustaci, K., Masalehdan, A., et al., 2001. Sex differences in brain maturation during childhood and adolescence. Cerebr. Cortex 11 (6), 552e557. De Bolle, M., De Fruyt, F., McCrae, R.R., Lockenhoff, C.E., Costa, P.T., Aguilar-Vafaie, M.E., et al., 2015. The emergence of sex differences in personality traits in early adolescence: a cross-sectional, cross-cultural study. J. Personal. Soc. Psychol. 108 (1), 171e185. De Bondt, T., Jacquemyn, Y., Van Hecke, W., Sijbers, J., Sunaert, S., Parizel, P.M., 2013. Regional gray matter volume differences and sex-hormone correlations as a function of menstrual cycle phase and hormonal contraceptives use. Brain Res. 1530, 22e31. De Lacoste-Utamsing, C., Holloway, R.L., 1982. Sexual dimorphism in the human corpus callosum. Science 216 (4553), 1431e1432. De Vries, G.J., 2004. Minireview: sex differences in adult and developing brains: compensation, compensation, compensation. Endocrinology 145 (3), 1063e1068.

624 PART | II Cognitive development

Dennison, M., Whittle, S., Yucel, M., Vijayakumar, N., Kline, A., Simmons, J., Allen, N.B., 2013. Mapping subcortical brain maturation during adolescence: evidence of hemisphere- and sex-specific longitudinal changes. Dev. Sci. 16 (5), 772e791. Derntl, B., Habel, U., Windischberger, C., Robinson, S., Kryspin-Exner, I., Gur, R.C., Moser, E., 2009. General and specific responsiveness of the amygdala during explicit emotion recognition in females and males. BMC Neurosci. 10. Dessens, A.B., Slijper, F.M.E., Drop, S.L.S., 2005. Gender dysphoria and gender change in chromosomal females with congenital adrenal hyperplasia. Arch. Sex. Behav. 34, 389e397. Devine, A., Fawcett, K., Szucs, D., Dowker, A., 2012. Gender differences in mathematics anxiety and the relation to mathematics performance while controlling for test anxiety. Behav. Brain Funct. 8, 9. Diamond, A., 2013. Executive functions. In: Fiske, S.T. (Ed.), Annual Review of Psychology, vol. 64, pp. 135e168. Diamond, L.M., 2008. Female bisexuality from adolescence to adulthood: results from a 10-year longitudinal study. Dev. Psychol. 44 (1), 5e14. Diamond, L.M., 2014. Gender and same-sex sexuality. In: Tolman, D.L., Diamond, L.M. (Eds.), APA Handbook of Sexuality and Psychology, vol. 1. American Psychological Association. Diamond, L.M., Dickenson, J.A., Blair, K.L., 2017. Stability of sexual attractions across different timescales: the roles of bisexuality and gender. Arch. Sex. Behav. 46 (1), 193e204. Dietrich, T., Krings, T., Neulen, J., Willmes, K., Erberich, S., Thron, A., Sturm, W., 2001. Effects of blood estrogen level on cortical activation patterns during cognitive activation as measured by functional MRI. Neuroimage 13 (3), 425e432. Dodge, K.A., Coie, J.D., Lynam, D., 2006. Aggression and antisocial behavior in youth. In: Eisenberg, N. (Ed.), Handbook of Child Psychology, Social, Emotional, and Personality Development, sixth ed., vol. 3. Wiley, Hoboken, NJ, pp. 719e788. Donnelly, K.M., Allendorfer, J.B., Szaflarski, J.P., 2011. Right hemispheric participation in semantic decision improves performance. Brain Res. 1419, 105e116. Dorn, L.D., Biro, F.M., 2011. Puberty and its measurement: a decade in review. J. Res. Adolesc. 21 (1), 180e195. Dorn, L.D., Dahl, R.E., Woodward, H.R., Biro, F., 2006. Defining the boundaries of early adolescence: a user’s guide to assessing pubertal status and pubertal timing in research with adolescents. Appl. Dev. Sci. 10 (1), 30e56. Dosenbach, N.U.F., Koller, J.M., Earl, E.A., Miranda-Dominguez, O., Klein, R.L., Van, A.N., et al., 2017. Real-time motion analytics during brain MRI improve data quality and reduce costs. Neuroimage 161, 80e93. Dosenbach, N.U.F., Nardos, B., Cohen, A.L., Fair, D.A., Power, J.D., Church, J.A., et al., 2010. Prediction of individual brain maturity using fMRI. Science 329 (5997), 1358e1361. Douw, L., Schoonheim, M.M., Landi, D., van der Meer, M.L., Geurts, J.J.G., Reijneveld, J.C., et al., 2011. Cognition is related to resting-state smallworld network topology: an magnetoencephalographic study. Neuroscience 175, 169e177. Dubb, A., Gur, R., Avants, B., Gee, J., 2003. Characterization of sexual dimorphism in the human corpus callosum. Neuroimage 20 (1), 512e519. Ducharme, S., Albaugh, M.D., Nguyen, T.V., Hudziak, J.J., Mateos-Perez, J.M., Labbe, A., et al., 2016. Trajectories of cortical thickness maturation in normal brain development: the importance of quality control procedures. Neuroimage 125, 267e279. Duke, S.A., Balzer, B.W., Steinbeck, K.S., 2014. Testosterone and its effects on human male adolescent mood and behavior: a systematic review. J. Adolesc. Health 55 (3), 315e322. Dumas, J.A., Kutz, A.M., Naylor, M.R., Johnson, J.V., Newhouse, P.A., 2010. Increased memory load-related frontal activation after estradiol treatment in postmenopausal women. Horm. Behav. 58 (5), 929e935. Eagly, A.H., 2018. The shaping of science by ideology: how feminism inspired, led, and constrained scientific understanding of sex and gender. J. Soc. Issues 1e18. Eagly, A.H., Wood, W., 1999. The origins of sex differences in human behavior: evolved dispositions versus social roles. Am. Psychol. 54 (6), 408e423. Eaton, N.R., Keyes, K.M., Krueger, R.F., Balsis, S., Skodol, A.E., Markon, K.E., et al., 2012. An invariant dimensional liability model of gender differences in mental disorder prevalence: evidence from a national sample. J. Abnorm. Psychol. 121 (1), 282e288. Eccles, J.S., Freedman-Doan, C., Frome, P., Jacobs, J., Yoon, K.S., 2000. Gender-role socialization in the family: a longitudinal approach. In: Eckes, T., Trautner, H.M. (Eds.), The Developmental Social Psychology of Gender. Erlbaum, Mahwah, NJ. Egan, S.K., Perry, D.G., 2001. Gender identity: a multidimensional analysis with implications for psychosocial adjustment. Dev. Psychol. 37 (4), 451e463. Eisenberg, N., Spinrad, T.L., Sadovsky, A., 2006. Empathy-related responding in children. In: Killen, M., Smetana, J.G. (Eds.), Handbook of Moral Development. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 517e549. Eklund, A., Nichols, T.E., Knutsson, H., 2016. Cluster failure: why fMRI inferences for spatial extent have inflated false-positive rates. Proc. Natl. Acad. Sci. U.S.A. 113 (33). E4929-E4929. Eklund, P.L.E., Gooren, L.J.G., Bezemer, P.D., 1988. Prevalence of transsexualism in The Netherlands. Br. J. Psychiatry 152, 638e640. Eliot, L., Richardson, S.S., 2016. Sex in context: limitations of animal studies for addressing human sex/gender neurobehavioral health disparities. J. Neurosci. 36 (47), 11823e11830. Else-Quest, N.M., Hyde, J.S., Goldsmith, H.H., Van Hulle, C.A., 2006. Gender differences in temperament: a meta-analysis. Psychol. Bull. 132 (1), 33e72. Else-Quest, N.M., Hyde, J.S., Linn, M.C., 2010. Cross-national patterns of gender differences in mathematics: a meta-analysis. Psychol. Bull. 136 (1), 103e127. Eluvathingal, T.J., Hasan, K.M., Kramer, L., Fletcher, J.M., Ewing-Cobbs, L., 2007. Quantitative diffusion tensor tractography of association and projection fibers in normally developing children and adolescents. Cerebr. Cortex 17 (12), 2760e2768.

Sex differences in brain and behavioral development Chapter | 27

625

Endendijk, J.J., Beltz, A.M., McHale, S.M., Bryk, K., Berenbaum, S.A., 2016. Linking prenatal androgens to gender-related attitudes, identity, and activities: evidence from girls with congenital adrenal hyperplasia. Arch. Sex. Behav. 45 (7), 1807e1815. Engman, J., Linnman, C., Van Dijk, K.R.A., Milad, M.R., 2016. Amygdala subnuclei resting-state functional connectivity sex and estrogen differences. Psychoneuroendocrinology 63, 34e42. Erickson, K.I., Voss, M.W., Prakash, R.S., Chaddock, L., Kramer, A.F., 2010. A cross-sectional study of hormone treatment and hippocampal volume in postmenopausal women: evidence for a limited window of opportunity. Neuropsychology 24 (1), 68e76. Ernst, M., Maheu, F.S., Schroth, E., Hardin, J., Golan, L.G., Cameron, J., et al., 2007. Amygdala function in adolescents with congenital adrenal hyperplasia: a model for the study of early steroid abnormalities. Neuropsychologia 45 (9), 2104e2113. Fabes, R.A., Martin, C.L., Hanish, L.D., Anders, M.C., Madden-Derdich, D.A., 2003. Early school competence: the roles of sex-segregated play and effortful control. Dev. Psychol. 39 (5), 848e858. Fair, D.A., Cohen, A.L., Dosenbach, N.U.F., Church, J.A., Miezin, F.M., Barch, D.M., et al., 2008. The maturing architecture of the brain’s default network. Proc. Natl. Acad. Sci. U.S.A. 105 (10), 4028e4032. Fair, D.A., Cohen, A.L., Power, J.D., Dosenbach, N.U.F., Church, J.A., Miezin, F.M., et al., 2009. Functional brain networks develop from a "local to distributed" organization. PLoS Comput. Biol. 5 (5). Fair, D.A., Dosenbach, N.U.F., Church, J.A., Cohen, A.L., Brahmbhatt, S., Miezin, F.M., et al., 2007. Development of distinct control networks through segregation and integration. Proc. Natl. Acad. Sci. U.S.A. 104 (33), 13507e13512. Falk, D., Froese, N., Sade, D.S., Dudek, B.G., 1999. Sex differences in brain/body relationships of Rhesus monkeys and humans. J. Hum. Evol. 36 (2), 233e238. Fama, R., Sullivan, E.V., 2015. Thalamic structures and associated cognitive functions: relations with age and aging. Neurosci. Biobehav. Rev. 54, 29e37. Faravelli, C., Scarpato, M.A., Castellini, G., Lo Sauro, C., 2013. Gender differences in depression and anxiety: the role of age. Psychiatry Res. 210 (3), 1301e1303. Fausto-Sterling, A., 2008. Myths of Gender: Biological Theories about Women and Men, revised edition. Basic Books. Feng, J., Spence, I., Pratt, J., 2007. Playing an action video game reduces gender differences in spatial cognition. Psychol. Sci. 18 (10), 850e855. Fernández, G., Weis, S., Stoffel-Wagner, B., Tendolkar, I., Reuber, M., Beyenburg, S., et al., 2003. Menstrual cycle-dependent neural plasticity in the adult human brain is hormone, task, and region specific. J. Neurosci. 23 (9), 3790e3795. Ferri, J., Bress, J.N., Eaton, N.R., Proudfit, G.H., 2014. The impact of puberty and social anxiety on amygdala activation to faces in adolescence. Dev. Neurosci. 36 (3e4), 239e249. Fiedler, K., 2011. Voodoo correlations are everywhere-not only in neuroscience. Perspect. Psychol. Sci. 6 (2), 163e171. Filipek, P.A., Richelme, C., Kennedy, D.N., Caviness, V.S., 1994. The young adult human brain: an MRI-based morphometric analysis. Cerebr. Cortex 4 (4), 344e360. Filippi, M., Valsasina, P., Misci, P., Falini, A., Comi, G., Rocca, M.A., 2013. The organization of intrinsic brain activity differs between genders: a resting-state fMRI study in a large cohort of young healthy subjects. Hum. Brain Mapp. 34, 1330e1334. Filkowski, M.M., Olsen, R.M., Duda, B., Wanger, T.J., Sabatinelli, D., 2017. Sex differences in emotional perception: meta analysis of divergent activation. Neuroimage 147, 925e933. Fine, C., 2010. Delusions of Gender: How Our Minds, Society, and Neurosexism Create Difference. W.W. Norton & Company, New York. Fine, C., 2017. Testosterone Rex: Myths of Sex, Science, and Society. W. W. Norton & Company. Fischer, A., LaFrance, M., 2015. What drives the smile and the tear: why women are more emotionally expressive than men. Emot. Rev. 7 (1), 22e29. Fivush, R., 1998. Gendered narratives: elaboration, structure, and emotion in parent-child reminiscing across the preschool years. In: Thompson, C.P., Herrmann, D.J. (Eds.), Autobiographical Memory: Theoretical and Applied Perspectives. Lawrence Erlbaum Associates, Mahwah, NJ, pp. 79e103. Fivush, R., Buckner, J.P., 2000. Gender, sadness, and depression: the development of emotional focus through gendered discourse. In: Fischer, A.H. (Ed.), Gender and Emotion: Social Psychological Perspectives. Studies in Emotion and Social Interaction. Cambridge University Press, New York, pp. 232e253. Flashman, L.A., Andreasen, N.C., Flaum, M., Swayze, V.W., 1997. Intelligence and regional brain volumes in normal controls. Intelligence 25 (3), 149e160. Flatters, I., Hill, L.J., Williams, J.H., Barber, S.E., Mon-Williams, M., 2014. Manual control age and sex differences in 4 to 11 year old children. PLoS One 9 (2), e88692. Forbes, E.E., Dahl, R.E., 2010. Pubertal development and behavior: hormonal activation of social and motivational tendencies. Brain Cogn. 72 (1), 66e72. Forbes, E.E., Ryan, N.D., Phillips, M.L., Manuck, S.B., Worthman, C.M., Moyles, D.L., et al., 2010. Healthy adolescents’ neural response to reward: associations with puberty, positive affect, and depressive symptoms. J. Am. Acad. Child Adolesc. Psychiatry 49 (2), 162e172. Frank, C.K., Baron-Cohen, S., Ganzel, B.L., 2015. Sex differences in the neural basis of false-belief and pragmatic language comprehension. Neuroimage 105, 300e311. Frisen, L., Nordenstrom, A., Falhammar, H., Filipsson, H., Holmdahl, G., Janson, P.O., et al., 2009. Gender role behavior, sexuality, and psychosocial adaptation in women with congenital adrenal hyperplasia due to CYP21A2 deficiency. J. Clin. Endocrinol. Metab. 94 (9), 3432e3439. Friston, K.J., 2004. Human Brain Function, second ed. Elsevier, San Diego, CA. Frost, J.A., Binder, J.R., Springer, J.A., Hammeke, T.A., Bellgowan, P.S.F., Rao, S.M., Cox, R.W., 1999. Language processing is strongly left lateralized in both sexes: evidence from functional MRI. Brain 122, 199e208.

626 PART | II Cognitive development

Fusar-Poli, P., Placentino, A., Carletti, F., Landi, P., Allen, P., Surguladze, S., et al., 2009. Functional atlas of emotional faces processing: a voxel-based meta-analysis of 105 functional magnetic resonance imaging studies. J. Psychiatry Neurosci. 34 (6), 418e432. Galvan, A., 2013. The teenage brain: sensitivity to rewards. Curr. Dir. Psychol. Sci. 22 (2), 88e93. Gao, W., Alcauter, S., Smith, J.K., Gilmore, J.H., Lin, W.L., 2015. Development of human brain cortical network architecture during infancy. Brain Struct. Funct. 220 (2), 1173e1186. Gates, G.J., 2011. How Many People Are Lesbian, Gay, Bisexual and Transgender? Retrieved from. https://williamsinstitute.law.ucla.edu/wp-content/ uploads/Gates-How-Many-People-LGBT-Apr-2011.pdf. Gauthier, C.T., Duyme, M., Zanca, M., Capron, C., 2009. Sex and performance level effects on brain activation during a verbal fluency task: a functional magnetic resonance imaging study. Cortex 45 (2), 164e176. Geary, D.C., 1998. Male, Female: The Evolution of Human Sex Differences. American Psychological Association, Washington, DC. Geiser, C., Lehmann, W., Eid, M., 2008. A note on sex differences in mental rotation in different age groups. Intelligence 36 (6), 556e563. Gennatas, E.D., Avants, B.B., Wolf, D.H., Satterthwaite, T.D., Ruparel, K., Ciric, R., et al., 2017. Age-related effects and sex differences in gray matter density, volume, mass, and cortical thickness from childhood to young adulthood. J. Neurosci. 37 (20), 5065e5073. Gentile, B., Grabe, S., Dolan-Pascoe, B., Twenge, J.M., Wells, B.E., Maitino, A., 2009. Gender differences in domain-specific self-esteem: a metaanalysis. Rev. Gen. Psychol. 13 (1), 34e45. Gershoni, M., Pietrokovski, S., 2017. The landscape of sex-differential transcriptome and its consequent selection in human adults. BMC Biol. 15, 15. Gibber, J., Goy, R.W., 1985. Infant directed behavior in young rhesus monkey: sex difference and effects of prenatal androgen. Am. J. Primatol. 8, 235e237. Giedd, J.N., Blumenthal, J., Jeffries, N.O., Castellanos, F.X., Liu, H., Zijdenbos, A., et al., 1999. Brain development during childhood and adolescence: a longitudinal MRI study. Nat. Neurosci. 2 (10), 861e863. Giedd, J.N., Castellanos, F.X., Rajapakse, J.C., Vaituzis, A.C., Rapoport, J.L., 1997. Sexual dimorphism of the developing human brain. Prog. Neuro Psychopharmacol. Biol. Psychiatry 21 (8), 1185e1201. Giedd, J.N., Clasen, L.S., Lenroot, R., Greenstein, D., Wallace, G.L., Ordaz, S., et al., 2006. Puberty-related influences on brain development. Mol. Cell. Endocrinol. 254, 154e162. Giedd, J.N., Denker, A.H., 2015. The adolescent brain: insights from neuroimaging. In: Bourguignon, J.P., Carel, J.C., Christen, Y. (Eds.), Brain Crosstalk in Puberty and Adolescence, vol. 13. Springer-Verlag Berlin, Berlin, pp. 85e96. Giedd, J.N., Rapoport, J.L., 2010. Structural MRI of pediatric brain development: what have we learned and where are we going? Neuron 67 (5), 728e734. Giedd, J.N., Raznahan, A., Alexander-Bloch, A., Schmitt, E., Gogtay, N., Rapoport, J.L., 2015. Child Psychiatry Branch of the National Institute of Mental Health longitudinal structural magnetic resonance imaging study of human brain development. Neuropsychopharmacology 40 (1), 43e49. Giorgio, A., Watkins, K.E., Douaud, G., James, A.C., James, S., De Stefano, N., et al., 2008. Changes in white matter microstructure during adolescence. Neuroimage 39 (1), 52e61. Goddings, A.L., Beltz, A.M., Peper, J.S., Crone, E.A., Braams, B.R., 2019. Understanding the role of puberty in structural and functional development of the adolescent brain. J. Res. Adolesc. 29 (1), 32e53. Goddings, A.L., Mills, K.L., Clasen, L.S., Giedd, J.N., Viner, R.M., Blakemore, S.J., 2014. The influence of puberty on subcortical brain development. Neuroimage 88, 242e251. Gogos, A., 2013. Natural and synthetic sex hormones: effects on higher-order cognitive function and prepulse inhibition. Biol. Psychol. 93 (1), 17e23. Gogos, A., Wu, Y.C., Williams, A.S., Byrne, L.K., 2014. The effects of ethinylestradiol and progestins (“the pill”) on cognitive function in premenopausal women. Neurochem. Res. 39 (12), 2288e2300. Gogtay, N., Giedd, J.N., Lusk, L., Hayashi, K.M., Greenstein, D., Vaituzis, A.C., et al., 2004. Dynamic mapping of human cortical development during childhood through early adulthood. Proc. Natl. Acad. Sci. U.S.A. 101 (21), 8174e8179. Gold, E.B., Sternfeld, B., Kelsey, J.L., Brown, C., Mouton, C., Reame, N., et al., 2000. Relation of demographic and lifestyle factors to symptoms in a multi-racial/ethnic population of women 40-55 years of age. Am. J. Epidemiol. 152 (5), 463e473. Goldstein, J.M., Seidman, L.J., Horton, N.J., Makris, N., Kennedy, D.N., Caviness Jr., V.S., et al., 2001. Normal sexual dimorphism of the adult human brain assessed by in vivo magnetic resonance imaging. Cerebr. Cortex 11 (6), 490e497. Gorski, R.A., Gordon, J.H., Shryne, J.E., Southam, A.M., 1978. Evidence for a morphological sex difference within the medial preoptic area of the rat brain. Brain Res. 148 (2), 333e346. Goto, M., Abe, O., Miyati, T., Inano, S., Hayashi, N., Aoki, S., et al., 2011. 3 tesla MRI detects accelerated hippocampal volume reduction in postmenopausal women. J. Magn. Reson. Imaging 33 (1), 48e53. Goy, R.W., Bercovitch, F.B., McBrair, M.C., 1988. Behavioral masculinization is independent of genital masculinization in prenatally androgenized female rhesus macaques. Horm. Behav. 22, 552e571. Goy, R.W., McEwen, B.S., 1980. Sexual Differentiation of the Brain. MIT Press, Cambridge. Graber, J.A., 2013. Pubertal timing and the development of psychopathology in adolescence and beyond. Horm. Behav. 64 (2), 262e269. Grayson, D.S., Fair, D.A., 2017. Development of large-scale functional networks from birth to adulthood: a guide to the neuroimaging literature. Neuroimage 160, 15e31. Green, T., Chromik, L.C., Mazaika, P.K., Fierro, K., Raman, M.M., Lazzeroni, L.C., et al., 2014. Aberrant parietal cortex developmental trajectories in girls with turner syndrome and related visual-spatial cognitive development: a preliminary study. Am. J. Med. Genet. B-Neuropsy. Genet. 165 (6), 531e540.

