Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations [1 ed.] 0128166606, 9780128166604

Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations focuses on the neurophysiological b

375 59 5MB

English Pages 306 [291] Year 2020

Report DMCA / Copyright

DOWNLOAD FILE

Polecaj historie

Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations [1 ed.]
 0128166606, 9780128166604

Table of contents :
Dedication
Front Matter
Copyright
Contributors
Preface
References
Acknowledgments
Beyond negativity: Motivational relevance as cause of attentional bias to positive stimuli
Introduction
Main theories of attention to positive information
Phylogenetic relevance
Ontogenetic relevance
Current relevance
Arousal as cause of attentional bias to phylogenetic and ontogenetic relevant events
Methods
Cueing paradigms
Exogenous spatial cueing paradigm
Dot probe paradigm
Emotional Stroop paradigm
Emotional flanker paradigm
Visual search
Brain regions involved in the emergence of the Bias
Somatovisceral responses related to the Bias
Similarities and differences between healthy and clinical populations
Limitations and future directions
Summary
References
Attention bias toward negative stimuli
Introduction
Major theories on the nature and underlying causes of attention biases
Prevalent paradigms and findings in attention bias
Brain regions involved in the emergence of attention bias
Autonomic responses related to attention bias
Similarities and differences between healthy and clinical populations
Limitations and future directions
Summary
References
Further Reading
The neurophysiological basis of optimism bias
Introduction
Major theories in the field
When does optimism bias manifest?
How and why does optimism bias emerge?
Methods used to investigate the optimism bias
Brain regions involved in the emergence and maintenance of the bias
Somatovisceral responses related to optimism bias
Similarities and differences between healthy and clinical populations
Limitations
Future directions
Summary
References
Negative expectancy biases in psychopathology
Introduction
Methods used to investigate expectancy biases
Development of expectancy biases
Negative expectancy biases
Pathways to expectancy bias
Factors that contribute to the robustness of expectancy bias
Avoidance behaviors
Modulatory effects of expectancies
Uncontrollability of expected outcomes
Uncertainty and illusory correlations
Postevent “validation” processes (rumination and immunization)
The neural basis of expectancy biases
Expectations influencing perception
Neural correlates of overgeneralization of fear stimuli
Neural correlates of regulatory responses
Somatovisceral responses related to negative expectancy biases
Similarities and differences between healthy and clinical populations
Recapitulation, limitations, and some future directions
References
Further reading
Positive interpretation bias across the psychiatric disorders☆
Introduction
Major theories in the field
Methods used to investigate a positive interpretation bias
Brain regions involved in the emergence of a positive interpretation biases
Positive interpretation bias specifically
Constructs related to a positive interpretation bias
Summary
Somatovisceral (e.g., autonomic) responses involved in the emergence of a positive interpretation bias
Similarities and differences between healthy and clinical populations in the biological basis of positive interpretation bi ...
Depression
Social anxiety disorder
Comorbid depression and social anxiety disorder
Other anxiety disorder
Summary and debate
Limitations
Future directions
Chapter summary
References
Resolving ambiguity: Negative interpretation biases
Introduction
Major theories in the field
Methods used to investigate the biases
Direct measures of interpretation biases
Indirect measures of interpretation biases
Brain regions involved in the emergence of the bias
Somatovisceral responses related to the biases
Similarities and differences between healthy and clinical populations
Negative interpretation biases: A risk factor for anxiety and depression
Limitations
Future directions
Summary
References
A “rosy view” of the past: Positive memory biases
Methods used to investigate the bias
Major theories in the field
Self-protecting and self-enhancing mechanisms
Avoiding negative information and assimilating positive information
Distancing negative memories
Emotion regulation strategies
Social disclosure
Self-consistency
Consistency with current knowledge
Consistency with current emotions
Consistency with expectations
Variables that may moderate the bias—Individual differences
Cultural differences
Age-related differences
Personality differences
Brain regions involved in the emergence of the bias
Similarities and differences between healthy and clinical populations
Summary, limitations, and future directions
References
Further reading
Negative memory biases in health and psychiatric disorders
Introduction
Major theories of affective memory biases
Semantic/associative network models
Beck’s theory of emotional disorders
Elaboration/priming hypotheses
Overgeneral memory and executive functioning
The combined cognitive bias hypothesis
Methods for examining memory biases
Measuring negative encoding biases
Measuring negative recall biases
Brain regions involved in the emergence of memory biases
Somatovisceral contributions
Endocrine correlates
Similarities and differences between healthy and clinical populations
Limitations
Future directions
Summary
References
Further reading
The interplay among attention, interpretation, and memory biases in depression: Revisiting the combined cognitive bias hyp ...
Introduction
Major theories in the field
The causal loop diagram of depression dynamics
The attention-memory bias-interaction research framework
Conceptualizing the combined influence of cognitive biases on depression over time
Methods used to investigate the CCBH
Association questions
Causal questions
Predictive magnitude questions
Empirical research on the CCBH
Association questions
Causal questions
Predictive magnitude questions
Limitations and future directions
Summary
References
Further reading
The impact of top-down factors on threat perception biases in health and anxiety
Introduction
Major theories and evidence supporting top-down threat processing
Theoretical considerations
Methods used to investigate threat perception
Neural mechanisms involved in threat processing
Neural mechanisms of bottom-up threat processing
Neural mechanisms of top-down threat processing
Peripheral mechanisms of threat processing
Threat processing in anxiety versus healthy populations
Neural mechanisms of threat processing in anxiety versus healthy populations
Limitations
Future directions
Summary and conclusions
References
Cognitive biases across development: A detailed examination of research in fear learning
Introduction
Cognitive biases related to fear learning across development
Fear acquisition and extinction
Retrieval of fear extinction memories
Biases related to overgeneralization of fear
Conclusions and future directions
References
Further reading
Attentional control and cognitive biases as determinants of vulnerability and resilience in anxiety and depression
Introduction
Cognitive biases and emotional vulnerability
Neurocognitive mechanisms of attentional control in emotional vulnerability
A mediating role for attentional control in cognitive biases
Exercising attentional control to reduce emotional vulnerability
Implications for real world, educational, and clinical settings
Can adaptive cognitive training change cognitive bias?
Concluding remarks and future directions
References
Further reading
Index
A
B
C
D
E
F
G
I
M
N
O
P
R
S
T
U
V
W

Citation preview

“Don’t believe everything you think.” Louise Penny, A Great Reckoning

To Guido, Maximilian, Benjamin, and Anna Noam, Yaron, and Shachar

Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations

Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations Edited by

Tatjana Aue Hadas Okon-Singer

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 © 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: 978-0-12-816660-4 For information on all Academic Press publications visit our website at https://www.elsevier.com/books-and-journals

Publisher: Nikki Levy Acquisitions Editor: Joslyn Chaiprasert-Paguio Editorial Project Manager: Samantha Allard Production Project Manager: Surya Narayanan Jayachandran Cover Designer: Mark Rogers Typeset by SPi Global, India

Contributors Numbers in parenthesis indicate the pages on which the authors’ contributions begin.

Elinor Abado  (19), Department of Psychology; The Integrated Brain and Behavior Research Center (IBBR), University of Haifa, Haifa, Israel Orly Adler (139), Department of Psychology and The Institute of Information Processing and Decision Making, University of Haifa, Haifa, Israel Umkalthoom Alzubaidi  (1), School of Psychology and Clinical Language Sciences, Reading, University of Reading, United Kingdom Tatjana Aue (41), Department of Psychology, University of Bern, Bern, Switzerland Yasmene Bajandouh  (1), School of Psychology and Clinical Language Sciences, Reading, University of Reading, United Kingdom Anna Benedict  (119), Department of Psychology, University of Notre Dame, Notre Dame, IN, United States Matthew Burke (99), University of British Columbia, Vancouver, BC, Canada Judith K. Daniels  (71), Department of Clinical Psychology and Experimental Psychopathology, University of Groningen, Groningen, The Netherlands Nazanin Derakhshan  (261), Department of Psychological Sciences, Birkbeck University of London, London, United Kingdom Mihai Dricu (41), Department of Psychology, University of Bern, Bern, Switzerland Jonas Everaert  (193), Department of Experimental-Clinical and Health Psychology, Ghent University, Ghent, Belgium J.A. Faunce (173), Department of Psychology, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States B.H. Friedman  (173), Department of Psychology, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States Rivkah Ginat-Frolich (243), Department of Psychology and the Integrated Brain and Behavior Research Center, University of Haifa, Haifa, Israel S.S. Grant (173), Department of Psychology, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States A.M. Huskey (173), Department of Psychology, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States Jingwen Jin (215), Department of Psychology; The State Key Laboratory of Brain and Cognitive Sciences, The University of Hong Kong, Pok Fu Lam, Hong Kong

xiii

xiv  Contributors Peter J. de Jong  (71), Department of Clinical Psychology and Experimental Psychopathology, University of Groningen, Groningen, The Netherlands Ellen Jopling (99), University of British Columbia, Vancouver, BC, Canada Ernst H.W. Koster (193), Department of Experimental-Clinical and Health Psychology, Ghent University, Ghent, Belgium Laura Kress (41), Department of Psychology, University of Bern, Bern, Switzerland; Department of Psychology, University of Uppsala, Uppsala, Sweden Joelle LeMoult (99), University of British Columbia, Vancouver, BC, Canada Aprajita Mohanty  (215), Department of Psychology, Stony Brook University, Stony Brook, NY, United States Hadas Okon-Singer  (19), Department of Psychology; The Integrated Brain and Behavior Research Center (IBBR), University of Haifa, Haifa, Israel Ainat Pansky  (139), Department of Psychology and The Institute of Information Processing and Decision Making, University of Haifa, Haifa, Israel Thalia Richter  (19), Department of Psychology; The Integrated Brain and Behavior Research Center (IBBR), University of Haifa, Haifa, Israel Victoria Shaffer (119), Department of Psychology, University of Notre Dame, Notre Dame, IN, United States Tomer Shechner  (243), Department of Psychology and the Integrated Brain and Behavior Research Center, University of Haifa, Haifa, Israel Tamara J. Sussman  (215), Department of Psychiatry, Columbia University Irving Medical Center; Department of Psychiatry, New York State Psychiatric Institute, New York, NY, United States Alison Tracy (99), University of British Columbia, Vancouver, BC, Canada Julia Vogt  (1), School of Psychology and Clinical Language Sciences, Reading, University of Reading, United Kingdom Jessica Wilson (99), University of British Columbia, Vancouver, BC, Canada K. Lira Yoon (119), Department of Psychology, University of Notre Dame, Notre Dame, IN, United States

Preface Biases in information processing have attracted major research interest in examining psychiatric disorders such as anxiety and depression, as well as in studying populations at risk of developing psychopathology. Even among healthy individuals, these biases have been receiving increasing attention. This is b­ ecause our feelings about our own lives and about other people are strongly influenced by how we attend to our environment, what we expect, and how we interpret or memorize a given situation. Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations offers a comprehensive and in-depth description of these critical biases, provides examples from relevant ­cutting-edge research, and discusses methodological developments, implications, future directions, and open questions. The traditional view is that biased information processing is categorically unhealthy and disadvantageous. This book reveals evidence contradicting this outdated view by delineating adaptive and maladaptive forms of positive and negative biases. Such evidence shows that no bias is “bad” or “good” per se. Rather, often the strength of a bias is what determines its degree of adaptability. Identifying and applying the correct dose of a given bias may therefore be crucial to leading a healthy and satisfactory life. The fact that biases can interact (e.g., reinforce each other) has also been disregarded until lately. This omission may play a major role in the difficulties in accessing and distinguishing between different mental states or psychiatric diagnoses (e.g., anxiety vs depression) and in assessing the origins of such states. Indeed, understanding the mechanisms underlying mental disorders (or well-being in general) may not always be as simple as has often been proposed in the literature. Consideration of causal influences between biases can create a picture of human functioning that is not only more complex and mentally challenging but also more realistic. A better understanding of any given phenomenon usually goes along with its facilitated convertibility. Hence, in the long run, interrelated investigations have the potential to foster quality of life more rapidly and more effectively. The chapters of this book provide a comprehensive summary of studies examining biases in attention, expectancy, interpretation, and memory. Each chapter features information regarding leading theoretical views, experimental paradigms for studying the specific bias, and behavioral, neural, and somatovisceral findings. Moreover, in each chapter, the authors refer to limitations in existing literature and suggest future research directions to achieve a better understanding of each bias. Finally, individual chapters are dedicated to developmental aspects of processing biases, causal relations between the different biases, and implications of the scientific and theoretical research for everyday life. xv

xvi  Preface

To date, most research on processing biases has focused on attention biases. In Chapter  1, Vogt, Bajandouh, and Alzubaidi (2020) underscore the importance of attention biases to positive stimuli. Biased attention to positiverewarding cues is thought to be based on their relevance, which originates in their evolutionary significance, in the positive valence they acquired via learning processes, or in their relevance to current goals. Attention biases to positive stimuli are associated with neural activation in regions related to reward, attention, and control, including the amygdala, the anterior cingulate cortex, and the dorsolateral prefrontal cortex. Autonomic measures are rather inconsistent, although changes in facial muscle activity, heart rate, and skin conductance have been associated with attention biases to positive stimuli. Attention biases toward positive stimuli have been commonly considered an attribute of healthy behavior, with clinical and subclinical populations showing decreased strength of such biases. Yet, in certain situations, these biases can be problematic, as in the case of orienting attention toward high caloric tasty food among obese participants. In Chapter 2, Abado, Richter, and Okon-Singer (2020) focus on biased attention toward negative stimuli (here, threat cues), which includes facilitated engagement of attention to threat, followed by difficulty in disengaging attention from the threat cue and subsequently by enhanced attentional avoidance. These three attention bias components have been studied extensively using various methods, paradigms, stimuli, and populations. Corresponding studies point to consistent biases among participants with anxiety disorders, while evidence regarding attention bias among participants with depression is mixed. Biases in such clinical populations are accompanied by abnormal neural activation in the amygdala, the thalamic pulvinar nucleus, and prefrontal and parietal regions, as well as abnormal autonomic reactivity that includes eye movements, heart rate, blood pressure, and skin conductance. Attention biases also appear in individuals with subclinical levels of anxiety, although disengagement and avoidance are less robust in subclinical populations, possibly due to large interindividual differences. In Chapter  3, Dricu, Kress, and Aue (2020) focus on positive expectancy biases, and specifically on optimism bias. Optimism bias (also termed overoptimism) refers to an individual’s tendency to estimate positive events as more likely and negative events as less likely than what would be predicted by a rational consideration of the information at hand. This type of responding is hard to overcome and may even stabilize over time, as demonstrated by the fact that information in the environment supporting an initial optimism bias is more easily integrated than information that challenges overoptimistic expectancies. It is worth noting that overoptimism not only does exist for the self (i.e., personal optimism bias) but extends to others with whom we identify (social optimism bias). Moreover, the bias reverses into a pessimism bias toward unpopular ­out-groups. Whether or not optimism bias can be reliably revealed depends on its specific definition and the paradigm used for its investigation. Among the brain regions

Preface xvii

involved in optimism bias are the anterior cingulate cortex, the ventromedial prefrontal cortex, the insula, the precuneus/posterior cingulate cortex, and the striatum. To date, somatovisceral evidence is missing. Whereas slight overoptimism is considered a precondition for mental health and success, the absence of optimism bias (or even a pessimism bias) as well as extreme optimism bias have been associated with psychopathology (reduced optimism bias/pessimism bias: obsessive compulsive disorder, depression, borderline personality disorder; increased optimism bias: mania, substance abuse). In Chapter 4, de Jong and Daniels (2020) discuss the opposite pattern, i.e., negatively biased expectancies. The authors outline four qualitatively different types of bias (consequential expectations, response expectations, cue expectations, and self-efficacy expectations) and discuss the potential underlying mechanisms in depth. Existing data suggest the existence of a downward negativity spiral, with felt fear increasing negative expectancy bias and associated avoidance behaviors. In turn, these behaviors, along with strong resistance to (or ignoring of) contradicting evidence, prevent experiences that question a preexisting negative expectancy bias, thereby making this negative bias even more resistant to change (e.g., because people interpret their [unnecessary] avoidance behavior as proof of the dangerous nature of a situation). Whether or not expectancies are biased and how this deviation from rationality is expressed in behavior, brain responses, and peripheral physiology may depend not only on factors such as controllability and predictability of an aversive outcome but also on the nature of the cues used to predict the aversive outcome. Brain regions involved in the biasing of negative expectancies include lowlevel sensory areas as well as areas associated with executive control (in particular the dorsolateral prefrontal cortex). Yet, more research is needed to identify the exact brain and somatovisceral mechanisms underlying the d­ ifferent types of expectancy bias. In Chapter  5, Jopling, Wilson, Burke, Tracy, and LeMoult (2020) discuss positive interpretation biases (i.e., interpreting ambiguous situations in a positive manner as in “seeing the glass as half full”). Like the other biases discussed in this book, most research has focused on negative interpretation biases. Yet, accumulating evidence suggests that healthy individuals exhibit positive interpretation biases of ambiguous situations, while individuals with anxiety or depression do not exhibit these positive biases. Hence, a lack of positive interpretation biases can serve as a psychopathological risk factor. Interpretation biases are thought to be rooted in schemas that develop during childhood and are based on past experiences and social interactions. Abnormal somatovisceral responses to ambiguous situations emerge in psychopathology. Furthermore, evidence suggests that the ventromedial prefrontal cortex, the rostral anterior cingulate cortex, and connections between reward-related and occipital areas play a major role in positive interpretation biases. Ongoing research in this area will enhance our understanding of the biological and cognitive mechanisms ­underlying these biases.

xviii  Preface

In Chapter  6, Yoon, Shaffer, and Benedict (2020) summarize evidence regarding negative interpretation biases of ambiguous situations in clinical disorders, especially social anxiety and depression, as well as in subclinical states. Although negative interpretation is common across different disorders, the specific interpretative contents are unique to each disorder (e.g., “people think I’m boring” in social anxiety, and catastrophic interpretations of bodily symptoms in panic disorder). Negative interpretation of ambiguous situations also characterizes subclinical states. For example, interpreting physical symptoms as serious and uncontrollable is thought to exacerbate conditions of chronic fatigue. Direct measures (e.g., subjective reports of the interpretation) have yielded consistent findings of biases, while evidence from indirect measures (e.g., assessed by reaction time) is mixed. Neuroimaging studies suggest that the left prefrontal cortex plays a role in resolving lexical ambiguity. Connectivity between the dorsolateral prefrontal cortex and the amygdala may be specifically associated with negative interpretation biases. More research is needed to understand the association between these negative interpretation biases and somatovisceral responses, as well as their relation to other types of cognitive biases. In Chapter 7, Adler and Pansky (2020) discuss prevalent “rosy” views of the past, termed positive memory bias(es). Healthy individuals tend to remember the past as better than it truly was and to remember more pleasant than unpleasant events. Furthermore, memories associated with negative affect fade faster than memories associated with positive affect. Similar to other cognitive biases, positive memory biases are modulated by individual differences related to culture, age, and personality characteristics. Positive memory biases are common mainly in autobiographical memory and for self-relevant information. Theoretical views posit that positive memory biases are adaptive processes that emerge to maintain well-being and self-consistency. In line with these views, clinical populations usually exhibit the opposite pattern (i.e., negative memory biases), with one interesting exception: pathological gamblers. Neuroimaging studies suggest that the prefrontal cortex, the amygdala, the hippocampus, and sensory regions are involved in memory biases. More research is needed to link positive memory biases to somatovisceral responses, as well as to understand clinical conditions such as mania that may be characterized by abnormal positive memory biases. In Chapter  8, Grant, Huskey, Faunce, and Friedman (2020) summarize evidence regarding negative memory biases. Negative biases exist in different memory processes, including working memory, encoding, and retrieval. Diverse theoretical views have been developed to explain these memory biases, mainly by adherence to existing schemas and moods that lead to negative biases in psychopathologies such as depression and anxiety. As in the other biases summarized earlier, evidence points to the involvement of prefrontal-limbic inhibitory pathways. It is further suggested that dysregulation of the h­ ypothalamus-pituitary-adrenal (HPA) axis is associated with negative memory biases. Supplementary

Preface xix

investigations are required to understand the relations to other biases, as well as to shed light on memory biases in other clinical and subclinical states. In Chapter  9, Everaert and Koster (2020) summarize evidence regarding correlational and causal links between different processing biases. The authors focus on associations between attention bias, interpretation bias, and memory bias. Specifically, they conceptualize the combined influence of these biases on depression and aggregate existing research in the field with respect to the type of question investigated (and the conclusions to be drawn). Associative questions are concerned with detecting correlations and possible relationships between cognitive biases. Causal questions go a step further and additionally consider the direction of the influence between biases. Finally, predictive magnitude questions address the predictive power of the biases with respect to pathological symptoms. The authors emphasize that additional consideration of the time factor is essential in pinpointing relevant feedback and feed-forward interactions between different types of biased information processing. Furthermore, the neural and somatovisceral mechanisms involved need to be identified. In Chapter  10, Sussman, Jin, and Mohanty (2020) describe influences and interactions between top-down and bottom-up processes in biased threat perception. Among other things, the authors focus on the causal influence of (biased) expectancies on deployment of attention to subsequently presented stimuli. Prior expectancies may influence (facilitate or complicate) attention deployment by means of establishing mental templates upon which incoming sensory information can be paired. The neural substrates involved in this type of attention gating can, for example, encompass higher-order prefrontal and parietal regions. Currently, very little is known about the potential involvement of somatovisceral mechanisms. The evidence suggests that because earlier studies frequently neglected the effect of endogenous top-down factors such as prior expectancies on threat perception, the validity and predictive power of research in the area is limited. Differences between anxious and healthy populations are also highlighted. In Chapter 11, Ginat-Frolich and Shechner (2020) discuss developmental aspects of processing biases. The chapter begins by highlighting challenges that are specific to studying biases among children and adolescents. Then, the authors discuss biases related to different aspects of fear learning, generalization, and extinction as well as to threat appraisal in typically developing and anxious children and adolescents. Accumulating evidence points to age-­ related differences between typically developing and anxious children and adolescents in the context of fear learning. Specifically, although more research is needed, the relatively few studies on fear learning in anxious children and adolescents suggest enhanced fear in response to threat, greater difficulty in differentiating between threat and safety cues, and abnormalities in activation and connectivity of the amygdala, the hippocampus, and several prefrontal regions.

xx  Preface

Finally, in Chapter 12, Derakhshan (2020) discusses potential clinical applications of research on processing biases. The chapter focuses on cognitive training that is designed to strengthen attentional control, which in turn is hypothesized to alleviate symptoms of anxiety and depression and enhance ­resilience. Derakhshan provides an overview of cognitive training interventions in different populations at high risk of developing psychiatric disorders or symptoms. She suggests that attentional control may mediate the influence of cognitive biases on emotional vulnerability and resilience. Accordingly, studies involving cognitive training targeting attentional control may reveal causal relations between cognitive biases, attentional control, and resilience. Applications of attentional control training in educational and clinical settings are discussed. Together, these chapters provide an exhaustive review of research investigating positive and negative processing biases in the domains of attention, expectancy, interpretation, and memory over the last decades. For each domain, the chapters review positive as well as negative biases. Most biases represent normal phenomena. However, in their pathological appearance, these biases are extreme, persistent, and impair everyday functioning. We tend to view positive biases as healthy, often overlooking that they also characterize maladaptive states, such as gambling, obesity, or addiction. Moreover, negative biases that characterize depression and anxiety have been the subject of much research, while hardly any studies have examined patients with mania, who may show the opposite bias. The fact that examining the phasic manic state is challenging may explain why evidence is missing for mania but not for other disorders. A general conclusion arising from reviewing the book’s chapters is that very little somatovisceral data have been collected. Specifically, although the importance of adding implicit and/or autonomic measures when examining the different cognitive biases is clear, studies that measure somatovisceral data are not common. By contrast, neural correlates have recently begun to be addressed. Identification of the critical neurocognitive processes involved in biased information processing (in the positive as well as the negative direction) may ultimately help in the development of therapeutic venues or everyday applications that enhance quality of life. Importantly, different ­biases appear to share several key neural substrates pointing to the e­ xistence of shared underlying neural mechanisms. Such shared mechanisms may in fact be related to interactions between biases. Accordingly, some chapters review evidence for correlational and causal relations between these different biases. Nevertheless, more information is required to understand critical causal relations involved in biased information processing. Finally, the authors suggest future directions for research and therapy, such as studying the developmental trajectories of cognitive biases and ­developing clinical and subclinical applications based on the accumulating

Preface xxi

scientific findings reviewed. The aim of this book is to enable readers to better understand and access affect-based human cognitive biases and associated mental states. Tatjana Auea, Hadas Okon-Singerb,c Department of Psychology, University of Bern, Bern, Switzerland b Department of Psychology, University of Haifa, Haifa, Israel c The Integrated Brain and Behavior Research Center (IBBR), University of Haifa, Haifa, Israel a

References Abado, E., Richter, T., & Okon-Singer, H. (2020). Attention bias toward negative stimuli. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 19–40). San Diego: Elsevier. Adler, O., & Pansky, A. (2020). A “rosy view” of the past: Positive memory biases. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 139–171). San Diego: Elsevier. de Jong, P. J., & Daniels, J. (2020). Negative expectancy biases in psychopathology. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 71–97). San Diego: Elsevier. Derakhshan, N. (2020). Attentional control and cognitive biases as determinants of vulnerability and resilience in anxiety and depression. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 261–274). San Diego: Elsevier. Dricu, M., Kress, L., & Aue, T. (2020). The neurophysiological basis of optimism bias. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 41–70). San Diego: Elsevier. Everaert, J., & Koster, E. H. W. (2020). The interplay among attention, interpretation, and memory biases in depression: Revisiting the combined cognitive bias hypothesis. In T.  Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 193–213). San Diego: Elsevier. Ginat-Frolich, R., & Shechner, T. (2020). Cognitive biases across development: A detailed examination of research in fear learning. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 243–260). San Diego: Elsevier. Grant, S. S., Huskey, A. M., Faunce, J. A., & Friedman, B. H. (2020). Negative memory biases in health and psychiatric disorders. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 173–191). San Diego: Elsevier. Jopling, E., Wilson, J., Burke, M., Tracy, A., & LeMoult, J. (2020). Positive interpretation bias across the psychiatric disorders. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 99–117). San Diego: Elsevier. Sussman, T. J., Jin, J., & Mohanty, A. (2020). The impact of top-down factors on threat perception biases in health and anxiety. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 215–241). San Diego: Elsevier. Vogt, J., Bajandouh, Y., & Alzubaidi, U. (2020). Beyond negativity: Motivational relevance as cause of attentional bias to positive stimuli. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 1–18). San Diego: Elsevier.

Yoon, K. L., Shaffer, V., & Benedict, A. (2020). Resolving ambiguity: Negative interpretation biases. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 119–138). San Diego: Elsevier.

Acknowledgments We would like to express our appreciation to the authors of the different chapters of this book for their enthusiastic commitment throughout the entire publishing process. We are also indebted to the editorial board and staff of Elsevier Publications. Finally, we acknowledge the support of the Israel Science Foundation grant #823/18 (H O-S) and the Swiss National Science Foundation (grants PP00P1_150492 and PP00P1_183709, TA). The authors declare no ­conflicts of interest.

xxiii

Chapter 1

Beyond negativity: Motivational relevance as cause of attentional bias to positive stimuli Julia Vogt, Yasmene Bajandouh, Umkalthoom Alzubaidi School of Psychology and Clinical Language Sciences, Reading, University of Reading, United Kingdom

Introduction Imagine that you arrive at a party and you enter a room full of people: Who will attract your attention? The person smiling at you or someone who looks upset and angry? And now, imagine that you approach the buffet: Which food will grab your attention? The chocolate cake that you love or a healthier but less loved option such as the vegetables? These scenarios illustrate the phenomenon of attentional bias. Research on attentional bias to emotional information has been a focus of attention research for about 30 years (MacLeod, Mathews, & Tata, 1986; Yiend, 2010). Attention is a mechanism that allows observers to focus on a subset of possible sensory inputs (Luck & Vecera, 2002). In almost any given situation, people are surrounded by so much information that it is not possible to process all available information—such as at the party where you cannot pay attention to everybody and everything. Additionally, not all information is relevant to the ongoing behavior of an individual. Attention describes the processes and mechanisms that determine how sensory input, perceptual objects, trains of thought, or courses of action are selected from an array of concurrent possible stimuli, objects, thoughts, and actions (Talsma, Senkowski, Soto-Faraco, & Woldorff, 2010). Various fields of psychology and neuroscience have studied attentional bias to emotional information, including vision (neuro)science, clinical, or social psychology. In this chapter, we define attentional bias as increased allocation of attention to information that often occurs automatically, which means quick, efficient, unintentional, and/or uncontrollable (Moors & De Houwer, 2006). Much research on attentional bias has illustrated that negative and particularly threatening stimuli such as angry faces (Kuhn, Pickering, & Cole, 2016) evoke attentional bias especially when observers are high in state or trait ­anxiety Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00001-5 © 2020 Elsevier Inc. All rights reserved.

1

2  Cognitive biases in health and psychiatric disorders

(Bar-Haim, Lamy, Pergamin, Bakermans-Kranenburg, & van IJzendoorn, 2007; see also Chapter 2). However, positive stimuli attract attention as well (Pool, Brosch, Delplanque, & Sander, 2016), and smiling faces are faster detected than angry faces (Becker, Anderson, Mortensen, Neufeld, & Neel, 2011) (Fig. 1). Like the findings on the negativity bias, attention to positive information is enhanced or dependent on the observer’s current state or her personality. For instance, optimists (Kress, Bristle, & Aue, 2018; Segerstrom, 2001) display a positivity bias as well as observers who have positive thoughts activated on their mind (Smith et al., 2006; but see Van Dessel & Vogt, 2012). In the present chapter, we will discuss theories and evidence investigating when and why positive information such as the smiling person or a chocolate cake will grab attention. Specifically, we will highlight recent work that emphasizes how temporary goals can not only induce but also override attentional bias. We will proceed to discuss how attentional bias to positive stimuli can be measured and which brain regions and psychophysiological responses are associated with attention to positive input. While attention to positive events has been highlighted as characteristic of healthy populations, we will also discuss why it can be problematic; for instance, obesity seems to be related to an attentional bias to high-caloric but tasty food. We will finish the chapter by highlighting limitations and suggestions for future research.

Main theories of attention to positive information Most theories of attentional bias assume that the bias originates in the relevance of information in people’s environment. For instance, threatening events could represent potential dangers to survival, whereas beautiful people might offer possibilities for reproduction (Lang, Bradley, & Cuthbert, 1997; Neuberg, Kenrick, Maner, & Schaller, 2004). However, existing theories diverge in what they mean by relevance and in what kind of relevance they consider necessary for a stimulus to possess to be capable of attracting attention automatically. In what follows, we will identify three classes of theories proposing that positive events cause attentional bias because they are relevant. First, a wide range of theories assume that attentional bias to emotional events originates in the evolutionary relevance of these events, that is, because they were relevant during

FIG. 1  Examples of positive stimuli that evoke attentional bias.

Attention to positive information  Chapter | 1  3

the evolution of the human species to the survival or reproduction motive. A second set of theories assume that also stimuli that acquired positive valence via learning processes during the lifetime of the observer will attract attention. A third set of theories propose that the bias is driven by the current goals of the individual and positive (or any) information will attract attention when it is relevant to an active need or goal of the observer.

Phylogenetic relevance Evolutionary accounts are well-known for suggesting that only threats to survival that were present during the evolution of the human species such as angry facial expressions (Kuhn et  al., 2016) or dangerous animals like snakes and spiders (Lipp & Derakshan, 2005) will evoke attentional bias (Öhman & Mineka, 2001). According to these theories, attentional bias to emotional information evolved in the evolution because it was highly adaptive to become aware of these stimuli. Consequently, the bias is assumed to be hard-wired by now (LeDoux, 1996). Some of these accounts argue that biologically relevant negative and threatening stimuli will receive attentional priority because the fast detection of these events was more critical for survival than the detection of positive stimuli (Aarts & Dijksterhuis, 2003; Pratto & John, 1991). Moreover, Öhman and Mineka (2001; but see Pessoa & Adolphs, 2010) suggested that biologically relevant stimuli are more automatic and robust in biasing attention than other events. According to other authors (Lang et al., 1997; Neuberg et al., 2004), a system that only responds to negative events would be maladaptive because functional behavior also requires responding to stimuli that offer positive consequences. These theories suggest that the automatic allocation of attention is guided by three primary motivations: survival, sexual needs, and hunger. Indeed, attention is automatically directed to biologically relevant positive stimuli that correspond to the reproduction motive such as erotica (Most, Smith, Cooter, Levy, & Zald, 2007; Sennwald et al., 2016) and nonerotic images of beautiful people (Maner et  al., 2003). Further, infant faces displaying perceptual features of the baby schema such as large eyes and rounded cheeks evoke an attentional bias (Brosch, Sander, & Scherer, 2007). This can be interpreted as evidence that vulnerable offspring grabs attention to ensure successful caretaking. Finally, hungry participants, compared with satiated participants, show a stronger attentional bias to food-related stimuli (Tapper, Pothos, & Lawrence, 2010). The latter findings suggest that attentional bias to motivationally relevant events is sensitive to context and reflects changes in the strength of a need or motivation. We will come back to this observation in the third section of this review of major theories.

Ontogenetic relevance The preceding section discussed how both negative and positive stimuli of phylogenetic relevance attract attention. In this section, we will review ­theories

4  Cognitive biases in health and psychiatric disorders

s­uggesting that attentional bias is not limited to phylogenetically relevant events. Specifically, events that acquired negative or positive valence during the lifetime of an observer also appear to evoke attentional bias (Le Pelley, Mitchell, Beesley, George, & Wills, 2016). For instance, knives are comparable to phylogenetic threats in their capacity to evoke attentional bias in adults but not in children (Blanchette, 2006; LoBue, 2010; see also discussion in Chapter 2). Such modern threats were not present until quite late in the history of the human species. Therefore, the appraisal of relevance for such stimuli cannot be caused by a mechanism responding to inborn relevance but must be caused by a mechanism that responds to the learned relevance or valence of such stimuli. In an early demonstration of this effect, stimuli associated with positive attitudes for participants (e.g., a bike) were faster detected than stimuli that evoked neutral evaluations (Roskos-Ewoldsen & Fazio, 1992). Roskos-Ewoldsen and Fazio (1992) argued that attitudes, that is, positive or negative evaluations, have an orienting value. Consequently, they attract an observer’s attention to make them aware of stimuli that they like or dislike. Participants display attentional bias toward various stimuli that are associated with learned positive attitudes such as their partner (Dewitte, De Houwer, Koster, & Buysse, 2007) or stimuli related to a cherished hobbies like birds for bird lovers (Dalgleish, 1995) or exercise-related words for people who enjoy physical activity (Calitri, Lowe, Eves, & Bennett, 2009). Finally, many accounts attribute attentional bias to drugs in addictions to the learned liking of the drug (Franken, 2003; Mogg, Bradley, Field, & De Houwer, 2003). In recent years, various studies have shown how positive stimuli evoke an attentional bias even when their valence was only learned in the experimental session. For instance, when participants learn that a certain stimulus feature is related to winning money in an experiment, this feature grabs attention even after it stopped being associated with reward (Le Pelley et al., 2016; Raymond & O’Brien, 2009).

Current relevance Some studies have suggested that positive stimuli or any stimuli will only attract attention when they are currently relevant to an active goal of the individual (Gronau, Cohen, & Ben-Shakhar, 2003; Lichtenstein-Vidne, Henik, & Safadi, 2012; Vogt, Koster, & De Houwer, 2017). For instance, recent evidence suggests that drug-related cues bias attention in addicts but only when they pursue a goal of finding drug-related cues in a visual display (Brown, Duka, & Forster, 2018). Goal theories assume that goals are knowledge structures that represent desired end states and that goal pursuit is characterized by the heightened accessibility of these structures (Moskowitz, 2002). For instance, when the goal is to search for drug-related cues, the representation of drug-related cues will be highly accessible. According to these theories, the heightened accessibility of goal-relevant events in long-term or working memory will guide attention

Attention to positive information  Chapter | 1  5

automatically to stimuli in the environment that match the goal representation. Importantly, the goal representation often includes positive stimuli and means that are instrumental for goal pursuit such as food when hungry. Therefore, goal theories would suggest that positive information attracts attention when it is relevant to goal pursuit. Indeed, means that are instrumental to active goals evoke attentional bias that are automatic in the sense of unintended or fast (Vogt, De Hower, Moors, Van Damme, & Crombez, 2010; Wieber & Sassenberg, 2006). For instance, automatic attention to goal-relevant events emerges even when those stimuli are presented only briefly or when attention to goal-relevant events in the attention task is irrelevant for goal achievement (Lichtenstein-Vidne et al., 2012; Vogt et al., 2010). Further, goal-driven attention reflects the instrumentality of attended information for goal achievement. For instance, when the goal is to win as many tokens as possible, stimuli relevant to winning a high number of tokens attract attention over stimuli that are relevant to winning a low number of tokens (Vogt, De Houwer, & Crombez, 2011). Likewise, attention is not allocated to goal-related but goal-irrelevant information (e.g., “boat” when “ship” is goal relevant) indicating that the effect reflects relevance for goal pursuit rather than mere cognitive associations (Vogt, De Houwer, & Moors, 2011). Similarly, these studies suggest that goal-driven attention serves as a goal shielding mechanism by preventing attention to competing goals and other highly salient events (Vogt, Houwer, Crombez, & Van Damme, 2013). Importantly, these theories can explain why stimuli relevant to the survival and reproduction motive only attract attention when they also match an observer’s current concern (Gronau et al., 2003; Vogt et al., 2017). For instance, attentional bias to beautiful people is strongest in individuals looking for a partner and absent when the goal of being faithful to their current partner is activated in individuals who are in a monogamous relationship (Maner, Gailliot, & Miller, 2009). Relatedly, people are inattentive to tasty but high-caloric food when dieting goals are activated (Papies, Stroebe, & Aarts, 2008). Further, people attend to goal-­ relevant stimuli when goals are not fulfilled but not after completing (Moskowitz, 2002) or giving up on the goal. For instance, women who approached the childbearing deadline and wished to have a baby showed an increased attentional bias to baby pictures; in contrast, women who just passed the deadline and had given up their baby wish did not show this bias (Light & Isaacowitz, 2006). Based on these findings, researchers have suggested that emotion regulation goals could induce or prevent attention to emotional information. For instance, attention to positive events and away from negative events might reflect a person’s attempt to feel good or to suppress negative feelings. Indeed, attention to positive events has been linked to the desire to feel good (Segerstrom, 2001; Xing & Isaacowitz, 2006). Relatedly, suppressing negative feelings caused attentional avoidance of negative images but only when positive distractors are present (Vogt & De Houwer, 2014). However, at other times, aversive events might grab attention when the dominant emotion regulation goal indicates to

6  Cognitive biases in health and psychiatric disorders

fight the source of an emotional state. Supporting the latter assumption, we found that people attend to aversive situations when mastering those situations is possible (Vogt et al., 2017). Crucially, in this case, people also attend spontaneously to positive stimuli that allow them to alleviate the aversive situation directly. For example, they attend to stimuli such as water and soap when experiencing disgust (Vogt, Lozo, Koster, & De Houwer, 2011).

Arousal as cause of attentional bias to phylogenetic and ontogenetic relevant events Some researchers have suggested that attentional bias might be caused by high levels of arousal that characterizes most of the stimuli described earlier (Schimmack, 2005; Vogt, De Houwer, Koster, Van Damme, & Crombez, 2008). This would explain why attentional biases to phylogenetic and ontogenetic events are comparable without assuming multiple mechanisms underlying the bias. Similarly, emotion theories conceptualize arousal as an indicator of relevant events that should be selected by attentional processes for further processing (Lang et al., 1997). Indeed, various studies have found that high levels of arousal attract attention independent of valence (Schimmack, 2005; Vogt et al., 2008) or that it enhances the bias to positive stimuli (Pool et al., 2016). Future research is still needed to see whether goal-relevant events evoke high levels of arousal that could underlie the bias to goal-relevant events.

Methods In this section, we will review some of the most prominent paradigms that have been used to measure attentional bias to positive information. It is important to note that attention is not regarded as a unitary concept, but as an umbrella concept for a variety of processes (Luck & Vecera, 2002). For instance, some attentional processes refer to the selection of stimuli, which means how attention is focused on information or how people become consciously aware of it. Other processes explain how attention inhibits irrelevant stimuli and whether some stimuli cannot be ignored.

Cueing paradigms Exogenous spatial cueing paradigm Cueing paradigms are useful for measuring attentional orienting to peripheral cues (Posner, 1980; Vogt et al., 2008). In a spatial cueing paradigm, participants are asked to detect visual targets presented at two locations on the screen. The target is preceded by a visual cue at the same location (validly cued trials) or opposite location (invalidly cued trials). Valid cues typically lead to response time benefits (due to engagement of attention at the validly cued location), whereas invalid cues lead to response time costs (due to delayed disengagement

Attention to positive information  Chapter | 1  7

of ­attention from the invalidly cued location), a difference referred to as cue validity effect. Emotional cues lead to a larger cue validity effect than neutral cues and to both enhanced engagement and impaired disengagement, which implies that they engage and hold attention (Fox, Russo, Bowles, & Dutton, 2001; but see Mogg, Holmes, Garner, & Bradley, 2008).

Dot probe paradigm The dot-probe paradigm is like the cueing paradigm but presents two cues simultaneously at two different spatial locations on the screen (Dodd & Porter, 2010; Johnson, 2009) (Fig. 2). It does therefore not allow to differentiate between attentional engagement and disengagement (but see Koster, Crombez, Van Damme, Verschuere, & De Houwer, 2004). In contrast, it is useful to measure whether a stimulus attracts attention to its location in competition with other stimuli that might be a more realistic reflection of real environments and thus capture the true function of attention (Desimone & Duncan, 1995). Emotional Stroop paradigm The emotional Stroop paradigm has been used to measure attentional interference by positive stimuli (Gantiva, Araujo, Aragão, & Hewitt, 2018). In this task, participants must name the color of words or images (Fig. 2). The emotional Stroop effect describes the finding that it takes longer to name emotional than neutral stimuli (Williams, Mathews, & MacLeod, 1996). The effect reflects the capacity of a stimulus to interfere with a participant’s main task and to impair inhibition of task-­ irrelevant inputs. However, results in such tasks might reflect differences situated at the response stage (e.g., emotional stimuli interrupt the response selection mechanism and slow down motor responses) rather than differences in the allocation of attention (i.e., attention is directed toward stimuli; see Algom, Chajut, & Lev, 2004).

Emotional flanker paradigm The (emotional) Flanker paradigm is another paradigm used to capture interference during selective attention (Horstmann, Borgstedt, & Heumann, 2006).

FIG. 2  Schematic overviews of dot-probe task (A) and emotional Stroop task (B).

8  Cognitive biases in health and psychiatric disorders

In the emotional variant of the paradigm, participants must indicate the valence of a target stimulus while trying to ignore flanking distracters. In relation to the target stimulus, flankers may be congruent (flankers are of the same valence), incongruent (flankers are of opposite valence), or neutral (flankers are of neutral valence; Fenske & Eastwood, 2003; Horstmann et al., 2006). Participants cannot completely ignore the flankers resulting in slower responses on incongruent trials (interference effect) and faster responses on congruent trials (facilitation effect) (Horstmann et al., 2006).

Visual search In a visual search task, participants search for a discrepant target within a varying number of stimuli, for instance, a happy face among angry faces (see Fig. 3; Hickey, Chelazzi, & Theeuwes, 2010). Search time is measured. By varying the number of distractors, researchers can investigate whether a stimulus “pops out,” which indicates that search is independent of set size. Pop out suggests that a stimulus attracts attention fast and in a very efficient way without that observers must scan all stimuli to find the target. The visual search paradigm is therefore best suited to measure attentional capture.

Brain regions involved in the emergence of the Bias Traditionally, the limbic system has been characterized as the emotional brain (e.g., LeDoux, 1996). More precisely, the amygdala has been considered as the structure that is responsible for the fast and preferred processing of emotional and especially negative and threatening events (see also Chapter 2). Dominant theories assumed for a long time that sensory modalities project information

FIG. 3  Schematic overviews of flanker task (C) and visual search task (D).

Attention to positive information  Chapter | 1  9

on emotional input to the amygdala via subcortical thalmo-amygdala pathways and visual cortices (Tamietto & de Gelder, 2010). Indeed, those with amygdala lesions show impaired emotion processing (Vuilleumier, Richardson, Armony, Driver, & Dolan, 2004). However, recent evidence suggests that attentional bias to positive stimuli emerges as early as attentional bias to negative events. For instance, negative and positive stimuli evoke increased P1 amplitudes that reflect enhanced perceptual processing in the visual cortex (Brosch, Sander, Pourtois, & Scherer, 2008). This is in line with evidence showing that the amygdala reacts to relevant stimuli in general (Fitzgerald, Angstadt, Jelsone, Nathan, & Phan, 2006). Further, various networks and pathways in the brain are involved in attention allocation to positive events (Pessoa & Adolphs, 2010; Vuillemier & Huang, 2009). For instance, striate and occipitotemporal extrastriate regions have been associated with attentional bias to babies (Brosch et al., 2008) and with valuedriven attentional capture (Anderson, Laurent, & Yantis, 2014). Areas related to the processing of reward and to the control of actions, thoughts, and attention are also associated with attentional bias to positive information (see Kress & Aue, 2017, for an overview). For instance, activity in the anterior cingulate cortex (ACC) is associated with the magnitude of attentional bias to reward in visual search (Hickey et  al., 2010). Models of drug addiction propose that attentional bias to substance related cues is a consequence of a dopaminergic activity in the dorsal anterior cingulate cortex (dACC) and dorsolateral prefrontal cortex (DLPFC) that raises the salience of those cues (Franken, 2003). For instance, smokers in a placebo condition displayed attentional bias and showed elevated brain activation in the dACC and the right dorsolateral prefrontal cortex (r-DLPFC) in response to smoking cues. By contrast, those who were given a dopamine antagonist (haloperidol) did not demonstrate enhanced brain activation in these regions or an attentional bias (Luijten et  al., 2012). Finally, attentional bias to food in hungry observers is related to stronger functional coupling between the posterior parietal cortex and the posterior cingulate cortex (Mohanty, Gitelman, Small, & Mesulam, 2008). In sum, a variety of brain regions and networks are correlated with attentional bias to positive events. Recent studies have therefore argued that the amygdala appears to coordinate cortical networks and pathways like the prefrontal and visual cortices that convey and modulate information on the significance of the identified emotional stimulus to effectively recruit adaptive responses (Pessoa & Adolphs, 2010). This implies that the brain areas responsible for processing and responding to the emotional value of events (e.g., amygdala, orbitofrontal cortex (OFC), superior colliculus, and pulvinar) interact with brain regions such as the PFC and ACC that coordinate processes like decision-making and the control of attention, thoughts, and actions (Brown & Braver, 2005; Pourtois, Schettino, & Vuilleumier, 2012). Consequently, top-down processes such as goals may influence reactions to emotional events (Vuilleumier & Huang, 2009).

10  Cognitive biases in health and psychiatric disorders

Somatovisceral responses related to the Bias Various somatovisceral responses have been associated with attentional bias to positive input. For instance, attentional biases toward smoking-related stimuli in smokers are strongly associated with enhanced activity in the zygomaticus facial muscles that are responsible for raising the corners of the mouth upward enabling the face to smile (Waters et al., 2003). However, heart rate and skin conductance responses (SCRs) did not increase in response to the smoking cues. The findings suggest that attentional bias in addiction is a reaction to positive attitudes toward and liking of the drug. Inconsistent with the Waters et al.’s (2003) findings, other studies suggest that attentional bias to positive information is linked to increases in SCRs that result from enhanced sweat gland secretions. The production of SCRs is associated with an increased activation of the right medial PFC (Critchley, Melmed, Featherstone, Mathias, & Dolan, 2002). Heightened SCRs indicate heightened arousal independent of valence. They occur in response to both negative and positive events of motivational relevance, which means to stimuli that should be avoided or approached. For instance, a recent study (Gantiva et al., 2018) compared behavioral and physiological responses in an emotional Stroop task between dysphoric and nondysphoric participants. Nondysphoric participants showed an enhanced attentional bias and SCRs toward positive stimuli, whereas participants with dysphoria demonstrated increased attentional bias in response to negative stimuli; the latter also displayed heart rate deceleration that is associated with sustained attention. Relatedly, alcoholics showed larger SCRs and lower heart rate acceleration to alcohol words (Stormark, Laberg, Nordby, & Hugdahl, 2000). In sum, visceral responses that reflect not only liking but also motivational responses such as wanting to approach or avoid stimuli appear to be involved in attentional bias to positive stimuli. Finally, effective cardiac vagal tone regulation that predicts the ability to rapidly alter the cardiac autonomic reactivity has been associated with increased PFC reactivity and improved control of attention and emotional regulation, including adaptive attentional responses to positive stimuli (Porges, 1992; Thayer & Lane, 2000). Supporting this notion, faster and higher relapse rates in alcoholdependent patients were related to increased attentional bias and elevated high frequency heart rate variability in response to alcohol-related cues (Garland, Franken, & Howard, 2012). In sum, various somatovisceral reactions are associated with (dys)functional attentional bias to both pleasant and unpleasant stimuli, and with mechanisms that allow to control the bias.

Similarities and differences between healthy and clinical populations Attention to positive stimuli is significantly lowered or absent in many affective disorders. For instance, depressed individuals show an absence of attentional

Attention to positive information  Chapter | 1  11

capture by stimuli that are associated with reward (Anderson, Leal, Hall, Yassa, & Yantis, 2014). Impaired attention to positive information also characterizes people that are only at risk for depression. For instance, female adolescents who are more vulnerable to the development of depression orient attention away from positive stimuli after a negative mood induction (Joormann, Talbot, & Gotlib, 2007). In contrast, healthy samples show attentional bias toward positive information that seems to serve adaptive emotion regulation. Indeed, when participants were instructed to attend to positive stimuli after a stress induction, they reported to be less frustrated than participants who were instructed to attend to negative images (Johnson, 2009). Interestingly, the capacity to attend to positive or negative stimuli appears to be related to genetic variations. A study by Fox, Ridgewell, and Ashwin (2009) examined attention allocation to negative and positive events and its relation to allelic variations in the promotor region of the serotonin transporter gene (5-HTTLPR). Those with S (short) allele score higher on measures of neuroticism (Lesch et al., 1996). Interestingly, individuals homozygous for an S allele showed increased vigilance toward threatening and negative stimuli, whereas those with an L (long) allele oriented attention toward positive stimuli and shifted away from the negative stimuli. This suggests that serotonin transporter allelic variation that impacts the neuroendocrine system underlies (mal)adaptive emotional processing including attentional bias. However, attention to positive stimuli is not inevitably adaptive (Dodd & Porter, 2010). For instance, attentional bias toward sexual stimuli has been observed in females who had low sexual functioning compared with females who possessed high sexual functioning (Beard & Amir, 2010). This suggests that sexual dysfunctions lead to an oversensitivity toward relevant sexual contents. Dysfunctional bias also characterizes obesity and addictions (Garland et al., 2012; Luijten et al., 2012; Stormark et al., 2000; Waters et al., 2003). For example, both healthy and obese participants direct their gaze toward food-related cues compared with nonfood cues when hungry; however, when satiated, obese participants continued to divert their gaze toward food-related cues in contrast to healthy participants who shifted their gaze away from food-related stimuli (Castellanos et al., 2009). Attentional bias to food found in those with heightened weight-to-height ratio also predicts future weight gain (Yokum, Ng, & Stice, 2011). Similar results have been obtained in addiction. Increases in subjective motivational states such as substance craving are associated with increases in attention allocation to substance related cues that, in turn, are supposed to enhance craving (Franken, 2003). For instance, Field and Eastwood (2005) trained heavy drinkers either to attend or to avoid attending to images of alcohol. To this end, participants were presented with pairs of alcohol and neutral images in a dotprobe task. In the attend alcohol condition, the probes almost always appeared in the location of the alcohol images whereas in the avoid-alcohol condition the probes were in the location of the neutral images. The participants in the attend

12  Cognitive biases in health and psychiatric disorders

alcohol condition showed increased alcohol-related attentional bias compared with their baseline scores, whereas those in the avoid-alcohol condition showed reductions in attentional bias following training. Increases in attentional bias in the alcohol attend group were strongly associated with increases in subjective craving and alcohol consumption as measured by a taste evaluation task. In comparison, those in the avoid-alcohol group did not show any differences in both attentional bias and subjective craving measured at baseline and after training. Subsequently, the avoid-alcohol group consumed significantly less alcohol than the attend alcohol group. This study suggests a causal role of attentional bias in craving and the development of addictions. In sum, both healthy and clinical samples attend to positive stimuli. Whereas healthy samples seem to attend to positive stimuli when it supports adaptive emotion regulation, attention to positive stimuli in clinical samples seems to characterize dysfunctional processes such as craving (Field et  al., 2016; Shechner & Bar-Haim, 2016). However, anxious individuals allocate attention to positive events relevant to a current goal even in the presence of threatening information (Vogt et al., 2013). Making positive events goal relevant thus seems to be a tool to induce attentional bias to positive information even in clinical samples.

Limitations and future directions Though various studies have shown that positive stimuli attract attention, several questions remain unsolved. For instance, it is unclear whether all kinds of positive stimuli evoke attentional bias in similar ways, for example, whether they differ in their capacity to capture or hold attention. Additionally, it remains to be clarified to what extent the bias varies depending on individual differences. We hope that future studies will compare a variety of stimuli in the same experimental design while also measuring individual differences and testing clinical populations. Further, combining behavioral measures with neuroscientific and psychophysiological methods will help to understand which mechanism(s) underlie the bias. Importantly, highlighting the role of contextual factors will allow the field to shift the research focus from asking whether positive stimuli bias attention automatically to why and when people attend to them. Attention to positive stimuli varies not only across individuals but also across situations. For instance, dieters or addicts attend to high-caloric food or drugs when craving but are inattentive to them when goals to abstain are activated (Field et  al., 2016; Papies et  al., 2008). We hope that future research will continue to study how contextual factors such as temporary goals or expectations (Kress & Aue, 2017) prevent and induce (dys)functional attentional bias to positive information. Taking context into account will also help to clarify under which circumstances attention to positive stimuli is adaptive. For instance, attention to positive stimuli might be adaptive when it serves, for instance, an emotion regulation

Attention to positive information  Chapter | 1  13

goal. In contrast, if it enhances craving or turns people blind to relevant negative information such as error feedback or signs of real threat, it will be maladaptive. We believe that combining attentional measures with behavioral outcomes or indicators of relevant regulatory processes will allow researchers to gain insight into when attention to positive information is (dys)functional. Importantly, if attentional bias is context dependent, then attempts to train observers to acquire an attentional bias toward or away from positive stimuli must take relevant contextual factors into account. Only then will these trainings be efficiently transferrable to relevant real-life situations. For instance, it might be necessary to train people to attend toward (or away from) positive stimuli in response to the specific situation that usually evokes a dysfunctional attentional bias (cf. Salemink, Woud, Roos, Wiers, & Lindgren, 2019).

Summary The present chapter outlined when and why positive information attracts attention in both healthy and clinical samples, how attention can be measured, and which neuroscientific and psychophysiological measures reflect it. We suggested that attention to positive stimuli is highly context dependent with temporary goals and subsequent top-down processes being one factor that not only causes but also erases attention to positive information. Ultimately, this line of work can help to improve the prevention and treatment of psychiatric disorders, for instance, by better tailoring attentional trainings with respect to individualistic eliciting (i.e., contextual) factors.

References Aarts, H., & Dijksterhuis, A. (2003). On wildebeests and humans: The preferential detection of negative stimuli. Psychological Science, 14, 14–18. Algom, D., Chajut, E., & Lev, S. (2004). A rational look at the emotional Stroop phenomenon: A generic slowdown, not a Stroop effect. Journal of Experimental Psychology: General, 133, 323–338. Anderson, B. A., Laurent, P. A., & Yantis, S. (2014). Value-driven attentional priority signals in human basal ganglia and visual cortex. Brain Research, 1587, 88–96. Anderson, B. A., Leal, S. L., Hall, M. G., Yassa, M. A., & Yantis, S. (2014). The attribution of value-based attentional priority in individuals with depressive symptoms. Cognitive, Affective, & Behavioral Neuroscience, 14, 1221–1227. Bar-Haim, Y., Lamy, D., Pergamin, L., Bakermans-Kranenburg, M. J., & van IJzendoorn, M. H. (2007). Threat-related attentional bias in anxious and nonanxious individuals: A meta-analytic study. Psychological Bulletin, 133, 1–24. Beard, C., & Amir, N. (2010). Attention bias for sexual words in females with low sexual functioning: A brief report. US National Library of Medicine, 36, 216–226. Becker, D. V., Anderson, U. S., Mortensen, C. R., Neufeld, S. L., & Neel, R. (2011). The face in the crowd effect unconfounded: Happy faces, not angry faces, are more efficiently detected in single- and multiple-target visual search tasks. Journal of Experimental Psychology: General, 140, 637–659.

14  Cognitive biases in health and psychiatric disorders Blanchette, I. (2006). Snakes, spiders, guns, and syringes: How specific are evolutionary constraints on the detection of threatening stimuli? Quarterly Journal of Experimental Psychology, 59, 1484–1504. Brosch, T., Sander, D., Pourtois, G., & Scherer, K. R. (2008). Beyond fear: Rapid spatial orienting toward positive emotional stimuli. Psychological Science, 19, 362–370. Brosch, T., Sander, D., & Scherer, K. R. (2007). That baby caught my eye... Attention capture by infant faces. Emotion, 7, 685–689. Brown, J. W., & Braver, T. S. (2005). Learned predictions of error likelihood in the anterior cingulate cortex. Science, 307, 1118–1121. Brown, C. R. H., Duka, T., & Forster, S. (2018). Attentional capture by alcohol related stimuli may be activated involuntarily by top-down search goals. Psychopharmacology, 235, 2087–2099. Calitri, R., Lowe, R., Eves, F. F., & Bennett, P. (2009). Associations between visual attention, implicit and explicit attitude and behaviour for physical activity. Psychology & Health, 24, 1105–1123. Castellanos, E. H., Charboneau, E. J., Dietrich, M. S., Park, S., Bradley, B. P., Mogg, K., et  al. (2009). Obese adults have visual attention bias for food cue images: Evidence for altered reward system function. International Journal of Obesity, 33, 1063–1073. Critchley, H. D., Melmed, R. N., Featherstone, E., Mathias, C. J., & Dolan, R. J. (2002). Volitional control of autonomic arousal: A functional magnetic resonance study. NeuroImage, 16, 909–919. Dalgleish, T. (1995). Performance on the emotional Stroop task in groups of anxious, expert, and control subjects: A comparison of computer and card presentation formats. Cognition and Emotion, 9, 341–362. Desimone, R., & Duncan, J. (1995). Neural mechanisms of selective visual attention. Annual Review of Neuroscience, 18, 193–222. Dewitte, M., De Houwer, J., Koster, E. H. W., & Buysse, A. (2007). What’s in a name? Attachmentrelated attentional bias. Emotion, 7, 535–545. Dodd, H. F., & Porter, M. A. (2010). I see happy people: Attention bias towards happy but not angry facial expressions in Williams syndrome. Cognitive Neuropsychiatry, 15, 549–567. Fenske, M. J., & Eastwood, J. D. (2003). Modulation of focused attention by faces expressing emotion: Evidence from flanker tasks. Emotion, 3, 327–343. Field, M., & Eastwood, B. (2005). Experimental manipulation of attentional bias increases the motivation to drink alcohol. Psychopharmacology, 183, 350–357. Field, M., Werthmann, J., Franken, I., Hofmann, W., Hogarth, L., & Roefs, A. (2016). The role of attentional bias in obesity and addiction. Health Psychology, 35, 767–780. Fitzgerald, D. A., Angstadt, M., Jelsone, L. M., Nathan, P. J., & Phan, K. L. (2006). Beyond threat: Amygdala reactivity across multiple expressions of facial affect. NeuroImage, 30, 1441–1448. Fox, E., Ridgewell, A., & Ashwin, C. (2009). Looking on the bright side: Biased attention and the human serotonin transporter gene. Proceedings of the Royal Society B: Biological Sciences, 276, 1747–1751. Fox, E., Russo, R., Bowles, R. J., & Dutton, K. (2001). Do threatening stimuli draw or hold visual attention in sub-clinical anxiety? Journal of Experimental Psychology: General, 130, 681–700. Franken, I. H. (2003). Drug craving and addiction: Integrating psychological and neuropsychopharmacological approaches. Progress in Neuro-Psychopharmacology & Biological Psychiatry, 27, 563–579. Gantiva, C., Araujo, A., Aragão, N., & Hewitt, N. (2018). Modulation of physiological responses as indices of attentional bias in dysphoria. International Journal of Mental Health and Addiction, 16, 328–338.

Attention to positive information  Chapter | 1  15 Garland, E. L., Franken, I. H. A., & Howard, M. O. (2012). Cue-elicited heart rate variability and attentional bias predict alcohol relapse following treatment. Psychopharmacology, 222, 17–26. Gronau, N., Cohen, A., & Ben-Shakhar, G. (2003). Dissociations of personally significant and taskrelevant distractors inside and outside the focus of attention: A combined behavioral and psychophysiological study. Journal of Experimental Psychology: General, 132, 512–529. Hickey, C., Chelazzi, L., & Theeuwes, J. (2010). Reward guides vision when it’s your thing: Trait reward-seeking in reward-mediated visual priming. PLoS One, 5, e14087. Horstmann, G., Borgstedt, K., & Heumann, M. (2006). Flanker effects with faces may depend on perceptual as well as emotional differences. Emotion, 6, 28–39. Johnson, D. R. (2009). Goal-directed attentional deployment to emotional faces and individual differences in emotional regulation. Journal of Research in Personality, 43, 8–13. Joormann, J., Talbot, L., & Gotlib, I. H. (2007). Biased processing of emotional information in girls at risk for depression. Journal of Abnormal Psychology, 116, 135–143. Koster, E. H. W., Crombez, G., Van Damme, S., Verschuere, B., & De Houwer, J. (2004). Does imminent threat capture and hold attention? Emotion, 4, 312–317. Kress, L., & Aue, T. (2017). The link between optimism bias and attention bias: A neurocognitive perspective. Neuroscience and Biobehavioral Reviews, 80, 688–702. Kress, L., Bristle, M., & Aue, T. (2018). Seeing through rose-colored glasses: How optimistic expectancies guide visual attention. PLoS One, 13, e0193311. Kuhn, G., Pickering, A., & Cole, G. G. (2016). “Rare” emotive faces and attentional orienting. Emotion, 16, 1–5. Lang, P. J., Bradley, M. M., & Cuthbert, B. (1997). Motivated attention: Affect, activation and action. In P. J. Lang, R. F. Simons, & M. T. Balaban (Eds.), Attention and orienting: Sensory and motivational processes (pp. 97–135). Hillsdale, NJ: Lawrence Erlbaum Associates, Inc. Le Pelley, M. E., Mitchell, C. J., Beesley, T., George, D. N., & Wills, A. J. (2016). Attention and associative learning in humans: An integrative review. Psychological Bulletin, 142, 1111–1140. LeDoux, J. E. (1996). The emotional brain: The mysterious underpinnings of emotional life. New York, NY: Simon & Schuster. Lesch, K., Bengel, D., Heils, A., Sabol, S., Greenberg, B., Petri, S., et al. (1996). Association of anxiety-related traits with a polymorphism in the serotonin transporter gene regulatory region. Science, 274, 1527–1531. Lichtenstein-Vidne, L., Henik, A., & Safadi, Z. (2012). Task relevance modulates processing of distracting emotional stimuli. Cognition and Emotion, 26, 42–52. Light, J., & Isaacowitz, D. (2006). The effect of developmental regulation on visual attention: The example of the “biological clock”. Cognition and Emotion, 20, 623–645. Lipp, O. V., & Derakshan, N. (2005). Attentional bias to pictures of fear-relevant animals in a dot probe task. Emotion, 5, 365–369. LoBue, V. (2010). What’s so scary about needles and knives? Examining the role of experience in threat detection. Cognition and Emotion, 24, 180–187. Luck, S. J., & Vecera, S. P. (2002). Attention. [Series Ed.). (Volume Ed.]In Pashler, H. & Yantis, S. (Eds.), Stevens’ Handbook of Experimental Psychology: Vol. 1. Sensation and Perception. (3rd ed.)(pp. 235–286). New York: Wiley. Luijten, M., Veltman, D. J., Hester, R., Smits, M., Pepplinkhuizen, L., & Franken, I. H. (2012). Brain activation associated with attentional bias in smokers is modulated by a dopamine antagonist. Neuropsychopharmacology, 37, 2772–2779. MacLeod, C., Mathews, A., & Tata, P. (1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology, 95, 15–20.

16  Cognitive biases in health and psychiatric disorders Maner, J. K., Gailliot, M. T., & Miller, S. L. (2009). The implicit cognition of relationship maintenance: Inattention to attractive alternatives. Journal of Experimental Social Psychology, 45, 174–179. Maner, J., Kenrick, D., Becker, D., Delton, A., Hofer, B., Wilbur, C., et al. (2003). Sexually selective cognition: Beauty captures the mind of the beholder. Journal of Personality and Social Psychology, 85, 1107–1120. Mogg, K., Bradley, B. P., Field, M., & De Houwer, J. (2003). Eye movements to smoking-related pictures in smokers: Relationship between attentional biases and implicit and explicit measures of stimulus valence. Addiction, 98, 825–836. Mogg, K., Holmes, A., Garner, M., & Bradley, B. P. (2008). Effects of threat cues on attentional shifting, disengagement and response slowing in anxious individuals. Behaviour Research and Therapy, 46, 656–667. Mohanty, A., Gitelman, D. R., Small, D. M., & Mesulam, M. M. (2008). The spatial attention network interacts with limbic and monoaminergic systems to modulate motivation-induced attention shifts. Cerebral Cortex, 18, 2604–2613. Moors, A., & De Houwer, J. (2006). Automaticity: A theoretical and conceptual analysis. Psychological Bulletin, 132, 297–326. Moskowitz, G. B. (2002). Preconscious effects of temporary goals on attention. Journal of Experimental Social Psychology, 38, 397–404. Most, S. B., Smith, S. D., Cooter, A. B., Levy, B. N., & Zald, D. H. (2007). The naked truth: Positive, arousing distractors impair rapid target perception. Cognition & Emotion, 21, 964–981. Neuberg, S. L., Kenrick, D. T., Maner, J. K., & Schaller, M. (2004). From evolved motives to everyday mentation: Evolution, goals, and cognition. In J. Forgas & K. Williams (Eds.), Social motivation: Conscious and unconscious processes (pp. 133–152). New York: Cambridge University Press. Öhman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108, 483–522. Papies, E. K., Stroebe, W., & Aarts, H. (2008). The allure of forbidden food: On the role of attention in self-regulation. Journal of Experimental Social Psychology, 44, 1283–1292. Pessoa, L., & Adolphs, R. (2010). Emotion processing and the amygdala: From a “low road” to “many roads” of evaluating biological significance. Nature Reviews Neuroscience, 11, 773–783. Pool, E., Brosch, T., Delplanque, S., & Sander, D. (2016). Attentional bias for positive emotional stimuli: A meta-analytic investigation. Psychological Bulletin, 142, 79–106. Porges, S. (1992). Vagal tone: A physiologic marker of stress vulnerability. Pediatrics, 90, 498–504. Posner, M. I. (1980). Orienting of attention. The Quarterly Journal of Experimental Psychology, 32, 3–25. Pourtois, G., Schettino, A., & Vuilleumier, P. (2012). Brain mechanisms for emotional influences on perception and attention: What is magic and what is not. Biological Psychology, 92, 492–512. Pratto, F., & John, O. P. (1991). Automatic vigilance: The attention-grabbing power of negative social information. Journal of Personality and Social Psychology, 61, 380–391. Raymond, J. E., & O’Brien, J. L. (2009). Selective visual attention and motivation: The consequences of value learning in an attentional blink task. Psychological Science, 20, 981–988. Roskos-Ewoldsen, D. R., & Fazio, R. H. (1992). On the orienting value of attitudes: Attitude accessibility as a determinant of an object’s attraction of visual attention. Journal of Personality and Social Psychology, 63, 198–211. Salemink, E., Woud, M. L., Roos, M., Wiers, R., & Lindgren, K. P. (2019). Reducing alcoholrelated interpretive bias in negative affect situations: Using a scenario-based cognitive Bias modification training paradigm. Addictive Behaviors, 88, 106–113.

Attention to positive information  Chapter | 1  17 Schimmack, U. (2005). Attentional interference effects of emotional pictures: Threat, negativity, or arousal? Emotion, 5, 55–66. Segerstrom, S. C. (2001). Optimism and attentional bias for negative and positive stimuli. Personality and Social Psychology Bulletin, 27, 1334–1343. Sennwald, V., Pool, E., Brosch, T., Delplanque, S., Bianchi‐Demicheli, F., & Sander, D. (2016). Emotional attention for erotic stimuli: Cognitive and brain mechanisms. The Journal of Comparative Neurology, 524, 1668–1675. Shechner, T., & Bar-Haim, Y. (2016). Threat monitoring and attention bias modification in anxiety and stress-related disorders. Current Directions in Psychological Science, 25, 431–437. Smith, N. K., Larsen, J. T., Chartrand, T. L., Cacioppo, J. T., Katafiasz, H. A., & Moran, K. E. (2006). Being bad isn’t always that good: Affective context moderates the attention bias toward negative information. Journal of Personality and Social Psychology, 90, 210–220. Stormark, K. M., Laberg, J. C., Nordby, H., & Hugdahl, K. (2000). Alcoholics’ selective attention to alcohol stimuli: Automated processing? Journal of Studies on Alcohol, 61, 18–23. Talsma, D., Senkowski, D., Soto-Faraco, S., & Woldorff, M. G. (2010). The multifaceted interplay between attention and multisensory integration. Trends in Cognitive Sciences, 14, 400–410. Tamietto, M., & de Gelder, B. (2010). Neural bases of the non-conscious perception of emotional signals. Nature Reviews Neuroscience, 11, 697–709. Tapper, K., Pothos, E. M., & Lawrence, A. D. (2010). Feast your eyes: Hunger and trait reward drive predict attentional bias for food cues. Emotion, 10, 949–954. Thayer, J. F., & Lane, R. D. (2000). A model of neurovisceral integration in emotion regulation and dysregulation. Journal of Affective Disorders, 61, 201–216. Van Dessel, P., & Vogt, J. (2012). When does hearing laughter draw attention to happy faces? Task relevance determines the influence of a cross-modal affective context on emotional attention. Frontiers in Human Neuroscience, 6, 294. Vogt, J., & De Houwer, J. (2014). Emotion regulation meets emotional attention: The influence of emotion suppression on emotional attention depends on the nature of the distracters. Emotion, 14, 840–845. Vogt, J., De Houwer, J., & Crombez, G. (2011). Multiple goal management starts with attention: goal prioritizing affects the allocation of spatial attention to goal relevant events. Experimental Psychology, 58, 55–61. Vogt, J., De Houwer, J., Koster, E. H. W., Van Damme, S., & Crombez, G. (2008). Allocation of spatial attention to emotional stimuli depends upon arousal and not valence. Emotion, 8, 880–885. Vogt, J., De Houwer, J., & Moors, A. (2011). Unintended allocation of spatial attention to goal relevant but not to goal related events. Social Psychology, 42, 48–55. Vogt, J., De Hower, J., Moors, A., Van Damme, S., & Crombez, G. (2010). The automatic orienting of attention to goal-relevant stimuli. Acta Psychologica, 134, 61–69. Vogt, J., Houwer, J., Crombez, G., & Van Damme, S. (2013). Competing for attentional priority: Temporary goals versus threats. Emotion, 13, 587–598. Vogt, J., Koster, E. H. W., & De Houwer, J. (2017). Safety first: Instrumentality for reaching safety determines attention allocation under threat. Emotion, 17, 528–537. Vogt, J., Lozo, L., Koster, E. H. W., & De Houwer, J. (2011). On the role of goal relevance in emotional attention: Disgust evokes early attention to cleanliness. Cognition and Emotion, 25, 466–477. Vuilleumier, P., & Huang, Y.-M. (2009). Emotional attention: Uncovering the mechanisms of affective biases in perception. Current Directions in Psychological Science, 18, 148–152. Vuilleumier, P., Richardson, M. P., Armony, J. L., Driver, J., & Dolan, R. J. (2004). Distant influences of amygdala lesion on visual cortical activation during emotional face processing. Nature Neuroscience, 7, 1271–1278.

18  Cognitive biases in health and psychiatric disorders Waters, A. J., Shiffman, S., Sayette, M. A., Paty, J. A., Gwaltney, C. J., & Balabanis, M. H. (2003). Attentional bias predicts outcome in smoking cessation. Health Psychology, 22, 378–387. Wieber, F., & Sassenberg, K. (2006). I can’t take my eyes off of it—attention attraction effects of implementation intentions. Social Cognition, 24, 723–752. Williams, J. M. G., Mathews, A., & MacLeod, C. (1996). The emotional Stroop task and psychopathology. Psychological Bulletin, 120, 3–24. Xing, C., & Isaacowitz, D. M. (2006). Aiming at happiness: How motivation affects attention to and memory for emotional images. Motivation and Emotion, 30, 249–256. Yiend, J. (2010). The effects of emotion on attention: A review of attentional processing of emotional information. Cognition and Emotion, 24, 3–47. Yokum, S., Ng, J., & Stice, E. (2011). Attentional bias to food images associated with elevated weight and future weight gain: An fMRI study. Obesity, 19, 1775–1783.

Chapter 2

Attention bias toward negative stimuli Elinor Abadoa,b, Thalia Richtera,b, Hadas Okon-Singera,b a

Department of Psychology, University of Haifa, Haifa, Israel, bThe Integrated Brain and Behavior Research Center (IBBR), University of Haifa, Haifa, Israel

Introduction Attention bias occurs when an individual exhibits abnormal orienting of attention toward negative stimuli, compared with neutral stimuli. Such a bias can be manifested both in covert and in overt attention by heightened vigilance to threat, difficulty in disengaging attention from threat, and/or avoidance of threatening stimuli (for review, see Aue & Okon-Singer, 2015). For instance, individuals who suffer from spider phobia present attention bias while seeing a spider on a computer screen, as they immediately orient their attention toward the spider and they are unable to rapidly withdraw attention from the spider, as they fixate on it and on its location. In addition, during the later stages of processing, phobic participants avoid looking at the spider. Attention bias can also occur toward positive stimuli (for further elaboration, see Chapter 1 on attention bias toward positive stimuli). The present chapter will focus on threat-related attention bias, and it will simply be referred to as “attention bias.” Attention bias can be found in various populations, such as individuals suffering from depression (for meta-analysis, see Peckham, McHugh, & Otto, 2010), attention-deficit hyperactivity disorder (ADHD; e.g., Hommer et  al., 2014) and posttraumatic stress disorder (PTSD; for review, see Lazarov et al., 2018). It is important to note that the content of attention bias differs between disorders (e.g., Hankin, Gibb, Abela, & Flory, 2010; Williams, Watts, MacLeod, & Mathews, 1997). Individual differences can also mediate and regulate the bias (for review, see Cisler & Koster, 2010). However, most research on attention bias focuses on populations with anxiety disorders. It is a characteristic of many anxiety disorders, such as generalized anxiety disorder, social anxiety disorder, and specific phobias, regardless of age or severity of symptoms (for meta-analysis, see Bar-Haim, Lamy, Pergamin, Bakermans-Kranenburg, & Van Ijzendoorn, 2007). Attention bias is believed to play a significant role in the etiology and maintenance of Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00002-7 © 2020 Elsevier Inc. All rights reserved.

19

20  Cognitive biases in health and psychiatric disorders

anxiety disorders (for reviews, see Mathews & MacLeod, 2002; Okon-Singer, 2018; Van Bockstaele et al., 2014; Williams, Watts, MacLeod, & Matthews, 1988). While this bias is also found among healthy individuals, it is more pronounced, severe, and persistent among anxious individuals (Aue, Guex, Chauvigné, & Okon-Singer, 2013; for review, see Aue & Okon-Singer, 2015). Therefore, research on attention bias typically involves anxious populations. In the present chapter, we use the term “anxious” to refer to both clinically anxious populations (i.e., individuals who have been diagnosed according to DSM criteria) and subclinical populations (i.e., individuals who have not been diagnosed with an anxiety disorder but were evaluated using questionnaires or other measures). Cognitive biases in general and attention bias in particular differ among disorders, as biases reflect abnormal processing of certain disorder-related stimuli. Two highly related disorders are anxiety and depression. According to Beck (Beck, 1967, 1976; Beck, Emery, & Greenberg, 1985), both disorders are characterized by a negative triad: a negative view of the self, of the world, and of the future. These views develop from childhood via negative experiences. As these beliefs are internalized, they become schemas: an unconscious set of beliefs that affects the way we interpret certain situations, usually the same situations that have caused these negative views to develop. Cognitive biases (i.e., abnormally negative processing of certain stimuli; Kendall & Ingram, 1989) are formed through these negative schemas. Thus, the difference between anxiety and depression lies within the content of the bias (e.g., anxiety is more characterized by uncertainty, while depression is more characterized by loss or failure; for review, see Richter, Aue, Richter-Levin, & Okon-Singer, under review). It is important to note that, in terms of attention bias, there are several differences between participants with anxiety and participants with depression. First, participants with anxiety usually display attention bias toward threatening stimuli that are relevant to their specific anxiety disorder (e.g., spider phobia and spiders, social anxiety, and social situations), while participants with depression exhibit bias toward self-relevant stimuli (e.g., participants with dysphoria and adjectives such as “useless”; Koster, De Raedt, Goeleven, Franck, & Crombez, 2005). In addition, as will be discussed extensively throughout the chapter, different components of attention bias are exhibited in different disorders. In addition to theoretical importance, these differences have clinical importance as they can help differentiating between anxiety and depression. In the current chapter, we will introduce attention bias across several disorders, the main theories revolving around the bias and the primary methods used to study it. Furthermore, we will discuss brain regions and peripheral responses that are involved in the emergence of attention bias and autonomic responses related to it. We will also refer to similarities and differences between healthy and clinical populations. Lastly, we will discuss the limitations of current literature, methods, theories, and practices and offer future directions for further research.

Attention bias toward negative stimuli  Chapter | 2  21

Major theories on the nature and underlying causes of attention biases Attention bias has been researched for decades. Thus, many recent reviews and meta-analyses draw conclusions and offer models regarding the nature of attention bias, based on the vast amount of research accumulated over the years. In this section, we will present some of the prominent models and theories. It was suggested that attention bias includes three main components: faster engagement with feared stimuli compared with neutral stimuli, slower disengagement from feared stimuli compared with neutral stimuli, and avoidance. Following this initial vigilance, whereby fearful participants rapidly notice fearful stimuli, an avoidance component is exhibited, whereby participants avoid the same fearful stimuli (e.g., Koster, Crombez, Verschuere, Van Damme, & Wiersema, 2006; Mogg, Bradley, Miles, & Dixon, 2004; Williams et al., 1988). These feared stimuli are typically threat related (e.g., spiders, snakes, blood, and angry faces). The fast engagement component is characterized by rapid engagement with threatening stimuli, compared with neutral stimuli. Typically, anxious individuals are more easily, or more quickly, drawn to threatening stimuli, compared with neutral stimuli and with nonanxious individuals (for review, see Cisler & Koster, 2010). The second component of attention bias is slower disengagement from threatening stimuli, compared with neutral stimuli. Anxious individuals exhibit difficulty in disengagement, meaning that their attention is strongly captured by threatening stimuli compared with neutral stimuli and with nonanxious individuals. Difficulty in disengagement hinders the ability to relocate attention to new stimuli (for review, see Cisler & Koster, 2010). Lastly, attention bias is characterized by a vigilance-avoidance pattern, in which anxious participants initially exhibit hypervigilance toward the threatening stimulus, as reviewed earlier, but then avoid it at later stages of processing (Mogg et al., 2004). Anxious participants avoid threatening stimuli by allocating their attention toward nonthreatening stimuli. In other words, while participants outwardly avoid looking at the fearful stimulus, their attentional resources inwardly remain focused on the stimulus. In their review on mechanisms of attention biases toward threat in anxiety disorders, Cisler and Koster (2010) argue that malfunctions in the threat detection mechanism, in attention control abilities, and in emotion regulation strategies can lead to facilitated engagement, difficulty in disengagement, and attentional avoidance, respectively. Further, they claim that threat detection is an automatic process (i.e., does not require intent, control, or awareness) and that attention avoidance is a strategic process (i.e., requires intent, control, and awareness), while difficulty in disengagement is both automatic and strategic (see also Shiffrin & Schneider, 1977, for the differentiation between automatic and strategic processing). For reviews on more models of attentional biases toward threat in anxiety disorders, see Cisler and Koster (2010), Van Bockstaele

22  Cognitive biases in health and psychiatric disorders

et al. (2014), and Yiend (2010). In their meta-analysis on attention bias in anxious and nonanxious individuals, Bar-Haim et  al. (2007) offer an integrative model. They argue that there are four stages of processing in anxiety disorders and that different disorders could result from malfunctioning on different stages or combinations thereof. The first stage involves a preattentive threat evaluation system as it preattentively (i.e., does not require attentional resources) evaluates stimuli in terms of threat (low to high threat). As stage is an unconscious one, in case of anxiety, mildly threatening stimuli could be automatically perceived as highly threatening stimuli. In the second stage, a resource allocation system is involved. If stimuli were evaluated as highly threatening in the first stage, in this stage, resources are allocated to threatening stimuli, and one becomes conscious and alert of the threat. In case of anxiety, attentional resources are allocated to the mildly threatening stimulus. In the third stage, a guided threat evaluation system is involved. A conscious evaluation occurs, in which prior knowledge regarding the nature of the stimulus is taken into account. In case of anxiety, one consciously evaluates the stimulus as highly threatening, even if prior knowledge indicates otherwise. Lastly, a goal engagement system is involved: if the stimulus is indeed evaluated as threatening, the alert state continues. However, if the stimulus is not evaluated as a threat, then one relaxes and no longer feels alert. In case of anxiety, the latter outcome of the evaluation does not ease the anxiety, as conscious understanding of the situation and of the illogical evaluation of it does not lead to the termination of the anxious state. The second stage of this model (i.e., resource allocation system) implies causality between interpretation bias (i.e., the interpretation of ambiguous stimuli as threatening ones; Becker & Rinck, 2004) and allocation of attention resources. Along these lines, in her review on attention processing of attentional information, Yiend (2010) suggests that anxious and nonanxious individuals have a different “evaluative system,” which leads to anxious individuals’ oversensitivity to threatening stimuli. A more integrative view was offered by Everaert, Duyck, and Koster (2014): their view suggests causal interactions between several biases: attention, memory, and interpretation biases. Specifically, they suggest that there is a covariance between attention bias and interpretation bias, which in turn affects memory bias. Note, however, that these findings were found among participants with subclinical depression and the sample did not include other disorders. In their review on the causal relationship between attention bias and fear/ anxiety, Van Bockstaele et al. (2014) argue that there is a causal, circular, and mutual relationship between the two factors. In other words, attention bias and fear/anxiety feed and maintain each other, resulting in a never-ending cycle. This approach has great therapeutic relevance, in the sense that this cycle can be broken if attention bias is treated properly, resulting in reduction of fear and anxiety levels. Similar claims were made by Yiend (2010); according to her approach, treating anxious individuals’ attentional oversensitivity to threatening stimuli is key to ease the negative effects of the bias.

Attention bias toward negative stimuli  Chapter | 2  23

As for the reason for the existence of attention biases, the biological preparedness hypothesis claims that humans are predisposed to fear certain “biologically relevant” stimuli (i.e., phylogenetic stimuli such as blood, heights, and spiders) and that this fear is necessary for survival (Seligman, 1971). Along these lines, Öhman (1996) refers to attention bias as a way to survive and to avoid harmful stimuli. According to Öhman, humans are prepared to notice certain biologically relevant features and thus avoid potentially threatening animals. The process leading to behavioral avoidance of certain stimuli begins as unconscious processing of threat. Then, relevant stimuli become integrated in a conscious processing system that includes expectancy and prior learning, which can have great impact on anxiety (for review on attention-expectancy interactions, see Aue & Okon-Singer, 2015). While Seligman’s evolution-based theory is highly prominent, there remain mixed findings in the literature. Several studies did not find differences in attention allocated to phylogenetic (ancient threat, e.g., spiders and snakes) versus ontogenetic (recent threat, e.g., guns and knives) stimuli and offer alternative factors that could play a role in fear. These include spatiotemporal unpredictability (Merckelbach, Van den Hout, Jansen, & van der Molen, 1988) and fear relevance (i.e., potential danger imposed by a stimulus, regardless of whether it is phylogenetic or not; e.g., Brosch & Sharma, 2005).

Prevalent paradigms and findings in attention bias Attention bias has been researched extensively using many different stimuli and populations. Mainly, comparisons are made between negative (i.e., threatening to sadness evoking) and neutral stimuli. Types of stimuli usually studied include visual stimuli or lexical stimuli. For instance, in the case of anxiety disorders, attention bias has been exhibited toward pictures of mutilated faces (e.g., Koster, Crombez, Verschuere, & De Houwer, 2004), pictures of phylogenetic stimuli (i.e., snakes and spiders), pictures of ontogenetic stimuli (i.e., guns and syringes; e.g., Brosch & Sharma, 2005), and pictures of angry faces (e.g., Brosch, Sander, Pourtois, & Scherer, 2008). Similarly, attention bias was displayed toward words related to physical or social threat (e.g., Mogg, Bradley, De Bono, & Painter, 1997). In a meta-analysis comparing attention bias between anxious and nonanxious individuals, it was found that both visual and lexical stimuli produce attention bias in anxious individuals but not in nonanxious individuals, with no difference between the two types of stimuli in terms of combined effect sizes across studies included in the meta-analysis (see Bar-Haim et al., 2007). In the case of depressed participants, stimuli used to reveal attention bias may differ from those used in anxious participants. For instance, in depressed participants, verbal stimuli are self-relevant. In a study that found attentional differences between nondysphoric and dysphoric participants, positive stimuli included words like “strong,” “friendly,” and “optimistic,” while negative stimuli included words like “rejected,” “desperate,” and “useless” (Koster et al., 2005).

24  Cognitive biases in health and psychiatric disorders

The authors claim that this effect is “mood congruent,” as these negative adjectives reflect the way participants with dysphoria feel. Similarly, different facial expressions are used in the study of attention bias in anxiety versus depression. For instance, depressed participants exhibit attention bias toward depressionrelevant faces (e.g., sad facial expressions), while anxious participants exhibit bias toward threat-relevant faces (e.g., angry faces; Hankin et al., 2010). Attention bias can be measured using reaction time (RT) and accuracy rates when comparing emotional and neutral stimuli. Several tasks have been commonly employed to assess attention biases in subclinical and clinical populations (see also elaboration in Chapter  1 in this book). These include the dot-probe task, in which two cue stimuli, one emotional and one neutral, are simultaneously presented. When these stimuli disappear, a probe appears in one of the two previously occupied locations. Participants are asked to indicate which location has been replaced by a probe (see Fig. 1A). Typically, anxious participants are faster to respond when the probe appears in the same location as the threatening

+

Time

Sad

Chair

Time

:

(A)

(B)

Anger

(C)

(D)

FIG. 1  Prevalent paradigms used to study attention bias. (A) An example of a congruent trial on the dot-probe paradigm. (B) An example of a valid trial on the emotional variation of the spatial cuing paradigm. (C) An example of the emotional variation of the Stroop task (threatening word). (D) An example of a visual search array that includes eight neutral distractors (butterflies) and one threatening deviant (spider; third row, middle column). The picture for (B) was taken from Open Affective Standardized Image Set (OASIS; Kurdi, Lozano, & Banaji, 2017), while pictures for (D) were taken from “google images.” These pictures were labeled for reuse under “usage rights.”

Attention bias toward negative stimuli  Chapter | 2  25

stimulus, compared with when it appears in the location of the neutral stimulus. Thus, this paradigm allows for the measurement of spatial attention allocation. For instance, anxious participants responded faster to a dot probe when it appeared in the location of a socially or physically threatening word, compared with when it appeared in the location of a neutral word. This effect was not significant in the control group (MacLeod, Mathews & Tata, 1986). The dotprobe task is also used to study attention bias in depressed populations, and it shows significant differences between depressed and nondepressed groups: a medium effect was found for the relation between depression and attention bias for negative stimuli, and a smaller effect was found for attention bias away from positive stimuli. This latter effect was larger in depressed participants compared with nondepressed participants (for a meta-analysis, see Peckham et al., 2010). Another common method is the emotional modification of the spatial cuing paradigm, in which an emotional or a neutral cue appears, followed by a target in the same or in a different location. If the cue and the target appeared in the same location, the cue is valid. If the cue and the target appeared in different locations, the cue is invalid (see Fig.  1B). Typically, anxious participants respond faster to valid threat-cued trials, compared with valid neutral-cued trials. Similarly, they respond slower to invalid threat-cued trials, compared with invalid neutral-cued trials. For instance, following a threatening cue (threat words or angry faces), anxious participants responded slower when the target appeared in the different location, compared with conditions in which the cue was positive or neutral (Fox, Russo, Bowles, & Dutton, 2001). This method allows for the differentiation between engagement and disengagement components (e.g., Posner, 1980). On the emotional Stroop task, participants are required to name the ink color of a word while ignoring its semantic content, which is either threat-related or neutral (see Fig. 1C). For anxious participants, it is harder to disregard the negative semantic content, and as a result, it takes them longer to respond, compared with healthy participants. For instance, in a study by Mathews and MacLeod (1985), the ability of anxious participants to name the color of the words presented was slower than control participants but even more so when the words were related to physical or social threat. Control participants took the same amount of time to react to threatening and nonthreatening words (Mathews & MacLeod, 1985). These results were replicated and extended to various threatening stimuli and anxiety disorders (for review, see Teachman, Joormann, Steinman, & Gotlib, 2012). The emotional Stroop task is also used to study attention bias in depressed participants, and it shows marginally significant differences between depressed and nondepressed participants. This effect is most pronounced when the Stroop task includes self-relevant words (for a meta-­analysis, see Peckham et al., 2010). Lastly, attention bias can be studied using an emotional variation of the visual search task, in which participants search for an emotional target among neutral distractors, or vice versa (see Fig. 1D). Typically, anxious participants’

26  Cognitive biases in health and psychiatric disorders

reactions to threatening stimuli among neutral stimuli are faster, compared with neutral stimuli among neutral stimuli and to healthy participants’ reactions to the same stimuli. Similarly, anxious participants’ reactions to neutral stimuli among threatening stimuli are typically slower, compared with neutral stimuli among neutral stimuli. For instance, when participants were required to look for a threatening target (i.e., spiders) or a nonthreatening target (i.e., birds) in an array of neutral distractors (i.e., butterflies), both low- and high-spiderfearful participants responded faster to spider targets than to bird targets, while high-fearful participants exhibited a stronger bias than low-fearful participants (Aue et  al., 2013). In addition, spider targets were also identified accurately more often than bird targets by both phobic and nonfearful participants (Aue, Chauvigné, Bristle, Okon-Singer, & Guex, 2016). However, among depressed participants, visual search tasks yield mixed findings (for review, see De Raedt & Koster, 2010; see also Lichtenstein-Vidne et al., 2017, who found attention bias among anxious participants, but not among depressed ones). The aforementioned paradigms can also be used to study a specific component of attention bias. For instance, in an experiment where exposure duration was manipulated in a dot-probe task, participants with high fear of spiders exhibited faster engagement with spider pictures than with control pictures, compared with low-fear participants. This finding was true only in the condition with the shortest exposure time (200 ms). In longer exposure times (i.e., 500 ms and 2000 ms), no significant biases were found in either group of participants (Mogg & Bradley, 2006). Thus, the authors conclude that high levels of fear are linked with an enhanced early bias toward fear-relevant stimuli, which subsides at longer exposure times. Regarding slower disengagement, Yiend and Mathews (2001) used the spatial cuing paradigm to examine disengagement in anxiety. In Experiment 2, it was found that where a target (i.e., a neutral or a threatening picture) was previously spatially cued, both high- and low-trait anxious participants exhibited a slowing of RT on invalid trials, compared with valid trials. However, while low-anxiety participants reacted slower regardless of the type of target, high-anxiety participants reacted slower only when threatening pictures were involved. The slower detection times in invalid trials were attributed to the cost of shifting attention to the noncued location or to the benefit from engaging with the cue preceding the target (Yiend & Mathews, 2001; Exp. 2). Thus, the authors conclude that threatening pictures interfere with the performance of anxious participants in invalid trials, likely due to task-irrelevant processing, which occurs at the time of viewing the threatening stimulus. Regarding the vigilance-avoidance pattern, Koster et al. (2006) found that, when exposure time is short (i.e., 100 ms), high-trait anxious individuals (HTA) exhibited more rapid engagement and difficulties in disengagement, compared with low-trait anxious individuals (LTA). When exposure times increased (i.e., 200 and 500 ms), HTA participants exhibited stronger attentional avoidance, defined as negative engagement with and negative disengagement from threatening pictures, compared with LTA participants. In other words, HTA p­ articipants

Attention bias toward negative stimuli  Chapter | 2  27

attended more strongly (i.e., faster) to neutral pictures compared with threatening pictures, and they attended more strongly (i.e., faster) to neutral pictures compared with LTA participants. Weierich, Treat, and Hollingworth (2008) argue that disengagement and avoidance occur simultaneously, covertly, and overly, respectively. To determine whether stimulus-onset asynchrony (SOA; i.e., the time between the onset of the cue and the onset of the target) plays a role in disengagement, Yiend and Mathews (2001; Exp. 3) conducted an additional experiment in which SOA was increased from 500 to 2000 ms, allowing participants more time to shift their attention from the cue to the target. Results showed both (high and low) anxiety groups benefited from prolonged SOA on invalid trials when nonthreatening targets appeared, but not when threatening targets appeared. Following invalid threatening pictures, both groups exhibited slower disengagement. However, in both of the aforementioned experiments, only the high-­ anxiety group displayed a positive correlation between the degree of interference and the severity of the threat posed in the pictures. The authors argue that this finding could reflect a universal difficulty in disengaging attention from threatening stimuli. This difficulty is more easily exhibited in the absence of competition, as it becomes easier for threatening stimuli to hold attention regardless of anxiety levels, due to longer SOAs (Yiend & Mathews, 2001; Exp. 3). Each of the aforementioned paradigms has its advantages and disadvantages. The use of either one depends on the purpose of the study, whether it focuses on spatial attention and whether it wishes to distinguish between engagement and disengagement components and the type of population studied (e.g., fearful, phobic, anxious, and depressed). Nonetheless, each paradigm has been used extensively and has proven effective in revealing attention bias. Also, it is important to consider additional factors that influence emotional processing, such as the properties of the stimuli presented, levels of valence and arousal, relevance, and ecological value, and whether the stimulus encourages approach or avoidance behaviors (for review, see Okon-Singer, LichtensteinVidne, & Cohen, 2013). To summarize, faster engagement is exhibited through faster RTs in valid/ congruent trials (i.e., trials in which the locations of the cue and the target\probe match), while slower disengagement is exhibited through slower RTs in invalid\ incongruent trials (i.e., trials in which the locations of the cue and the target\ probe do not match). Avoidance is displayed when participants attend more strongly to neutral stimuli than they do to threatening stimuli. These phenomena are moderated by exposure times and SOAs, the types and the intensity of stimuli used, the paradigm used, the population of participants, and the severity of their symptoms. For further discussion on factors that may affect emotional processing, see Okon-Singer et al. (2013). In conclusion, several tasks have been commonly employed to reveal attention biases in different populations. Nevertheless, mixed findings do exist. For instance, a therapeutic method called attention bias modification (ABM)

28  Cognitive biases in health and psychiatric disorders

aims at reducing anxiety levels by reducing attention bias via means of training. A recent review on ABM yielded mixed results regarding the existence of attention bias in anxious participants’ pretraining, putting in question whether attention bias toward threatening stimuli exists in anxious participants in the first place (Mogg, Waters, & Bradley, 2017). Further mixed findings can be found in Becker and Rinck (2004), who did not find attention bias in spider phobics. Rather, they found an interpretation bias, in which phobics interpret ambiguous stimuli as fear-related ones (i.e., a beetle is interpreted as a spider). The authors suggest that faster engagement in phobics is not due to better detection abilities or to lower perceptual threshold to certain stimuli, but rather due to a more liberal response criterion. Similar interpretation biases were found in participants with panic disorder (Windmann & Krüger, 1998), social anxiety (Winton, Clark, & Edelmann, 1995), and depression (for review, see Everaert et  al., 2014). Therefore, it is important to consider how we define attention bias, how we measure it, and how we differentiate it from other biases (see also Okon-Singer, 2018, for elaboration). Finally, Shani, Tal, Zilcha-Mano, and Okon-Singer (2019) suggest that the high variance in the field of cognitive training may serve as a basis for employing machine-learning algorithms for optimizing training gains.

Brain regions involved in the emergence of attention bias Many works suggest that the amygdala is crucial for the rapid detection and reorienting of attention to aversive cues. The amygdala reacts to a broad range of motivationally significant stimuli, including faces (e.g., Minxha et al., 2017), words (e.g., Weisholtz et al., 2015), and scenarios (e.g., Sanchez et al., 2015). It was suggested that the amygdala can prioritize the processing of motivationally significant stimuli via direct and indirect projections to sensory cortex (see Okon-Singer et al., 2018 for details; see Fig. 2 for normal and abnormal brain connectivity between the amygdala and other regions based on attention-related tasks). Recent models further highlight the role of the thalamic pulvinar nucleus in attention biases to threat (Pessoa, 2013, 2017). The pulvinar has extensive connections to diverse cortical and subcortical regions, including connections to visual regions, frontoparietal attention-related areas (such as the prefrontal cortex, the cingulate cortex, and the intraparietal sulcus; Buchsbaum et al., 2006; Yuan et  al., 2016) and the amygdala (Tamietto, Pullens, de Gelder, Weiskrantz, & Goebel, 2012). The pulvinar plays a role in filtering visual distractors (Fischer & Whitney, 2012) and specifically in selective attention to emotional stimuli (Padmala, Lim, & Pessoa, 2010). Furthermore, enhanced connectivity between the pulvinar and higher-order visual and frontal areas was found among participants with social anxiety, compared with control participants (Tadayonnejad, Klumpp, Ajilore, Leow, & Phan, 2016). Hakamata et  al. (2016) manipulated attention to fearful faces by asking participants to attend to judge the similarity

Attention bias toward negative stimuli  Chapter | 2  29

IPS

PFC Pul

Visual Amy

(A)

IPS

PFC Pul Visual Amy

(B) FIG. 2  (A) Brain connectivity among healthy participants. (B) Brain connectivity in anxiety, as implicated from existing literature. Thick solid lines indicate enhanced connectivity, while thin solid lines indicate typical connectivity. PFC, prefrontal cortex; Pul, pulvinar; Amy, amygdala; Visual, visual cortex; IPS, intraparietal sulcus.

between a pair of two faces and a pair of two houses. They showed that participants with attention bias to aversive information exhibited higher pulvinar activation to unattended fearful than to unattended neutral faces and enhanced effective connectivity from the pulvinar to frontoparietal areas. These authors suggested that the pulvinar may be involved in gating unattended aversive information depending on individual threat-related attention bias. Finally, based on data from patients with brain injury, Arend, Henik, and Okon-Singer (2015) suggested that the pulvinar may perform “emotional tagging”—a control process that determines whether a certain stimulus will be considered emotional and therefore receive prioritized processing (see Fig. 2 for normal and abnormal brain connectivity between the pulvinar and other regions). As for depressed populations, De Raedt and Koster (2010) offer a framework for understanding vulnerability in recurrent depression. Their proposed framework includes two main components: a cognitive component, which includes negative schemata, diminished attentional control and rumination, and a biological component, which includes the hypothalamic-pituitary-adrenal

30  Cognitive biases in health and psychiatric disorders

(HPA) axis, serotonin, the amygdala, and prefrontal and subcortical areas. The authors claim that the HPA axis becomes more impaired during each depressive episode, making it more sensitive to stressors. This in turn leads to decreased activity in prefrontal areas, as they are associated with serotonin metabolism. Decreased activity in prefrontal areas is also associated with increased activity in the amygdala in reaction to external stressors, which leads to prolonged negative affect. Furthermore, decreased activity in prefrontal areas leads to attentional impairment, further contributing to a sustained negative affect.

Autonomic responses related to attention bias While most of the studies examining attention bias have used behavioral measurements, such as RT and accuracy, somatovisceral responses provide objective measures, over which participants have less control or awareness. While such methods are complicated to interpret, they can reflect some of the participants’ subjective feelings and arousal levels. These methods vary on the scale of how voluntary they are. One common measure for attention bias is derived by recording eye movements, which—early after stimulus onset—occur reflexively and will be the focus of the present section. Eye movements can also be considered behavioral measures, since unlike heart rate or blood pressure participants can exert some control over their eye movements. Eye movement measures are widely used because they explicitly show where participants look and for how long (for review on eye movements in attention and decision making, see Orquin & Loose, 2013). For instance, spider phobics exhibited speeded threat detection and increased distraction by threat, compared with nonanxious controls in three variations of the visual search task (Rinck, Reinecke, Ellwart, Heuer, & Becker, 2005). These findings are consistent with faster engagement and slower disengagement, respectively. Similarly, spider-fearful participants fixated earlier on spider pictures and for shorter durations, compared with nonfearful participants (Rinck & Becker, 2006). These findings are consistent with faster engagement and avoidance, respectively. Eye movement measures are often considered more reliable than RT measures in the research of attention bias. These claims have been argued in the study of PTSD (for review, see Lazarov, Suarez-Jimenez, et al., 2018) and in the study of depression (Lazarov, Ben-Zion, Shamai, Pine, & Bar-Haim, 2018). For instance, using eye-tracking methods, one study found that depressed participants dwell for longer periods on sad faces and show no bias toward happy faces, compared with nondepressed participants (Lazarov, Ben-Zion, et al., 2018). The authors claim that the eye-tracking method used in the study has better psychometric properties than RT-based measures, which often yield mixed findings (Lazarov, Ben-Zion, et al., 2018; Lazarov, Suarez-Jimenez, et al., 2018). Attention bias is characterized by increased arousal levels, and thus, many studies have measured arousal levels when displaying emotional stimuli using autonomic measures, such as heart rate (e.g., Sartory, Rachman, & Grey,

Attention bias toward negative stimuli  Chapter | 2  31

1977), blood pressure (e.g., McNally, Enock, Tsai, & Tousian, 2013; OkonSinger et al., 2014), pupil diameter (e.g., Aue et al., 2016; de Gee, Knapen, & Donner, 2014), and skin conductance measures (e.g., Öhman & Soares, 1994). Thus, these measures capture both arousal and valence levels. Please note, however, that autonomic measures do signal not only arousal and valence but also variables such as concentration, interest, and tiredness. For instance, elevated heart rate is often an index of increased arousal, and it would be expected in threatening situations. Thus, it is considered a measure of fear in fearful/phobic participants (Sartory et al., 1977). In other words, participants with spider phobia would exhibit increased heart rate toward spider-related film clips, but not toward other threatening stimuli, such as blood-related film clips, and the opposite pattern would be found in participants with blood-injection-injury phobia (Sarlo, Palomba, Angrilli, & Stegagno, 2002). To study the relationship between attention bias and autonomic measures, studies often measure autonomic measures while presenting one of the paradigms mentioned earlier. For instance, to examine the effects of prior expectancy on attention bias using autonomic measures, one study (Aue et al., 2016) presented a visual search array that contained eight distractors and one target: a threatening target (i.e., spiders) or a neutral target (i.e., birds) to participants with and without spider phobia. Both behavioral measures (RT and accuracy) and autonomic measures (i.e., pupil diameter and heart rate) found similar results: expectancy to encounter a certain stimulus affected detection of the neutral target, but only little the detection of the threatening target. In other words, both behavioral and autonomic measures converged and showed that detection of negative stimuli was virtually not affected by prior expectancy. Autonomic measures are widely used in the study of psychology in general and in emotions in particular. However, as noted earlier, it is important to remember the interpretation of such data needs to be done carefully, as it does not have any inherent fixed meaning. Specifically, it is difficult to isolate a component that is associated with attentional processes per se and does not touch upon other components of the stimulus processing. In addition, while conclusions regarding levels of arousal can be made, conclusions regarding valence are harder to reach. Rather, we can only tentatively reach conclusions about attention bias from such measurements. Thus, specific meanings and interpretations are often attached to data according to the context of the experiment (e.g., it is unlikely for a snake phobic to be happy at the sight of snakes).

Similarities and differences between healthy and clinical populations Attention bias is a hallmark of anxiety disorders. Apart from their deviant performance in the dot-probe, emotional spatial cueing, and Stroop tasks, anxious individuals have been shown to exhibit attention bias toward emotional task-irrelevant distractors, which worsens their performance on many

32  Cognitive biases in health and psychiatric disorders

tasks (e.g., Lichtenstein-Vidne et al., 2017; Okon-Singer, Alyagon, Kofman, Tzelgov, & Henik, 2011). However, as we have shown in the previous sections, attention bias is not specific to anxiety. In their meta-analysis, Peckham et al. (2010) showed that attention bias is strong and consistent in depressed participants across many studies and using various methods. This bias was not moderated by age, sex, type of depressed sample, duration of stimuli presentation, or type of stimuli. For meta-analysis on different types of stimuli (e.g., lexical, visual, and auditory) and populations used in the study of attention bias (e.g., clinical vs. nonclinical populations and adults vs. youths and children), see Bar-Haim et al. (2007) and Peckham et al. (2010) for anxiety disorders and depression, respectively. As mentioned earlier, attention bias includes three main components: faster engagement, slower disengagement, and avoidance (Koster et al., 2006). Both healthy and clinical populations exhibit attention bias (e.g. Aue et  al., 2013, 2016), but not all components of it. Specifically, while healthy, subclinical and clinical individuals exhibit early attention engagement, only anxious clinical and subclinical individuals exhibit difficulty in early disengagement and enhanced later avoidance (for review, see Aue & Okon-Singer, 2015). For instance, in a spatial cuing paradigm, while low-anxiety participants reacted slower on invalid trials regardless of the type of target, high-anxiety participants exhibited an interaction between validity and threat in that they reacted slower on invalid trials when threatening pictures were involved, compared with neutral pictures (Yiend & Mathews, 2001; Exp. 2). Thus, only anxious participants exhibited slower disengagement from threatening stimuli. Furthermore, anxious participants exhibit a more lasting and persistent bias than healthy participants (Aue et al., 2013). In addition, among participants with depression (De Raedt & Koster, 2010) and PTSD (Lazarov, Suarez-Jimenez, et al., 2018), difficulty in disengagement is the most dominant component, while enhanced engagement and vigilance are not exhibited. For instance, depressed participants exhibit difficulty in disengagement from sad faces but not from happy faces, while nondepressed participants do not exhibit this bias toward sad faces (Lazarov, Ben-Zion, et al., 2018). These findings also indicate an absence of positive bias among depressed participants, as nondepressed participants exhibit a bias toward happy faces. Thus, different populations exhibit different components of the bias, with varying degrees of severity. Not all components of attention bias are seen across all disorders. For instance, both depressed participants (for review, see De Raedt & Koster, 2010) and participants with PTSD (for review, see Lazarov, Suarez-Jimenez, et  al., 2018) exhibit difficulty in disengagement, but not enhanced engagement or vigilance-avoidance. For example, depressed participants show impaired disengagement with negative self-relevant stimuli, while nondysphoric participants maintain their attention more strongly on positive self-relevant words (e.g., “beloved,” “skillful,” and “powerful”), dysphoric participants exhibit difficulty in disengagement from negative self-relevant words (e.g., “loser,” “failure,” and

Attention bias toward negative stimuli  Chapter | 2  33

“hopeless”; Koster et al., 2005). This effect in dysphoric participants was enhanced during longer stimuli presentation (i.e., 1500 ms compared with 250 and 500 ms), which indicates the existence of the bias in later processing stages (for review, see De Raedt & Koster, 2010). Furthermore, there are individual differences that could mediate pathological as opposed to healthy reactions. For instance, Mathews and MacLeod (2002) suggest that anxiety-prone or vulnerable individuals have a lower threshold of sensitivity to threat and thus they move from avoidance to vigilance easily and more readily than less vulnerable individuals. This occurs regardless of the severity or the relevance of the supposedly threatening stimulus (see also Mogg & Bradley, 1998; Yiend & Mathews, 2001). These claims are consistent with interpretation bias, manifested by anxious individuals’ tendency to perceive ambiguous stimuli as threatening (e.g., a beetle perceived as a spider: Becker & Rinck, 2004). Other individual differences that could mediate attention bias include attentional control ability (e.g., ability to actively disengage from emotional stimuli; Peers & Lawrence, 2009) and emotional regulation strategies (e.g., the intentional allocation of attention toward neutral stimuli in the presence of threatening stimuli; Johnson, 2009; for review, see Cisler & Koster, 2010). Interestingly, it has been found that motivation in the form of monetary reward can enhance cognitive control of negative and positive emotional distractors in a perceptual task among healthy participants (Walsh, Carmel, Harper, & Grimshaw, 2018). Thus, proactive control strategies can reduce the effect of emotionally distracting stimuli among healthy individuals. It remains to be investigated whether clinical populations are equally responsive to such rewards. In conclusion, even though there are many differing models that attempt to describe what lies at the heart of anxiety and why some people are more prone to it than others, there seems to be a general consensus: differences in the evaluation system of threat leads to abnormal threat detection in anxious individuals. This oversensitivity to threat is manifested through attention bias and other biases, such as interpretation bias (e.g., Mogg & Bradley, 1998; Öhman, 1996; for reviews, see Mathews & MacLeod, 2002; Yiend, 2010; for meta-analysis, see Bar-Haim et al., 2007).

Limitations and future directions Although not addressed in the current chapter, it is important to remember that biases do not exist in a void. As mentioned, several other biases (e.g., expectancy and interpretation) are also related to attention bias. While most of the aforementioned studies and reviews focused solely on attention bias, recent studies have also examined the causal relationship between attention bias and expectancy bias and how attention bias can be modified using manipulation of expectancies (e.g., Abado, Sagi, Silber, de Houwer, Aue, & Okon-Singer, under review; Aue et al., 2013, 2016; Aue, Guex, Chauvigné, Okon-Singer, & Vuilleumier, 2018; for review, see Aue & Okon-Singer, 2015). For instance, it

34  Cognitive biases in health and psychiatric disorders

was found that attention bias in participants with high levels of fear of spiders can be lessened by lowering expectations regarding the appearance of threatening pictures (i.e., of spiders) versus nonthreatening pictures (i.e., birds; Abado et al., under review; see also Öhman, 1996 for the role of prior expectancy in attention to threat, from an evolutionary point of view). Furthermore, in their review on the causal effects of biases on anxiety, Mathews and MacLeod (2002) integrate findings about the causal relationship between emotional processing biases (i.e., attention bias and interpretation bias) and the development of vulnerability to anxiety. They conclude that these biases have causal effects on vulnerability to anxiety and that this relationship is mediated by vigilant emotional processing. Similar work has been done on the interactions between other biases (i.e., attention, interpretation, and memory) in depression (Everaert et al., 2014). Thus, it is important to examine how different biases interact and influence each other and how they influence disorders in a broader context. It is also important to study attention bias in different disorders, how the bias varies between disorders, and what the implications are for such variance. For instance, some studies show that the early engagement and vigilance components of the bias are less dominant in depression (for review, see De Raedt & Koster, 2010) and in PTSD (for review, see Lazarov, Suarez-Jimenez, et al., 2018) than they are in anxiety, while other studies show that, in some cases, bias toward negative stimuli does not even exist in depression, but it does exist in anxiety (Lichtenstein-Vidne et al., 2017). Studies that show different patterns of the bias can help identify different underlying cognitive mechanisms of each disorder and help differentiate between them. In addition, it is crucial to consider the construct validity of the measure at hand. As mentioned before, interpretation bias can sometimes be mistakenly construed as attention bias (Becker & Rinck, 2004). Thus, future research should focus on refining the differentiation between biases and between disorders, when measuring attention bias. A key to improve the research of attention bias could be found in the methods that are being used in different attentional studies. As mentioned in the prevalent paradigms and findings section, most studies of attention bias rely on behavioral measures and most commonly RT. Some recent studies have called these measures unreliable due to poor psychometric properties and instead offer the use of eye-tracking measures, such as in the study of depression (Lazarov, Ben-Zion, et  al., 2018) and in the study of PTSD (Lazarov, Suarez-Jimenez, et al., 2018). Lastly, future research should consider attention bias in the therapeutic context. Most of the aforementioned reviews and studies focused on the theoretical meanings of findings, but it is also crucial to draw clinical conclusions and to apply theoretical knowledge in clinical practice. For instance, it has been found that attention bias and anxiety disorders are linked, and it has been found that fear, anxiety, and attention bias contribute to each other, maintaining a vicious cycle (for review, see Van Bockstaele et al., 2014). In addition, the fact attention

Attention bias toward negative stimuli  Chapter | 2  35

bias can be manipulated and is not rigid can be used in attention bias modification therapy (for review, see Shechner et al., 2012). Similarly, it has been found that, when participants learn to fear a specific stimulus, generalization of this fear can be decreased when they learn how to better distinguish between relevant and irrelevant stimuli (Ginat-Frolich, Klein, Katz, & Shechner, 2017). Furthermore, it was shown that expectancies can—under certain conditions— attenuate attention bias (Abado et al., under review) and that motivation in the form of monetary reward (Walsh et al., 2018) can increase cognitive control of emotional distractors. Hence, attention bias is not fixed and can be modified, as humans are able to learn and to reorient their expectancies, motivation, and attention.

Summary In summary, there are varying theories that try to explain attention bias from different perspectives: evolutionary, cognitive, neurocognitive, etc. In the case of anxiety, most theories refer to a low sensitivity threshold that makes anxious individuals more vulnerable to fear mildly threatening stimuli. Furthermore, it is agreed upon that attention bias includes three main components: facilitated engagement and difficulty in disengagement (both components constitute vigilance), followed by avoidance. These components have been studies extensively, using various methods, paradigms, stimuli, and populations, all of which are factors that may moderate some of the biased reactions. In addition, some individual differences, such as attentional control ability and emotion regulation strategies, may also modulate attentional biases. Of note, some of the components of attention bias are found in healthy populations as well but to a lesser extent. The studies and reviews discussed in this chapter focus almost exclusively on attention bias, and thus, future research should examine the interaction between different biases and the interaction between biases and different disorders, as cognition and emotion are highly dynamic. Furthermore, implementation of the theoretical conclusions on the clinical field may strongly contribute to the development of targeted therapies for different mental disorders.

References Arend, I., Henik, A., & Okon-Singer, H. (2015). Dissociating emotion and attention functions in the pulvinar nucleus of the thalamus. Neuropsychology, 29(2), 191. Aue, T., Chauvigné, L. A., Bristle, M., Okon-Singer, H., & Guex, R. (2016). Expectancy influences on attention to threat are only weak and transient: Behavioral and physiological evidence. Biological Psychology, 121, 173–186. Aue, T., Guex, R., Chauvigné, L. A., & Okon-Singer, H. (2013). Varying expectancies and attention bias in phobic and non-phobic individuals. Frontiers in Human Neuroscience, 7, 418. Aue, T., Guex, R., Chauvigné, L., Okon-Singer, H., & Vuilleumier, P. (2018). Expectancies influence attention to neutral but not to threatening stimuli: An fMRI study. Emotion.

36  Cognitive biases in health and psychiatric disorders Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. Bar-Haim, Y., Lamy, D., Pergamin, L., Bakermans-Kranenburg, M. J., & Van Ijzendoorn, M. H. (2007). Threat-related attentional bias in anxious and nonanxious individuals: A meta-analytic study. Psychological Bulletin, 133(1), 1–24. Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. University of Pennsylvania Press. Beck, A. T. (1976). Cognitive therapy and the emotional disorders. New York: International Universities Press. Beck, A. T., Emery, G., & Greenberg, R. I. (1985). Anxiety disorders and phobias. New York: Basic Books. Becker, E., & Rinck, M. (2004). Sensitivity and response bias in fear of spiders. Cognition and Emotion, 18(7), 961–976. Brosch, T., Sander, D., Pourtois, G., & Scherer, K. R. (2008). Beyond fear: Rapid spatial orienting toward positive emotional stimuli. Psychological Science, 19, 362–370. Brosch, T., & Sharma, D. (2005). The role of fear-relevant stimuli in visual search: A comparison of phylogenetic and ontogenetic stimuli. Emotion, 5(3), 360–364. Buchsbaum, M. S., Buchsbaum, B. R., Chokron, S., Tang, C., Wei, T. C., & Byne, W. (2006). Thalamocortical circuits: fMRI assessment of the pulvinar and medial dorsal nucleus in normal volunteers. Neuroscience Letters, 404(3), 282–287. Cisler, J. M., & Koster, E. H. W. (2010). Mechanisms of attentional biases towards threat in anxiety disorders: An integrative review. Clinical Psychology Review, 30, 203–216. de Gee, J. W., Knapen, T., & Donner, T. H. (2014). Decision-related pupil dilation reflects upcoming choice and individual bias. Proceedings of the National Academy of Sciences, 111(5), E618– E625. De Raedt, R., & Koster, E. H. (2010). Understanding vulnerability for depression from a cognitive neuroscience perspective: A reappraisal of attentional factors and a new conceptual framework. Cognitive, Affective, & Behavioral Neuroscience, 10(1), 50–70. Everaert, J., Duyck, W., & Koster, E. H. (2014). Attention, interpretation, and memory biases in subclinical depression: A proof-of-principle test of the combined cognitive biases hypothesis. Emotion, 14(2), 331. Fischer, J., & Whitney, D. (2012). Attention gates visual coding in the human pulvinar. Nature Communications, 3, 1051. Fox, E., Russo, R., Bowles, R., & Dutton, K. (2001). Do threatening stimuli draw or hold visual attention in subclinical anxiety? Journal of Experimental Psychology: General, 130, 681–700. Ginat-Frolich, R., Klein, Z., Katz, O., & Shechner, T. (2017). A novel perceptual discrimination training task: Reducing fear overgeneralization in the context of fear learning. Behaviour Research and Therapy, 93, 29–37. Hakamata, Y., Sato, E., Komi, S., Moriguchi, Y., Izawa, S., Murayama, N., et al. (2016). The functional activity and effective connectivity of pulvinar are modulated by individual differences in threat-related attentional bias. Scientific Reports, 6, . Article No.: 34777. Hankin, B. L., Gibb, B. E., Abela, J. R., & Flory, K. (2010). Selective attention to affective stimuli and clinical depression among youths: Role of anxiety and specificity of emotion. Journal of Abnormal Psychology, 119(3), 491. Hommer, R. E., Meyer, A., Stoddard, J., Connolly, M. E., Mogg, K., Bradley, B. P., et al. (2014). Attention bias to threat faces in severe mood dysregulation. Depression and Anxiety, 31(7), 559–565.

Attention bias toward negative stimuli  Chapter | 2  37 Johnson, D. R. (2009). Goal-directed attentional deployment to emotional faces and individual differences in emotion regulation. Journal of Research in Personality, 43, 8–13. Kendall, P. C., & Ingram, R. E. (1989). Cognitive-behavioral perspectives: Theory and research on depression and anxiety. In P. C. Kendall & D. Watson (Eds.), Personality, psychopathology, and psychotherapy. Anxiety and depression: Distinctive and overlapping features (pp. 27–53). San Diego, CA: Academic Press. Koster, E. H. W., Crombez, G., Verschuere, B., & De Houwer, J. (2004). Selective attention to threat in the dot probe paradigm: Differentiating vigilance and difficulty to disengage. Behaviour Research and Therapy, 42, 1183–1192. Koster, E. H., Crombez, G., Verschuere, B., Van Damme, S., & Wiersema, J. R. (2006). Components of attentional bias to threat in high trait anxiety: Facilitated engagement, impaired disengagement, and attentional avoidance. Behaviour Research and Therapy, 44(12), 1757–1771. Koster, E. H., De Raedt, R., Goeleven, E., Franck, E., & Crombez, G. (2005). Mood-congruent attentional bias in dysphoria: Maintained attention to and impaired disengagement from negative information. Emotion, 5(4), 446. Kurdi, B., Lozano, S., & Banaji, M. R. (2017). Introducing the open affective standardized image set (OASIS). Behavior Research Methods, 49(2), 457–470. Lazarov, A., Ben-Zion, Z., Shamai, D., Pine, D. S., & Bar-Haim, Y. (2018). Free viewing of sad and happy faces in depression: A potential target for attention Bias modification. Journal of Affective Disorders, 238, 94–100. Lazarov, A., Suarez-Jimenez, B., Tamman, A., Falzon, L., Zhu, X., Edmondson, D. E., et al. (2018). Attention to threat in posttraumatic stress disorder as indexed by eye-tracking indices: a systematic review. Psychological Medicine, 1–22. https://doi.org/10.1017/S0033291718002313. Lichtenstein-Vidne, L., Okon-Singer, H., Cohen, N., Todder, D., Aue, T., Nemets, B., et al. (2017). Attentional bias in clinical depression and anxiety: The impact of emotional and non-emotional distracting information. Biological Psychology, 122, 4–12. MacLeod, C., Mathews, A., & Tata, P. (1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology, 95(1), 15. Mathews, A., & MacLeod, C. (1985). Selective processing of threat cues in anxiety states. Behaviour Research and Therapy, 23(5), 563–569. Mathews, A., & MacLeod, C. (2002). Induced processing biases have causal effects on anxiety. Cognition & Emotion, 16, 331–354. McNally, R. J., Enock, P. M., Tsai, C., & Tousian, M. (2013). Attention bias modification for reducing speech anxiety. Behaviour Research and Therapy, 51(12), 882–888. Merckelbach, H., Van den Hout, M. A., Jansen, A., & van der Molen, G. M. (1988). Many stimuli are frightening, but some are more frightening than others: The contributions of preparedness, dangerousness, and unpredictability to making a stimulus fearful. Journal of Psychopathology and Behavioral Assessment, 10(4), 355–366. Minxha, J., Mosher, C., Morrow, J. K., Mamelak, A. N., Adolphs, R., Gothard, K. M., et al. (2017). Fixations gate species-specific responses to free viewing of faces in the human and macaque amygdala. Cell Reports, 18(4), 878–891. Mogg, K., & Bradley, B. P. (1998). A cognitive-motivational analysis of anxiety. Behaviour Research and Therapy, 36(9), 809–848. Mogg, K., & Bradley, B. P. (2006). Time course of attentional bias for fear-relevant pictures in spider-fearful individuals. Behaviour Research and Therapy, 44(9), 1241–1250. Mogg, K., Bradley, B. P., De Bono, J., & Painter, M. (1997). Time course of attentional bias for threat information in non-clinical anxiety. Behaviour Research and Therapy, 35(4), 297–303.

38  Cognitive biases in health and psychiatric disorders Mogg, K., Bradley, B. P., Miles, F., & Dixon, C. (2004). Time course of attentional bias for threat scenes: Testing the vigilance-avoidance hypothesis. Cognition and Emotion, 18, 689–700. Mogg, K., Waters, A. M., & Bradley, B. P. (2017). Attention bias modification (ABM): Review of effects of multisession ABM training on anxiety and threat-related attention in high-anxious individuals. Clinical Psychological Science, 5(4), 698–717. Öhman, A. (1996). Preferential preattentive processing of threat in anxiety: Preparedness and attentional biases. Current Controversies in the Anxiety Disorders, 2, 253–290. Öhman, A., & Soares, J. J. (1994). “Unconscious anxiety”: Phobic responses to masked stimuli. Journal of Abnormal Psychology, 103(2), 231. Okon-Singer, H. (2018). The role of attention bias to threat in anxiety: Mechanisms, modulators and open questions. Current Opinion in Behavioral Sciences, 19, 26–30. Okon-Singer, H., Alyagon, U., Kofman, O., Tzelgov, J., & Henik, A. (2011). Fear-related pictures deteriorate the performance of university students with high fear of snakes or spiders. Stress, 14(2), 185–193. Okon-Singer, H., Lichtenstein-Vidne, L., & Cohen, N. (2013). Dynamic modulation of emotional processing. Biological Psychology, 92(3), 480–491. Okon-Singer, H., Mehnert, J., Hoyer, J., Hellrung, L., Schaare, H. L., Dukart, J., et  al. (2014). Neural control of vascular reactions: Impact of emotion and attention. Journal of Neuroscience, 34(12), 4251–4259. Okon-Singer, H., Stout, D. M., Stockbridge, M. D., Gamer, M., Fox, A. S., & Shackman, A. J. (2018). The interplay of emotion and cognition. In A. S.  Fox, R. C.  Lapate, A. J.  Shackman, & R. J. Davidson (Eds.), The nature of emotion: Fundamental questions. (2nd edition) (pp. 181–186). New York: Oxford University Press. Orquin, J. L., & Loose, S. M. (2013). Attention and choice: A review on eye movements in decision making. Acta Psychologica, 144(1), 190–206. Padmala, S., Lim, S., & Pessoa, L. (2010). Pulvinar and affective significance: Responses track moment-to-moment stimulus visibility. Frontiers in Human Neuroscience, 4, 64. Peckham, A. D., McHugh, R. K., & Otto, M. W. (2010). A meta-analysis of the magnitude of biased attention in depression. Depression and Anxiety, 27(12), 1135–1142. https://doi.org/10.1002/ da.20755. Peers, P. V., & Lawrence, A. D. (2009). Attentional control of emotional distraction in rapid serial visual presentation. Emotion, 9, 140–145. Pessoa, L. (2013). The cognitive-emotional brain: From interactions to integration. MIT Press. Pessoa, L. (2017). A network model of the emotional brain. Trends in Cognitive Sciences, 1(5), 357–371. Posner, M. I. (1980). Orienting of attention. Quarterly Journal of Experimental Psychology, 32, 3–25. Rinck, M., & Becker, E. S. (2006). Spider fearful individuals attend to threat, then quickly avoid it: Evidence from eye movements. Journal of Abnormal Psychology, 115(2), 231. Rinck, M., Reinecke, A., Ellwart, T., Heuer, K., & Becker, E. S. (2005). Speeded detection and increased distraction in fear of spiders: Evidence from eye movements. Journal of Abnormal Psychology, 114(2), 235. Sanchez, T. A., Mocaiber, I., Erthal, F. S., Joffily, M., Volchan, E., Pereira, M. G., et al. (2015). Amygdala responses to unpleasant pictures are influenced by task demands and positive affect trait. Frontiers in Human Neuroscience, 9, 107. Sarlo, M., Palomba, D., Angrilli, A., & Stegagno, L. (2002). Blood phobia and spider phobia: Two specific phobias with different autonomic cardiac modulations. Biological Psychology, 60(2– 3), 91–108.

Attention bias toward negative stimuli  Chapter | 2  39 Sartory, G., Rachman, S., & Grey, S. J. (1977). An investigation of the relation between reported fear and heart rate. Behaviour Research and Therapy, 15, 435–438. Seligman, M. E. (1971). Phobias and preparedness. Behavior Therapy, 2(3), 307–320. Shani, R., Tal, S., Zilcha-Mano, S., & Okon-Singer, H. (2019). Can machine learning approaches lead toward personalized cognitive training? Frontiers in Behavioral Neuroscience, 13, 64. Shechner, T., Britton, J. C., Pérez‐Edgar, K., Bar‐Haim, Y., Ernst, M., Fox, N. A., et al. (2012). Attention biases, anxiety, and development: Toward or away from threats or rewards? Depression and Anxiety, 29(4), 282–294. Shiffrin, R. M., & Schneider, W. (1977). Controlled and automatic human information processing: II. Perceptual learning, automatic attending and a general theory. Psychological Review, 84, 127–190. Tadayonnejad, R., Klumpp, H., Ajilore, O., Leow, A., & Phan, K. L. (2016). Aberrant pulvinar effective connectivity in generalized social anxiety disorder. Medicine, 95(45). Tamietto, M., Pullens, P., de Gelder, B., Weiskrantz, L., & Goebel, R. (2012). Subcortical connections to human amygdala and changes following destruction of the visual cortex. Current Biology, 22(15), 1449–1455. Teachman, B. A., Joormann, J., Steinman, S. A., & Gotlib, I. H. (2012). Automaticity in anxiety disorders and major depressive disorder. Clinical Psychology Review, 32(6), 575–603. Van Bockstaele, B., Verschuere, B., Tibboel, H., De Houwer, J., Crombez, G., & Koster, E. H. (2014). A review of current evidence for the causal impact of attentional bias on fear and anxiety. Psychological Bulletin, 140(3), 682. Walsh, A. T., Carmel, D., Harper, D., & Grimshaw, G. M. (2018). Motivation enhances control of positive and negative emotional distractions. Psychonomic Bulletin & Review, 25(4), 1556– 1562. Weierich, M. R., Treat, T. A., & Hollingworth, A. (2008). Theories and measurement of visual attentional processing in anxiety. Cognition and Emotion, 22, 985–1018. Weisholtz, D. S., Root, J. C., Butler, T., Tüscher, O., Epstein, J., Pan, H., et al. (2015). Beyond the amygdala: Linguistic threat modulates peri-sylvian semantic access cortices. Brain and Language, 151, 12–22. Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1997). Cognitive psychology and emotional disorders (2nd ed.). New York, NY: Wiley. Williams, J. M. G., Watts, F. N., MacLeod, C., & Matthews, A. (1988). Cognitive psychology and emotional disorders. Chichester, England: Wiley. Windmann, S., & Krüger, T. (1998). Subconscious detection of threat as reflected by an enhanced response bias. Consciousness and Cognition, 7(4), 603–633. Winton, E. C., Clark, D. M., & Edelmann, R. J. (1995). Social anxiety, fear of negative evaluation and the detection of negative emotion in others. Behaviour Research and Therapy, 33(2), 193–196. Yiend, J. (2010). The effects of emotion on attention: A review of attentional processing of emotional information. Cognition and Emotion, 24(1), 3–47. Yiend, J., & Mathews, A. (2001). Anxiety and attention to threatening pictures. The Quarterly Journal of Experimental Psychology: Section A, 54(3), 665–681. Yuan, R., Di, X., Taylor, P. A., Gohel, S., Tsai, Y., & Biswal, B. B. (2016). Functional topography of the thalamocortical system in human. Brain Structure and Function, 221(4), 1971–1984.

40  Cognitive biases in health and psychiatric disorders

Further Reading Abado, Sagi, Silber, de Houwer, Aue, & Okon-Singer, Attention bias and expectancy bias in fear of spiders: Context modulates biological preparedness. Under review. Richter, T., Aue, T., Richter-Levin, G., & Okon-Singer, H. (under review). Cognitive biases in anxiety and depression through the lens of automatic processing.