Sex differences in brain and behavioral development Chapter | 27

627

Griksiene, R., Monciunskaite, R., Arnatkeviciute, A., Ruksenas, O., 2018. Does the use of hormonal contraceptives affect the mental rotation performance? Horm. Behav. 100, 29e38. Griksiene, R., Ruksenas, O., 2011. Effects of hormonal contraceptives on mental rotation and verbal fluency. Psychoneuroendocrinology 36 (8), 1239e1248. Grimshaw, G.M., Bryden, M.P., Finegan, J.A.K., 1995. Relations between prenatal testosterone and cerebral lateralization in children. Neuropsychology 9 (1), 68e79. Grön, G., Wunderlich, A.P., Spitzer, M., Tomczak, R., Riepe, M.W., 2000. Brain activation during human navigation: gender-different neural networks as substrate of performance. Nat. Neurosci. 3 (4), 404e408. Gronenschild, E., Habets, P., Jacobs, H.I.L., Mengelers, R., Rozendaal, N., van Os, J., Marcelis, M., 2012. The effects of FreeSurfer version, workstation type, and Macintosh operating system version on anatomical volume and cortical thickness measurements. PLoS One 7 (6), 13. Guiso, L., Monte, F., Sapienza, P., Zingales, L., 2008. Culture, gender, and math. Science 320 (5880), 1164e1165. Gur, R.C., Alsop, D., Glahn, D., Petty, R., Swanson, C.L., Maldjian, J.A., et al., 2000. An fMRI study of sex differences in regional activation to a verbal and a spatial task. Brain Lang. 74 (2), 157e170. Gur, R.C., Gur, R.E., 2017. Complementarity of sex differences in brain and behavior: from laterality to multimodal neuroimaging. J. Neurosci. Res. 95 (1e2), 189e199. Gur, R.C., Richard, J., Calkins, M.E., Chiavacci, R., Hansen, J.A., Bilker, W.B., et al., 2012. Age group and sex differences in performance on a computerized neurocognitive battery in children age 8-21. Neuropsychology 26 (2), 251e265. Gur, R.C., Turetsky, B.I., Matsui, M., Yan, M., Bilker, W., Hughett, P., Gur, R.E., 1999. Sex differences in brain gray and white matter in healthy young adults: correlations with cognitive performance. J. Neurosci. 19 (10), 4065e4072. Guyer, A.E., Silk, J.S., Nelson, E.E., 2016. The neurobiology of the emotional adolescent: from the inside out. Neurosci. Biobehav. Rev. 70, 74e85. Habel, U., Windischberger, C., Derntl, B., Robinson, S., Kryspin-Exner, I., Gur, R.C., Moser, E., 2007. Amygdala activation and facial expressions: explicit emotion discrimination versus implicit emotion processing. Neuropsychologia 45 (10), 2369e2377. Halari, R., Sharma, T., Hines, M., Andrew, C., Simmons, A., Kumari, V., 2006. Comparable fMRI activity with differential behavioural performance on mental rotation and overt verbal fluency tasks in healthy men and women. Exp. Brain Res. 169 (1), 1e14. Hall, J.J., Neal, T.J., Dean, R.S., 2008. Lateralization of cerebral functions. In: Horton, J.A.M., Wedding, D. (Eds.), The Neuropsychology Handbook, third ed. Springer, New York, pp. 183e214. Halpern, D.F., 1986. A different answer to the question, "Do sex-related differences in spatial abilities exist? Am. Psychol. 41 (9), 1014e1015. Halpern, D.F., 2012. Sex Differences in Cognitive Abilities, fourth ed. Psychology Press, New York. Halpern, D.F., Benbow, C.P., Geary, D.C., Gur, R.C., Hyde, J.S., Gernsbacher, M.A., 2007. The science of sex differences in science and mathematics. Psychol. Sci. Public Interest 8 (1), 1e51. Halpern, D.F., Eliot, L., Bigler, R.S., Fabes, R.A., Hanish, L.D., Hyde, J., et al., 2011. The pseudoscience of single-sex schooling. Science 333 (6050), 1706e1707. Hamann, S., 2005. Sex differences in the responses of the human amygdala. The Neuroscientist 11 (4), 288e293. Hamann, S., Stevens, J., Vick, J.H., Bryk, K., Quigley, C.A., Berenbaum, S.A., Wallen, K., 2014. Brain responses to sexual images in 46,XY women with complete androgen insensitivity syndrome are female-typical. Horm. Behav. 66 (5), 724e730. Hampson, E., 1990. Variations in sex-related cognitive abilities across the menstrual cycle. Brain Cogn. 14 (1), 26e43. Hampson, E., 2007. Endocrine contributions to sex differences in visuospatial perception and cognition. In: Becker, J.B., Berkley, K.J., Geary, N., Hampson, E., Herman, J., Young, E. (Eds.), Sex Differences in the Brain: From Genes to Behavior. Oxford University Press, New York. Hampson, E., Rovet, J.F., 2015. Spatial function in adolescents and young adults with congenital adrenal hyperplasia: clinical phenotype and implications for the androgen hypothesis. Psychoneuroendocrinology 54, 60e70. Hampson, E., Rovet, J.F., Altmann, D., 1998. Spatial reasoning in children with congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Dev. Neuropsychol. 14, 299e320. Hannagan, R.J., 2008. Gendered political behavior: a Darwinian feminist approach. Sex Roles 59 (7e8), 465e475. Harasty, J., Double, K.L., Halliday, G.M., Kril, J.J., McRitchie, D.A., 1997. Language-associated cortical regions are proportionally larger in the female brain. Arch. Neurol. 54 (2), 171e176. Hardee, J.E., Cope, L.M., Munier, E.C., Welsh, R.C., Zucker, R.A., Heitzeg, M.M., 2017. Sex differences in the development of emotion circuitry in adolescents at risk for substance abuse: a longitudinal fMRI study. Soc. Cogn. Affect. Neurosci. 12 (6), 965e975. Harden, K.P., Mann, F.D., Grotzinger, A.D., Patterson, M.W., Steinberg, L., Tackett, J.L., Tucker-Drob, E.M., 2018. Developmental differences in reward sensitivity and sensation seeking in adolescence: testing sex-specific associations with gonadal hormones and pubertal development. J. Personal. Soc. Psychol. 115 (1), 161e178. Hariri, A.R., Bookheimer, S.Y., Mazziotta, J.C., 2000. Modulating emotional responses: effects of a neocortical network on the limbic system. Neuroreport 11 (1), 43e48. Hartung, C.M., Lefler, E.K., 2019. Sex and gender in psychopathology: DSMe5 and beyond. Psychol. Bull. 145 (4), 390e409. Hartung, C.M., Widiger, T.A., 1998. Gender differences in the diagnosis of mental disorders: conclusions and controversies of the DSM-IV. Psychol. Bull. 123 (3), 260e278. Hedges, L.V., Nowell, A., 1995. Sex differences in mental test scores, variability, and numbers of high-scoring individuals. Science 269 (5235), 41e45. Heil, M., Kavsek, M., Rolke, B., Beste, C., Jansen, P., 2011. Mental rotation in female fraternal twins: evidence for intra-uterine hormone transfer? Biol. Psychol. 86 (1), 90e93.

628 PART | II Cognitive development

Henderson, V.W., 2008. Cognitive changes after menopause: influence of estrogen. Clin. Obstet. Gynecol. 51 (3), 618e626. Henderson, V.W., Popat, R.A., 2011. Effects of endogenous and exogenous estrogen exposures in midlife and late-life women on episodic memory and executive functions. Neuroscience 191, 129e138. Herlitz, A., Loven, J., 2013. Sex differences and the own-gender bias in face recognition: a meta-analytic review. Vis. Cogn. 21. Herlitz, A., Lovén, J., Thilers, P., Rehnman, J., 2010. Sex differences in episodic memory: the where but not the why. In: Bäckman, L., Nyberg, L. (Eds.), Memory, Aging and the Brain. Psychology Press, New York, pp. 132e143. Herlitz, A., Rehnman, J., 2008. Sex differences in episodic memory. Curr. Dir. Psychol. Sci. 17 (1), 52e56. Herlitz, A., Reuterskiold, L., Loven, J., Thilers, P.P., Rehnman, J., 2013. Cognitive sex differences are not magnified as a function of age, sex hormones, or puberty development during early adolescence. Dev. Neuropsychol. 38 (3), 167e179. Herting, M.M., Gautam, P., Chen, Z.H., Mezher, A., Vetter, N.C., 2018. Test-retest reliability of longitudinal task-based fMRI: implications for developmental studies. Dev. Cogn. Neurosci. 33, 17e26. Herting, M.M., Gautam, P., Spielberg, J.M., Dahl, R.E., Sowell, E.R., 2015. A longitudinal study: changes in cortical thickness and surface area during pubertal maturation. PLoS One 10 (3), 14. Herting, M.M., Gautam, P., Spielberg, J.M., Kan, E., Dahl, R.E., Sowell, E.R., 2014. The role of testosterone and estradiol in brain volume changes across adolescence: a longitudinal structural MRI study. Hum. Brain Mapp. 35 (11), 5633e5645. Herting, M.M., Maxwell, E.C., Irvine, C., Nagel, B.J., 2012. The impact of sex, puberty, and hormones on white matter microstructure in adolescents. Cerebr. Cortex 22 (9), 1979e1992. Herting, M.M., Sowell, E.R., 2017. Puberty and structural brain development in humans. Front. Neuroendocrinol. 44, 122e137. Hier, D.B., Crowley, W.F., 1982. Spatial ability in androgen-deficient men. N. Engl. J. Med. 302, 1202e1205. Hines, M., 2010. Sex-related variation in human behavior and the brain. Trends Cogn. Sci. 14 (10), 448e456. Hines, M., 2011. Gender development and the human brain. Annu. Rev. Neurosci. 34, 69e88. Hines, M., Ahmed, F., Hughes, I.A., 2003. Psychological outcomes and gender-related development in complete androgen insensitivity syndrome. Arch. Sex. Behav. 32 (2), 93e101. Hines, M., Constantinescu, M., Spencer, D., 2015. Early androgen exposure and human gender development. Biol. Sex Differ. 6, 10. Hines, M., Fane, B.A., Pasterski, V.L., Mathews, G.A., Conway, G.S., Brook, C., 2003. Spatial abilities following prenatal androgen abnormality: targeting and mental rotations performance in individuals with congenital adrenal hyperplasia. Psychoneuroendocrinology 28, 1010e1026. Hines, M., McAdams, L.A., Chiu, L., Bentler, P.M., Lipcamon, J., 1992. Cognition and the corpus callosum: verbal fluency, visuospatial ability, and language lateralization related to midsaggital surface areas of callosal subregions. Behav. Neurosci. 106 (1), 3e14. Hines, M., Spencer, D., Kung, K.T.F., Browne, W.V., Constantinescu, M., Noorderhaven, R.M., 2016. The early postnatal period, mini-puberty, provides a window on the role of testosterone in human neurobehavioural development. Curr. Opin. Neurobiol. 38, 69e73. Hirnstein, M., Hugdahl, K., Hausmann, M., 2019. Cognitive sex differences and hemispheric asymmetry: a critical review of 40 years of research. Laterality 24 (2), 204e252. Hirnstein, M., Westerhausen, R., Korsnes, M.S., Hugdahl, K., 2013. Sex differences in language asymmetry are age-dependent and small: a large-scale, consonantevowel dichotic listening study with behavioral and fMRI data. Cortex 49 (7), 1910e1921. Hjelmervik, H., Hausmann, M., Osnes, B., Westerhausen, R., Specht, K., 2014. Resting states are resting traits - an fMRI study of sex differences and menstrual cycle effects in resting state cognitive control networks. PLoS One 9 (7), 10. Hoff, K.A., Briley, D.A., Wee, C.J.M., Rounds, J., 2018. Normative changes in interests from adolescence to adulthood: a meta-analysis of longitudinal studies. Psychol. Bull. 144 (4), 426e451. Holloway, R.L., 1980. Within-species brain-body weight variability: a reexamination of the Danish data and other primate species. Am. J. Phys. Anthropol. 53 (1), 109e121. Hong, D.S., Reiss, A.L., 2014. Cognitive and neurological aspects of sex chromosome aneuploidies. Lancet Neurol. 13 (3), 306e318. Hopkins, E.J., Weisberg, D.S., Taylor, J.C.V., 2016. The seductive allure is a reductive allure: people prefer scientific explanations that contain logically irrelevant reductive information. Cognition 155, 67e76. Hugdahl, K., Thomsen, T., Ersland, L., 2006. Sex differences in visuo-spatial processing: an fMRI study of mental rotation. Neuropsychologia 44 (9), 1575e1583. Hulshoff Pol, H.E., Cohen-Kettenis, P.T., Van Haren, N.E.M., Peper, J.S., Brans, R.G.H., Cahn, W., et al., 2006. Changing your sex changes your brain: influences of testosterone and estrogen on adult human brain structure. Eur. J. Endocrinol. 155, S107eS114. Hyde, J.S., 2005. The gender similarities hypothesis. Am. Psychol. 60, 581e592. Hyde, J.S., Bigler, R.S., Joel, D., Tate, C.C., van Anders, S.M., 2019. The future of sex and gender in psychology: five challenges to the gender binary. Am. Psychol. 74 (2), 171e193. Im, K., Lee, J.M., Lee, J., Shin, Y.W., Kim, I.Y., Kwon, J.S., Kim, S.I., 2006. Gender difference analysis of cortical thickness in healthy young adults with surface-based methods. Neuroimage 31 (1), 31e38. Im, K., Lee, J.M., Lyttelton, O., Kim, S.H., Evans, A.C., Kim, S.I., 2008. Brain size and cortical structure in the adult human brain. Cerebr. Cortex 18 (9), 2181e2191. Ingalhalikar, M., Smith, A., Parker, D., Satterthwaite, T.D., Elliott, M.A., Ruparel, K., et al., 2014. Sex differences in the structural connectome of the human brain. Proc. Natl. Acad. Sci. U.S.A. 111 (2), 823e828. Ioannidis, J.P.A., 2016. Evidence-based medicine has been hijacked: a report to David Sackett. J. Clin. Epidemiol. 73, 82e86.

Sex differences in brain and behavioral development Chapter | 27

629

Jacobs, E.G., Weiss, B., Makris, N., Whitfield-Gabrieli, S., Buka, S.L., Klibanski, A., Goldstein, J.M., 2017. Reorganization of functional networks in verbal working memory circuitry in early midlife: the impact of sex and menopausal status. Cerebr. Cortex 27 (5), 2857e2870. Jacobs, E.G., Weiss, B.K., Makris, N., Whitfield-Gabrieli, S., Buka, S.L., Klibanski, A., Goldstein, J.M., 2016. Impact of sex and menopausal status on episodic memory circuitry in early midlife. J. Neurosci. 36 (39), 10163e10173. Jernigan, T.L., Brown, S.A., Dowling, G.J., 2018. The adolescent brain cognitive development study. J. Res. Adolesc. 28 (1), 154e156. Joel, D., Berman, Z., Tavor, I., Wexler, N., Gaber, O., Stein, Y., et al., 2015. Sex beyond the genitalia: the human brain mosaic. Proc. Natl. Acad. Sci. U.S.A. 112 (50), 15468e15473. Joel, D., Fausto-Sterling, A., 2016. Beyond sex differences: new approaches for thinking about variation in brain structure and function. Philos. Trans. R. Soc. Biol. Sci. 371 (1688), 9. Johnson, W., Bouchard, T.J., 2007. Sex differences in mental abilities: g masks the dimensions on which they lie. Intelligence 35 (1), 23e39. Jordan-Young, R., Rumiati, R.I., 2012. Hardwired for sexism? Approaches to sex/gender in neuroscience. Neuroethics 5 (3), 305e315. Jordan, K., Wurstenberg, T., Heinze, H.J., Peters, M., Jancke, L., 2002. Women and men exhibit different cortical activation patterns during mental rotation tasks. Neuropsychologia 40 (13), 2397e2408. Kaltiala-Heino, R., Sumia, M., Tyolajarvi, M., Lindberg, N., 2015. Two years of gender identity service for minors: overrepresentation of natal girls with severe problems in adolescent development. Child Adolesc. Psychiatr. Ment. Health 9 (1), 9. Kansaku, K., Kitazawa, S., 2001. Imaging studies on sex differences in the lateralization of language. Neurosci. Res. 41 (4), 333e337. Keller, J.B., Hedden, T., Thompson, T.W., Anteraper, S.A., Gabrieli, J.D.E., Whitfield-Gabrieli, S., 2015. Resting-state anticorrelations between medial and lateral prefrontal cortex: association with working memory, aging, and individual differences. Cortex 64, 271e280. Kelly, A.M.C., Di Martino, A., Uddin, L.Q., Shehzad, Z., Gee, D.G., Reiss, P.T., et al., 2009. Development of anterior cingulate functional connectivity from late childhood to early adulthood. Cerebr. Cortex 19 (3), 640e657. Kendler, K.S., Gardner, C.O., 2014. Sex differences in the pathways to major depression: a study of opposite-sex twin pairs. Am. J. Psychiatry 171 (4), 426e435. Keyes, K.M., Cheslack-Postava, K., Westhoff, C., Heim, C.M., Haloossim, M., Walsh, K., Koenen, K., 2013. Association of hormonal contraceptive use with reduced levels of depressive symptoms: a national study of sexually active women in the United States. Am. J. Epidemiol. 178 (9), 1378e1388. Khorashad, B.S., Khazai, B., Roshan, G.M., Hiradfar, M., Afkhamizadeh, M., van de Grift, T.C., 2018. Prenatal testosterone and theory of mind development: findings from disorders of sex development. Psychoneuroendocrinology 89, 250e255. Khorashad, B.S., Roshan, G.M., Reid, A.G., Aghili, Z., Moghadam, M.D., Khazai, B., et al., 2018. Childhood sex-typed behavior and gender change in individuals with 46,XY and 46,XX disorders of sex development: an Iranian Multicenter Study. Arch. Sex. Behav. 47 (8), 2287e2298. Kidwell, M.C., Lazarevic, L.B., Baranski, E., Hardwicke, T.E., Piechowski, S., Falkenberg, L.S., et al., 2016. Badges to acknowledge open practices: a simple, low-cost, effective method for increasing transparency. PLoS Biol. 14 (5), 15. Kim, G.W., Jeong, G.W., 2017. Menopause-related brain activation patterns during visual sexual arousal in menopausal women: an fMRI pilot study using time-course analysis. Neuroscience 343, 449e458. Kimura, D., 1999. Sex and Cognition. MIT Press, Cambridge, MA. Kinnish, K.K., Strassberg, D.S., Turner, C.W., 2005. Sex differences in the flexibility of sexual orientation: a multidimensional retrospective assessment. Arch. Sex. Behav. 34 (2), 173e183. Kitazawa, S., Kansaku, K., 2005. Sex difference in language lateralization may be task-dependent. Brain 128, E30. Klein, S.L., Schiebinger, L., Stefanick, M.L., Cahill, L., Danska, J., de Vries, G.J., et al., 2015. Opinion: sex inclusion in basic research drives discovery. Proc. Natl. Acad. Sci. U.S.A. 112 (17), 5257e5258. Kling, K.C., Hyde, J.S., Showers, C.J., Buswell, B.N., 1999. Gender differences in self-esteem: a meta-analysis. Psychol. Bull. 125 (4), 470e500. Knickmeyer, R.C., Davenport, M., 2011. Turner Syndrome and sexual differentiation of the brain: implications for understanding male-biased neurodevelopmental disorders. J. Neurodev. Disord. 3, 293e306. Kodiweera, C., Alexander, A.L., Harezlak, J., McAllister, T.W., Wu, Y.C., 2016. Age effects and sex differences in human brain white matter of young to middle-aged adults: a DTI, NODDI, and q-space study. Neuroimage 128, 180e192. Kokstejn, J., Musálek, M., Tufano, J.J., 2017. Are sex differences in fundamental motor skills uniform throughout the entire preschool period? PLoS One 12 (4), e0176556. Konrad, C., Engelien, A., Schöning, S., Zwitserlood, P., Jansen, A., Pletziger, E., et al., 2008. The functional anatomy of semantic retrieval is influenced by gender, menstrual cycle, and sex hormones. J. Neural Transm. 115 (9), 1327e1337. Koolschijn, P., Crone, E.A., 2013. Sex differences and structural brain maturation from childhood to early adulthood. Dev. Cogn. Neurosci. 5, 106e118. Korpaisarn, S., Safer, J.D., 2019. Etiology of gender identity. Endocrinol Metab. Clin. N. Am. 48 (2), 323-þ. Koscik, T., O’Leary, D., Moser, D.J., Andreasen, N.C., Nopoulos, P., 2009. Sex differences in parietal lobe morphology: relationship to mental rotation performance. Brain Cogn. 69 (3), 451e459. Kranz, G.S., Hahn, A., Kaufmann, U., Kublbock, M., Hummer, A., Ganger, S., et al., 2014. White matter microstructure in transsexuals and controls investigated by diffusion tensor imaging. J. Neurosci. 34 (46), 15466e15475. Kreukels, B.P.C., Kohler, B., Nordenstrom, A., Roehle, R., Thyen, U., Bouvattier, C., et al., 2018. Gender dysphoria and gender change in disorders of sex development/intersex conditions: results from the dsd-LIFE Study. J. Sex. Med. 15 (5), 777e785. Krogsrud, S.K., Fjell, A., Tamnes, C.K., Grydeland, H., Mork, L., Due-Tonnessen, P., et al., 2016. Changes in white matter microstructure in the developing brain-A longitudinal diffusion tensor imaging study of children from 4 to 11 years of age. Neuroimage 124, 473e486. Kyratzis, A., 2001. Emotion talk in preschool same-sex friendship groups: fluidity over time and context. Early Educ. Dev. 12 (3), 359e392.