Chapter 3

The neurophysiological basis of optimism bias Mihai Dricua, Laura Kressa,b, Tatjana Auea a

Department of Psychology, University of Bern, Bern, Switzerland, bDepartment of Psychology, University of Uppsala, Uppsala, Sweden

Introduction The Oxford English Dictionary defines the term optimism as “hopefulness and confidence about the future or the success of something” (Optimism, n.d. para. 3). Such hopefulness and confidence about the future are essential in many aspects of everyday life. For instance, optimistic expectancies about the future can motivate people to engage in self-relevant situations and initiate goaldirected behavior, which, in turn, provides important opportunities to obtain beneficial rewards (Armor & Taylor, 1998; Shepperd, Waters, Weinstein, & Klein, 2015). Psychological research has revealed that people who look positively at their environment are more likely to experience positive emotions and to be satisfied with their life (Davidson, 2004; Fox, 1993; Norris, Larsen, Crawford, & Cacioppo, 2011). An optimistic outlook on the future may even promote well-being and protect mental and physical health (Garrett et al., 2014; Scheier & Carver, 1992). While the strength of optimism displayed varies from person to person (thus reflecting a stable personality trait; (Scheier, Carver, & Bridges, 1994)), most people tend to be optimistic rather than pessimistic about their own future. Seminal work by Neil Weinstein in the 1980s has revealed that this optimism exceeds the level of what would be justified by a rational consideration of situational characteristics, meaning that people are overly optimistic when it comes to making predictions about their future. This phenomenon, named unrealistic optimism, has been suggested to represent a prevalent cognitive bias in the general population (Sharot, 2011; Windschitl & Stuart, 2015). Researchers distinguish between two types of unrealistic optimism that differ in the standard against, which personal predictions are compared (Shepperd et al., 2015). Unrealistic comparative optimism indicates the expectation that one’s personal outcomes will be more favorable than the outcomes of one’s Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00003-9 © 2020 Elsevier Inc. All rights reserved.

41

42  Cognitive biases in health and psychiatric disorders

peers (i.e., more positive outcomes and/or less negative outcomes will happen compared with those of others) (Weinstein, 1980). Unrealistic absolute optimism indicates that people predict a better outcome than they will likely have in reality; that is, the standard of comparison is a quantitative objective measure such as epidemiological or base-rate data (Shepperd, Ouellette, & Fernandez, 1996). Regardless of the standard of comparison, optimism bias appears in a plethora of domains, including health (Weinstein, 1982), profession (Hoch, 1985), sports (Aue, Nusbaum, & Cacioppo, 2011), and politics (Babad, 1995). In fact, individuals are strongly motivated to maintain a positive outlook of their future even if they receive contradictory information. When receiving such information, respondents more readily update their beliefs in response to information that is better than expected than they do to information that is worse than expected (Sharot, Korn, & Dolan, 2011). For example, individuals promptly decrease their estimated chance of developing cancer if they are informed that the base rate of cancer in the general population is lower than initially assumed for themselves (i.e., better than expected information), but they do not increase their estimated chance of being pickpocketed when the presented base rate of pickpocketing in the general population is higher than initially assumed for themselves (i.e., worse than expected information). This asymmetry in belief updating has been proposed to maintain unrealistic optimism over time. To date, the vast majority of studies that investigated belief updating have sampled only future negative outcomes such as having heart failure or missing a flight. A similar mechanism is assumed in response to positive outcomes (Chowdhury, Sharot, Wolfe, Düzel, & Dolan, 2014; Kuzmanovic, Jefferson, & Vogeley, 2015, 2016; Marks & Baines, 2017; Sharot et al., 2011; Sharot et al., 2012), although this assumption has been challenged empirically (Shah, Harris, Bird, Catmur, & Hahn, 2016). For a clear definition of optimism bias, one has to distinguish it from other closely related concepts in the field of optimism research. Dispositional ­optimism, for instance, describes a relatively stable disposition of having an o­ ptimistic yet not necessarily unrealistic life orientation that is usually assessed by personality inventories (such as the Life Orientation Test; (Scheier et al., 1994)). These personality inventories measure optimistic expectancies about the future in a general way (e.g., “I’m always optimistic about my future”). Optimism bias, instead, represents overly optimistic expectancies about specific future situations in various areas. It is usually measured by asking people about the probability of experiencing a certain event (e.g., “living past 80”) in their future compared with the chances of a person similar in age and gender or by comparing a person’s expectancy with an objective outcome such as base rates (Sharot, 2011; Shepperd et al., 1996; Weinstein, 1980). Even though optimism bias and dispositional optimism are not the same (correlation coefficients between the two vary between .01 and .59; (Aspinwall & Brunhart, 1996; Radcliffe & Klein, 2002; Shepperd et al., 2015; Taylor et al., 1992), dispositional optimism is likely to influence an individual’s readiness to display optimism bias. The focus of this chapter is on

The neurophysiological basis of optimism bias  Chapter | 3  43

optimism bias, but we additionally refer to findings on dispositional optimism where no or few findings on optimism bias exist (e.g., in section “Somatovisceral responses related to optimism bias” on somatovisceral responses). It is furthermore important to note that research on optimism bias suffers from a lack of consistent terminology. Specifically, different terms (e.g., wishful thinking, unrealistic optimism, motivated reasoning, and overoptimism) have been used to refer to the same psychological phenomenon (or at least highly similar phenomena) in previous research. Prior attempts to distinguish different terms used in the field have not prevailed in the literature (Krizan & Windschitl, 2007). Therefore, one has to be aware of the problem that, on the one hand, different terms have been used for the same phenomenon, while, on the other hand, the same terms have been used for different phenomena. To make this chapter more accessible, we consistently refer to the broad term optimism bias throughout the following sections. Whenever possible, we clearly indicate whether the research that we reviewed investigated one of the most important subconcepts, unrealistic comparative optimism and unrealistic absolute optimism, as proposed by Shepperd et al. (2015). Furthermore, we refer to asymmetric updating of future predictions, suggested to strengthen optimism bias, as optimistic belief updating. In the present chapter, we provide an overview on influential theories in the field (section “Major theories in the field”); review methodologies (e.g., questionnaires and cognitive tasks) used to examine optimism bias (section “Methods used to investigate the optimism bias”); identify neural (section “Brain regions involved in the emergence and maintenance of the bias”) and somatovisceral (section “Somatovisceral responses related to optimism bias”) responses suggested to underlie the emergence of optimism bias, outline differences in the amount of optimism bias displayed by healthy people and clinical populations (section “Similarities and differences between healthy and clinical populations”); highlight limitations of previous research on optimism bias (section “Limitations”); and point out future directions for the field (section “Future directions”).

Major theories in the field When does optimism bias manifest? Since Neil Weinstein’s seminal work in the 1980s, researchers have debated the true prevalence of the optimism bias. At one end of the spectrum, some authors have argued that the optimism bias reflects an evolutionary advantage that promotes well-being and encourages risk taking (Taylor & Brown, 1988; Trimmer, 2016). In this sense, optimism bias reflects a basic heuristic that humans apply when thinking about their future, their skills, or their personality traits and is thus present in all healthy individuals (e.g., Sharot, 2011). Further support for this hypothesis comes from studies on asymmetrical updating of beliefs, which

44  Cognitive biases in health and psychiatric disorders

show that overoptimistic beliefs are particularly resilient to disconfirming evidence (Chowdhury et  al., 2014; Kuzmanovic, Jefferson, & Vogeley, 2016; Moutsiana, Charpentier, Garrett, Cohen, & Sharot, 2015; Sharot et al., 2011). At the other end of the spectrum, several authors have argued that a genuine optimism bias might not exist because its manifestation is either the result of methodological confounds or because it masquerades as other established cognitive and motivational biases. Regarding the latter, comparative optimism bias is seen by some authors not as a stand-alone phenomenon, but rather as the result of impression management under social (un) desirability (Shepperd, Carroll, Grace, & Terry, 2002; Sweldens, Puntoni, Paolacci, & Vissers, 2014) or an illusion of control (Horswill & McKenna, 1999; McKenna, 1993). Many events featured in the comparative optimism bias literature carry a social stigma, from “attempting suicide” and “having a drinking problem” (e.g., Weinstein, 1980) to “contracting genital warts” and “having a sexual dysfunction” (e.g., Sharot, 2011). Publicly admitting high personal likelihoods compared with those of peers might have negative social repercussions. Consequently, respondents downplay their perceived risk in these scenarios. Strong evidence for the role of impression management in optimism bias comes from a series of studies performed by Sweldens et al. (2014). One meta-analysis concluded that the greater the social desirability of an event, the greater the reported comparative optimism. In one empirical study, the authors manipulated the perceived anonymity of the research setting and showed that the effect of social desirability on comparative optimism significantly diminished with greater emphasis on anonymity. In two other studies, the authors found that the effect of social desirability is larger among people with stronger impression management tendencies and that explicit instructions to make a good impression increase comparative optimism, whereas instructions to make a negative impression reverse the effect (Sweldens et al., 2014). Some authors reduce unrealistic optimism to an illusion of control, suggesting that perceiving an event as controllable should be both necessary and sufficient to generate optimism bias about that event (McKenna, 1993). It follows that both unrealistic optimism and the illusion of control could operate for events in which individuals think they can exercise high control. However, for events that are perceived as low in controllability, expecting positive outcomes could only be the result of an unrealistic optimism bias. Consequently, to truly dissociate an optimistic bias from an illusion of control, investigations should focus predominantly on events that are perceived as low in controllability. In this regard, McKenna found a significant comparative optimism bias for events that are perceived as high in controllability but not for events that are low in controllability (McKenna, 1993). The author concluded, unsurprisingly, that the manifestation of optimism bias is simply the unwarranted illusion of control over minimizing negative outcomes and maximizing positive outcomes. Another argument against the existence of a genuine optimism bias relates to potential methodological confounds (Harris, Shah, Catmur, Bird, & Hahn, 2013; Klar & Ayal, 2004; Shepperd, Klein, Waters, & Weinstein,

The neurophysiological basis of optimism bias  Chapter | 3  45

2013; Windschitl & Stuart, 2015). The biased sampling of events is arguably the most detrimental confounding factor. In the simplest terms, it suggests that most of the events featured and analyzed in studies on optimism bias have been questionably selected without having been balanced for key characteristics. For example, studies that simultaneously investigated positive and negative events did not match these events for absolute valence scores (i.e., distance from a neutral point score), frequency, controllability, emotional intensity, or personal experience (e.g., Blair et al., 2013; Chambers, Windschitl, & Suls, 2003; Lench & Ditto, 2008). Studies that investigated either negative events or positive events rarely ensured that the sample included both rare and frequent events and controllable and uncontrollable events (e.g., Sharot, Riccardi, Raio, & Phelps, 2007; Weinstein, 1980). These event characteristics significantly influence the magnitude and direction of overoptimistic beliefs (Helweg-Larsen & Shepperd, 2001; Klein & Helweg-Larsen, 2002). If left unbalanced, either of these characteristics might confound results and bias their interpretation (Shepperd et al., 2013). The method of asking the question also considerably influences the magnitude of the comparative optimism bias. On the one hand, there is great variability in the subjective interpretation of the same verbal probability expressions. For example, chance terms such as “probably” and “likely” may have different meanings for different people (Budescu & Wallsten, 1985; Jenkins, Harris, & Lark, 2018). More important, chance terms are interpreted as denoting a higher probability when used to describe the likelihood of pleasant events in one’s own future than when used to describe someone else’s future (Smits & Hoorens, 2005). Conversely, chance terms are interpreted as denoting a lower probability when they are used to describe the likelihood of unpleasant events in one’s own future than when similarly used for someone else’s future (Smits & Hoorens, 2005). Clearly, the very content of the question is understood differently by different people. On the other hand, framing the question directly versus indirectly can drive the results (Helweg-Larsen & Shepperd, 2001; Rose, Suls, & Windschitl, 2011). In the indirect method, respondents are asked to make an absolute judgment separately about themselves and about another. The difference between the two types of absolute judgments serves as an indirect measure of comparative optimism. In the direct method, respondents are asked to indicate their likelihood of experiencing an event compared with that of others, or, in broader terms, how their standing on a target dimension compares with that of others (e.g., how creative are you compared with your coworkers?). It has been argued that it is the nature of the direct task that creates the phenomenon of comparative optimism, that is, respondents are prompted to think about and generate differences between themselves and others that might not have existed prior to their being asked (Chambers & Windschitl, 2004). The indirect measure of assessing comparative optimism would then be preferred to assess the true prevalence of the phenomenon. Most of the research on comparative optimism bias has been conducted on the assumption that a checklist of characteristics must be fulfilled for the bias

46  Cognitive biases in health and psychiatric disorders

to manifest (Helweg-Larsen & Shepperd, 2001; Rose et  al., 2011; Shepperd et al., 2013). Indeed, most studies that use future events as the unit of analysis have shown that optimism bias manifests for some events more than for others, clearly suggesting that not all events are assessed in the same manner by respondents (e.g., Campbell, Greenauer, Macaluso, & End, 2007; Weinstein, 1980). An optimism bias has been shown for events that are highly controllable (Harris, Cikara, & Fiske, 2008), are unfamiliar (Chambers et al., 2003; Price, Pentecost, & Voth, 2002), have a big emotional impact (Krizan & Windschitl, 2009), are temporally distant (Shepperd et  al., 1996), and in which the comparison target is psychologically/interpersonally distant (Harris, Middleton, & Joiner, 2000; Shepperd et  al., 2002). The magnitude of the bias in all other combinations is significantly reduced or completely absent (Chambers, 2008; Helweg-Larsen & Shepperd, 2001; Klein & Helweg-Larsen, 2002). In addition, the mood, personality traits, and cultural background of the respondents play a significant role (Norris et al., 2011; Rose, Endo, Windschitl, & Suls, 2008). Aside from overoptimistic beliefs for oneself compared with others, there are also instances of an optimism bias displayed toward groups and individuals that one identifies with or evaluates positively. For example, sports fans consistently display overoptimistic beliefs about the likelihood of their favorite team winning the game (Aue et  al., 2011; Babad, 1987; Price, 2000; Simmons & Massey, 2012), and voters overestimate the chances of their preferred political candidate winning elections (Babad, 1995, 1997). These instances of social optimism bias for in-group members compared with out-group members can be understood as a natural extension of the optimism bias for personal outcomes (e.g., Aue et al., 2011).

How and why does optimism bias emerge? Among the authors who recognize the optimism bias as an authentic phenomenon, two lines of explanation have been put forth, that is, cognitive accounts and motivational accounts. Traditionally, these two accounts have been investigated separately, with cognitive psychologists favoring the former and affective psychologists placing more weight on the latter. Cognitive accounts postulate that optimism biases arise from the differential manner in which people process information about themselves and about others at several stages, including perception, memory retrieval, and decision-making (Chambers et al., 2003; Price et al., 2002; Weinstein, 1980). Furthermore, cognitive accounts propose that the optimism bias would inadvertently be reduced or eliminated if experimenters ensured that self-relevant information was not processed preferentially compared with information about others (Chambers & Windschitl, 2004). For example, optimism bias is greatly reduced when participants are provided with diagnostic information concerning others’ behavior in risky situations or when the attentional focus of respondents on others’ risk likelihoods is experimentally boosted (Weinstein & Lachendro, 1982).

The neurophysiological basis of optimism bias  Chapter | 3  47

A subset of cognitive biases is represented by egocentric biases, an umbrella term used for a set of specific cognitive mechanisms. All egocentric biases state that individuals evaluate themselves differently from how they evaluate others (Chambers & Windschitl, 2004). The consequence is that self-relevant information, relative to other-relevant information, has disproportionate weight in the judgment process. Some of the manifestations of egocentric biases include differences in the amount of available information, in the quality of this information, or in the accessibility of self-relevant and other-relevant knowledge. The differential accessibility account posits that, when two entities are being compared on a given dimension, any difference in the accessibility of relevant information can lead to a biased comparative judgment. Consequently, trait and likelihood information about the self and close others may simply be more accessible from memory than is information about other people. According to the account of differential sources of information, respondents rely on episodic memories and idiosyncrasies that are diagnostic of a trait or a behavior when forming judgments about a familiar entity (e.g., self and friend). However, individuals rely on social stereotypes and the prevalence of a behavior or an event in the general population when thinking about an unfamiliar entity (Klar, Medding, & Sarel, 1996). A related concept is the representativeness heuristic, according to which people judge their likelihood of experiencing an event from how well they match their own stereotype of the individual who usually experiences such an event (Tversky & Kahneman, 1973). For example, when people are asked to estimate their risk of getting in an automobile accident relative to the average driver, the question itself prompts respondents to think about someone who drives too fast, mixes alcohol and driving, and is inattentive to other drivers (Perloff & Fetzer, 1986). In comparison with this stereotypical individual, people naturally conclude that their risk is lower. People may also use different standards to answer a comparative question. Asked if one is sportive, one person might think of how often they go to the gym or run in the morning. Instead, another person may think of how often they play organized sports such as football or basketball (Chambers & Windschitl, 2004). Motivational accounts of optimism bias postulate that some form of affect or motivation is the underlying biasing factor (Krizan & Windschitl, 2007; Lench & Bench, 2012). Consequently, manipulating the underlying affect or motive would substantially reduce or eliminate the optimism bias. For example, inducing negative affective states increases the perception of personal risks, leading to a significantly lower optimism bias for undesirable outcomes compared with that of controls who received no affect induction (Helweg-Larsen & Shepperd, 2001). One of the many forms of motivated reasoning is the desirability bias, that is, being overoptimistic about a future outcome as a result of the respondent’s preference or desire for that outcome (Lench, 2009). Consequently, people end up expecting positive outcomes to occur significantly more often than they expect negative outcomes to occur. There are several mechanisms through which desires can lead to a desirability bias. One such mechanism is

48  Cognitive biases in health and psychiatric disorders

valence ­priming, in which the valence of an outcome activates similarly valenced knowledge about that preferred outcome. For example, in determining the likelihood of obtaining a passing grade, that is, a positive outcome, a student might differentially activate more positive information that is congruent with the valence of the desired outcome, such as episodic memories of studying the course material throughout the academic year or incidents of classmates being less prepared than the student is. Valence priming thus works similarly to semantic priming, in which the activation of a mental concept spreads more readily to similar semantic concepts, for example, hearing the word “chair” activates the concept of “table.” Valence priming can also explain a reduced optimism bias when the target outcome is undesirable and the underlying mood of the respondent is congruently negative (Krizan & Windschitl, 2007). Another important mechanism of motivated reasoning that might bias respondents is repeated simulation. The desirability of an outcome influences the extent to which people passively or actively simulate that event’s occurrence. Repeated simulation of a desired outcome can occur because the very process of imagining that outcome is enjoyable. Since this simulation could include both the outcome and its precursors, the enhanced simulation makes the precipitating causes of that outcome more available in memory than the causes of alternative outcomes are (Krizan & Windschitl, 2007). In this sense, desirability bias can manifest not only for positive but also for negative outcomes. Repeated simulation of undesirable events can make some people pessimistic. Perhaps, out of a strategy to prepare for potential undesirable outcomes, people can simulate negative scenarios more vividly than they can positive scenarios. For example, individuals who have generalized anxiety disorder (GAD) describe worrying, the tenet of their illness, as a positive mechanism that helps them cope with future uncertainty (Newman, Llera, Erickson, Przeworski, & Castonguay, 2013). Interestingly, when these individuals imagine their future as having few positive outcomes, reward-processing regions in the brain become more active (Blair et al., 2017). The source of the desirability bias can stem not only from the future outcome’s affective quality but also from the perception of the reference group, a perception that might succumb to stereotypes and prejudices (Aue et al., 2011; Babad, 1997). In this sense, respondents are motivated to generate unflattering views of the reference group, particularly if this is an out-group. Consequently, individuals may evaluate themselves and in-group members in an overoptimistic manner compared with how they evaluate out-group members (Babad & Katz, 1991). For example, Dricu and colleagues found that student respondents manifested the greatest desirability bias (indexed as the magnitude of the difference between likelihood estimates of positive and negative future outcomes) when they assessed a similar student character but not when they assessed an out-group member such as an elderly character or a businessperson (Dricu et al., 2018). More important, the desirability bias changed direction and increased in magnitude when respondents assessed an alcoholic character, such that they

The neurophysiological basis of optimism bias  Chapter | 3  49

expected significantly more negative future outcomes than future positive outcomes. The alcoholic character was also perceived by the respondents as an extreme outgroup member, strongly suggesting that desires in the form of prejudice toward the reference group dictate the nature of the optimism bias (Dricu et al., 2018). Regarding the predictive power of desires on optimism bias, studies have shown that respondents manifest a larger bias for severe outcomes (either positive or negative) because such outcomes more readily generate affective reactions, which, in turn, create motivations to approach or avoid the respective event (Gold & Brown, 2009; Krizan & Windschitl, 2007; Lench, 2009; Lench & Bench, 2012). These motivations then translate cognitively into low likelihoods for future negative outcomes and high likelihoods for future positive outcomes. Similarly, individuals are motivated to view the risk likelihood of close friends as comparable with their own risk because it would be anxiety provoking to assume that close friends are at risk for negative events (Perloff & Fetzer, 1986). Although authors have traditionally preferred one account over the other, modern theories of optimism bias agree that cognitive and motivational accounts are not mutually exclusive. Instead, different causal mechanisms exert independent and interdependent effects in the production of comparative optimism (Chambers & Windschitl, 2004; Shepperd et al., 2002). Such a pluralistic view on the causes of optimism bias allows for the possibility of unique (either-or) and interactive effects of cognitive and motivational mechanisms in the expression of overoptimistic beliefs. This flexible perceptual system is able to exchange inferential tactics in response to changing goals and situational demands. Intricate interactions between cognitive and motivational accounts also make possible the emergence of social optimism bias, such as those displayed by sports fans and voters (e.g., Stiers & Dassonneville, 2018). Overoptimistic beliefs in these cases arise, on the one hand, from egocentric biases such as differential accessibility, as well as amounts of information that make it easier to retrieve information relevant to in-group members and in greater quantities than it is to retrieve information about out-group members (Babad, 1995; Castano, Yzerbyt, Bourguignon, & Seron, 2002; Hollander, 2004). On the other hand, by evaluating groups and individuals that one identifies with more favorably than others on a given attribute, one is able to maintain positive self-esteem via group membership and a distinct social identity (Braga, Mata, Ferreira, & Sherman, 2017; Krizan, Miller, & Johar, 2010; Tajfel, Billig, Bundy, & Flament, 1971). Research into the mechanisms and causes of optimistic belief updating is still in the early stages. At least partially overlapping mechanisms with comparative optimism bias might be responsible for why individuals disproportionately prefer to update their beliefs in the direction that is favorable to them whenever prompted with new information (Kuzmanovic et al., 2015; Sharot et al., 2011; Sharot, Kanai, et al., 2012). Curiously, the magnitude of this optimistic update bias fluctuates throughout life. Children, young adults, and older adults display optimistic tendencies in belief update, but the magnitude of the update bias is

50  Cognitive biases in health and psychiatric disorders

exacerbated in children (Moutsiana et al., 2013) and older adults (Chowdhury et  al., 2014) compared with that in young adults. This pattern is exclusively driven by a greater unwillingness to accept and learn from unfavorable information. In other words, people start by being less willing to adapt their beliefs when confronted with unfavorable information in childhood and then learn to counteract this tendency as young adults only to revert to their initial behavior as older adults. The mechanisms for these longitudinal differences are not yet well understood (Chowdhury et al., 2014; Moutsiana et al., 2013).

Methods used to investigate the optimism bias As described in section “Introduction” of this chapter, unrealistic optimism can be investigated in several ways, depending on the standard of reference (Shepperd et al., 2013; Shepperd et al., 2015). Unrealistic absolute optimism refers to a personal estimate for a future event that is optimistically inaccurate compared with an objective standard, such as a normative statistic provided by a research institute or the event itself. In a typical unrealistic comparative optimism experiment, participants are asked directly or indirectly about their perceived standing in regard to a personality trait, a skill, or the likelihood of experiencing a future event. In the direct method, respondents indicate their likelihood of experiencing an event compared with that of others or how their position on a target dimension (e.g., creativity) compares with that of others. In the indirect method, respondents make an absolute judgment separately about themselves and about another. The difference between the two absolute judgments then serves as an indirect measure. Studies that ask direct questions use ratio scores such as “80% less” or “40% more” than the average peer (Price et al., 2002; Weinstein, 1980) or, more often, ambiguous expressions such as “much below average” to “more above average,” which can be construed by respondents in terms of either ratio or difference (Klar & Ayal, 2004). In contrast, in the indirect method, the two separate likelihoods are commonly analyzed as difference scores. This has prompted some authors to suggest that the direct and indirect methods tap into distinct types of optimism (Shepperd et  al., 2013), which have their own separate mechanisms (Price et  al., 2002). One reason for this divergence might be that the method of collecting data interacts with an event’s perceived frequency; that is, questions in a ratio format may better capture comparative optimism for rare events, whereas methods that use difference scores may better capture optimism bias for common events (Klar & Ayal, 2004). In terms of methods of analysis, the standard practice of investigating the moderators of optimism bias has been (a) to use the event as the unit of analysis and (b) to perform correlations between the likelihood estimates averaged across participants and various event characteristics such as perceived controllability and personal experience (e.g., Price et  al., 2002; Weinstein, 1980). If a significant correlation is found, it is assumed that the event characteristic is

The neurophysiological basis of optimism bias  Chapter | 3  51

moderating the optimism bias. Of course, correlation is not causation, and in fact, the reverse relation is equally possible (i.e., participants perceive that they have more control over things about which they feel optimistic) (McKenna, 1993). Alternatively, multiple regression analyses between event characteristics and likelihood estimates are performed to causally investigate the drivers behind optimism bias (e.g., Chambers et al., 2003; Dricu et al., 2018). Much less often, analyses of variance are used with the participants as the unit of analysis and with the event characteristics as within-subject factors (e.g., Harris & Middleton, 1994; Lin, Lin, & Raghubir, 2004). It is not yet known how the method of analysis (e.g., regression vs analysis of variance) could affect the magnitude and direction of the optimism bias, as studies that simultaneously use multiple methods are scarce. For instance, analyses of variance make inferences about the respondents, whereas regression analyses allow inferences about the event characteristics, thus informing us about different underlying causes of optimism bias.

Brain regions involved in the emergence and maintenance of the bias From the generous behavioral data collected over the past 40 years, various hypotheses have been put forth regarding the brain structures and networks associated with the optimism bias. One prominent prediction is that optimism bias is associated with brain regions involved in autobiographical memory and prospective thinking (Sharot et al., 2007). This prediction rests on the assumption that assessing future likelihoods for oneself involves prospective thinking and imagination, which, in return, piggybacks on brain structures that are also involved in the processing of episodic memories (e.g., the ventromedial frontal cortex [vmPFC] extending into the anterior cingulate cortex [ACC], the precuneus extending into the posterior cingulate cortex [PCC], and the hippocampus; (Addis, Wong, & Schacter, 2007; Maddock, Garrett, & Buonocore, 2001; Nielsen, Balslev, & Hansen, 2005; Spreng, Mar, & Kim, 2009)). Perhaps unsurprisingly, the same regions are associated with self-referential processing, by which a person becomes aware that specific contents are related to his or her own self (Chavez, Heatherton, & Wagner, 2016; Meffert, Blanken, Blair, White, & Blair, 2013). Sharot and colleagues directly compared remembering the past and imagining one’s future and found that four regions in the brain were engaged, that is, the vmPFC/ACC, the precuneus/PCC, the hippocampus, and the dorsomedial frontal cortex (Sharot et al., 2007). Of these regions, the right hippocampus and the right vmPFC/ACC were significantly more active and functionally more connected during remembering the past than during imagining the future (main effect of task); they were also more active during desirable events than during undesirable events (main effect of valence). The main effect of task is likely caused by the vividness of the recalled information compared with the imagined scenarios (e.g., Addis et al., 2007), whereas the main effect

52  Cognitive biases in health and psychiatric disorders

of valence is likely the manifestation of a repeated simulation; that is, positive events, whether recalled or imagined, are in themselves rewarding and invite repeated simulation (Krizan & Windschitl, 2007). More important, activity in the vmPFC/ACC positively correlated with scores of trait optimism (Sharot et al., 2007), suggesting that optimistic respondents engaged more in remembering and imagining positive scenarios than they did in negative scenarios (for a review on the links between memory processes and optimism bias, see Kress & Aue, 2017). Although these findings, taken at face value, seem unrelated to the optimism bias phenomenon per se, they readily explain why these brain structures consistently appear in functional imaging contrasts between deliberating over desirable events versus undesirable events and between estimating event likelihoods for oneself versus others (Blair et  al., 2013; Blair et  al., 2017; Kuzmanovic, Jefferson, & Vogeley, 2016). For example, in an indirect comparative optimism task, Kuzmanovic, Jefferson, and Vogeley (2016) asked participants to estimate likelihoods of experiencing positive and negative events separately for oneself and for others and found a main effect of task (estimates for oneself vs others) characterized by enhanced activation in the bilateral vmPFC extending into the ACC. In a direct comparative optimism experiment, Blair et al. (2013) showed that participants engaged the precuneus extending into the PCC and the vmPFC extending into the ACC when deliberating over positive versus negative scenarios. The pattern of neural responses in the vmPFC/ACC was driven by reduced activity (compared with baseline) when participants assessed their likelihood for future negative events (positive events did not engage this region relative to baseline), whereas the pattern of neural responses in the precuneus/PCC was driven by increased activity (compared with baseline) when participants assessed their likelihoods for future positive events (negative events did not engage this region relative to baseline). Interestingly, these brain regions were unrelated to the optimism bias per se, that is, overestimating positive events and underestimating negative events. When the analysis included the estimate likelihoods as parametric modulators, the authors identified other regions: The right ventral ACC showed increased response to positive events as a function of the respondent’s increasing probability estimates, whereas the left insula and left dorsomedial frontal cortex showed increased deactivation to negative events as a function of decreasing probability estimates (Blair et al., 2013). The authors remarked that different brain structures might be involved, on the one hand, in setting up the blueprint for the optimism bias (e.g., providing the “raw materials” in the form of prospective thinking and episodic memories) and, on the other hand, generating the optimism bias itself. In a similar direct comparative optimism task, Blair et al. (2017) observed that, at the behavioral level, healthy participants manifested optimism bias for positive and negative events, whereas individuals with GAD did not. At a neural level, both groups similarly recruited the vmPFC/ACC and the precuneus/PCC when deliberating over positive future events compared with negative events, suggesting that both groups used

The neurophysiological basis of optimism bias  Chapter | 3  53

e­ pisodic memories and prospective thinking to perform the task. However, the use of the likelihood estimates as a parametric modulator in the analysis revealed that individuals with GAD displayed significantly decreased neural activity in the dorsal ACC compared with that in healthy respondents, indicating that optimism bias per se is generated elsewhere in the brain. In a structural imaging study, Chowdhury et al. (2014) found that the volume of dorsal and ventral ACC in healthy older adults correlated with the magnitude of optimistic update bias. This result is directly comparable with the functional findings of Blair et al. (2013) and Blair et al. (2017), thus bringing together two forms of optimism bias, that is, comparative optimism and optimistic belief update. Optimistic belief update, introduced in sections “Introduction” and “Major theories in the field”, manifests as a result of an asymmetrical update of beliefs in response to new information; that is, individuals are more willing to update their personal estimates of experiencing an event if the information received is favorable (e.g., incidence of a negative event in the general population is lower than estimated for oneself) than if it is unfavorable (e.g. incidence of a negative event in the general population is higher than estimated for oneself). In other words, the asymmetry is driven by the significantly smaller updates of belief in response to undesirable information. In a functional magnetic resonance imaging (fMRI) experiment on optimistic belief update, Sharot et al. (2011) found that blood oxygen level–dependent (BOLD) activity in the right inferior frontal cortex (IFC) (pars opercularis) decreased as a function of the magnitude of the undesirable information received (i.e., incidence of a negative event in the general population minus initial personal estimate for that event), whereas BOLD activity in the left IFC (pars orbitalis) and the left superior frontal cortex/medial frontal cortex increased. Furthermore, the authors correlated the BOLD signal change in these three regions with the size of the subsequent updates and found that the more reluctant the respondents were to update their initial estimates for negative events, the more the BOLD signal decreased in the right IFC (pars opercularis). No correlations were found between the size of the updates and the BOLD signal change in the left IFC (pars orbitalis) and the left superior frontal cortex/medial frontal cortex. In a subsequent transcranial magnetic stimulation experiment, Sharot, Kanai, et al. (2012) temporarily disrupted the activity of either the left or the right pars opercularis of the IFC while respondents performed a task of belief update. There were no significant differences in belief updates following the disruption of activity in the right pars opercularis, whereas disruption of the left pars opercularis led to the disappearance of the optimistic belief update altogether by boosting the update of beliefs in response to bad news (there was no effect on updates in response to good news). It remains an open question as to why disrupting the neural activity in the left but not the right pars opercularis (known for its general role in inhibitory control, e.g., Aron, Robbins, & Poldrack, 2014) led to these findings (Sharot, Kanai, et al., 2012). However, white-matter connectivity also supports an important role of the left

54  Cognitive biases in health and psychiatric disorders

but not the right pars opercularis in the optimistic update of beliefs (Moutsiana et al., 2015). Based on the assumptions that expecting/imagining low likelihoods of negative events and high likelihoods of positive events constitutes a reward in itself, another prediction is that optimism bias depends on the brain’s reward and valuation system (e.g., Sharot, Guitart-Masip, Korn, Chowdhury, & Dolan, 2012). In a simplistic model of how the brain processes rewards (e.g., the metaanalysis by Knutson and Greer (2008)), the orbitofrontal cortex (OFC) typically encodes the value of things (how desirable or undesirable something is; e.g., (Kringelbach & Rolls, 2004)), while activity in the striatum reflects the experience of receiving the reward and the instrumental learning that follows (e.g., Delgado, Locke, Stenger, & Fiez, 2003; O’doherty et  al., 2004). Optionally, the insula might be recruited to reflect the awareness of the valuation process, particularly for negative valuations (e.g., (Craig & Craig, 2009)). Neuroimaging studies have implicated these three regions in the manifestation of the optimism bias (striatum: (Aue et  al., 2011; Blair et  al., 2017; Charpentier, BrombergMartin, & Sharot, 2018; Guitart-Masip et al., 2012; Sharot, Guitart-Masip, et al., 2012); OFC and insula: (Blair et al., 2013; Chavez et al., 2016; Kuzmanovic, Rigoux, & Tittgemeyer, 2018)). In an elegant experiment, Chavez et al. (2016) found that evaluating oneself and evaluating positive stimuli share a common neural representation. Specifically, participants performed two tasks: rate the valence of a series of images and rate the degree to which a presented positive or negative adjective described either themselves or their close friend. Using multivariate pattern analysis, a classifier trained to dissociate neural responses to positive images from responses to negative images could also accurately dissociate thinking about oneself from thinking about a close friend. The performance for decoding the evaluation of oneself was most accurate from neural responses to positive images in the OFC. The authors speculated that the OFC might underlie positivity bias when evaluating own performance and abilities, that is, the motivation to hold favorable views of oneself. In two experiments on optimistic belief update (Kuzmanovic et  al., 2018; Kuzmanovic, Jefferson, & Vogeley, 2016), neural activity in the OFC tracked favorable updating; that is, the BOLD signal in the OFC positively correlated with the size of updates following good news (i.e., the base rate of the negative event was lower than initially assumed) but negatively correlated with the size of updates following bad news (i.e., the base rate of the negative event was higher than initially assumed). In other words, not only improving beliefs but also avoiding worsening beliefs (or refraining from updating) triggered the OFC activity. Moreover, respondents with the strongest optimistic belief update exhibited the greatest valence-tracking effect in the OFC (Kuzmanovic et al., 2018). Concerning the striatum, Sharot et  al. (2007) found that simply imaging and remembering positive events, compared with negative events, triggered higher activity in the dorsal striatum. In studies on optimistic belief update,

The neurophysiological basis of optimism bias  Chapter | 3  55

being confronted with favorable information (i.e., incidence of a negative event in the general population is lower than initially estimated for oneself) leads to extensive activation in the dorsal striatum (Kuzmanovic, Jefferson, & Vogeley, 2016). More important, Blair et  al. (2017) found that, compared with that in healthy participants, the BOLD signal in the dorsal striatum of respondents with GAD increased as a function of their decreasing likelihood estimates of these events. As respondents with GAD assessed their chances of experiencing positive events in the future as lower than their peers, activity in the dorsal striatum increased. Interestingly, individuals with GAD view worrying, the tenet of their disorder, in a positive manner as a preemptive compensatory strategy that prepares them for the future (Newman et al., 2013). If the low estimates of positive events are a proxy for worrying, one may be tempted to interpret the significant activation of the dorsal striatum as a function of such a compensatory strategy, reflecting the positive state that respondents with GAD experience when worrying about their chances of experiencing future positive events. This would mirror the behavioral results in which healthy participants displayed a comparative optimism bias, whereas respondents with GAD did not. In a study on comparative optimism bias, Blair et al. (2013) found that the respondents’ bias for negative events (i.e., decreased likelihoods of negative events occurring to oneself relative to others) was inversely correlated with activity in the insula; that is, as respondents considered future negative events more likely to occur to them, activity in the insula increased. Given the involvement of the insula in the awareness and anticipation of negative information (Craig & Craig, 2009; Knutson & Greer, 2008), the authors interpreted the respondents’ comparative optimism bias for negative events as a strategy to avoid thinking about negative outcomes (see also the behavioral work on how optimistic expectancies guide attention away from undesirable stimuli (Kress, Bristle, & Aue, 2018; Peters, Vieler, & Lautenbacher, 2016)). A similar relation between the BOLD signal change in the insula and optimism bias was also found by Kuzmanovic, Jefferson, and Vogeley (2016) in an experiment on belief updates. When respondents updated their initial estimates for negative events favorably (i.e., downward toward the received base rate), activity in the insula increased as the estimation error decreased (difference between one’s original estimate and the estimate received). In other words, the more the base rates corroborated the respondents’ initial comparative optimism bias, the higher the activity in the insula during the subsequent updates. Thus, it seems that (re) assessing negative events, especially when such a process concludes low likelihoods, triggers activation in the insula. In summary, several brain regions have been proposed to underlie comparative optimism bias and optimism bias in belief updates (see Fig. 1 for details). Key structures involved in remembering the past and imagining the future, such as the vmPFC/ACC and the precuneus/PCC, might provide the necessary input for deliberating over future events in the form of episodic memories and prospective thinking (Sharot et al., 2007; Spreng et al., 2009). Other brain

x = 12

x = –11

x = –5

x = 51

y = 13

y = –16

x = –51

Comparative optimism bias

Optimistic belief updating

FIG. 1  Summary of the most prominent brain regions underlying optimism bias. As an illustration, we created 4-mm regions of interest around the peak coordinates reported by the neuroimaging studies described in the text. All peaks have been transformed into the Montreal Neurological Institute (MNI) atlas with the Yale BioImage Suite Application (http://sprout022.sprout.yale.edu/mni2tal/mni2tal.html). Red corresponds to regions underlying general cognitive mechanisms necessary for comparative optimism bias, whereas purple depicts regions involved in generating the comparative optimism bias per se. Similarly, cyan corresponds to brain regions underlying general cognitive mechanisms necessary for optimistic belief updating, whereas blue depicts regions involved in generating the optimistic bias in belief updating. vACC, ventral anterior cingulate cortex; dACC, dorsal anterior cingulate cortex; vmPFC, ventromedial prefrontal cortex; OFC, orbitofrontal cortex; Prec/PCC, precuneus extending into the posterior cingulate cortex; IFC, inferior frontal cortex; dStr, dorsal striatum; Ins, insula.

56  Cognitive biases in health and psychiatric disorders

x=7

The neurophysiological basis of optimism bias  Chapter | 3  57

regions might then be responsible for generating the optimism bias per se by interfering with how one represents the subjective value of stimuli or how one thinks about oneself (Blair et al., 2013; Blair et al., 2017; Kuzmanovic et al., 2018; Kuzmanovic, Jefferson, & Vogeley, 2016). More research is needed into the neural substrates of social optimism bias. However, one study asked participants to evaluate the likelihood estimates of their favorite team, of a neutral team, and of their least favorite team winning an important game and found that the dorsal striatum played a crucial role in biasing respondents’ attention and in generating the social optimism bias (Aue et  al., 2011). Specifically, the magnitude of the optimism bias in their behavior correlated to the degree of differential connectivity between the dorsal striatum and regions of visual perception and attention.

Somatovisceral responses related to optimism bias To our knowledge, no studies have so far directly investigated somatovisceral responses related to optimism bias. However, dispositional optimism has been linked to cardiovascular health and improved cardiac function (e.g., parasympathetic nervous system activation, which may protect against various forms of cardiac disease; see (Boehm & Kubzansky, 2012)). Furthermore, dispositional optimism has been associated with stressor-related immune changes (Segerstrom, 2001). Specifically, in the absence of goal conflict, optimists display lower immune responses than pessimists do. By contrast, when facing two conflicting goals, optimists who are likely to engage in both goals and therefore experience short-term stress display higher immune responses than pessimists do. Thus, even though optimism may generally protect cardiovascular health and contribute to immunological resilience, optimists may face short-term physiological costs in their persistence to gain long-term rewards. Although these findings on dispositional optimism do not directly translate to optimism bias, they can inspire future investigations of potential somatovisceral responses associated with optimism bias. Similarly, the anticipation of positive emotional and rewarding information resembles processes that are present during the formation of optimistic expectancies. Therefore, somatovisceral responses associated with the anticipation of positive and rewarding information can reveal first hints on the physiological mechanisms underlying optimism bias. For instance, anticipating positive emotional pictures elicits higher activity of the M. zygomaticus major and stronger startle responses measured through electromyography than anticipating neutral pictures does (Sabatinelli, Bradley, & Lang, 2001; Schumacher et  al., 2015; Sege, Bradley, & Lang, 2014). Whereas zygomaticus responses (associated with raising the lip corners) are typically enhanced when processing positive information (Dimberg, 1982, 1986) and may thus play an important role when anticipating them, startle responses are usually reduced when viewing positive emotional pictures (Sabatinelli et al., 2001).

58  Cognitive biases in health and psychiatric disorders

Furthermore, skin conductance—a sensitive measure of arousal (Bradley, Cuthbert, & Lang, 1990)—is enhanced when humans anticipate positive compared with neutral pictures (Schumacher et al., 2015) and when monkeys anticipate reward (Amiez, Procyk, Honoré, Sequeira, & Joseph, 2003). However, heart rate (which usually accelerates during reward processing; (Fowles, Fisher, & Tranel, 1982)) does not display a consistent pattern during the anticipation of positive information. Whereas the anticipation of positive emotional pictures was associated with heart rate acceleration in one study (Schumacher et  al., 2015), it was associated with heart rate deceleration in another (Poli, Sarlo, Bortoletto, Buodo, & Palomba, 2007). In sum, enhanced zygomaticus, startle, and skin conductance responses, as well as altered heart rate underlying the anticipation of positive information, may inform about potential processes that are present in optimism bias. Interestingly, Schumacher et al. (2015) point out that when people were uncertain about the valence of an anticipated picture in their study, the pattern of physiological responses resembled the anticipation of negative instead of positive pictures and might thus support the idea of a pessimism bias (i.e., a tendency to anticipate a negative outcome in uncertain situations; note, however, that neither optimism nor pessimism bias was directly investigated in their study). This idea contradicts behavioral findings that showed that people are often overly optimistic in uncertain situations. Thus, it is essential to directly investigate somatovisceral responses underlying optimism bias and compare corresponding behavioral and physiological responses in future studies.

Similarities and differences between healthy and clinical populations Whereas comparative optimism bias and asymmetrical optimistic belief update have been demonstrated in nonclinical populations, patients with mental disorders show diverging patterns (see Fig. 2 for details). When compared with the average other, individuals with obsessive compulsive disorder (OCD) predominantly manifest a pessimism bias concerning general positive and negative events, as well as specific OCD-related negative events (Moritz & Jelinek, 2009; Zetsche, Rief, & Exner, 2015). Interestingly, this pessimism bias only emerges in the indirect method of framing the question, that is, personal likelihoods and others’ likelihoods are estimated separately, suggesting that the bias is mostly an unconscious phenomenon in OCD individuals. Patients with clinical depression are more often reported with a pessimistic bias (Strunk, Lopez, & DeRubeis, 2006), although instances of comparative “realism” have also been reported, that is, expecting similar likelihoods for oneself and for others of experiencing positive and negative outcomes (Alloy & Ahrens, 1987). A meta-analysis has found evidence that both state and trait anxiety are generally associated with decreased optimism bias (Helweg-Larsen &

The neurophysiological basis of optimism bias  Chapter | 3  59

FIG. 2  Schematic overview of the findings on optimism bias in clinical populations. The top figure plots clinical populations along a continuum from pessimism bias (i.e., expecting more negative outcomes and/or less positive outcomes for oneself) to optimism bias (i.e., expecting more positive outcomes and/or less negative outcomes for oneself). The middle point reflects absence of bias. The bottom figure plots clinical populations along a continuum from pessimistic update of beliefs (i.e., individuals update beliefs more readily in response to bad news than to good news) to optimistic update of beliefs (i.e., individuals update beliefs more readily in response to good news than to bad news). The middle point reflects a symmetrical update of beliefs.

Shepperd, 2001), suggesting that optimism bias might be absent in anxiety disorders. However, the picture is more nuanced. For example, Blair and colleagues simultaneously assessed comparative optimism bias in individuals with social anxiety disorder (SAD), in individuals with GAD and in healthy controls (Blair et al., 2017). The majority of the future events used by the authors were not social in nature, for example, having a heart attack or winning the lottery, thus giving the individuals with SAD the opportunity to assess their likelihoods in a similar manner to that of healthy controls. Indeed, both healthy controls and individuals with SAD were characterized by optimistically biased responses. By contrast, responses in individuals with GAD depended on the valence of the scenarios presented. These individuals displayed comparative optimism for negative events only. For the positive events, instead, respondents with GAD expected similar likelihoods for themselves as for others of the same age and gender. More studies are needed to disentangle this surprising pattern of results, but one can speculate that it might be driven by the preferential manner in which the respondents with GAD in that study processed low-impact (e.g., getting a hug) versus high-impact events (e.g., winning the lottery). Such responding may reflect these individuals’ tendency to worry over mundane events that may otherwise be considered neutrally valenced (Newman et al., 2013).

60  Cognitive biases in health and psychiatric disorders

Regarding belief updating, depressed individuals exhibit a symmetrical update of beliefs in response to good news and bad news (Garrett et  al., 2014; Korn, Sharot, Walter, Heekeren, & Dolan, 2014), as do individuals with highfunctioning autism spectrum disorder (De Martino, Harrison, Knafo, Bird, & Dolan, 2008; Kuzmanovic, Rigoux, & Vogeley, 2016). However, the underlying cognitive processes are likely different. Individuals with an autism spectrum disorder are known for solving tasks in a literal manner and failing to incorporate emotional context into the decision-making process (De Martino et al., 2008; A. Harris et al., 2013). By contrast, the symmetrical update of beliefs in depressed individuals might stem from a pronounced readiness to update beliefs about negative outcomes. In the healthy population, the asymmetrical update of beliefs in the optimistic direction is driven exclusively by the reluctance to update in the face of bad news (Kuzmanovic & Rigoux, 2017; Sharot, Kanai, et al., 2012). Depressed individuals do not exhibit this reluctance. The update of beliefs in the face of good news is intact in them (Garrett et al., 2014); however, when faced with bad news regarding their predictions, they promptly adapt their initial estimates to the feedback (Garrett et al., 2014; Korn et al., 2014). An interesting pattern of findings that connects the research on optimism bias and optimistic update of beliefs concerns patients with borderline personality disorder. Despite showcasing a pronounced pessimism bias, these individuals are able to overcome the pessimism when provided with relevant information by promptly lowering their expectations for negative outcomes following good news (Korn, La Rosée, Heekeren, & Roepke, 2016). In sum, only a few studies have investigated the presence of optimism bias in clinical populations. Even though the evidence so far suggests that several clinical populations largely do not display optimism bias, further investigations are needed to corroborate this evidence. Future clinical studies could additionally examine patients who likely display exaggerated forms of optimism bias. For instance, exaggerated optimism bias may be present during the manic phases of patients with bipolar disorder. These exaggerated forms may also be critical in individuals with substance abuse who justify their continued dysfunctional behavior via exaggerated perception of short-term benefits of substance use compared with the lower perception of associated long-term risks (Dillard, Midboe, & Klein, 2009; Goldberg & Fischhoff, 2000; Weinstein, Marcus, & Moser, 2005).

Limitations As with any domain of research, the literature on optimism bias has several limitations that need to be highlighted so that appropriate steps can be taken to address them. On the one hand, different terms have been used to describe the same psychological phenomenon (e.g., wishful thinking, unrealistic optimism, comparative optimism, and overoptimism), while, on the other hand, the same terms have been used for slightly different phenomena (Kress & Aue, 2017;

The neurophysiological basis of optimism bias  Chapter | 3  61

Shepperd et al., 2013). To ensure comparability among experiments and interpretations, investigators need a common framework to discuss the optimism bias. Furthermore, studies on optimism bias have almost exclusively focused on negative events. Incorporating positive and neutral events would better delineate the extent and mechanisms of the optimism bias. Several other limitations and methodological issues have been voiced (Chambers & Windschitl, 2004; Harris & Hahn, 2011; Shepperd, Pogge, & Howell, 2017; Windschitl & Stuart, 2015). In section “Major theories in the field”, we discussed biased sampling of events, for example, the disproportionate selection of rare and frequent events and controllable and uncontrollable events. The type of response format that respondents use to give their answers can also influence the findings. A narrow scale (e.g., a seven-point scale from “much less likely = -3” to “much more likely = +3”) can artificially inflate the magnitude of overoptimistic beliefs by not considering more nuanced differences in likelihoods (Harris & Hahn, 2011; Shepperd et al., 2017). Furthermore, the explicit nature of the direct method of assessing optimism bias may overstate the incidence of optimism bias compared with that found with the indirect method (Chambers & Windschitl, 2004). Minority underrepresentation refers to the nonrepresentative sampling of respondents that could lead, in combination with using too many rare events, to an overall disproportionate number of respondents who objectively do not experience the target events. Consequently, the overall pool of respondents might seem biased toward being overoptimistic, when, in fact, they could be accurate in their predictions (Windschitl & Stuart, 2015). Above and beyond these limitations, neuroimaging studies suffer from additional shortcomings. To begin with, such studies have assigned unequal focuses on some types of optimism bias while neglecting others. To our knowledge, only one study has looked into social optimism bias (Aue et al., 2011), and two studies have investigated comparative optimism bias (Blair et al., 2013; Blair et al., 2017). Six studies have investigated optimistic belief updating (Chowdhury et al., 2014; Kuzmanovic et al., 2015; Kuzmanovic, Jefferson, & Vogeley, 2016; Moutsiana et al., 2015; Sharot et al., 2011; Sharot, Kanai, et al., 2012). Clearly, a diverse portfolio of studies would provide a better understanding of the phenomena of optimism bias. Similarly, there is often unwarranted focus on particular brain regions in the interpretation of the results with considerable neglect of other brain structures that are simultaneously reported (Blair et  al., 2017; Kuzmanovic, Jefferson, & Vogeley, 2016; Sharot et  al., 2011). Furthermore, the majority of neuroimaging studies on optimism bias have reported deactivations, that is, BOLD signal changes downward away from the baseline. While researchers know with comfortable certainty that activation in the brain (i.e., increase of the BOLD signal) roughly corresponds to an increase in neuronal activity in the respective regions (Logothetis, 2008), the interpretation of deactivation is far more contentious (Frankenstein et al., 2003; Hayes & Huxtable, 2012). Therefore, studies encountering deactivations should be more grounded

62  Cognitive biases in health and psychiatric disorders

with their interpretation since multiple explanations may exist. A related limitation of neuroimaging studies, in general, is the convoluted method of analysis. The human mind is a complex processing machine, and complex analyses are understandably required to further the understanding of it. However, such analyses can also lead to results that are difficult to interpret or, worse, that can lead to spurious findings. The fMRI remains an indirect measure of brain activity that limits the conclusions drawn from its findings (Farah, 2014; Logothetis, 2008). Using a convoluted method of analysis further abstracts the level of information, increasing the gap between the underlying brain activity and the interpretations drawn from it. For example, studies on asymmetrical belief updates routinely regress brain activity on sophisticated formulas that serve as proxies for the optimistic update of beliefs, such as the product of the estimation error (difference between the initial estimate and feedback received), the learning rate (general tendency of the respondent to update the initial estimate, regardless of the valence of the feedback received), and the personal relevance of the target events (Kuzmanovic et al., 2018). Consequently, researchers should be careful with their interpretations, and the reader should view those findings only within their specific analyses. Lastly, inconsistent or improper anatomical labeling of brain activation can also be detrimental. When in doubt, the researcher should use the label that is more confined, such as the OFC instead of the medial frontal cortex or the pars opercularis instead of the IFC. For example, activation in the IFC has been found in several neuroimaging studies on optimistic belief updating (Moutsiana et al., 2015; Sharot et al., 2011; Sharot, Kanai, et al., 2012). The IFC is a gyral complex consisting of eight structural and functional subregions in the left and the right hemispheres (Clos, Amunts, Laird, Fox, & Eickhoff, 2013; Foundas, Eure, Luevano, & Weinberger, 1998; Petrides & Pandya, 2002). However, the heterogeneity of the IFC has been overlooked in the literature on optimistic belief updating, which has favored the generic label “inferior frontal gyrus.” It is imperative that researchers locate and label their activations more precisely, as their choice will influence the search for precedent in the literature and all subsequent interpretations. Readers are also advised to go beyond the verbal labels chosen by the original authors when bridging results across studies. Whenever possible, the size and the position of the entire cluster of activation should be considered instead of the reported peak coordinate.

Future directions On the basis of the heterogeneous definitions of optimism bias used by the various theories in the field (section “Major theories in the field”) and the methodological shortcomings of previous empirical research on optimism bias (section “Limitations”), it is important that future research use consistent terminology and standardized methods to examine the bias (Garrett et al., 2014; Shah et al., 2016). The use of a large sample of both positive and negative events that has

The neurophysiological basis of optimism bias  Chapter | 3  63

been matched on key characteristics such as perceived controllability, frequency, and emotional impact is crucial to determine when and how optimism bias manifests (Dricu et al., 2018). To improve future theories and methodologies in the field, researchers should, for instance, clearly describe how optimism bias has been measured or induced in a particular study (e.g., which exact items were used). Furthermore, future research needs to directly investigate causal relations between optimism bias and mental health/psychopathology. Such research is critical to identifying causes and consequences and can therefore greatly advance the development of interventions that aim to change optimism bias (e.g., to induce it in patients with depression or reduce it in patients with mania). First attempts to influence optimism bias through pharmacological interventions (such as L-DOPA; (Sharot, Guitart-Masip, et al., 2012)) or neurostimulation (such as transcranial magnetic stimulation; (Sharot, Kanai, et al., 2012)) have already proven to be successful. Furthermore, behavioral interventions comprising cognitive trainings (e.g., mental imagery training) should be further developed to advance noninvasive and easily deliverable methods of modulating optimism (Blackwell et al., 2013; Murphy et al., 2015). Most important, it is essential that future research does not examine optimism bias and its neurophysiological mechanism in isolation. Instead, ­ additional positive cognitive biases (e.g., interpretation bias, attention bias, and memory bias) should be considered when investigating optimism bias to uncover cognitive mechanisms involved in the emergence and maintenance of overly optimistic expectancies (Kress & Aue, 2017). Different cognitive biases have been suggested to interact and mutually enforce each other by the combined cognitive biases hypothesis (Hirsch, Clark, & Mathews, 2006); see Aue and Okon-Singer (2015) for an extension with respect to expectancies. Notably, empirical research has already revealed that positive attention bias dynamically interacts with optimism bias, thus providing first empirical support for cognitive bias interactions in the positive domain (Kress et al., 2018). Similar causal relations to positive biases in memory or interpretation can yield a more elaborative view on multiple cognitive bias interactions underlying healthy information processing.

Summary Optimism bias describes people’s tendency to overestimate their likelihood to experience positive events and underestimate their likelihood to experience negative events in the future. Such an optimistic outlook on the future can enhance their motivation to engage in self-relevant and difficult situations and make it more likely to obtain rewards. Theoretical considerations vary in the extent to which they claim the existence of an optimism bias in the general population. Even though some researchers suggest that optimism bias exists in up to 80% of the population, others argue that it may merely represent other,

64  Cognitive biases in health and psychiatric disorders

closely related cognitive phenomena (e.g., illusion of control). Whereas neuroimaging studies on optimism bias have revealed some involvement of cingulate and prefrontal brain areas in the emergence of the bias, there are no studies on the somatovisceral processes related to optimism bias so far. Interestingly, first attempts to compare the extent of optimism bias displayed by healthy people and patients with mental disorders reveal an absence of optimism bias in various patient groups (e.g., patients with depression, borderline personality disorder, and OCD). However, the few findings on optimism bias in patients with mental disorders must be further supported by future research. Such research would greatly benefit from the use of more consistent terminology and more reliable and rigorous methods to target some of the important limitations that present and previous research on optimism bias suffers from. In addition, future research on optimism bias should take into account potential interactions with other positive cognitive biases (e.g., in attention or memory) to yield a more elaborative view on healthy information processing.

References Addis, D. R., Wong, A. T., & Schacter, D. L. (2007). Remembering the past and imagining the future: Common and distinct neural substrates during event construction and elaboration. Neuropsychologia, 45(7), 1363–1377. Alloy, L. B., & Ahrens, A. H. (1987). Depression and pessimism for the future: Biased use of statistically relevant information in predictions for self versus others. Journal of Personality and Social Psychology, 52(2), 366. Amiez, C., Procyk, E., Honoré, J., Sequeira, H., & Joseph, J.-P. (2003). Reward anticipation, cognition, and electrodermal activity in the conditioned monkey. Experimental Brain Research, 149(3), 267–275. Armor, D. A., & Taylor, S. E. (1998). Situated optimism: Specific outcome expectancies and selfregulation. In Advances in Experimental Social Psychology (Vol. 30, pp. 309–379): Elsevier. Aron, A. R., Robbins, T. W., & Poldrack, R. A. (2014). Inhibition and the right inferior frontal cortex: One decade on. Trends in Cognitive Sciences, 18(4), 177–185. Aspinwall, L. G., & Brunhart, S. M. (1996). Distinguishing optimism from denial: Optimistic beliefs predict attention to health threats. Personality and Social Psychology Bulletin, 22(10), 993–1003. Aue, T., Nusbaum, H. C., & Cacioppo, J. T. (2011). Neural correlates of wishful thinking. Social Cognitive and Affective Neuroscience, 7(8), 991–1000. Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. Babad, E. (1987). Wishful thinking and objectivity among sports fans. Social Behaviour. Babad, E. (1995). Can accurate knowledge reduce wishful thinking in voters’ predictions of election outcomes? The Journal of Psychology, 129(3), 285–300. Babad, E. (1997). Wishful thinking among voters: Motivational and cognitive influences. International Journal of Public Opinion Research, 9(2), 105–125. Babad, E., & Katz, Y. (1991). Wishful thinking—Against all odds. Journal of Applied Social Psychology, 21(23), 1921–1938. Blackwell, S. E., Rius-Ottenheim, N., Schulte-van Maaren, Y. W., Carlier, I. V., Middelkoop, V. D., Zitman, F. G., et al. (2013). Optimism and mental imagery: A possible cognitive marker to promote well-being? Psychiatry Research, 206(1), 56–61.

The neurophysiological basis of optimism bias  Chapter | 3  65 Blair, K. S., Otero, M., Teng, C., Geraci, M., Ernst, M., Blair, R., et al. (2017). Reduced optimism and a heightened neural response to everyday worries are specific to generalized anxiety disorder, and not seen in social anxiety. Psychological Medicine, 47(10), 1806–1815. Blair, K. S., Otero, M., Teng, C., Jacobs, M., Odenheimer, S., Pine, D. S., et al. (2013). Dissociable roles of ventromedial prefrontal cortex (vmPFC) and rostral anterior cingulate cortex (rACC) in value representation and optimistic bias. NeuroImage, 78, 103–110. Boehm, J. K., & Kubzansky, L. D. (2012). The heart’s content: The association between positive psychological well-being and cardiovascular health. Psychological Bulletin, 138(4), 655. Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1990). Startle reflex modification: Emotion or attention? Psychophysiology, 27(5), 513–522. Braga, J. P., Mata, A., Ferreira, M. B., & Sherman, S. J. (2017). Motivated reasoning in the prediction of sports outcomes and the belief in the “hot hand”. Cognition and Emotion, 31(8), 1571–1580. Budescu, D. V., & Wallsten, T. S. (1985). Consistency in interpretation of probabilistic phrases. Organizational Behavior and Human Decision Processes, 36(3), 391–405. Campbell, J., Greenauer, N., Macaluso, K., & End, C. (2007). Unrealistic optimism in internet events. Computers in Human Behavior, 23(3), 1273–1284. Castano, E., Yzerbyt, V., Bourguignon, D., & Seron, E. (2002). Who may enter? The impact of in-group identification on in-group/out-group categorization. Journal of Experimental Social Psychology, 38(3), 315–322. Chambers, J. R. (2008). Explaining false uniqueness: Why we are both better and worse than others. Social and Personality Psychology Compass, 2(2), 878–894. Chambers, J. R., & Windschitl, P. D. (2004). Biases in social comparative judgments: The role of nonmotivated factors in above-average and comparative-optimism effects. Psychological Bulletin, 130(5), 813. Chambers, J. R., Windschitl, P. D., & Suls, J. (2003). Egocentrism, event frequency, and comparative optimism: When what happens frequently is “more likely to happen to me”. Personality and Social Psychology Bulletin, 29(11), 1343–1356. Charpentier, C. J., Bromberg-Martin, E. S., & Sharot, T. (2018). Valuation of knowledge and ignorance in mesolimbic reward circuitry. Proceedings of the National Academy of Sciences. 201800547. Chavez, R. S., Heatherton, T. F., & Wagner, D. D. (2016). Neural population decoding reveals the intrinsic positivity of the self. Cerebral Cortex, 27(11), 5222–5229. Chowdhury, R., Sharot, T., Wolfe, T., Düzel, E., & Dolan, R. J. (2014). Optimistic update bias increases in older age. Psychological Medicine, 44(9), 2003–2012. Clos, M., Amunts, K., Laird, A. R., Fox, P. T., & Eickhoff, S. B. (2013). Tackling the multifunctional nature of Broca’s region meta-analytically: Co-activation-based parcellation of area 44. NeuroImage, 83, 174–188. Craig, A. D., & Craig, A. (2009). How do you feel--now? The anterior insula and human awareness. Nature Reviews Neuroscience, 10(1). Davidson, R. J. (2004). Well-being and affective style: Neural substrates and biobehavioural correlates. Philosophical Transactions of the Royal Society, B: Biological Sciences, 359(1449), 1395. De Martino, B., Harrison, N. A., Knafo, S., Bird, G., & Dolan, R. J. (2008). Explaining enhanced logical consistency during decision making in autism. Journal of Neuroscience, 28(42), 10746–10750. Delgado, M., Locke, H., Stenger, V. A., & Fiez, J. (2003). Dorsal striatum responses to reward and punishment: Effects of valence and magnitude manipulations. Cognitive, Affective, & Behavioral Neuroscience, 3(1), 27–38.

66  Cognitive biases in health and psychiatric disorders Dillard, A. J., Midboe, A. M., & Klein, W. M. (2009). The dark side of optimism: Unrealistic optimism about problems with alcohol predicts subsequent negative event experiences. Personality and Social Psychology Bulletin, 35(11), 1540–1550. Dimberg, U. (1982). Facial reactions to facial expressions. Psychophysiology, 19(6), 643–647. Dimberg, U. (1986). Facial reactions to fear-relevant and fear-irrelevant stimuli. Biological Psychology, 23(2), 153–161. Dricu, M., Bührer, S., Hesse, F., Eder, C., Andres, P., & Aue, T. (2018). Warmth and competence predict overoptimistic beliefs for out-group but not in-group members. PLoS One, 13(11), e0207670. Farah, M. J. (2014). Brain images, babies, and bathwater: Critiquing critiques of functional neuroimaging. Hastings Center Report, 44(s2), S19–S30. Foundas, A. L., Eure, K. F., Luevano, L. F., & Weinberger, D. R. (1998). MRI asymmetries of Broca’s area: The pars triangularis and pars opercularis. Brain and Language, 64(3), 282–296. Fowles, D. C., Fisher, A. E., & Tranel, D. T. (1982). The heart beats to reward: The effect of monetary incentive on heart rate. Psychophysiology, 19(5), 506–513. Fox, E. (1993). Allocation of visual attention and anxiety. Cognition and Emotion, 7(2), 207–215. Frankenstein, U., Wennerberg, A., Richter, W., Bernstein, C., Morden, D., Rémy, F., et al. (2003). Activation and deactivation in blood oxygenation level dependent functional magnetic resonance imaging. Concepts in Magnetic Resonance Part A: An Educational Journal, 16(1), 63–70. Garrett, N., Sharot, T., Faulkner, P., Korn, C. W., Roiser, J. P., & Dolan, R. J. (2014). Losing the rose tinted glasses: Neural substrates of unbiased belief updating in depression. Frontiers in Human Neuroscience, 8, 639. Gold, R. S., & Brown, M. G. (2009). Explaining the effect of event valence on unrealistic optimism. Psychology, Health & Medicine, 14(3), 262–272. Goldberg, J., & Fischhoff, B. (2000). The long-term risks in the short-term benefits: Perceptions of potentially addictive activities. Health Psychology, 19(3), 299. Guitart-Masip, M., Chowdhury, R., Sharot, T., Dayan, P., Duzel, E., & Dolan, R. J. (2012). Action controls dopaminergic enhancement of reward representations. Proceedings of the National Academy of Sciences. 201202229. Harris, L. T., Cikara, M., & Fiske, S. T. (2008). Envy as predicted by the stereotype content model: Volatile ambivalence. In Envy: Theory and research. New York, NY, US: Oxford University Press. Harris, A. J., & Hahn, U. (2011). Unrealistic optimism about future life events: A cautionary note. Psychological Review, 118(1), 135. Harris, P., & Middleton, W. (1994). The illusion of control and optimism about health: On being less at risk but no more in control than others. British Journal of Social Psychology, 33(4), 369–386. Harris, P., Middleton, W., & Joiner, R. (2000). The typical student as an in‐group member: Eliminating optimistic bias by reducing social distance. European Journal of Social Psychology, 30(2), 235–253. Harris, A., Shah, P., Catmur, C., Bird, G., & Hahn, U. (2013). Autism, optimism and positive events: Evidence against a general optimistic bias. Paper presented at the proceedings of the annual meeting of the cognitive science society. Hayes, D. J., & Huxtable, A. G. (2012). Interpreting deactivations in neuroimaging. Frontiers in Psychology, 3, 27. Helweg-Larsen, M., & Shepperd, J. A. (2001). Do moderators of the optimistic bias affect personal or target risk estimates? A review of the literature. Personality and Social Psychology Review, 5(1), 74–95.

The neurophysiological basis of optimism bias  Chapter | 3  67 Hirsch, C. R., Clark, D. M., & Mathews, A. (2006). Imagery and interpretations in social phobia: Support for the combined cognitive biases hypothesis. Behavior Therapy, 37(3), 223–236. Hoch, S. J. (1985). Counterfactual reasoning and accuracy in predicting personal events. Journal of Experimental Psychology: Learning, Memory, and Cognition, 11(4), 719. Hollander, B. (2004). People think like me: Religion and wishful thinking in the 2000 US presidential election. Journal of Media and Religion, 3(4), 187–197. Horswill, M. S., & McKenna, F. P. (1999). The effect of perceived control on risk taking. Journal of Applied Social Psychology, 29(2), 377–391. Jenkins, S. C., Harris, A. J., & Lark, R. (2018). Understanding ‘unlikely (20% likelihood)‘or ‘20% likelihood (unlikely)‘ outcomes: The robustness of the extremity effect. Journal of Behavioral Decision Making. Klar, Y., & Ayal, S. (2004). Event frequency and comparative optimism: Another look at the indirect elicitation method of self-others risks. Journal of Experimental Social Psychology, 40(6), 805–814. Klar, Y., Medding, A., & Sarel, D. (1996). Nonunique invulnerability: Singular versus distributional probabilities and unrealistic optimism in comparative risk judgments. Organizational Behavior and Human Decision Processes, 67(2), 229–245. Klein, C. T., & Helweg-Larsen, M. (2002). Perceived control and the optimistic bias: A meta-­ analytic review. Psychology and Health, 17(4), 437–446. Knutson, B., & Greer, S. M. (2008). Anticipatory affect: Neural correlates and consequences for choice. Philosophical Transactions of the Royal Society of London B: Biological Sciences, 363(1511), 3771–3786. Korn, C. W., La Rosée, L., Heekeren, H. R., & Roepke, S. (2016). Processing of information about future life events in borderline personality disorder. Psychiatry Research, 246, 719–724. Korn, C., Sharot, T., Walter, H., Heekeren, H., & Dolan, R. (2014). Depression is related to an absence of optimistically biased belief updating about future life events. Psychological Medicine, 44(3), 579–592. Kress, L., & Aue, T. (2017). The link between optimism bias and attention bias: A neurocognitive perspective. Neuroscience & Biobehavioral Reviews, 80, 688–702. Kress, L., Bristle, M., & Aue, T. (2018). Seeing through rose-colored glasses: How optimistic expectancies guide visual attention. PLoS One, 13(2), e0193311. Kringelbach, M. L., & Rolls, E. T. (2004). The functional neuroanatomy of the human orbitofrontal cortex: Evidence from neuroimaging and neuropsychology. Progress in Neurobiology, 72(5), 341–372. Krizan, Z., Miller, J. C., & Johar, O. (2010). Wishful thinking in the 2008 US presidential election. Psychological Science, 21(1), 140–146. Krizan, Z., & Windschitl, P. D. (2007). The influence of outcome desirability on optimism. Psychological Bulletin, 133(1), 95. Krizan, Z., & Windschitl, P. D. (2009). Wishful thinking about the future: Does desire impact optimism? Social and Personality Psychology Compass, 3(3), 227–243. Kuzmanovic, B., Jefferson, A., & Vogeley, K. (2015). Self‐specific optimism Bias in belief updating is associated with high trait optimism. Journal of Behavioral Decision Making, 28(3), 281–293. Kuzmanovic, B., Jefferson, A., & Vogeley, K. (2016). The role of the neural reward circuitry in selfreferential optimistic belief updates. NeuroImage, 133, 151–162. Kuzmanovic, B., & Rigoux, L. (2017). Valence-dependent belief updating: Computational validation. Frontiers in Psychology, 8, 1087. Kuzmanovic, B., Rigoux, L., & Tittgemeyer, M. (2018). Influence of vmPFC on dmPFC predicts valence-guided belief formation. Journal of Neuroscience, 38(37), 7996–8010.

68  Cognitive biases in health and psychiatric disorders Kuzmanovic, B., Rigoux, L., & Vogeley, K. (2016). Brief report: Reduced optimism bias in selfreferential belief updating in high-functioning autism. Journal of Autism and Developmental Disorders, 1–9. Lench, H. C. (2009). Automatic optimism: The affective basis of judgments about the likelihood of future events. Journal of Experimental Psychology: General, 138(2), 187. Lench, H. C., & Bench, S. W. (2012). Automatic optimism: Why people assume their futures will be bright. Social and Personality Psychology Compass, 6(4), 347–360. Lench, H. C., & Ditto, P. H. (2008). Automatic optimism: Biased use of base rate information for positive and negative events. Journal of Experimental Social Psychology, 44(3), 631–639. Lin, C.-H., Lin, Y.-C., & Raghubir, P. (2004). The interaction between order of elicitation and event controllability on the self-positivity bias. In NA – Advances in Consumer Research (pp. 523– 529). Valdosta, GA, US: Association for Consumer Research. Logothetis, N. K. (2008). What we can do and what we cannot do with fMRI. Nature, 453(7197), 869. Maddock, R. J., Garrett, A. S., & Buonocore, M. H. (2001). Remembering familiar people: The posterior cingulate cortex and autobiographical memory retrieval. Neuroscience, 104(3), 667–676. Marks, J., & Baines, S. (2017). Optimistic belief updating despite inclusion of positive events. Learning and Motivation, 58, 88–101. McKenna, F. P. (1993). It won’t happen to me: Unrealistic optimism or illusion of control? British Journal of Psychology, 84(1), 39–50. Meffert, H., Blanken, L., Blair, K. S., White, S. F., & Blair, J. R. (2013). The influence of valence and decision difficulty on self-referential processing. Frontiers in Human Neuroscience, 7, 46. Moritz, S., & Jelinek, L. (2009). Inversion of the “unrealistic optimism” bias contributes to overestimation of threat in obsessive-compulsive disorder. Behavioural and Cognitive Psychotherapy, 37(2), 179–193. Moutsiana, C., Charpentier, C. J., Garrett, N., Cohen, M. X., & Sharot, T. (2015). Human frontal–­ subcortical circuit and asymmetric belief updating. Journal of Neuroscience, 35(42), 14077–14085. Moutsiana, C., Garrett, N., Clarke, R. C., Lotto, R. B., Blakemore, S.-J., & Sharot, T. (2013). Human development of the ability to learn from bad news. Proceedings of the National Academy of Sciences, 110(41), 16396–16401. Murphy, S. E., O’Donoghue, M. C., Drazich, E. H., Blackwell, S. E., Nobre, A. C., & Holmes, E. A. (2015). Imagining a brighter future: The effect of positive imagery training on mood, prospective mental imagery and emotional bias in older adults. Psychiatry Research, 230(1), 36–43. Newman, M. G., Llera, S. J., Erickson, T. M., Przeworski, A., & Castonguay, L. G. (2013). Worry and generalized anxiety disorder: A review and theoretical synthesis of evidence on nature, etiology, mechanisms, and treatment. Annual Review of Clinical Psychology, 9, 275–297. Nielsen, F.Å., Balslev, D., & Hansen, L. K. (2005). Mining the posterior cingulate: Segregation between memory and pain components. NeuroImage, 27(3), 520–532. Norris, C. J., Larsen, J. T., Crawford, L. E., & Cacioppo, J. T. (2011). Better (or worse) for some than others: Individual differences in the positivity offset and negativity bias. Journal of Research in Personality, 45(1), 100–111. O’doherty, J., Dayan, P., Schultz, J., Deichmann, R., Friston, K., & Dolan, R. J. (2004). Dissociable roles of ventral and dorsal striatum in instrumental conditioning. Science, 304(5669), 452–454. Optimism. ((n.d.) para. 3). Oxford English dictionary. Retrieved from http: //www.oed.com/view/ Entry/132073?redirected From=optimism#eid Perloff, L. S., & Fetzer, B. K. (1986). Self–other judgments and perceived vulnerability to victimization. Journal of Personality and Social Psychology, 50(3), 502. Peters, M. L., Vieler, J. S., & Lautenbacher, S. (2016). Dispositional and induced optimism lead to attentional preference for faces displaying positive emotions: An eye-tracker study. The Journal of Positive Psychology, 11(3), 258–269.

The neurophysiological basis of optimism bias  Chapter | 3  69 Petrides, M., & Pandya, D. (2002). Comparative cytoarchitectonic analysis of the human and the macaque ventrolateral prefrontal cortex and corticocortical connection patterns in the monkey. European Journal of Neuroscience, 16(2), 291–310. Poli, S., Sarlo, M., Bortoletto, M., Buodo, G., & Palomba, D. (2007). Stimulus-preceding negativity and heart rate changes in anticipation of affective pictures. International Journal of Psychophysiology, 65(1), 32–39. Price, P. C. (2000). Wishful thinking in the prediction of competitive outcomes. Thinking & Reasoning, 6(2), 161–172. Price, P. C., Pentecost, H. C., & Voth, R. D. (2002). Perceived event frequency and the optimistic bias: Evidence for a two-process model of personal risk judgments. Journal of Experimental Social Psychology, 38(3), 242–252. Radcliffe, N. M., & Klein, W. M. (2002). Dispositional, unrealistic, and comparative optimism: Differential relations with the knowledge and processing of risk information and beliefs about personal risk. Personality and Social Psychology Bulletin, 28(6), 836–846. Rose, J. P., Endo, Y., Windschitl, P. D., & Suls, J. (2008). Cultural differences in unrealistic optimism and pessimism: The role of egocentrism and direct versus indirect comparison measures. Personality and Social Psychology Bulletin, 34(9), 1236–1248. Rose, J. P., Suls, J., & Windschitl, P. D. (2011). When and why people are comparatively optimistic about future health risks: The role of direct and indirect comparison measures. Psychology, Health & Medicine, 16(4), 475–483. Sabatinelli, D., Bradley, M. M., & Lang, P. J. (2001). Affective startle modulation in anticipation and perception. Psychophysiology, 38(4), 719–722. Scheier, M. F., & Carver, C. S. (1992). Effects of optimism on psychological and physical wellbeing: Theoretical overview and empirical update. Cognitive Therapy and Research, 16(2), 201–228. Scheier, M. F., Carver, C. S., & Bridges, M. W. (1994). Distinguishing optimism from neuroticism (and trait anxiety, self-mastery, and self-esteem): A reevaluation of the life orientation test. Journal of Personality and Social Psychology, 67(6), 1063. Schumacher, S., Herwig, U., Baur, V., Mueller-Pfeiffer, C., Martin-Soelch, C., Rufer, M., et  al. (2015). Psychophysiological responses during the anticipation of emotional pictures. Journal of Psychophysiology. Sege, C. T., Bradley, M. M., & Lang, P. J. (2014). Startle modulation during emotional anticipation and perception. Psychophysiology, 51(10), 977–981. Segerstrom, S. C. (2001). Optimism, goal conflict, and stressor-related immune change. Journal of Behavioral Medicine, 24(5), 441–467. Shah, P., Harris, A. J., Bird, G., Catmur, C., & Hahn, U. (2016). A pessimistic view of optimistic belief updating. Cognitive Psychology, 90, 71–127. Sharot, T. (2011). The optimism bias. Current Biology, 21(23), R941–R945. Sharot, T., Guitart-Masip, M., Korn, C. W., Chowdhury, R., & Dolan, R. J. (2012). How dopamine enhances an optimism bias in humans. Current Biology, 22(16), 1477–1481. Sharot, T., Kanai, R., Marston, D., Korn, C. W., Rees, G., & Dolan, R. J. (2012). Selectively altering belief formation in the human brain. Proceedings of the National Academy of Sciences, 109(42), 17058–17062. Sharot, T., Korn, C. W., & Dolan, R. J. (2011). How unrealistic optimism is maintained in the face of reality. Nature Neuroscience, 14(11), 1475. Sharot, T., Riccardi, A. M., Raio, C. M., & Phelps, E. A. (2007). Neural mechanisms mediating optimism bias. Nature, 450(7166), 102. Shepperd, J. A., Carroll, P., Grace, J., & Terry, M. (2002). Exploring the causes of comparative optimism. Psychologica Belgica, 42(1/2), 65–98.

70  Cognitive biases in health and psychiatric disorders Shepperd, J. A., Klein, W. M., Waters, E. A., & Weinstein, N. D. (2013). Taking stock of unrealistic optimism. Perspectives on Psychological Science, 8(4), 395–411. Shepperd, J. A., Ouellette, J. A., & Fernandez, J. K. (1996). Abandoning unrealistic optimism: Performance estimates and the temporal proximity of self-relevant feedback. Journal of Personality and Social Psychology, 70(4), 844. Shepperd, J. A., Pogge, G., & Howell, J. L. (2017). Assessing the consequences of unrealistic optimism: Challenges and recommendations. Consciousness and Cognition, 50, 69–78. Shepperd, J. A., Waters, E. A., Weinstein, N. D., & Klein, W. M. (2015). A primer on unrealistic optimism. Current Directions in Psychological Science, 24(3), 232–237. Simmons, J. P., & Massey, C. (2012). Is optimism real? Journal of Experimental Psychology: General, 141(4), 630. Smits, T., & Hoorens, V. (2005). How probable is probably? It depends on whom you’re talking about. Journal of Behavioral Decision Making, 18(2), 83–96. Spreng, R. N., Mar, R. A., & Kim, A. S. (2009). The common neural basis of autobiographical memory, prospection, navigation, theory of mind, and the default mode: A quantitative metaanalysis. Journal of Cognitive Neuroscience, 21(3), 489–510. Stiers, D., & Dassonneville, R. (2018). Affect versus cognition: Wishful thinking on election day: An analysis using exit poll data from Belgium. International Journal of Forecasting, 34(2), 199–215. Strunk, D. R., Lopez, H., & DeRubeis, R. J. (2006). Depressive symptoms are associated with unrealistic negative predictions of future life events. Behaviour Research and Therapy, 44(6), 861–882. Sweldens, S., Puntoni, S., Paolacci, G., & Vissers, M. (2014). The bias in the bias: Comparative optimism as a function of event social undesirability. Organizational Behavior and Human Decision Processes, 124(2), 229–244. Tajfel, H., Billig, M. G., Bundy, R. P., & Flament, C. (1971). Social categorization and intergroup behaviour. European Journal of Social Psychology, 1(2), 149–178. Taylor, S. E., & Brown, J. D. (1988). Illusion and well-being: A social psychological perspective on mental health. Psychological Bulletin, 103(2), 193. Taylor, S. E., Kemeny, M. E., Aspinwall, L. G., Schneider, S. G., Rodriguez, R., & Herbert, M. (1992). Optimism, coping, psychological distress, and high-risk sexual behavior among men at risk for acquired immunodeficiency syndrome (AIDS). Journal of Personality and Social Psychology, 63(3), 460. Trimmer, P. C. (2016). Optimistic and realistic perspectives on cognitive biases. Current Opinion in Behavioral Sciences, 12, 37–43. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5(2), 207–232. Weinstein, N. D. (1980). Unrealistic optimism about future life events. Journal of Personality and Social Psychology, 39(5), 806. Weinstein, N. D. (1982). Unrealistic optimism about susceptibility to health problems. Journal of Behavioral Medicine, 5(4), 441–460. Weinstein, N. D., & Lachendro, E. (1982). Egocentrism as a source of unrealistic optimism. Personality and Social Psychology Bulletin, 8(2), 195–200. Weinstein, N. D., Marcus, S. E., & Moser, R. P. (2005). Smokers’ unrealistic optimism about their risk. Tobacco Control, 14(1), 55–59. Windschitl, P. D., & Stuart, J. O. R. (2015). Optimism biases. In The Wiley Blackwell handbook of judgment and decision making (pp. 431–455). Chichester, UK: Wiley-Blackwell. Zetsche, U., Rief, W., & Exner, C. (2015). Individuals with OCD lack unrealistic optimism bias in threat estimation. Behavior Therapy, 46(4), 510–520.

Chapter 4

Negative expectancy biases in psychopathology Peter J. de Jong, Judith K. Daniels Department of Clinical Psychology and Experimental Psychopathology, University of Groningen, Groningen, The Netherlands

Introduction In an everchanging environment, it is crucial to extract regularities based on past events to understand the present and to predict the future. For effective adaptation, it is important to identify predictors and/or causes of pertinent outcomes such as bodily harm, food deprivation, or particular mental states (e.g., happiness, sadness, and fear). Such knowledge about the covariation of critical events allows to anticipate future outcomes, thereby increasing the odds to obtain desired ones (i.e., rewards) and avoid the aversive ones (i.e., punishment). Consistent with the major functional importance of (valid) expectations, the brain has been conceptualized as a “prediction machine” that constantly compares predictions based on mental representations of expected associations ­(so-called priors) with actual outcomes (e.g., Fernández, Pedreira, & Boccia, 2017). The specific predictions coded in the priors are derived from probabilistic models, which rest on previous experiences with the current situation (and are thus highly learning dependent), or on inferred expectations (e.g., triggered by preceding cues). Incoming information is contrasted with the predicted information to enable rapid detection of unexpected events. Sensory input that is not consistent with these prior representations is thought to immediately become very salient. Unexpected events (i.e., prediction errors) are therefore also more likely to be remembered than predictable events, indicating that the human brain is particularly well adapted to react to a rapidly changing world (Proulx, Sleegers, & Tritt, 2017) (Fig. 1). Together, this seems to imply that erroneous expectations will be readily corrected, thereby immunizing people against the development of robust expectancy biases. So how can we then explain that persistent and invalidating negative expectancy biases are very common and represent a core feature of virtually all mental disorders? What is at the core of anxiety, depression, and other Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00004-0 © 2020 Elsevier Inc. All rights reserved.

71

72  Cognitive biases in health and psychiatric disorders

FIG. 1  Illustration of how outcome experiences may correct or reinforce generalized expectations (priors).

mental disorders that prevents proper updating of negative expectations? Which processes are involved in the apparent failure of learning from experience? In the following section, we will first shortly describe the most common methods and measures that are used to investigate expectancy biases. Next, we will focus on the features of negative expectancy biases and how negative expectancies may emerge. Subsequently, we address in more depth why once established generalized negative expectancies are so resistant to change and how neural processes may be involved in this apparent insensitivity to disconfirming information.

Methods used to investigate expectancy biases Studies that are designed to examine the role of expectancy biases in mental disorders typically rely on self-report measures to assess people’s expectancies of particular events or outcomes. To investigate people’s general inclination to expect particular (negative) outcomes, disorder-specific questionnaire measures have been developed. The items of such questionnaires typically take the form of conditional statements that specify the cue and/or context that may give rise to particular outcomes. For example, the Spider Phobia Beliefs Questionnaire (Arntz, Lavy, van den Berg, & van Rijsoort, 1993) asks people to indicate the probability of particular outcomes in situations involving the presence of a spider (e.g., during confrontation with a spider, the spider will bite me; if the spider does not go away and crawls on me, I will get a heart attack). Similarly, the Blushing Phobia Beliefs Questionnaire (Dijk, de Jong, Müller, & Boersma, 2010) asks people to indicate the probability of negative evaluations by others when they would display a blush (e.g., when I blush, others will think I am not socially skillful). As an alternative approach, some studies employed short vignettes describing disorder-relevant scenarios (e.g., an ambiguous interpersonal situation) and ask participants to indicate the probability of particular outcomes (e.g., I will be rejected; others will think I am incompetent) (e.g., Dijk & de Jong, 2012). These questionnaire or vignette-based measures of people’s expectations all rely on abstract or imagined stimuli or situations. As an alternative approach to investigate expectancy biases, lab-based methods have been designed in which people are exposed to concrete stimuli that are followed by concrete outcomes. Stimuli can, for example, represent disorder-relevant or

Negative expectancy biases in psychopathology  Chapter | 4  73

neutral (disorder-irrelevant) pictures or written scenarios, whereas aversive sounds, tactile (heat or electrical) stimuli, or rejecting faces are often used as negative outcomes (e.g., Duits et al., 2016; Hermann, Ofer, & Flor, 2004). During and after such procedure, participants can be asked to indicate their outcome expectancy given the presentation of a particular type of stimulus (e.g., de Jong, Merckelbach, & Arntz, 1995). Importantly, these lab-based procedures allow to experimentally manipulate the type of stimuli that are associated with aversive outcomes, the type and intensity of the aversive outcomes, and the objective stimulus/outcome contingencies. In other words, this type of measurement procedure allows to bring the factual experiences under experimental control, thereby setting the stage for investigating factors involved in the development or persistence of factually unjustified (dysfunctional) expectancies and for examining what factors may help correct/undermine biased expectancies. As an additional asset, these lab-based procedures allow to concurrently assess (neuro)physiological responses, which may help to further unravel the processes involved in the persistence of biased expectancies (e.g., Aue & Okon-Singer, 2015; Wiemer et al., 2015). The following sections will provide a series of concrete examples of how such procedures have been used in studies on negative expectancy biases.

Development of expectancy biases Negative expectancy biases There is ample evidence that unrealistic expectations are a core feature of psychopathology and may contribute to the development and chronicity of mental disorders (for a review, see Rief et  al., 2015). These biased expectations can concern various aspects of future events (Fig. 2). First of all, expectancy biases often concern the probability that a particular negative outcome would

Associative learning experiences

Generalized expectations (prior) -Consequential expectations

Observation of associations

(e.g., harm, rejection)

-Response expectations (e.g., fear, pain)

-Cue expectations Instructed associations

(e.g., probability of encountering cue)

-Self-efficacy expectations (e.g., inability to cope with fear response)

FIG. 2  Schematic illustration of the various pathways that may give rise to the development of various types of generalized expectations.

74  Cognitive biases in health and psychiatric disorders

o­ ccur (e.g., when encountering a dog, it is very likely that the dog will bite me). However, expectancy bias may also concern the intensity/aversiveness of the anticipated negative outcome (e.g., such a dog bite will be extremely painful), and/or the costs that are associated with the occurrence of the anticipated negative outcome (e.g., I may get seriously wounded or may even die). In addition to these expectancies regarding specific consequences (consequential expectations), expectancy bias may concern the probability that a particular source of threat will occur in the first place (e.g., Aue & Hoeppli, 2012). In other words, people might overestimate the probability that one will encounter a threat cue such as a dog or will start blushing (i.e., cue expectations). Importantly, biased outcome expectancies not only are restricted to negative external events (such as a bite of a dog or signs of social rejection by others) but also may involve undesirable internal experiences such as pain and fear/panic (response expectations such as I will experience extreme fear/disgust/pain/despair) (see Peerdeman, van Laarhoven, Peters, & Evers, 2016). Finally, people might have biased expectations regarding their ability to cope with such internal or external threats (i.e., self-efficacy expectations). The exact content and characteristics of biased expectations vary across disorders (Rief et al., 2015). For example, individuals with social anxiety disorder tend to overestimate the probability of displaying a blush (e.g., it is very likely that I will blush when I present my work) and are inclined to overestimate the negative consequences of their blushing (e.g., observers will think I am incompetent) (Dijk & de Jong, 2012). People with chronic pain disorder tend to expect that particular movements will damage their back (Vlaeyen, Crombez, & Linton, 2009); patients with Anorexia Nervosa expect that they will become fat (and therefore rejected) when they stop their excessive dieting (Levinson, Rapp, & Riley, 2014); people with spider phobia overestimate the probability of encountering a spider (Aue & Hoeppli, 2012; de Jong & Muris, 2002) and anticipate extreme fear and losing control when confronted with a spider (Arntz et al., 1993); patients with major depression typically exhibit negative expectations regarding their ability to deal with future stressful events (De Raedt & Hooley, 2016; for more disorder-specific examples, see Rief & Glombiewski, 2016). A striking feature of expectancy biases is their tendency to persist even in the face of contradictory experiences. For example, patients with panic disorder typically hold on to the expectation that the next attack of palpitations will be fatal despite dozens of past panic attacks that turned out harmless. Clearly, such robustness of expectancy biases may help explain why mental disorders often run a chronic course. Yet, it also points to the crucial puzzle of why these expectations are so resistant to correction. To understand and effectively address the refractoriness of mental disorders, it is crucial to know why dysfunctional expectations tend to persist despite the availability of disconfirming evidence. To solve this conundrum, it may be helpful to first consider how negative expectations may generally arise.

Negative expectancy biases in psychopathology  Chapter | 4  75

Pathways to expectancy bias Various pathways may foster individuals’ (negative) expectancies (see Fig. 2). First, people may acquire expectancies via incidental learning experiences. On the basis of concrete experiences, people may learn that particular stimuli or behaviors are associated with particular (undesirable) consequences or outcomes. These associative learning experiences (e.g., being bitten by this particular dog) may give rise to more abstract generalized expectations or schemas (all dogs are dangerous and will bite me). These schemas are proposed to be formed by system consolidation, which is reflected in a gradual process of information reorganization and migration from hippocampus to neocortex (Fernández et al., 2017). Similar generalized expectations may also arise via observational (­social) learning or modeling. For example, seeing one’s mother panicking when confronted with a spider may result in a generalized expectation that spiders are dangerous animals. Third, negative expectancies may also be formed on the basis of instruction. For example, a parent may express concerns with regard to touching particular stimuli: “Don’t pick up that cookie from the ground; it’s dirty and will make you sick!,” which may give rise to generalized contamination concerns. To allow functional adaptation, emerging schemas should update their content and relations in the face of inconsistent (safe) experiences (accommodation). Importantly, the comparator (prediction error detection) function of our brain seems especially sensitive for detecting deviation in the more dangerous (aversive) direction and is relatively insensitive for detecting deviation below the expected level (Arntz, 1997). Accordingly, studies that experimentally manipulated participants’ pain expectations showed that underpredictions resulted in a very persistent heightened pain expectancy in the face of a series of disconfirming experiences (lower pain than expected), whereas such insensitivity to disconfirmation was absent in case of experimentally induced overpredictions of pain. A single instance of higher pain than predicted was already sufficient to correct the apparently overoptimistic (i.e., underpredicted) pain expectations (e.g., Arntz, van Eck, & de Jong, 1992). Thus, the comparator system seems geared to a better safe than sorry heuristic. Unexpected deviations in a more threatening direction may be taken as a red flag signaling that the next confrontation might exceed bearable limits and may be life threatening (Arntz, 1997). Such relatively high impact of underpredictions of threats can be seen as an adaptive mechanism since the implications of underpredictions seem far more critical for survival than overpredictions. Although this asymmetry with regard to the impact of over- vs underpredictions of threat may help avoid negative outcomes, it also implies that unexpected negative experiences have a disproportionally strong influence on the development of people’s generalized expectancies and may thus set the stage for robust but objectively unjustified (i.e., biased) negative expectancies.

76  Cognitive biases in health and psychiatric disorders

As soon as generalized negative expectations have been developed, several mechanisms may come into play that hamper correction of these expectations. In the following, we will subsequently highlight a series of mechanisms that have been put forward to explain the robustness of once acquired negative expectancy bias.

Factors that contribute to the robustness of expectancy bias Avoidance behaviors First of all, negative outcome or response expectations may motivate avoidance and other safety behaviors to ward off the anticipated threatening outcome. Convergently, several studies have shown that the strength of expectancy bias has predictive value for individuals’ avoidance behaviors (e.g., Olatunji, Cisler, Meunier, Connolly, & Lohr, 2008). Avoidance behaviors hamper correction of unjustified generalized expectancies by preventing experiences that can disconfirm the validity of one’s expectations. For example, although excessive hand washing in obsessive compulsive disorder (OCD) may be intended to avoid negative outcomes (e.g., the transmission of disease), it also serves to prevent the correction of inaccurate danger expectancies (e.g., touching a door knob is an important health risk). Accordingly, it has been shown that clinical interventions are less effective if patients are allowed to use this type of safety behaviors during exposure exercises (e.g., Sloan & Telch, 2002). There is also experimental evidence within the context of aversive Pavlovian conditioning that points to avoidance as an important pathway to preserve once acquired negative expectations. For example, it has been shown that acquired shock expectancies (i.e., the CS+ will be followed by a shock) persist if participants are subsequently given the opportunity to push a button upon the presentation of the CS+ as a means to avoid the (expected) shock from being delivered. Using this button during the CS+ only extinction procedure (avoidance) prevented participants to learn that in fact also without pressing the button, the shock was no longer delivered. Thus, avoidance interfered with updating (correction) of earlier acquired negative (shock) expectancies (e.g., Lovibond, Mitchell, Minard, Brady, & Menzies, 2009). Importantly, negative outcome or response expectancies have been found to be strong predictors not only of avoidance but also of anticipatory anxiety (Rachman, 1990). For example, it has been shown that experimentally induced heightened pain expectations resulted in heightened anticipatory anxiety and physiological responding (as indexed by skin conductance and heart rate responses) (e.g., ; Arntz et al., 1992). Such heightened emotional responding in anticipation of a threatening outcome may not only contribute to avoidance motivation but also may be used as confirmatory information for the threat value of the anticipated outcome: If I feel anxious, there must be

Negative expectancy biases in psychopathology  Chapter | 4  77

danger (Arntz, Rauner, & van den Hout, 1995; Verwoerd, de Jong, Wessel, & van Hout, 2013). In other words, the anticipatory responses may be taken as further evidence for the validity of the biased outcome expectations, whereas subsequent avoidance will prevent the occurrence of disconfirming experiences that could have helped to correct the prior. Subsequently, also the avoidance/escape behaviors themselves may be taken as evidence that indeed the occurrence of a negative outcome was prevented. In other words, employing safety behaviors not only may interfere with the modification of already existing expectations but also may promote the development of negative expectancies. Participants who were instructed to engage in OCD safety behaviors for 2 weeks showed an increase in danger estimates and fear of contamination (Deacon & Maack, 2008). One way to explain these findings is that these participants may have concluded that their safety behaviors might have prevented the occurrence of dangerous outcomes. This might have led to a threatening appraisal of initially harmless stimuli as dangerous. Thus, they might have inferred danger on the basis of their avoidance behaviors (if I feel the urge to wash, then there must be danger). In the context of nonexistent threats, such an “if I avoid, there must be danger,” heuristic will logically contribute to the persistence of negative expectancy biases. To test whether indeed patients with anxiety disorders infer danger on the basis of their own acts of avoidance, a group of participants with panic disorder, OCD, or social anxiety disorder and a nonclinical control group were asked to rate the danger in scenarios that systematically varied in the absence/presence of objective danger and the absence/presence of safety behavior (Gangemi, Mancini, & van den Hout, 2012). Interestingly, whereas nonanxious controls only relied on objective danger information, the danger estimates of anxiety patients were also influenced by safety behavior information. This tendency to infer danger on the basis of their avoidance behaviors might contribute to the development and persistence of phobic beliefs. As soon as one believes that one’s safety behaviors imply danger, people may enter a downward spiral in which safety behaviors strengthen the perception of danger and vice versa. All in all, there is consistent evidence that avoidance behaviors not only may contribute to the robustness of biased expectancies by preventing the experience of correcting events but also can contribute by functioning as input that strengthens prior (biased) expectations (see also Fig. 3).

Modulatory effects of expectancies Negative expectancies may have an impact not only on strategically controlled behavior (e.g., avoidance) and reflective processes (e.g., emotional reasoning) but also on automatic, data-driven processes (e.g., saliency/perception threshold). Although people may generally conceive sensation and perception as passive bottom-up processes, as a mere registration of relevant internal or external stimuli, this is clearly not the case. There is ample evidence that our perception

78  Cognitive biases in health and psychiatric disorders

Generalized expectations

Executive monitoring Emotional reasoning

(Prior) Avoidance reasoning

Cue (CS+/CS-)

Avoidance

Sitation-specific expectation

Anticipatory response

Outcome experience

Postevent processing

Anticipated controllability outcome

FIG.  3  Schematic illustration of how the proposed factors that are described in this section may contribute to robustness of (biased) generalized expectations. The concrete cue-based ­situation-specific expectations are moderated by the prior and may elicit avoidance and anticipatory responses. The type of anticipatory responses (e.g., fear: sensory preprocessing) will have an impact on the actual outcome experience. The type of impact will depend on the anticipated controllability of the outcome. Independent of the outcome, anticipatory emotional responses may give rise to emotional reasoning (e.g., if I feel anxious, it must be dangerous), which may confirm the prior. The concrete expectation may also directly motivate avoidance, thereby preventing the outcome to occur; because this precludes expectation-violating experiences, this may consolidate the prior, possibly also indirectly via avoidance-based reasoning bias (if I avoid, it must be dangerous). The concrete outcome experience is monitored (e.g., with regard to cue-outcome contingencies), which will either result in consolidation or modification of the prior. Finally, postevent processing of the actual experience will provide input to consolidate or modify the prior.

is influenced by our expectations, even already during very early processing steps. Moreover, there is increasing evidence indicating that also the process of sensation can be modulated by higher order cognitions such as expectations (e.g., Sussman, Mohanty, & Jin, 2020). The concrete type of modulation depends on the functional meaning or goal relevance of the expected events. If, for example, it has functional value to readily detect the movement of a particular stimulus (e.g., a spider) when it is heading for a particular direction (e.g., toward oneself), expectancy-based sensory processing in the primary visual cortex can be a means to facilitate sensory processing of the moving stimulus (cf. Kok, Brouwer, van Gerven, & de Lange, 2013). Prior expectancies may also have an inhibitory influence on the processing of sensory information. For example, in the context of placebo analgesia research, it has been shown that when participants expect less pain due to an “analgesic” procedure (e.g., application of a creme), they also show less activity

Negative expectancy biases in psychopathology  Chapter | 4  79

in the primary somatosensory brain areas during cued anticipation of the pain stimulus (e.g., Elsenbruch et  al., 2012). This makes sense from the perspective that the placebo-induced lowered pain expectation rendered the cued pain stimulus less important/less motivationally salient. Because the pain stimulus was expected to be reduced in pain intensity, it was no longer critical to invest restricted resources in processing the sensory information that was elicited by the cued pain stimulus. If, on the other hand, participants expect more intense pain due to a particular procedure, such nocebo-induced heightened pain expectancy is associated with heightened activation of the salience network or pain matrix (insula, anterior cingulate cortex, primary, and secondary sensory cortext; Legrain, Iannetti, Plaghki, & Mouraux, 2011). In this case, the modulation of sensory processing is consistent with the goal to monitor whether the increased pain remains within acceptable limits. These differential modulatory effects of heightened versus lowered (pain) expectations help explain how the same sensory input may result in very different subjective experiences. This not only is relevant within the context of placebo/nocebo effects but also may more generally help explain how systematic differences in priors may give rise to differential perception and emotional processing of both the predictive cues and the actual aversive outcomes.

Uncontrollability of expected outcomes At this stage, it seems important to differentiate between the impact of expectations that concern outcomes, which people tend to perceive as controllable, and the impact of expectations of outcomes, which might be perceived as less or not at all controllable (and thus accompanied by low self-efficacy expectations). Most studies that are designed to investigate the moderating influence of expectations on the (emotional) impact of aversive outcomes such as Pavlovian conditioning and placebo studies, rely on very specific and narrowly circumscribed aversive stimuli that are presented for a very short and predictable duration (e.g., 1-s aversive tactile electrical shock, 500-ms burst of 100 dB white noise, 3-s heat stimulation of particular preset intensity, and 1-s presentation of facial expression signaling rejection). Moreover, consistent with ethical constraints, participants are instructed that they can always stop their cooperation, which further strengthens the perceived controllability of the aversive outcome (e.g., if it becomes too intense, I just quit). Under these circumstances, the emotional impact of an aversive outcome is typically lower when it is expected (e.g., due to a signal such as a CS+) than when it is presented unexpectedly. Thus, if cued, participants’ subjective and physiological responding to the aversive outcome (as indexed by fMRI, skin conductance, and heart rate) is typically reduced compared with noncued or invalidly cued aversive outcomes (e.g., Knight, Waters, King, & Bandettini, 2010). This expectancy-elicited reduction in responding is known as unconditioned response (UCR) diminution (Kimmel, 1966). In the

80  Cognitive biases in health and psychiatric disorders

meantime, the response to the predicting cue (CS+) is typically heightened and typically shows a negative relationship with the magnitude of the responses to the outcome (UCR) (Goodman, Harnett, & Knight, 2018). This heightened conditional response (CR) to the predicting cues is thought to reflect the effort associated with the preparatory processes that assist in coping with the aversive outcome. The UCR diminution is taken to reflect the reduced impact due to these coping mechanisms that could be employed because of the predictive quality of the aversive outcome (Goodman et al., 2018). The finding that people’s emotional responding toward cued (and thus expected) aversive outcomes is reduced is not restricted to inherently negative stimuli. Also when spider fearful participants were presented with validly cued pictures of spiders, they responded with less intense emotional responses than when presented with uncued spider stimuli or when spiders were presented when they expected to see neutral stimuli on the screen (Sebastiani, D’Alessandro, & Gemignani, 2014). Thus, when it concerns controllable threats/aversive events, heightened outcome expectations result in lowered emotional responding to the cued aversive events, probably because individuals are then in a position to effectively prepare for the upcoming aversive outcome. The resultant lowered subjective and physiological responding to the aversive outcome may in fact signal a reduction of associated costs: Heightened expectations of its occurrence reduce the aversive properties of the outcome and heighten the perceived self-efficacy to cope with this type of aversive outcomes. Thus, in the context of controllable threats, heightened outcome expectancies tend to result in lowered emotional responding, which in turn may correct inflated expectancies about the associated costs. Although the impact of controllable/narrowly circumscribed aversive outcomes is generally reduced when expected, the opposite seems true for outcomes that involve some uncertainty about the outcome (e.g., with regard to the exact intensity, exact timing, or duration) and/or that are perceived as relatively threatening or uncontrollable (Goodman, Harnett, & Knight, 2018; Goodman, Harnett, Wheelock, et al., 2018). For example, it has been shown that a cued pain stimulus of high intensity resulted in persistently heightened subjective pain and physiological responding to subsequently presented pain stimuli of medium intensity if it was suddenly presented following a series of cued pain stimuli of medium intensity (Arntz et  al., 1992). In addition, the sudden increase in pain intensity resulted in heightened (defensive) heart rate responsivity to subsequently presented cues. Thus, it seems that the induced uncertainty about the exact nature of the outcome resulted in heightened fear (as indexed by defensive responsivity to the cue) together with heightened attention for the pain stimulus, which in turn might have modulated the promulgation of painful somatosensory input throughout the sensory system, thereby enhancing neural gain and pain perception (cf. Almarzouki, Brown, Brown, Leung, & Jones, 2017; Fardo et al., 2017). More generally, there is evidence that cued prestimulus processing of predictable though (highly) feared outcomes may directly

Negative expectancy biases in psychopathology  Chapter | 4  81

modulate sensory processing in lower sensory cortices (Reynolds & Chelazzi, 2004). This enhancement of sensory processing of feared outcomes can be mediated by direct projections from the higher cortex to the sensory cortices or to the sensory thalamus (Wimmer et al., 2015). There is further evidence suggesting that the impact of cued expectations on the regulation of emotional responses to aversive outcomes might be moderated by (un)controllability. Wood et al. (2015) specifically designed a study to determine the independent and joint impact of predictability and controllability on the human brain response. To test the influence of predictability, half of the aversive outcomes were cued. To test the impact of controllability, half of the participants were provided with the opportunity to use a button to terminate the aversive stimulation. When predictable outcomes were uncontrollable, the neural activity of the ventromedial prefrontal cortex (vmPFC) and hippocampus was enhanced. When, on the other hand, predictable outcomes were controllable, neural activity of these areas was reduced. The vmPFC and hippocampus are proposed to play an important role in the emotion regulation process (e.g., Goodman, Harnett, & Knight, 2018; Goodman, Harnett, Wheelock, et al., 2018). The heightened neural activity may thus reflect increased efforts to cope with uncontrollable aversive events. Dysfunction of this neural circuitry may result in a failure to successfully regulate the emotional response to uncontrollable stressors. In turn, this may fuel negative outcome expectancies and promote avoidance of situations/behaviors that are linked to this type of uncontrollable aversive events. Thus, this type of expectancies may, for example, promote avoidance of places where “uncontrollable” dogs may appear or medical interventions that may give rise to uncontrollable aversive side effects. From such perspective, it would be interesting to see in future research whether individual differences in participants’ brain responses to uncontrollable outcomes have prognostic value for people’s more general ability to cope with future stressors. All in all, the available evidence may be taken to indicate that high controllability of cued outcomes may help prevent the development and persistence of negative outcome expectancies (via reduced physiological and emotional responding), whereas low (perceived) control of cued outcomes may promote the development of negative expectancy biases (via heightened physiological and emotional responding).

Uncertainty and illusory correlations In most studies discussed so far, there was a clear contingency between the predicting stimulus (cue/CS+) and the aversive outcome (unconditional stimulus, US). In a typical Pavlovian (differential aversive) conditioning procedure, the CS− is never and the CS+ is always followed by the negative US. However, in real life, the covariation between particular cues and particular outcomes is typically less straightforward/more ambiguous. This can have major implications. First of all, cues that have limited predictive power leave room for uncertainty

82  Cognitive biases in health and psychiatric disorders

about whether or not the outcome will occur and whether or not one can cope with this situation. Accordingly, people generally respond with larger stress/ anticipatory fear responses to cues with lower predictive power (e.g., cues that signal 20% or 60% probability of US occurrence) than to cues that are always (100%) followed by the US (e.g., Hefner & Curtin, 2012); this seems especially the case for high anxious individuals. In turn, these heightened anticipatory physiological fear responses may give rise to emotional reasoning and promote escape and avoidance, thereby contributing to the persistence of heightened expectancy bias. Second, there is ample evidence that the perceived covariation between stimuli/cues and outcomes varies as a function of current situational information and prior expectations (e.g., Alloy & Tabachnik, 1984). Thus, when the current situational information is ambiguous, the assessment of covariations is typically determined by people’s expectations; the stronger the expectations, the more are people inclined to perceive covariations that are consistent with their prior expectations, which in turn reinforces their prior expectations, etc. In line with this, it has been shown that once expectations about covariations between a particular cue and the aversive outcome are induced (stimulus X is more often followed by shock than stimulus Y), these expectations are highly robust against refutation by new information that is factually inconsistent with this earlier acquired expectation (de Jong, Merckelbach, & Arntz, 1990). More specifically, it was shown that when two categories of cue stimuli were followed by an aversive outcome equally often (50%), the differential shock expectancies persisted and tended to become even stronger over trials. Following the experiment, participants generally overestimated the probability of shock given the target slide. This illusory correlation was also expressed in lowered skin conductance responses to the shock outcome on trials where the cue slides preceded the shock (UCR diminution), and relatively strong “omission” responses following the target slide when the cue slide was not followed by a shock outcome. In other words, participants seemed to display an orienting response upon the nonoccurrence of the expected outcome. Importantly, illusory correlations between cues and aversive outcomes have been shown to be most easily elicited when the outcome was relatively aversive (Wiemer, Mühlberger, & Pauli, 2014). Aversive outcomes might elicit a particularly strong motivation to detect covariations between cues and outcomes to help avoid future encounters. Also, aversive outcomes are typically the target of negative expectancies in patients with mental disorders. Thus, it seems reasonable to propose that the perception of illusory correlations may also contribute to the persistence of disorderspecific negative expectancy biases. In line with this, there is consistent evidence that illusory correlations (IC) are involved in various mental disorders including spider phobia (e.g., de Jong, Merckelbach, Arntz, & Nijman, 1992), panic disorder (Pauli, Montoya, & Martz, 1996), social anxiety disorder (e.g., Hermann et al., 2004), posttraumatic stress disorder (Engelhard, de Jong, van den Hout, & van Overveld, 2009),

Negative expectancy biases in psychopathology  Chapter | 4  83

­ lood-injury fears (van Overveld, de Jong, & Peters, 2010), eating disorder b (Mayer, Muris, Kramer-Freher, Stout, & Polak, 2012), and body dissatisfaction (Alleva, Martijn, & Jansen, 2016). In the typical IC study, participants are presented with a series of various types of stimuli (e.g., pictures of spiders, flowers, and mushrooms) that are followed by various outcomes (e.g., shock, tone, and nothing) (e.g., Tomarken, Sutton, & Mineka, 1995). Importantly, each stimulusoutcome combination is presented equally often. In spite of the absence of a systematic relationship between particular stimuli and particular outcomes, participants typically perceive a covariation between disorder-relevant stimuli and aversive outcomes. Disorder-specific ICs are reflected in both online US expectations and postexperimentally reported covariation estimates. Consistent with the view that these ICs may contribute to the persistence of mental disorders, ICs tend to reduce following successful treatment (e.g., van Overveld, de Jong, Huijding, & Peters, 2010), whereas residual ICs following treatment showed prognostic value for a return of fear (de Jong et al., 1995). As further indirect evidence for their causal influence, ICs showed predictive value for the persistence of PTSD symptomatology (Engelhard et al., 2009). In apparent conflict with findings related to ICs between neutral stimuli and aversive outcomes, ICs between disorder-relevant stimuli (e.g., spiders) and aversive outcomes are paralleled with heightened instead of lowered UCRs to the aversive outcomes that are preceded by disorder-relevant “predictive” stimuli. Accordingly, it has been shown that in spite of the heightened shock expectancy on spider cue trials, skin conductance responses to shock outcomes were not reduced (no UCR diminution) but were in fact increased (e.g., de Jong et al., 1995; de Jong & Merckelbach, 1991). One explanation might be that under these circumstances, fearful individuals experience the aversive outcome as relatively uncontrollable. The heightened UCR may then reflect the increased effort to regulate the emotional response to uncontrollable stressors (see also the previous section). As a second explanation, it might be that the increased responding to the aversive outcome is due to affective response matching (vanOyen Witvliet & Vrana, 2000). In contrast to the typical UCR diminution studies, the predictive cues in the IC studies are not intrinsically neutral but elicit distress and defensive arousal also independent of their perceived predictive validity for shock outcome. Thus, the defensive response that is elicited by the disorder-relevant stimuli may in fact prime the defensive response to the aversive outcome stimulus (e.g., a shock or burst of white noise), resulting in a potentiated UCR. Consistent with such view, it has been found that the IC between spider pictures and shock outcome was paralleled with relatively strong activity of the primary sensory motor cortex (PSMc) in response to aversive outcomes that were cued by spider pictures (Wiemer et al., 2015). In individuals with spider phobia, PSMc activity showed a positive correlation with the experienced aversiveness of the shock outcome. Thus, these findings may indicate that the spider cues primed the sensory processing of the subsequently presented shock resulting in

84  Cognitive biases in health and psychiatric disorders

more intense processing and heightened aversiveness of the shock outcome. In its turn, the heightened aversiveness of the shock following spider pictures may contribute to the persistence of heightened shock expectancies on spider trials (cf. Wiemer et al., 2014). Together, the available evidence indicates that also within a more ambiguous context of low or zero contingencies, persistent expectancy biases may arise. This seems especially the case when the US is relatively aversive and when the cue per se is already perceived as threatening; exactly, the type of features that is also typically involved in disorder-relevant expectations. Thus, it seems that also within a relatively ambiguous context, which is closer to the covariations in real life than modeled in a prototypical differential conditioning paradigm, robust expectancy biases may arise.

Postevent “validation” processes (rumination and immunization) Finally, there are a series of processes that immunize against correction of expectancy biases, which take place in the aftermath of actual experiences. For example, people may start ruminating about earlier CS-US co-occurrences and/ or about the aversiveness of the US. Experimental induction of such ruminative processes within the context of aversive conditioning procedures have been shown to result in both heightened fear of the predictive cue (CS+) and relatively persistent US expectancies (e.g., Gazendam & Kindt, 2012; Joos, Vansteenwegen, Vervliet, & Hermans, 2013). It would be interesting to see in future research whether such experimental procedures would also result in stronger connectivity between the dorsolateral prefrontal cortex (dlPFC) and somatosensory cortex activity as a possible underlying neurophysiological process (cf. Greening, Lee, & Mather, 2016). Postevent processes may also immunize against apparently straightforward refutations of prior expectations. For example, people may discard the invalidating evidence as an exceptional, nonrepresentative event and the exception that proofs the rule (Rief et al., 2015) or challenge the validity/relevance of the event for their expectation or put extremely heavy weight on past events to reduce the value of this single current event. This type of processes may thus result in an assimilation of the refuting evidence within the prior (the generalized expectancies), instead of accommodating the prior in response to the new (conflicting) evidence. To aggravate matters, people generally tend to rely on confirmation biased heuristics such as “what I believe is true” (e.g., Vroling, Glashouwer, Lange, Allart, & de Jong, 2016), which further hampers the correction of expectations on the basis of disconfirming evidence. In addition, people are generally inclined to selectively search for additional information that confirms prior expectations together with ignoring information that is at odds with their prior views (e.g., de Jong & Vroling, 2014). Such selective search for expectancyconfirming information will further hamper correction of distorted expectancies

Negative expectancy biases in psychopathology  Chapter | 4  85

and may thus contribute to the persistence of “pathogenic” expectancy biases (de Jong, 2015). Thus, within the context of treatment, it seems important to design strategies that help optimizing the impact of prediction errors and to minimize the opportunity for postevent immunization processes.

The neural basis of expectancy biases In the previous section(s), we already alluded to brain processes that may be involved in the development and persistence of (biased) expectancies. This section zooms in on the neural basis of expectancy biases and aims to provide a global outline of the neural mechanisms that may be involved in the failure to correct generalized expectancies. Consistent with the conceptualization of the brain as a “prediction machine,” which constantly compares predictions based on mental representations of expected associations with actual outcomes (e.g., Fernández et al., 2017), several brain regions have been implicated in these processes—ranging from regions subserving the low-level processing of sensory stimuli to evolutionary younger brain regions involved in executive functions.

Expectations influencing perception In his book, Principles of psychology, James (1890) already suggested that perception is more than a direct registration of sensations and argued that although a part of what we perceive comes through our senses, another part (and it may be the larger part) always comes out of our head. Since then, many studies have shown that perception is far from a passive reflection of available sensory information. For example, recent neuroimaging studies indicate that the recognition of emotions and behavioral response times to targets in visual search paradigms might be influenced by prior expectations, likely due to a strategic deployment of attentional processes triggered by the preceding cues (Aue, Guex, Chauvigné, Okon-Singer, & Vuilleumier, 2019; Barbalat, Bazargani, & Blakemore, 2013; Dzafic, Martin, Hocking, Mowry, & Burianová, 2016; Sussman et al., 2020). Moreover, there is now ample evidence indicating that even the process of sensation can be modulated by higher order cognitions (e.g., expectations). For example, if the brain expects sensory information stemming from the touch of the finger tips, the processing for this type of information is preactivated and thus sensitized (Fiorio & Haggard, 2005). This even works across sensory domains (facilitated by multimodal processing in the posterior parietal cortex) so that the visual presentation of an image of a hand enhances the subsequent sensory acuity for touch of the fingertips (Konen & Haggard, 2014). This enhancement functions not only at the central level but also even all the way down to the sensory organs. While interactions between sensory pathways such as the visual and auditory systems have long been known to occur in the sensory cortices (for a review Choi, Lee, & Lee, 2018), it was recently shown that prior knowledge of the directionality of a subsequent sound leads to adjustment of the eardrums

86  Cognitive biases in health and psychiatric disorders

to optimize sound perception from this direction and that this adjustment is synchronized with the eye movements (Gruters et al., 2018). Together, this type of studies seems to indicate that integration of prior expectations and sensory input is already evident in the earliest stages of bottom-up sensory processing.

Neural correlates of overgeneralization of fear stimuli Overgeneralization of fear stimuli is one of the mechanisms related to the development of biased expectations. The dlPFC has been associated with contingency awareness during fear conditioning (Carter, O’Doherty, Seymour, Koch, & Dolan, 2006), and a recent study indicates that an underactivation of the dlPFC could be associated with overgeneralization of fear cues in anxiety patients (Balderston, Hsiung, Ernst, & Grillon, 2017; cf. Wiemer et  al., 2015). In addition, two recent studies indicated that overgeneralization of fear stimuli might be associated with an underactivation of the medial PFC during safe trials in both trauma exposed individuals (Harnett et al., 2018) and patients with generalized anxiety (Greenberg, Carlson, Cha, Hajcak, & Mujica-Parodi, 2013a), which might indicate that limited downregulation of limbic structures via prefrontal structures might be driving this process. Convergently, overgeneralization of fear has been repeatedly associated with an increased activation of the insula (Greenberg et al., 2013a; Greenberg, Carlson, Cha, Hajcak, & Mujica-Parodi, 2013b; Kaczkurkin et al., 2017; Morey et al., 2015; Tuominen et al., 2019).

Neural correlates of regulatory responses Top-down regulation generally facilitates adaptive coping in stressful situations. However, there are large individual differences in how effective this top-down modulation is, and limitations in stress regulation are thought to play a central role in the development and clinical course of several mental disorders (Compas et al., 2017). When stressors are predictable, the adaptive regulatory response starts right after the cue and entails preparatory processes of up- or downregulation of several neurocircuits before the stressful event occurs. This preparatory response is associated with activations in the right dlPFC, which in turn is causally related to behavioral performance (Vanderhasselt et al., 2007). The elicited dlPFC activation also modulates the generation of autonomic (e.g., skin conductance) responses (see Remue et al., 2016, for an experimental TMS study), likely mediated via the ventromedial PFC (vmPFC), which has direct anatomical connections to the amygdala. As a result, nonreinforced trials that were cued as CS+ trials are associated with an overactivation of the somatosensory cortex, which is functionally connected to the dlPFC at that moment (Greening et al., 2016). The right dlPFC thus plays a vital role both during the preparatory stage and contingency monitoring and the detection of conflict (i.e., prediction error). The preparatory response is known to be strongly moderated by expectations. Under conditions of controllability, cued fear stimuli seem to elicit less

Negative expectancy biases in psychopathology  Chapter | 4  87

preparatory activation, particularly in the vmPFC and the hippocampus (Wood et al., 2015), which may reflect diminished efforts to prepare for the impending aversive event as the event can be avoided. If the anticipated aversive outcome cannot be avoided, the preparatory responses will probably be heightened as a reflection of the increased effort to cope with the upcoming threatening outcome (cf. Wiemer et  al., 2015). However, whether indeed such increased preparatory response activation occurs may depend on people’s self-efficacy expectations. It has been argued that if people have low expectations of successful coping, this could lead to an underdeployment of preparatory activation (De Raedt & Hooley, 2016). More generally, this seems to imply that expectations about one’s options to deal with an upcoming stressful event may already shape the regulatory response before its onset. Illusory correlations have been found to be associated with an overactivation of the left dlPFC directly during the presentation of phobia-specific cues (Wiemer et al., 2015, see Wiemer & Pauli, 2016b for results of a related functional connectivity analysis). In a group of phobia patients, the overactivation of the left dlPFC during the presentation of pictures with spiders predicted the illusory correlation between spider pictures and the likelihood of subsequent shocks. Interestingly, Aue et al. (2015) did not find a significant association between brain activity during the presentation of phobia-relevant cues and biased encounter expectancies. Analyzing brain activation during the subsequent rating phase (i.e., following stimulus processing) indicated that activation in the dlPFC correlated with the encounter expectancies: Phobic patients showed differential underactivation of the right dlPFC, the precuneus, and visual areas that were correlated with the strength of their encounter expectancy bias. In conjunction, it seems clear that expectancy biases are associated with differential under- or overactivation of prefrontal regulatory brain structures. The exact patterns, however, still are somewhat inconsistent and need further exploration.

Somatovisceral responses related to negative expectancy biases Evidence regarding autonomic responses that are associated with biased expectations comes mainly from lab studies using classical conditioning-like procedures. In the prototypical study, an originally neutral stimulus (CS+) is paired with an aversive or painful outcome (e.g., electrical stimulation), whereas another type of stimulus (CS−) is never paired with such an outcome. Most studies assessing autonomic responding in this type of research relied on participants’ sweat response. The sweat response is under sympathetic control and can be easily assessed by putting constant voltage between two electrodes attached to participants’ hand and to measure the current that runs between both electrodes (which will vary as a function of the enhanced conductivity of the hand when people show a sweat response) (for details, see Cacioppo, Tassinary, &

88  Cognitive biases in health and psychiatric disorders

Berntson, 2017). The so-called skin conductance response (SCR) can be best seen as an index of general arousal, and depending on the context and type of stimulation, it may reflect a defensive reflex, an orientation response, cognitive effort, or preparatory efforts to cope with a certain upcoming situation. Clearly, then, it is not a “process-pure” measure and what exactly heightened SCRs reflect is not inherently evident from the response per se. Accordingly, heightened responding to the CS+ may be interpreted not only as reflecting a defensive response but also as reflecting emotion regulation processes in anticipation of the upcoming aversive outcome (e.g., Grings, 1969). Perhaps most relevant for the current context, it has been shown that people also show an SCR when following a series of reinforced CS+ trials, the aversive outcome is incidentally omitted. Thus, at the offset of the CS+, participants then show a so-called omission response, which may be interpreted as an orientation response elicited by the surprise that the aversive outcome did not occur (Seligman, Maier, & Solomon, 1971). Such an omission response may thus be used as an implicit measure of a negative outcome expectancy. As already discussed in a previous section, also the SCR to the outcome itself may be affected by repeated pairings of a CS+ and an aversive outcome. This so-called UCR diminution may reflect a weakened impact of predictable compared with unpredictable aversive outcomes (Goodman, Harnett, & Knight, 2018; Kimmel, 1966; Merckelbach & de Jong, 1988). Thus, also the SCR to the aversive outcome may be used as an implicit measure of negative outcome expectancies (the stronger the expectancy, the weaker the SCR to the outcome) (cf. Grings & Sukoneck, 1971). However, this UCR diminution may only be evident when it concerns a controllable/predictable outcome and when the signaling stimulus per se is affectively neutral. If the signal itself is considered as a threat (e.g., a spider) and the outcome as relatively uncontrollable (e.g., I would not know how to cope when the spider would really touch me), a cued aversive outcome would elicit a potentiated instead of an attenuated SCR (e.g., de Jong et al., 1995; vanOyen Witvliet & Vrana, 2000). Under these circumstances, it has been shown that relatively strong outcome expectations were accompanied by relatively large SCRs to the aversive event (e.g., de Jong et al., 1995). Thus, for using outcome elicited SCRs as an implicit measure of negative outcome expectations, it is crucial to take the type of cue (neutral or threatening) and type of outcome (controllable vs uncontrollable) into consideration. As another way to index autonomic responding within the context of negative outcome expectancies, some studies included cardiovascular parameters with a focus on heart rate acceleration and deceleration. Because heart rate (HR) is under both sympathetic and parasympathetic control, HR acceleration can be due to an increase of sympathetic and/or a reduction of parasympathetic activity. Similarly, HR deceleration can be elicited by increased parasympathetic and/or reduced sympathetic activity. Most relevant to the current context, orientation responses are reflected in a HR deceleration, whereas defensive responses are reflected in a HR acceleration. Thus, if a neutral cue (CS+) becomes

Negative expectancy biases in psychopathology  Chapter | 4  89

predictive of aversive outcomes, this will typically result in a CS+-elicited HR deceleration, whereas the CS−, which is never followed by an aversive outcome, will typically elicit weaker orientation responses (and thus a weaker HR deceleration) than the CS+ (Arntz et al., 1992). However, if participants are uncertain about the intensity of the outcome (and thus with regard to their self-efficacy to cope with such outcomes), the CS+ elicits a defensive attitude as is reflected in a HR acceleration instead of the typical HR deceleration that occurs in response to warning signals (CS+) of outcomes that are of predictable intensity (e.g., Arntz et al., 1992). Thus, also for the cardiovascular measures it depends on the type of outcome (uncontrollable/uncertain intensity vs controllable/fixed intensity) whether expectancy biases may be evident in a relatively strong HR deceleration, or instead in a relatively strong HR acceleration in response to the CS+. Finally, also for the cardiovascular measures there is evidence for UCR diminution. This is reflected in a weakened HR acceleration in response to the aversive outcome if the outcome was expected. However, also for HR responsivity yields that when the intensity of the outcome is uncertain and uncontrollable, cued outcomes elicit a HR acceleration (response potentiation) instead of a HR deceleration (response attenuation) (e.g., Arntz et al., 1992).

Similarities and differences between healthy and clinical populations Negative expectancy biases represent common phenomena that are not restricted to people with mental disorders. For example, research using an illusory correlation paradigm showed that not only high but also low socially anxious individuals displayed persistently heightened expectancies that ambiguous social events (e.g., upon arrival on a party, an unknown person approaches you) or negative social events (e.g., at a party, you hear people gossiping about you) would be followed by social rejection (de Jong, de Graaf-Peters, van Hout, & van Wees, 2009). Similarly, both high and low spider fearful individuals indicated to expect that especially spider slides would be followed by an aversive shock outcome (e.g., de Jong, 1993). In the meantime, it seems that inflated cue expectancies (i.e., the probability that one would encounter a source of threat such as a spider) is restricted to fearful individuals (Aue & Okon-Singer, 2015), although it should be noted that research within this domain is thus far very limited. Healthy and clinical populations are both subject to similar types of processes that counterforce modification of strong beliefs, especially if these beliefs concern the likelihood that particular signals predict harmful outcomes (e.g., de Jong, 2015). Such resistance to correct negative expectancies can be generally considered as an adaptive strategy because in case of life or death, it seems wise to play it safe. Moreover, in many cases, overly negative expectancies do not interfere with daily life. For example, for most people with negatively biased expectancies about what would happen upon encountering a spider, this bias

90  Cognitive biases in health and psychiatric disorders

has no impact on their quality of life because they can easily avoid or escape situations where they may encounter a spider. Under such circumstances, there is little to gain by testing whether indeed one’s negative expectancies are factual correct. Obviously, the situation is fundamentally different for people with mental disorders. By definition, their outcome expectancies strongly interfere with their daily activities and have a highly negative impact on their quality of life. In these cases, the common insensitivity to threat-disconfirming information may render patients’ immune for information that could help correct their invalidating expectancy biases. On the basis of the currently available evidence, it remains unclear whether the differences between clinical and nonclinical populations are due to differences in the strength of the expectancies (e.g., due to differences in learning history) or to fundamental differences in neurological or cognitive functioning that render patients especially prone to threat-confirming information processing. Irrespective of the ultimate source of the differences in expectancy biases between clinical and nonclinical populations, it is clear that patients with mental disorders fail to correct disorder-relevant expectancy biases in spite of their detrimental influence on their quality of life. Thus, from a clinical perspective, it is critical to design interventions that promote an “advocate of the devil” perspective and that help counterforce the various processes that could otherwise contribute to the robustness of patients’ dysfunctional and factually unjustified negative expectancies.

Recapitulation, limitations, and some future directions Negative expectancy biases can be conceived as inadvertent by-products of otherwise functional processes that help prevent the occurrence of negative (harmful) outcomes. Negative expectancy biases represent common phenomena, and are at the core of many mental disorders. To correct dysfunctional expectancy biases within the context of mental disorders, common treatment procedures such as exposure in  vivo and behavioral experiments can be conceptualized as opportunities to elicit salient prediction errors. Starting point of such an approach is thus to elicit expectation-violating experiences, which in turn would lead to adjustment of the generalized expectations. Yet, disorder-specific expectancies appear highly robust against correction and even tend to be immune against straightforward refutations. Thus, to improve treatment efficacy, it would be important to identify factors that critically contribute to the robustness of patients’ expectancies (cf. Kang, Vervliet, Engelhard, van Dis, & Hagenaars, 2018; Kube, Rief, Gollwitzer, Gärtner, & Glombiewski, 2018). This chapter discussed a series of candidate mechanisms that may help explain why once acquired generalized negative expectations are so robust against refutation. These factors included avoidance of possible expectation-violating situations, unjustified inferences on the basis of anticipatory distress and avoidance behaviors (if I avoid something, it must be dangerous), expectancy-based

Negative expectancy biases in psychopathology  Chapter | 4  91

modulation of sensory input and perceptual processing, the perception of ­expectancy-congruent illusory correlations, and postevent “immunizing” processes such as rumination. The reviewed literature not only provided helpful starting points for explaining the refractoriness of expectancy biases but also pointed to some important gaps in the available evidence that await further scrutiny. In the following, we outline just a couple of leads that seem worth of further investigation. Importantly, patients’ expectations are not only robust but also overgeneral and sensitive for illusory correlations. In the previous section, several neural circuits have been proposed to be involved in the persistence of negative expectancy biases. Most prominently, it was discussed that regulatory regions in the dorsolateral PFC show altered activation depending on the cued expectations as well as disorder-specific aberrations. Since the dlPFC has been shown to be involved in contingency awareness, it is tempting to assume that altered prefrontal control could also counterforce the development of objectively unjustified cueoutcome associations. In addition, it was discussed that cued expectations of aversive outcomes resulted in anticipatory somatosensory cortex activity that was associated with heightened aversiveness of the actual outcome (e.g., Wiemer et  al., 2015). Interestingly, the strength of cued sensory preprocessing showed a positive correlation with trait anxiety (Greening et  al., 2016). Individuals with high trait anxiety may thus experience cued aversive events as more aversive than people low in trait anxiety. This may help explain why trait anxiety is associated with a heightened chance of developing biased outcome expectancies and anxiety disorders (e.g., Chan & Lovibond, 1996). It should be emphasized, however, that most of the cited neuroimaging studies are purely correlational and crosssectional and thus do not allow causal inferences. Therefore, an important and exciting next step for future research would be to bring those structures showing diverging activation patterns under experimental control, for example, via transcranial magnetic stimulation (TMS). In addition, it would be interesting and clinically relevant to test whether the strength of cue-based processing has prognostic value for treatment success and long-term prognosis (cf. Duits et al., 2016). Finally, from the perspective that generalized expectancies would exert topdown influence on all levels of the system, it would also be relevant to see how exactly expectancy biases relate to other types of information processing biases that are often found in mental disorders such as attentional bias, interpretation bias, and memory bias. For understanding how to best modify robust expectancy bias, it would be especially relevant to know if these other types of biases represent relatively independent phenomena or can be better understood as functionally related biases that are in fact reflections of the modulating influence of biased expectancies. As a first step, it would be helpful to examine how the various biases relate and covary with regard to particular concerns (see, e.g., Aue & Okon-Singer, 2015).