630 PART | II Cognitive development

Lamminmaki, A., Hines, M., Kuiri-Hanninen, T., Kilpelainen, L., Dunkel, L., Sankilampi, U., 2012. Testosterone measured in infancy predicts subsequent sex-typed behavior in boys and in girls. Horm. Behav. 61 (4), 611e616. Lansdell, H., Davie, J.C., 1972. Massa intermedia: possible relation to intelligence. Neuropsychologia 10 (2), 207e210. Largo, R.H., Caflisch, J.A., Hug, F., Muggli, K., Molnar, A.A., Molinari, L., 2001. Neuromotor development from 5 to 18 years. Part 2: associated movements. Dev. Med. Child Neurol. 43 (7), 444e453. Largo, R.H., Caflisch, J.A., Hug, F., Muggli, K., Molnar, A.A., Molinari, L., et al., 2001. Neuromotor development from 5 to 18 years. Part 1: timed performance. Dev. Med. Child Neurol. 43 (7), 436e443. Lawton, C.A., 2010. Gender, spatial abilities, and wayfinding. In: Chrisler, J.C., McCreary, D.R. (Eds.), Handbook of Gender Research in Psychology, Gender Research in General and Experimental Psychology, vol. 1. Springer Science þ Business Media, New York, pp. 317e341. Lebel, C., Walker, L., Leemans, A., Phillips, L., Beaulieu, C., 2008. Microstructural maturation of the human brain from childhood to adulthood. Neuroimage 40 (3), 1044e1055. Lenroot, R.K., Giedd, J.N., 2010. Sex differences in the adolescent brain. Brain Cogn. 72 (1), 46e55. Lenroot, R.K., Gogtay, N., Greenstein, D.K., Wells, E.M., Wallace, G.L., Clasen, L.S., et al., 2007. Sexual dimorphism of brain developmental trajectories during childhood and adolescence. Neuroimage 36 (4), 1065e1073. Lentini, E., Kasahara, M., Arver, S., Savic, I., 2013. Sex differences in the human brain and the impact of sex chromosomes and sex hormones. Cerebr. Cortex 23 (10), 2322e2336. Leonard, C.M., Towler, S., Welcome, S., Halderman, L.K., Otto, R., Eckert, M.A., Chiarello, C., 2008. Size matters: cerebral volume influences sex differences in neuroanatomy. Cerebr. Cortex 18 (12), 2920e2931. Lepage, J.F., Mazaika, P.K., Hong, D.S., Raman, M., Reiss, A.L., 2013. Cortical brain morphology in young, estrogen-naive, and adolescent, estrogentreated girls with Turner Syndrome. Cerebr. Cortex 23 (9), 2159e2168. Levay, S., 1991. A difference in hypothalamic structure between heterosexual and homosexual men. Science 253 (5023), 1034e1037. Levine, S.C., Vasilyeva, M., Lourenco, S.F., Newcombe, N.S., Huttenlocher, J., 2005. Socioeconomic status modifies the sex difference in spatial skill. Psychol. Sci. 16 (11), 841e845. Li, K., Huang, X.Y., Han, Y.P., Zhang, J., Lai, Y.H., Yuan, L., et al., 2015. Enhanced neuroactivation during working memory task in postmenopausal women receiving hormone therapy: a coordinate-based meta-analysis. Front. Hum. Neurosci. 9, 9. Liben, L.S., 1995. Psychology meets geography: exploring the gender gap on the national geography bee. Psychol. Sci. Agenda 8, 8e9. Lim, S., Han, C.E., Uhlhaas, P.J., Kaiser, M., 2015. Preferential detachment during human brain development: age- and sex-specific structural connectivity in diffusion tensor imaging (DTI) data. Cerebr. Cortex 25 (6), 1477e1489. Lindberg, S.M., Hyde, J.S., Petersen, J.L., Linn, M.C., 2010. New trends in gender and mathematics performance: a meta-analysis. Psychol. Bull. 136 (6), 1123e1135. Lippa, R.A., 2008. Sex differences and sexual orientation differences in personality: findings from the BBC internet survey. Arch. Sex. Behav. 37 (1), 173e187. Lisofsky, N., Martensson, J., Eckert, A., Lindenberger, U., Gallinat, J., Kuhn, S., 2015. Hippocampal volume and functional connectivity changes during the female menstrual cycle. Neuroimage 118, 154e162. Liu, H.S., Stufflebeam, S.M., Sepulcre, J., Hedden, T., Buckner, R.L., 2009. Evidence from intrinsic activity that asymmetry of the human brain is controlled by multiple factors. Proc. Natl. Acad. Sci. U.S.A. 106 (48), 20499e20503. Lombardo, M.V., Ashwin, E., Auyeung, B., Chakrabarti, B., Taylor, K., Hackett, G., et al., 2012. Fetal testosterone influences sexually dimorphic gray matter in the human brain. J. Neurosci. 32 (2), 674e680. Lombardo, M.V., Auyeung, B., Pramparo, T., Quartier, A., Courraud, J., Holt, R.J., et al., 2019. Sex-specific impact of prenatal androgens on social brain default mode subsystems. Mol. Psychiatry. Lopez-Larson, M.P., Anderson, J.S., Ferguson, M.A., Yurgelun-Todd, D., 2011. Local brain connectivity and associations with gender and age. Dev. Cogn. Neurosci. 1 (2), 187e197. Lord, C., Buss, C., Lupien, S.J., Pruessner, J.C., 2008. Hippocampal volumes are larger in postmenopausal women using estrogen therapy compared to past users, never users and men: a possible window of opportunity effect. Neurobiol. Aging 29 (1), 95e101. Luders, E., Gaser, C., Narr, K.L., Toga, A.W., 2009. Why sex matters: brain size independent differences in gray matter distributions between men and women. J. Neurosci. 29 (45), 14265e14270. Luders, E., Harr, K.L., Thompson, P.M., Rex, D.E., Woods, R.P., DeLuca, H., et al., 2006. Gender effects on cortical thickness and the influence of scaling. Hum. Brain Mapp. 27 (4), 314e324. Luders, E., Narr, K.I., Thompson, P.M., Rex, D.E., Jancke, L., Steinmetz, H., Toga, A.W., 2004. Gender differences in cortical complexity. Nat. Neurosci. 7 (8), 799e800. Luders, E., Narr, K.L., Thompson, P.M., Toga, A.W., 2009. Neuroanatomical correlates of intelligence. Intelligence 37 (2), 156e163. Luders, E., Narr, K.L., Zaidel, E., Thompson, P.M., Toga, A.W., 2006. Gender effects on callosal thickness in scaled and unscaled space. Neuroreport 17 (11), 1103e1106. Luders, E., Steinmetz, H., Jancke, L., 2002. Brain size and grey matter volume in the healthy human brain. Neuroreport 13 (17), 2371e2374. Luders, E., Thompson, P.M., Narr, K.L., Toga, A.W., Jancke, L., Gaser, C., 2006. A curvature-based approach to estimate local gyrification on the cortical surface. Neuroimage 29 (4), 1224e1230. Luders, E., Thompson, P.M., Narr, K.L., Zamanyan, A., Chou, Y.Y., Gutman, B., et al., 2011. The link between callosal thickness and intelligence in healthy children and adolescents. Neuroimage 54 (3), 1823e1830.

Sex differences in brain and behavioral development Chapter | 27

631

Luders, E., Thompson, P.M., Toga, A.W., 2010. The development of the corpus callosum in the healthy human brain. J. Neurosci. 30 (33), 10985e10990. Luders, E., Toga, A.W., 2010. Sex differences in brain anatomy. Prog. Brain Res. 186, 3e12. Luders, E., Toga, A.W., Thompson, P.M., 2014. Why size matters: differences in brain volume account for apparent sex differences in callosal anatomy the sexual dimorphism of the corpus callosum. Neuroimage 84, 820e824. Luna, B., Velanova, K., Geier, C.F., 2010. Methodological approaches in developmental neuroimaging studies. Hum. Brain Mapp. 31 (6), 863e871. Lungu, O., Potvin, S., Tikasz, A., Mendrek, A., 2015. Sex differences in effective fronto-limbic connectivity during negative emotion processing. Psychoneuroendocrinology 62, 180e188. Lust, J.M., Geuze, R.H., Van de Beek, C., Cohen-Kettenis, P.T., Groothuis, A.G.G., Bouma, A., 2010. Sex specific effect of prenatal testosterone on language lateralization in children. Neuropsychologia 48 (2), 536e540. Lv, B., Li, J., He, H.G., Li, M., Zhao, M.C., Ai, L.K., et al., 2010. Gender consistency and difference in healthy adults revealed by cortical thickness. Neuroimage 53 (2), 373e382. Maccoby, E.E., 1998. The Two Sexes: Growing up Apart, Coming Together. Harvard University Press, Cambridge, MA. Maeda, Y., Yoon, S.Y., 2013. A meta-analysis on gender differences in mental rotation ability measured by the Purdue Spatial Visualization Tests: visualization of rotations (PSVT:R). Educ. Psychol. Rev. 25 (1), 69e94. Maeda, Y., Yoon, S.Y., 2016. Are gender differences in spatial ability real or an artifact? Evaluation of measurement invariance on the Revised PSVT:R. J. Psychoeduc. Assess. 34 (4), 397e403. Maguire, E.A., Gadian, D.G., Johnsrude, I.S., Good, C.D., Ashburner, J., Frackowiak, R.S., Frith, C.D., 2000. Navigation-related structural change in the hippocampi of taxi drivers. Proc. Natl. Acad. Sci. 97, 4398e4403. Maguire, E.A., Spiers, H.J., Good, C.D., Hartley, T., Frackowiak, R.S.J., Burgess, N., 2003. Navigation expertise and the human hippocampus: a structural brain imaging analysis. Hippocampus 13 (2), 250e259. Mak, L.E., Minuzzi, L., MacQueen, G., Hall, G., Kennedy, S.H., Milev, R., 2017. The default mode network in healthy individuals: a systematic review and meta-analysis. Brain Connect. 7 (1), 25e33. Maki, P.M., Henderson, V.W., 2012. Hormone therapy, dementia, and cognition: the Women’s Health Initiative 10 years on. Climacteric 15 (3), 256e262. Maki, P.M., Resnick, S.M., 2000. Longitudinal effects of estrogen replacement therapy on PET cerebral blood flow and cognition. Neurobiol. Aging 21 (2), 373e383. Maki, P.M., Resnick, S.M., 2001. Effects of estrogen on patterns of brain activity at rest and during cognitive activity: a review of neuroimaging studies. Neuroimage 14 (4), 789e801. Maki, P.M., Rich, J.B., Rosenbaum, R.S., 2002. Implicit memory varies across the menstrual cycle: estrogen effects in young women. Neuropsychologia 40 (5), 518e529. Maki, P.M., Sundermann, E., 2009. Hormone therapy and cognitive function. Hum. Reprod. Update 15, 667e681. Malouf, M.A., Migeon, C.J., Carson, K.A., Petrucci, L., Wisniewski, A.B., 2006. Cognitive outcome in adult women affected by congenital adrenal hyperplasia due to 21-hydroxylase deficiency. Horm. Res. 65, 142e150. Mandy, W., Chilvers, R., Chowdhury, U., Salter, G., Seigal, A., Skuse, D., 2012. Sex differences in autism spectrum disorder: evidence from a large sample of children and adolescents. J. Autism Dev. Disord. 42 (7), 1304e1313. Marrelec, G., Fransson, P., 2011. Assessing the influence of different ROI selection strategies on functional connectivity analyses of fMRI data acquired during steady-state conditions. PLoS One 6 (4), 14. Martel, M.M., 2013. Sexual selection and sex differences in the prevalence of childhood externalizing and adolescent internalizing disorders. Psychol. Bull. 139 (6), 1221e1259. Martel, M.M., Klump, K., Nigg, J.T., Breedlove, S.M., Sisk, C.L., 2009. Potential hormonal mechanisms of attention-deficit/hyperactivity disorder and major depressive disorder: a new perspective. Horm. Behav. 55 (4), 465e479. Martin, C.L., Andrews, N.C., England, D.E., Zosuls, K., Ruble, D.N., 2017. A dual identity approach for conceptualizing and measuring children’s gender identity. Child Dev. 88 (1), 167e182. Martin, C.L., Fabes, R.A., 2001. The stability and consequences of young children’s same-sex peer interactions. Dev. Psychol. 37 (3), 431e446. Martin, C.L., Ruble, D., 2004. Children’s search for gender cues: cognitive perspectives on gender development. Curr. Dir. Psychol. Sci. 13 (2), 67e70. Martin, C.L., Ruble, D.N., 2010. Patterns of gender development. Annu. Rev. Psychol. 61, 353e381. Martin, C.L., Ruble, D.N., Szkrybalo, J., 2002. Cognitive theories of early gender development. Psychol. Bull. 128 (6), 903e933. Marwha, D., Halari, M., Eliot, L., 2017. Meta-analysis reveals a lack of sexual dimorphism in human amygdala volume. Neuroimage 147, 282e294. Maylor, E.A., Reimers, S., Choi, J., Collaer, M.L., Peters, M., Silverman, I., 2007. Gender and sexual orientation differences in cognition across adulthood: age is kinder to women than to men regardless of sexual orientation. Arch. Sex. Behav. 36 (2), 235e249. Mazzone, L., Mueller, S.C., Maheu, F., VanRyzin, C., Merke, D.P., Ernst, M., 2011. Emotional memory in early steroid abnormalities: an fMRI study of adolescents with congenital adrenal hyperplasia. Dev. Neuropsychol. 36 (4), 473e492. McCabe, D.P., Castel, A.D., 2008. Seeing is believing: the effect of brain images on judgments of scientific reasoning. Cognition 107 (1), 343e352. McCarthy, M.M., 2016. Multifaceted origins of sex differences in the brain. Philos. Trans. R. Soc. Biol. Sci. 371 (1688), 11. McCarthy, M.M., Arnold, A.P., 2011. Reframing sexual differentiation of the brain. Nat. Neurosci. 14 (6), 677e683. McCarthy, M.M., Woolley, C.S., Arnold, A.P., 2017. Incorporating sex as a biological variable in neuroscience: what do we gain? Nat. Rev. Neurosci. 18, 707e708.

632 PART | II Cognitive development

McClure, E.B., 2000. A meta-analytic review of sex differences in facial expression processing and their development in infants, children, and adolescents. Psychol. Bull. 126 (3), 424e453. McCormick, C.M., Teillon, S.M., 2001. Menstrual cycle variation in spatial ability: relation to salivary cortisol levels. Horm. Behav. 39 (1), 29e38. McEwen, B.S., 2016. In pursuit of resilience: stress, epigenetics, and brain plasticity. In: Sequeira, S. (Ed.), Special Issue: Advances in Meditation Research, vol. 1373. Blackwell Science Publ, Oxford, pp. 56e64. McGlone, J., 1980. Sex differences in human brain asymmetry: a critical survey. Behav. Brain Sci. 3 (2), 215e227. McLean, C.P., Asnaani, A., Litz, B.T., Hofmann, S.G., 2011. Gender differences in anxiety disorders: prevalence, course of illness, comorbidity and burden of illness. J. Psychiatr. Res. 45 (8), 1027e1035. Merikangas, K.R., He, J., Burnstein, M., Swanson, S.A., Avenevoli, S., Cui, L., et al., 2010. Lifetime prevalence of mental disorders in US adolescents: results from the national comorbidity study-adolescent supplement (NCS-A). J. Am. Acad. Child Adolesc. Psychiatry 49 (10). Merke, D.P., Fields, J.D., Keil, M.F., Vaituzis, A.C., Chrousos, G.P., Giedd, J.N., 2003. Children with classic congenital adrenal hyperplasia have decreased amygdala volume: potential prenatal and postnatal hormonal effects. J. Clin. Endocrinol. Metab. 88 (4), 1760e1765. Meyer-Bahlburg, H.F.L., 2005. Gender identity outcome in female-raised 46,XY persons with penile agenesis, cloacal exstrophy of the bladder, or penile ablation. Arch. Sex. Behav. 34 (4), 423e438. Meyer-Bahlburg, H.F.L., Dolezal, C., Baker, S.W., Ehrhardt, A.A., New, M.I., 2006. Gender development in women with congenital adrenal hyperplasia as a function of disorder severity. Arch. Sex. Behav. 35, 667e684. Meyer, G.J., Finn, S.E., Eyde, L.D., Kay, G.G., Moreland, K.L., Dies, R.R., et al., 2001. Psychological testing and psychological assessment: a review of evidence and issues. Am. Psychol. 56 (2), 128e165. Miller, D.I., Halpern, D.F., 2014. The new science of cognitive sex differences. Trends Cogn. Sci. 18 (1), 37e45. Miller, L.R., Marks, C., Becker, J.B., Hurn, P.D., Chen, W.J., Woodruff, T., et al., 2017. Considering sex as a biological variable in preclinical research. FASEB J. 31 (1), 29e34. Mills, K.L., Goddings, A.L., Clasen, L.S., Giedd, J.N., Blakemore, S.J., 2014. The developmental mismatch in structural brain maturation during adolescence. Dev. Neurosci. 36 (3e4), 147e160. Mills, K.L., Goddings, A.L., Herting, M.M., Meuwese, R., Blakemore, S.J., Crone, E.A., et al., 2016. Structural brain development between childhood and adulthood: convergence across four longitudinal samples. Neuroimage 141, 273e281. Moffitt, T.E., 1993. Adolescence-limited and life-course-persistent antisocial-behavior - a developmental taxonomy. Psychol. Rev. 100 (4), 674e701. Moffitt, T.E., Caspi, A., Harrington, H., Milne, B.J., 2002. Males on the life-course-persistent and adolescence-limited antisocial pathways: follow-up at age 26 years. Dev. Psychopathol. 14 (1), 179e207. Moffitt, T.E., Caspi, A., Rutter, M., Silva, P.A., 2001. Sex Differences in Antisocial Behaviour: Conduct Disorder, Delinquency, and Violence in the Dunedin Longitudinal Study. Cambridge University Press, New York. Mondschein, E.R., Adolph, K.E., Tamis-LeMonda, C.S., 2000. Gender bias in mothers’ expectations about infant crawling. J. Exp. Child Psychol. 77 (4), 304e316. Money, J., Ehrhardt, A.A., 1972. Man and Woman, Boy and Girl. Johns Hopkins University Press, Baltimore. Moore, D.S., Johnson, S.P., 2008. Mental rotation in human infants: a sex difference. Psychol. Sci. 19 (11), 1063e1066. Moore, W.E., Pfeifer, J.H., Masten, C.L., Mazziotta, J.C., Iacoboni, M., Dapretto, M., 2012. Facing puberty: associations between pubertal development and neural responses to affective facial displays. Soc. Cogn. Affect. Neurosci. 7 (1), 35e43. Morrongiello, B.A., Matheis, S., 2007. Understanding children’s injury-risk behaviors: the independent contributions of cognitions and emotions. J. Pediatr. Psychol. 32 (8), 926e937. Mota, J., Esculcas, C., 2002. Leisure-time physical activity behavior: structured and unstructured choices according to sex, age, and level of physical activity. Int. J. Behav. Med. 9 (2), 111e121. Mueller, S.C., Mandell, D., Leschek, E.W., Pine, D.S., Merke, D.P., Ernst, M., 2009. Early hyperandrogenism affects the development of hippocampal function: preliminary evidence from a functional magnetic resonance imaging study of boys with familial male precocious puberty. J. Child Adolesc. Psychopharmacol. 19 (1), 41e50. Mueller, S.C., Merke, D.P., Leschek, E.W., Fromm, S., Grillon, C., Cornwell, B.R., et al., 2011. Grey matter volume correlates with virtual water maze task performance in boys with androgen excess. Neuroscience 197, 225e232. Mueller, S.C., Merke, D.P., Leschek, E.W., Fromm, S., VanRyzin, C., Ernst, M., 2011. Increased medial temporal lobe and striatal grey-matter volume in a rare disorder of androgen excess: a voxel-based morphometry (VBM) study. Int. J. Neuropsychopharmacol. 14 (4), 445e457. Mueller, S.C., Temple, V., Oh, E., VanRyzin, C., Williams, A., Cornwell, B., et al., 2008. Early androgen exposure modulates spatial cognition in congenital adrenal hyperplasia (CAH). Psychoneuroendocrinology 33, 973e980. Mueller, S.C., Verwilst, T., Van Branteghem, A., T’Sjoen, G., Cools, M., 2016. The contribution of the androgen receptor (AR) in human spatial learning and memory: a study in women with complete androgen insensitivity syndrome (CAIS). Horm. Behav. 78, 121e126. Mutlu, A.K., Schneider, M., Debbane, M., Badoud, D., Eliez, S., Schaer, M., 2013. Sex differences in thickness, and folding developments throughout the cortex. Neuroimage 82, 200e207. Nass, R., Heier, L., Moshang, T., Oberfield, S., George, A., New, M.I., Speiser, P.W., 1997. Magnetic resonance imaging in the congenital adrenal hyperplasia population: increased frequency of white-matter abnormalities and temporal lobe atrophy. J. Child Neurol. 12 (3), 181e186. Negriff, S., Susman, E.J., 2011. Pubertal timing, depression, and externalizing problems: a framework, review, and examination of gender differences. J. Res. Adolesc. 21 (3), 717e746. Nesselroade, J.R., 2011. Psychologists doing empirical research d selection at work! Perspect. Psychol. Sci. 6 (4), 319e320.