92  Cognitive biases in health and psychiatric disorders

As one example of such approach, a recent study was designed to test how expectancy bias and memory bias are related within the context of social anxiety (Caouette et  al., 2015). Participants underwent a two-visit task that measured expectations about (visit 1) and memory of (visit 2) social feedback from unknown peers. In this study, it was found that the relationship between social anxiety and biased memory about the (negative) feedback was mediated by negative expectancy bias. Thus, these findings are consistent with the view that memory bias might be better conceptualized as a consequence of expectancy bias than as an independent agent. Further work along these lines including also attentional bias and interpretation bias and targeting also other types of concerns would be extremely welcome. To more directly test to what extent heightened expectancies can indeed causally influence the strength of memory bias, it would be critical to experimentally vary the strength of participants’ expectancies (e.g., by varying the objective contingencies) and to examine how this affects participants’ memory bias. Recent work of Aue and colleagues exactly used such type of approach to test the impact of cued expectancies on attentional bias within the context of spider phobia (e.g., Aue, Chauvigné, Bristle, Okon-Singer, & Guex, 2016; Aue et al., 2019). Another approach could be to design a cognitive bias modification procedure that can specifically target expectancy bias and to examine the impact of such expectancy bias modification on other biases such as attentional, interpretation, and memory biases (cf. Everaert, Duyck, & Koster, 2014). Clearly, lot of future work is still needed to fully solve the critical puzzle of how exactly the various biases are interrelated and to what extent these relationships may vary across disorders. To conclude, negative expectancy biases are assumed to play a critical role in the chronicity of mental disorders. Understanding what factors contribute to the robustness of expectancy biases is thus key for improving currently available treatment procedures. The available literature not only provided some important clues about relevant candidate mechanisms but also pointed to important questions that require further thought and further research. We hope that this chapter is helpful in inspiring future research to arrive at more final answers about the mechanisms involved in the persistence of negative expectancy bias and about the procedures that are most effective in adjusting invalidating negative expectancy biases in patients with mental disorders.

References Alleva, J. M., Martijn, C., & Jansen, A. (2016). Covariation bias in women with a negative body evaluation: How is it expressed and can it be diminished? Journal of Behavior Therapy and Expermental Psychiatry, 50, 33–39. Alloy, L. B., & Tabachnik, N. (1984). Assessment of covariation by humans and animals: The joint influence of prior expectations and current situational information. Psychological Review, 91, 112–149. Almarzouki, A. F., Brown, C. A., Brown, R. J., Leung, M. H. K., & Jones, A. K. P. (2017). Negative expectations interfere with the analgesic effect of safety cues on pain perception by priming the cortical representation of pain in the midcingulate cortex. PLoS One, 12, e0180006.

Negative expectancy biases in psychopathology  Chapter | 4  93 Arntz, A. (1997). The match-mismatch model of phobia acquisition. In G. C. L. Davey (Ed.), Phobias: A handbook of theory, research and treatment (pp. 375–395): Wiley. Arntz, A., Lavy, E., van den Berg, G., & van Rijsoort, S. (1993). Negative beliefs of spider phobics: A psychometric evaluation of the Spider Phobia Beliefs Questionnaire. Advances in Behavior Research and Therapy, 15, 257–277. Arntz, A., Rauner, M., & van den Hout, M. (1995). “If I feel anxious, there must be danger”: Ex-consequentia reasoning in inferring danger in anxiety disorders. Behaviour Research and Therapy, 33, 917–925. Arntz, A., van Eck, M., & de Jong, P. J. (1992). Unpredictable sudden increase of pain intensity and acquired fear. Journal of Psychophysiology, 6, 54–64. Aue, T., Chauvigné, L. A. S., Bristle, M., Okon-Singer, H., & Guex, R. (2016). Expectancy influences on attention to threat are only weak and transient: Behavioral and physiological evidence. Biological Psychology, 121, 173–186. Aue, T., Guex, R., Chauvigné, L. A. S., Okon-Singer, H., & Vuilleumier, P. (2019). Expectancies influence attention to neutral but not necessarily to threatening stimuli: An fMRI study. Emotion, 19, 1244–1258. Aue, T., & Hoeppli, M. E. (2012). Evidence for an encounter expectancy bias in fear of spiders. Cognition and Emotion, 26, 727–736. Aue, T., Hoeppli, M. E., Piguet, C., Hofstetter, C., Rieger, S. W., & Vuilleumier, P. (2015). Brain systems underlying encounter expectancy bias in spider phobia. Cognitive and Affective Behavioral Neuroscience, 15, 335–348. Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. Balderston, N. L., Hsiung, A., Ernst, M., & Grillon, C. (2017). Effect of threat on right dlPFC activity during behavioral pattern separation. Journal of Neuroscience, 37, 9160–9171. Barbalat, G., Bazargani, N., & Blakemore, S. J. (2013). The influence of prior expectations on emotional face perception in adolescence. Cerebral Cortex, 23, 1542–1551. Cacioppo, J. T., Tassinary, L. G., & Berntson, G. G. (2017). Handbook of psychophysiology. New York, NY: Cambridge University Press. Caouette, J. D., Ruiz, S. K., Lee, C. C., Anbari, Z., Schriber, R. A., & Guyer, A. E. (2015). Expectancy bias mediates the link between social anxiety and memory bias for social evaluation. Cognition and Emotion, 29, 945–953. Carter, R. M., O’Doherty, J. P., Seymour, B., Koch, C., & Dolan, R. J. (2006). Contingency awareness in human aversive conditioning involves the middle frontal gyrus. Neuroimage, 29, 1007–1012. Chan, C. K., & Lovibond, P. F. (1996). Expectancy bias and trait anxiety. Journal of Abnormal Psychology, 105, 637–647. Choi, I., Lee, J.-Y., & Lee, S.-H. (2018). Bottom-up and top-down modulation of multisensory integration. Current Opinion in Neurobiology, 52, 115–122. Compas, B. E., Jaser, S. S., Bettis, A. H., Watson, K. H., Gruhn, M. A., Dunbar, J. P., et  al. (2017). Coping, emotion regulation, and psychopathology in childhood and adolescence: A ­meta-analysis and narrative review. Psychological Bulletin, 143, 939–991. de Jong, P. J. (1993). Covariation bias in phobia: Mere resistance to preexperimental expectancies? Behavior Therapy, 24, 447–454. de Jong, P. J. (2015). Danger-confirming reasoning and the persistence of phobic beliefs. In N. Galbraith (Ed.), Aberrant beliefs and reasoning (pp. 132–153). London: Psychology Press. de Jong, P. J., de Graaf-Peters, V., van Hout, W.J.P.J., & van Wees, R. (2009). Covariation bias for social events and signs of rejection in high and low socially anxious individuals. Journal of Behavior Therapy and Experimental Psychiatry, 40, 359–373.

94  Cognitive biases in health and psychiatric disorders de Jong, P. J., & Merckelbach, H. (1991). Covariation bias and electrodermal responding before and after behavioural treatment. Behaviour Research and Therapy, 29, 307–314. de Jong, P. J., Merckelbach, H., & Arntz, A. (1990). Illusory correlation, on-line probability estimates, and electrodermal responding in a (quasi)-conditioning paradigm. Biological Psychology, 31, 201–212. de Jong, P. J., Merckelbach, H., & Arntz, A. (1995). Covariation bias in phobic women: The relationship between a priori expectancy, on-line expectancy, autonomic responding, and a posteriori contingency judgment. Journal of Abnormal Psychology, 104, 55–62. de Jong, P. J., Merckelbach, H., Arntz, A., & Nijman, H. (1992). Covaria­tion detection in treated and untreated spider phobics. Journal of Abnor­mal Psychology, 101, 724–727. de Jong, P. J., & Muris, P. (2002). Spider phobia interaction of disgust and perceived likelihood of involuntary physical contact. Journal of Anxiety Disorders, 16, 51–65. de Jong, P. J., & Vroling, M. S. (2014). Better safe than sorry: Threat-confirming reasoning bias in anxiety disorders. In I. Blanchette (Ed.), Emotion and reasoning (pp. 22–43). London: Psychology Press. De Raedt, R., & Hooley, J. M. (2016). The role of expectancy and proactive control in stress regulation: A neurocognitive framework for regulation expectation. Clinical Psychology Review, 45, 45–55. Deacon, B. J., & Maack, D. J. (2008). The effects of safety behaviors on the fear of contamination: An experimental investigation. Behaviour Research and Therapy, 46, 537–547. Dijk, C., & de Jong, P. J. (2012). Blushing-fearful individuals overestimate the costs and probability of their blushing. Behaviour Research and Therapy, 50, 158–162. Dijk, C., de Jong, P. J., Müller, E., & Boersma, W. (2010). Blushing-fearful individuals’ judgmental biases and conditional cognitions: An internet inquiry. Journal of Psychopathology and Behavioral Assessment, 32, 264–270. Duits, P., Klein Hofmeijer-Sevink, M., Engelhard, I. M., Baas, J. M. P., Ehrismann, W. A. M., & Cath, D. C. (2016). Threat expectancy bias and treatment outcome in patients with panic disorder and agoraphobia. Journal of Behavior Therapy and Experimental Psychiatry, 52, 99–104. Dzafic, I., Martin, A. K., Hocking, J., Mowry, B., & Burianová, H. (2016). Dynamic emotion perception and prior expectancy. Neuropsychologia, 86, 131–140. Elsenbruch, S., Kotsis, V., Benson, S., Rosenberger, C., Reidick, D., Schedlowski, M., et al. (2012). Neural mechanisms mediating the effects of expectation in visceral placebo analgesia: An fMRI study in healthy placebo responders and nonresponders. Pain, 153, 382–390. Engelhard, I. M., de Jong, P. J., van den Hout, M. A., & van Overveld, M. (2009). Expectancy bias and the persistence of posttraumatic stress. Behaviour Research and Therapy, 47, 887–892. Everaert, J., Duyck, W., & Koster, E. H. W. (2014). Attention, interpretation, and memory biases in subclinical depression: A proof-of-principle test of the combined cognitive biases hypothesis. Emotion, 14, 331–340. Fardo, F., Auksztulewicz, R., Allen, M., Dietz, M. J., Roepstorff, A., & Friston, K. J. (2017). Expectation violation and attention to pain jointly modulate neural gain in somatosensory cortex. Neuroimage, 153, 109–121. Fernández, R. S., Pedreira, M. E., & Boccia, M. M. (2017). Does reconsolidation occur in natural settings? Memory reconsolidation and anxiety disorders. Clinical Psychology Review, 57, 45–58. Fiorio, M., & Haggard, P. (2005). Viewing the body prepares the brain for touch: Effects of TMS over somatosensory cortex. European Journal of Neuroscience, 22, 773–777. Gangemi, A., Mancini, F., & van den Hout, M. (2012). Behavior as information: “If I avoid, then there must be a danger”. Journal of Behavior Therapy and Experimental Psychiatry, 43, 1032– 1038.

Negative expectancy biases in psychopathology  Chapter | 4  95 Gazendam, F. J., & Kindt, M. (2012). Worrying affects associative fear learning: A startle fear conditioning study. PLoS One, 7, e34882. Goodman, A. M., Harnett, N. G., & Knight, D. C. (2018). Pavlovian conditioned diminution of the neurobehavioral response to threat. Neuroscience & Biobehavioral Reviews, 84, 218–224. Goodman, A. M., Harnett, N. G., Wheelock, M. D., Hurst, D. R., Orem, T. R., Gossett, E. W., et al. (2018). Anticipatory prefrontal cortex activity underlies stress-induced changes in Pavlovian fear conditioning. Neuroimage, 174, 237–247. Greenberg, T., Carlson, J. M., Cha, J., Hajcak, G., & Mujica-Parodi, L. R. (2013a). Ventromedial prefrontal cortex reactivity is altered in generalized anxiety disorder during fear generalization. Depression & Anxiety, 30, 242–250. Greenberg, T., Carlson, J. M., Cha, J., Hajcak, G., & Mujica-Parodi, L. R. (2013b). Neural reactivity tracks fear generalization gradients. Biological Psychology, 92, 2–8. Greening, S. G., Lee, T. H., & Mather, M. (2016). Individual differences in anticipatory somatosensory cortex activity for shock is positively related with trait anxiety and multisensory integration. Brain Sciences, 6, 2. Grings, W. W. (1969). Anticipatory and preparatory electrodermal behavior in paired stimulation situations. Psychophysiology, 5, 597–611. Grings, W. W., & Sukoneck, H. I. (1971). Prediction probability as a determiner of anticipatory and preparatory electrodermal behavior. Journal of Experimental Psychology, 91, 310–317. Gruters, K. G., Murphy, D. L. K., Jenson, C. D., Smith, D. W., Shera, C. A., & Groh, J. M. (2018). The eardrums move when the eyes move: A multisensory effect on the mechanics of hearing. Proceedings of the National Academy of Sciences of the United States of America, 115, E1309– E1318. Harnett, N. G., Ference, E. W., Wood, K. H., Wheelock, M. D., Knight, A. J., & Knight, D. C. (2018). Trauma exposure acutely alters neural function during Pavlovian fear conditioning. Cortex, 109, 1–13. Hefner, K. R., & Curtin, J. J. (2012). Alcohol stress response dampening: Selective reduction of anxiety in the face of uncertain threat. Journal of Psychopharmacology, 26, 232–244. Hermann, C., Ofer, J., & Flor, H. (2004). Covariation bias for ambiguous social stimuli in generalized social phobia. Journal of Abnormal Psychology, 113, 646–653. James, W. (1890). Principles of psychology. New York: Holt. Joos, E., Vansteenwegen, D., Vervliet, B., & Hermans, D. (2013). Repeated activation of a CS-UScontingency memory results in sustained conditioned responding. Frontiers in Psychology, 4, 305. Kaczkurkin, A. N., Burton, P. C., Chazin, S. M., Manbeck, A. B., Espensen-Sturges, T., Cooper, S. E., et al. (2017). Neural substrates of overgeneralized conditioned fear in PTSD. American Journal of Psychiatry, 174, 125–134. Kang, S., Vervliet, B., Engelhard, I. M., van Dis, E. A. M., & Hagenaars, M. A. (2018). Reduced return of threat expectancy after counterconditioning versus extinction. Behaviour Research and Therapy, 108, 78–84. Kimmel, H. D. (1966). Inhibition of the unconditional response in classical conditioning. Psychological Review, 73, 232–240. Knight, D. C., Waters, N. S., King, M. K., & Bandettini, P. A. (2010). Learning-related diminution of unconditioned SCR and fMRI signal responses. Neuroimage, 49, 843–848. Kok, P., Brouwer, G. J., van Gerven, M. A. J., & de Lange, F. P. (2013). Prior expectations bias sensory representations in visual cortex. Journal of Neuroscience, 33, 16275–16284. Konen, C. S., & Haggard, P. (2014). Multisensory parietal cortex contributes to visual enhancement of touch in humans: A single-pulse TMS study. Cerebral Cortex, 24, 501–507.

96  Cognitive biases in health and psychiatric disorders Kube, T., Rief, W., Gollwitzer, M., Gärtner, T., & Glombiewski, J. A. (2018). Why dysfunctional expectations in depression persist—Results from two experimental studies investigating cognitive immunization. Psychological Medicine, 80, 535–543. Legrain, V., Iannetti, G. D., Plaghki, L., & Mouraux, A. (2011). The pain matrix reloaded: A salience detection system for the body. Progress in Neurobiology, 93, 111–124. Levinson, C. A., Rapp, J., & Riley, E. N. (2014). Addressing the fear of fat: Extending imaginal exposure therapy for anxiety disorders to anorexia nervosa. Eating and Weight Disorders, 19, 521–552. Lovibond, P. F., Mitchell, C. J., Minard, E., Brady, A., & Menzies, R. G. (2009). Safety behaviours preserve threat beliefs: Protection from extinction of human fear conditioning by an avoidance response. Behaviour Research and Therapy, 47, 716–720. Mayer, B., Muris, P., Kramer-Freher, N., Stout, J., & Polak, M. (2012). Covariation bias for foodrelated control is associated with eating disorders symptoms in normal adolescents. Journal of Behavior Therapy and Experimental Psychiatry, 43, 1008–1013. Merckelbach, H., & de Jong, P. J. (1988). Conditioned inhibition of psychophysiological responses to painful stimuli: Facts and suggestions. Psychologica Belgica, 28, 139–155. Morey, R. A., Dunsmoor, J. E., Haswell, C. C., Brown, V. M., Vora, A., Weiner, J., et al. (2015). Fear learning circuitry is biased toward generalization of fear associations in posttraumatic stress disorder. Translational Psychiatry, 5, e700. Olatunji, B. O., Cisler, J. M., Meunier, S., Connolly, K., & Lohr, J. M. (2008). Expectancy bias for fear and disgust and behavioral avoidance in spider fearful individuals. Cognitive Therapy and Research, 32, 460–469. Pauli, P., Montoya, P., & Martz, G. E. (1996). Covariation bias in panic-prone individuals. Journal of Abnormal Psychology, 105, 658–662. Peerdeman, K. J., van Laarhoven, A. I. M., Peters, M. L., & Evers, A. W. M. (2016). An integrative review of the influence of expectancies on pain. Frontiers of Psychology, 7, 1270. Proulx, T., Sleegers, W., & Tritt, S. M. (2017). The expectancy bias: Expectancy-violating faces evoke earlier pupillary dilation than neutral or negative faces. Journal of Experimental Social Psychology, 70, 69–79. Rachman, S. (1990). Fear and courage (2nd ed). New York: Freeman. Remue, J., Vanderhasselt, M. A., Baeken, C., Rossi, V., Tullo, J., & De Raedt, R. (2016). The effect of a single HF-rTMS session over the left DLPFC on the physiological stress response as measured by heart rate variability. Neuropsychology, 30, 756–766. Reynolds, J. H., & Chelazzi, L. (2004). Attentional modulation of visual processing. Annual Review of Neuroscience, 27, 611–647. Rief, W., & Glombiewski, J. A. (2016). Expectation-focused psychological interventions (EFPI). Verhaltenstherapie, 26, 47–54. Rief, W., Glombiewski, J. A., Gollwitzer, M., Schubo, A., Schwarting, R., & Thorwart, A. (2015). Expectancies as core features of mental disorders. Current Opinion in Psychiatry, 28, 378–385. Sebastiani, L., D’Alessandro, L., & Gemignani, A. (2014). Does fear expectancy prime fear? An autonomic study in spider phobics. International Journal of Psychophysiology, 91, 178–185. Seligman, M. E. P., Maier, S. F., & Solomon, R. L. (1971). Unpredictable and uncontrollable aversive events). In F. R. Brush (Ed.), Aversive conditioning and learning (pp. 347–400). London: Academic Press. Sloan, T., & Telch, M. J. (2002). The effects of safety-seeking behavior and guided threat reappraisal on fear reduction during exposure: An experimental investigation. Behaviour Research and Therapy, 40, 235–251.

Negative expectancy biases in psychopathology  Chapter | 4  97 Sussman, T. J., Jin, J., & Mohanty, A. (2020). The impact of top-down factors on threat perception biases in health and anxiety. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 215–241). San Diego: Elsevier. Tomarken, A. J., Sutton, S. K., & Mineka, S. (1995). Fear-relevant illusory correlations: What types of associations promote judgmental bias? Journal of Abnormal Psychology, 104, 312–326. Tuominen, L., Boeke, E., DeCross, S., Wolthusen, R. P., Nasr, S., Milad, M., et al. (2019). The relationship of perceptual discrimination to neural mechanisms of fear generalization. Neuroimage, 188, 445–455. van Overveld, M., de Jong, P. J., Huijding, J., & Peters, M. L. (2010). Contamination and harm relevant UCS-expectancy bias in spider phobic individuals: Influence of treatment. Clinical Psychology and Psychotherapy, 17, 510–518. van Overveld, M., de Jong, P. J., & Peters, M. (2010). Disgust and harm related UCS expectancy bias in blood fearful individuals. Clinical Psychology and Psychotherapy, 17, 100–109. Vanderhasselt, M. A., De Raedt, R., Baeken, C., Leyman, L., Clerinx, P., & D’Haenen, H. (2007). The influence of rTMS over the right dorsolateral prefrontal cortex on top-down attentional processes. Brain Research, 1137, 111–116. vanOyen Witvliet, C., & Vrana, S. R. (2000). Emotional imagery, the visual startle, and covariation bias: An affective matching account. Biological Psychology, 52, 187–204. Verwoerd, J., de Jong, P. J., Wessel, I., & van Hout, W.J.P.J. (2013). “If I feel disgusted, I must be getting ill”: Emotional reasoning in the context of contamination fear. Behaviour Research and Therapy, 51, 122–127. Vlaeyen, J. W., Crombez, G., & Linton, S. J. (2009). The fear-avoidance model of pain. Pain, 157, 1588–1589. Vroling, M. E., Glashouwer, K. A., Lange, W.-G., Allart, E., & de Jong, P. J. (2016). “What I believe is true”: Belief-confirming reasoning bias in social anxiety disorder. Journal of Behavior Therapy and Experimental Psychiatry, 53, 9–16. Wiemer, J., Mühlberger, A., & Pauli, P. (2014). Illusory correlations between neutral and aversive stimuli can be induced by outcome aversiveness. Cognition and Emotion, 28, 193–207. Wiemer, J., & Pauli, P. (2016b). Enhanced functional connectivity between sensorimotor and visual cortex predicts covariation bias in spider phobia. Biological Psychology, 121, 128–137. Wiemer, J., Schulz, S. M., Reicherts, P., Glotzbach-Schoon, E., Andreatta, M., & Pauli, P. (2015). Brain activity associated with illusory correlations in animal phobia. Social Cognitive and Affective Neuroscience, 10, 969–977. Wimmer, R. D., Schmitt, L. I., Davidson, T. J., Nakajima, M., Deisseroth, K., & Halassa, M. M. (2015). Thalamic control of sensory selection in divided attention. Nature, 526, 705–709. Wood, K. H., Wheelock, M. D., Shumen, J. R., Bowen, K. H., Ver Hoef, L. W., & Knight, D. C. (2015). Controllability modulates the neural response to predictable but not unpredictable threat in humans. Neuroimage, 119, 371–381.

Further reading Grupe, D. W., & Nitschke, J. B. (2013). Uncertainty and anticipation in anxiety: An integrated neurobiological and psychological perspective. Nature Reviews Neuroscience, 14, 488–501. Wiemer, J., & Pauli, P. (2016a). How fear-relevant illusory correlations might develop and persist in anxiety disorders: A model of contributing factors. Journal of Anxiety Disorders, 44, 55–62. Wood, K. H., Ver Hoef, L. W., & Knight, D. C. (2012). Neural mechanisms underlying the conditioned diminution of the unconditioned fear response. Neuroimage, 60, 787–799.

Chapter 5

Positive interpretation bias across the psychiatric disorders☆ Ellen Jopling, Jessica Wilson, Matthew Burke, Alison Tracy, Joelle LeMoult University of British Columbia, Vancouver, BC, Canada

Introduction Do you perpetually see your glass as half empty or as half full? An interpretation bias—the tendency to interpret ambiguity in a negative or positive way— has been researched in both healthy and psychiatric populations. Although the majority of this research has focused on interpretation biases toward negative information, researchers have increasingly focused on interpretation biases toward positive information. Researchers assessing interpretation biases have found that healthy controls tend to demonstrate a positive interpretation bias in response to ambiguous information (Alloy & Abramson, 1979; Canli et  al., 2004; Gotlib, Jonides, Buschkuehl, & Joormann, 2011; McKendree-Smith & Scogin, 2000). In contrast, clinical samples—particularly samples of participants with major depressive disorder (MDD), social anxiety disorder (SAD), or generalized anxiety disorder (GAD)—fail to show this bias (Amin, Foa, & Coles, 1998; McKendree-Smith & Scogin, 2000). Consistent with this evidence, researchers have concluded that individuals with psychopathology hold more accurate interpretations than healthy controls, who see the world through “rose-colored glasses” (McKendree-Smith & Scogin, 2000). Importantly, a positive interpretation bias is associated with both higher trait resilience and a decreased risk of depressive symptomatology (Kleim, Thörn, & Ehlert, 2014). Thus, it is likely a powerful buffer against stress and adversity. In fact, researchers have suggested that the lack of a positive interpretation bias may be a more important predictor of symptomatology than the presence of a negative interpretation bias (e.g., Moser, Hajcak, Huppert, Foa, & Simons, 2008). ☆ Preparation of this chapter was facilitated by Michael Smith Foundation for Health Research Scholar Award, SSHRC Insight Development Grant, NSERC Discovery Grant, and CIHR Project Grant to Joelle LeMoult.

Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00005-2 © 2020 Elsevier Inc. All rights reserved.

99

100  Cognitive biases in health and psychiatric disorders

In the current chapter, we focus on findings related to a positive interpretation bias in particular. We summarize the major theories in the field, describe the methods used to investigate the bias, and then describe the neurophysiological (i.e., neural and somatovisceral) correlates of a positive interpretation bias in both nonpsychiatric (i.e., healthy control) and psychiatric populations; finally, we discuss limitations in the existing literature and suggest future directions for the field.

Major theories in the field Theories of child development posit that information processing biases, such as the positive interpretation bias, are rooted in large part in infants’ mental models of their world, called schemas (Piaget, 1952, 1955; Wadsworth, 1996). Schemas are formed based on individuals’ past experiences, interactions, and communications, and they are used both to understand and respond to situations. As such, schemas influence even young children’s interpretations. There are two categories of developmental models that explain how information processing biases form and how they influence risk for psychopathology (for a review see Field & Lester, 2010). Whereas moderation models propose that information processing biases are present in all young children, acquisition models propose that biases emerge later with developmental sophistication. As such, temperament plays a different role in each of the models. According to moderation models, the strength of an information processing bias changes over time depending on individual factors, such as a child’s temperament. Although the inherent positivity bias observed in children (Heyman & Giles, 2004) tends to attenuate with age (Heyman & Legare, 2005), the extent of attenuation is influenced by temperament. In contrast, according to acquisition models, not only can temperament influence the evolution of information processing biases (as is also posited in the moderation model), but also information processing biases can influence the development of anxious or depressive personality traits. Thus, the acquisition model is consistent with cognitive formulations of psychopathology (Beck, 1967; Beck, Rush, Shaw, & Emery, 1979; Holmes, Mathews, Dalgleish, & Mackintosh, 2006; Mathews & MacLeod, 2005), which assert that interpretation biases play a central role in the onset and maintenance of psychiatric disorders. Despite sharing this core tenet, there exist important differences between the cognitive models that explain the unique facets of each psychiatric disorder. Here, we focus on the psychiatric disorders in which a positive interpretation bias has been investigated (i.e., MDD, SAD, and GAD), and we present the seminal cognitive model(s) for each disorder. For a more comprehensive description of cognitive models of psychopathology, we invite the readers to see other excellent reviews (Everaert, Podina, & Koster, 2017b; Mathews & MacLeod, 2005; Williams, Watts, MacLeod, & Mathews, 1997). Within depression, there are several major cognitive theories that provide the foundation for our understanding of interpretation biases. Among the earliest of

Positive interpretation bias across the psychiatric disorders  Chapter | 5  101

these theories was Beck’s cognitive theory of depression (Beck, 1967). Beck asserted that individuals with depression or at risk for depression hold negative schemas that consist of themes of loss and failure and that provoke negative biases in attention and memory. These attention and memory biases cause new information to be processed through a negative lens, resulting in more negative and fewer positive interpretations. In turn, interpretation biases reinforce negative schemas and memories, thereby perpetuating a cycle of negative thinking and worsening psychopathology. More recently, Disner, Beevers, Haigh, and Beck (2011) added to Beck’s original model by proposing the neurobiological correlates of cognition. Here, they advance that cognitive biases in depression are driven by the combination of increased activation of subcortical regions (e.g., the thalamus and amygdala) and attenuated top-down control from regions such as the dorsolateral prefrontal cortex and the anterior cingulate cortex. Ingram (1984) put forth an alternative cognitive model of depression that draws from Bower’s network theory (Bower, 1981). Ingram posited that interpretations of life events activate an interconnected network of nodes, each of which contains specific cognitive sets. Activation of any one node activates adjacent nodes and results in an automatic spreading of activation that strengthens the connection between adjacent nodes. In the context of depression, negative interpretations trigger adjacent nodes that contain negative cognitive sets, thereby priming negative cognitions and increasing vulnerability to depression over time. The dominant cognitive models of social anxiety disorder (SAD; Clark & Wells, 1995; Hirsch & Clark, 2004; Rapee & Heimberg, 1997) posit that individuals with SAD interpret ambiguous information—particularly self-performance in social situations—as negative rather than as positive and then preferentially attend to negative aspects of their self-impression. Moreover, Hirsch and Clark (2004) offer six possible explanations for the fact that individuals with SAD interpret social interactions in a negative vs positive light. Specifically, they assert that biased interpretation of social interactions exist because individuals with SAD exhibit: (1) negative predictions about social interactions (i.e., negative expectancies), (2) negative interpretations of ongoing social interactions, (3) negative memory biases about past social interactions, (4) negative interpretations of their own performance in social interactions, (5) inhibited processing of social cues, and (6) negative interpretation of and attention to social cues. At its core, GAD is typified by worry, which leads to somatic and cognitive symptoms. Cognitive models of GAD (e.g., Wells, 1999) posit that biased interpretation of these somatic and cognitive symptoms (i.e., interpreted as signs of an inability to cope or a loss of control) supports negative beliefs and perpetuates worry. In turn, this perpetuates and exacerbates the somatic and cognitive symptoms and worsens the course of GAD. Building on these models, research has begun to examine the presence and function of positive interpretation biases across the psychiatric disorders. Before reviewing the empirical evidence, however, we first present the most common methods that have been used to investigate positive interpretation biases.

102  Cognitive biases in health and psychiatric disorders

Methods used to investigate a positive interpretation bias Two broad categories of methods have been used in the positive-interpretationbias literature: (1) measures that assess the presence of a positive (or negative) interpretation bias and (2) measures that induce a positive interpretation bias. Researchers interested in positive interpretation biases have relied on a number of self-report questionnaires, information processing tasks, and training paradigms; here, we present the most commonly used or psychometrically sound methods used to investigate a positive interpretation bias. Originally, researchers investigating interpretation biases tended to use selfreport measures, one of the most common of which is the ambiguous scenarios test (AST; Berna, Lang, Goodwin, & Holmes, 2011; Butler & Mathews, 1983; Mathews & Mackintosh, 2000). In the AST, participants are presented with short vignettes (e.g., “You made an appointment with an acquaintance to go to the cinema. Shortly before the appointment, this person leaves a message on your answering machine that the appointment has been canceled.”). Each vignette is followed by a negative interpretation (e.g., “This acquaintance doesn’t like me.”) and a positive or benign interpretation (e.g., “This acquaintance feels sick.”), and participants select the interpretation that best describes the situation. As an alternative to the AST, Wenzlaff and colleagues created the scrambled sentence test (SST; Rude, Wenzlaff, Gibbs, Vane, & Whitney, 2002; Wenzlaff & Bates, 1998). In the SST, participants are presented with six words in a random order (e.g., “interesting life my boring generally is”) and are asked to use five of the words to form a grammatically correct sentence. Importantly, the words can be unscrambled to form either a positive sentence (e.g., “My life is generally interesting.”) or a negative sentence (e.g., “My life is generally boring.”). Positive interpretation biases are calculated based on either the time needed to unscramble positive vs negative sentences or the number of each sentence type successfully unscrambled. Other researchers have used homophones (i.e., words that sound the same but have different meanings, e.g., won and one) or homographs (i.e., words that are spelled the same but have different meanings; e.g., just) to investigate interpretation biases (Byrne & Eysenck, 1993; Hertel, Mathews, Peterson, & Kintner, 2003). In a homophone task, researchers present each homophone auditorily, and participants are asked to write down the spelling of the word. When assessing young children (who may have limited spelling skills), researchers have used a pictorial homophone task, in which participants are asked to choose which of two pictures best captures the meaning of the word (Hadwin, Frost, French, & Richards, 1997). In a homograph test, researchers present each homograph visually, and participants are asked to construct a sentence using it. The valence of the sentences can then be coded by independent judges or rated by participants themselves to index participants’ positive interpretation bias. To overcome the limitations of self-report measures (e.g., response and demand biases; Althubaiti, 2016) and reaction time-based measures (e.g.,

Positive interpretation bias across the psychiatric disorders  Chapter | 5  103

­delayed and variable reaction time in psychiatric samples; Moretti et  al., 1996), a limited number of studies have used an eyeblink test to measure interpretation biases (e.g., Lawson, MacLeod, & Hammond, 2002). This work is based on the premise that the magnitude of individuals’ startle reflex and thus their eyeblink reflex is larger when imagining negatively valenced stimuli compared with neutral stimuli (cf. Bradley, Cuthbert, & Lang, 1999). In the eyeblink test, blink reflexes are elicited while participants are presented with ambiguous stimuli (e.g., acoustically merged valenced vs neutral auditory words that differ in only one phoneme, stress and dress). Based on the magnitude of the eyeblink reflex, the valence of participants’ interpretation of the word can be inferred. Relatedly, when examining the neural or somatovisceral correlates of positive interpretation biases, researchers have presented participants with ambiguous stimuli (e.g., surprised facial expressions or ambiguous sentences; Kim et al., 2004; Kim, Somerville, Johnstone, Alexander, & Whalen, 2003); in this context, measures of biological activity when viewing the stimuli can be coupled with participants’ self-reported interpretation of it. The second category of methods used in the positive-interpretation-bias field leverages cognitive bias modification procedures to induce a positive interpretation bias. Cognitive bias modification for interpretation (CBM-I; Mathews & Mackintosh, 2000) paradigms train a positive (or negative) interpretation bias through a two-phase training approach. In phase one, the training phase, participants view ambiguous scenarios containing a word fragment that disambiguates the scenario in the valence consistent with their training condition (i.e., in a positive, negative, or “no-training” neutral direction). In phase two, the test phase, participants view ambiguous scenarios and then, following a brief filler task, rate the similarity of four sentences (two positive sentences and two negative sentences) to the original scenarios. Training effectiveness is determined based on participants’ similarity ratings of positive vs negative sentences during the test phase. Holmes and colleagues (Blackwell & Holmes, 2010; Holmes et al., 2006; Holmes, Lang, & Shah, 2009; Lang, Moulds, & Holmes, 2009) developed an alternative to Mathews and Mackintosh’s (2000) CBM paradigm that targets interpretation biases through positive imagery. To train a positive interpretation bias, participants hear training scenarios that resolve in a positive direction toward the end of the scenario. For example, “You have started an evening class which is tough going. You are determined to succeed, and after a while, it becomes much easier and more enjoyable.” Although all participants listen to the same scenarios, they are randomly assigned to process them using imagery or a verbal approach. To promote imagery vs verbal processing, after each scenario, participants rate the vividness of the situation (imagery condition) or their understanding of it (verbal condition). Pre- to posttraining changes in interpretation bias are then assessed via an independent measure of interpretation biases (e.g., the SST).

104  Cognitive biases in health and psychiatric disorders

Brain regions involved in the emergence of a positive interpretation biases The current section reviews evidence on the neurological correlates of a positive interpretation bias (see Fig.  1). Because this area of inquiry has been significantly understudied, this section also considers findings on the related topic of an optimism bias (i.e., the extent to which an individual expects positive outcomes). Given that optimism and positive interpretation biases both involve a pattern of thinking in which individuals move toward positive and away from negative conclusions (Sharot, Riccardi, Raio, & Phelps, 2007), research on the optimism bias offers candidate brain regions that might be relevant to future research investigating positive interpretation biases. Here, we highlight a few of these candidate regions and also direct readers to Chapter 3 of Dricu, Kress, and Aue (2020) for an extensive review of patterns of neural activation associated with optimism bias.

Positive interpretation bias specifically To investigate the neural correlates of a positive interpretation bias, researchers have conducted studies examining participants’ fMRI activation patterns when interpreting ambiguous facial expressions and sentences (Kim et  al., 2003, 2004). These studies documented that individuals with a propensity to positively interpret ambiguous (e.g., surprised) stimuli show greater activation of the ventromedial prefrontal cortex (vmPFC), which has been associated with

FIG. 1  Patterns of neural activity and connectivity associated with the positive interpretation bias and the optimism bias. rACC, rostral anterior cingulate cortex; vmPFC, ventromedial prefrontal cortex. Note: Reward system structures include bilateral claustrum, insula, putamen, thalamus, and right caudate and posterior cingulate cortex. Occipital cluster includes left inferior occipital gyrus and fusiform gyrus.

Positive interpretation bias across the psychiatric disorders  Chapter | 5  105

the inhibition of emotional responses (Santos, Seixas, Brandão, & Moutinho, 2011) and positive affective processing (Harrison et al., 2017). In their initial study, Kim et al. (2003) asked participants to passively view presentations of surprised facial expressions during fMRI scans. Results demonstrated that greater activation in the vmPFC was observed among individuals who interpreted the surprised expressions as positive. The role of the vmPFC in positive interpretation biases was further supported in a subsequent study that examined participants’ neural responses to surprised faces that were preceded (i.e., primed) by either a positive or negative sentence (Kim et al., 2004). Consistent with findings described earlier, greater vmPFC activation occurred in response to faces preceded by positive sentences.

Constructs related to a positive interpretation bias Evidence for the involvement of occipital regions (including the inferior occipital and fusiform gyri) comes from work examining the neural correlates of the optimism bias. Aue, Nusbaum, and Cacioppo (2012), for example, documented that the optimism bias is associated with increased functional connectivity between an occipital cluster (including the left inferior occipital gyrus and fusiform gyrus) and key structures of the human reward system (including limbic and dorsal striatal regions). Conceptually, these results indicate that the optimism bias is associated with increased functional connectivity of brain regions responsible for visual processing and attention with reward-related behaviors (Mangun, Buonocore, Girelli, & Jha, 1998; Robbins, Cador, Taylor, & Everitt, 1989; Rossion, Schiltz, & Crommelinck, 2003). Given that optimism is related to a positive interpretation bias (Kress & Aue, 2017; Tran, Hertel, & Joormann, 2011), the role of various occipital areas might be examined in future work on positive interpretation biases. There is also evidence for the involvement of both the rostral anterior cingulate cortex (rACC) and the amygdala in the optimism bias. For instance, Sharot et al. (2007) demonstrated that individuals with more optimism about the future showed enhanced activation of the rACC and the amygdala when imagining positive future events, relative to negative future events. Similarly, Sharot (2011) documented increased functional connectivity between the rACC and the amygdala in individuals who display an optimism bias. These findings suggest that an optimism bias is related to increased connectivity between regions responsible for affective error responding (the rACC; Blair et  al., 2013) and emotional responding (the amygdala; Phelps & LeDoux, 2005).

Summary Although the neurological correlates of a positive interpretation bias have been understudied, preliminary evidence suggests that a number of brain regions are involved in the emergence of this bias, namely, the vmPFC. Specifically, the

106  Cognitive biases in health and psychiatric disorders

literature suggests that increased activation in the vmPFC is associated with an increased likelihood of interpreting ambiguous information in a positive manner (Kim et al., 2003, 2004). A related literature concerned with optimism bias provides evidence for the role of increased functional connectivity between occipital areas and reward system structures among individuals who demonstrate a greater optimism bias (Aue et al., 2012). Finally, work completed by Sharot et al. (2007) and Sharot (2011) suggests that both the rACC and the amygdala are also involved in the emergence of an optimism bias; specifically, individuals who tend to expect positive future events show increased activity in the rACC and the amygdala and stronger functional connectivity between them. Although only a few studies have investigated the neural correlates of a positive interpretation bias, the existing literature has identified a number of candidate mechanisms whose roles ought to be further elucidated in future research.

Somatovisceral (e.g., autonomic) responses involved in the emergence of a positive interpretation bias Psychophysiological measures, also referred to as somatovisceral or autonomic responding, can augment self-report measures of positive interpretation bias and can elucidate the physiological changes associated with positive interpretation biases. In this section, we summarize the small, but growing, body of research that examines the influence of positive interpretation biases on human physiology in both healthy and psychiatric populations. Because this literature is relatively small and because no examination of positive interpretation biases in the absence of training has occurred, we also consider studies that induced a positive interpretation bias via CBM-I or imagery rescripting procedures. In the first study to find somatovisceral effects of positive interpretation biases, Joormann, Waugh, and Gotlib (2015) used CBM-I to investigate the effects of training positive or negative interpretation biases on autonomic responses to stress in participants with depression and healthy controls. Participants in the active training group were taught to complete ambiguous sentences with their assigned positive or negative interpretation condition, whereas the control group received no training. After training, the authors measured participants’ heart rate both in anticipation of and during a psychosocial stressor. Regardless of diagnostic status, participants in the positive interpretation condition had smaller increases in heart rate in anticipation of the stressor compared with participants in the negative interpretation condition. Thus, this evidence suggests that positive interpretation biases mitigate somatovisceral activation in anticipation of stress in both healthy and depressed groups. In contrast to these findings, however, other researchers have found no significant differences in psychophysiology between individuals who completed positive CBM-I training or imagery rescripting and controls who received negative CBM-I or no training at all (Lester, Field, & Muris, 2011;

Positive interpretation bias across the psychiatric disorders  Chapter | 5  107

Tolgou et  al.,  2018). Lester et  al. (Lester et  al., 2011), for example, used CBM-I to train either positive or negative interpretation biases and then tested mean heart rate responses to a behavioral avoidance stressor in a sample of children 6–11 years old. As expected, compared with children in the negative training condition, children in the positive training condition demonstrated more positive interpretations of ambiguous scenarios and more adaptive behavioral responses to the stressor post training, confirming the effects of the CBM-I induction. However, children in the negative and positive training conditions did not differ significantly in their mean heart rate responses to the stressor, suggesting that positive interpretation biases may not be associated with subsequent somatovisceral responses to stress in children. In a separate study, Tolgou et al. (2018) investigated the association between positive imagery rescripting and biological responses to an aversive film clip in a sample of adults with high levels of health-related anxiety. After participants watched the aversive film clip, they were randomly assigned to one of four conditions. In the first condition, positive imagery, participants were instructed to recall a positive event from their past. In the second condition, imagery reexperiencing, participants were asked to reexperience the most stressful scene from the film. In the third condition, imagery rescripting, participants were instructed to reimagine the film with a positive interpretation and a desirable outcome. Finally, in the control condition, no training was provided. Salivary cortisol levels were measured at baseline, after the film, and after the intervention; heart rate was measured continuously. The authors found that the effects of positive imagery rescripting were not detectable by biological measures, as participants in the positive imagery rescripting condition did not differ from the other conditions in either cortisol or heart rate responses to the films. There are a number of possible reasons why Lester et al. (Lester et al., 2011) and Tolgou et al. (Tolgou et al., 2018) did not find significant effects of positive interpretation biases on psychophysiology while others reported significant findings (Joormann et al., 2015). First, whereas Joormann et al. (Joormann et al., 2015) focused on depressed and healthy adults, Lester et al. and Tolgou et al. focused on healthy children and anxious adults, respectively. It is possible that these groups have different patterns of physiological responding or that the positive interpretation training has different effectiveness in these disparate samples. Second, the stress tests and procedures for collecting psychophysiological data in the three studies differed from one another. The behavioral avoidance task and aversive film used by Lester et al. and Tolgou et al., respectively, may not have been strong enough to elicit the changes in psychophysiology that were observed in the psychosocial stress task used by Joormann et al. Third, it is possible that the CBM-I and imagery rescripting procedures used by Lester et al. and Tolgou et al. were not sufficiently strong to modify implicit psychophysiological responding relative to the other training conditions. It is possible that different forms of training or training over multiple sessions may be necessary to elicit positive interpretations detectable by psychophysiology.

108  Cognitive biases in health and psychiatric disorders

Similarities and differences between healthy and clinical populations in the biological basis of positive interpretation biases Substantial research has documented that differences exist in the way healthy and clinical populations interpret ambiguous information (see the review by Hirsch, Meeten, Krahé, & Reeder, 2016). Findings suggest that healthy controls exhibit a positive interpretation bias, whereas clinical populations fail to do so. Compared with the amount of research that provides behavioral data on positive interpretation biases, there is relatively little research examining the biological correlates of positive interpretation biases in healthy vs clinical samples. In fact, to date, researchers have compared the biological correlates of positive interpretation biases in healthy vs clinical samples exclusively in the context of depression, SAD, and comorbid depression SAD.

Depression Within the depression literature, there is growing evidence that individuals with depression are less likely than healthy controls to hold a positive interpretation bias. McKendree-Smith and Scogin (2000), for example, examined positive interpretation biases among participants with varying levels of depressive symptomatology. The authors asked participants to rate the desirability of neutral feedback about their personality and found that nondepressed and mildly depressed participants were more likely to hold a positive interpretation of their personality profile than were severely depressed participants. Similarly, Alloy and Abramson (1979) observed that nondepressed individuals exhibited cognitive biases that facilitated positive interpretations of themselves and the world, whereas depressed persons maintained a realistic—albeit negative—perspective. Despite data documenting behavioral differences between depressed and control participants’ positive interpretation biases (Alloy & Abramson, 1979; Canli et  al., 2004; Everaert et  al., 2017a; Gotlib et  al., 2011), there is a shortage of research examining the biological correlates of positive interpretation bias in depressed vs control samples. In fact, the only study to examine the biological correlates of positive interpretation bias in participants with depression compared with healthy controls was conducted by Joormann et al. (2015). The authors examined autonomic responses to stress in participants following CBM-I training. Interestingly, both depressed and control participants who received positive CBM-I training demonstrated attenuated heart rate responses in anticipation of the stressor compared with depressed and control participants who received negative CBM-I training. This study offers preliminary evidence that, regardless of diagnostic status, positive interpretation biases are associated with attenuated physiological responses to stress.

Positive interpretation bias across the psychiatric disorders  Chapter | 5  109

Social anxiety disorder Moser and colleagues recently investigated the neurophysiological correlates of a positive interpretation bias in social anxiety (Moser et al., 2008). In this study, Moser et al. (2008) examined electroencephalogram recordings while individuals with either low or high symptoms of social anxiety completed a grammatical decision task. Specifically, participants viewed a series of ambiguous sentence stems that were resolved with either a positive or a negative terminal word. For instance, the ambiguous sentence stem, “As you give a speech, you see a person in the crowd smiling, which means that your speech is…,” was ended with either a positive (e.g., “funny”) or negative (e.g., “stupid”) word. In response to sentences with negative endings, low-anxious controls exhibited increased P600 magnitudes. Given that the P600 is a language-relevant eventrelated potential that fires in response to syntactic violations, this finding suggests that ­low-anxious controls found negative sentence endings unexpected, which Moser and colleagues interpreted as indicative of a positive interpretation bias. In contrast, individuals with high levels of social anxiety symptoms exhibited equally large P600 amplitudes to sentences that concluded positively or negatively, suggesting that high-anxiety participants failed to demonstrate a positive interpretation bias. Biological findings in samples of SAD are consistent with behavioral data that use questionnaire and reaction time-based tasks to compare the way socially anxious and healthy individuals interpret ambiguous scenarios. For example, Amin et al. (1998) found that individuals with SAD chose positive interpretations significantly less often than did individuals with obsessive compulsive disorder or than healthy controls. Interestingly, these findings may only hold true for socially relevant scenarios. Huppert, Foa, Furr, Filip, and Mathews (2003), for example, documented that socially anxious individuals showed neither a positive nor negative interpretation bias to nonsocial prompts. Researchers have also attempted to induce a positive interpretation bias in socially anxious individuals through CBM-I paradigms. Murphy, Hirsch, Mathews, Smith, and Clark (2007), for example, demonstrated that individuals scoring high in social anxiety could be trained to hold more positive interpretations of potentially threatening scenarios. To date, however, researchers have not yet examined the biological correlates of inducing a positive interpretation bias in SAD.

Comorbid depression and social anxiety disorder The biological correlates of positive interpretation biases have also been compared in individuals with depression, SAD, comorbid depression SAD, or neither depression nor SAD (Moser, Huppert, Foa, & Simons, 2012). Specifically, Moser et  al. (2012) measured electrophysiological activity while participants completed a grammatical decision task, in which social scenarios were disambiguated with a negative or positive ending. Results indicated that healthy

110  Cognitive biases in health and psychiatric disorders

controls exhibited an increased N400 amplitude in response to sentences with negative endings. Given that N400 activation is related to the semantic processing of unexpected words and events, Moser et al. (Moser et al., 2012) concluded that healthy control participants did not expect negative sentences, suggesting that they hold a positive interpretation bias. Conversely, all psychiatric groups (i.e., depression, SAD, and comorbid depression SAD) demonstrated enhanced N400 amplitude for sentences that ended positively. Thus, Moser et al. (Moser et al., 2012) postulated that clinical participants did not expect positive endings, reflecting a lack of a positive interpretation bias.

Other anxiety disorder To date, there are no studies comparing the biological correlates of positive interpretation biases in healthy controls vs participants with any other anxiety disorder. This gap in the literature is surprising given behavioral evidence of group differences in positive interpretation biases between other anxiety disorders (especially individuals with GAD) and healthy controls (e.g., Eysenck, Mogg, May, Richards, & Mathews, 1991).

Summary and debate Neurophysiological, somatovisceral, and behavioral research has demonstrated that there are differences in the way healthy individuals and psychiatric populations interpret ambiguous information. Researchers have found increased activation in brain regions such as the vmPFC, which is thought to be involved in interpreting ambiguous information and specifically when resolving ambiguous scenarios as positive in healthy populations. Comparatively, researchers have documented decreased activation of the vmPFC in clinical samples (e.g., Blair et al., 2013). Thus, the vmPFC could play a role in maintaining symptoms of depression. This conclusion is strengthened by behavioral research documenting that depressed individuals do not show a positive interpretation bias when presented with ambiguous situations (Vestre & Caulfield, 1986). Findings are also consistent with cognitive theories of depression, which suggest that the way individuals interpret emotion-eliciting situations determines subjective affect. Beck’s model of depression, for example, underscores that individuals who interpret situations in a negatively biased manner are more likely to hold a negatively biased worldview, which can lead to depressed mood (Beck, 1976, 1987). While the vmPFC may account for differences in the way healthy vs depressed individuals interpret information, less is known about biological correlates of the positive interpretation bias in other psychiatric disorders. For example, there is not yet enough research to conclude which brain region(s) play a role in positive interpretation biases in SAD, though electroencephalogram data suggests that participants with SAD and MDD have similar brain activation patterns. Specifically, Moser et al. (2008, 2012) demonstrated similar

Positive interpretation bias across the psychiatric disorders  Chapter | 5  111

electrophysiological activity among individuals with SAD and MDD when they were exposed to ambiguous sentences using electroencephalogram recordings, with psychiatric groups having enhanced activation for positive endings providing evidence for the lack of a positive interpretation bias in these individuals compared with the healthy controls. If brain activation during positive interpretation biases is similar for individuals with SAD and MDD, what accounts for the differences in symptoms across disorders? This question can be approached by looking at how psychiatric individuals interpret ambiguous information. Whereas individuals with depression have been found to interpret ambiguous information as negative compared with healthy controls, socially anxious individuals interpret ambiguous information as threatening. It is of no surprise that researchers have found that individuals with SAD tend to choose positive interpretations significantly less often than healthy controls as social contexts are full of ambiguity, leaving many interpretations of a situation open to the anxious individual (e.g., fearing how a situation will turn out or what others might think of them). This example paints a picture of how the neurophysiological findings can be ingrained in the behavioral research. To determine whether this is a bottom-up or top-down relationship, further research will need to be conducted. By training positive interpretation biases in psychiatric populations and examining changes in neural activity, we might better understand the biological correlates of positive interpretation biases.

Limitations There has been an increasing focus in recent years on understanding the causes, correlates, and consequences of positive interpretation biases. Despite this progress, the field continues to be limited in several ways. The clearest limitation stems from the shortage of research on positive interpretation biases. Historically, the focus has been on how clinical and healthy samples differ in their negative interpretation biases, and it is only in recent years that researchers have examined group differences in positive interpretation biases specifically. With this in mind, research on the biological correlates of positive interpretation biases is still in its infancy. A second limitation is that we continue to rely predominantly on self-report or reaction time-based methods to assess positive interpretation biases. This limits the accuracy and reliability of our findings given that self-report measures are subject to demand and response bias, and reaction time-based measures are limited by slowed and or variable response latencies, particularly in psychiatric samples (Azorin, Benhaïm, Hasbroucq, & Possamaï, 1995; Byrne, 1976). In addition, to date, most of the research examining the relation between interpretation biases and biological functioning has included only one measure of biology (e.g., neural or somatovisceral). Not only will it be important for future studies to include additional measures of biological functioning (e.g., cortisol),

112  Cognitive biases in health and psychiatric disorders

but also it will be critical to include multiple biological measures simultaneously. Doing so will allow for a more complete understanding of the biological bases of positive interpretation biases.

Future directions As we look to the future, we hope that this field will grow in a number of exciting ways. First and foremost, we are enthusiastic about the fact that an increasing number of studies are assessing positive interpretation biases explicitly, and we hope that this trend continues. We find assessments of positive interpretation biases particularly strong when done independently of assessments of negative interpretation biases, for example, by comparing positive vs neutral interpretations rather than positive vs negative interpretations. Some studies have begun to examine the biological correlates of positive interpretation biases in child or adolescent samples (Lester et al., 2011); however, this remains an area of considerable opportunity. Specifically, studying positive interpretation biases in children and adolescents has the potential to help elucidate the process through which the bias develops and the biological foundations that contribute to it. This might be accomplished, for example, by examining children at risk for psychopathology before the onset of clinical symptoms. As we look forward, we also believe that it will be critical to increasingly examine multiple cognitive biases simultaneously. To date, most cognitive biases are still examined in isolation despite the fact that cognitive theories posit that they influence, and are influenced by, one another (Aue & Okon-Singer, 2015; Beck, 1967, 1978; Everaert, Koster, & Derakshan, 2012; Hirsch, Clark, & Mathews, 2006; Kress & Aue, 2017). Taking this one step farther, it will also be important to use experimental designs to identify whether manipulating one cognitive bias will lead to changes in other cognitive biases and in biological function. Doing so will enable researchers to identify those cognitive biases that lie at the core of psychiatric dysfunction and that would be the most promising targets in intervention. Along this line, it will be invaluable to identify the neurobiological markers that underlie positive interpretation biases and that predict the course of psychiatric symptoms. With this information, we may begin to target these specific mechanisms in treatment by pursuing brain-based personalized treatment.

Chapter summary Taken together, increasing evidence documents that positive interpretation biases play a critical role in the pathogenesis of psychiatric risk and resilience (Kleim et  al., 2014). This conclusion is supported by preliminary evidence from somatovisceral data that show positive interpretation biases are associated with diminished physiological response to stress (Joormann et  al., 2015) and that multiple psychiatric disorders—MDD, SAD, and GAD—are typified by

Positive interpretation bias across the psychiatric disorders  Chapter | 5  113

the lack of a positive interpretation bias (LeMoult & Gotlib, 2018). Although results are preliminary and future research is needed, empirical results implicate several key brain regions in the emergence of a positive interpretation bias. Namely, research specifically examining the neurological correlates of the positive interpretation bias has provided evidence for the role of increased activation in the vmPFC. Related research from the field of study concerned with the optimism bias provides evidence for the role of increased functional connectivity between occipital areas and reward system structures and increased activity in both the rACC and amygdala, as well as stronger functional connectivity between them (Aue et al., 2012; Kim et al., 2003, 2004; Sharot, 2011; Sharot et al., 2007). The fact that clinical samples show opposite activation in some of these same brain regions suggests that anomalous neural functioning may contribute to the lack of positive interpretation biases documented in clinical samples. Despite the promise of these preliminary findings, additional research is needed to better understand how positive interpretation biases first emerge, how they are maintained, and how they might be propagated.

References Alloy, L. B., & Abramson, L. Y. (1979). Judgment of contingency in depressed and nondepressed students: Sadder but wiser? Journal of Experimental Psychology: General, 108(4), 441–485. Althubaiti, A. (2016). Information bias in health research: Definition, pitfalls, and adjustment methods. Journal of Multidisciplinary Healthcare, 9, 211–217. https://doi.org/10.2147/JMDH.S104807. Amin, N., Foa, E. B., & Coles, M. E. (1998). Negative interpretation bias in social phobia. Behaviour Research and Therapy, 36(10), 945–957. https://doi.org/10.1016/S0005-7967(98)00060-6. Aue, T., Nusbaum, H. C., & Cacioppo, J. T. (2012). Neural correlates of wishful thinking. Social Cognitive and Affective Neuroscience, https://doi.org/10.1093/scan/nsr081. Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, https://doi.org/10.1016/j.cpr.2015.08.005. Azorin, J. M., Benhaïm, P., Hasbroucq, T., & Possamaï, C. A. (1995). Stimulus preprocessing and response selection in depression: A reaction time study. Acta Psychologica, 89(2), 95–100. https://doi.org/10.1016/0001-6918(94)00024-B. Beck, A. T. (1967). Depression: Clinical, experimental, and theoretical aspects. The Journal of The American Medicial Association, https://doi.org/10.1001/jama.1968.03140130056023. Beck, A. T. (1976). Cognitive therapy and the emotional disorders. International Universities Press. Beck, A. T. (1978). Beck depression inventory. Depression, 2006, 2–4. Beck, A. T. (1987). Cognitive models of depression. International Journal of Cognitive Therapy, 1, 5–37. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression. Cognitive Therapy of Depression, https://doi.org/10.1016/S0272-7358(02)00106-X. Berna, C., Lang, T. J., Goodwin, G. M., & Holmes, E. A. (2011). Developing a measure of interpretation bias for depressed mood: An ambiguous scenarios test. Personality and Individual Differences, https://doi.org/10.1016/j.paid.2011.04.005. Blackwell, S. E., & Holmes, E. A. (2010). Modifying interpretation and imagination in clinical depression: A single case series using cognitive bias modification. Applied Cognitive Psychology, 24(3), 338–350. https://doi.org/10.1002/acp.1680.

114  Cognitive biases in health and psychiatric disorders Blair, K. S., Otero, M., Teng, C., Jacobs, M., Odenheimer, S., Pine, D. S., et al. (2013). Dissociable roles of ventromedial prefrontal cortex (vmPFC) and rostral anterior cingulate cortex (rACC) in value representation and optimistic bias. NeuroImage, 78, 103–110. https://doi.org/10.1016/j. neuroimage.2013.03.063. Bower, G. H. (1981). Mood and Memory. The American Psychologist, 36(2), 129–148. https://doi. org/10.1016/0005-7967(87)90052-0. Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1999). Affect and the startle reflex. In Startle modification: Implications for neuroscience, cognitive science and clinical science. https://doi. org/10.1017/cbo9780511665523.010. Butler, G., & Mathews, A. (1983). Cognitive processes in anxiety. Advances in Behaviour Research and Therapy, 5(1), 51–62. https://doi.org/10.1016/0146-6402(83)90015-2. Byrne, A., & Eysenck, M. W. (1993). Individual differences in positive and negative interpretive biases. Personality and Individual Differences, 14(6), 849–851. https://doi.org/10.1016/01918869(93)90100-H. Byrne, D. G. (1976). Choice reaction times in depressive states. The British Journal of Social and Clinical Psychology, 15(2), 149–156. Canli, T., Sivers, H., Thomason, M. E., Whitfield-Gabrieli, S., Gabrieli, J. D., & Gotlib, I. H. (2004). Brain activation to emotional words in depressed vs healthy subjects. Neuroreport, 15(17), 2585–2588. https://doi.org/10.1097/00001756-200412030-00005. Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In Heimberg R. G., Liebowitz M., Hope D. A., & Schneier F. R. (Eds.), Social phobia: Diagnosis, assessment and treatment. Guilford Press. Disner, S. G., Beevers, C. G., Haigh, E. A. P., & Beck, A. T. (2011). Neural mechanisms of the cognitive model of depression. Nature Reviews. Neuroscience, 12(8), 467–477. https://doi. org/10.1038/nrn3027. Dricu, M., Kress, L., & Aue, T. (2020). The neurophysiological basis of optimism bias. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 41–70). San Diego: Elsevier. Everaert, J., Grahek, I., Duyck, W., Buelens, J., Van den Bergh, N., & Koster, E. H. W. (2017a). Mapping the interplay among cognitive biases, emotion regulation, and depressive symptoms. Cognition and Emotion, 31(4), 726–735. https://doi.org/10.1080/02699931.2016.1144561. Everaert, J., Koster, E. H. W., & Derakshan, N. (2012). The combined cognitive bias hypothesis in depression. Clinical Psychology Review, https://doi.org/10.1016/j.cpr.2012.04.003. Everaert, J., Podina, I. R., & Koster, E. H. W. (2017b). A comprehensive meta-analysis of interpretation biases in depression. Clinical Psychology Review, 58(September), 33–48. https://doi. org/10.1016/j.cpr.2017.09.005. Eysenck, M. W., Mogg, K., May, J., Richards, A., & Mathews, A. (1991). Bias in interpretation of ambiguous sentences related to threat in anxiety. Journal of Abnormal Psychology, 100(2), 144–150. https://doi.org/10.1037/0021-843X.100.2.144. Field, A. P., & Lester, K. J. (2010). Is there room for “Development” in developmental models of information processing biases to threat in children and adolescents? Clinical Child and Family Psychology Review, https://doi.org/10.1007/s10567-010-0078-8. Gotlib, I. H., Jonides, J., Buschkuehl, M., & Joormann, J. (2011). Memory for affectively valenced and neutral stimuli in depression: Evidence from a novel matching task. Cognition & Emotion, 25(7), 1246–1254. https://doi.org/10.1080/02699931.2010.538374. Hadwin, J., Frost, S., French, C. C., & Richards, A. (1997). Cognitive processing and trait anxiety in typically developing children: Evidence for an interpretation bias. Journal of Abnormal Psychology, 106(3), 486–490.

Positive interpretation bias across the psychiatric disorders  Chapter | 5  115 Harrison, B. J., Fullana, M. A., Via, E., Soriano-Mas, C., Vervliet, B., Martínez-Zalacaín, I., et al. (2017). Human ventromedial prefrontal cortex and the positive affective processing of safety signals. NeuroImage, https://doi.org/10.1016/j.neuroimage.2017.02.080. Hertel, P. T., Mathews, A., Peterson, S., & Kintner, K. (2003). Transfer of training emotionally biased interpretations. Applied Cognitive Psychology, 17(7), 775–784. https://doi.org/10.1002/acp.905. Heyman, G. D., & Giles, J. W. (2004). Valence effects in reasoning about evaluative traits. ­Merrill-Palmer Quarterly: Journal of Developmental Psychology, https://doi.org/10.1353/mpq.2004.0004. Heyman, G. D., & Legare, C. H. (2005). Children’s evaluation of sources of information about traits. Developmental Psychology, https://doi.org/10.1037/0012-1649.41.4.636. Hirsch, C. R., & Clark, D. M. (2004). Information-processing bias in social phobia. Clinical Psychology Review, 24(7), 799–825. https://doi.org/10.1016/j.cpr.2004.07.005. Hirsch, C. R., Clark, D. M., & Mathews, A. (2006). Imagery and interpretations in social phobia: Support for the combined cognitive biases hypothesis. Behavior Therapy, https://doi. org/10.1016/j.beth.2006.02.001. Hirsch, C. R., Meeten, F., Krahé, C., & Reeder, C. (2016). Resolving ambiguity in emotional disorders: The nature and role of interpretation biases. Annual Review of Clinical Psychology, https://doi.org/10.1146/annurev-clinpsy-021815-093436. Holmes, E. A., Lang, T. J., & Shah, D. M. (2009). Developing interpretation bias modification as a “cognitive vaccine” for depressed mood: Imagining positive events makes you feel better than thinking about them verbally. Journal of Abnormal Psychology, 118(1), 76–88. https://doi. org/10.1037/a0012590. Holmes, E. A., Mathews, A., Dalgleish, T., & Mackintosh, B. (2006). Positive interpretation training: Effects of mental imagery versus verbal training on positive mood. Behavior Therapy, 37(3), 237–247. https://doi.org/10.1016/j.beth.2006.02.002. Huppert, J. D., Foa, E. B., Furr, J. M., Filip, J. C., & Mathews, A. (2003). Interpretation bias in social anxiety: A dimensional perspective. Cognitive Therapy and Research, 27(5), 569–577. Ingram, R. E. (1984). Toward an information-processing analysis of depression. Cognitive Therapy and Research, 8(5), 443–477. https://doi.org/10.1007/BF01173284. Joormann, J., Waugh, C. E. E., & Gotlib, I. H. (2015). Cognitive bias modification for interpretation in major depression: Effects on memory and stress reactivity. Clinical Psychological Science, 3(1), 126–139. https://doi.org/10.1177/2167702614560748. Kim, H., Somerville, L. H., Johnstone, T., Alexander, A. L., & Whalen, P. J. (2003). Inverse amygdala and medial prefrontal cortex responses to surprised faces. Neuroreport, 14(18), 2317– 2322. https://doi.org/10.1097/01.wnr.0000101520.44335.20. Kim, H., Somerville, L. H., Johnstone, T., Polis, S., Alexander, A. L., Shin, L. M., et al. (2004). Contextual modulation of amygdala responsivity to surprised faces. Journal of Cognitive Neuroscience, 16(10), 1730–1745. https://doi.org/10.1162/0898929042947865. Kleim, B., Thörn, H. A., & Ehlert, U. (2014). Positive interpretation bias predicts well-being in medical interns. Frontiers in Psychology, 5(June), 1–6. https://doi.org/10.3389/fpsyg.2014.00640. Kress, L., & Aue, T. (2017). The link between optimism bias and attention bias: A neurocognitive perspective. Neuroscience and Biobehavioral Reviews, https://doi.org/10.1016/j.neubiorev.2017.07.016. Lang, T. J., Moulds, M. L., & Holmes, E. A. (2009). Reducing depressive intrusions via a computerized cognitive bias modification of appraisals task: Developing a cognitive vaccine. Behaviour Research and Therapy, 47(2), 139–145. https://doi.org/10.1016/j.brat.2008.11.002. Lawson, C., MacLeod, C., & Hammond, G. (2002). Interpretation revealed in the blink of an eye: Depressive bias in the resolution of ambiguity. Journal of Abnormal Psychology, 111(2), 321– 328. https://doi.org/10.1037//0021-843X.111.2.321.

116  Cognitive biases in health and psychiatric disorders LeMoult, J., & Gotlib, I. H. (2018). Depression: A cognitive perspective. Clinical Psychology Review, https://doi.org/10.1016/j.cpr.2018.06.008. Lester, K. J., Field, A. P., & Muris, P. (2011). Experimental modification of interpretation bias about animal fear in young children: Effects on cognition, avoidance behavior, anxiety vulnerability, and physiological responding. Journal of Clinical Child & Adolescent Psychology, 40(6), 864–877. https://doi.org/10.1080/15374416.2011.618449. Mangun, G. R., Buonocore, M. H., Girelli, M., & Jha, A. P. (1998). ERP and fMRI measures of visual spatial selective attention. Human Brain Mapping. http://dx.doi.org/10.1002/(SICI)10970193(1998)6:5/63.0.CO;2-Z. Mathews, A., & Mackintosh, B. (2000). Induced emotional interpretation bias and anxiety. Journal of Abnormal Psychology, 109(4), 602–615. https://doi.org/10.1037/0021-843X.109.4.602. Mathews, A., & MacLeod, C. (2005). Cognitive vulnerability to emotional disorders. Annual Review of Clinical Psychology, https://doi.org/10.1146/annurev.clinpsy.1.102803.143916. McKendree-Smith, N., & Scogin, F. (2000). Depressive realism: Effects of depression severity and interpretation time. Journal of Clinical Psychology, 56(12), 1601–1608. http://dx.doi. org/10.1002/1097-4679(200012)56:123.0.CO;2-K. Moretti, M. M., Segal, Z. V., McCann, C. D., Shaw, B. F., Miller, D. T., & Vella, D. (1996). Selfreferent versus other-referent information processing in dysphoric, clinically depressed, and remitted depressed subjects. Personality and Social Psychology Bulletin, 22(1), 68–80. https:// doi.org/10.1177/0146167296221007. Moser, J. S., Hajcak, G., Huppert, J. D., Foa, E. B., & Simons, R. F. (2008). Interpretation bias in social anxiety as detected by event-related brain potentials. Emotion, 8(5), 693–700. https://doi. org/10.1037/a0013173. Moser, J. S., Huppert, J. D., Foa, E. B., & Simons, R. F. (2012). Interpretation of ambiguous social scenarios in social phobia and depression: Evidence from event-related brain potentials. Biological Psychology, 89(2), 387–397. https://doi.org/10.1016/j.biopsycho.2011.12.001. Murphy, R., Hirsch, C. R., Mathews, A., Smith, K., & Clark, D. M. (2007). Facilitating a benign interpretation bias in a high socially anxious population. Behaviour Research and Therapy, https://doi.org/10.1016/j.brat.2007.01.007. Phelps, E. A., & LeDoux, J. E. (2005). Contributions of the amygdala to emotion processing: From animal models to human behavior. Neuron, https://doi.org/10.1016/j.neuron.2005.09.025. Piaget, J. (1952). Chapter III. The third stage: The “Secondary Circular Reactions” and the procedures destined to make interesting sights last. In The origins of intelligence in children. Piaget, J. (1955). The construction of reality in the child. Journal of Consulting Psychology, https:// doi.org/10.1037/h0038817. Rapee, R. M., & Heimberg, R. G. (1997). A cognitive-behavioral model of anxiety in social phobia. Behaviour Research and Therapy, 35(8), 741–756. https://doi.org/10.1016/S00057967(97)00022-3. Robbins, T. W., Cador, M., Taylor, J. R., & Everitt, B. J. (1989). Limbic-striatal interactions in reward-related processes. Neuroscience and Biobehavioral Reviews, https://doi.org/10.1016/ S0149-7634(89)80025-9. Rossion, B., Schiltz, C., & Crommelinck, M. (2003). The functionally defined right occipital and fusiform “face areas” discriminate novel from visually familiar faces. NeuroImage, https://doi. org/10.1016/S1053-8119(03)00105-8. Rude, S. S., Wenzlaff, R. M., Gibbs, B., Vane, J., & Whitney, T. (2002). Negative processing biases predict subsequent depressive symptoms. Cognition & Emotion, 16(3), 423–440. https://doi. org/10.1080/02699930143000554.

Positive interpretation bias across the psychiatric disorders  Chapter | 5  117 Santos, J. P., Seixas, D., Brandão, S., & Moutinho, L. (2011). Investigating the role of the ventromedial prefrontal cortex in the assessment of brands. Frontiers in Neuroscience, https://doi. org/10.3389/fnins.2011.00077. Sharot, T. (2011). The optimism bias. Current Biology, 21(23), R941–R945. https://doi. org/10.1016/j.cub.2011.10.030. Sharot, T., Riccardi, A. M., Raio, C. M., & Phelps, E. A. (2007). Neural mechanisms mediating optimism bias. Nature, https://doi.org/10.1038/nature06280. Tolgou, T., Rohrmann, S., Stockhausen, C., Krampen, D., Warnecke, I., & Reiss, N. (2018). Physiological and psychological effects of imagery techniques on health anxiety. Psychophysiology, 55(2), e12984. https://doi.org/10.1111/psyp.12984. Tran, T. B., Hertel, P. T., & Joormann, J. (2011). Cognitive bias modification: Induced interpretive biases affect memory. Emotion, https://doi.org/10.1037/a0021754. Vestre, N. D., & Caulfield, B. P. (1986). Perception of neutral personality descriptions by depressed and nondepressed subjects. Cognitive Therapy and Research, https://doi.org/10.1007/ BF01173380. Wadsworth, B. J. (1996). Piaget’s theory of cognitive and affective development: Foundations of constructivism. In Piaget’s theory of cognitive and affective development: Foundations of constructivism (5th ed.). https://doi.org/http://doi.apa.org/psycinfo/1996-97227-000. Wells, A. (1999). A cognitive model of generalized anxiety disorder. Behavior Modification, https:// doi.org/10.1177/0145445599234002. Wenzlaff, R. M., & Bates, D. E. (1998). Unmasking a cognitive vulnerability to depression: How lapses in mental control reveal depressive thinking. Journal of Personality and Social Psychology, 75(6), 1559–1571. https://doi.org/10.1037/0022-3514.75.6.1559. Williams, J. M. G., Watts, F. N., MacLeod, C. M., & Mathews, A. (1997). Cognitive psychology and emotional disorders. Vol. 226. The Wiley Series in Clinical Psychology.

Chapter 6

Resolving ambiguity: Negative interpretation biases K. Lira Yoon, Victoria Shaffer, Anna Benedict Department of Psychology, University of Notre Dame, Notre Dame, IN, United States

Introduction Ambiguity is abundant in everyday life, and individuals face the challenge of resolving ambiguity to understand what is happening and to deal with situations. Thus, accurate resolution of ambiguity is critical in one’s everyday functioning. Interpretation is the process of assigning meaning to ambiguous information, thereby resolving the ambiguity. This chapter focuses on negative interpretation biases in which ambiguous information is interpreted consistently in a negative manner. Individuals’ interpretations of a situation or information can guide their decision-making and behavior and influence their emotions. Thus, negative interpretation biases may play a critical role in psychological and health conditions, specifically in anxiety and depression (Matthews & MacLeod, 2002).

Major theories in the field Most cognitive models of anxiety and depression posit that information processing biases play a crucial role in the development and maintenance of anxiety and depression (e.g., Beck, Emery, & Greenberg, 1985; Hirsch, Clark, & Mathews, 2006; Ingram, Miranda, & Segal, 1999). Information processing biases refer to the notion that an individual’s processing of information is systematically distorted so that processing of information consistent with the person’s framework of preconceived ideas and representations of the world (i.e., a schema) is facilitated. In particular, negative interpretation biases refer to the tendencies in which ambiguous information is perceived in a negative manner. When an individual perceives neutral or ambiguous information as being more negative than is actually the case (i.e., negative interpretation biases), illness-related or threat-related/negative information may become overrepresented in cognition. As a result, subjective sense of threat in the environment is increased, thereby maintaining or worsening the disorder. It is important to note, however, that these cognitive models do not posit that individuals with psychopathology exhibit negative interpretation biases Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00006-4 © 2020 Elsevier Inc. All rights reserved.

119

120  Cognitive biases in health and psychiatric disorders

for all ambiguous information. Instead, (a) which ambiguous information an individual consistently interprets in a negative manner and (b) the content of their interpretations is dictated by the themes of their schemas (the contentspecificity hypothesis; Beck & Clark, 1997). For example, both an individual with depression and an individual with panic disorder may resolve ambiguity in a negative manner (i.e., negative interpretation biases). However, a person with panic disorder may interpret ambiguous bodily sensations as signs of a heart attack but not necessarily interpret a coworker’s ambiguous comment as rejection, whereas someone with depression may interpret the same comment as rejection but not perceive ambiguous bodily sensations in a negative manner. Unlike biases in some other cognitive processes, most notably biases in attention and imagery, there are no theories or models specifically on negative interpretation biases. As mentioned earlier, most cognitive theories include negative interpretation biases as one of the factors involved in psychopathology, with varying specifics (e.g., which ambiguous information is interpreted negatively and the relative importance of negative interpretation biases in the etiology of disorders). These theories are typically specific to one type of disorder, and reviewing details of each specific theory is beyond the scope of the chapter. Furthermore, similarities across different theories are more evident than the differences. Thus, a few theories on different disorders are reviewed in the succeeding text as examples to illustrate the proposed role of negative interpretation biases in their etiology. According to Clark and Wells’ (1995) cognitive model of social anxiety disorder, individuals process the self and social situations based on their negative assumptions (e.g., “People think I’m boring,” “I sound foolish,” etc.) resulting in increased anxiety and distress in social situations. Negative interpretation biases are even more emphasized in cognitive models of panic disorders (e.g., Casey, Oei, & Newcombe, 2004; Clark, 1986). These theories posit that catastrophic misinterpretation of benign bodily sensations (i.e., negative interpretation biases) results in increases in anxiety and autonomic responses that provide further evidence to the catastrophic interpretations, eventually leading to panic attacks. Negative interpretation biases have also been implicated in conditions other than psychological disorders, such as health and illness behavior. The common-sense model of illness representation (Diefenbach & Leventhal, 1996), for example, postulates that illness representations, or schemas, guide the way individuals assign personal meanings to symptoms and their appraisals of the outcome of their coping strategies (e.g., “I will feel better once I get some sleep,” “Nothing will stop this pain,” etc.). The tendency to interpret ambiguous somatic symptoms in an illness-related manner may perpetuate and worsen symptoms and distress in individuals with health conditions (e.g., pain disorder). Similarly, cognitive models of chronic fatigue syndrome (e.g., Fry & Martin, 1996) suggest that physical symptoms and associated disability are in part exacerbated by the tendency to interpret the symptoms to be serious, damaging, and uncontrollable.

Resolving ambiguity: Negative interpretation biases  Chapter | 6  121

In sum, cognitive models share the premise that negative interpretation biases play an important role in various psychological disorders and health conditions. In general, cognitive theories have hypothesized that individuals with (sub)clinical conditions tend to assign negative meanings to resolve disorderrelevant ambiguous information. These negative interpretation biases, in turn, would reinforce their existing negative or illness-related schemas and increase distress, thereby worsening negative thinking and symptoms.

Methods used to investigate the biases A wide range of paradigms have been used to examine negative interpretation biases, which can be categorized according to the type of stimuli used (e.g., words, scenarios, and facial pictures), the component of interpretation biases assessed (e.g., generation vs selection of interpretations), whether the biases are assessed directly (i.e., participants are asked to provide their interpretations of a stimulus) or indirectly (e.g., response times), and whether online (i.e., initial, immediate interpretations of stimuli) or offline (i.e., later, more reflective interpretations influenced by elaborative processes) interpretations are assessed. In all these tasks, participants are presented with ambiguous stimuli; tasks then assess whether the participants consistently resolve the ambiguity in a positive/ neutral or negative manner. Reviewing all the paradigms is beyond the scope of this chapter (see Schoth & Liossi, 2017, for review), and only some of the paradigms are reviewed in the succeeding text.

Direct measures of interpretation biases Paradigms that directly assess interpretation biases ask individuals to report on their interpretations of ambiguous information. These paradigms use a wide array of stimuli and differ greatly regarding the task that participants are asked to do. For example, participants might be presented with ambiguous scenarios and be asked to select (or rank order) their interpretations of the scenarios from a set of provided options (e.g., Butler & Mathews, 1983). In the scrambled sentence task, participants are presented with a series of words in random order that can be combined to form a benign or negative sentence (e.g., Wenzlaff & Bates, 1998). Typically, the number of sentences resolved in a negative manner is used as an index of negative interpretation biases, but the time it takes participants to unscramble the words to form negative sentences can also be used as a measure of interpretation biases. The latter index, however, is an indirect measure of interpretation biases, with shorter reaction times (RTs) to form negative (vs positive) sentences indicating negative interpretation biases. In the word sentence association task (e.g., Beard & Amir, 2009), a single prime word is presented, followed by an ambiguous sentence to which participants respond by indicating whether the sentence is related to the initial prime. For example, the sentence, “Someone is in your way,” appears following either the negative

122  Cognitive biases in health and psychiatric disorders

prime “Inconsiderate” or the benign prime “Unaware.” If the ambiguous sentence is interpreted in a negative manner, the person will endorse the sentence to be related to the negative primes. Both in the homophone (words with different spellings that sound identical) and homograph (a word with multiple meanings with the same spelling) tasks, words are presented auditorily. Participants are asked to write down the word (e.g., pane vs pain) in the homophone task (e.g., Mathews, Richards, & Eysenck, 1989), and they are asked to write down the first word that comes to their mind (e.g., terminal-airport vs disease) in the homograph task (e.g., Holmes, Lang, Moulds, & Steele, 2008). In both tasks, the number of responses that reflect the negative version of the words is used as an index of negative interpretation biases. Images, typically facial pictures, have also been used to assess interpretation biases. In an emotion recognition task, participants are asked to identify the emotion depicted by the face (e.g., Winton, Clark, & Edelmann, 1995). When prototypical facial expressions are used, the number of negative emotions assigned to neutral faces represents negative interpretation biases. For morphed faces that are created by blending two facial expressions (e.g., 50% angry/50% neutral, or 30% angry/70% neutral), the point at which a participant detects negative emotions is examined to assess negative interpretation biases (e.g., Joormann & Gotlib, 2006). A lower intensity of negative emotion required to detect the emotion suggests the presence of more negative interpretation biases. In the similarity rating task, participants are presented with pairs of faces expressing different intensity of emotions and asked to rate the similarity of each pair (e.g., Gebhardt & Mitte, 2014). The idea is that if, for example, socially anxious individuals have negative interpretation biases, moderate angry faces would be interpreted in a negative manner, and thus, moderate and extremely angry faces would be rate as more similar than by their less anxious counterparts. In general, paradigms that directly measure individuals’ interpretation biases or assess offline interpretation biases are prone to alternative explanations. Because these paradigms generally allow individuals enough time to reflect on their responses, individuals may not provide their first interpretation of a stimulus. Instead, their responses might reflect demand characteristics or response bias. Despite these limitations, offline or direct measures of interpretation biases, such as the tasks mentioned earlier, are easier to administer, and the outcomes are more straightforward.

Indirect measures of interpretation biases Based on the priming principle that exposure to an initial stimulus (i.e., a prime) facilitates the processing of related material, most indirect measures of interpretation biases rely on differences in RTs to different types of stimuli. As such, ambiguous primes are presented first, and participants are asked to

Resolving ambiguity: Negative interpretation biases  Chapter | 6  123

respond to a subsequent unambiguous target. Individuals’ RTs to targets that are related to the negative versus benign meaning of the prime are compared with assess the occurrence of negative interpretation biases (see Fig. 1). To the extent that participants cannot easily manipulate the indicators of interpretations (e.g., differences in RTs to various prime-target combinations), indirect measures of interpretation biases may reflect initial automatic interpretations rather than demand characteristics, response bias, or later interpretations following deliberation. On the lexical decision task with homographs (Richards & French, 1992), for example, a homograph is presented as a prime, followed by a target letter string that is either a real word or a nonword. Participants’ task is to indicate whether the target is a word or not. Importantly, real words are either related to the negative meaning (e.g., growth—tumor) or benign (e.g., growth—height) meaning of the prime. Faster RTs to negative (vs benign) targets suggest that the initial ambiguous primes have been interpreted in a negative manner. In a variant of lexical decision tasks, ambiguous scenarios have been used (e.g., Hirsch & Mathews, 1997). Typically, the last word of a scenario is presented by itself, and participants are either asked to indicate whether the word is (a) grammatically correct or logical or (b) a real word. The last word resolves the ambiguous scenario in either a negative or a benign manner; faster RTs to words that disambiguate the scenario in a negative (vs benign) manner indicate negative interpretation biases (Fig. 1A). Pairs of faces have also been used as a prime and a target (Yoon & Zinbarg, 2007). In this version, trials with neutral faces as primes serve as critical trials examining interpretation biases. Similar to the lexical decision task, faster RTs to negative (vs positive) target faces following neutral faces suggest the presence of negative interpretation biases (Fig. 1B). In the modified version of the word sentence association task (Fig. 1C), ambiguous sentences are presented before single words, and the participants are asked to judge whether the sentences are related to the subsequently presented words. Participants’ interpretations of the ambiguous sentences will affect their judgment of the word (direct measures) and the speed (i.e., RTs) with which they make the ratings (indirect measures). In sum, indirect measures of interpretation biases are more likely to reflect initial, automatic interpretations. However, because interpretations are inferred from differential behavioral or psychophysiological responses to negative versus positive stimuli, these measures have lower face validity than the direct measures. In contrast, direct measures are straightforward. Responses on the direct measures, however, might reflect demand characteristics, response bias, or later interpretations rather than individuals’ initial interpretations of ambiguous information. As will be reviewed in section “Similarities and differences between healthy and clinical populations,” studies using direct measures have largely demonstrated the presence of negative interpretation biases in (sub)clinical populations. Findings from studies using indirect measures, however, have been inconsistent.

124  Cognitive biases in health and psychiatric disorders

You wonder if, when you are in the interview, all of your preparation will be...

????????

Forgotten

If it is important to remember a particular detail, then it is annoying if it is...

????????

Forgotten

The interviewer asks you to tell them more about your last job and you think this means that they are...

????????

Agreeing

(A)

(B) Someone is in your way.

Someone is in your way.

Inconsiderate

Was the word related to the sentence?

Unaware

Was the word related to the sentence?

(C) FIG. 1  Examples of indirect measures of interpretation biases. Reaction times (RTs) to the last slides are compared to index negative interpretation biases. (A) A variant of a lexical decision task using ambiguous sentences. (B) A priming task with pictures of faces. Note that schematic faces, instead of facial pictures, are used in this figure for illustrative purposes. (C) A modified version of the word sentence association task.

Resolving ambiguity: Negative interpretation biases  Chapter | 6  125

Brain regions involved in the emergence of the bias Research on brain regions involved in resolving ambiguity examined brain activations to ambiguous sentences. It is important to note from the outset that this line of research is mostly on resolution of nonemotional stimuli: extant research has not examined negative interpretation biases per se. Furthermore, to our best knowledge, no studies have examined brain regions involved in the resolution of lexical ambiguity in clinical populations. Ambiguous sentences were constructed by including a homograph (e.g., bank), which was disambiguated by the context in a manner that was either consistent with the dominant (i.e., money) or the subordinate (i.e., river) meaning of the homograph. Although multiple regions have been implicated, the left prefrontal cortex, including the left inferior frontal gyrus, seems to be the area most consistently being involved when evaluating ambiguity (e.g., Rodd, Johnsrude, & Davis, 2012; Zempleni, Renken, Hoeks, Hoogduin, & Stowe, 2007). In particular, greater activation in the left inferior frontal gyrus is associated with the need to suppress an initial selection of dominant, highly expected meaning, in favor of less expected, subordinate meaning. A recent study used “strongly biased homographs” that had one meaning strongly dominant over the other meaning (95% preference for the dominant meaning) to create ambiguous sentences, and participants were asked to judge whether each sentence was logical or not (Mestres-Missé, Trampel, Turner, & Kotz, 2016). For ambiguous sentences that included homographs, the logical sentences were always consistent with the strong dominant meaning of the homographs. Thus, interpretations of both sentences with strongly biased homographs and unambiguous words were consistent with expected meanings of the words. Nevertheless, participants were less accurate when judging ambiguous, compared with unambiguous, congruent sentences, reflecting a degree of uncertainty due to the subordinate meaning (albeit extremely low) for strongly biased homographs. In addition, ambiguous congruent sentences activated cognitive control and performance monitoring networks such as the left lateral prefrontal cortex and insula more than unambiguous sentences, suggesting that the brain represents a “probability distribution” of all possible outcomes. Thus, brain areas implicated in context processing and selecting among competing representations seem to be involved in the resolution of lexical ambiguity. As mentioned earlier, this line of research did not examine brain activations of (sub)clinical populations nor examined negative interpretation biases. Given the involvement of the prefrontal cortex in resolving (lexical) ambiguity, one may speculate that the prefrontal cortex may be involved in negative interpretation biases. In this regard, it is of note that individuals with psychological disorders (vs healthy controls) tend to exhibit lower prefrontal cortex activation when processing emotional information (e.g., Bishop, 2007). The amygdala plays a critical role in emotion processing, in particular threat processing (e.g., Ledoux, 1998). Along this line, (sub)clinical

126  Cognitive biases in health and psychiatric disorders

p­ opulations, ­compared with healthy controls, exhibit greater amygdala activations in response to ambiguous stimuli that can be interpreted as negative (e.g., see Hattingh et al., 2012, for review). In addition, the connectivity between the amygdala and anterior cingulate cortex has been suggested as an aversive amplification circuit in which increased coupling between these two regions is related to greater threat processing under stress (Robinson et  al., 2013; Robinson, Charney, Overstreet, Vytal, & Grillon, 2012). Furthermore, increased amygdala activity, coupled with less activation of the dorsolateral prefrontal cortex (DLPFC) that serves to regulate or inhibit amygdala activity (e.g., Motzkin, Philippi, Wolf, Baskaya, & Koenigs, 2015), may be associated with a greater tendency to perceive ambiguous stimuli as more emotionally significant than they actually are (i.e., negative interpretation biases). Negative interpretation biases may result in a negative affective state in response to the ambiguous stimuli (e.g., Phillips, Drevets, Rauch, & Lane, 2003a, 2003b). Increased affective states, in turn, may strengthen amygdala activity, thereby augmenting threat/negative evaluations and making the selection of negative interpretations more likely (Bishop, 2007). When faced with ambiguous information, competition among alternative interpretations occurs to resolve the ambiguity, which is reflected in higher prefrontal cortex activity (e.g., Mestres-Missé et al., 2016). One may therefore speculate that lower prefrontal cortex activity in (sub)clinical samples goes along with competing alternatives not being as strongly represented and/or alternative meanings not being entertained. Limited and fragmented evidence, thus, points to a possibility that a combination of an overactive amygdala and a weak prefrontal cortex, DLPFC in particular, that fails to regulate the limbic system may be associated with negative interpretation biases (e.g., Bishop, 2007; Disner, Beevers, Haigh, & Beck, 2011). These patterns of brain activity (Fig. 2) suggest that negative representations might be more activated when processing ambiguous information. These speculations, however, need to be empirically examined. To better understand brain mechanisms underlying negative interpretation biases, studies should examine differences, if any, in brain activations, including connectivity between different regions of the brain, when ambiguity is resolved in a negative versus positive manner. Future studies should also examine neural underpinnings of negative interpretation biases in (sub)clinical populations.

Somatovisceral responses related to the biases As mentioned earlier, findings from research relying on direct measures of negative interpretation biases might reflect demand characteristics and/or response bias rather than genuine interpretation biases. One way to eliminate these alternative explanations is to employ psychophysiological measures by themselves or in conjunction with self-reports or behavioral data (e.g., RTs). Despite benefits of using psychophysiological measures, there is paucity in research examining somatovisceral responses in relation to interpretation biases.

Resolving ambiguity: Negative interpretation biases  Chapter | 6  127

FIG. 2  Brain regions possibly involved in interpretation biases. Potentially negative information overly activates the limbic system, including the amygdala, which sends strong signals to the prefrontal cortex (PFC). Hypoactivity in the PFC is associated with less top-down cognitive control from the dorsolateral prefrontal cortex (DLPFC) over the amygdala (shown in gray). Greater connectivity between the amygdala and anterior cingulate cortex (ACC), coupled with less connectivity between the DLPFC and amygdala, may result in negative interpretations of ambiguous information.

In one such rare study (Kirschner, Hilbert, Hoyer, Lueken, & Beesdo-Baum, 2016), participants’ skin conductance responses (SCRs) were assessed while they were presented with pictures of dangerous, safe, or ambiguous content. These pictures were preceded by an anticipatory cue that always correctly indicated the content of a subsequent picture (i.e., a certain cue) or an uncertain cue that did not provide information about the content of a subsequent picture. High worriers exhibited negative interpretation biases in that they rated ambiguous pictures, which were always preceded by uncertain cues, as more dangerous, anxiety provoking, and difficult to tolerate than the less worried counterparts. High worriers’ SCRs were elevated during both the uncertain and certain anticipation phase, but SCRs did not differ between the high and the low worriers while actually viewing the pictures. Thus, high worriers negatively interpreted ambiguous stimuli (as reflected in their ratings) without exhibiting greater autonomic arousal than the low worriers while viewing the pictures. These findings are in line with cognitive theories of worry and generalized anxiety disorder that posit worrying is being negatively reinforced because it restrains physiological arousal that could be aversive (Borkovec, Alcaine, & Behar, 2004). The results, however, may alternatively indicate that high worriers’ seemingly negative interpretation biases may reflect demand characteristics and/or response biases. Affective startle responses have also been used in a few studies as an objective, indirect measure of potential negative interpretation biases. The magnitude

128  Cognitive biases in health and psychiatric disorders

of the startle responses is greater for negative and smaller for positive, than neutral stimuli (e.g., Larson, Ruffalo, Nietert, & Davidson, 2005; Vrana & Lang, 1990). Thus, negative (vs benign) interpretations of ambiguous information would result in larger startle responses. In a college student sample, participants with higher levels of depressive symptoms exhibited greater startle responses to ambiguous auditory stimuli that were created by blending neutral and negative word pairs that were acoustically identical except for one phoneme (e.g., dress and stress), suggesting the presence of negative interpretation biases (Lawson, MacLeod, & Hammond, 2002). A more recent study with a clinical sample with depression, however, failed to replicate the finding (Käse et al., 2013). Clearly, more research is needed to better understand psychophysiological responses associated with negative interpretation biases.

Similarities and differences between healthy and clinical populations Negative interpretation biases have been most extensively examined in social anxiety disorder and to a lesser degree in depression, but they may characterize individuals with various other conditions (see Hirsch, Meeten, Krahé, & Reeder, 2016, for review). Indeed, research has found negative interpretation biases in various (sub)clinical populations, including social anxiety (e.g., Coles, Heimberg, & Schofield, 2008), depression (see Everaert, Podina, & Koster, 2017, for review), eating disorders (e.g., Brockmeyer et al., 2018), somatoform and related disorders (e.g., Woud, Zhang, Becker, McNally, & Margraf, 2014), and personality disorders (e.g., Lobbestael & McNally, 2016). This is especially true when direct measures were used, but studies with indirect measures have yielded mixed results (e.g., Hindash & Amir, 2012; Lawson & MacLeod, 1999). In one of the first studies, participants in two clinical groups (generalized anxiety disorder and depression) chose negative options more frequently than did controls when presented with ambiguous social scenarios (Butler & Mathews, 1983), suggesting the presence of negative interpretation biases. Similarly, when participants were asked to write down the word they heard in a homophone task (e.g., die/dye), participants diagnosed with depression wrote more negative words than their never-depressed counterparts (Mogg, Bradbury, & Bradley, 2006). Similarly, when presented with ambiguous social vignettes that included ambiguous statements about interpersonal (e.g., When meeting her date, Lisa said “You’re certainly not what I expected”) and noninterpersonal (e.g., Upon entering the restaurant, Lisa said “This is an unusual place”) evaluation, college students with higher levels of social anxiety endorsed more negative interpretations of ambiguous, interpersonal events compared with the low anxiety group (Constans, Penn, Ihen, & Hope, 1999). Socially anxious participants, however, did not differ from their less anxious counterparts in their interpretations of noninterpersonal statements, supporting the content specificity in their biases (Beck & Clark, 1997). Ambiguous social scenarios elicited

Resolving ambiguity: Negative interpretation biases  Chapter | 6  129

similar negative interpretation biases in individuals with social anxiety disorder, but not in a clinical control group with individuals suffering from other anxiety disorders (specific phobia, panic disorder, and agoraphobia), demonstrating once more content specificity (Stopa & Clark, 2000). Negative interpretation biases have also been noted in psychological disorders other than emotional disorders. Women with anorexia nervosa, for example, exhibited body-related negative interpretation biases on a scrambled sentences task (Brockmeyer et al., 2018). In this study, participants were given only 10 s to unscramble each sentence while concurrently remembering a sixdigit number. Despite the use of a direct measure, participants’ responses likely reflected their automatic, initial interpretations due to the added cognitive load and time pressure. People with health conditions also exhibit negative interpretation biases. When asked to write down the first word that came to mind after hearing homophones that are illness related (e.g., week/weak), individuals with chronic fatigue syndrome were more likely to produce illness-related words (MossMorris & Petrie, 2003). Similarly, individuals with chronic pain were more likely to interpret ambiguous stimuli in a pain- or health-related manner. When patients with chronic pain were asked to write down the first word that came to mind, they produced more pain-related words in response to ambiguous cues on the homograph task (Pincus, Pearce, McClelland, Farley, & Vogel, 1994), suggesting the presence of negative interpretation biases. In sum, previous studies using direct measures have largely demonstrated the presence of negative interpretation biases. Studies with indirect measures of interpretation biases, however, have yielded mixed results. Although some studies found that (sub)clinical samples exhibited shorter RTs to negative targets compared with positive targets following ambiguous primes, suggesting the presence of interpretation bias (e.g., Hindash & Amir, 2012; Yoon & Zinbarg, 2008), other studies have failed to find differences in RTs between individuals with high versus low levels of depressive symptoms (e.g., Lawson & MacLeod, 1999). Thus, (sub)clinical populations may not differ from healthy controls in their actual, initial interpretations of ambiguous information. A major line of research on interpretation biases in social anxiety, and to some extent depression, has examined the perception of neutral facial expressions. Studies asking participants to categorize facial expressions demonstrated individuals with clinical depression to be slower and less accurate (e.g., Leppanen, Milders, Bell, Terriere, & Hietanen, 2004) in identifying a neutral expression than controls. Studies that used morphed faces also indicated the presence of interpretation biases in (sub)clinical samples. For example, women diagnosed with depression exhibited negative interpretation biases as reflected by the fact that they required less intensity in the morphed sad faces to correctly identify the expression than controls (Bento de Souza, Barbosa, Lacerda, dos Santos, & Torro-Alves, 2014). Interpreting subtle or ambiguous facial expressions in a negative manner may be due to (socially) anxious individuals’ use

130  Cognitive biases in health and psychiatric disorders

of a more liberal criterion (i.e., response bias) to detect negative cues (Frenkel, Lamy, Algom, & Bar-Haim, 2009; Yoon, Yang, Chong, & Oh, 2014). Anxious individuals, however, were also more sensitive to subtle, but actually present, negative cues than their less anxious counterparts (e.g., Yang, Yoon, Chong, & Oh, 2013; Yoon et al., 2014). Along this line, findings from studies employing indirect measures also indicated that individuals with high levels of social anxiety tended to perceive neutral faces in a negative manner (e.g., Yoon & Zinbarg, 2007, 2008). Some studies used event-related brain potentials (ERPs) to assess interpretation biases. Given that ERPs are direct measures of online neural activity with excellent temporal resolution, they are particularly suited to assess individuals’ online, real-time, initial interpretations of ambiguous information. In a study with a subclinical sample, individuals with high levels of social anxiety performed faster when ambiguous social sentences were completed with a negative compared with positive terminal word, suggesting the presence of negative interpretation biases (Moser, Hajcak, Huppert, Foa, & Simons, 2008). In this study, the P600 component of ERP, which reflects violations of expectations, was larger when ambiguous sentences ended in a negative (vs positive) manner in individuals with low levels of social anxiety, suggesting that they interpreted ambiguous sentences in a positive manner. In contrast, the amplitude of P600 did not differ in response to negative versus positive endings in individuals with high levels of social anxiety. That is, despite behavioral data indicating the presence of negative biases in individuals with high levels of social anxiety, ERP data did not suggest such biases. Instead, ERP data indicated positive biases in individuals with low levels of social anxiety. Similarly, in a clinical sample with social anxiety disorder, behavioral data indicated the presence of negative interpretation biases, whereas ERP data indicated a lack of positive bias that was present in healthy controls (Moser, Huppert, Foa, & Simons, 2012). Healthy individuals, however, are not immune to negative interpretation biases. College students without psychopathology or complaints regarding acute or chronic pain, for example, showed negative interpretation biases on the lexical decision task (Vancleef, Peters, & de Jong, 2009). Participants were presented with ambiguous scenarios, in which the probe (i.e., the last word) could disambiguate the sentence in a health-threatening or safe manner. Participants were faster to respond to probes that resolved ambiguity in a health-threatening manner. Thus, even healthy people may at times be more likely to make negative inferences when processing health- or illness-related ambiguity.

Negative interpretation biases: A risk factor for anxiety and depression According to most cognitive theories (e.g., Beck & Clark, 1997; Mathews & MacLeod, 2005), negative interpretation biases play a causal role in the development of psychopathology, most notably anxiety and depression. That is,

Resolving ambiguity: Negative interpretation biases  Chapter | 6  131

a person who interprets ambiguous situations in a negative manner will likely experience more negative emotions and stress, thereby increasing likelihood of developing emotional disorders. To examine whether interpretation biases play a role in the genesis of emotional disorders, researchers used cognitive bias modification paradigms in which cognitive biases were induced or trained. In cognitive bias modification programs that target interpretation biases, individuals are repeatedly presented with ambiguous information whose interpretation is constrained in a particular direction by consistently resolving the ambiguity in a positive/benign or in a negative manner, depending on the training condition. Pioneering work by Andrew Mathews (e.g., Grey & Mathews, 2000; Mathews & Mackintosh, 2000) demonstrated that inducing interpretation biases affect emotions. Following the training designed to induce negative interpretation biases, healthy participants demonstrated an increase in their anxiety levels compared with their baseline (e.g., Mathews & Mackintosh, 2000) and in response to a subsequent stressor (Mackintosh, Mathews, Yiend, Ridgeway, & Cook, 2006; Wilson, MacLeod, Mathews, & Rutherford, 2006). Similarly, following a training in which ambiguous scenarios were resolved in either a negative self-relevant way or a positive self-relevant way, participants in the negative training condition demonstrated a decrease in self-esteem after a subsequent stress task, whereas participants in the positive training condition did not exhibit such change in their self-esteem (Tran, Siemer, & Joormann, 2011). Findings from a recent interpretation bias training study (Hertel, Mor, Ferrari, Hunt, & Agrawal, 2014) further suggest that negative interpretation biases play a causal role in rumination, a known risk factor for depression that is also associated with other psychological disorders (see Nolen-Hoeksema, Wisco, & Lyubomirsky, 2008, for review). Importantly, symptom reduction in a subclinical sample with elevated levels of social anxiety was mediated by increases in benign interpretations following benign interpretation training (Beard & Amir, 2008). Together, these findings suggest that interpretation biases play a causal role in anxiety and depression. Other studies using longitudinal designs to examine the relation between interpretation biases and prospective symptoms also support the notion that negative interpretation biases play a causal role in emotional disorders. For example, negative interpretation biases assessed by the scrambled sentences test under cognitive load predicted diagnoses of depression 18–28 months later (Rude, Wenzlaff, Gibbs, Vane, & Whitney, 2002), suggesting that the presence of negative interpretation biases may confer risk for future depression. Another line of evidence that relates interpretation biases to psychopathology comes from research with high-risk populations. For example, young daughters of depressed mothers who themselves have never experienced depression were more likely to select negative options for ambiguous words that were created by acoustically blending word pairs (e.g., sad and sand) and faster to respond to negative (vs positive) words following ambiguous scenarios (Dearing & Gotlib, 2009). That is, girls at increased risk for developing depression

132  Cognitive biases in health and psychiatric disorders

e­ xhibited negative i­nterpretation biases even though they were not (yet) depressed. Thus, negative interpretation biases do not seem to be mere symptoms or sequelae of emotional disorders but a cognitive tendency that precedes the onset of psychopathology.

Limitations Individuals with psychological disorders and other health conditions (e.g., pain disorder) have been hypothesized to consistently resolve ambiguity in a negative manner. Although research using direct measures of interpretation biases largely supports this notion, evidence from research using indirect measures is mixed. Therefore, it would be important for future research to distinguish whether psychological and health conditions are associated with genuine initial negative interpretation biases or whether the previous findings using direct measures reflect response bias and/or more elaborate, later interpretations. That being said, both genuine initial interpretation biases and response bias/ later interpretations are likely to have the same consequences in the real world. A socially anxious individual, for example, might first interpret other people’s neutral facial expressions as neutral (i.e., no bias in her initial interpretation). Either due to her response bias or changes in her interpretation following deliberation or elaboration of information (i.e., later interpretation), she may conclude that the neutral expression is a negative one. The consequence will be the same as if she interpreted neutral expression as negative in the first place. That is, the socially anxious person may still try to avoid interacting with the other person with a neutral expression, interact in a defensive manner, and/or experience an increase in anxiety and distress. In other words, the sequelae of response bias and later interpretations based on more elaborate processing would be similar to those of initial automatic negative interpretation biases. Nevertheless, these two biases could have different implications for treatment. Employing ERP may be useful in differentiating the two biases and provide better understanding of the underlying mechanisms of negative interpretation biases. The next two limitations may have contributed to mixed findings in the literature using indirect measures of interpretation biases. First, a wide range of tasks has been used to assess negative interpretation biases. Using different tasks allows conceptual replication of previous findings and provides confidence that the findings are not task dependent. At the same time, however, it is challenging to synthesize findings and to pinpoint the source of mixed findings. Methodological consistency and efforts to replicate findings using the same tasks would facilitate a systematic examination of the exact conditions under which negative interpretation biases are observed, thus helping us draw more firm conclusions. Second, psychometric properties of interpretation bias measures are virtually unknown. Given the revelation that tasks designed to assess attentional biases (e.g., dot-probe task) suffer from low reliability (e.g., Okon-Singer, 2018), it would not be too surprising to discover that reliability

Resolving ambiguity: Negative interpretation biases  Chapter | 6  133

of interpretation bias tasks is also inadequate. Researchers should examine and report psychometric properties of the tasks used in their studies.

Future directions Despite extensive research on negative interpretation biases, a number of key questions still remain unanswered. In addition to the issues already discussed earlier, most previous studies examined the biases without any consideration of other information processing biases (e.g., attention and memory) that have also been implicated in psychopathology. It is, however, clear that these biased cognitive processes do not operate in isolation (e.g., Everaert et al., 2017): different biases reinforce each other, thereby creating a vicious cycle leading to increases in distress and symptoms. Negative interpretation biases, for example, play a key role in negative memory biases in that individuals’ memory of ambiguous information is consistent with the manner with which the individuals resolved ambiguity (Hertel, Brozovich, Joormann, & Gotlib, 2008). Similarly, interpretation bias training can affect biases in attention (Amir, Bomyea, & Beard, 2010) and memory (Tran et al., 2011). Thus, future research should assess multiple cognitive processes in a single sample to examine how interpretation biases interact with other biases to affect the onset and maintenance of psychopathology. The difference between clinical and healthy populations may be that the connections between various cognitive biases are much weaker in healthy individuals and, thus, the vicious cycle can be easily broken through (Aue & Okon-Singer, 2015). Although clinical populations as a group seem to exhibit negative interpretation biases (at least with direct measures) compared with healthy controls, it is important to keep in mind that there is large variability even within clinical populations. Thus, it would be important to understand differences between individuals with a (sub)clinical condition who do and do not exhibit such biases. Relatedly, it would be important to use within-person approaches to understand the relation between changes in the interpretation biases and changes in the symptoms in a given person. Thus, instead of (or in addition to) focusing on between-group differences, future research should strive to understand factors related to within-person changes using intensive longitudinal designs. Cognitive restructuring, the core procedure of cognitive behavior therapy (CBT), which is effective in treating a number of conditions (e.g., Hofmann, Asnaani, Vonk, Sawyer, & Fang, 2012), essentially targets negative interpretation biases by helping clients to generate alternative, more realistic interpretations of situations. As such, changes in the biases might be one mechanism by which CBT leads to favorable outcomes (Morrison & Heimberg, 2013). Thus, assessing changes in the biases throughout the course of treatment to examine if baseline, and/or changes in, negative interpretation biases predict treatment outcomes would be another important next step. Such information could help us eventually identify who will benefit from CBT and/or interpretation bias modification training, which could save resources and provide more individually tailored care.

134  Cognitive biases in health and psychiatric disorders

Summary Research employing direct measures of interpretation biases provides evidence that individuals with elevated symptoms of various psychological disorders, most notably social anxiety disorder and depression, and health conditions exhibit negative interpretation biases. Although evidence is mixed when indirect measures are used, negative interpretation biases seem to play an important role in the maintenance and exacerbation of distress and symptoms. More research is warranted to address numerous gaps noted earlier to advance our understanding and treatment of psychopathology.

References Amir, N., Bomyea, J., & Beard, C. (2010). The effect of single-session interpretation modification on attention bias in socially anxious individuals. Journal of Anxiety Disorders, 24(2), 178–182. https://doi.org/10.1016/j.janxdis.2009.10.005. Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. https://doi.org/10.1016/j.cpr.2015.08.005. Beard, C., & Amir, N. (2008). A multi-session interpretation modification program: Changes in interpretation and social anxiety symptoms. Behaviour Research and Therapy, 46(10), 1135–1141. Beard, C., & Amir, N. (2009). Interpretation in social anxiety: When meaning precedes ambiguity. Cognitive Therapy and Research, 33, 406–415. Beck, A. T., & Clark, D. A. (1997). An information processing model of anxiety: Automatic and strategic processes. Behavior Research and Therapy, 35(1), 49–58. Beck, A. T., Emery, G., & Greenberg, R. (1985). Anxiety disorders and phobias: A cognitive perspective. New York, NY: Basic Books. Bento de Souza, I. B. M., Barbosa, F. F., Lacerda, A. M., dos Santos, N. A., & Torro-Alves, N. (2014). Evaluation of facial expressions in women with major depression: Is there a negative bias? Psychology & Neuroscience, 7(4), 513–519. https://doi.org/10.3922/j.psns.2014.4.10. Bishop, S. J. (2007). Neurocognitive mechanisms of anxiety: An integrative account. Trends in Cognitive Sciences, 11(7), 307–316. https://doi.org/10.1016/j.tics.2007.05.008. Borkovec, T. D., Alcaine, O. M., & Behar, E. (2004). Avoidance theory of worry and generalized anxiety disorder. In R. G. Heimberg, C. L. Turk, & D. S. Mennin (Eds.), Generalized anxiety disorder: Advances in research and practice (pp. 77–108). New York: Guilford Press. Brockmeyer, T., Anderle, A., Schmidt, H., Febry, S., Wünsch-Leiteritz, W., Leiteritz, A., et  al. (2018). Body image related negative interpretation bias in anorexia nervosa. Behaviour Research and Therapy, 104, 69–73. https://doi.org/10.1016/j.brat.2018.03.003. Butler, G., & Mathews, A. (1983). Cognitive processes in anxiety. Advances in Behaviour Research and Therapy, 5(1), 51–62. https://doi.org/10.1016/0146-6402(83)90015-2. Casey, L. M., Oei, T. P. S., & Newcombe, P. A. (2004). An integrated cognitive model of panic disorder: The role of positive and negative cognitions. Clinical Psychology Review, 24(5), 529–555. https://doi.org/10.1016/j.cpr.2004.01.005. Clark, D. M. (1986). A cognitive approach to panic. Behaviour Research and Therapy, 24(4), 461–470. https://doi.org/10.1016/0005-7967(86)90011-2. Clark, D. M., & Wells, A. (1995). A cognitive model of social phobia. In R. G.  Heimberg, M. R. Liebowitz, D. A. Hope, & F. R. Schneier (Eds.), Social phobia: Diagnosis, assessment, and treatment (pp. 69–93). New York: The Guilford Press.

Resolving ambiguity: Negative interpretation biases  Chapter | 6  135 Coles, M. E., Heimberg, R. G., & Schofield, C. A. (2008). Interpretation of facial expressions and social anxiety: Specificity and source of biases. Cognition and Emotion, 22(6), 1159–1173. Constans, J. I., Penn, D. L., Ihen, G. H., & Hope, D. A. (1999). Interpretive biases for ambiguous stimuli in social anxiety. Behaviour Research and Therapy, 37(7), 643–651. https://doi. org/10.1016/S0005-7967(98)00180-6. Dearing, K. F., & Gotlib, I. H. (2009). Interpretation of ambiguous information in girls at risk for depression. Journal of Abnormal Child Psychology, 37(1), 79–91. https://doi.org/10.1007/ s10802-008-9259-z. Diefenbach, M. A., & Leventhal, H. (1996). The common-sense model of illness representation: Theoretical and practical considerations. Journal of Social Distress and the Homeless, 5(1), 11–38. Disner, S. G., Beevers, C. G., Haigh, E. A., & Beck, A. T. (2011). Neural mechanisms of the cognitive model of depression. Nature Reviews Neuroscience, 12(8), 467. Everaert, J., Grahek, I., Duyck, W., Buelens, J., Van den Bergh, N., & Koster, E. H. (2017). Mapping the interplay among cognitive biases, emotion regulation, and depressive symptoms. Cognition and Emotion, 31(4), 726–735. Everaert, J., Podina, I. R., & Koster, E. H. W. (2017). A comprehensive meta-analysis of interpretation biases in depression. Clinical Psychology Review, 58, 33–48. https://doi.org/10.1016/j. cpr.2017.09.005. Frenkel, T. I., Lamy, D., Algom, D., & Bar-Haim, Y. (2009). Individual differences in perceptual sensitivity and response bias in anxiety: Evidence from emotional faces. Cognition & Emotion, 23(4), 688–700. https://doi.org/10.1080/02699930802076893. Fry, A. M., & Martin, M. (1996). Fatigue in the chronic fatigue syndrome: A cognitive phenomenon? Journal of Psychosomatic Research, 41(5), 415–426. https://doi.org/10.1016/S00223999(96)00190-0. Gebhardt, C., & Mitte, K. (2014). Seeing through the eyes of anxious individuals: An investigation of anxiety-related interpretations of emotional expressions. Cognition and Emotion, 28(8), 1367–1381. https://doi.org/10.1080/02699931.2014.881328. Grey, S., & Mathews, A. (2000). Effects of training on interpretation of emotional ambiguity. The Quarterly Journal of Experimental Psychology, 53A(4), 1143–1162. Hattingh, C. J., Ipser, J., Tromp, S. A., Syal, S., Lochner, C., Brooks, S. J., et  al. (2012). Functional magnetic resonance imaging during emotion recognition in social anxiety disorder: An activation likelihood meta-analysis. Frontiers in Human Neuroscience, 6, 347. https://doi. org/10.3389/fnhum.2012.00347. Hertel, P. T., Brozovich, F., Joormann, J., & Gotlib, I. H. (2008). Biases in interpretation and memory in generalized social phobia. Journal of Abnormal Psychology, 117(2), 278–288. https://doi. org/10.1037/0021-843X.117.2.278. Hertel, P., Mor, N., Ferrari, C., Hunt, O., & Agrawal, N. (2014). Looking on the dark side: Rumination and cognitive-bias modification. Clinical Psychological Science, 2(6), 714–726. Hindash, A. H. C., & Amir, N. (2012). Negative interpretation bias in individuals with depressive symptoms. Cognitive Therapy and Research, 36(5), 502–511. https://doi.org/10.1007/s10608011-9397-4. Hirsch, C. R., Clark, D. M., & Mathews, A. (2006). Imagery and interpretations in social phobia: Support for the combined cognitive biases hypothesis. Behavior Therapy, 37(3), 223–236. https://doi.org/10.1016/j.beth.2006.02.001. Hirsch, C. R., & Mathews, A. (1997). Interpretative inferences when reading about emotional events. Behaviour Research and Therapy, 35(12), 1123–1132. https://doi.org/10.1016/S00057967(97)80006-X.

136  Cognitive biases in health and psychiatric disorders Hirsch, C. R., Meeten, F., Krahé, C., & Reeder, C. (2016). Resolving ambiguity in emotional disorders: The nature and role of interpretation biases. Annual Review of Clinical Psychology, 12, 281–305. Hofmann, S. G., Asnaani, A., Vonk, I. J. J., Sawyer, A. T., & Fang, A. (2012). The efficacy of cognitive behavioral therapy: A review of meta-analyses. Cognitive Therapy and Research, 36(5), 427–440. https://doi.org/10.1007/s10608-012-9476-1. Holmes, E. A., Lang, T. J., Moulds, M. L., & Steele, A. M. (2008). Prospective and positive mental imagery deficits in dysphoria. Behaviour Research and Therapy. Ingram, R. E., Miranda, J., & Segal, Z. V. (1999). Cognitive vulnerability to depression. New York, NY: Guilford Press. Joormann, J., & Gotlib, I. H. (2006). Is this happiness I see? Biases in the identification of emotional facial expressions in depression and social phobia. Journal of Abnormal Psychology, 115(4), 705–714. https://doi.org/10.1037/0021-843x.115.4.705. Käse, M., Dresler, T., Andreatta, M., Ehlis, A.-C., Wolff, B., Kittel-Schneider, S., et al. (2013). Is there a negative interpretation bias in depressed patients? An affective startle modulation study. Neuropsychobiology, 67(4), 201–209. Kirschner, H., Hilbert, K., Hoyer, J., Lueken, U., & Beesdo-Baum, K. (2016). Psychophysiological reactivity during uncertainty and ambiguity processing in high and low worriers. Journal of Behavior Therapy and Experimental Psychiatry, 50, 97–105. Larson, C. L., Ruffalo, D., Nietert, J. Y., & Davidson, R. J. (2005). Stability of emotion-modulated startle during short and long picture presentation. Psychophysiology, 42(5), 604–610. Lawson, C., & MacLeod, C. (1999). Depression and the interpretation of ambiguity. Behaviour Research and Therapy, 37(5), 463–474. Lawson, C., MacLeod, C., & Hammond, G. (2002). Interpretation revealed in the blink of an eye: Depressive bias in the resolution of ambiguity. Journal of Abnormal Psychology, 111, 321–328. Ledoux, J. (1998). The emotional brain: The mysterious underpinnings of emotional life. New York: Simon and Schuster. Leppanen, J. M., Milders, M., Bell, J. S., Terriere, E., & Hietanen, J. K. (2004). Depression biases the recognition of emotionally neutral faces. Psychiatry Research, 128, 123–133. Lobbestael, J., & McNally, R. J. (2016). An empirical test of rejection-and anger-related interpretation bias in borderline personality disorder. Journal of Personality Disorders, 30(3), 307–319. Mackintosh, B., Mathews, A., Yiend, J., Ridgeway, V., & Cook, E. (2006). Induced biases in emotional interpretation influence stress vulnerability and endure despite changes in context. Behavior Therapy, 37(3), 209–222. https://doi.org/10.1016/j.beth.2006.03.001. Mathews, A., & Mackintosh, B. (2000). Induced emotional interpretation bias and anxiety. Journal of Abnormal Psychology, 109(4), 602. Mathews, A., & MacLeod, C. (2005). Cognitive vulnerability to emotional disorders. Annual Review of Clinical Psychology, 1, 167–195. https://doi.org/10.1146/annurev.clinpsy.1.102803.143916. Mathews, A., Richards, A., & Eysenck, M. (1989). Interpretation of homophones related to threat in anxiety states. Journal of Abnormal Psychology, 98(1), 31. Matthews, A., & MacLeod, C. (2002). Induced processing biases have casual effects on anxiety. Cognition and Emotion, 16(3), 331–354. Mestres-Missé, A., Trampel, R., Turner, R., & Kotz, S. A. (2016). In favor of general probability distributions: Lateral prefrontal and insular cortices respond to stimulus inherent, but irrelevant differences. Brain Structure and Function, 221(3), 1781–1786. Mogg, K., Bradbury, K. E., & Bradley, B. P. (2006). Interpretation of ambiguous information in clinical depression. Behaviour Research and Therapy, 44(10), 1411–1419.

Resolving ambiguity: Negative interpretation biases  Chapter | 6  137 Morrison, A. S., & Heimberg, R. G. (2013). Social anxiety and social anxiety disorder. Annual Review of Clinical Psychology, 9, 249–274. https://doi.org/10.1146/annurev-clinpsy-050212-185631. Moser, J. S., Hajcak, G., Huppert, J. D., Foa, E. B., & Simons, R. F. (2008). Interpretation bias in social anxiety as detected by event-related brain potentials. Emotion, 8(5), 693. Moser, J. S., Huppert, J. D., Foa, E. B., & Simons, R. F. (2012). Interpretation of ambiguous social scenarios in social phobia and depression: Evidence from event-related brain potentials. Biological Psychology, 89(2), 387–397. https://doi.org/10.1016/j.biopsycho.2011.12.001. Moss-Morris, R., & Petrie, K. J. (2003). Experimental evidence for interpretive but not attention biases towards somatic information in patients with chronic fatigue syndrome. British Journal of Health Psychology, 8(2), 195–208. Motzkin, J. C., Philippi, C. L., Wolf, R. C., Baskaya, M. K., & Koenigs, M. (2015). Ventromedial prefrontal cortex is critical for the regulation of amygdala activity in humans. Biological Psychiatry, 77(3), 276–284. https://doi.org/10.1016/j.biopsych.2014.02.014. Nolen-Hoeksema, S., Wisco, B. E., & Lyubomirsky, S. (2008). Rethinking rumination. Perspectives on Psychological Science, 3, 400–424. Okon-Singer, H. (2018). The role of attention bias to threat in anxiety: Mechanisms, modulators and open questions. Current Opinion in Behavioral Sciences, 19, 26–30. Phillips, M. L., Drevets, W. C., Rauch, S. L., & Lane, R. (2003a). Neurobiology of emotion perception I: The neural basis of normal emotion perception. Biological Psychiatry, 54(5), 504–514. Phillips, M. L., Drevets, W. C., Rauch, S. L., & Lane, R. (2003b). Neurobiology of emotion perception II: Implications for major psychiatric disorders. Biological Psychiatry, 54(5), 515–528. Pincus, T., Pearce, S., McClelland, A., Farley, S., & Vogel, S. (1994). Interpretation bias in responses to ambiguous cues in pain patients. Journal of Psychosomatic Research, 38(4), 347–353. https://doi.org/10.1016/0022-3999(94)90039-6. Richards, A., & French, C. C. (1992). Anxiety-related bias in semantic activation when processing threat/neutral homographs. The Quarterly Journal of Experimental Psychology, 45A(3), 503–525. Robinson, O. J., Charney, D. R., Overstreet, C., Vytal, K., & Grillon, C. (2012). The adaptive threat bias in anxiety: Amygdala–dorsomedial prefrontal cortex coupling and aversive amplification. Neuroimage, 60(1), 523–529. https://doi.org/10.1016/j.neuroimage.2011.11.096. Robinson, O. J., Overstreet, C., Allen, P. S., Letkiewicz, A., Vytal, K., Pine, D. S., et al. (2013). The role of serotonin in the neurocircuitry of negative affective bias: Serotonergic modulation of the dorsal medial prefrontal-amygdala ‘aversive amplification’ circuit. NeuroImage, 78, 217–223. https://doi.org/10.1016/j.neuroimage.2013.03.075. Rodd, J. M., Johnsrude, I. S., & Davis, M. H. (2012). Dissociating frontotemporal contributions to semantic ambiguity resolution in spoken sentences. Cerebral Cortex (New York, NY: 1991), 22(8), 1761–1773. https://doi.org/10.1093/cercor/bhr252. Rude, S. S., Wenzlaff, R. M., Gibbs, B., Vane, J., & Whitney, T. (2002). Negative processing biases predict subsequent depressive symptoms. Cognition & Emotion, 16(3), 423–440. Schoth, D. E., & Liossi, C. (2017). A systematic review of experimental paradigms for exploring biased interpretation of ambiguous information with emotional and neutral associations. Frontiers in Psychology, 8, https://doi.org/10.3389/fpsyg.2017.00171. Stopa, L., & Clark, D. M. (2000). Social phobia and interpretation of social events. Behaviour Research and Therapy, 38(3), 273–283. Tran, T. B., Siemer, M., & Joormann, J. (2011). Implicit interpretation biases affect emotional vulnerability: A training study. Cognition and Emotion, 25(3), 546–558. https://doi.org/10.1080/ 02699931.2010.532393.

138  Cognitive biases in health and psychiatric disorders Vancleef, L. M., Peters, M. L., & De Jong, P. J. (2009). Interpreting ambiguous health and bodily threat: Are individual differences in pain-related vulnerability constructs associated with an online negative interpretation bias? Journal of Behavior Therapy and Experimental Psychiatry, 40(1), 59–69. Vrana, S. R., & Lang, P. J. (1990). Fear imagery and the startle-probe reflex. Journal of Abnormal Psychology, 99(2), 189. Wenzlaff, R. M., & Bates, D. E. (1998). Unmasking a cognitive vulnerability to depression: How lapses in mental control reveal depressive thinking. Journal of Personality and Social Psychology, 75(6), 1559–1571. https://doi.org/10.1037/0022-3514.75.6.1559. Wilson, E. J., MacLeod, C., Mathews, A., & Rutherford, E. M. (2006). The causal role of interpretive bias in anxiety reactivity. Journal of Abnormal Psychology, 115(1), 103–111. https://doi. org/10.1037/0021-843X.115.1.103. Winton, E. C., Clark, D. M., & Edelmann, R. J. (1995). Social anxiety, fear of negative evaluation and detection of negative emotion in others. Behaviour Research and Therapy, 33, 193–196. Woud, M. L., Zhang, X. C., Becker, E. S., McNally, R. J., & Margraf, J. (2014). Don’t panic: Interpretation bias is predictive of new onsets of panic disorder. Journal of Anxiety Disorders, 28(1), 83–87. https://doi.org/10.1016/j.janxdis.2013.11.008. Yang, J. W., Yoon, K. L., Chong, S. C., & Oh, K. J. (2013). Accurate but pathological: Social anxiety and ensemble coding of emotion. Cognitive Therapy and Research, 37(3), 572–578. https://doi. org/10.1007/s10608-012-9500-5. Yoon, K. L., Yang, J.-W., Chong, S. C., & Oh, K. J. (2014). Perceptual sensitivity and response bias in social anxiety: An application of signal detection theory. Cognitive Therapy and Research, 38(5), 551–558. https://doi.org/10.1007/s10608-014-9619-7. Yoon, K. L., & Zinbarg, R. E. (2007). Threat is in the eye of the beholder: Social anxiety and the interpretation of neutral facial expressions. Behaviour Research and Therapy, 45, 839–847. Yoon, K. L., & Zinbarg, R. E. (2008). Interpreting neutral faces as threatening: A default mode for social anxiety. Journal of Abnormal Psychology, 117, 680–685. Zempleni, M.-Z., Renken, R., Hoeks, J. C. J., Hoogduin, J. M., & Stowe, L. A. (2007). Semantic ambiguity processing in sentence context: Evidence from event-related fMRI. NeuroImage, 34(3), 1270–1279. https://doi.org/10.1016/j.neuroimage.2006.09.048.

Chapter 7

A “rosy view” of the past: Positive memory biases Orly Adler, Ainat Pansky Department of Psychology and The Institute of Information Processing and Decision Making, University of Haifa, Haifa, Israel

It is well established that the human memory does not work like a video recording, but is rather a dynamic process of continuous constructions and reconstructions, resulting in recollections that are never the exact reproductions of the initial experiences (Lane, Ryan, Nadel, & Greenberg, 2015; Loftus, 2003; Radvansky, 2017; Schacter, 2002; Schacter, Guerin, & St Jacques, 2011). With regard to memories for past emotions—whether due to fast attenuation (Levine, Schmidt, Kang, & Tinti, 2012) or because feelings cannot be stored as memory traces to begin with (Robinson & Clore, 2002)—they too need to be reconstructed (Kaplan, Levine, Lench, & Safer, 2016; see Levine & Pizarro, 2004; Levine & Safer, 2002; Levine, Safer, & Lench, 2006). In the current chapter, we review the theoretical mechanisms proposed to underlie emotional memory biases, focusing on those inclined toward recollecting positive information. We shall describe the purpose they serve, common methodologies by which they are examined, the mechanisms that underlie their occurrence, the manner by which they are reflected in neural activations, and their manifestations in clinical populations. A plethora of research has confirmed the existence of positive memory biases, by which people tend to regard their past more favorably than it actually was. For example, college students remembered having more A’s on their high school transcripts than they actually had (Bahrick, Hall, & Berger, 1996; Bahrick, Hall, & Da costa, 2008), and individuals remembered their medicaltest results as better than they really were (Croyle et al., 2006; see Christensen, Wood, & Barrett, 2003). Moreover, pleasant life events were found to be better recalled (Mather, 2006; Skowronski, Betz, Thompson, & Shannon, 1991; Walker, Skowronski, & Thompson, 2003; see Thompson, Skowronski, Larsen, & Betz, 2013), to come to mind more readily (Levine & Bluck, 2004; Master, Lishman, & Smith, 1983), and to be subjectively judged as more clearly remembered than unpleasant life events (Matlin & Stang, 1978; Walker, Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00007-6 © 2020 Elsevier Inc. All rights reserved.