Sex differences in brain and behavioral development Chapter | 27

633

Newcombe, N., Bandura, M.M., Taylor, D.G., 1983. Sex differences in spatial ability and spatial activities. Sex Roles 9, 377e386. Newcombe, N., Dubas, J.S., 1992. A longitudinal study of predictors of spatial ability in adolescent females. Child Dev. 63, 37e46. Newman, S.D., Hansen, M.T., Gutierrez, A., 2016. An fMRI study of the impact of block building and board games on spatial ability. Front. Psychol. 7, 9. Nguyen, T.V., McCracken, J., Ducharme, S., Botteron, K.N., Mahabir, M., Johnson, W., et al., 2013. Testosterone-related cortical maturation across childhood and adolescence. Cerebr. Cortex 23 (6), 1424e1432. Nolen-Hoeksema, S., Hilt, L., 2009. Gender differences in depression. In: Gotlieb, I.H., Hammen, C.L. (Eds.), Handbook of Depression, second ed. Guilford, NewYork, pp. 386e404. Nopoulos, P., Flaum, M., O’Leary, D., Andreasen, N.C., 2000. Sexual dimorphism in the human brain: evaluation of tissue volume, tissue composition and surface anatomy using magnetic resonance imaging. Psychiatry Res. Neuroimaging 98 (1), 1e13. Nordenström, A., Servin, A., Bohlin, G., Larsson, A., Wedell, A., 2002. Sex-typed toy play behavior correlates with the degree of prenatal androgen exposure assessed by CYP21 genotype in girls with congenital adrenal hyperplasia. J. Clin. Endocrinol. Metab. 87 (11), 5119e5124. Ohnishi, T., Matsuda, H., Hirakata, M., Ugawa, Y., 2006. Navigation ability dependent neural activation in the human brain: an fMRI study. Neurosci. Res. 55 (4), 361e369. Oinonen, K.A., Mazmanian, D., 2002. To what extent do oral contraceptives influence mood and affect? J. Affect. Disord. 70 (3), 229e240. Op de Macks, Z.A., Bunge, S.A., Bell, O.N., Wilbrecht, L., Kriegsfeld, L.J., Kayser, A.S., Dahl, R.E., 2016. Risky decision-making in adolescent girls: the role of pubertal hormones and reward circuitry. Psychoneuroendocrinology 74, 77e91. Op de Macks, Z.A., Moor, B.G., Overgaauw, S., Guroglu, B., Dahl, R.E., Crone, E.A., 2011. Testosterone levels correspond with increased ventral striatum activation in response to monetary rewards in adolescents. Dev. Cogni. Neurosci. 1 (4), 506e516. Ortigue, S., Thut, G., Landis, T., Michel, C.M., 2005. Time-resolved sex differences in language lateralization. Brain 128, E28. Ossewaarde, L., van Wingen, G.A., Rijpkema, M., Backstrom, T., Hermans, E.J., Fernandez, G., 2013. Menstrual cycle-related changes in amygdala morphology are associated with changes in stress sensitivity. Hum. Brain Mapp. 34 (5), 1187e1193. Pargulski, J.R., Reynolds, M.R., 2017. Sex differences in achievement: distributions matter. Personal. Individ. Differ. 104, 272e278. Pasterski, V., Acerini, C.L., Dunger, D.B., Ong, K.K., Hughes, I.A., Thankamony, A., Hines, M., 2015. Postnatal penile growth concurrent with minipuberty predicts later sex-typed play behavior: evidence for neurobehavioral effects of the postnatal androgen surge in typically developing boys. Horm. Behav. 69, 98e105. Pasterski, V.L., Geffner, M.E., Brain, C., Hindmarsh, P., Brook, C., Hines, M., 2005. Prenatal hormones and postnatal socialization by parents as determinants of male-typical toy play in girls with congenital adrenal hyperplasia. Child Dev. 76, 264e278. Patrick, C.J., Hajcak, G., 2016. RDoC: translating promise into progress. Psychophysiology 53 (3), 415e424. Paus, T., 2010. Sex differences in the human brain: a developmental perspective. Prog. Brain Res. 186, 13e28. Paus, T., Nawaz-Khan, I., Leonard, G., Perron, M., Pike, G.B., Pitiot, A., et al., 2010. Sexual dimorphism in the adolescent brain: role of testosterone and androgen receptor in global and local volumes of grey and white matter. Horm. Behav. 57 (1), 63e75. Peper, J.S., Hulshoff Pol, H.E., Crone, E.A., van Honk, J., 2011. Sex steroids and brain structure in pubertal boys and girls: a mini-review of neuroimaging studies. Neurosci. Biobehav. Rev. 191, 28e37. Perrin, J.S., Hervé, P.-Y., Leonard, G., Perron, M., Pike, G.B., Pitiot, A., et al., 2008. Growth of white matter in the adolescent brain: role of testosterone and androgen receptor. J. Neurosci. 28 (38), 9519e9524. Persad, C.C., Zubieta, J.K., Love, T., Wang, H., Tkaczyk, A., Smith, Y.R., 2009. Enhanced neuroactivation during verbal memory processing in postmenopausal women receiving short-term hormone therapy. Fertil. Steril. 92 (1), 197e204. Persson, J., Spreng, R.N., Turner, G., Herlitz, A., Morell, A., Stening, E., et al., 2014. Sex differences in volume and structural covariance of the anterior and posterior hippocampus. Neuroimage 99, 215e225. Pessoa, L., Adolphs, R., 2010. Emotion processing and the amygdala: from a ’low road’ to ’many roads’ of evaluating biological significance. Nat. Rev. Neurosci. 11 (11), 773e782. Peters, M., 1991. Sex differences in human brain size and the general meaning of differences in brain size. Can. J. Psychol. 45 (4), 507e522. Peters, M., Battista, C., 2008. Applications of mental rotation figures of the Shepard and Metzler type and description of a mental rotation stimulus library. Brain Cogn. 66 (3), 260e264. Peters, M., Jäncke, L., Staiger, J.F., Schlaug, G., Huang, Y., Steinmetz, H., 1998. Unsolved problems in comparing brain sizes in Homo sapiens. Brain Cogn. 37 (2), 254e285. Petersen, J., 2018. Gender difference in verbal performance: a meta-analysis of United States state performance assessments. Educ. Psychol. Rev. 30 (4), 1269e1281. Petersen, J.L., Hyde, J.S., 2010. A meta-analytic review of research on gender differences in sexuality, 1993-2007. Psychol. Bull. 136 (1), 21e38. Petersen, N., Cahill, L., 2015. Amygdala reactivity to negative stimuli is influenced by oral contraceptive use. Soc. Cogn. Affect. Neurosci. 10 (9), 1266e1272. Petersen, N., Kilpatrick, L.A., Goharzad, A., Cahill, L., 2014. Oral contraceptive pill use and menstrual cycle phase are associated with altered resting state functional connectivity. Neuroimage 90, 24e32. Pfeifer, J.H., Allen, N.B., 2012. Arrested development? Reconsidering dual-systems models of brain function in adolescence and disorders. Trends Cogn. Sci. 16 (6), 322e329. Phoenix, C.H., Goy, R.W., Gerall, A.A., Young, W.C., 1959. Organizing action of prenatally administered testosterone propionate on the tissues mediating mating behavior in the female Guinea pig. Endocrinology 65, 369e382.

634 PART | II Cognitive development

Phoenix, C.H., Slob, A.K., Goy, R.W., 1973. Effects of castration and replacement therapy on sexual behavior of adult male rhesuses. J. Comp. Physiol. Psychol. 84 (3), 472e481. Plante, E., Schmithorst, V.J., Holland, S.K., Byars, A.W., 2006. Sex differences in the activation of language cortex during childhood. Neuropsychologia 44 (7), 1210e1221. Pletzer, B., Kronbichler, M., Aichhorn, M., Bergmann, J., Ladurner, G., Kerschbaum, H.H., 2010. Menstrual cycle and hormonal contraceptive use modulate human brain structure. Brain Res. 1348, 55e62. Pletzer, B., Kronbichler, M., Kerschbaum, H., 2015. Differential effects of androgenic and anti-androgenic progestins on fusiform and frontal gray matter volume and face recognition performance. Brain Res. 1596, 108e115. Pletzer, B., Kronbichler, M., Nuerk, H.C., Kerschbaum, H., 2014. Hormonal contraceptives masculinize brain activation patterns in the absence of behavioral changes in two numerical tasks. Brain Res. 1543, 128e142. Polderman, T., Kreukels, B., Irwig, M., Beach, L., Chan, Y.M., Derks, E., et al., 2018. The biological contributions to gender identity and gender diversity: bringing data to the table. Behav. Genet. 48 (2), 95e108. Poldrack, R.A., 2000. Imaging brain plasticity: conceptual and methodological issues - a theoretical review. Neuroimage 12 (1), 1e13. Poldrack, R.A., Baker, C.I., Durnez, J., Gorgolewski, K.J., Matthews, P.M., Munafo, M.R., et al., 2017. Scanning the horizon: towards transparent and reproducible neuroimaging research. Nat. Rev. Neurosci. 18 (2), 115e126. Ponsoda, V., Martinez, K., Pineda-Pardo, J.A., Abad, F.J., Olea, J., Roman, F.J., et al., 2017. Structural brain connectivity and cognitive ability differences: a multivariate distance matrix regression analysis. Hum. Brain Mapp. 38 (2), 803e816. Power, J.D., Barnes, K.A., Snyder, A.Z., Schlaggar, B.L., Petersen, S.E., 2012. Spurious but systematic correlations in functional connectivity MRI networks arise from subject motion. Neuroimage 59 (3), 2142e2154. Power, J.D., Fair, D.A., Schlaggar, B.L., Petersen, S.E., 2010. The development of human functional brain networks. Neuron 67 (5), 735e748. Protopopescu, X., Butler, T., Pan, H., Root, J., Altemus, M., Polanecsky, M., et al., 2008. Hippocampal structural changes across the menstrual cycle. Hippocampus 18 (10), 985e988. Puts, D.A., Cárdenas, R.A., Bailey, D.H., Burriss, R.P., Jordan, C.L., Breedlove, S.M., 2010. Salivary testosterone does not predict mental rotation performance in men or women. Horm. Behav. 58 (2), 282e289. Puts, D.A., McDaniel, M.A., Jordan, C.L., Breedlove, N.J., 2008. Spatial ability and prenatal androgens: meta-analyses of congenital adrenal hyperplasia and digit ratio (2D:4D) studies. Arch. Sex. Behav. 37 (1), 100e111. Quinn, P.C., Liben, L.S., 2008. A sex difference in mental rotation in young infants. Psychol. Sci. 19 (11), 1067e1070. Rabinowicz, T., Dean, D.E., Petetot, J.M.C., de Courten-Myers, G.M., 1999. Gender differences in the human cerebral cortex: more neurons in males; More processes in females. J. Child Neurol. 14 (2), 98e107. Rabl, R., 1958. Strukturstudien an der aassa intermedia des thalamus opticus. J. fur Hirnforschung 4 (78e112). Ramtekkar, U.P., Reiersen, A.M., Todorov, A.A., Todd, R.D., 2010. Sex and age differences in attention-deficit/hyperactivity disorder symptoms and diagnoses: implications for DSM-V and ICD-11. J. Am. Acad. Child Adolesc. Psychiatry 49 (3), 217e228. Raznahan, A., Lee, Y., Stidd, R., Long, R., Greenstein, D., Clasen, L., et al., 2010. Longitudinally mapping the influence of sex and androgen signaling on the dynamics of human cortical maturation in adolescence. Proc. Natl. Acad. Sci. U.S.A. 107 (39), 16988e16993. Raznahan, A., Shaw, P.W., Lerch, J.P., Clasen, L.S., Greenstein, D., Berman, R., et al., 2014. Longitudinal four-dimensional mapping of subcortical anatomy in human development. Proc. Natl. Acad. Sci. U.S.A. 111 (4), 1592e1597. Reber, J., Tranel, D., 2017. Sex differences in the functional lateralization of emotion and decision making in the human brain. J. Neurosci. Res. 95 (1e2), 270e278. Reed, J.L., Gallagher, N.M., Sullivan, M., Callicott, J.H., Green, A.E., 2017. Sex differences in verbal working memory performance emerge at very high loads of common neuroimaging tasks. Brain Cogn. 113, 56e64. Reilly, D., Neumann, D.L., Andrews, G., 2015. Sex differences in mathematics and science achievement: a meta-analysis of national assessment of educational progress assessments. J. Educ. Psychol. 107 (3), 645e662. Reinius, B., Jazin, E., 2009. Prenatal sex differences in the human brain. Mol. Psychiatry 14 (11), 988e989. Resnick, S.M., Berenbaum, S.A., Gottesman, I.I., Bouchard, T.J., 1986. Early hormonal influences on cognitive functioning in congenital adrenal hyperplasia. Dev. Psychol. 22, 191e198. Resnick, S.M., Espeland, M.A., Jaramillo, S.A., Hirsch, C., Stefanick, M.L., Murray, A.M., et al., 2009. Postmenopausal hormone therapy and regional brain volumes the WHIMS-MRI Study. Neurology 72 (2), 135e142. Resnick, S.M., Maki, P.M., Golski, S., Kraut, M.A., Zonderman, A.B., 1998. Effects of estrogen replacement therapy on PET cerebral blood flow and neuropsychological performance. Horm. Behav. 34 (2), 171e182. Resnick, S.M., Metter, E.J., Zonderman, A.B., 1997. Estrogen replacement therapy and longitudinal decline in visual memory - a possible protective effect? Neurology 49 (6), 1491e1497. Robertson, D., Craig, M., van Amelsvoort, T., Daly, E., Moore, C., Simmons, A., et al., 2009. Effects of estrogen therapy on age-related differences in gray matter concentration. Climacteric 12 (4), 301e309. Robinson, J.P., Lubienski, S.T., 2011. The development of gender achievement gaps in mathematics and reading during elementary and middle School: examining direct cognitive assessments and teacher ratings. Am. Educ. Res. J. 48 (2), 268e302. Roivainen, E., 2011. Gender differences in processing speed: a review of recent research. Learn. Individ. Differ. 21 (2), 145e149. Rose, A.J., Smith, R.L., Glick, G.C., Schwartz-Mette, R.A., 2016. Girls’ and boys’ problem talk: implications for emotional closeness in friendships. Dev. Psychol. 52 (4), 629e639.

Sex differences in brain and behavioral development Chapter | 27

635

Ross, J., Roeltgen, D., Zinn, A., 2006. Cognition and the sex chromosomes: studies in turner syndrome. Horm. Res. 65, 47e56. Rubia, K., Hyde, Z., Halari, R., Giampietro, V., Smith, A., 2010. Effects of age and sex on developmental neural networks of visual-spatial attention allocation. Neuroimage 51 (2), 817e827. Ruble, D.N., Martin, C.L., Berenbaum, S.A., 2006. Gender development. In: Eisenberg, N. (Ed.), Handbook of Child Psychology: Vol. 3. Social, Emotional, and Personality Development, sixth ed. Wiley, Hoboken, NJ, pp. 858e932. Ruigrok, A.N.V., Salimi-Khorshidi, G., Lai, M.C., Baron-Cohen, S., Lombardo, M.V., Tait, R.J., Suckling, J., 2014. A meta-analysis of sex differences in human brain structure. Neurosci. Biobehav. Rev. 39, 34e50. Ryan, B.C., Vandenbergh, J.G., 2002. Intrauterine position effects. Neurosci. Biobehav. Rev. 26, 665e678. Sánchez-Martín, J.R., Azurmendi, A., Pascual-Sagastizabal, E., Cardas, J., Braza, F., Braza, P., et al., 2011. Androgen levels and anger and impulsivity measures as predictors of physical, verbal and indirect aggression in boys and girls. Psychoneuroendocrinology 36 (5), 750e760. Sanislow, C.A., Pine, D.S., Quinn, K.J., Kozak, M.J., Garvey, M.A., Heinssen, R.K., et al., 2010. Developing constructs for psychopathology research: research domain criteria. J. Abnorm. Psychol. 119 (4), 631e639. Satterthwaite, T.D., Vandekar, S., Wolf, D.H., Ruparel, K., Roalf, D.R., Jackson, C., et al., 2014. Sex differences in the effect of puberty on hippocampal morphology. J. Am. Acad. Child Adolesc. Psychiatry 53 (3), 341e350. Satterthwaite, T.D., Wolf, D.H., Loughead, J., Ruparel, K., Elliott, M.A., Hakonarson, H., et al., 2012. Impact of in-scanner head motion on multiple measures of functional connectivity: relevance for studies of neurodevelopment in youth. Neuroimage 60 (1), 623e632. Satterthwaite, T.D., Wolf, D.H., Roalf, D.R., Ruparel, K., Erus, G., Vandekar, S., et al., 2015. Linked sex differences in cognition and functional connectivity in youth. Cerebr. Cortex 25 (9), 2383e2394. Savic, I., Frisen, L., Manzouri, A., Nordenstrom, A., Hirschberg, A.L., 2017. Role of testosterone and Y chromosome genes for the masculinization of the human brain. Hum. Brain Mapp. 38 (4), 1801e1814. Savic, I., Garcia-Falgueras, A., Swaab, D.F., 2010. Sexual differentiation of the human brain in relation to gender identity and sexual orientation. Prog. Brain Res. 186, 41e62. Scheinost, D., Finn, E.S., Tokoglu, F., Shen, X.L., Papademetris, X., Hampson, M., Constable, R.T., 2015. Sex differences in normal age trajectories of functional brain networks. Hum. Brain Mapp. 36 (4), 1524e1535. Scheuringer, A., Pletzer, B., 2017. Sex differences and menstrual cycle dependent changes in cognitive strategies during spatial navigation and verbal fluency. Front. Psychol. 8, 381. Schmitt, J.E., Neale, M.C., Fassassi, B., Perez, J., Lenroot, R.K., Wells, E.M., Giedd, J.N., 2014. The dynamic role of genetics on cortical patterning during childhood and adolescence. Proc. Natl. Acad. Sci. U.S.A. 111 (18), 6774e6779. Schneider, S., Peters, J., Bromberg, U., Brassen, S., Menz, M.M., Miedl, S.F., et al., 2011. Boys do it the right way: sex-dependent amygdala lateralization during face processing in adolescents. Neuroimage 56 (3), 1847e1853. Schöning, S., Engelien, A., Kugel, H., Schäfer, S., Schiffbauer, H., Zwitserlood, P., et al., 2007. Functional anatomy of visuo-spatial working memory during mental rotation is influenced by sex, menstrual cycle, and sex steroid hormones. Neuropsychologia 45 (14), 3203e3214. Schuch, J.J.J., Roest, A.M., Nolen, W.A., Penninx, B., de Jonge, P., 2014. Gender differences in major depressive disorder: results from The Netherlands study of depression and anxiety. J. Affect. Disord. 156, 156e163. Schulz, K.M., Zehr, J.L., Salas-Ramirez, K.Y., Sisk, C.L., 2009. Testosterone programs adult social behavior before and during, but not after, adolescence. Endocrinology 150 (8), 3690e3698. Seldon, H.L., 2005. Does brain white matter growth expand the cortex like a balloon? Hypothesis and consequences. Laterality 10 (1), 81e95. Sergerie, K., Chochol, C., Armony, J.L., 2008. The role of the amygdala in emotional processing: a quantitative meta-analysis of functional neuroimaging studies. Neurosci. Biobehav. Rev. 32 (4), 811e830. Shaw, P., Gogtay, N., Rapoport, J., 2010. Childhood psychiatric disorders as anomalies in neurodevelopmental trajectories. Hum. Brain Mapp. 31 (6), 917e925. Shaw, P., Greenstein, D., Lerch, J., Clasen, L., Lenroot, R., Gogtay, N., et al., 2006. Intellectual ability and cortical development in children and adolescents. Nature 440 (7084), 676e679. Shaywitz, B.A., Shaywitz, S.E., Pugh, K.R., Constable, R.T., Skudlarski, P., Fulbright, R.K., et al., 1995. Sex differences in the functional organization of the brain for language. Nature 373 (6515), 607e609. Shaywitz, S.E., Shaywitz, B.A., Pugh, K.R., Fulbright, R.K., Skudlarski, P., Mencl, V.E., et al., 1999. Effect of estrogen on brain activation patterns in postmenopausal women during working memory tasks. JAMA 281 (13), 1197e1202. Shen, H.H., 2015. Core concept: resting-state connectivity. Proc. Natl. Acad. Sci. U.S.A. 112 (46), 14115e14116. Sherwin, B.B., 2012. Estrogen and cognitive functioning in women: lessons we have learned. Behav. Neurosci. 126 (1), 123e127. Shulman, E.P., Smith, A.R., Silva, K., Icenogle, G., Duell, N., Chein, J., Steinberg, L., 2016. The dual systems model: review, reappraisal, and reaffirmation. Dev. Cogn. Neurosci. 17, 103e117. Silverman, I., Choi, J., Peters, M., 2007. The Hunter-Gatherer theory of sex differences in spatial abilities: data from 40 countries. Arch. Sex. Behav. 36 (2), 261e268. Simmonds, D.J., Hallquist, M.N., Asato, M., Luna, B., 2014. Developmental stages and sex differences of white matter and behavioral development through adolescence: a longitudinal diffusion tensor imaging (DTI) study. Neuroimage 92, 356e368. Simpkins, S.D., Davis-Kean, P.E., Eccles, J.S., 2005. Parents’ socializing behavior and children’s participation in math, science, and computer out-ofschool activities. Appl. Dev. Sci. 9 (1), 14e30. Sisk, C.L., Zehr, J.L., 2005. Pubertal hormones organize the adolescent brain and behavior. Front. Neuroendocrinol. 26, 163e174.

636 PART | II Cognitive development

Skogli, E.W., Teicher, M.H., Andersen, P.N., Hovik, K.T., Oie, M., 2013. ADHD in girls and boys - gender differences in co-existing symptoms and executive function measures. BMC Psychiatry 13, 12. Skovlund, C., Morch, L.S., Kessing, L.V., Lidegaard, O., 2016. Association of hormonal contraception with depression. JAMA Psychiatry 73 (11), 1154e1162. Smith, D.V., Utevsky, A.V., Bland, A.R., Clement, N., Clithero, J.A., Harsch, A.E.W., et al., 2014. Characterizing individual differences in functional connectivity using dual-regression and seed-based approaches. Neuroimage 95, 1e12. Smith, S.M., Miller, K.L., Salimi-Khorshidi, G., Webster, M., Beckmann, C.F., Nichols, T.E., et al., 2011. Network modelling methods for FMRI. Neuroimage 54 (2), 875e891. Sohn, W.S., Yoo, K., Lee, Y.B., Seo, S.W., Na, D.L., Jeong, Y., 2015. Influence of ROI selection on resting state functional connectivity: an individualized approach for resting state fMRI analysis. Front. Neurosci. 9, 10. Sole-Padulles, C., Castro-Fornieles, J., de la Serna, E., Calvo, R., Baeza, I., Moya, J., et al., 2016. Intrinsic connectivity networks from childhood to late adolescence: effects of age and sex. Dev. Cogn. Neurosci. 17, 35e44. Somerville, L.H., 2016. Searching for signatures of brain maturity: what are we searching for? Neuron 92 (6), 1164e1167. Sommer, I.E., Aleman, A., Somers, M., Boks, M.P., Kahn, R.S., 2008. Sex differences in handedness, asymmetry of the Planum Temporale and functional language lateralization. Brain Res. 1206, 76e88. Sommer, I.E.C., Aleman, A., Bouma, A., Kahn, R.S., 2004. Do women really have more bilateral language representation than men? A meta-analysis of functional imaging studies. Brain 127, 1845e1852. Song, M., Zhou, Y., Li, J., Liu, Y., Tian, L.X., Yu, C.S., Jiang, T.Z., 2008. Brain spontaneous functional connectivity and intelligence. Neuroimage 41 (3), 1168e1176. Soto, C.J., John, O.P., Gosling, S.D., Potter, J., 2011. Age differences in personality traits from 10 to 65: big five domains and facets in a large crosssectional sample. J. Personal. Soc. Psychol. 100 (2), 330e348. Sowell, E.R., Peterson, B.S., Kan, E., Woods, R.P., Yoshii, J., Bansal, R., et al., 2007. Sex differences in cortical thickness mapped in 176 healthy individuals between 7 and 87 years of age. Cerebr. Cortex 17 (7), 1550e1560. Sowell, E.R., Thompson, P.M., Tessner, K.D., Toga, A.W., 2001. Mapping continued brain growth and gray matter density reduction in dorsal frontal cortex: inverse relationships during postadolescent brain maturation. J. Neurosci. 21 (22), 8819e8829. Sowell, E.R., Trauner, D.A., Gamst, A., Jernigan, T.L., 2002. Development of cortical and subcortical brain structures in childhood and adolescence: a structural MRI study. Dev. Med. Child Neurol. 44 (1), 4e16. Spencer, D., Pasterski, V., Neufeld, S., Glover, V., O’Connor, T.G., Hindmarsh, P.C., et al., 2017. Prenatal androgen exposure and children’s aggressive behavior and activity level. Horm. Behav. 96, 156e165. Spielberg, J.M., Forbes, E.E., Ladouceur, C.D., Worthman, C.M., Olino, T.M., Ryan, N.D., Dahl, R.E., 2015. Pubertal testosterone influences threatrelated amygdala-orbitofrontal cortex coupling. Soc. Cogn. Affect. Neurosci. 10 (3), 408e415. Spielberg, J.M., Galarce, E.M., Ladouceur, C.D., McMakin, D.L., Olino, T.M., Forbes, E.E., et al., 2015. Adolescent development of inhibition as a function of SES and gender: converging evidence from behavior and fMRI. Hum. Brain Mapp. 36 (8), 3194e3203. Sporns, O., 2011. Networks of the Brain. MIT Press, Cambridge, MA. Steele, C., 1997. A threat in the air. How stereotypes shape intellectual identity and performance. Am. Psychol. 52, 613e629. Steensma, T.D., Cohen-Kettenis, P.T., Zucker, K., 2018. Evidence for a change in the sex ratio of children referred for gender dysphoria: data from the center of expertise on gender dysphoria in Amsterdam (1988e2016). J. Sex Marital Ther. 1e3. Steinberg, L., 2008. A social neuroscience perspective on adolescent risk-taking. Dev. Rev. 28 (1), 78e106. Steinberg, L., 2010. A dual systems model of adolescent risk-taking. Dev. Psychobiol. 52 (3), 216e224. Stevens, J.S., Hamann, S., 2012. Sex differences in brain activation to emotional stimuli: a meta-analysis of neuroimaging studies. Neuropsychologia 50 (7), 1578e1593. Stoet, G., Geary, D.C., 2013. Sex differences in mathematics and reading achievement are inversely related: within- and across-nation assessment of 10 years of PISA data. PLoS One 8 (3), 10. Stone, M.H., 2015. Mass murder, mental illness, and men. Violence Gend. 2 (1), 51e86. Strandqvist, A., Herlitz, A., Nordenskjold, A., Ortqvist, L., Frisen, L., Hirschberg, A.L., Nordenstrom, A., 2018. Cognitive abilities in women with complete androgen insensitivity syndrome and women with gonadal dysgenesis. Psychoneuroendocrinology 98, 233e241. Su, R., Rounds, J., Armstrong, P.I., 2009. Men and things, women and people: a meta-analysis of sex differences in interests. Psychol. Bull. 135 (6), 859e884. Tan, A., Ma, W., Vira, A., Marwha, D., Eliot, L., 2016. The human hippocampus is not sexually-dimorphic: meta-analysis of structural MRI volumes. Neuroimage 124, 350e366. Telzer, E.H., McCormick, E.M., Peters, S., Cosme, D., Pfeifer, J.H., van Duijvenvoorde, A.C.K., 2018. Methodological considerations for developmental longitudinal fMRI research. Dev. Cogn. Neurosci. 33, 149e160. Thambisetty, M., Wan, J., Carass, A., An, Y., Prince, J.L., Resnick, S.M., 2010. Longitudinal changes in cortical thickness associated with normal aging. Neuroimage 52 (4), 1215e1223. Thomas, J., Metereau, E., Dechaud, H., Pugeat, M., Dreher, J.C., 2014. Hormonal treatment increases the response of the reward system at the menopause transition: a counterbalanced randomized placebo-controlled fMRI study. Psychoneuroendocrinology 50, 167e180. Thompson, A.E., Voyer, D., 2014. Sex differences in the ability to recognise non-verbal displays of emotion: a meta-analysis. Cognit. Emot. 28 (7), 1164e1195.

Sex differences in brain and behavioral development Chapter | 27

637

Thompson, P.M., Giedd, J.N., Woods, R.P., MacDonald, D., Evans, A.C., Toga, A.W., 2000. Growth patterns in the developing brain detected by using continuum mechanical tensor maps. Nature 404 (6774), 190e193. Thomsen, T., Hugdahl, K., Ersland, L., Barndon, R., Lundervold, A., Smievoll, A.I., et al., 2000. Functional magnetic resonance imaging (fMRI) study of sex differences in a mental rotation task. Med. Sci. Monit. 60 (6), 1186e1196. Tian, L.X., Wang, J.H., Yan, C.G., He, Y., 2011. Hemisphere- and gender-related differences in small-world brain networks: a resting-state functional MRI study. Neuroimage 54 (1), 191e202. Tiemeier, H., Lenroot, R.K., Greenstein, D.K., Tran, L., Pierson, R., Giedd, J.N., 2010. Cerebellum development during childhood and adolescence: a longitudinal morphometric MRI study. Neuroimage 49 (1), 63e70. Tomasi, D., Volkow, N.D., 2012. Gender differences in brain functional connectivity density. Hum. Brain Mapp. 33 (4), 849e860. Trabzuni, D., Ramasamy, A., Imran, S., Walker, R., Smith, C., Weale, M.E., et al., 2013. Widespread sex differences in gene expression and splicing in the adult human brain. Nat. Commun. 4, 7. Tracy, D.M., 1990. Toy-playing behavior, sex-role orientation, spatial ability, and science achievement. J. Res. Sci. Teach. 27 (7), 637e649. Turban, J.L., Ehrensaft, D., 2018. Research Review: gender identity in youth: treatment paradigms and controversies. J. Child Psychol. Psychiatry 59 (12), 1228e1243. Uttal, D.H., Miller, D.I., Newcombe, N.S., 2013. Exploring and enhancing spatial thinking: links to achievement in science, technology, engineering, and mathematics? Curr. Dir. Psychol. Sci. 22 (5), 367e373. van den Heuvel, M.P., Stam, C.J., Kahn, R.S., Pol, H.E.H., 2009. Efficiency of functional brain networks and intellectual performance. J. Neurosci. 29 (23), 7619e7624. van Duijvenvoorde, A.C.K., Achterberg, M., Braams, B.R., Peters, S., Crone, E.A., 2016. Testing a dual-systems model of adolescent brain development using resting-state connectivity analyses. Neuroimage 124, 409e420. van Hemmen, J., Cohen-Kettenis, P.T., Steensma, T.D., Veltman, D.J., Bakker, J., 2017. Do sex differences in CEOAEs and 2D:4D ratios reflect androgen exposure? A study in women with complete androgen insensitivity syndrome. Biol. Sex Differ. 8, 11. van Hemmen, J., Veltman, D.J., Hoekzema, E., Cohen-Kettenis, P.T., Dessens, A.B., Bakker, J., 2016. Neural activation during mental rotation in complete androgen insensitivity syndrome: the influence of sex hormones and sex chromosomes. Cerebr. Cortex 26 (3), 1036e1045. Vasta, R., Liben, L.S., 1996. The water-level task: an intriguing puzzle. Curr. Dir. Psychol. Sci. 5 (6), 171e177. Vawter, M.P., Evans, S., Choudary, P., Tomita, H., Meador-Woodruff, J., Molnar, M., et al., 2004. Gender-specific gene expression in post-mortem human brain: localization to sex chromosomes. Neuropsychopharmacology 29 (2), 373e384. Vendetti, M.S., Bunge, S.A., 2014. Evolutionary and developmental changes in the lateral frontoparietal network: a little goes a long way for higher-level cognition. Neuron 84 (5), 906e917. Vijayakumar, N., Allen, N.B., Youssef, G., Dennison, M., Yucel, M., Simmons, J.G., Whittle, S., 2016. Brain development during adolescence: a mixedlongitudinal investigation of cortical thickness, surface area, and volume. Hum. Brain Mapp. 37 (6), 2027e2038. Vogel, J.J., Bowers, C.A., Vogel, D.S., 2003. Cerebral lateralization of spatial abilities: a meta-analysis. Brain Cogn. 52 (2), 197e204. Voineskos, A.N., Winterburn, J.L., Felsky, D., Pipitone, J., Rajji, T.K., Mulsant, B.H., Chakravarty, M.M., 2015. Hippocampal (subfield) volume and shape in relation to cognitive performance across the adult lifespan. Hum. Brain Mapp. 36 (8), 3020e3037. Voyer, D., 2011. Sex differences in dichotic listening. Brain Cogn. 76 (2), 245e255. Voyer, D., Bryden, M.P., 1993. Masking and visual field effects on a lateralized rod-and-frame test. Can. J. Exp. Psychol. 47 (1), 26e37. Voyer, D., Nolan, C.L., Voyer, S., 2000. The relation between experience and spatial performance in men and women. Sex Roles 43, 891e915. Voyer, D., Postma, A., Brake, B., Imperato-McGinley, J., 2007. Gender differences in object location memory: a meta-analysis. Psychon. Bull. Rev. 14 (1), 23e38. Voyer, D., Voyer, S.D., 2014. Gender differences in scholastic achievement: a meta-analysis. Psychol. Bull. 140 (4), 1174e1204. Vul, E., Harris, C., Winkielman, P., Pashler, H., 2009. Puzzlingly high correlations in fMRI studies of emotion, personality, and social cognition. Perspect. Psychol. Sci. 4 (3), 274e290. Vuoksimaa, E., Kaprio, J., Eriksson, C.J.P., Rose, R.J., 2012. Pubertal testosterone predicts mental rotation performance of young adult males. Psychoneuroendocrinology 37, 1791e1800. Vuoksimaa, E., Kaprio, J., Kremen, W.S., Hokkanen, L., Viken, R.J., Tuulio-Henriksson, A., Rose, R.J., 2010. Having a male co-twin masculinizes mental rotation performance in females. Psychol. Sci. 21 (8), 1069e1071. Waber, D.P., 1976. Sex differences in cognition: a function of maturation rate? Am. Assoc. Adv. Sci. 192 (4239). Wager, T.D., Phan, K.L., Liberzon, I., Taylor, S.F., 2003. Valence, gender, and lateralization of functional brain anatomy in emotion: a meta-analysis of findings from neuroimaging. Neuroimage 19 (3), 513e531. Wai, J., Cacchio, M., Putallaz, M., Makel, M.C., 2010. Sex differences in the right tail of cognitive abilities: a 30 year examination. Intelligence 38 (4), 412e423. Wallen, K., 1996. Nature needs nurture: the interaction of hormonal and social influences on the development of behavioral sex differences in rhesus monkeys. Horm. Behav. 30, 364e378. Wallen, K., 2005. Hormonal influences on sexually differentiated behavior in nonhuman primates. Front. Neuroendocrinol. 26, 7e26. Wallen, K., 2009. The organizational hypothesis: reflections on the 50th anniversary of the publication of Phoenix, Goy, Gerall, and young (1959). Horm. Behav. 55 (5), 561e565. Wallentin, M., 2009. Putative sex differences in verbal abilities and language cortex: a critical review. Brain Lang. 108 (3), 175e183.