139

140  Cognitive biases in health and psychiatric disorders

Vogl, & Thompson, 1997; see Thompson et al., 2013). With respect to memory for emotions, the same trend is apparent in the form of the fading affect bias (Skowronski, Walker, Henderson, & Bond, 2014; Walker et al., 1997; Walker, Skowronski, & Thompson, 2003)—a well-documented phenomenon according to which affect associated with unpleasant past events fades in memory faster than affect associated with pleasant past events. Thus, comparisons between participants’ reports at event occurrence and upon event recall revealed a larger emotional intensity drop for the unpleasant than for the pleasant events (Landau & Gunter, 2009; Ritchie et al., 2015; Ritchie, Skowronski, Hartnett, Wells, & Walker, 2009; Walker et al., 1997). Certainly, it is possible that one of the contributors to the advantage of positive over negative recollections is merely their larger base rate in people’s memory, with positive experiences occurring about twice as frequently as negative experiences (Walker, Skowronski, & Thompson, 2003). However, as we discuss extensively in the succeeding text, profound and sophisticated mechanisms are more likely to underlie people’s tendency to view their past through rosy glasses. Although the comprehension that memory is prone to errors may be disturbing, positive memory biases actually reflect adaptive processes that operate in the service of maintaining well-being (Nørby, 2015; Schacter et al., 2011). Thus, by retaining the intensity of positive emotions for a longer time, the fading affect bias preserves the pleasantness and minimizes the unpleasantness of life events, thereby functioning as a healthy mechanism (Ritchie, Batteson, et al., 2015; Walker, Skowronski, & Thompson, 2003). Re-experiencing past events in a manner that reinforces a positive sense of self enhances positive self-identity, which is also crucial for maintaining well-being (Conway, 2005; Greenwald, 1980; Holland & Kensinger, 2010; Loftus, 1982; see also Ross, 1989; Ross & Wilson, 2003; Schacter et al., 2011; Wilson & Ross, 2003). Indeed, older adults, who commonly report better moods than young adults, show enhanced memory for positive as compared with negative past experiences (Mather & Carstensen, 2005; see Reed, Chan, & Mikels, 2014) and are more likely to falsely recognize positive items as having been seen before (Fernandes, Ross, Wiegand, & Schryer, 2008; Piguet, Connally, Krendl, Huot, & Corkin, 2008). Among young adults, there is a correlation between the probability of wrongly recalling positive life events and life satisfaction (Koo & Oishi, 2009). Taken together, not surprisingly, the growing field of positive psychology, which focuses on the positive subjective experience (Lyubomirsky, Sheldon, & Schkade, 2005; Seligman & Csikszentmihalyi, 2000), views positive memory retrieval and positive self-reflection among the interventions aimed at bringing about enhanced well-being (Burton & King, 2004; Seligman, Steen, Park, & Peterson, 2005; see Duckworth, Steen, & Seligman, 2005). Despite the burgeoning literature with regard to positive memory biases, a considerable amount of studies has shown the opposite pattern of superior memory for negative information (Charles, Mather, & Carstensen, 2003). In their review, Baumeister, Bratslavsky, Finkenauer, and Vohs (2001) concluded that,

A “rosy view” of the past: Positive memory biases  Chapter | 7  141

because negative materials evoke more conscious activity and receive deeper operations than positive materials, negative memories are better recalled than positive ones. One explanation for this apparent memory discrepancy could be that emotional valence has different impacts on memory over time. More specifically, as negative stimuli hold survival-relevant information, when facing a negative event, people immediately activate coping processes for dealing with this information, which in turn enhance the memory for the event (Pratto & John, 1991). Yet, to maintain well-being, people activate opponent processes that minimize the impact of the event (the mobilization-­minimization hypothesis: Taylor, 1991; see also Toyama, Katsuhara, Sakurai, & Ohira, 2014; Walker, Skowronski, & Thompson, 2003). Thus, due to adaptive functions, there exists an immediate superiority for negative memories that diminishes with time. An alternative account for the contradictory findings is that any apparent advantage for positive memories is bogus. Given that the assessment of an autobiographical memory is commonly performed by the rememberer himself and not by an extrinsic observer, such a judgment might be prone to personal motivations (for the potential underlying mechanisms, see section in the succeeding text on Major Theories in the Field). In other words, people may believe they remember positive events better than they really do (Levine & Bluck, 2004). In any case, both possible explanations imply that Baumeister et al.’s (2001) suggestion of negative memory superiority may be appropriate for experimental studies, which mostly use objective tools and short retention intervals between encoding and retrieval of the stimuli, but is unwarranted in the domain of autobiographical memory, which is relatively long term and subjectively measured. Indeed, memory bias for positive information is mostly demonstrated in the domain of autobiographical memory research (Kensinger & Schacter, 2008). Yet, because autobiographical memory is unique, exceptional, and often unverifiable, the conditions driving distortions of autobiographical content are difficult to manipulate or control using experimental methodologies (Bahrick et  al., 2008). With the understanding that examining the implications of extreme emotional experiences in the laboratory setting is quite challenging, we turn next to review the main methods used to study autobiographical memory biases.

Methods used to investigate the bias One of the more prevalent modes to elicit autobiographical memories in the laboratory is the word-cued memory technique designed by Galton (1879), in which the participants are presented with cue words (e.g., car and love), one at a time, and are asked to report a specific personal life event in response to each word (Crovitz & Schiffman, 1974; Galton, 1879; see Rubin, 2002; Rubin & Wenzel, 2004). Variations of the paradigm employ criteria (e.g., positive vs. negative; Berntsen, 2002; Bohanek, Fivush, & Walker, 2005; Ritchie, Batteson, et  al., 2015), or lifetime periods (e.g., the first week of classes in college;

142  Cognitive biases in health and psychiatric disorders

Holland, Tamir, & Kensinger, 2010; Ritchie, Batteson, et al., 2015) as retrieval cues. This common technique permits gauging both objective measures (e.g., retrieval latencies of negative vs. positive events and recent vs. remote episodes) and subjective measures (e.g., ratings of emotions, confidence in accuracy, and vividness) for autobiographical memories (Brewer, 1996). One obvious drawback of this technique is that the researcher has no access to what had actually occurred in the past, rendering the accuracy assessment and age determination of the reported memories impossible. These drawbacks are overcome in a different method, the diary paradigm, in which the participants are asked to record everyday events and to rate their affiliated feelings as they are being experienced. At a later time, the participants are asked to recount their memory for some of these events and feelings (Barclay & Wellman, 1986; Burt, Kemp, & Conway, 2003; Kemp, Burt, & Furneaux, 2008; Mill, Realo, & Allik, 2016; Stone & Broderick, 2007; Todd, Tennen, Carney, Armeli, & Affleck, 2004; Wirtz, Kruger, Scollon, & Diener, 2003). The comparison between the online reports and the retrospective accounts constitutes a measure of reconstruction and has been conceptualized as the memory-­experience gap (Miron-Shatz, Stone, & Kahneman, 2009)—the discrepancy between emotions as they are experienced and emotions as they are remembered. Another popular technique for examining the relationship between memory and emotion is mood induction (Blaney, 1986; Bower, 1981; Singer & Salovey, 1988). After exposing participants to positive, negative, or neutral stimuli like music (Miranda & Kihlstrom, 2005; Västfjäll, 2001), pictures (Buchanan, 2007), or videos (Fitzgerald et al., 2011), the participants are asked to retrieve either autobiographical information or items studied in an earlier phase of the experiment. Comparing the retrieval latencies or the amount of remembered events in the different affective states may shed light on the impact of emotion on memory. In other methodologies that are in extensive use in the domain of autobiographical memory, participants are asked to recall unique, emotional, public events (flashbulb memories; Brown & Kulik, 1977) or to perform guided interviews about events from the participants’ lives (Levine, Svoboda, Hay, Winocur, & Moscovitch, 2002; Ritchie, Batteson, et al., 2015). Also, recent years have witnessed an upsurge of studies combining neuroimaging technologies with behavioral methodologies such as those mentioned earlier, to elucidate the neural networks engaged during remembering (Bonnici & Maguire, 2018; Cabeza & St Jacques, 2007; Josselyn, Köhler, & Frankland, 2015; Moscovitch, Cabeza, Winocur, & Nadel, 2016). Acknowledging the integral methodological obstacles in autobiographical memory research, we next present major theories in the field and the mechanisms they propose to underlie positive memory biases. It is important to note that biases may potentially occur at either the encoding, the storage, or the retrieval stage of the memory. In many studies, the precise stage in which the bias occurs is not specified. Yet, in the cases in which it is, we will point it out.

A “rosy view” of the past: Positive memory biases  Chapter | 7  143

Major theories in the field Since long ago, theoreticians have addressed people’s inclination to view their pasts overfavorably. For example, in his theoretical framework of repression, Freud proposed a set of defense mechanisms, postulated to inhibit the recollection of self-threatening events, sometimes to the extent of becoming inaccessible, as if forgotten (Freud, 1957). In contemporary psychological theories, the notion of the “self” as a cornerstone of positive biases in memory is also salient. In fact, to maintain well-being, preserving a positive view of the self is considered to be fundamental (D’Argembeau & Van der Linden, 2008; Davidson, 2004; Rathbone, Holmes, Murphy, & Ellis, 2015; Sedikides & Alicke, 2012; Sedikides, Gaertner, & Cai, 2015). Indeed, people mostly possess positive schemas of themselves, portraying themselves as worthwhile, competent, warm, moral, attractive, and loveable (Alicke & Sedikides, 2009; Sedikides & Gregg, 2008). Thus, various techniques may be employed in the name of self-­ protection, self-enhancement, and self-consistency, each of which may result in a positive memory bias (Alicke & Sedikides, 2009; Libby & Eibach, 2007; Ritchie, Sedikides, & Skowronski, 2017; Skowronski, 2011). Next, we shall review these mechanisms and techniques, as well as the variables that may moderate their operation (see Fig. 1).

Self-protecting and self-enhancing mechanisms Before reviewing the manners by which the self-protective and self-enhancing mechanisms are employed, we wish to assert that the emergence of the positivity bias does not only depend on the event’s valence but also depend on the extent to which the event is relevant to oneself. When the information is not selfrelevant, the bias will not be apparent (Skowronski, 2011; Skowronski et al., 1991; see Holland & Kensinger, 2010). Avoiding negative information

-

Assimilating positive information

Distancing negative memories Consistency with current knowledge

-

Wellbeing

Emotion-regulation

Consistency with current emotions

-

Social disclosure

Consistency with expectations

FIG. 1  Moderated by individual differences, the following techniques (in italics) are employed in the name of self-protecting, self-enhancing, and self-consistency mechanisms that underlie positive memory biases in support of well-being.

144  Cognitive biases in health and psychiatric disorders

Avoiding negative information and assimilating positive information As mentioned earlier, people commonly possess favorable schemas of themselves (Alicke & Sedikides, 2009; Sedikides & Gregg, 2008). To preserve or even enhance positive self-evaluations, different processing modes may be employed when encoding negative or positive information. More specifically, it has been suggested that negative information often leads to attention narrowing to enable item-specific processing (Schwarz, 1990; Wegner & Vallacher, 1986). By contrast, positive information was proposed to elicit a gist-processing mode in reference to activated stored schemas (Clore et  al., 2001; see also the assimilation approach; Fiedler, 2001). Because, as stated, self-schemas are mostly positive, pleasant events are more likely to be readily assimilated into one’s collection of personal experiences that form one’s autobiographical memory (Holland & Kensinger, 2010; Matlin & Stang, 1978). This integration with existing knowledge provides the positive information with more elaboration, more retrieval routes, and superior recall (the mnemonic neglect model; Green, Pinter, & Sedikides, 2005; Pinter, Green, Sedikides, & Gregg, 2011). To illustrate, following short and long presentation durations of negative and positive feedbacks, not only did participants demonstrate poorer memory for the negative feedbacks compared with the positive ones, but their memory performance seemed to depend on a valence-duration interaction. Specifically, whereas twice as many positive feedbacks were recollected in the long-duration condition than in the short-duration condition, memory for the negative feedbacks was not affected by presentation duration. Most importantly, a recall advantage of positive feedbacks was evident only in the long-duration condition, whereas in the short-duration condition better recall was evident for the negative feedbacks (Sedikides & Green, 2000, 2009). The authors interpreted these results as suggesting that negative information entails a superficial mode of processing that requires minimal resources (e.g., time).a These findings have also been taken to indicate that self-protecting mechanisms operate already at the stage of memory encoding, as was evident in the effect of presentation duration. Distancing negative memories Another technique that helps to maintain a positive sense of self is to remember or construe the past in a self-flattering manner (Alicke & Govorun, 2005; Sedikides & Gregg, 2003, 2008). Thus, people tend to perceive flattering past events as recent and to dissociate themselves from embarrassing events by perceiving them as having occurred further in the past (Ross, Heine, Wilson, & Sugimori, 2005; Ross & Wilson, 2002). In their study, Ross and Wilson (2002) had students report the past-semester course in which they had received either a. Notably, this interpretation is inconsistent with Baumeister et al.’s (2001) view that deep operations are engaged in the processing of negative stimuli (mentioned earlier).

A “rosy view” of the past: Positive memory biases  Chapter | 7  145

their best grade or their worst grade and to rate their subjective distance from that course (i.e., “feels like yesterday” and “feels far away”). The results showed that the students felt farther from courses in which they obtained lower grades, even though the actual retention interval in the two conditions was c­ omparable. Thus, by distancing a failure event, one can render it less relevant to one’s ­current self.

Emotion regulation strategies To counteract negative feelings when facing unpleasant experiences, people may execute different emotion regulation strategies. They may choose to suppress their emotional expressions when encoding the event, to suppress their thoughts when recollecting it, to deliberately retrieve positive memories, to reappraise an event to change its meaning, and so forth (see Gross, 1998). The effectiveness of each emotion regulation strategy on the emotional outcome may vary (see Webb, Miles, & Sheeran, 2012), and so does their potential for causing positive memory biases. Thus, whereas retrieving memories (as in positive reminiscence) strengthens their traces and thereby enhances them (Roediger & Karpicke, 2006), expressive suppression was found to impair the memory for the event and thereby reduce its recalled negativity (Dillon, Ritchey, Johnson, & LaBar, 2007; Richards & Gross, 1999). Engaging in reappraisal may entail a positive memory bias as it involves the construction of the unpleasant event in less negative terms. To illustrate, Levine et al. (2012) compared the emotional reports of students before and after taking the final-year exam. Their findings revealed that, compared with other emotion regulation strategies, students who engaged in reappraisal to cope with the stressful preexam period demonstrated a positive memory bias in later remembering (i.e., having felt more positive and less negative emotions than they had reported originally). Social disclosure People claim to share their memories with others for many reasons—to maintain the memory of the event, to re-experience its associated emotion, to better understand it, or simply for the purpose of social communication (Walker, Skowronski, Gibbons, Vogl, & Ritchie, 2009). Whatever their conscious rationale may be, the self-protective-enhancive mechanisms that are constantly at work are likely to yield a positivity bias. To illustrate, unpleasant memories tend to be recounted less or with minimized negative elements (Mather, 2006; Skowronski et al., 1991; Skowronski & Walker, 2004). As rehearsed information is strengthened whereas omitted information is left nonrehearsed and consequently less memorable (Roediger & Karpicke, 2006), scarcely recounting the negative information may lead to the loss of the negative content and to the reinforcement of the positive content. Indeed, retrieving a past experience with a certain perspective was found to create a particular schema that guides the subsequent memory of the experience to be consistent with that schema (Alba & Hasher, 1983; Anderson & Pichert, 1978; Tversky & Marsh, 2000).

146  Cognitive biases in health and psychiatric disorders

Furthermore, even when unpleasant experiences are recounted, this generally entails expressions of support and new perceptions, whereas recounting pleasant experiences commonly entails expressions of joy and praise. Hence, social disclosure is most likely to alter the rememberer’s affective response to the recalled event (Skowronski & Walker, 2004) and to reinforce the fading affect bias (mentioned earlier). Indeed, stronger fading affect bias effects were found for frequently disclosed compared with infrequently disclosed experiences (Skowronski, Gibbons, Vogl, & Walker, 2004).

Self-consistency Apart from preserving a positive sense of self, owning a coherent and a stable record of the self over time is also vital for the maintenance of well-being (Conway, 2005; Greenwald, 1980). Thus, when changes in knowledge or feelings are marked, people tend to modify their retrieved memories to bring them to better accord with the present (Bahrick et al., 2008; Holland & Kensinger, 2010; Levine, 1997; Levine, Lench, & Safer, 2009; Levine & Safer, 2002; Ross, 1989; Ross, Blatz, & Schryer, 2008; Toglia, Read, Ross, & Lindsay, 2017; Wilson & Ross, 2003).

Consistency with current knowledge As people turn to their current knowledge to infer what they had thought and felt in the past, acquiring postevent information may lead to a distortion of the memory for the original event. For example, Safer, Levine, and Drapalski (2002) compared between students’ affective reports before and after taking an exam. They found that students who learned that they had done well on the exam before the retrospective report underestimated how anxious they had felt prior to the exam, compared with those who recalled their preexam emotions before learning their grade (Safer et al., 2002). Consistency with current emotions Changes in one’s feelings toward something or someone else may also produce memory distortion. For example, McFarland and Ross (1987) found that participants who became more favorable of their dating partners over time recalled having evaluated them a month earlier more positively than they actually did, compared with participants who became less favorable of their partners. Importantly, drawing on current feelings to reconstruct memories of past emotions may also lead to the mood-congruent memory (MCM) effect—the well-replicated finding that one’s memory is biased to become congruent in valence with one’s current mood (Blaney, 1986; Bower, 1981; Singer & Salovey, 1988). Thus, following a happy (vs sad) mood induction, more positive memories were recalled (Eich, Macaulay, & Ryan, 1994; Natale & Hantas, 1982; Teasdale, Taylor, & Fogarty, 1980), and the latency to retrieve them was shorter (Lloyd & Lishman, 1975; Riskind, 1983),

A “rosy view” of the past: Positive memory biases  Chapter | 7  147

compared with negative memories. In addition, a happy (vs sad) mood induction was found to influence the affect associated with autobiographical memories, such that the memories were rated as more positive (Madigan & Bollenbach, 1982; Snyder & White, 1982). One possibility is that the MCM effect results from spreading activation through a network of emotion nodes that are linked to related memories (the associative network theory; Bower, 1981; see also Blaney, 1986; Forgas, 1995; Rusting, 1998; Singer & Salovey, 1988). Alternatively, the MCM could occur due to the activation of self-schemas that facilitates the retrieval of memories that are congruent with that schema (schema models; reviewed by Rusting, 1998). Regardless of its origins, MCM is highly prevalent, with stronger effects of positive affect on mood-congruent retrieval than of negative affect (Brewin, Andrews, & Gotlib, 1993; Rusting, 1998; Singer & Salovey, 1988; see Matt, Vázquez, & Campbell, 1992). This bias asymmetry may be explained by the predisposition of people to regulate negative mood. Reappraising the situation or retrieving pleasant thoughts and memories with the purpose of reducing the sad affect that is experienced may result in mood incongruency, with superior memory for positive information despite the negative mood induction (Holland & Kensinger, 2010; Isen, 1985, 1987).

Consistency with expectations Another factor that may bias people’s evaluations of their past is their anticipation prior to the event occurrence (Holland & Kensinger, 2010; Levine & Safer, 2002). Anderson (1983) has long stated that people’s expectancies of an upcoming event may be based in part on the ease or difficulty of imagining that future event’s possible scenarios. In that context, research has shown that compared with negative or neutral simulations, positive simulations of future events are more easily constructed, contain a larger amount of details, and tend to be better recalled over time (D’Argembeau, Renaud, & Van der Linden, 2011; D’Argembeau & Van der Linden, 2004; de Vito, Neroni, Gamboz, Della Sala, & Brandimonte, 2015; Sharot, Riccardi, Raio, & Phelps, 2007; Szpunar, Addis, & Schacter, 2012; see also Dricu, Kress, & Aue, 2020, Chapter 3). As imagining an event in detail creates memory traces for that possible scenario and these traces are reactivated and assimilated into the memory of that event after its actual occurrence, obtaining a positive expectancy frequently results in a rosy memory (Devitt & Schacter, 2018). To illustrate, in Devitt and Schacter’s (2018) study, participants read several neutral plausible events, not before simulating them in either a negative or a positive manner. A subsequent memory test revealed that negative simulations did not impact the memory for the narratives, whereas positive simulations produced a memory bias such that both the narratives themselves (which were neutral) and their affiliated affect were remembered as positive. One may suggest that the privileged position of positive expectations influences the subjective experience of the individual during the event occurrence (e.g., Rasmussen & Berntsen, 2013; Sharot, 2011), yet research has shown that, when one expects a positive experience, the ­recollection

148  Cognitive biases in health and psychiatric disorders

of the event may be rosy even when its moment-to-moment report indicates a less favorable experience. For example, in a study that surveyed cyclists before, during, and after a 3-week tour, although their online reports mentioned heavy rain, dull companions, and physical exhaustion, their memories of the tour were generally positive, very much like their prior expectations (Mitchell, Thompson, Peterson, & Cronk, 1997). This phenomenon, of prospects bringing about a more favorable retrospective outlook on the event than when the event occurred, termed by Mitchell and his collogues (1997) the “rosy view,” has been observed in many other activities, like marathons, bicycle tours, and vacations (Mitchell et al., 1997; Sutton, 1992; Wirtz et al., 2003).

Variables that may moderate the bias—Individual differences When reviewing the mechanisms that distort recall in a positive direction, it is important to note the ample evidence indicating that the mechanisms mentioned earlier do not necessarily operate independently, but, rather, may interact with another factor—individual differences (Levine & Safer, 2002; see also Holland & Kensinger, 2010).

Cultural differences To maintain well-being, a coherent and consistent sense of self must also comprise coherency with one’s social group (Wang, 2016). With the need for a shared sense of reality, people accentuate aspects of the self that typically correspond to norms and values of the group they belong to or wish to belong to (Baumeister & Leary, 2017; Leary & Kowalski, 1990; Siibak, 2009). Memories are thus distorted in accordance with cultural values (Oishi, 2002; Wang, 2016). For instance, consistent with the Western individualistic value of personal happiness (Markus & Kitayama, 1994, 2010; Oishi, 2002), European-American participants remembered experiencing more pleasant than unpleasant emotions in their daily lives, whereas Japanese and Asian-American participants, whose cultures do not idolize individualism, remembered experiencing an equal number of positive and negative emotions (Markus & Kitayama, 1994; Oishi, 2002). Another study, which compared the difference between online and retrospective reports, showed that, although European-American and Asian-American participants did not differ in their daily diary documentation of emotions, EuropeanAmericans retrospectively recalled greater satisfaction than Asian-Americans (Oishi et al., 2007). In addition, and in line with Asian’s motivation to promote social harmony and maintain positive views of important others, Asian participants were found to remember interpersonal harms as less severe compared with American participants (Song & Wang, 2014). Age-related differences Another individual difference that plays a prominent role in this context is age. According to the socioemotional selectivity theory (Carstensen,

A “rosy view” of the past: Positive memory biases  Chapter | 7  149

Isaacowitz, & Charles, 1999), as people get older, they commence to view time as limited and hence prioritize their emotional well-being. Thus, in comparison with young adults, older adults tend to process information in a manner that provides them with more emotional fulfillment, displaying superior abilities in regulating their emotions than young adults (Carstensen, Pasupathi, Mayr, & Nesselroade, 2000; Holland & Kensinger, 2010; Mather & Carstensen, 2005). Indeed, older adults were found to commonly use positive reappraisals as a coping strategy (Folkman, Lazarus, Pimley, & Novacek, 1987) and demonstrate an attention bias by which they are inclined to attend to positive over negative information (Carstensen & Mikels, 2005; Mather & Carstensen, 2003, 2005; see Reed et al., 2014(. As these emotion-regulation strategies are associated with a reduction in negative emotion and with enhanced recall of positive memories (Levine et  al., 2012; Rusting & DeHart, 2000), it should come as no surprise that research shows a greater predisposition for positive memory biases among older adults than among young adults (Mather, 2006; SamanezLarkin & Carstensen, 2011). Indeed, older adults are inclined to forget unpleasant events, or the negative feelings associated with them (Berntsen & Rubin, 2002; Levine & Bluck, 1997), and to remember the past more favorably than it actually was (Comblain, D’Argembeau, & Van der Linden, 2005; Kennedy, Mather, & Carstensen, 2004; Levine & Bluck, 1997; Wagenaar & Groeneweg, 1990). In more controlled laboratory experiments, older adults (versus young adults) were found to remember a higher proportion of positive stimuli and a lower proportion of negative stimuli (Charles et al., 2003; Mather & Carstensen, 2003; Mather, Knight, & McCaffrey, 2005) and to falsely endorse more positive than negative items related to those they had studied earlier (Fernandes et al., 2008; Piguet et al., 2008). As a matter of fact, older adults’ predisposition to reconstruct their memory positively may thus account for the well-documented association between age and increased subjective well-being (Carstensen & Mikels, 2005; Charles & Carstensen, 2010; Reed et al., 2014).

Personality differences Individual differences in personality traits may also moderate the reconstructions of memories. To illustrate, participants who scored high on self-esteem measures were found to recall more positive autobiographical memories than lower self-esteem participants (Christensen et al., 2003; see Robinson & Clore, 2002). In MCM studies, participants with high self-esteem demonstrated mood-incongruent recall, retrieving positive memories following a sad induction (Smith & Petty, 1995). Extraversion was also found to be associated with emotional memory biases, as people high in extraversion demonstrated enhanced recall of positive memories (Denkova, Dolcos, & Dolcos, 2012; Mayo, 1983) and increased mood-congruent recall for positive information (Rusting, 1999). In fact, the link between extraversion and the positivity bias is very commonsensical. Extraverts are regularly orientated toward positive experiences (Larsen & Ketelaar, 1991) and are characterized by high levels of ­positive

150  Cognitive biases in health and psychiatric disorders

­affectivity (Barrett, 1997; Morrone, Depue, Scherer, & White, 2000)—a trait that is correlated with short latencies to retrieve positive autobiographical memories (MacLeod, Andersen, & Davies, 1994). To summarize, the positive memory bias is a widespread phenomenon, with ubiquitous documentation throughout the memory literature of positive reconstructions and fading affect biases. Nonetheless, it is important to note that, as the prime goal of the self is to maintain well-being, the tendency to remember the past in an overly positive manner characterizes situations in which the emotional event was successfully coped with and no longer elicits distress. By contrast, when recalling ongoing situations with which the individual is still coping, the memory seems to be biased toward exaggerating past negativity, so as to perceive improvement over time (Levine & Safer, 2002; Ross, 1989; Wilson & Ross, 2003). To illustrate, psychotherapy patients who failed to improve wellbeing exaggerated their pretherapy distress, whereas those who improved the most underestimated their pretherapy distress (Safer & Keuler, 2002). Widows and widowers who were still mourning 5 years after the deaths of their spouses exaggerated their earlier level of distress, whereas those who had succeeded to proceed with their lives underestimated it (Safer, Bonanno, & Field, 2001).

Brain regions involved in the emergence of the bias There is ample evidence that memories are not stored within a single region of the brain, but rather involve a distributed network of activations, mainly in the prefrontal cortex, the amygdala, and sensory cortices within the occipital and temporal lobes, with the hippocampus serving as the system’s “hub” (see Cabeza & St Jacques, 2007; Josselyn et al., 2015; Moscovitch et al., 2016). The amygdala has been long recognized as a key element of the neural basis of emotion (Klüver & Bucy, 1939; see Cardinal, Parkinson, Hall, & Everitt, 2002; Kragel & LaBar, 2016). People with amygdala lesions show deficits in emotional learning (Bechara et al., 1995), in the perception of emotions (e.g., in facial expressions; Adolphs, Tranel, Damasio, & Damasio, 1994; Young et al., 1995), and in the memory for emotional events (Cahill, 2000). Indeed, as one of the most extensively connected subcortical regions of the brain, with links to numerous cortical and subcortical regions (Amaral, 2003; Amaral, Price, Pitkänen, & Carmichael, 1992; LeDoux, 2000), the amygdala has a significant role in all the stages of memory. Not only is the amygdala involved in the encoding (Kensinger & Corkin, 2004; Mickley Steinmetz & Kensinger, 2009; see Dolcos et al., 2017) and consolidation (McGaugh, 2004; McIntyre, McGaugh, & Williams, 2012) of emotional experiences, but also it is a dominant player in the retrieval of emotional memories (Denkova, Dolcos, & Dolcos, 2013b; Dolcos, LaBar, & Cabeza, 2005; see Dolcos et al., 2017). Importantly, damage to the amygdala has been found to selectively impair the memory of unpleasant (but not pleasant) events. As shown by Buchanan, Tranel, and Adolphs (2005), compared with patients who suffered from damage only to the hippocampus,

A “rosy view” of the past: Positive memory biases  Chapter | 7  151

patients whose damage included the hippocampus and the amygdala recalled less negative predamage events. Moreover, these patients described predamage autobiographical memories in fewer negative (but not fewer positive) words and rated their negative (but not their positive) autobiographical memories as less intense and less vivid. The authors suggested that, due to its involvement in the neural circuitry necessary for the vivid recollection of unpleasant emotional events, the amygdala contributes to the emergence of positive memory biases (Buchanan et al., 2005). In addition to the large focus on the amygdala’s involvement in emotional autobiographical memories, intriguing findings have associated the activation of distinct brain regions with the valence of memories (but see Lindquist, Wager, Kober, Bliss-Moreau, & Barrett, 2012). More specifically, circuits involving temporal and posterior regions, like the amygdala, are more activated during the encoding and the recollection of negative events (as mentioned earlier), whereas circuits involving frontal and parietal regions, like the prefrontal cortex, are more activated during the encoding and the recollection of positive events (Denkova, Dolcos, & Dolcos, 2013a; Mickley Steinmetz, Addis, & Kensinger, 2010; Piefke, Weiss, Zilles, Markowitsch, & Fink, 2003; Ritchey, LaBar, & Cabeza, 2011; see Dolcos et  al., 2017). Importantly, the prefrontal cortex is related to executive functions (Friedman & Miyake, 2017; Shimamura, 2000) and is strongly associated with self-referential processing (Delgado et al., 2016; Roy, Shohamy, & Wager, 2012). Thus, enhanced activation of the prefrontal cortex during positive recollection may result from one employing self-referential processing modes to refer positive experiences to one’s concept of self (Holland & Kensinger, 2010; Tsukiura & Cabeza, 2008). This suggestion supports the notion that people tend to reconstruct their autobiographical memory in a manner that enhances their positive sense of self (Conway, 2005; Greenwald, 1980; Ross, 1989). We shall next briefly review brain-activity patterns that were suggested to correspond to some of the self-protection, selfenhancement, and self-consistency techniques that we described in this chapter, resulting in positive memory biases. Converging evidence from neuroimaging research has supported Rusting and DeHart’s (2000) suggestion, according to which the MCM effect (mentioned earlier) arises from the mood experienced during memory retrieval activating similar brain networks that were engaged during the encoding of that memory (Rusting & DeHart, 2000). Specifically, the ease of retrieving a memory consistent with one’s current mood has been associated with neural correlates within the limbic system (Buchanan, 2007; Haas & Canli, 2008). With regard to the retrieval facilitation of positive memories, enhanced activity within the anterior cingulate in response to positive mood was proposed to be involved (Lewis, Critchley, Smith, & Dolan, 2005). Interestingly, Lewis et  al. (2005) further showed that, regardless of retrieval success, when stimulus valence matched the state of mood at retrieval, activity was greater in regions associated with the attempt to retrieve, such as the dorsolateral prefrontal cortex. These findings

152  Cognitive biases in health and psychiatric disorders

support the notion that affect guides the search process for the solicited memory (Buchanan, 2007; Lewis et al., 2005). Dissociable neural activities were also found to reflect different emotion regulation strategies. For example, when encoding a negative stimulus, increased connectivity between the hippocampus and the prefrontal cortex was associated with reappraisal (Hayes et al., 2010), whereas decreased connectivity between the hippocampus and the prefrontal cortex was associated with suppression (Binder et al., 2012). These opposing neural patterns correspond to the memory enhancement and memory impairment commonly manifested following reappraisal and suppression, respectively (Ahn et  al., 2015; Dillon et  al., 2007; Dunn, Billotti, Murphy, & Dalgleish, 2009; Kim & Hamann, 2012; Liu, Cui, & Zhang, 2015). Furthermore, when suppressing a negative memory while it is being retrieved, inhibition of the visual cortex was exhibited, followed by inhibition of the amygdala and the hippocampus by prefrontal cortex regions (Depue, Curran, & Banich, 2007). Deliberate positive reminiscence, on the other hand, was associated with activity in the striatum (Speer, Bhanji, & Delgado, 2014), which is regularly involved in reward processing (Delgado, Jou, LeDoux, & Phelps, 2009; Haber & Knutson, 2010; O’Doherty, 2004; Tamir & Mitchell, 2012). Individual differences, proposed to moderate emotional memory biases, are also reflected in particular neural activations. Thus, age-by-valence interactions during encoding may underlie the positivity effect in older adults’ memory (see Dolcos et al., 2017; Samanez-Larkin & Carstensen, 2011). Specifically, studies have demonstrated that, during the encoding of negative events, older adults display reduced connectivity between the amygdala and the hippocampus and enhanced connectivity between the amygdala and the prefrontal cortex, compared with young adults (Murty et al., 2009; St Jacques, Dolcos, & Cabeza, 2009). During the encoding of positive events, older adults exhibit greater positive modulation of the medial temporal lobe by the PFC and an enhanced connectivity within PFC regions, compared with young adults (Addis, Leclerc, Muscatell, & Kensinger, 2010; Waring, Addis, & Kensinger, 2013). Furthermore, older adults, not young adults, who showed greater connectivity between the amygdala and the prefrontal cortex at rest, also remembered more positive than negative stimuli in a subsequent memory recognition test (Sakaki, Nga, & Mather, 2013). Taken together, these findings corroborate with the posterior-anterior shift in aging model (Davis, Dennis, Daselaar, Fleck, & Cabeza, 2007), which posits among older adults that there is a decline in the involvement of posterior regions compensated by greater engagement of the prefrontal cortex. In addition, these findings converge with the neural activity that characterizes the efficient emotion regulation strategies that older adults were found to employ (discussed earlier). Another individual difference that we mentioned earlier as associated with emotional memory biases is personality. For example, extraverts’ tendency to display positive memory biases (Haas & Canli, 2008) was associated with their heightened amygdala activity during the encoding of positive information

A “rosy view” of the past: Positive memory biases  Chapter | 7  153

(Canli et al., 2001; Canli, Sivers, Whitfield, Gotlib, & Gabrieli, 2002; Dolcos et al., 2017). Furthermore, high scores in extraversion are correlated with low levels of epinephrine acting upon the amygdala (Miller, Cohen, Rabin, Skoner, & Doyle, 1999) and with high levels of cortisol produced by the adrenal gland (LeBlanc & Ducharme, 2005). As stress hormones are thought to impact the consolidation of emotional memories (e.g., Bryant, McGrath, & Felmingham, 2013), the rosy memory extraverts usually pertain could also arise from their characteristic stress-induced levels of hormones (Haas & Canli, 2008). To summarize, numerous neuroimaging studies that have investigated interactions between cognition and affect have provided evidence for brain circuits underlying emotional memory biases, pointing to enhanced activity in the prefrontal cortex regions and reduced activity in the amygdala regions during the encoding and the retrieval of negative information as possible contributors to the formation of positive biases of memory.

Similarities and differences between healthy and clinical populations As discussed extensively earlier, memory biases for positive information are prevalent in the general population. People distort their memories to form a positive view of the past, with the goals of enhancing self-evaluation and maintaining well-being. Indeed, positive correlations were found between the magnitude of the fading affect bias and the extent to which the self is more positive, stable, and secure (Ritchie, Sedikides, & Skowronski, 2016; Ritchie, Skowronski, Cadogan, & Sedikides, 2014). It should thus come as no surprise that, with regard to clinical and subclinical populations, memory is commonly biased in the opposite direction. For example, in dysphoria, memory is commonly biased toward negative information (Gotlib et al., 2004; Koster, De Raedt, Leyman, & De Lissnyder, 2010; Matt et al., 1992; Walker, Skowronski, Gibbons, Vogl, & Thompson, 2003). Negative memory biases are considered a hallmark feature in cognitive models of depression and posttraumatic stress disorder (LeMoult & Gotlib, 2019; Lin, Hofmann, Qian, & Li, 2015; Nolen-Hoeksema, 2000; Rubin, Berntsen, & Bohni, 2008; Schacter, 2002; Williams & Moulds, 2007; see Denkova et  al., 2012). Although the findings with regard to anxiety are ambiguous and inconsistent (Mathews & MacLeod, 1994; Zlomuzica et al., 2014), increased levels of anxiety are associated with reduced affective fading for both positive and negative events (Walker, Yancu, & Skowronski, 2014). Even narcissists, although characterized by self-enhancing tendencies thought to spur positive memory reconstructions, were found to be prone to a disruption of the fading affect bias, manifesting the fading of positive affect instead of negative affect (Ritchie, Walker, Marsh, Hart, & Skowronski, 2015). Finally, with regard to the manic disorder, one would expect this illness to be associated with a positively biased memory. After all, according to the DSM-5 (American Psychiatric Association, 2013), mania is characterized by states of abnormal

154  Cognitive biases in health and psychiatric disorders

elevated moods. Indeed, previous studies have shown positive attention biases among bipolar patients in manic episodes (Elliott et al., 2004; Murphy et al., 1999). Yet, in the few studies that have examined positive memories among bipolar participants, no positivity bias was apparent in either the vividness of the memories or their affective impact, compared with healthy participants (Gruber, 2011; Gruber, Harvey, & Johnson, 2009). It is important to note, though, that the bipolar participants in these studies were in a euthymic state (i.e., neither in a manic, a depressed, nor a mixed-mood state). Indeed, difficulties in examining patients during a manic state could be the reason for the dearth in empirical investigations in this area (Henry, Weingartner, & Murphy, 1971). An exceptional disorder in this context is pathological gambling, of which cognitive biases are considered one of the foundations, with gamblers regularly manifesting superior recollection of their winning experiences than of their losing experiences (Ciccarelli, Griffiths, Nigro, & Cosenza, 2017; Goodie & Fortune, 2013; Wagenaar, 1988; see Fortune & Goodie, 2012). This may sound equivalent to people’s general tendency of remembering more pleasant than unpleasant events (Walker, Skowronski, & Thompson, 2003), but it seems that, in this case, additional processes contribute to an enhancement of the positive memory bias. It has been suggested that gamblers are inclined to store their gambling memories as a string of wins and losses (Rachlin, 1990), which they tend to reassess after every win. Appearing as a vague series of losses followed by a vivid win and in correspondence with the peak-and-end effect, by which the most intense or the final moments of an experience disproportionately bias recall (Fredrickson, 2000), the string may be misinterpreted as a representation of success. In addition to this memory storage deficit, pathological gamblers tend to apply common heuristics to a greater extent than the normal population. To illustrate, basing win probabilities on the ease of retrieving winning events from memory (the availability heuristic; Tversky & Kahneman, 1973) not only may be inappropriate to apply in chance events such as gambling (Baboushkin, Hardoon, Derevensky, & Gupta, 2001) but also may be based on an initial misperception of the recalled gambling episodes as success representatives (Toneatto, Blitz-Miller, Calderwood, Dragonetti, & Tsanos, 1997). In fact, considering the consistency with expectation technique (mentioned earlier, Holland & Kensinger, 2010; Levine & Safer, 2002), the mere anticipation for a positive experience such as a win could perhaps in itself positively bias the memory of the gambling strings, which might keep gamblers captured in a cognitive loop. Indeed, the degree of a gambler’s optimism was found to be positively related to the degree to which the gambler overestimated the success of past gambling experiences (Gibson & Sanbonmatsu, 2004). Another maladaptive behavior that is strongly related to positive memory biases is alcoholism. As previously discussed, having emotions associated with negative events fading from memory faster than emotions associated with positive events, the fading affect bias clearly serves to improve one’s well-being (Walker, Skowronski, & Thompson, 2003). Nonetheless, because just like for

A “rosy view” of the past: Positive memory biases  Chapter | 7  155

any other event negative emotions associated with drinking episodes (e.g., hangovers, violence, and arguments) quickly fade too, the fading affect bias may also serve to encourage future alcohol consumption or even alcohol abuse. Indeed, Gibbons et al. (2013) found that alcoholism is a variable that may moderate positive memory biases. Whereas participants who consume alcohol at low levels demonstrated a greater fading affect bias for ordinary past events than for past drinking events, heavy alcohol consumers showed the opposite pattern, with a greater fading affect bias for drinking events than for events not involving alcohol. The authors suggested that, for low-frequency drinkers, an alcohol event is so exceptional that the negative emotions associated with it may be retained for a while and eventually fade nearly at the same rate as the positive emotions associated with that episode. For heavy consumers, on the other hand, drinking episodes regularly serve to abolish unpleasant feelings (Labouvie & Bates, 2002); thus, the negative emotions associated with alcohol events fade even more quickly, resulting in a larger positive memory bias than the regular one characterizing daily nonalcohol events. To summarize, given their strong association with well-being, positive memory biases are most prevalent in the general population. By contrast, within clinical populations, for whom well-being is frequently impaired, distortions of memory toward a positive direction are found in a very limited number of disorders (such as pathological gambling and alcoholism).

Summary, limitations, and future directions In the current chapter, we reviewed the mechanisms that underlie the prevalent phenomenon of positivity bias in memory. We described how by means of selfprotecting, self-enhancing, and preserving self-consistency people’s recollections frequently depict a much rosier past than it really was. Thus, with the prime goal of maintaining well-being and attuned by individual differences, people turn to social support, selectively recount only the positive aspects of experiences, reappraise negative situations, subjectively distance themselves from unflattering events, and even turn to current feelings or prior expectations to reconstruct the memories of past events and of their affiliated emotions (McFarland & Ross, 1987; Ross & Wilson, 2002; Rusting & DeHart, 2000; Safer et al., 2002; Skowronski et al., 2004). It goes without saying that, as autobiographical memory plays an important role in the construction of personal identity (Wilson & Ross, 2003), the memory advantage for positive information is especially strong when the information is self-relevant (Holland & Kensinger, 2010). As we mentioned before, memory biases for positive information were mostly exhibited in the context of autobiographical memory (Kensinger & Schacter, 2008), which is unique by its nature, exceptional, and often unverifiable. Because the conditions driving distortions of autobiographical content are difficult to manipulate or control in experimental settings (Bahrick et al., 2008), various paradigms (e.g., cue-word, mood induction, and diary paradigms) were

156  Cognitive biases in health and psychiatric disorders

designed in an attempt to deal with possible artifacts. In addition, findings from studies that have integrated neuroimaging with behavioral methods revealed the neural networks (mainly those including the prefrontal cortex and the amygdala) involved in the positivity bias of memory. We are certain that this avenue of research will continue to flourish in the future and will contribute additional important insights. Another avenue of research that we hope will advance in the near future is the validation of the emotional change associated with positive memory biases using physiological measures of arousal (e.g., GSR), in addition to the selfreport data and brain correlates of valence. This line of research is suitable for examining events experienced in the laboratory, regarding which such somatovisceral responses can be collected both at encoding and at retrieval (in contrast to spontaneous autobiographical events experienced outside the lab that allow the collection of somatovisceral responses only at retrieval). In fact, in a research project currently conducted in our lab (Adler & Pansky, in preparation), we are taking this approach in our investigation of the emotional attenuation involved in recounting a negative memory compared with retrieving it covertly or not retrieving it at all. The emphasis on one’s well-being as an adaptive motivational engine that directs the reconstruction of memories and the rarity of positive biases within clinical populations directly concerns the growing field of positive psychology. Positive psychology, which concentrates on one’s well-being instead of on illness, aims for understanding and fostering the factors that allow the individual to flourish (Seligman & Csikszentmihalyi, 2000). Toward these aims, positive psychology theoreticians have developed various interventions in which people can intentionally engage and thereby enhance their positive emotions (Duckworth et al., 2005; Seligman et al., 2005). One of these elevating activities is retrieving positive memories. To illustrate, in one study, in contrast to participants who were asked to recall their memories every night for 1 week with no specific instructions, participants who were asked to recall positive episodes that had occurred to them on the same day were found to be happier and less depressed at 1-, 3-, and 6-month follow-ups, compared with how they had felt before the retrieval exercise (Seligman et al., 2005). Considering the previously discussed effect of strengthening certain memory traces through rehearsal, it would be of great interest to examine whether positive memory biases contribute to the improvement in well-being that was apparent following positive psychology interventions. In fact, by virtue of modern technology, people today are commonly engaged in intentional positive reminiscence. For example, in their study, Good, Ancient, Postolache, Socianu, and Afghan (2013) demonstrated that, among Facebook activities, looking back on one’s photos and wall posts are the most frequent and have the strongest mood-improving influence, compared with updating status, using messenger, or even playing games. Indeed, since mobile phones have become so widespread, people are constantly documenting their

A “rosy view” of the past: Positive memory biases  Chapter | 7  157

feelings, thoughts, and photos of their experiences. When reviewed at a later point in time or shared via social networks, these posted texts and photo collections could be an effective means by which to support reminiscing. Yet, an asymmetry was found with regard to the valence of the content that people are inclined to capture and post, with a positive-content predominance. For example, on Facebook, positive posts are twice more frequent than negative posts (Kramer, Guillory, & Hancock, 2014). On Twitter, positive emojis are used three times more often than negative emojis (Novak, Smailović, Sluban, & Mozetič, 2015). Positive posts are also retweeted more often than the negative ones (Gruzd, Doiron, & Mai, 2011). Research has also revealed that people are inclined to post especially attractive versions of themselves by selecting photos in which they think they look good or are having fun (Siibak, 2009; Strano, 2008). This should come as no surprise, as Spence and Holland (1991) have long noted that family albums have a strong bias toward conveying an overly positive impression of family life, showing its members at happy times (Frohlich, Kuchinsky, Pering, Don, & Ariss, 2002). Undoubtedly, this selective documenting and posting, alongside the social support that social networks bring about, resemble the aforementioned recounting techniques that people regularly employ. Thus, considering the previously discussed findings with respect to social disclosure (Roediger & Karpicke, 2006; Skowronski & Walker, 2004; Tversky & Marsh, 2000), together with the fact that archived content may stabilize memory in the manner in which it was posted, there is reason to expect that, paradoxically, despite the veracity that is attributed to documentation, it will eventually result in a positive bias of memory. Furthermore, as correlations between posting positive content and well-being were found (Kramer, 2010; Schwartz et al., 2016), it would be interesting to examine whether the positivity bias in memory plays a mediating role in the relationship between documentation and well-being. Given the immense and ever-growing involvement of social media in today’s (and tomorrow’s) world, we believe that this type of research has a huge potential for yielding both theoretical and applied contributions. To conclude and as Bernstein and Loftus (2009) have stated, given that they are based on inherently reconstructive processes, all of our memories are, in essence, false to some degree. Indeed, positive biases in recollections are extremely prevalent. Given that the main mechanisms that underlie these biases are self-protection, self-enhancement, and self-consistency, the common human tendency to view life through rosy glasses mainly reflects an adaptive feature— an indicator of psychological health and well-being.

References Addis, D. R., Leclerc, C. M., Muscatell, K. A., & Kensinger, E. A. (2010). There are age-related changes in neural connectivity during the encoding of positive, but not negative, information. Cortex, 46, 425–433. Adolphs, R., Tranel, D., Damasio, H., & Damasio, A. (1994). Impaired recognition of emotion in facial expressions following bilateral damage to the human amygdale. Nature, 372, 669–672.

158  Cognitive biases in health and psychiatric disorders Ahn, H. M., Kim, S. A., Hwang, I. J., Jeong, J. W., Kim, H. T., Hamann, S., et al. (2015). The effect of cognitive reappraisal on long‐term emotional experience and emotional memory. Journal of Neuropsychology, 9, 64–76. Alba, J. W., & Hasher, L. (1983). Is memory schematic? Psychological Bulletin, 93, 203–231. Alicke, M. D., & Govorun, O. (2005). The better-than-average effect. The Self in Social Judgment, 1, 85–106. Alicke, M. D., & Sedikides, C. (2009). Self-enhancement and self-protection: What they are and what they do. European Review of Social Psychology, 20, 1–48. Amaral, D. G. (2003). The amygdala, social behavior, and danger detection. Annals of the New York Academy of Sciences, 1000, 337–347. Amaral, D. G., Price, J. L., Pitkänen, A., & Carmichael, S. (1992). Anatomical organization of the primate amygdaloid complex. In J. P. Aggleton (Ed.), The amygdala: Neurobiological aspects of emotion, memory, and mental dysfunction (pp. 1–66). New York: Wiley-Liss. American Psychiatric Association. (2013). Diagnostic and statistical manual of mental disorders (5th ed.). Washington, DC: American Psychiatric Association. Anderson, C. A. (1983). Imagination and expectation: The effect of imagining behavioral scripts on personal influences. Journal of Personality and Social Psychology, 45, 293–305. Anderson, R. C., & Pichert, J. W. (1978). Recall of previously unrecallable information following a shift in perspective. Journal of Verbal Learning and Verbal Behavior, 17, 1–12. Baboushkin, H. R., Hardoon, K. K., Derevensky, J. L., & Gupta, R. (2001). Underlying cognitions in gambling behavior among university students. Journal of Applied Social Psychology, 31, 1409–1430. Bahrick, H. P., Hall, L. K., & Berger, S. A. (1996). Accuracy and distortion in memory for high school grades. Psychological Science, 7, 265–271. Bahrick, H. P., Hall, L. K., & Da Costa, L. A. (2008). Fifty years of memory of college grades: Accuracy and distortions. Emotion, 8, 13–22. Barclay, C. R., & Wellman, H. M. (1986). Accuracies and inaccuracies in autobiographical memories. Journal of Memory and Language, 25, 93–103. Barrett, L. F. (1997). The relationships among momentary emotion experiences, personality descriptions, and retrospective ratings of emotion. Personality and Social Psychology Bulletin, 23, 1100–1110. Baumeister, R. F., Bratslavsky, E., Finkenauer, C., & Vohs, K. D. (2001). Bad is stronger than good. Review of General Psychology, 5, 323–370. Baumeister, R. F., & Leary, M. R. (2017). The need to belong: Desire for interpersonal attachments as a fundamental human motivation. In B.  Laursen & R.  Zukauskiene (Eds.), Interpersonal Development (pp. 57–89). London: Routledge. Bechara, A., Tranel, D., Damasio, H., Adolphs, R., Rockland, C., & Damasio, A. R. (1995). Double dissociation of conditioning and declarative knowledge relative to the amygdala and hippocampus in humans. Science, 269, 1115–1118. Bernstein, D. M., & Loftus, E. F. (2009). How to tell if a particular memory is true or false. Perspectives on Psychological Science, 4, 370–374. Berntsen, D. (2002). Tunnel memories for autobiographical events: Central details are remembered more frequently from shocking than from happy experiences. Memory & Cognition, 30, 1010–1020. Berntsen, D., & Rubin, D. C. (2002). Emotionally charged autobiographical memories across the life span: The recall of happy, sad, traumatic and involuntary memories. Psychology and Aging, 17, 636–652. Binder, J., Dominique, J. F., Friese, M., Luechinger, R., Boesiger, P., & Rasch, B. (2012). Emotion suppression reduces hippocampal activity during successful memory encoding. Neuroimage, 63, 525–532.

A “rosy view” of the past: Positive memory biases  Chapter | 7  159 Blaney, P. H. (1986). Affect and memory: A review. Psychological Bulletin, 99, 229–246. Bohanek, J. G., Fivush, R., & Walker, E. (2005). Memories of positive and negative emotional events. Applied Cognitive Psychology, 19, 51–66. Bonnici, H. M., & Maguire, E. A. (2018). Two years later—Revisiting autobiographical memory representations in vmPFC and hippocampus. Neuropsychologia, 110, 159–169. Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129–148. Brewer, W. F. (1996). What is recollective memory? In D. C. Rubin (Ed.), Remembering our past: Studies in autobiographical memory (pp. 19–66). Cambridge, England: Cambridge University Press. Brewin, C. R., Andrews, B., & Gotlib, I. H. (1993). Psychopathology and early experience: A reappraisal of retrospective reports. Psychological Bulletin, 113, 82–98. Brown, R., & Kulik, J. (1977). Flashbulb memories. Cognition, 5, 73–99. Bryant, R. A., McGrath, C., & Felmingham, K. L. (2013). The roles of noradrenergic and glucocorticoid activation in the development of intrusive memories. PLoS One, 8, e62675. Buchanan, T. W. (2007). Retrieval of emotional memories. Psychological Bulletin, 133, 761–779. Buchanan, T. W., Tranel, D., & Adolphs, R. (2005). Emotional autobiographical memories in amnesic patients with medial temporal lobe damage. Journal of Neuroscience, 25, 3151–3160. Burt, C. D., Kemp, S., & Conway, M. A. (2003). Themes, events, and episodes in autobiographical memory. Memory & Cognition, 31, 317–325. Burton, C. M., & King, L. A. (2004). The health benefits of writing about intensely positive experiences. Journal of Research in Personality, 38, 150–163. Cabeza, R., & St Jacques, P. (2007). Functional neuroimaging of autobiographical memory. Trends in Cognitive Sciences, 11, 219–227. Cahill, L. (2000). Emotional modulation of long-term memory storage in humans: Adrenergic activation and the amygdala. In J. Aggleton (Ed.), The amygdala: A functional analysis (pp. 425–445). New York: Oxford University Press. Canli, T., Sivers, H., Whitfield, S. L., Gotlib, I. H., & Gabrieli, J. D. (2002). Amygdala response to happy faces as a function of extraversion. Science, 296, 2191. Canli, T., Zhao, Z., Desmond, J. E., Kang, E., Gross, J., & Gabrieli, J. D. E. (2001). An fMRI study of personality influences on brain reactivity to emotional stimuli. Behavioral Neuroscience, 115, 33–42. Cardinal, R. N., Parkinson, J. A., Hall, J., & Everitt, B. J. (2002). Emotion and motivation: The role of the amygdala, ventral striatum, and prefrontal cortex. Neuroscience & Biobehavioral Reviews, 26, 321–352. Carstensen, L. L., Isaacowitz, D. M., & Charles, S. T. (1999). Taking time seriously: A theory of socioemotional selectivity. American Psychologist, 54, 165–181. Carstensen, L. L., & Mikels, J. A. (2005). At the intersection of emotion and cognition: Aging and the positivity effect. Current Directions in Psychological Science, 14, 117–121. Carstensen, L. L., Pasupathi, M., Mayr, U., & Nesselroade, J. R. (2000). Emotional experience in everyday life across the adult life span. Journal of Personality and Social Psychology, 79, 644–655. Charles, S. T., & Carstensen, L. L. (2010). Social and emotional aging. Annual Review of Psychology, 61, 383–409. Charles, S. T., Mather, M., & Carstensen, L. L. (2003). Aging and emotional memory: The forgettable nature of negative images for older adults. Journal of Experimental Psychology: General, 132, 310–324. Christensen, T. C., Wood, J. V., & Barrett, L. F. (2003). Remembering everyday experience through the prism of self-esteem. Personality and Social Psychology Bulletin, 29, 51–62.

160  Cognitive biases in health and psychiatric disorders Ciccarelli, M., Griffiths, M. D., Nigro, G., & Cosenza, M. (2017). Decision making, cognitive distortions and emotional distress: A comparison between pathological gamblers and healthy controls. Journal of Behavior Therapy and Experimental Psychiatry, 54, 204–210. Clore, G. L., Wyer, R. S., Dienes, B., Gasper, K., Gohm, C., & Isbell, L. (2001). Affective feelings as feedback: Some cognitive consequences. In L. L. Martin & G. L. Clore (Eds.), Theories of mood and cognition: A user’s handbook (pp. 27–62). Mahwah, NJ: Erlbaum. Comblain, C., D’Argembeau, A., & Van der Linden, M. (2005). Phenomenal characteristics of autobiographical memories for emotional and neutral events in older and younger adults. Experimental Aging Research, 31, 173–189. Conway, M. A. (2005). Memory and the self. Journal of Memory and Language, 53, 594–628. Crovitz, H. F., & Schiffman, H. (1974). Frequency of episodic memories as a function of their age. Bulletin of the Psychonomic Society, 4, 517–518. Croyle, R. T., Loftus, E. F., Barger, S. D., Sun, Y. C., Hart, M., & Gettig, J. (2006). How well do people recall risk factor test results? Accuracy and bias among cholesterol screening participants. Health Psychology, 25, 425–432. D’Argembeau, A., Renaud, O., & Van der Linden, M. (2011). Frequency, characteristics and functions of future-oriented thoughts in daily life. Applied Cognitive Psychology, 25, 96–103. D’Argembeau, A., & Van der Linden, M. (2004). Phenomenal characteristics associated with projecting oneself back into the past and forward into the future: Influence of valence and temporal distance. Consciousness and Cognition, 13, 844–858. D’Argembeau, A., & Van der Linden, M. (2008). Remembering pride and shame: Self-enhancement and the phenomenology of autobiographical memory. Memory, 16, 538–547. Davidson, R. J. (2004). Well-being and affective style: Neural substrates and biobehavioural correlates. Philosophical Transactions of the Royal Society of London. Series B: Biological Sciences, 359, 1395–1411. Davis, S. W., Dennis, N. A., Daselaar, S. M., Fleck, M. S., & Cabeza, R. (2007). Que PASA? The posterior-anterior shift in aging. Cerebral Cortex, 18, 1201–1209. de Vito, S., Neroni, M. A., Gamboz, N., Della Sala, S., & Brandimonte, M. A. (2015). Desirable and undesirable future thoughts call for different scene construction processes. The Quarterly Journal of Experimental Psychology, 68, 75–82. Delgado, M. R., Beer, J. S., Fellows, L. K., Huettel, S. A., Platt, M. L., Quirk, G. J., et al. (2016). Viewpoints: Dialogues on the functional role of the ventromedial prefrontal cortex. Nature Neuroscience, 19, 1545–1552. Delgado, M. R., Jou, R. L., LeDoux, J., & Phelps, L. (2009). Avoiding negative outcomes: Tracking the mechanisms of avoidance learning in humans during fear conditioning. Frontiers in Behavioral Neuroscience, 3, 33. Denkova, E., Dolcos, S., & Dolcos, F. (2012). Reliving emotional personal memories: Affective biases linked to personality and sex-related differences. Emotion, 12, 515–528. Denkova, E., Dolcos, S., & Dolcos, F. (2013a). The effect of retrieval focus and emotional valence on the inferior frontal cortex activity during autobiographical recollection. Frontiers in Behavioral Neuroscience, 7, 192. Denkova, E., Dolcos, S., & Dolcos, F. (2013b). The effect of retrieval focus and emotional valence on the medial temporal lobe activity during autobiographical recollection. Frontiers in Behavioral Neuroscience, 7, 109. Depue, B. E., Curran, T., & Banich, M. T. (2007). Prefrontal regions orchestrate suppression of emotional memories via a two-phase process. Science, 317, 215–219. Devitt, A. L., & Schacter, D. L. (2018). An optimistic outlook creates a rosy past: The impact of episodic simulation on subsequent memory. Psychological Science, 29, 936–946.

A “rosy view” of the past: Positive memory biases  Chapter | 7  161 Dillon, D. G., Ritchey, M., Johnson, B. D., & LaBar, K. S. (2007). Dissociable effects of conscious emotion regulation strategies on explicit and implicit memory. Emotion, 7, 354–365. Dolcos, F., Katsumi, Y., Weymar, M., Moore, M., Tsukiura, T., & Dolcos, S. (2017). Emerging directions in emotional episodic memory. Frontiers in Psychology, 8, 1867. Dolcos, F., LaBar, K. S., & Cabeza, R. (2005). Remembering one year later: Role of the amygdala and the medial temporal lobe memory system in retrieving emotional memories. Proceedings of the National Academy of Sciences, 102, 2626–2631. Dricu, M., Kress, L., & Aue, T. (2020). The neurophysiological basis of optimism bias. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological foundations (pp. 41–70). San Diego: Elsevier. Duckworth, A. L., Steen, T. A., & Seligman, M. E. P. (2005). Positive psychology in clinical practice. Annual Review of Clinical Psychology, 1, 629–651. Dunn, B. D., Billotti, D., Murphy, V., & Dalgleish, T. (2009). The consequences of effortful emotion regulation when processing distressing material: A comparison of suppression and acceptance. Behavior Research and Therapy, 47, 761–773. Eich, E., Macaulay, D., & Ryan, L. (1994). Mood dependent memory for events of the personal past. Journal of Experimental Psychology: General, 123, 201–215. Elliott, R., Ogilvie, A., Rubinsztein, J. S., Calderon, G., Dolan, R. J., & Sahakian, B. J. (2004). Abnormal ventral frontal response during performance of an affective go/no go task in patients with mania. Biological Psychiatry, 55, 1163–1170. Fernandes, M., Ross, M., Wiegand, M., & Schryer, E. (2008). Are the memories of older adults positively biased? Psychology and Aging, 23, 297–306. Fiedler, K. (2001). Affective states trigger processes of assimilation and accommodation. In L. L.  Martin & G. L.  Clore (Eds.), Theories of mood and cognition: A user’s guidebook (pp. 85–98). Mahwah, NJ: Erlbaum. Fitzgerald, D. A., Arnold, J. F., Becker, E. S., Speckens, A. E., Rinck, M., Rijpkema, M., et  al. (2011). How mood challenges emotional memory formation: An fMRI investigation. Neuroimage, 56, 1783–1790. Folkman, S., Lazarus, R. S., Pimley, S., & Novacek, J. (1987). Age differences in stress and coping processes. Psychology and Aging, 2, 171–184. Forgas, J. P. (1995). Mood and judgment: The affect infusion model (AIM). Psychological Bulletin, 117, 39–66. Fortune, E. E., & Goodie, A. S. (2012). Cognitive distortions as a component and treatment focus of pathological gambling: A review. Psychology of Addictive Behaviors, 26, 298–310. Fredrickson, B. L. (2000). Extracting meaning from past affective experiences: The importance of peaks, ends, and specific emotions. Cognition and Emotion, 14, 577–606. Freud, S. (1957). Repression. (1915). Standard Edition, Vol. 14 (pp. 141–158). London: Hogarth Press. Friedman, N. P., & Miyake, A. (2017). Unity and diversity of executive functions: Individual differences as a window on cognitive structure. Cortex, 86, 186–204. Frohlich, D., Kuchinsky, A., Pering, C., Don, A., & Ariss, S. (2002). Requirements for photoware. In Proceedings of the 2002 ACM conference on computer supported cooperative work (pp. 166–175). New York: ACM Press. Galton, F. (1879). Psychometric experiments. Brain, 2, 149–162. Gibbons, J. A., Toscano, A., Kofron, S., Rothwell, C., Lee, S. A., Ritchie, T. D., et al. (2013). The fading affect bias across alcohol consumption frequency for alcohol-related and non-alcoholrelated events. Consciousness and Cognition, 22, 1340–1351. Gibson, B., & Sanbonmatsu, D. M. (2004). Optimism, pessimism, and gambling: The downside of optimism. Personality and Social Psychology Bulletin, 30, 149–160.

162  Cognitive biases in health and psychiatric disorders Good, A., Ancient, C., Postolache, G., Socianu, A., & Afghan, A. (2013). Self-soothing by reviewing favorite memories: An exploration of mobile application prototypes, which facilitate positive wellbeing via reminiscing. In International conference on human-computer interaction (pp. 417–421). Heidelberg: Springer. Goodie, A. S., & Fortune, E. E. (2013). Measuring cognitive distortions in pathological gambling: Review and meta-analyses. Psychology of Addictive Behaviors, 27, 730–743. Gotlib, I. H., Kasch, K. L., Traill, S., Joormann, J., Arnow, B. A., & Johnson, S. L. (2004). Coherence and specificity of information-processing biases in depression and social phobia. Journal of Abnormal Psychology, 113, 386–398. Green, J. D., Pinter, B., & Sedikides, C. (2005). Mnemic neglect and self‐threat: Trait modifiability moderates self‐protection. European Journal of Social Psychology, 35, 225–235. Greenwald, A. G. (1980). The totalitarian ego: Fabrication and revision of personal history. American Psychologist, 35, 603–618. Gross, J. J. (1998). Antecedent- and response-focused emotion regulation: Divergent consequences for experience, expression, and physiology. Journal of Personality and Social Psychology, 74, 224–237. Gruber, J. (2011). Can feeling too good be bad? Positive emotion persistence (PEP) in bipolar disorder. Current Directions in Psychological Science, 20, 217–221. Gruber, J., Harvey, A. G., & Johnson, S. L. (2009). Reflective and ruminative processing of positive emotional memories in bipolar disorder and healthy controls. Behaviour Research and Therapy, 47, 697–704. Gruzd, A., Doiron, S., & Mai, P. (2011). Is happiness contagious online? A case of Twitter and the 2010 Winter Olympics. In Hawaii international conference on system sciences (pp. 1–9). Haas, B. W., & Canli, T. (2008). Emotional memory function, personality structure and psychopathology: A neural system approach to the identification of vulnerability markers. Brain Research Reviews, 58, 71–84. Haber, S. N., & Knutson, B. (2010). The reward circuit: Linking primate anatomy and human imaging. Neuropsychopharmacology, 35, 4–26. Hayes, J. P., Morey, R. A., Petty, C. M., Seth, S., Smoski, M. J., McCarthy, G., et al. (2010). Staying cool when things get hot: Emotion regulation modulates neural mechanisms of memory encoding. Frontiers in Human Neuroscience, 4, 230. Henry, G. M., Weingartner, H., & Murphy, D. L. (1971). Idiosyncratic patterns of learning and word association during mania. American Journal of Psychiatry, 128, 564–574. Holland, A. C., & Kensinger, E. A. (2010). Emotion and autobiographical memory. Physics of Life Reviews, 7, 88–131. Holland, A. C., Tamir, M., & Kensinger, E. A. (2010). The effect of regulation goals on emotional event-specific knowledge. Memory, 18, 504–521. Isen, A. M. (1985). Asymmetry of happiness and sadness in effects on memory in normal college students. Journal of Experimental Psychology: General, 114, 388–391. Isen, A. M. (1987). Positive affect, cognitive processes, and social behavior. In L. Berkowitz (Ed.), Advances in experimental social psychology (pp. 203–253). New York: Academic Press. Josselyn, S. A., Köhler, S., & Frankland, P. W. (2015). Finding the engram. Nature Reviews Neuroscience, 16, 521–534. Kaplan, R. L., Levine, L. J., Lench, H. C., & Safer, M. A. (2016). Forgetting feelings: Opposite biases in reports of the intensity of past emotion and mood. Emotion, 16, 309–319. Kemp, S., Burt, C. D., & Furneaux, L. (2008). A test of the peak-end rule with extended autobiographical events. Memory & Cognition, 36, 132–138.