638 PART | II Cognitive development

Wang, Z.Q., Jia, X.Q., Liang, P.P., Qi, Z.G., Yang, Y.H., Zhou, W.D., Li, K.C., 2012. Changes in thalamus connectivity in mild cognitive impairment: evidence from resting state fMRI. Eur. J. Radiol. 81 (2), 277e285. Warren, A.M., Gurvich, C., Worsley, R., Kulkarni, J., 2014. A systematic review of the impact of oral contraceptives on cognition. Contraception 90 (2), 111e116. Webb, E.A., Elliott, L., Carlin, D., Wilson, M., Hall, K., Netherton, J., et al., 2018. Quantitative brain MRI in congenital adrenal hyperplasia: in vivo assessment of the cognitive and structural impact of steroid hormones. J. Clin. Endocrinol. Metab. 103 (4), 1330e1341. Weis, S., Hodgetts, S., Hausmann, M., 2019. Sex differences and menstrual cycle effects in cognitive and sensory resting state networks. Brain Cogn. 131, 66e73. Weisberg, D.S., Keil, F.C., Goodstein, J., Rawson, E., Gray, J.R., 2008. The seductive allure of neuroscience explanations. J. Cogn. Neurosci. 20 (3), 470e477. Weisberg, Y.J., DeYoung, C.G., Hirsh, J.B., 2011. Gender differences in personality across the ten aspects of the Big Five. Front. Psychol. 2. Weiss, E., Siedentopf, C.M., Hofer, A., Deisenhammer, E.A., Hoptman, M.J., Kremser, C., et al., 2003. Sex differences in brain activation pattern during a visuospatial cognitive task: a functional magnetic resonance imaging study in healthy volunteers. Neurosci. Lett. 344 (3), 169e172. Weiss, E.M., Siedentopf, C., Hofer, A., Deisenhammer, E.A., Hoptman, M.J., Kremser, C., et al., 2003. Brain activation pattern during a verbal fluency test in healthy male and female volunteers: a functional magnetic resonance imaging study. Neurosci. Lett. 352 (3), 191e194. Werling, D.M., Geschwind, D.H., 2013. Sex differences in autism spectrum disorders. Curr. Opin. Neurol. 26 (2), 146e153. Wierenga, L., Langen, M., Ambrosino, S., van Dijk, S., Oranje, B., Durston, S., 2014. Typical development of basal ganglia, hippocampus, amygdala and cerebellum from age 7 to 24. Neuroimage 96, 67e72. Willcutt, E.G., 2012. The prevalence of DSM-IV Attention-Deficit/Hyperactivity Disorder: a meta-analytic review. Neurotherapeutics 9 (3), 490e499. Wood, W., Eagly, A.H., 2002. A cross-cultural analysis of the behavior of women and men: implications for the origins of sex differences. Psychol. Bull. 128 (5), 699e727. Wraga, M., Helt, M., Jacobs, E., Sullivan, K., 2007. Neural basis of stereotype-induced shifts in women’s mental rotation performance. Soc. Cogn. Affect. Neurosci. 2 (1), 12e19. Xie, S., Yang, J.T., Zhang, Z.X., Zhao, C.X., Bi, Y.C., Zhao, Q.L., et al., 2017. The effects of the X chromosome on intrinsic functional connectivity in the human brain: evidence from turner syndrome patients. Cerebr. Cortex 27 (1), 474e484. Yang, S.C., Lin, S.J., Tsai, C.Y., 2015. Effect of sex, age, and BMI on the development of locomotor skills and object control skills among preschool children. Percept. Mot. Skills 121 (3), 873e888. Young, K.D., Bellgowan, P.S.F., Bodurka, J., Drevets, W.C., 2013. Functional neuroimaging of sex differences in autobiographical memory recall. Hum. Brain Mapp. 34 (12), 3320e3332. Yu, Q.B., Tang, Y.Y., Li, J., Lu, Q.L., Wang, H.L., Sui, D., et al., 2009. Sex differences of event-related potential effects during three-dimensional mental rotation. Neuroreport 20 (1), 43e47. Yu, V.Y., MacDonald, M.J., Oh, A., Hua, G.N., De Nil, L.F., Pang, E.W., 2014. Age-related sex differences in language lateralization: a magnetoencephalography study in children. Dev. Psychol. 50 (9), 2276e2284. Zahn-Waxler, C., 1993. Warriors and worriers: gender and psychopathology. Dev. Psychopathol. 5 (1e2), 79e89. Zahn-Waxler, C., Shirtcliff, E.A., Marceau, K., 2008. Disorders of childhood and adolescence: gender and psychopathology. Annu. Rev. Clin. Psychol. 4, 275e303. Zhang, K., Sejnowski, T.J., 2000. A universal scaling law between gray matter and white matter of cerebral cortex. Proc. Natl. Acad. Sci. U.S.A. 97 (10), 5621e5626. Zhang, T.H., Casanova, R., Resnick, S.M., Manson, J.E., Baker, L.D., Padual, C.B., et al., 2016. Effects of hormone therapy on brain volumes changes of postmenopausal women revealed by optimally-discriminative voxel-based morphometry. PLoS One 11 (3), 16. Zhao, C.X., Gong, G.L., 2017. Mapping the effect of the X chromosome on the human brain: neuroimaging evidence from turner syndrome. Neurosci. Biobehav. Rev. 80, 263e275. Zhou, J.N., Hofman, M.A., Gooren, L.J.G., Swaab, D.F., 1995. A sex difference in the human brain and its relation to transsexuality. Nature 378 (6552), 68e70. Zilles, K., Armstrong, E., Schleicher, A., Kretschmann, H.J., 1988. The human pattern of gyrification in the cerebral cortex. Anat. Embryol. 179 (2), 173e179. Zubiaurre-Elorza, L., Junque, C., Gomez-Gil, E., Guillamon, A., 2014. Effects of cross-sex hormone treatment on cortical thickness in transsexual individuals. J. Sex. Med. 11 (5), 1248e1261. Zucker, K.J., 2017. Epidemiology of gender dysphoria and transgender identity. Sex. Health 14 (5), 404e411. Zucker, K.J., Bradley, S.J., Sanikhani, M., 1997. Sex differences in referral rates of children with gender identity disorder: some hypotheses. J. Abnorm. Child Psychol. 25 (3), 217e227. Zuo, X.N., Ehmke, R., Mennes, M., Imperati, D., Castellanos, F.X., Sporns, O., Milham, M.P., 2012. Network centrality in the human functional connectome. Cerebr. Cortex 22 (8), 1862e1875. Zuo, X.N., Kelly, C., Di Martino, A., Mennes, M., Margulies, D.S., Bangaru, S., et al., 2010. Growing together and growing apart: regional and sex differences in the lifespan developmental trajectories of functional homotopy. J. Neurosci. 30 (45), 15034e15043.

Index Note: ‘Page numbers followed by “f ” indicate figures and “b” indicate boxes’.

A ABRs. See Auditory brain stem responses (ABRs) ACC. See Anterior cingulate cortex (ACC) Accessory olfactory bulb, 4 Acetylcholine (ACh), 34 3-Acetylpyridine (3AP), 249 ACTH hormone. See Adrenocorticotropic hormone (ACTH hormone) Activity activity-dependent developmental refinement of corticospinal connectivity, 185e186 remodeling of inhibitory synapses, 95 columns, 108e109 interests, 591e592 ACtx, 46 ADCY1 gene. See Adenylate cyclase 1 gene (ADCY1 gene) Adenosine triphosphate (ATP), 33e34 Adenylate cyclase 1 gene (ADCY1 gene), 153 ADHD. See Attention-deficit hyperactivity disorder (ADHD) Adolescent face processing in, 447e450 handling children and adolescents process faces, 447e448 handling children and adolescents process facial expressions, 448 neural substrates, 448e449 hormone influences on human behavior, 598 Adrenocorticotropic hormone (ACTH hormone), 562 Adult circuitry organization Cajal and cerebellar circuit, 243e245 of cerebellar cortex, 243e245 Adult neurogenesis, 17e20 adult-born interneurons, 18e20 regeneration of sensory input, 17e18 Adult-born interneurons, 18e20 Adults face processing in, 436e439 handling, 436 models, 436e438 neural substrates of face processing, 438e439 plasticity, 159 Advanced diffusion MRI techniques, 303e304 Affective neuroscience, 485

Afferent cholinergic inputs from basal forebrain, 188 Afferent connectivity of motor cortex, 178e180 basal forebrain afferent projections to motor cortex, 179e180 intracortical motor cortex connections, 178e179 Afferent peripheral circuits, 34e38 Aggression, 592 AGl. See Lateral agranular cortex (AGl) AGm. See Medial agranular cortex (AGm) Agranular cortex, 170 AIS. See Axon initial segment (AIS) Alerting function in attention, 507 ALM cortex. See Anterior lateral motor cortex (ALM cortex) American Sign Language (ASL), 476 AMPA receptors (AMPARs), 133 Animal models, 561 Ankyrins, 93 Anterior cingulate cortex (ACC), 487e488, 509 Anterior lateral motor cortex (ALM cortex), 225e226 Anterioreventral cochlear nucleus (AVCN), 36 Anti-Hebbian STDP, 129e130 properties, 131e132 3AP. See 3-Acetylpyridine (3AP) Apical dendritic bundles, 115e117 Apolipoprotein E receptor 2 (Apoer2), 246 Arginine vasopressin (AVP), 562 Arkypallidal GPe neurons, 233e234 Ascending somatosensory pathways, 154e156 ASD. See Autism spectrum disorder (ASD) ASL. See American Sign Language (ASL) Atonal homolog 1 (Atoh1), 81, 245 ATP. See Adenosine triphosphate (ATP) Attention, 505e507 brain networks, 507e509 development of brain and behavior, 509e513 childhood, 511e513 infancy, 509e510 toddlerhood, 510e511 individual differences, 513e516 environment, 515e516 genes, 514e515 temperament, 513e514

plasticity of attention networks, 516e517 and self-regulation, 506e507 Attention-deficit hyperactivity disorder (ADHD), 539, 568 Atypical development of face processing, 450e453 Audiotopy, 59e60 Audiovisual integration in early speech perception, 416 Auditory brain stem responses (ABRs), 28 Auditory midbrain and forebrain circuits, 44e52 afferent regulation of higher auditory circuit development, 47e48 developmental regulation over reinstating hearing in deaf, 51 experience-dependent influences on circuit development, 48e50 local cortical circuits, 46e47 thalamocortical subplate circuitry, 45e46 Auditory system, 27e31 auditory circuitry, 29e30 cochlear transduction, 30e31 functional circuit development auditory midbrain and forebrain circuits, 44e52 auditory system, 27e31 brain stem circuits, 39e44 peripheral circuits, 31e39 neurobiological approach to studying, 27e29 Autism spectrum disorder (ASD), 424, 450e453, 498e499, 539 Autobiographical memory, 401e402 AVCN. See Anterioreventral cochlear nucleus (AVCN) AVP. See Arginine vasopressin (AVP) Axon guidance of subcerebral, including corticospinal projections, 182e184 Axon initial segment (AIS), 80

B Backpropagation process, 134 Barrel cortex, critical periods for inhibition in, 157e159 Barrels formation, 153e154 Basal forebrain afferent projections to motor cortex, 179e180 Basal ganglia, 221, 235e236

639

640 Index

Basal ganglia (Continued ) direct and indirect striatal output pathways, 232e233 effect of direct and indirect striatal output pathways on behavior, 232 general organization, 221e222 GPe, 233e234 organization of corticostriatal projections, 222e228 STN, 234e235 striatum, 228e232 Basic helix-loop-helix domain-containing, class B5 (Bhlhb5), 182 Basilar membrane (BM), 30 Basket cells, 257e259 basket cell-PC synapse formation, 93e94 BCM learning rule. See BienenstockCooper-Munro learning rule (BCM learning rule) BDA. See Biotinylated dextran amine (BDA) BDNF. See Brain-derived neurotrophic factor (BDNF) Bed nucleus of stria terminalis (BNST), 563e564 Betz cells, 168 Bhlhb5. See Basic helix-loop-helix domain-containing, class B5 (Bhlhb5) Bienenstock-Cooper-Munro learning rule (BCM learning rule), 127e128 Biotinylated dextran amine (BDA), 84 “Black reaction” method, 255 Blood oxygen leveledependent response (BOLD), 269, 600 BM. See Basilar membrane (BM) BMP4. See Bone morphogenetic protein 4 (BMP4) BNST. See Bed nucleus of stria terminalis (BNST) BOLD. See Blood oxygen leveledependent response (BOLD) Bone morphogenetic protein 4 (BMP4), 251e253 Brain networks, 507e509 alerting, 507 executive attention, 508e509 orienting, 507e508 Brain sex differences, 600e618. See also Psychological sex differences genetic perspectives, 614 hormone perspectives, 615e618 implications, 611e613 issues in studying brain, 600e601 sex differences in brain structure and development, 601e606 socialization perspectives, 613 Brain stem circuits, 39e44 afferent regulation of cochlear nucleus development, 41e42 of third-order brain stem nuclei, 42e43 fine-scale connectivity to lateral superior olive, 40e41 in medial superior olive, 39e40 functional circuit assembly in, 39

influence of source and pattern of afferent activity, 43e44 Brain structure, implications of sex differences in, 605e606 Brain volume, 601e602 Brain-derived neurotrophic factor (BDNF), 83 Brainebehavior relationships, 305e309

C C-ANT. See Child Attention Network Task (C-ANT) CA1 pyramidal cells, postnatal development of, 213 CA3 pyramidal cells, postnatal development of, 212e213 Cadherin-7 (Cad-7), 253 CAH. See Congenital adrenal hyperplasia (CAH) CAIS. See Complete androgen insensitivity syndrome (CAIS) Cajal, 243e245 Calmodulin-dependent protein kinase II (CaMKIIa), 88 Calmodulin-dependent protein kinase IV (CaMKIV), 88 Calretinin (CR), 209, 213 Calyx of Held, 42 CaMKIIa. See Calmodulin-dependent protein kinase II (CaMKIIa) CaMKIV. See Calmodulin-dependent protein kinase IV (CaMKIV) Candelabrum cells, 256e257 Cannabinoid type 1 receptor (CB1R), 132e133 CAR. See Cortisol awakening response (CAR) Caregivers, 573e574 Caudal forelimb area (CFA), 168e169 Caudal ganglionic eminence (CGE), 209 Causal modeling, 282 CB1R. See Cannabinoid type 1 receptor (CB1R) CBG. See Cortisol-binding globulin (CBG) Cbln1. See Cerebellin-1 (Cbln1) CC. See Corpus callosum (CC) CC projecting neurons. See Corticocortically projecting neurons (CC projecting neurons) CCK. See Cholecystokinin (CCK) CDP. See Correlation-dependent plasticity (CDP) Cerebellar interneurons in cerebellar circuit, 253e259 excitatory interneurons, 253e255 inhibitory interneurons, 255e259 Cerebellar locomotor region (CLR), 172e174 Cerebellin-1 (Cbln1), 90e91, 253 Cerebellum afferent fibers, 79e81 cell types, 79e81 CFePC synapses, 82e89 compartmentalization, 81e82

generation of neurons, 81 inhibitory synapses from basket cells and stellate cells to PCs, 93e95 microcircuit in cerebellar cortex, 79e82 PFePC synapses, 89e92 Cerebral cortex, 221 CFA. See Caudal forelimb area (CFA) CFs. See Characteristic frequencies (CFs); Climbing fibers (CFs) CGE. See Caudal ganglionic eminence (CGE) Characteristic frequencies (CFs), 29e30 Chemosensation, 3e4 Child Attention Network Task (C-ANT), 512 Childhood, 511e513 Childhood, ToM development in, 470e474 CHL1. See Close homolog of L1 (CHL1) Cholecystokinin (CCK), 209 Circuit plasticity, 20e21 Climbing fibers (CFs), 79e80, 244e245 CFePC synapses, 82e89 dendritic translocation of single CFs, 83e85 early phase of CF synapse elimination, 85e86 functional differentiation of multiple CFs, 82e83 late phase of CF synapse elimination, 86e89 multiple innervation of PCs by CFs in early postnatal period, 82 development and refinement of climbing fiber projections, 247e249 Close homolog of L1 (CHL1), 94 CLR. See Cerebellar locomotor region (CLR) CM connections. See Cortico-motoneuronal connections (CM connections) CMM. See Congenital mirror movement disorder (CMM) CNC. See Cochlear nucleus complex (CNC) CNV. See Contingent-negative variation (CNV) CO. See Cytochrome oxidase (CO) Cochlear hair cells, development and maturation of, 31e34 Cochlear nucleus complex (CNC), 29e30, 34e36 Cochlear nucleus development, afferent regulation of, 41e42 Cochlear transduction, 30e31 Cognitive control development and neural basis, 523e528 control instantiation, 526e528 error monitoring, 524e526 individual differences in, 531e533 role in decision-making, motivation, and social behavior, 528e530 Cognitive development conceptual frameworks, 267e268 electrophysiology, 268e269 eye-tracking, 268 frameworks and methods, 267e270 MRI and other imaging methods, 269

Index

prenatal maternal biological stress signals and infant and child development, 570 psychosocial stress and infant and child development, 569 Cognitive skills, 587e590 mathematical skills, 588e589 memory, 589e590 perceptual speed, 590 spatial skills, 588 verbal skills, 589 Complete androgen insensitivity syndrome (CAIS), 595 Computed tomography (CT), 289 Conceptual frameworks, 267e268 Conduct problems (CPs), 496e498 Configural or synthetic perception, 3 Congenital adrenal hyperplasia (CAH), 596 Congenital cataract, 451 Congenital mirror movement disorder (CMM), 186e187 Contagious crying, 488e489 Contextual appraisal, 493 Contingent-negative variation (CNV), 507 Contrast enhancement, 11e12 Corpus callosum (CC), 603 Correlation-dependent plasticity (CDP), 127e128 Cortical columns, 103e107, 104t, 112 barrels, 113e117 function, 115 microcolumns and apical dendritic bundles, 115e117 system of interleaving modules in rodent layer VI, 114e115 columnar organization of afferent and efferent projections, 108e110 columnar structures in cortex, 105te106t complex relationship relations between minicolumns and dendritic bundles, 117e118 columns in nonmammals, 118 columns outside mammalian isocortex, 117e118 during development, 119e121 function, 118 gene expression in cortex in “columnar” fashion, 110e111 interleaving module system in rodent layer II, 110e111 overlap between columnar entities, 111 microscopic and macroscopic cell patterning defining cortical modules, 112 Montcastle’s definition, 107e108 neurons in, 112 in neuropathology, 118e119 physiological methods to revealing columns, 108 transient columnar domains during development, 121e122 Cortical maturation, 343e345 Cortical stimulation, 168e169 Cortico-motoneuronal connections (CM connections), 175e177, 184

Cortico-pontine axons, 171e172 Cortico-subcerebral circuits in control of facial movement, 174 of fine/skilled motor function, 172 for general locomotor and posture control, 172e174 in motor planning, 174 Corticobulbar axons, 171e172 Corticocortical inputs, 149e151 Corticocortically projecting neurons (CC projecting neurons), 114 Corticospinal axons, 171e172 Corticospinal connectivity in spinal cord, 184e185 Corticospinal projection neurons (CSN), 175 Corticospinal tract (CST), 175 Corticostriatal projection neurons (CStrPN), 177 Corticostriatal projections, organization of, 222e228 Corticothalamic inputs, 149e151 Corticothalamic projecting neurons (CT projecting neurons), 114 Corticotropin-releasing hormone (CRH), 562 Cortisol awakening response (CAR), 572 Cortisol-binding globulin (CBG), 566e567, 572 Couptf-interacting protein 1 (Ctip1), 182 Couptf-interacting protein 2 (Ctip2), 181 CPs. See Conduct problems (CPs) CR. See Calretinin (CR) Credit assignment problem, 132 CRH. See Corticotropin-releasing hormone (CRH) CRHR1 gene, 576 Critical or sensitive periods, 281 Critical periods, 342e343 Cross-cultural differences in cognitive control development, 532e533 Cross-modal stimuli, 69 Cross-situational learning, 326e327 CSN. See Corticospinal projection neurons (CSN) CST. See Corticospinal tract (CST) CStrPN. See Corticostriatal projection neurons (CStrPN) CT. See Computed tomography (CT) CT projecting neurons. See Corticothalamic projecting neurons (CT projecting neurons) Ctip1. See Couptf-interacting protein 1 (Ctip1) Ctip2. See Couptf-interacting protein 2 (Ctip2) Culture, 477e478 Cytochrome oxidase (CO), 110

D dACC. See Dorsal anteerior cingulate cortex (dACC) DAN. See Dorsal attention network (DAN) Dcc receptor, 186e187

641

DCN. See Deep cerebellar nuclei (DCN); Developmental cognitive neuroscience (DCN); Dorsal cochlear nucleus (DCN) DCX. See Doublecortin (DCX) Decision-making, cognitive control role in, 528e530 Declarative memory, 396e397 development of neural substrate supporting, 403e404 developmental changes in, 398e402 autobiographical memory, 401e402 episodic memory, 399e401 neural substrate of, 402e403 Deep cerebellar nuclei (DCN), 79 Deep layer sensory topographies, 66e67 Deep short-axon cell (dSAC), 15 Default mode network (DMN), 610 Dehydroepiandrosterone (DHEA), 562 Dehydroepiandrosterone sulfate (DHEAS), 567 Dendritic bundles, 118 excitability, 134 inhibition, 134 translocation of single CFs, 83e85 Dendrogenesis, 210 Dentate gyrus (DG), 202e203 postnatal development, 212 Deterministic epigenesis, 277 Developmental affective neuroscience, 485 Developmental cognitive neuroscience (DCN), 273, 468 assumptions underlying frameworks, 277e278 critical or sensitive periods, 281 frameworks for human functional brain development, 275e276 functional brain imaging, 279e281 from genetics to behavior in, 282 predictions and evidence, 278e279 theories, 274e275 Developmental language disorders, 423e424 Developmental science, 268e269 DG. See Dentate gyrus (DG) DHEA. See Dehydroepiandrosterone (DHEA) DHEAS. See Dehydroepiandrosterone sulfate (DHEAS) Diffusion kurtosis imaging (DKI), 303e304 Diffusion magnetic resonance imaging, 300e304 advanced diffusion MRI techniques, 303e304 diffusion parameters in development, 300e302 diffusion tensor imaging theory, 300 fiber tractography, 302e303 sex differences, 303 Diffusion tensor imaging (DTI), 474, 600 theory, 300 Direct striatal output pathways, 232e233 Disorder/difference of sex development (DSD), 596

642 Index

Diurnal cortisol, 574 DKI. See Diffusion kurtosis imaging (DKI) DLPFC. See Dorsolateral prefrontal cortex (DLPFC) DMN. See Default mode network (DMN) Dopamine, 226e228 Dorsal anteerior cingulate cortex (dACC), 524 Dorsal attention network (DAN), 369, 372 Dorsal cochlear nucleus (DCN), 36 Dorsal medial striatum, 223 Dorsal stream processes, 366e371 mental rotation, 369e371 neurodevelopmental disorders of visuospatial processing, 372e379 spatial attention, 368e369 spatial localization, 366e367 trajectories of dorsal and ventral stream development, 371e372 Dorsolateral prefrontal cortex (DLPFC), 361, 495e496 Dorsolateral striatum, 223 Doublecortin (DCX), 209 Downstream decoding, 14 dSAC. See Deep short-axon cell (dSAC) DSD. See Disorder/difference of sex development (DSD) DTI. See Diffusion tensor imaging (DTI) Dynamic mapping, 277e278

E E/I ratio. See Excitation/inhibition ratio (E/I ratio) EAE. See Enhanced acoustic environments (EAE) Early left anterior negativity (ELAN), 423 EC. See Entorhinal cortex (EC) eCB. See Endocannabinoid (eCB) EEG. See Electroencephalography (EEG) EF. See Executive function (EF) Efferent connectivity of motor cortex, 171e178 anatomical organization of subcerebral projections, 171e172 functional organization of subcerebral projections, 172e174 organization of motor cortex-spinal cord connectivity, 175 of motor cortex-striatum connectivity, 177 of reciprocal connectivity between motor cortex and thalamus, 177e178 species-specific differences of corticospinal connectivity, 175e177 Efferent peripheral circuits, 34e38 Efna. See Ephrin A (Efna) EGFP. See Enhanced green fluorescent protein (EGFP) ELAN. See Early left anterior negativity (ELAN) Electroencephalogram, 337 Electroencephalography (EEG), 268, 509, 524e525

Electrophysiological method in face processing, 446b Electrophysiology, 268e269 Emotion emotion-related processing, 609e610 recognition, 489e490 regulation, 491e493 Emotional development, 569 Empathy, 485e486 clearing up definitional issues, 486e488 development, 488e493 affect sharing and physiological synchrony, 488e489 emotion recognition, 489e490 emotion regulation, 491e493 emotion understanding, 490 motivation to care, 493 perspective-taking and theory of mind, 490e491 maladaptive alterations in developmental trajectories of empathy, 496e499 neurodevelopmental changes in empathic responding, 493e496 evidence from event-related potential, 494 evidence from functional magnetic resonance imaging, 494e496 Endocannabinoid (eCB), 133 Endolymph, 33 Enhanced acoustic environments (EAE), 49 Enhanced green fluorescent protein (EGFP), 120 Entopeduncular nucleus (EP nucleus), 221e222 Entorhinal cortex (EC), 203, 209 layer II stellate cells, 212 EP nucleus. See Entopeduncular nucleus (EP nucleus) Eph receptor A (EphA), 119e120 Ephrin A (Efna), 119e120 Ephrin B3, 184 Epigenetics, 570e571 Episodic memory, 399e401 EPL. See External plexiform layer (EPL) EPSCs. See Excitatory postsynaptic currents (EPSCs) EPSP. See Excitatory postsynaptic potential (EPSP) ERN. See Error-related negativity (ERN) ERPs. See Event-related potentials (ERPs) Error monitoring, 524e526 Error-related negativity (ERN), 524e525 ETCs. See External tufted cells (ETCs) Event-related potentials (ERPs), 268, 416, 439, 494, 507, 524e525 Excitation/inhibition ratio (E/I ratio), 157 Excitatory interneurons, 253e255 granule cells, 253e255 UBCs, 255 Excitatory postsynaptic currents (EPSCs), 82 Excitatory postsynaptic potential (EPSP), 130e131 Executive attention, 508e509 Executive function (EF), 539

from biological to environmental predictors of individual differences, 545e551 clinical insights, from infancy to adolescence, 543e545 early EF predicts academic, socio-cognitive and social success at school, 548e551 normative developmental trajectories for EF from infancy to adolescence, 540e543 Explicit memory, 396e397 External plexiform layer (EPL), 7e8 microcircuits, 9e11 External segment of globus pallidus (GPe), 221e222, 233e234 External tufted cells (ETCs), 8 Eye-tracking, 268

F FA. See Fractional anisotropy (FA) Face perception, 347, 348f Face processing, 435e436 in adults, 436e439 electrophysiological and neuroimaging methods, 446b in first year of life, 444 handling infants process facial expressions, 442 infants learning to see faces, 439e441 neural signatures, 443e444 neural substrates, 442e443 impairments and atypical development of, 450e453 Autism spectrum disorder and Williams syndrome, 451e453 congenital cataract, 451 prosopagnosia, 450e451 in school-age children and adolescents, 447e450 in toddlers and preschoolers, 444e447 Fast timescale synchronization, 13e14, 13f Fast-spiking neurons (FS neurons), 147 FC. See Frontal cortex (FC) fCOI approach. See Functional channel of interest approach (fCOI approach) FEF. See Frontal eye fields (FEF) Fetal adrenal development, 567 Fetal brain development, 567 Fetal programming, 565e566 Fezf2 (zinc finger transcription factor), 180e181 FFA. See Fusiform face area (FFA) Fiber tractography, 302e303 Fine-tuning preferential responses, 471e472 fMRI. See Functional magnetic resonance imaging (fMRI) fNIRS. See Functional near infra-red spectroscopy (fNIRS) Fog2, 187 Fractional anisotropy (FA), 605 Fragile X syndrome, 378 Frontal cortex (FC), 215 GPe neurons, 234 Frontal eye fields (FEF), 63, 369 FS neurons. See Fast-spiking neurons (FS neurons)

Index

Functional asymmetry. See Lateralization Functional brain imaging, 279e281 Functional channel of interest approach (fCOI approach), 469e470 Functional circuit development auditory midbrain and forebrain circuits, 44e52 auditory system, 27e31 brain stem circuits, 39e44 peripheral circuits, 31e39 Functional connectivity, 610 Functional magnetic resonance imaging (fMRI), 269, 405, 467e468, 489, 524e525, 600 Functional maturity, 472e473 Functional near infra-red spectroscopy (fNIRS), 269, 469 Fusiform face area (FFA), 280, 364, 438

G G-protein-coupled receptors (GPCRs), 3e4 GABAA receptor-mediated signaling, 95 GABAergic interneurons, 208e209, 243e244 Gbx2, 245 GCL. See Granule cell layer (GCL) GCs. See Granule cells (GCs) GDP. See Giant Depolarizing Potentials (GDP) Gender identity, 590e591 Genes, 514e515 Genetic perspectives, 595, 614 Gestalt theory, 337 Gestational stress influences human fetus, 567e568 GFAP. See Glial fibrillary acidic protein (GFAP) Giant Depolarizing Potentials (GDP), 212 GL. See Glomerular layer (GL) GLAST (glutamate transporter), 92 Glial fibrillary acidic protein (GFAP), 94 Glomerular layer (GL), 6 microcircuits, 8e9 Glomeruli, 4e5 Glucocorticoid receptors (GRs), 562 Glucocorticoid response elements (GREs), 562e563 GluD2 protein, 89, 92 Golgi cells, 80e81, 255e256 GPCRs. See G-protein-coupled receptors (GPCRs) GPe. See External segment of globus pallidus (GPe) GPi. See Internal segment of globus pallidus (GPi) Granule cell layer (GCL), 9 interneurons, 255e256 Golgi cells, 255e256 Lugaro cells, 256 Granule cells (GCs), 9, 79e80, 253e255 Graph theory, 283 Gray matter, 604e605

GREs. See Glucocorticoid response elements (GREs) GRs. See Glucocorticoid receptors (GRs)

H HARDI. See High angular resolution diffusion imaging (HARDI) HDB. See Horizontal limb of diagonal band of Broca (HDB) Hebbian learning, 122 Hebbian STDP, 128e129 properties, 131 Hemispheric specialization. See Lateralization Heterogeneity of Purkinje cells, 245e246 HF. See Hippocampal formation (HF) High angular resolution diffusion imaging (HARDI), 303e304 High-density arrays, 268e269 Hippocampal circuits, 201e206. See also Peripheral circuits adult organization, 202e206 subfield features and numbers, 203e204 zonal specializations, 204e206 cell autonomous organization, 209 coordinated network activity, 211e212 early stages, 208 EC, 209 functional backdrop, 206 developmental milestones in humans, 206 spatial navigation system in rodents, 206 maturational events, 210e211 neural activity, 209e210 neurogenesis, 208e209 postnatal development of electrophysiological patterns, 212e215 Hippocampal formation (HF), 202 Hippocampal rhythms, 214e215 Holistic processing, 436 Homo sapiens, 335 Horizontal limb of diagonal band of Broca (HDB), 15 Hormone perspectives, 595e599, 615e618 adolescent hormone influences on brain sex differences, 615e616 on human behavior, 598 circulating hormone influences on brain sex differences, 616e617 on human behavior, 599 early hormone influences on human behavior, 596e598 cognitive sex differences, 597 noncognitive sex differences, 597e598 evidence for hormone influences on nonhuman sex-typed behavior, 595e596 exogenous hormone influences on brain sex differences, 617e618 on human behavior, 599 prenatal hormone influences on brain sex differences, 615 HPA axis. See Hypothalamicepituitary eadrenocortical axis (HPA axis) 5-HT3A receptor (5-HT3AR), 148

643

Htr3a, 231e232 Human brain development, 289 Human corticospinal circuitry, molecular development of, 186e187 Human functional brain development. See also Structural brain development frameworks for, 275e276 interactive specialization, 276 maturational viewpoint, 276 skill learning, 276 11-beta Hydroxysteroid dehydrogenase (11b-HSD), 562 Hypothalamicepituitaryeadrenocortical axis (HPA axis), 562, 563f, 566e567 functioning, 569e570

I IC. See Inferior colliculus (IC) ICA. See Independent component analysis (ICA) IDED shift. See Intradimensional/ extradimensional shift (IDED shift) Idiopathic hypogonadotropic hypogonadism (IHH), 597 IDT. See Infant-directed talk (IDT) IFC. See Inferior frontal cortex (IFC) IFG. See Inferior frontal gyrus (IFG) IGF-1. See Insulin like growth factor I (IGF1) IGT. See Iowa gambling task (IGT) IHCs. See Inner hair cells (IHCs) IHH. See Idiopathic hypogonadotropic hypogonadism (IHH) Imaging methods, 269 Imitation-based technique, 399e400 Impairments of face processing, 450e453 Implicit memory, 397e398 Independent component analysis (ICA), 470 Indirect aggression, 592 Indirect striatal output pathways, 232e233 Individual differences in attention, 513e516 environment, 515e516 genes, 514e515 temperament, 513e514 in cognitive control cross-cultural differences in cognitive control development, 532e533 temperament, cognitive control, and psychopathology, 531e532 Infant-directed talk (IDT), 418 Infants/infancy, 509e510 learning words from co-occurrence statistics, 322e323 speech perception development in, 415e416 speech production in, 416e417 Inferior colliculi, 57 Inferior colliculus (IC), 29e30 Inferior frontal cortex (IFC), 527 Inferior frontal gyrus (IFG), 495e496 Infragranular excitatory neurons, 147e148 Inhibition of return (IOR), 368 Inhibitory control development, 527 Inhibitory interneurons, 255e259

644 Index

Inhibitory interneurons (Continued ) granule cell layer interneurons, 255e256 molecular layer interneurons, 257e259 Purkinje cell layer interneurons, 256e257 Inhibitory neuron connectivity, 148e149 Inhibitory synapses from basket cells and stellate cells to PCs, 93e95 activity-dependent remodeling of inhibitory synapses, 95 basket cell-PC synapse formation, 93e94 stellate Cell-PC synapse formation, 94e95 Inner hair cells (IHCs), 30e31 Insulin like growth factor I (IGF-1), 83 Integrated perspectives, 599e600 Integrin B1 (ItgB1), 87e88 Intentional communication, 418e419 Interactive specialization (IS), 276, 282 future challenges, 283e285 Interaural time differences (ITDs), 40 Interhemispheric commissures, sex differences in, 603 Internal plexiform layer (IPL), 8 Internal segment of globus pallidus (GPi), 221e222 Interneurons, 209 postnatal development of, 213 Intersubject correlation (ISC), 472 Intracortical connectivity, 187e188 excitatory plasticity, 156e157 motor cortex connections, 178e179 and subcortical afferent connectivity, 187e188 Intradimensional/extradimensional shift (IDED shift), 542 Intratelencephalic type (IT type), 147 IOR. See Inhibition of return (IOR) Iowa gambling task (IGT), 529 IP3 receptor, 134 IPL. See Internal plexiform layer (IPL) IS. See Interactive specialization (IS) ISC. See Intersubject correlation (ISC) IT type. See Intratelencephalic type (IT type) ITDs. See Interaural time differences (ITDs) ItgB1. See Integrin B1 (ItgB1)

K Kal1 gene, 187 KCC2 (K+-dependent intracellular chloride transporter), 46e47 Kenyon cells (KCs), 135e136 Kinocilium, 33 Kolliker’s organ (Ko), 33e34

L L1 cell adhesion molecule (L1CAM), 94 L4 excitatory neurons, 147 Language, 418, 475e477, 608e609 disorders developmental language disorders, 423e424 lexical development, 425 neural foundations of, 426

socialecognitive foundations of language, 425 speech perception, 424 speech production, 424e425 syntactic development, 426 impairment, 424 learning, 589 nature of, 414 prenatal perception of, 414 Lateral agranular cortex (AGl), 170 Lateral geniculate nucleus (LGN), 64, 342 Lateral olivocochlear neurons (LOC neurons), 34e36 Lateral superior olive (LSO), 36 fine-scale connectivity in, 40e41 Lateralization, 606e607 LC. See Locus coeruleus (LC) Learning. See also Memory co-occurrence statistics, 320e327 infants learning words from co-occurrence statistics, 322e323 relations between words and referents, 326e327 in speech, 321e322, 324e325 in visual domain, 323e324 probability distributions, 319e320 sex differences in structures involved in, 603 Lemniscal pathway, 152 Lexical development, 419e421, 425 developmental processes, 420e421 neural bases of word learning, 421 stages of, 419e420 LFP. See Local field potential (LFP) LGN. See Lateral geniculate nucleus (LGN) LIM domain only 4 (Lmo4), 182 b-Lobe neurons (b-LNs), 135e136 LOC neurons. See Lateral olivocochlear neurons (LOC neurons) Local cortical circuits, 46e47 Local field potential (LFP), 13 Localization, 279 Locus coeruleus (LC), 507 Long-term depression (LTD), 127e128 mechanisms, 132e134 Long-term memory, 396 Long-term potentiation (LTP), 127e128 mechanisms, 132e134 Low-threshold spiking cells (LTS cells), 148 LSO. See Lateral superior olive (LSO) LTD. See Long-term depression (LTD) LTP. See Long-term potentiation (LTP) LTS cells. See Low-threshold spiking cells (LTS cells) “Lugaro” cell, 255e256

M M2 muscarinic acetylcholine receptors, 110e111 Magnetic resonance imaging (MRI), 189, 269, 289, 600. See also Neuroimaging brain mapping approaches, 294e300 cortical thickness, 296e297

sex differences, 298e300 voxel-based strategies, 294e296 white matter, 297e298 volume analyses, 292e294 gray matter decreases in development, 292 regional and temporal dynamics, 292e293 sex differences, 293e294 white matter increases in development, 293 Magnetoencephalography (MEG), 269, 439 Main olfactory bulb (MOB), 4 Maladaptive alterations in developmental trajectories of empathy, 496e499 ASD, 498e499 conduct problems, 496e498 Mammalian visual system, 337e338 MAP2. See Microtubule-associated protein 2 (MAP2) Matching hypothesis, 249 Mathematical skills, 588e589 Maturational events, 210e211 MCL. See Mitral cell layer (MCL) Mean length of utterance (MLU), 546e547 Medial agranular cortex (AGm), 170 Medial ganglionic eminence (MGE), 209 Medial geniculate nucleus (MG), 29e30 Medial olivocochlear neurons (MOC neurons), 34e36 Medial prefrontal cortex (mPFC), 467, 487e488, 491, 573 Medial superior olive, fine-scale connectivity in, 39e40 Medialefrontal cortex (MFC), 524 MEG. See Magnetoencephalography (MEG) Membrane-spanning, 4-pass A odorantbinding receptors (MS4ARs), 6 Memory, 395, 589e590 development, 395 developmental changes in declarative memory, 398e402 different forms, 395e398 declarative and nondeclarative memory, 396 declarative or explicit memory, 396e397 nondeclarative, procedural, or implicit memory, 397e398 relations between different forms of memory, 398 short-and long-term memory, 396 mechanisms of developmental change, 402e408 basic cognitive and mnemonic processes, 405 consolidation and storage, 405e406 development of neural substrate supporting declarative memory, 403e404 encoding, 405 functional consequences of temporalcortical network development, 404e405 neural structures and processes, 402

Index

neural substrate of declarative memory, 402e403 retrieval, 406e408 sex differences in structures involved in, 603 Mental rotation, 350, 369e371 Mesencephalic locomotor region (MLR), 172e174, 177e178 Metabotropic glutamate receptor (mGluR), 132e133 mGluR1, 86e87 MFC. See Medialefrontal cortex (MFC) MFs. See Mossy fibers (MFs) MG. See Medial geniculate nucleus (MG) MGE. See Medial ganglionic eminence (MGE) mGluR. See Metabotropic glutamate receptor (mGluR) Microcolumns, 115e117 Microtubule-associated protein 2 (MAP2), 115 Microzones, 81 Mineralocorticoid receptors (MRs), 562 Minicolumn, 107e108 Mitral cell layer (MCL), 8 MLR. See Mesencephalic locomotor region (MLR) MLU. See Mean length of utterance (MLU) MOB. See Main olfactory bulb (MOB) MOC neurons. See Medial olivocochlear neurons (MOC neurons) Molecular layer interneurons, 257e259 Mossy fibers (MFs), 79e80, 244e245. See also Parallel fibers (PFs) development and refinement of mossy fiber projections, 250e253 Motivation, cognitive control role in, 528e530 Motor development, 72 map alignment, 61 skills, 591 Motor cortex, 167e168. See also Somatosensory cortex connectivity, 170e188 discovery, 168 evolution, 168 organization of motor cortex-spinal cord connectivity, 175 of motor cortex-striatum connectivity, 177 technological advances, 188e189 topographic organization, 168e170 Mototopic representation, 61 Movement disorders, 221 mPFC. See Medial prefrontal cortex (mPFC) MRI. See Magnetic resonance imaging (MRI) MRs. See Mineralocorticoid receptors (MRs) MS4ARs. See Membrane-spanning, 4-pass A odorant-binding receptors (MS4ARs) Multifactor STDP rule, 131 Multimodal imaging, 304e305 Multisensory integration, 60, 68

impact of sensory experience on maturation of, 68e73 Multisensory neurons, 67, 70f Multivariate pattern analyses, 475 Myelination, 291 Myosin Va, 88

N N-methyl-D-aspartate receptor (NMDA receptor), 87, 132e133, 152e153 NMDAR-dependent LTD, 132e133 Nativism, 273e274 NE. See Norepinephrine (NE) Neonate(s), 63 speech perception in, 415 Netrin-1, 183e184 Neural bases of word learning, 421 Neural computation, 11e14 contrast enhancement, 11e12 downstream decoding, 14 fast timescale synchronization, 13e14 slow timescale decorrelation, 12e13 Neural signatures of face processing in infants, 443e444 Neurite orientation dispersion and density imaging (NODDI), 303e304 Neurobehavioral development, 565e572 postnatal stress and, 572e577 prenatal stress and, 565e572 Neurodevelopmental changes in empathic responding, 493e496 disorders of visuospatial processing, 372e379 neurogenetic syndromes, 376e379 perinatal stroke, 373e376 spina bifida, 376 processes, 567 Neurofascin 186 (NF186), 93 Neurogenesis, 208e209 Neurogenetic syndromes, 376e379 fragile X syndrome, 378 Turner syndrome, 378e379 Williams syndrome, 376e378 Neurogliaform cells (NGFCs), 148 Neuroglioform (NGF), 231e232 Neuroimaging methods, 290 in face processing, 446b studies of young children and infants, 468e470 Neuromodulation of STDP, 132 Neuromodulators, 131 Neuronal circuitry of cerebellar cortex, 243 adult circuitry organization of cerebellar cortex, 243e245 cerebellar interneurons in cerebellar circuit, 253e259 development and refinement of climbing fiber projections, 247e249 of mossy fiber projections, 250e253 heterogeneity of Purkinje cells, 245e246 modular organization of cerebellar cortex, 245

645

Neuropilin-1 (NRP1), 94, 184 Neuroscience, 268 Neurosymphony of stress, 572 Newborn, visual perception in, 339e340 NF186. See Neurofascin 186 (NF186) NGF. See Neuroglioform (NGF) NGFCs. See Neurogliaform cells (NGFCs) a9-Nicotinic acetylcholine receptor, 44 NMDA receptor. See N-methyl-D-aspartate receptor (NMDA receptor) NODDI. See Neurite orientation dispersion and density imaging (NODDI) Noncognitive sex differences, 590e593 activity interests, 591e592 gender identity, 590e591 physical and motor skills, 591 psychological disorders, 593 sexual orientation, 591 social behaviors, 592 temperament and personality, 592 Nondeclarative memory, 396e398 Norepinephrine (NE), 507 NRP1. See Neuropilin-1 (NRP1)

O Object perception, 346e347 OC. See Oral contraceptive (OC) OCB. See Olivocochlear bundle (OCB) Occipital face area (OFA), 364 Occipital gyrus (OG), 280 Ocular dominance columns/stripes (OD columns/stripes), 110 Oculomotor development, 343e345 OD columns/stripes. See Ocular dominance columns/stripes (OD columns/stripes) Odorant receptor (OR), 5 Odorants, 3 OFA. See Occipital face area (OFA) OFC. See Orbitofrontal cortex (OFC) OG. See Occipital gyrus (OG) OHCs. See Outer hair cells (OHCs) OKN. See Optokinetic nystagmus (OKN) Olfaction, 3 Olfactory bulb neural computation, 11e14 plasticity in, 17e21 sensory inputs, 4e6 sensory processing modulation, 14e17 synaptic microcircuits, 6e11 synaptic organization, 4e17 Olfactory nerve layer (ONL), 7e8 Olfactory sensory neurons (OSNs), 5e6 Olfactory systems, 3e4 Olivocochlear bundle (OCB), 36 ONL. See Olfactory nerve layer (ONL) Onset of hearing, 28 Ontogenic units/columns, 119e120 Optic tectum (OT), 57 Optokinetic nystagmus (OKN), 346 OR. See Odorant receptor (OR) Oral contraceptive (OC), 599 Orbitofrontal cortex (OFC), 487e488 Orientation columns, 110

646 Index

Orienting function in attention, 507e508 OSNs. See Olfactory sensory neurons (OSNs) OT. See Optic tectum (OT) Otx1, 181e182 Otx2, 245 Outer hair cells (OHCs), 30e31

P P/Q-type VDCC, 88 Pancreatic transcription factor 1a (Ptf1a), 81, 245 Paralemniscal pathway, 152 Parallel fibers (PFs), 79e80 PFePC synapses, 89e92 developmental elimination, 91 formation, 89 heterosynaptic competition between PF and CF inputs, 91e92 stabilization and maintenance, 89e91 Parallel pathways, 8 Parasympathetic nervous system (PNS), 565 Parvalbumin (PV), 209 PC. See Precuneus (PC) PCA. See Principal component analysis (PCA) pCRH. See Placental CRH (pCRH) PCs. See Purkinje cells (PCs) PDS. See Pubertal Development Scale (PDS) “Perceptual narrowing” process, 415 Perceptual speed, 590 “Performance monitoring”, 524 Periglomerular cells (PGCs), 8 Perilymph, 33 Perinatal stroke (PS), 373e376 Perineuronal nets (PNNs), 47, 210 Peripheral circuits, 31e39. See also Hippocampal circuits afferent and efferent peripheral circuits, 34e38 ear opening and maturation of response to sound, 37e38 initial wiring of cochlea, 36 synapse formation and refinement, 36e37 development and maturation of cochlear hair cells, 31e34 place code, 34 Personality, 592 PES. See Posterror slowing (PES) PET. See Positron emission tomography (PET) PFC. See Prefrontal cortex (PFC) PFs. See Parallel fibers (PFs) PGCs. See Periglomerular cells (PGCs) Phenylketonuria (PKU), 543 Phonemes, 414 Phospholipase C (PLC), 133e134 PLC-b1, 153 PLCb4, 81e82 Physical skills, 591 Piagetian theory, 336e337 Pinwheels, 110 PKCg. See Protein kinase Cg (PKCg) PKU. See Phenylketonuria (PKU)

Placental CRH (pCRH), 566 Plasticity, 274, 278 of attention networks, 516e517 in olfactory bulb, 17e21 adult neurogenesis, 17e20 circuit and synaptic plasticity, 20e21 PLC. See Phospholipase C (PLC) Plexin A4 (PlxnA4), 83 Plexin C1 (PlxnC1), 87e88 PlxnA4. See Plexin A4 (PlxnA4) PNNs. See Perineuronal nets (PNNs) PNS. See Parasympathetic nervous system (PNS) Pom nucleus. See Posterior medial nucleus (Pom nucleus) Positron emission tomography (PET), 289, 539, 600 Posterior medial nucleus (Pom nucleus), 114, 149 Posterior temporal cortex (pSTS), 491 Posterioreventral cochlear nucleus (PVCN), 36 Posterror slowing (PES), 524 Postnatal development of electrophysiological patterns, 212e215 development of major hippocampal rhythms, 214e215 early developmental patterns of brain activity, 214 single cell electrophysiological properties, 212e213 Postnatal stress and neurobehavioral development, 572e577 early adversity, 574e575 diurnal cortisol, 574 early care effects on cortisol set points and reactivity, 574e575 individual differences in sensitivity to experience, 575e577 social regulation of HPA axis and role of caregivers, 573e574 Postnatal visual development, 340e349 cortical maturation and oculomotor development, 343e345 critical periods, 342e343 for development of holistic perception, 347e349 face perception, 347 object perception, 346e347 visual attention development, 343 visual memory development, 345 visual physiology, 342 visual stability development, 345e346 Postsynaptic density-93 (PSD-93), 89e90 Posttraumatic stress disorder (PTSD), 575 Pragmatic development, 419 Precerebellin, 90e91 Precuneus (PC), 467 Prefrontal cortex (PFC), 539 Pregnancy, stress regulation and, 566e567 Premotor area, 169e170 Prenatal hormone influences on brain sex differences, 615 Prenatal psychosocial stress, 569

Prenatal stress, 565e572 epigenetics, 570e571 fetal programming, 565e566, 571e572 gestational stress influences human fetus, 567e568 interactions with postnatal environment, 571 and neurobehavioral development, 565e572 prenatal maternal biological stress signals and infant and child development, 569e570 cognitive development, 570 HPA functioning, 570 social/emotional development, 569 prenatal maternal psychosocial stress and infant and child development, 568 cognitive development, 569 HPA functioning, 569 socioemotional development, 568 sex differences, 570 stress regulation and pregnancy, 566e567 Prenatal visual function, 338 Preschooler(s), 541, 543. See also Schoolage children face processing in, 444e447 functional signatures of face processing, 446e447 handling young children process faces, 444e445 facial expressions, 445e446 Primary somatosensory cortex, 143, 169 Principal component analysis (PCA), 437, 475 Probabilistic epigenesis, 277 Procedural memory, 397e398 Progenitor cells, 119 Prosopagnosia, 450e451 Protein kinase Cg (PKCg), 86 Protodeclaratives, 418e419 Protoimperatives, 418e419 Prototypical GPe neurons, 233 Pruning, 290 PS. See Perinatal stroke (PS) PSD-93. See Postsynaptic density-93 (PSD-93) pSTS. See Posterior temporal cortex (pSTS) Psychological disorders, 593 Psychological sex differences, 587e600. See also Brain sex differences cognitive skills, 587e590 genetic perspectives, 595 hormone perspectives, 595e599 integrated perspectives, 599e600 noncognitive sex differences, 590e593 socialization perspectives, 593e595 Psychopathology, 531e532 PT type. See Pyramidal tract type (PT type) Ptf1a. See Pancreatic transcription factor 1a (Ptf1a) PTSD. See Posttraumatic stress disorder (PTSD) Pubertal Development Scale (PDS), 294 Purkinje cells (PCs), 79, 118

Index

heterogeneity of, 245e246 layer interneurons, 256e257 PV. See Parvalbumin (PV) PVCN. See Posterioreventral cochlear nucleus (PVCN) Pyramidal tract type (PT type), 147, 222e223

Q Quadrigeminal plate. See Tectal plate Quantal EPSCs (qEPSCs), 83e84

R Rad51, 186e187 Radial glial cells (RGCs), 121 Radial unit hypothesis, 151e152 Randomized controlled trial (RCT), 546 Receptive fields of barrel cortex neurons, 154 Reelin, 210, 246 Region-of-interest (ROI), 280, 292, 600e601 Regional structure volume, 602e604 sex differences in interhemispheric commissures, 603 structures involved in learning and memory, 603 subcortical structures, 603e604 Relational aggression, 592 Retinocentric saccades, 345 Retinotectal inputs, 63e64 Rey Osterreith Complex Figure (ROCF), 365e366 RFA. See Rostral forelimb area (RFA) RGCs. See Radial glial cells (RGCs) rLS. See Rostral lateral suprasylvian sulcus (rLS) RMS. See Rostral migratory stream (RMS) ROCF. See Rey Osterreith Complex Figure (ROCF) ROI. See Region-of-interest (ROI) Rostral forelimb area (RFA), 168e169 Rostral lateral suprasylvian sulcus (rLS), 60 Rostral migratory stream (RMS), 18

S SAM system. See Sympatheticeadrenal medullary system (SAM system) Satb2. See Special AT-rich sequence binding protein 2 (Satb2) SBCs. See Single bouquet cells (SBCs) SC. See Superior colliculus (SC) Scala media, 33 Scheibel collaterals, 244e245 School age, 541e542, 544 School readiness, 540, 548 School-age children face processing in, 447e450 handling children and adolescents process faces, 447e448 facial expressions, 448 neural substrates, 448e449 Schwann cells, 88

SCPN. See Subcerebral projection neurons (SCPN) Semaphorin 3A (Sema3A), 83, 94, 184 Semaphorin 7A (Sema7A), 87e88 Sensorimotor cortex, 170 Sensory chronology, 63e67 Sensory inputs, olfactory bulb, synaptic organization, 4e6 Sensory map alignment, 61 Sensory processing modulation, 14e17 brain state and context, 16e17 local circuits and centrifugal innervation, 14e16 Sensory systems, 3 “Sensory training” paradigm, 72 SES. See Socioeconomic status (SES) Sex differences in brain and behavioral development brain sex differences, 600e613 interpreting sex differences, 587 issues in studying sex differences, 586 psychological sex differences, 587e593 in stress, 570 Sexual orientation, 591 SG. See Spiral ganglion (SG) SGNs. See Spiral ganglion neurons (SGNs) Sharp-wave ripples (SWR), 214e215 Short-term memory, 396 Sibling neuron circuits in developing columns, 120e121 Silent synapses, 152e153 Single bouquet cells (SBCs), 148 Sister neurons, 120 Skill learning, 276 Slow timescale decorrelation, 12e13, 12f SLR. See Subthalamic locomotor region (SLR) sMRI. See Structural MRI (sMRI) SNc. See Substantia nigra pars compacta (SNc) SNr. See Substantia nigra pars reticulata (SNr) SOA. See Stimulus onset asynchrony (SOA) SOC. See Superior olivary complex (SOC) Social aggression, 592 Social behaviors, 592 cognitive control role in, 530 Socialecognitive foundations of language, 418e419, 425 infant-directed talk, 418 intentional communication, 418e419 pragmatic development, 419 social engagement, 418 Socialization of cognitive sex differences, 593e594 of noncognitive sex differences, 594e595 perspectives, 593e595, 613 Socioeconomic status (SES), 515, 542, 593 Socioemotional development prenatal maternal biological stress signals and infant and child development, 569 prenatal maternal psychosocial stress and infant and child development, 568 SOM. See Somatostatin (SOM)

647

Somatosensory cortex adult plasticity, 159 ascending pathways, 144f connections during birth and migration, 151e152 developmental critical periods and barrel formation in, 153e159 mature cortical circuit, 144e151 thalamocortical innervation, 152e153 Somatostatin (SOM), 209, 213 Somatotopy, 59 Sox5. See SRY-box 5 (Sox5) Spatial skills, 588, 594, 607e608 Special AT-rich sequence binding protein 2 (Satb2), 181 Specialization, 279 Species-specific differences of corticospinal connectivity, 175e177 Species-specific organization, 168e170 b-Spectrins, 93 Speech perception, 414e416, 424 audiovisual integration in early, 416 development in infancy, 415e416 in neonates, 415 prenatal perception of speech, 414 production, 416e418, 424e425 phonological development, 417e418 relationship between speech and motor development, 417 speech production in infancy, 416e417 Spike timingedependent plasticity (STDP), 128, 129f cellular mechanisms, 132e134 and circuit homeostasis, 132 contributing to adult plasticity and learning, 136e137 contributing to development of neural circuits, 135e136 definition and forms, 128e130, 129f dendritic excitability and, 134 discovery, 128 functional properties, 131e132 mechanisms for LTP and LTD components, 132e134 multifactor plasticity rule, 130e131 neuromodulation, 132 realistic learning rule in vivo, 135 Spina bifida meningomyelocele, 376 Spiny projection neurons (SPNs), 221e222, 228e229 Spiral ganglion (SG), 29e30 Spiral ganglion neurons (SGNs), 29e30 SPL. See Superior parietal lobs (SPL) SPNs. See Spiny projection neurons (SPNs); Subplate neurons (SPNs) Spoken languages, 413e414 Spontaneous sharp-wave bursts (SPW bursts), 211 SPW bursts. See Spontaneous sharp-wave bursts (SPW bursts) SRY-box 5 (Sox5), 181 sSACs. See Superficial short-axon cells (sSACs)

648 Index

Static mapping, 277e278 Statistical learning mechanisms learning co-occurrence statistics, 320e327 learning probability distributions, 319e320 linking individual differences in statistical learning to language development, 327e328 scaling statistical learning to real-world challenges, 329e330 statistical learning in individuals with language delays and disorders, 328e329 STDP. See Spike timingedependent plasticity (STDP) Stellate cells, 259 stellate cell-PC synapse formation, 94e95 Stereopsis, 342e343 Stimulus onset asynchrony (SOA), 368 STN. See Subthalamic nucleus (STN) Stress anatomy and physiology, 562e565 future directions, 577e578 hormones, 567 postnatal stress and neurobehavioral development, 572e577 prenatal stress and neurobehavioral development, 565e572 Striatal interneurons, 230e232 Striatum, 221e223, 228e232 physiology, 230 striatal interneurons, 230e232 Stripes, 245 Structural brain development connecting different techniques brainebehavior relationships, 305e309 multimodal imaging, 304e305 diffusion magnetic resonance imaging, 300e304 MRI brain mapping approaches, 294e300 volume analyses, 292e294 postmortem studies and histology, 290e292 myelination, 291 sex-specific differences, 291 synaptogenesis and pruning, 290 Structural MRI (sMRI), 600 STS. See Superior temporal sulcus (STS) Subcerebral projection neurons (SCPN), 172 specification and differentiation of, 180e182 Subcerebral projections anatomical organization of, 171e172 functional organization, 172e174 Subcortical structures, sex differences in, 603e604 Subplate neurons (SPNs), 45, 152 Substantia nigra pars compacta (SNc), 226e228 Substantia nigra pars reticulata (SNr), 221e222, 226e228 Subthalamic locomotor region (SLR), 172e174 Subthalamic nucleus (STN), 221e222, 234e235

SUM. See Supramammillary nucleus (SUM) Superficial layer visuotopy, 63e64 Superficial short-axon cells (sSACs), 8e9 Superficialedeep layer maturational delay, 67 Superior colliculi, 57 Superior colliculus (SC), 57, 117 anatomical organization, general, 58, 58f functional role, 57e58 maturation, 63e73 motor output, 59e63 movement field, 62f multisensory enhancement, 64f multisensory integration, 59e63 spatial topographies, 59e63 Superior olivary complex (SOC), 29e30 Superior parietal lobs (SPL), 369 Superior temporal sulcus (STS), 280, 438e439, 470 Supragranular excitatory neurons, 145e146 Supramammillary nucleus (SUM), 215 SWR. See Sharp-wave ripples (SWR) Sympatheticeadrenal medullary system (SAM system), 564e565 Synaptic learning rules, 127e128 Synaptic organization, olfactory bulb, 4e17 Synaptic plasticity, 20e21, 127e128. See also Plasticity Synaptogenesis, 290 Syntactic development, 422e423, 426. See also Postnatal visual development developmental stages in, 422 early sentences, 422 grammatical morphology, 422 later grammatical development, 423 neural bases of grammatical development, 423

T TAARs. See Trace amine-associated receptors (TAARs) Targeted visual exploration, learning from, 349e350 Task-driven fMRI measurements, 469 TCs. See Tufted cells (TCs) Tectal plate, 57 Tectum, 57 Temperament, 592 Temporal-cortical network development, functional consequences of, 404e405 Temporoparietal junction (TPJ), 467 Thalamic connectivity, 187 Thalamocortical innervation, 152e153 subplate circuitry, 45e46 Theory of mind (ToM), 467, 490e491, 540 discovering reliable neural markers of individual differences in, 475 early sensitivity to mental states prior neural and behavioral evidence, 468 neuroimaging studies of young children and infants, 468e470

family role on, 478e479 neural correlates of ongoing ToM development in childhood, 470e474 integration and separation of functional networks, 473e474 reliable spontaneous (uninstructed) responses to movies, 472e473 response selectivity, 471e472 neural correlates of structural changes in, 474e475 neurophysiological approaches to understanding cognitive empathy, 491 role of developmental experience culture, 477e478 language, 475e477 Theta oscillations, 214e215 Third-order brain stem nuclei, afferent regulation of, 42e43 Toddlerhood, 510e511 Toddlers, face processing in, 444e447 functional signatures of face processing, 446e447 handling young children process faces, 444e445 handling young children process facial expressions, 445e446 ToM. See Theory of mind (ToM) TP. See Transitional probability (TP) TPJ. See Temporoparietal junction (TPJ) Trace amine-associated receptors (TAARs), 5 Transient columnar domains during development, 121e122 Transitional probability (TP), 321 Translaminar plasticity, 159 Transverse zones, 245 Tufted cells (TCs), 8 Turner syndrome, 378e379

U Unipolar brush cells (UBCs), 255

V VAN. See Ventral Attention Network (VAN) VAS. See Vibroacoustic stimulus (VAS) Vasoactive intestinal peptide (VIP), 209 VB. See Ventrobasal complex (VB) VBM. See Voxel-based morphometry (VBM) VDCCs. See Voltage-dependent Ca2+ channels (VDCCs) Ventral Attention Network (VAN), 369 Ventral nucleus of trapezoid body (VNTB), 36 Ventral posterior medial nucleus (VPM nucleus), 114, 144 Ventral stream processes, 361e366 perception of faces, 364 of global and local levels of visual pattern structure, 362e364 spatial construction, 365e366 Ventrobasal complex (VB), 113

Index

Ventromedial prefrontal cortex (vmPFC), 487e488 Ventroposterolateral nuclei (VPL nuclei), 152 Verbal skills, 589 Very low-density lipoprotein receptor (Vldlr), 246 Vesicular glutamate transporter 2 (VGluT2), 84e85 Vestibulo-ocular response (VOR), 345 VGluT2. See Vesicular glutamate transporter 2 (VGluT2) Vibroacoustic stimulus (VAS), 568 VIP. See Vasoactive intestinal peptide (VIP) Vision, 335 Visual attention development, 343 Visual cortex modules, 110 Visual domain, learning co-occurrence statistics in, 323e324 Visual memory development, 345 Visual stability development, 345e346 Visual system development classic theoretical accounts, 336e337 gestalt theory, 337 Piagetian theory, 336e337 hormonal and environmental influences on object perception, 352e353 learning from

associations between visible and occluded objects, 350 targeted visual exploration, 349e350 visual-manual exploration, 350e352 postnatal visual development, 340e349 prenatal development of visual system, 337e338 prenatal visual function, 338 structure development in visual system, 338 visual perception in newborn, 339e340 faces and objects, 340 visual behaviors at birth, 340 visual organization at birth, 340 Visual word form area (VWFA), 280, 471 Visuospatial processing development, 359e380 anatomical organizations of primary visual systems, 361 dorsal stream processes, 366e371 ventral stream processes, 361e366 Visuotopy, 59 Vldlr. See Very low-density lipoprotein receptor (Vldlr) vmPFC. See Ventromedial prefrontal cortex (vmPFC) VNTB. See Ventral nucleus of trapezoid body (VNTB)

649

Voltage-dependent Ca2+ channels (VDCCs), 79e80 Voltage-sensitive calcium channels (VSCCs), 133 VOR. See Vestibulo-ocular response (VOR) Voxel-based morphometry (VBM), 294 Voxel-based strategies, 294e296 VPL nuclei. See Ventroposterolateral nuclei (VPL nuclei) VPM nucleus. See Ventral posterior medial nucleus (VPM nucleus) VSCCs. See Voltage-sensitive calcium channels (VSCCs) VWFA. See Visual word form area (VWFA)

W White matter, 605 Williams syndrome (WS), 376e378, 450e453 face processing in, 453 Word neural bases of word learning, 421 segmentation, 321e322, 324e325

Z Zebrin I, 249 Zfpm2, 187 Zinc finger nucleases (ZFN), 189