A “rosy view” of the past: Positive memory biases  Chapter | 7  163 Kennedy, Q., Mather, M., & Carstensen, L. L. (2004). The role of motivation in the age-related positivity effect in autobiographical memory. Psychological Science, 15, 208–214. Kensinger, E. A., & Corkin, S. (2004). Two routes to emotional memory: Distinct neural processes for valence and arousal. Proceedings of the National Academy of Sciences, 101, 3310–3315. Kensinger, E. A., & Schacter, D. L. (2008). Memory and emotion. In M. Lewis, J. M. HavilandJones, & L. F. Barrett (Eds.), Handbook of emotions (pp. 601–617). New York: Guilford Press. Kim, S. H., & Hamann, S. (2012). The effect of cognitive reappraisal on physiological reactivity and emotional memory. International Journal of Psychophysiology, 83, 348–356. Klüver, H., & Bucy, P. C. (1939). Preliminary analysis of functions of the temporal lobes in monkeys. Archives of Neurology & Psychiatry, 42, 979–1000. Koo, M., & Oishi, S. (2009). False memory and the associative network of happiness. Personality and Social Psychology Bulletin, 35, 212–220. Koster, E. H., De Raedt, R., Leyman, L., & De Lissnyder, E. (2010). Mood-congruent attention and memory bias in dysphoria: Exploring the coherence among information-processing biases. Behaviour Research and Therapy, 48, 219–225. Kragel, P. A., & LaBar, K. S. (2016). Decoding the nature of emotion in the brain. Trends in Cognitive Sciences, 20, 444–455. Kramer, A. D. (2010). An unobtrusive behavioral model of gross national happiness. In Proceedings of the 28th conference on human factors in computing systems (pp. 287–290). Kramer, A. D., Guillory, J. E., & Hancock, J. T. (2014). Experimental evidence of massive-scale emotional contagion through social networks. Proceedings of the National Academy of Sciences, 111, 8788–8790. Labouvie, E., & Bates, M. E. (2002). Reasons for alcohol use in young adulthood: Validation of a three-dimensional measure. Journal of Studies on Alcohol, 63, 145–155. Landau, J. D., & Gunter, B. C. (2009). “Don’t worry; you really will get over it”: Methodological investigations of the fading affect bias. The American Journal of Psychology, 122, 209–217. Lane, R. D., Ryan, L., Nadel, L., & Greenberg, L. (2015). Memory reconsolidation, emotional arousal, and the process of change in psychotherapy: New insights from brain science. Behavioral and Brain Sciences, 38, 1–64. Larsen, R. J., & Ketelaar, T. (1991). Personality and susceptibility to positive and negative emotional states. Journal of Personality and Social Psychology, 61, 132–140. Leary, M. R., & Kowalski, R. M. (1990). Impression management: A literature review and twocomponent model. Psychological Bulletin, 107, 34–47. LeBlanc, J., & Ducharme, M. B. (2005). Influence of personality traits on plasma levels of cortisol and cholesterol. Physiology & Behavior, 84, 677–680. LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23, 155–184. LeMoult, J., & Gotlib, I. H. (2019). Depression: A cognitive perspective. Clinical Psychology ­Review, 69, 51–66. Levine, L. J. (1997). Reconstructing memory for emotions. Journal of Experimental Psychology: General, 126, 165–177. Levine, L. J., & Bluck, S. (1997). Experienced and remembered emotional intensity in older adults. Psychology and Aging, 12, 514–523. Levine, L. J., & Bluck, S. (2004). Painting with broad strokes: Happiness and the malleability of event memory. Cognition and Emotion, 18, 559–574. Levine, L. J., Lench, H. C., & Safer, M. A. (2009). Functions of remembering and misremembering emotion. Applied Cognitive Psychology, 23, 1059–1075. Levine, L. J., & Pizarro, D. A. (2004). Emotion and memory research: A grumpy overview. Social Cognition, 22, 530–554.

164  Cognitive biases in health and psychiatric disorders Levine, L. J., & Safer, M. A. (2002). Sources of bias in memory for emotions. Current Directions in Psychological Science, 11, 169–173. Levine, L. J., Safer, M. A., & Lench, H. C. (2006). Remembering and misremembering emotions. In L. J. Sanna & E. C. Chang (Eds.), Judgments over time: The interplay of thoughts, feelings, and behaviors (pp. 271–290). New York: Oxford University Press. Levine, L. J., Schmidt, S., Kang, H. S., & Tinti, C. (2012). Remembering the silver lining: Reappraisal and positive bias in memory for emotion. Cognition and Emotion, 26, 871–884. Levine, B., Svoboda, E., Hay, J. F., Winocur, G., & Moscovitch, M. (2002). Aging and autobiographical memory: Dissociating episodic from semantic retrieval. Psychology and Aging, 17, 677–689. Lewis, P. A., Critchley, H. D., Smith, A. P., & Dolan, R. J. (2005). Brain mechanisms for mood congruent memory facilitation. Neuroimage, 25, 1214–1223. Libby, L. K., & Eibach, R. P. (2007). How the self affects and reflects the content and subjective experience of autobiographical memory. In C.  Sedikides & S.  Spencer (Eds.), The self (pp. 75–91). New York: Psychology Press. Lin, M., Hofmann, S. G., Qian, M., & Li, S. (2015). Enhanced association between perceptual stimuli and trauma-related information in individuals with posttraumatic stress disorder symptoms. Journal of Behavior Therapy and Experimental Psychiatry, 46, 202–207. Lindquist, K. A., Wager, T. D., Kober, H., Bliss-Moreau, E., & Barrett, L. F. (2012). The brain basis of emotion: A meta-analytic review. Behavioral and Brain Sciences, 35, 121–143. Liu, F., Cui, L., & Zhang, Q. (2015). The influences of reappraisal and suppression instructions on memory for neutral words in negative background. Neuroreport, 26, 1023–1031. Lloyd, G. G., & Lishman, W. A. (1975). Effect of depression on the speed of recall of pleasant and unpleasant experiences. Psychological Medicine, 5, 173–180. Loftus, E. F. (1982). Memory and its distortions. In A. G. Kraut (Ed.), The G. Stanley Hall lecture series (pp. 123–154). Washington, DC: American Psychological Association. Loftus, E. F. (2003). Make-believe memories. American Psychologist, 58, 867–873. Lyubomirsky, S., Sheldon, K. M., & Schkade, D. (2005). Pursuing happiness: The architecture of sustainable change. Review of General Psychology, 9, 111–131. Macleod, A. K., Andersen, A., & Davies, A. (1994). Self-ratings of positive and negative affect and retrieval of positive and negative affect memories. Cognition and Emotion, 8, 483–488. Madigan, R. J., & Bollenbach, A. K. (1982). Effects of induced mood on retrieval of personal episodic and semantic memories. Psychological Reports, 50, 147–157. Markus, H. R., & Kitayama, S. (1994). A collective fear of the collective: Implications for selves and theories of selves. Personality and Social Psychology Bulletin, 20, 568–579. Markus, H. R., & Kitayama, S. (2010). Cultures and selves: A cycle of mutual constitution. Perspectives on Psychological Science, 5, 420–430. Master, D., Lishman, W. A., & Smith, A. (1983). Speed of recall in relation to affective tone and intensity of experience. Psychological Medicine, 13, 325–331. Mather, M. (2006). Why memories may become more positive as people age. In B. Uttl, N. Ohta, & A. Siegenthale (Eds.), Memory and emotion: Interdisciplinary perspectives (pp. 135–158). Malden, MA: Blackwell. Mather, M., & Carstensen, L. L. (2003). Aging and attentional biases for emotional face. Psychological Science, 14, 409–415. Mather, M., & Carstensen, L. L. (2005). Aging and motivated cognition: The positivity effect in attention and memory. Trends in Cognitive Sciences, 9, 496–502. Mather, M., Knight, M., & McCaffrey, M. (2005). The allure of the alignable: Younger and older adults’ false memories of choice features. Journal of Experimental Psychology: General, 134, 38–51.

A “rosy view” of the past: Positive memory biases  Chapter | 7  165 Mathews, A., & MacLeod, C. (1994). Cognitive approaches to emotion and emotional disorders. Annual Review of Psychology, 45, 25–50. Matlin, M. W., & Stang, D. J. (1978). The Pollyanna principle: Selectivity in language, memory, and thought. Cambridge, MA: Schenkman. Matt, G. E., Vázquez, C., & Campbell, W. K. (1992). Mood-congruent recall of affectively toned stimuli: A meta-analytic review. Clinical Psychology Review, 12, 227–255. Mayo, P. R. (1983). Personality traits and the retrieval of positive and negative memories. Personality and Individual Differences, 4, 465–471. McFarland, C., & Ross, M. (1987). The relation between current impressions and memories of self and dating partners. Personality and Social Psychology Bulletin, 13, 228–238. McGaugh, J. L. (2004). The amygdala modulates the consolidation of memories of emotionally arousing experiences. Annual Review of Neuroscience, 27, 1–28. McIntyre, C. K., McGaugh, J. L., & Williams, C. L. (2012). Interacting brain systems modulate memory consolidation. Neuroscience & Biobehavioral Reviews, 36, 1750–1762. Mickley Steinmetz, K. R., Addis, D. R., & Kensinger, E. A. (2010). The effect of arousal on the emotional memory network depends on valence. NeuroImage, 53, 318–324. Mickley Steinmetz, K. R., & Kensinger, E. A. (2009). The effects of valence and arousal on the neural activity leading to subsequent memory. Psychophysiology, 46, 1190–1199. Mill, A., Realo, A., & Allik, J. (2016). Retrospective ratings of emotions: The effects of age, daily tiredness, and personality. Frontiers in Psychology, 6, 2020. Miller, G. E., Cohen, S., Rabin, B. S., Skoner, D., & Doyle, W. J. (1999). Personality and tonic cardiovascular, neuroendocrine, and immune parameters. Brain, Behavior and Immunity, 13, 109–123. Miranda, R., & Kihlstrom, J. (2005). Mood congruence in childhood and recent autobiographical memory. Cognition and Emotion, 19, 981–998. Miron-Shatz, T., Stone, A., & Kahneman, D. (2009). Memories of yesterday’s emotions: Does the valence of experience affect the memory-experience gap? Emotion, 9, 885–891. Mitchell, T. R., Thompson, L., Peterson, E., & Cronk, R. (1997). Temporal adjustments in the evaluation of events: The “rosy view”. Journal of Experimental Social Psychology, 33, 421–448. Morrone, J. V., Depue, R. A., Scherer, A. J., & White, T. L. (2000). Film-induced incentive motivation and positive activation in relation to agentic and affiliative components of extraversion. Personality and Individual Differences, 29, 199–216. Moscovitch, M., Cabeza, R., Winocur, G., & Nadel, L. (2016). Episodic memory and beyond: The hippocampus and neocortex in transformation. Annual Review of Psychology, 67, 105–134. Murphy, F. C., Sahakian, B. J., Rubinsztein, J. S., Michael, A., Rogers, R. D., Robbins, T. W., et al. (1999). Emotional bias and inhibitory control processes in mania and depression. Psychological Medicine, 29, 1307–1321. Murty, V. P., Sambataro, F., Das, S., Tan, H. Y., Callicott, J. H., Goldberg, T. E., et al. (2009). Agerelated alterations in simple declarative memory and the effect of negative stimulus valence. Journal of Cognitive Neuroscience, 21, 1920–1933. Natale, M., & Hantas, M. (1982). Effect of temporary mood states on selective memory about the self. Journal of Personality and Social Psychology, 42, 927–934. Nolen-Hoeksema, S. (2000). The role of rumination in depressive disorders and mixed anxiety/ depressive symptoms. Journal of Abnormal Psychology, 109, 504–511. Nørby, S. (2015). Why forget? On the adaptive value of memory loss. Perspectives on Psychological Science, 10, 551–578. Novak, P. K., Smailović, J., Sluban, B., & Mozetič, I. (2015). Sentiment of emojis. PLoS One, 10, e0144296. O’Doherty, J. P. (2004). Reward representations and reward-related learning in the human brain: Insights from neuroimaging. Current Opinion in Neurobiology, 14, 769–776.

166  Cognitive biases in health and psychiatric disorders Oishi, S. (2002). The experiencing and remembering of well-being: A cross-cultural analysis. Personality and Social Psychology Bulletin, 28, 1398–1406. Oishi, S., Schimmack, U., Diener, E., Kim-Prieto, C., Scollon, C. N., & Choi, D. W. (2007). The value-congruence model of memory for emotional experiences: An explanation for cultural differences in emotional self-reports. Journal of Personality and Social Psychology, 93, 897–905. Piefke, M., Weiss, P. H., Zilles, K., Markowitsch, H. J., & Fink, G. R. (2003). Differential remoteness and emotional tone modulate the neural correlates of autobiographical memory. Brain, 126, 650–668. Piguet, O., Connally, E., Krendl, A. C., Huot, J. R., & Corkin, S. (2008). False memory in aging: Effects of emotional valence on word recognition accuracy. Psychology and Aging, 23, 307–314. Pinter, B., Green, J. D., Sedikides, C., & Gregg, A. P. (2011). Self-protective memory: Separation/ integration as a mechanism for mnemic neglect. Social Cognition, 29, 612–624. Pratto, F., & John, O. P. (1991). Automatic vigilance: The attention-grabbing power of negative social information. Journal of Personality and Social Psychology, 61, 380–391. Rachlin, H. (1990). Why do people gamble and keep gambling despite heavy losses? Psychological Science, 1, 294–297. Radvansky, G. A. (2017). Human memory. New York, NY: Routledge. Rasmussen, A. S., & Berntsen, D. (2013). The reality of the past versus the ideality of the future: Emotional valence and functional differences between past and future mental time travel. Memory & Cognition, 41, 187–200. Rathbone, C. J., Holmes, E. A., Murphy, S. E., & Ellis, J. A. (2015). Autobiographical memory and well-being in aging: The central role of semantic self-images. Consciousness and Cognition, 33, 422–431. Reed, A. E., Chan, L., & Mikels, J. A. (2014). Meta-analysis of the age-related positivity effect: Age differences in preferences for positive over negative information. Psychology and Aging, 29, 1–15. Richards, J. M., & Gross, J. J. (1999). Composure at any cost? The cognitive consequences of emotion suppression. Personality and Social Psychology Bulletin, 25, 1033–1044. Riskind, J. H. (1983). Nonverbal expressions and the accessibility of life experience memories: A congruence hypothesis. Social Cognition, 2, 62–86. Ritchey, M., LaBar, K. S., & Cabeza, R. (2011). Level of processing modulates the neural correlates of emotional memory formation. Journal of Cognitive Neuroscience, 23, 757–771. Ritchie, T. D., Batteson, T. J., Bohn, A., Crawford, M. T., Ferguson, G. V., Schrauf, R. W., et al. (2015). A pancultural perspective on the fading affect bias in autobiographical memory. Memory, 23, 278–290. Ritchie, T. D., Sedikides, C., & Skowronski, J. J. (2016). Emotions experienced at event recall and the self: Implications for the regulation of self-esteem, self-continuity and meaningfulness. Memory, 24, 577–591. Ritchie, T. D., Sedikides, C., & Skowronski, J. J. (2017). Does a person selectively recall the good or the bad from their personal past? It depends on the recall target and the person’s favourability of self-views. Memory, 25, 934–944. Ritchie, T. D., Skowronski, J. J., Cadogan, S., & Sedikides, C. (2014). Affective responses to selfdefining autobiographical events. Self and Identity, 13, 513–534. Ritchie, T., Skowronski, J. J., Hartnett, J., Wells, B., & Walker, W. R. (2009). The fading affect bias in the context of emotion activation level, mood, and personal theories of emotion change. Memory, 17, 428–444. Ritchie, T. D., Walker, W. R., Marsh, S., Hart, C., & Skowronski, J. J. (2015). Narcissism distorts the fading affect bias in autobiographical memory. Applied Cognitive Psychology, 29, 104–114.

A “rosy view” of the past: Positive memory biases  Chapter | 7  167 Robinson, M. D., & Clore, G. L. (2002). Belief and feeling: Evidence for an accessibility model of emotional self-report. Psychological Bulletin, 128, 934–960. Roediger, H. L., & Karpicke, J. D. (2006). Test-enhanced learning: Taking memory tests improves long-term retention. Psychological Science, 17, 249–255. Ross, M. (1989). Relation of implicit theories to the construction of personal histories. Psychological Review, 96, 341–357. Ross, M., Blatz, C. W., & Schryer, E. (2008). Social memory processes [Series Ed.). (Vol. Ed.]. In J. Byrne & H. Roediger (Eds.), Learning and memory: A comprehensive reference: Cognitive psychology of memory (pp. 911–926). Oxford, England: Elsevier. Ross, M., Heine, S. J., Wilson, A. E., & Sugimori, S. (2005). Cross-cultural discrepancies in selfappraisals. Personality and Social Psychology Bulletin, 31, 1175–1188. Ross, M., & Wilson, A. E. (2002). It feels like yesterday: Self-esteem, valence of personal past experiences, and judgments of subjective distance. Journal of Personality and Social Psychology, 82, 792–803. Ross, M., & Wilson, A. E. (2003). Autobiographical memory and conceptions of self: Getting better all the time. Current Directions in Psychological Science, 12, 66–69. Roy, M., Shohamy, D., & Wager, T. D. (2012). Ventromedial prefrontal-subcortical systems and the generation of affective meaning. Trends in Cognitive Sciences, 16, 147–156. Rubin, D. C. (2002). Autobiographical memory across the lifespan. In P. Graf & N. Ohta (Eds.), Lifespan development of human memory (pp. 159–184). Cambridge, MA: MIT Press. Rubin, D. C., Berntsen, D., & Bohni, M. K. (2008). A memory-based model of posttraumatic stress disorder: Evaluating basic assumptions underlying the PTSD diagnosis. Psychological Review, 115, 985–1011. Rubin, D. C., & Wenzel, A. (2004). Autobiographical memory tasks: Six common methods. In A. Wenzel & D. C. Rubin (Eds.), Cognitive methods and their application to clinical research (pp. 215–217). Washington, DC: American Psychological Association. Rusting, C. L. (1998). Personality, mood, and cognitive processing of emotional information: Three conceptual frameworks. Psychological Bulletin, 124, 165–196. Rusting, C. L. (1999). Interactive effects of personality and mood on emotion-congruent memory and judgment. Journal of Personality and Social Psychology, 77, 1073–1086. Rusting, C. L., & DeHart, T. (2000). Retrieving positive memories to regulate negative mood: Consequences for mood-congruent memory. Journal of Personality and Social Psychology, 78, 737–752. Safer, M. A., Bonanno, G. A., & Field, N. P. (2001). “It was never that bad”: Biased recall of grief and long-term adjustment to the death of a spouse. Memory, 9, 195–203. Safer, M. A., & Keuler, D. J. (2002). Individual differences in misremembering pre-psychotherapy distress: Personality and memory distortion. Emotion, 2, 162–178. Safer, M. A., Levine, L. J., & Drapalski, A. L. (2002). Distortion in memory for emotions: The contributions of personality and post-event knowledge. Personality and Social Psychology Bulletin, 28, 1495–1507. Sakaki, M., Nga, L., & Mather, M. (2013). Amygdala functional connectivity with medial prefrontal cortex at rest predicts the positivity effect in older adults’ memory. Journal of Cognitive Neuroscience, 25, 1206–1224. Samanez-Larkin, G. R., & Carstensen, L. L. (2011). Socioemotional functioning and the aging brain. In J. Decety & J. T. Cacioppo (Eds.), The Oxford handbook of social neuroscience (pp. 507–521). New York, NY: Oxford University Press. Schacter, D. L. (2002). The sevens sins of memory: How the mind forgets and remembers. Boston, MA: Houghton Mifflin.

168  Cognitive biases in health and psychiatric disorders Schacter, D. L., Guerin, S. A., & St Jacques, P. L. (2011). Memory distortion: An adaptive perspective. Trends in Cognitive Sciences, 15, 467–474. Schwartz, H. A., Sap, M., Kern, M. L., Eichstaedt, J. C., Kapelner, A., Agrawal, M., et al. (2016). Predicting individual well-being through the language of social media. Pacific Symposium on Biocomputing, 21, 516–527. Schwarz, N. (1990). Feelings as information: Informational and motivational functions of affective states. In E. T. Higgins & E. M. Sorrentino (Eds.), Handbook of motivation and cognition (pp. 527–561). New York: Guilford. Sedikides, C., & Alicke, M. D. (2012). Self-enhancement and self-protection motives. In R. M.  Ryan (Ed.), The Oxford handbook of human motivation (pp. 303–322). Oxford: Oxford University Press. Sedikides, C., Gaertner, L., & Cai, H. (2015). On the panculturality of self-enhancement and selfprotection motivation: The case for the universality of self-esteem. Advances in Motivation Science, 2, 185–241. Sedikides, C., & Green, J. D. (2000). On the self-protective nature of inconsistency-negativity management: Using the person memory paradigm to examine self-referent memory. Journal of Personality and Social Psychology, 79, 906–922. Sedikides, C., & Green, J. D. (2009). Memory as a self‐protective mechanism. Social and Personality Psychology Compass, 3, 1055–1068. Sedikides, C., & Gregg, A. P. (2003). Portraits of the self. In M. A. Hogg & J. Cooper (Eds.), Sage handbook of social psychology (pp. 110–138). London, England: Sage. Sedikides, C., & Gregg, A. P. (2008). Self-enhancement: Food for thought. Perspectives on Psychological Science, 3, 102–116. Seligman, M. E., & Csikszentmihalyi, M. (2000). Positive psychology: An introduction. American Psychologist, 55, 5–14. Seligman, M. E., Steen, T. A., Park, N., & Peterson, C. (2005). Positive psychology progress: Empirical validation of interventions. American Psychologist, 60, 410–421. Sharot, T. (2011). The optimism bias. Current Biology, 21, 941–945. Sharot, T., Riccardi, A. M., Raio, C. M., & Phelps, E. A. (2007). Neural mechanisms mediating optimism bias. Nature, 450, 102–105. Shimamura, A. P. (2000). The role of the prefrontal cortex in dynamic filtering. Psychobiology, 28, 207–218. Siibak, A. (2009). Constructing the self through the photo selection-visual impression management on social networking websites. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 3(1) [article 1]. Singer, J. A., & Salovey, P. (1988). Mood and memory: Evaluating the network theory of affect. Clinical Psychology Review, 8, 211–251. Skowronski, J. J. (2011). The positivity bias and the fading affect bias in autobiographical memory: A self-motives perspective. In M. D.  Alicke & C.  Sedikides (Eds.), Handbook of self-­ enhancement and self-protection (pp. 211–231). New York: Guilford Press. Skowronski, J. J., Betz, A. L., Thompson, C. P., & Shannon, L. (1991). Social memory in everyday life: Recall of self-events and other-events. Journal of Personality and Social Psychology, 60, 831–843. Skowronski, J., Gibbons, J., Vogl, R., & Walker, W. R. (2004). The effect of social disclosure on the intensity of affect provoked by autobiographical memories. Self and Identity, 3, 285–309. Skowronski, J. J., & Walker, W. R. (2004). How describing autobiographical events can affect autobiographical memories. Social Cognition, 22, 555–590.

A “rosy view” of the past: Positive memory biases  Chapter | 7  169 Skowronski, J. J., Walker, W. R., Henderson, D. X., & Bond, G. D. (2014). The fading affect bias: Its history, its implications, and its future. Advances in Experimental Social Psychology, 49, 163–218. Smith, S. M., & Petty, R. E. (1995). Personality moderators of mood congruency effects on cognition: The role of self-esteem and negative mood regulation. Journal of Personality and Social Psychology, 68, 1092–1107. Snyder, M., & White, P. (1982). Moods and memories: Elation, depression, and the remembering of the events of one’s life. Journal of Personality, 50, 149–167. Song, Q., & Wang, Q. (2014). Harmed or not harmed? Culture in interpersonal transgression memory and self-acceptance. International Journal of Applied Psychology, 4, 188–195. Speer, M. E., Bhanji, J. P., & Delgado, M. R. (2014). Savoring the past: Positive memories evoke value representations in the striatum. Neuron, 84, 847–856. Spence, J., & Holland, P. (1991). Family snaps: The meanings of domestic photography. London: Virago. St Jacques, P. L., Dolcos, F., & Cabeza, R. (2009). Effects of aging on functional connectivity of the amygdala for subsequent memory of negative pictures: A network analysis of functional magnetic resonance imaging data. Psychological Science, 20, 74–84. Stone, A. A., & Broderick, J. E. (2007). Real-time data collection for pain: Appraisal and current status. Pain Medicine, 8, S85–S93. Strano, M. M. (2008). User descriptions and interpretations of self-presentation through Facebook profile images. Cyberpsychology: Journal of Psychosocial Research on Cyberspace, 2. article 5. Sutton, R. I. (1992). Feelings about a Disneyland visit: Photography and the reconstruction of bygone emotions. Journal of Management Inquiry, 1, 278–287. Szpunar, K. K., Addis, D. R., & Schacter, D. L. (2012). Memory for emotional simulations: Remembering a rosy future. Psychological Science, 23, 24–29. Tamir, D. I., & Mitchell, J. P. (2012). Disclosing information about the self is intrinsically rewarding. Proceedings of the National Academy of Sciences, 109, 8038–8043. Taylor, S. E. (1991). Asymmetrical effects of positive and negative events: The mobilization-­ minimization hypothesis. Psychological Bulletin, 110, 67–85. Teasdale, J. D., Taylor, R., & Fogarty, S. J. (1980). Effects of induced elation-depression on the accessibility of memories of happy and unhappy experiences. Behaviour Research and Therapy, 18, 339–346. Thompson, C. P., Skowronski, J. J., Larsen, S. F., & Betz, A. L. (2013). Autobiographical memory: Remembering what and remembering when. New York: Psychology Press. Todd, M., Tennen, H., Carney, M. A., Armeli, S., & Affleck, G. (2004). Do we know how we cope? Relating daily coping reports to global and time-limited retrospective assessments. Journal of Personality and Social Psychology, 86, 310–319. Toglia, M. P., Read, J. D., Ross, D. F., & Lindsay, R. C. L. (2017). The handbook of eyewitness psychology: Volume I: Memory for events. New York, NY: Routledge. Toneatto, T., Blitz-Miller, T., Calderwood, K., Dragonetti, R., & Tsanos, A. (1997). Cognitive distortions in heavy gambling. Journal of Gambling Studies, 13, 253–266. Toyama, A., Katsuhara, M., Sakurai, Y., & Ohira, H. (2014). Effects of emotional context during encoding: An advantage for negative context in immediate recognition and positive context in delayed recognition. Psychology, 5, 994–1000. Tsukiura, T., & Cabeza, R. (2008). Orbitofrontal and hippocampal contributions to memory for face-name associations: The rewarding power of a smile. Neuropsychologia, 46, 2310–2319. Tversky, A., & Kahneman, D. (1973). Availability: A heuristic for judging frequency and probability. Cognitive Psychology, 5, 207–232.

170  Cognitive biases in health and psychiatric disorders Tversky, B., & Marsh, E. J. (2000). Biased retellings of events yield biased memories. Cognitive Psychology, 40, 1–38. Västfjäll, D. (2001). Emotion induction through music: A review of the musical mood induction procedure. Musicae Scientiae, 5, 173–211. Wagenaar, W. A. (1988). Paradoxes of gambling behaviour. London: Erlbaum. Wagenaar, W. A., & Groeneweg, J. (1990). The memory of concentration camp survivors. Applied Cognitive Psychology, 4, 77–87. Walker, W. R., Skowronski, J. J., Gibbons, J. A., Vogl, R. J., & Ritchie, T. D. (2009). Why people rehearse their memories: Frequency of use and relations to the intensity of emotions associated with autobiographical memories. Memory, 17, 760–773. Walker, W. R., Skowronski, J. J., Gibbons, J. A., Vogl, R. J., & Thompson, C. P. (2003). On the emotions that accompany autobiographical memories: Dysphoria disrupts the fading affect bias. Cognition and Emotion, 17, 703–723. Walker, W. R., Skowronski, J. J., & Thompson, C. P. (2003). Life is pleasant—and memory helps to keep it that way!. Review of General Psychology, 7, 203–210. Walker, W. R., Vogl, R. J., & Thompson, C. P. (1997). Autobiographical memory: Unpleasantness fades faster than pleasantness over time. Applied Cognitive Psychology, 11, 399–413. Walker, W. R., Yancu, C. N., & Skowronski, J. J. (2014). Trait anxiety reduces affective fading for both positive and negative autobiographical memories. Advances in Cognitive Psychology, 10, 81–89. Wang, Q. (2016). Remembering the self in cultural contexts: A cultural dynamic theory of autobiographical memory. Memory Studies, 9, 295–304. Waring, J. D., Addis, D. R., & Kensinger, E. A. (2013). Effects of aging on neural connectivity underlying selective memory for emotional scenes. Neurobiology of Aging, 34, 451–467. Webb, T. L., Miles, E., & Sheeran, P. (2012). Dealing with feeling: A meta-analysis of the effectiveness of strategies derived from the process model of emotion regulation. Psychological Bulletin, 138, 775–808. Wegner, D. M., & Vallacher, R. R. (1986). Action identification. In R. M. Sorrentino & E. T. Higgins (Eds.), Handbook of motivation and cognition: Foundations of social behavior (pp. 550–582). New York: Guilford Press. Williams, A. D., & Moulds, M. L. (2007). Cognitive avoidance of intrusive memories: Recall vantage perspective and associations with depression. Behavior Research and Therapy, 45, 1141–1153. Wilson, A., & Ross, M. (2003). The identity function of autobiographical memory: Time is on our side. Memory, 11, 137–149. Wirtz, D., Kruger, J., Scollon, C. N., & Diener, E. (2003). What to do on spring break? The role of predicted, on-line, and remembered experience in future choice. Psychological Science, 14, 520–524. Young, A. W., Aggleton, J. P., Hellawell, D. J., Johnson, M., Broks, P., & Hanley, J. R. (1995). Face processing impairments after amygdalotomy. Brain, 118, 15–24. Zlomuzica, A., Dere, D., Machulska, A., Adolph, D., Dere, E., & Margraf, J. (2014). Episodic memories in anxiety disorders: clinical implications. Frontiers in Behavioral Neuroscience, 8, 1–19.

Further reading Joormann, J., & Siemer, M. (2004). Memory accessibility, mood regulation, and dysphoria: Difficulties in repairing sad mood with happy memories? Journal of Abnormal Psychology, 113, 179–188.

A “rosy view” of the past: Positive memory biases  Chapter | 7  171 McIsaac, H. K., & Eich, E. (2004). Vantage point in traumatic memory. Psychological Science, 15, 248–253. Sharot, T., Martorella, E. A., Delgado, M. R., & Phelps, E. A. (2007). How personal experience modulates the neural circuitry of memories of September 11. Proceedings of the National Academy of Sciences, 104, 389–394.

Chapter 8

Negative memory biases in health and psychiatric disorders S.S. Grant, A.M. Huskey, J.A. Faunce, B.H. Friedman Department of Psychology, Virginia Polytechnic Institute and State University, Blacksburg, VA, United States

Introduction Cognition, emotions, and behavior are largely guided by subjective representations of past events, situations, and stimuli. Emotional memories are more likely to be recalled than neutral ones, due to numerous factors including autonomic activation and increased recruitment of attentional resources (Hamann, 2001). Cognitive prioritization and increased recall of emotional information is highly adaptive. From an evolutionary perspective, emotional events, whether positive or negative, have future relevance for survival and reproductive purposes. However, a wide body of research implicates memory biases in multiple psychiatric disorders. Memory biases may play a crucial role in the development and continuation of psychiatric symptoms. Emotion memory bias describes the tendency to selectively recall emotionor mood-congruent memories. Memory biases may be based on processes that are implicit (without conscious awareness) or explicit (conscious), automatic, or strategic. Negative memory biases (NMBs) in psychiatric disorders are often automatic and may be a central mechanism in the maintenance of emotion-related pathologies such as anxiety disorders and depression (Teasdale, 1983). In this chapter, we focus primarily on mood and anxious pathologies, because these reflect the most frequently studied psychiatric disorders pertaining to NMBs. We begin by describing some leading theories in the field of memory biases and psychiatry, followed by a discussion of methods used to study the biases. Memory processes occur in the context of mind-body interactions. Investigations have shown that high engagement of autonomic and endocrine stress responses related to emotional activation may impair memory (e.g., via heightened stress and release of cortisol; Lupien & McEwen, 1997; Reisberg & Heuer, 1995). Therefore, subsequent subsections will review literature describing physiological correlates of NMBs. Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00008-8 © 2020 Elsevier Inc. All rights reserved.

173

174  Cognitive biases in health and psychiatric disorders

The chapter concludes with a comparison of memory biases between healthy and clinical populations. Traditionally, investigations have examined groups of individuals with clinical diagnoses and compared them with “healthy” subjects, but in this subsection, we take a more dimensional approach in describing what differentiates adaptive from maladaptive memory processes. Finally, limitations and future directions are described. This survey of memory biases provides an organized synthesis of the literature, highlighting implications for future research directions and the development of new therapeutic avenues.

Major theories of affective memory biases The following theories focus on differing contributing aspects and outcomes of memory biases. While the varied accounts of NMBs in psychopathologies diverge, they are not mutually exclusive; they share a degree of overlap, which will also be discussed. Cognitive models of emotional disorders highlight various aspects of biased information processing that reinforce cognitive patterns underlying pathologies. Cognitive biases are both a product of and a contributor to emotional disorders. Common across major cognitive theories is a description of automatic negative biases in attention, interpretation, perception, and memory (Mathews & MacLeod, 2005). These cognitive processes do not operate in isolation. For example, among depressed individuals, attentional bias toward negative information leads to facilitated recall of negative information (see Chapter 9: Disner, Beevers, Haigh, & Beck, 2011; Everaert and Koster (2020)). The specific qualities of the biases in memory processes differ somewhat by psychopathology type, though the general presence of NMBs is transdiagnostic.

Semantic/associative network models One of the first theories describing selective memory in psychiatric disorders is Bower’s (1981) theory of associative/semantic networks (SNs). The theory is rooted in preexisting SN theories, which explain how semantic information (i.e., word meanings) is stored in memory (Collins & Quillian, 1972). In a cognitive SN, activation of a higher-order semantic representation enables the generation of new concepts or memories. In the case of a person with prepotent negative cognitive biases, the higher-order organizing process would be negative affect (e.g., threat and personal inadequacy), which governs the way that lower-order specifics of daily experiences are perceived and encoded. Encoding occurs when information enters the memory system from sensory input and is converted into a construct to be later recalled. In his influential work on mood congruent memory (MCM), Bower (1981) proposed that events, emotions, and concepts are represented as nodes or units within a SN (see Fig. 1). Both internal (e.g., physiological responses) and external stimuli elicit node activation, and activation of one node spreads to ­adjacent

Negative memory biases in health and psychiatric disorders  Chapter | 8  175

Expressive behaviors

Evoking appraisals

Autonomic patterns

Inhibits

Emotion 3

Verbal labels

Emotion 6

Event 19

Actors actions

Time place

FIG. 1  Theory of associative networks. Sample of the connections surrounding a specific emotion node or unit. Bidirectional arrows refer to the mutual exchange of activation between units. An inhibitory pathway from Emotion 3 to Emotion 6 is also shown. (Reproduced from Bower, G.H. (1981). Mood and memory. American Psychologist, 36(2), 129. Published by the American Psychological Association.)

nodes associated with related emotions. This process could explain why memory processes are facilitated when affective state during retrieval matches the information to be recalled. Accordingly, retrieval is more effective when mood or emotion and incoming information are congruent. Direct evidence for Bower’s SN theory is supported in studies using cognitive tasks to explore the mechanisms by which mood congruent or enhanced emotional memory takes place. In one study, a free recall task with mostly emotionally neutral items was employed (to minimize potential effects of mood induction and to more closely mimic naturalistic settings). Task performance demonstrated that participants more frequently recalled emotional items versus neutral items. Moreover, participants tended to consecutively recall or cluster items of the same emotional valence (Long, Danoff, & Kahana, 2015). Results from this study suggest that this tendency likely results from associative mechanisms (contextual encoding and retrieval mechanisms). Bower posits that NMBs occur through implicit priming of conceptual nodes associated with negative emotions. Imagine a person is dining at a restaurant when he or she receives bad news. This unfortunate news activates certain negative conceptual nodes, which then link the restaurant with negative emotions at subsequent visits. Subsequent neutral or mildly negative events at this restaurant (e.g., receiving an incorrect order) will then be interpreted more negatively (Barry, Naus, & Rehm, 2004). In this scenario, ambiguous situations or cues are thus assigned negative self-relevant meaning. A similar process is proposed to

176  Cognitive biases in health and psychiatric disorders

occur in depression, except that the cues needed to elicit the negative feelings are more generalized and the negative emotions are more likely to be elicited.

Beck’s theory of emotional disorders Beck’s theory of biased information processing in anxiety and depression is among the most influential in the field (Beck, Emery, & Greenberg, 1985; Beck, Rush, Shaw, & Emery, 1979). This theory involves schemas, which are stable cognitive structures that comprised representations of prior experiences (Mitte, 2008). Particular schemas are activated by a relevant stimulus, leading to schema-congruent processing of said stimulus. Schemas are disorder specific; those associated with anxiety deal with threat and vulnerability (Beck, Emery, et al., 1985), whereas in depression, schemas are related to loss (Beck, Rush, et al., 1979). Depression is characterized by a propensity for biased recall of negative self-relevant information (Beck, Rush, et al., 1979). The presence of this tendency is robustly supported and is one of the most prominent findings in the depression literature (Gotlib & Joormann, 2010; Mathews & MacLeod, 2005; Matt, Vázquez, & Campbell, 1992; Williams, Watts, MacLeod, & Mathews, 1997). For example, in a study of depressed and nondepressed college students, subjects were induced to focus internally or externally. Subjects were asked to recall 10 events that occurred in the prior 2 weeks. Half of the subjects were instructed to recall events that had happened to themselves, while the other half were asked to recall events that happened to other people. While nondepressed individuals tended to recall positive events, regardless of attentional focus, depressed individuals recalled negative information about themselves in the selffocus condition (Pyszczynski, Hamilton, Herring, & Greenberg, 1989). Three overarching schemas (the negative cognitive triad) are described in Beck’s theory of depression: view of self, world, and future (Beck, Rush, et al., 1979). For example, a negative self-view may engender thoughts of personal inadequacy such as “I am worthless and a failure” or “there are no likeable qualities about me.” A negative world view leads to the belief that all personal experiences result in failure or defeat and may be accompanied by thoughts such as “everyone hates me” or “every situation for me is ruled by bad luck.” A negative future view causes hopelessness for the future, and characteristic sentiments such as “nothing will ever go well for me.” Regarding memory processes specifically, depression is characterized by a propensity for biased recall of negative self-relevant information (Beck, Rush, et  al., 1979). This tendency is one of the most prominent findings in the depression literature (Gotlib & Joormann, 2010; Mathews & MacLeod, 2005; Matt et al., 1992; Williams et al., 1997). In fact, negative self-referent information processing biases prospectively predict depressive symptoms (Connolly, Abramson, & Alloy, 2016). In a study of nonclinical adolescents, participants completed a computerized self-referent encoding task (SRET; Derry & Kuiper,

Negative memory biases in health and psychiatric disorders  Chapter | 8  177

1981; Hammen & Zupan, 1984) and a measure of depressive symptoms at baseline and 9 months later. Each trial of the task involved presentation of a positive (e.g., attractive) or negative (e.g., awful) adjective above a self-referent (“Like Me?”) or structural (“Has an ‘E’?”) question. Within 8000 ms, participants selected “Yes” or “No” on the computer. Higher depressive symptoms were significantly correlated with slower response times when negative descriptors were judged not to be self-descriptive (Connolly et al., 2016). Additionally, this bias and decreased recall of endorsed positive self-referent adjectives predicted increased depressive symptoms at follow-up.

Elaboration/priming hypotheses Expanding upon prior work, Williams, Watts, MacLeod, and Mathews (1988) developed a theory that clearly differentiated among NMBs related to differing pathologies. They argued that negative cognitive biases in depression are caused by explicit assignment of additional meaning to negative information. This process of attaching meaning to a concept being encoded, called elaboration in memory research, facilitates memorization of this negative information relative to positive information. In contrast, anxious individuals shift cognitive resources away from negative threat cues, ostensibly leading to poorer encoding of threat-relevant information. This process is thought to be implicit; therefore, memories encoded from anxious attentional biases would be implicit as well. For example, anxious people may recall threat-relevant words more easily than healthy controls, without having explicit awareness of their feeling anxious (Mathews & MacLeod, 1986; Mogg, Bradley, & Hallowell, 1994). However, explicit NMBs have also been found in clinically anxious samples (Friedman, Thayer, & Borkovec, 2000), and the implicitness of NMBs in anxiety has not been consistently supported (see Williams et al., 1997, for review). The attentional effects of anxiety and depression are well documented (Hertel & El-Messidi, 2006; Koster, De Raedt, Leyman, & De Lissnyder, 2010; Mogg, Mathews, & Weinman, 1987; Sanchez, Duque, Romero, & Vazquez, 2017). Because attention affects memory (Koster et al., 2010; Tas, Luck, & Hollingworth, 2016), attentional biases predict memory biases. However, the effect of negative attentional biases in anxiety or depression on subsequent recall of pathology-related information is less clearly delineated by Williams et al. (1988). Williams et al. (1997) subsequently revised their model to distinguish between memory-based and non-memory-based elaboration (e.g., worrying; Mitte, 2008). The revised model consists of three central assumptions: (1) both encoding and retrieval involve automatic and strategic (focused) components, (2) bias in one process does not entail a bias in the other (i.e., distinct biases may occur at various stages of memory processing), and (3) different emotional states may affect different aspects of information processing (e.g., anxiety may have different effects on encoding than sadness).

178  Cognitive biases in health and psychiatric disorders

Overgeneral memory and executive functioning The following is an excerpt from a sample transcript from a psychotherapy interview in Williams (1996, pp. 245–246). A patient was asked to recall what sorts of things made him/her happy: “Therapist: When you think back, now, can you try to remember any particular time? I want you to try to recall any one of these times. Any time will do, it doesn’t have to be particularly important or special. Patient: I remember there used to be other children on the Common sometimes. Sometimes they would be friends of mine and I would stop and chat to them for a while. Therapist: Can you remember any particular time when you met any of your friends? Patient: If it was winter, there weren’t usually many people about.” Williams (1996) described an overgeneral retrieval style of negative autobiographical memories in both depression (corroborated in Raes et al., 2005) and posttraumatic stress disorder (PTSD) (Schönfeld & Ehlers, 2006). Overgeneral negative memory means that, when prompted to recall prior negative experiences, the individual may remember their feelings of distress at the time but be unable to recall specific details of the event. This phenomenon runs counter to a classical assumption in emotional memory literature that strong emotion facilitates recall (Bower, 1981). In healthy adults, negative emotion typically enhances memory accuracy (Kensinger, 2007). However, several studies show that psychopathologies may moderate the effect of emotions on memory. Theories concerning overgeneral memory bias generally propose that overgeneral memories of highly emotional negative events across psychopathologies reflect an emotion regulation strategy (1996) or a means of reducing executive demands (Conway & Pleydell-Pearce, 2000). Alternatively, Williams et al. (2007) proposed that overgeneral recall in depression results from individual differences in inhibitory control. Inhibitory dysfunction is a compelling explanation for overgeneral memory. Through an established relationship between inhibitory control and cognitive performance, the former helps explain findings that overgeneral memory predicts deficits in problem-solving and imagining future events (Gotlib & Joormann, 2010; Raes et al., 2005; Williams et al., 2007). Both severity (Williams, 1996) and duration (Raes et al., 2005) of depressive symptoms predict overgeneralized memory recall. Psychotic disorders such as schizophrenia are associated with a variety of NMBs such as memory bias for threatening information, which is associated with the presence of persecutory delusions (Lepage, Sergerie, Pelletier, & Harvey, 2007). Schizophrenia is also linked with overgeneral emotional memory bias. In a study by Neumann, Blairy, Lecompte, and Philippot (2007), participants with schizophrenia showed deficits in recalling specific details about their own previous exposure to affective pictures, particularly for negative pictures. Another empirically supported memory bias in schizophrenia is overconfidence in false memories. The liberal acceptance account (Moritz, Woodward, Jelinek, & Klinge, 2008) purports that people with schizophrenia believe they remember

Negative memory biases in health and psychiatric disorders  Chapter | 8  179

specific events with minimal evidence. Mere familiarity with the concept of a stimulus is adequate for the subject to believe that they remember the stimulus from prior experience. This bias may be a risk factor for the development of delusions and positive symptoms. In contrast, healthy control subjects tend to search available evidence before forming strong inferences about personal relevance of events (Moritz et al., 2008).

The combined cognitive bias hypothesis Recently formulated theories reflect greater appreciation for the interplay among various cognitive biases (e.g., links between expectancy biases and attention biases in fear and anxiety; Aue & Okon-Singer, 2015). Cognitive biases do not operate in isolation (see Chapter 9: Everaert and Koster (2020); Hirsch, Clark, & Mathews, 2006). As a result, increased understanding of memory biases follows from enhanced awareness of related and interacting biases. According to the combined cognitive bias hypothesis (CCBH; Hirsch et al., 2006), cognitive biases may interact with one another in contributing to negative emotion, which is often exaggerated versus either bias alone. For example, with social anxiety, self-focused negative interpretations of neutral or mildly negative social stimuli may arise from biases at initial encoding of social events (Hirsch et al., 2006), and from biased reconstruction of the events (Hackmann, Clark, & McManus, 2000). These interpretations may also become increasingly negative over time (Brendle & Wenzel, 2004). Such interpretations and recollections contribute to negative self-imagery, which significantly influences subsequent construal and memories of self-image and social scenarios. One can easily imagine how this sequence reflects a vicious cycle of interactions between interpersonal performance and self-concept. The CCBH theory can extend to other emotional problems such as dysphoria and depression (Everaert, Koster, & Derakshan, 2012). Evidence suggests a robust interplay between attention and explicit memory in dysphoria (Koster et al., 2010) and similar but more moderate effects in depression (e.g., attention bias to negative faces and negative interpretation of neutral faces; Sanchez et al., 2017). Experimentally induced self-focused attention distorts details and interpretation of memory in people with dysphoria and depression (Hertel & El-Messidi, 2006). Lastly, current interpretation biases influence memory for both past and future situations and stimuli (Tran, Hertel, & Joormann, 2011).

Methods for examining memory biases A range of laboratory tasks have been designed to investigate NMBs at the encoding, recognition, and recall stages of memory processing. Recognition is the rapid process of recognizing information as familiar and does not require deep processing. By contrast, recall involves retrieval of details about a past event, stimulus, or situation and requires deeper processing. Investigations

180  Cognitive biases in health and psychiatric disorders

e­ xamine how attentional and emotional biases influence memory formation and recall. For instance, studies of NMB formation have measured encoding biases toward specific emotions and used them to predict emotionally biased recall (Everaert, Duyck, & Koster, 2014; Everaert, Tierens, Uzieblo, & Koster, 2013; Hirsch et  al., 2006). Biases in perception, learning, and memory are transdiagnostic markers for clinical and subclinical depression and anxiety (Aue & Okon-Singer, 2015).

Measuring negative encoding biases Biases toward negative stimuli, facilitated by fast-acting attentional networks, are a potential mechanism underlying the development of NMBs. Information must be perceived and attended to to be encoded, and thus, attention represents the first phase of memory (Tas et al., 2016). This notion aligns with the CCBH, as the theory highlights the link between cognitive processes such as attention and memory. Emotional biases (distortions in cognition and decision-making due to emotional factors) can be examined in conjunction with cognitive biases (systematic errors in thinking that affect decision-making) by observing attention biases toward emotionally valenced information (e.g., positive or negative). Attention biases during encoding have been examined in laboratory settings during which individuals are asked to make rapid evaluations of stimuli often presented via a computer monitor. For example, photos of emotional faces are often used to study emotionally valenced attention and encoding biases, using response latency and/or neurophysiological activity as dependent variables. Such studies have shown that depressed participants fixate on sad faces more and happy faces less than their healthy counterparts. Depressed individuals also generally focus on emotional pictures for a longer period of time (Duque & Vázquez, 2015). Additionally, a recent metaanalysis reveals that individuals with an anxiety disorder attended longer to stimuli that were specifically threatening to them (Pergamin-Hight, Naim, Bakermans-Kranenburg, van IJzendoorn, & Bar-Haim, 2015). Meanwhile, individuals with depression display biases toward negative self-views and attributions (Phillips, Hine, & Thorsteinsson, 2010).

Measuring negative recall biases An experimental procedure, proposed along with the CCBH, allows researchers to examine the impact of attention biases on later interpretation of stimuli or situations. With this method, the impact of encoding biases can be used to predict NMBs during the recall task (Hirsch et al., 2006). For example, encoding and recall biases have been examined using the self-referent incidental recall task and the self-reference encoding task. The tasks measure affective biases toward the self, specifically, emotion words used to describe the self (Klein & Loftus, 1988). Throughout multiple trials, individuals are asked to indicate

Negative memory biases in health and psychiatric disorders  Chapter | 8  181

whether certain emotionally valenced adjectives described them or not so that researchers can examine biases toward certain self-descriptive emotions during encoding. Participants must respond within 2 s with a “yes” or “no” to examine an individual’s automatic bias toward emotional descriptions of the self. To examine the impact of affective encoding on negative biases during memory recall, individuals are asked to recall which words they remembered from the self-evaluation trials in the self-reference encoding task (Romero, Sanchez, Vázquez, & Valiente, 2016). As a measure of affective recall bias, the number of emotional words recalled (e.g., positive or negative) is compared with total words presented (e.g., Michalak, Rohde, & Troje, 2015). Recent investigations reveal that individuals with remitted major depressive disorder (MDD) still displayed greater bias toward negative words, which is suggested to contribute to the negative cognitive biases observed in depression (Romero, Sanchez, & Vazquez, 2014). Traditional associative learning paradigms may also measure slower, more deliberate responses along with the emotionally conditioned responses. After observing several presentations of neutral pictures paired with an aversive stimulus (e.g., a shock), participants are asked to rate how much they expect the aversive event to occur in the presence of the neutral picture that was paired with the shock. Because expectancies are formed in part based on prior information about the event (paired stimulus and aversive event), NMBs can be indirectly measured via this rating (Lovibond, Liu, Weidemann, & Mitchell, 2011). Measuring automatic and deliberate responses to emotionally relevant stimuli in the same study allows researchers to test for unique and interacting influences of these processes (Gawronski, Balas, & Hu, 2016). For example, individuals with PTSD show undifferentiated startle response to neutral versus threat cues, while expectation ratings illustrate that the participants were consciously aware of what conditions to expect (Jovanovic, Kazama, Bachevalier, & Davis, 2012).

Brain regions involved in the emergence of memory biases Neuroimaging studies have identified several neural regions consistently associated with memory. The limbic system supports various functions such as emotional processing and memory. Top-down control by frontal regions such as the prefrontal cortex (PFC) involves adaptive goal-directed behavior and supervision of lower-level cognitive processes. Bottom-up memory biases in depression are associated with increased activity in the amygdala, hippocampus, and anterior cingulate cortex. This increased amygdala activity is correlated with increased activation of the hippocampus, caudate, and putamen. Heightened activity in these regions has been shown to predict increased recall of negative events or information (Packard & Teather, 1998). Additionally, biased memory and rumination are linked with increased activity in the medial PFC and decreased activity in the dorsolateral PFC (Gotlib & Hamilton, 2008; Ray et al., 2005).

182  Cognitive biases in health and psychiatric disorders

A pattern of hyperactivity in limbic networks, in concert with limited top-down control, aligns with several neurocognitive theories of anxiety and depression. In additional support for the role of the limbic system in NMB, noradrenergic projections from the locus coeruleus (LC) to forebrain regions play a role in regulating attention and biased memory encoding and recall (see Sara & Bouret, 2012, for review). Norepinephrine (NE), synthesized primarily in the LC, acts in multiple brain regions to facilitate action, focused attention, and memory consolidation (Nielsen, Barber, Chai, Clewett, & Mather, 2015). LC projections to the PFC aid in attention to threat-relevant physiological arousal, whereas projections to the amygdala and hippocampus play a role in memory consolidation (Sara & Bouret, 2012). The LC is involved in threat-related encoding processes and selectively strengthens prioritized memory representations during autonomic arousal (Clewett, Huang, Velasco, Lee, & Mather, 2018). These processes underscore the importance of the LC in the development of NMBs.

Somatovisceral contributions Bodily (i.e., somatovisceral) activity in the peripheral nervous system and neuroendocrine systems are markers of perceptual encoding and retrieval of information. Not surprisingly, brain regions involved in such somatovisceral activity are similarly implicated in attention regulation, memory processing, and recall (D'esposito & Postle, 2015). Body state during memory encoding and later recall displays differences during baseline and task-related measurements between individuals with psychiatric conditions and healthy control groups (Belzung, Willner, & Philippot, 2015). Somatovisceral changes are of particular interest in the development of the NMBs linking mood disorders together (Belzung et al., 2015). Such models of information processing have been used to describe the development of posttraumatic symptoms following a traumatic event, such that trauma-related cues begin to automatically trigger similar responses in nonthreatening environments. Evidence suggests that greater emotional arousal and concomitant sympathetic activity during the encoding of an event or stimulus increases attention to that event, which is shown to increase more vivid recall later (Nielsen et al., 2015). These researchers found that increased sympathetic arousal predicted more negatively biased recall of emotional faces in young women, particularly those with lower estradiol and progesterone levels. These hormones modulate emotional memory; specifically, they decrease negative recall during heightened acute sympathetic activation. Cognitive and physiological inhibitory systems have received increased attention during the past few decades for their role in tonically regulating arousal systems (Thayer & Friedman, 2002; Thayer, Hansen, Saus-Rose, & Johnsen, 2009). Many of these perspectives developed in association with the growing finding that “frontolimbic” connectivity increases parasympathetic activity

Negative memory biases in health and psychiatric disorders  Chapter | 8  183

­ uring resting state and greater inhibitory responding during mentally demandd ing tasks (Smith, Thayer, Khalsa, & Lane, 2017). Conversely, the chronic hyperarousal states observed in PTSD and some anxiety disorders are linked with increased sympathetic arousal. According to the somatic marker hypothesis (Damasio, 1996), the arousal level an individual experiences during exposure to stimuli can in turn change the way the individual responds to the stimuli. Autonomic reactivity during encoding and retrieval of emotional information may affect neural processes mediating mood congruent memory (Critchley, Eccles, & Garfinkel, 2013). This notion is intuitive, considering that the recollection of an emotional memory involves remembering one’s emotional state (including interoceptive state) at the time of the given event. When heart rate (HR) accelerates in the absence of obvious external emotional cues, one can try to appraise which emotion contributed to the increased HR (e.g., fear or excitement). According to the hypothesis, frontal regulation plays a central role in regulating not only somatic responses in the body but also their mental representations of body state (Damasio, 1996). Increased HR, for example, may be evaluated negatively and that emotion is encoded along with the environmental cues or contexts. Repeated elevations in HR, followed by negative evaluations of related physical sensations, may contribute to the formation of generalized NMBs. For instance, mood disorders are often associated with aberrant autonomic functioning and reduced cardiac control (e.g., decreased baroreflex sensitivity; Garcia et al., 2012), which might relate to biased memory processes. One of many examples of a systematic association between autonomic responses and memory bias was demonstrated in a study concerning mood congruent memory bias in MDD (Garcia, Valenza, Tomaz, & Barbieri, 2016). In this study, high-frequency heart rate variability (a proxy measure of parasympathetic nervous system regulation of the heart) during an emotion-eliciting audiovisual stimulus was positively associated with recall performance in MDD patients; these effects were not found in healthy controls. These results highlight the role of cardiac regulation as a physiological marker of negative memory processing in depression.

Endocrine correlates Regarding pathogenic disease processes linked with NMBs, the chronic hyperarousal states observed in PTSD and some anxiety disorders are linked with increased sympathetic arousal in the body. Long-term changes in cognitive and physiological response mechanisms, linked with biased information processing, are proposed as pathogenic characteristics that facilitate disease processes within the nervous and endocrine systems (Ottaviani et  al., 2016). Specifically, the hypothalamic-pituitary-adrenal (HPA) axis is a complex neuroendocrine system that has been implicated as a necessary moderating factor of emotion regulation and immune function in humans (McVicar, Ravalier, & Greenwood, 2014).

184  Cognitive biases in health and psychiatric disorders

Sympathetic arousal co-occurs with and catalyzes the HPA’s role in increasing catecholamine activity in the heart, and resultant weakening of cardiac muscle tissue is a known precursor to the development of heart failure (Florea & Cohn, 2014). Dysregulation of HPA function is associated with increased risk for mood disorders and digestive dysfunction, particularly among women with histories of childhood sexual abuse (Stein, Yehuda, Koverola, & Hanna, 1997). Although norepinephrine (NE; neurotransmitter in the catecholamine family) is synthesized in the LC, its effector junctions are widespread across the central and peripheral nervous systems (Markovic, Anderson, & Todd, 2014). The widespread effects of NE are a major part of HPA functioning in activating the sympathetic nervous system (i.e., increased HR). Greater NE activity, indexed by increased sympathetic activity, has been linked with preferential encoding and recall biases of negative over neutral or positive information (Nielsen et al., 2015). The effect of increased sympathetic arousal on NMBs was enhanced in women with low estradiol and progesterone levels, suggesting an interactive effect between NE and sex hormones.

Similarities and differences between healthy and clinical populations In nonclinical and clinical populations alike, the development and maintenance of NMBs derive from similar physiological, emotional, and cognitive processes. Moreover, negativity bias appears to be a generalized effect. That is, there exists a broad tendency to recall negative stimuli more readily than positive stimuli, when presented with a series of affective photos (Koster, Crombez, Verschuere, & De Houwer, 2004). However, in some cases, healthy individuals show a tendency to recall positive information (i.e., positivity memory bias), whereas those with psychopathologies display NMBs (e.g., Johnson, Petzel, Hartney, & Morgan, 1983). Primarily, what appears to distinguish clinically significant NMBs from adaptive ones is the extent of bias and its appropriateness to context. For example, soldiers were found to develop fewer negative long-term effects of combat exposure when trained to prioritize (i.e., bias) threat information beforehand (Wald et al., 2017). However, carrying over prepotent negativity biases into nonthreatening contexts is not adaptive. Accordingly, threat-focused biases and related posttraumatic stress symptoms typically do not raise concerns for combat soldiers unless the symptoms continue after the soldiers return to civilian life (Prigerson, Maciejewski, & Rosenheck, 2001). Thus, clinicians and researchers must consider the context of NMBs—namely, whether they are pervasive or situation ­specific—before making assumptions about negativity bias as a stable feature of a disorder. Nevertheless, it is still worth reiterating that there are meaningful differences in NMB between controls and clinical populations. People with mood and anxiety disorders differ from healthy controls in the degree of negativity bias in recall of emotional words and other stimuli (Gotlib & Joormann, 2010; Mathews & MacLeod, 2005; Matt et al., 1992; Williams et al., 1997).

Negative memory biases in health and psychiatric disorders  Chapter | 8  185

Psychopathology does not always confer the same domain-general cognitive preference for negative stimuli. NMBs differ among psychopathologies. For example, NMBs in depression tend to be more explicit in nature and involve self-reference, compared with anxiety disorders, where NMBs are typically less explicit and more threat relevant. Additionally, some research suggests a negativity bias in word recall for depressed individuals and a positivity bias in both socially anxious and neurotypical individuals (Sanz, 1996). Moreover, negativity biases in attention and recall have been found in subclinical individuals with a family history of depression (Joormann, Talbot, & Gotlib, 2007) and bipolar disorder (Gotlib, Traill, Montoya, Joormann, & Chang, 2005), suggesting the role of negative biases in an underlying endophenotype for these disorders. If subclinical populations at risk for psychopathology display NMBs, this may warrant the analysis of psychopathology as different symptoms on a continuum rather than as discrete categories. Dimensional models are becoming increasingly popular in clinical psychology, and the NMB literature has begun to reflect this. Dimensional perspectives may help focus the pathophysiology of negative cognitive biases and aid in treatment models.

Limitations Although the literature paints a compelling picture of the biological and clinical relevance of negative cognitive bias in memory encoding and recall, much remains to be explored. One caution is the limited scope of psychopathology accounted for by current research in negativity biases. Despite the strong associations between negative cognitive biases and cognitive control issues, limited research tests the role of NMBs in disorders strongly associated with cognitive inhibitory failures, such as obsessive-compulsive disorder or bipolar disorder. Moreover, although NMB research has recently been examined in relation to severity of symptoms as opposed to diagnostic categories, this body of literature is still relatively small. Important limitations in the field of memory biases and psychiatric disorders also result from relatively isolated subfields of research. Although theories such as the CCBH have recently emerged to address this issue, these efforts are still in their relative infancy. In view of the complexity of memory processes, thorough understanding of memory bias involves converging evidence from multiple measures (e.g., clinical, behavioral, and neural). Unfortunately, few investigations or theories have incorporated methodologies or perspectives from varying fields of research.

Future directions Unanswered questions often concern the directionality of relationships between negative biases and psychiatric symptoms. Future researchers should attempt to elucidate the degree to which memory biases are caused by and/or cause psychiatric symptoms. Based on existing literature, it is likely that memory biases

186  Cognitive biases in health and psychiatric disorders

both contribute to and result from clinical disorders (Hankin, Wetter, Cheely, & Oppenheimer, 2008; Kéri & Kelemen, 2009; Tryon, 1999). Future research may help fill research gaps regarding how and under which circumstances NMBs might systematically contribute to pathologic states. This approach ultimately requires further research about aberrant memory processes in psychiatric conditions beyond depressive and anxious pathologies. Although several important questions remain concerning NMBs in MDD and generalized anxiety, knowledge and comprehensive theories about biases in conditions such as schizophrenia and bipolar disorder remain more limited. Future research must also continue to clarify pathological cognitive and physiological factors that influencing memory bias formation. This strategy includes the study of memory biases in individuals with psychiatric comorbidities, which occur commonly within psychiatric populations; however, current research in this domain is limited. For example, among individuals who are both anxious and depressed, how do memory biases operate, given the sometimes stark contrasts among these biases? Future methodological directions will likely reflect increased appreciation for the spectral nature of psychopathologies. Historically, investigations have tended to dichotomize groups (e.g., the use of healthy/pathological experimental groups); however, future work must increasingly be based on a “functional” approach to psychopathology, which emphasizes psychological dysfunctions in basic domains such as cognition and affect, rather than diagnostic labels (van Praag et al., 1990). Current NIH Research Domain Criteria (RDoC) frameworks reflect this perspective (Insel et  al., 2010). Lastly, future work must explore potential modulatory contributions of sex in memory bias research, since such research is limited.

Summary Memory biases are complex, at times adaptive, and exist across a wide spectrum of health versus psychopathology. Certain memory biases may reflect explicit or implicit emotion regulation strategies. For example, given that in some circumstances, negative biases can help inoculate against the negative consequences of subsequent trauma exposure, clinicians must consider the client’s life situation before deciding which interventions are in the client’s best interest. A wide variety of memory biases were identified in this chapter, most prominently better recognition of negatively valenced information, and overgeneral recall of negative memories. Included in these descriptions was persuasive theoretical and empirical support for the role of the biases in the generation and maintenance of pathologies. Rather than merely identifying memory biases in psychiatric disorders, the aim of this chapter is to explore these varied biases in an overall effort to (1) explicate healthy and adaptive memory processes and (2) use this knowledge to spur further investigations that may inform novel treatment approaches.

Negative memory biases in health and psychiatric disorders  Chapter | 8  187

References Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. Barry, E. S., Naus, M. J., & Rehm, L. P. (2004). Depression and implicit memory: Understanding mood congruent memory bias. Cognitive Therapy and Research, 28, 387–414. Beck, A. T., Emery, G., & Greenberg, R. L. (1985). Anxiety disorders and phobias: A cognitive perspective. New York: Basic Books. Beck, A. T., Rush, A. J., Shaw, B. F., & Emery, G. (1979). Cognitive therapy of depression (pp. 171–186). New York: Guilford Press. Belzung, C., Willner, P., & Philippot, P. (2015). Depression: From psychopathology to pathophysiology. Current Opinion in Neurobiology, 30, 24–30. Bower, G. H. (1981). Mood and memory. American Psychologist, 36, 129. Brendle, J. R., & Wenzel, A. (2004). Differentiating between memory and interpretation b­ iases in socially anxious and nonanxious individuals. Behaviour Research and Therapy, 42, 155–171. Clewett, D. V., Huang, R., Velasco, R., Lee, T. H., & Mather, M. (2018). Locus coeruleus activity strengthens prioritized memories under arousal. Journal of Neuroscience, 38(6), 1558–1574. Collins, A. M., & Quillian, M. R. (1972). Experiments on semantic memory and language comprehension. In L. W. Gregg (Ed.), Cognition in learning and memory. Oxford, England: John Wiley & Sons. Connolly, S. L., Abramson, L. Y., & Alloy, L. B. (2016). Information processing biases concurrently and prospectively predict depressive symptoms in adolescents: Evidence from a self-referent encoding task. Cognition and Emotion, 30(3), 550–560. Conway, M. A., & Pleydell-Pearce, C. W. (2000). The construction of autobiographical memories in the self-memory system. Psychological Review, 107, 261–288. Critchley, H. D., Eccles, J., & Garfinkel, S. N. (2013). Interaction between cognition, emotion, and the autonomic nervous system. Handbook of Clinical Neurology, 117, 59–77. Damasio, A. R. (1996). The somatic marker hypothesis and the possible functions of the prefrontal cortex. Philosophical Transactions of the Royal Society B, 351, 1413–1420. Derry, P. A., & Kuiper, N. A. (1981). Schematic processing and self-reference in clinical depression. Journal of Abnormal Psychology, 90(4), 286. D’esposito, M., & Postle, B. R. (2015). The cognitive neuroscience of working memory. Annual Review of Psychology, 66, 115–142. Disner, S. G., Beevers, C. G., Haigh, E. A., & Beck, A. T. (2011). Neural mechanisms of the cognitive model of depression. Nature Reviews Neuroscience, 12, 467–477. Duque, A., & Vázquez, C. (2015). Double attention bias for positive and negative emotional faces in clinical depression: Evidence from an eye-tracking study. Journal of Behavior Therapy and Experimental Psychiatry, 46, 107–114. Everaert, J., Duyck, W., & Koster, E. H. (2014). Attention, interpretation, and memory biases in subclinical depression: A proof-of-principle test of the combined cognitive biases hypothesis. Emotion, 14(2), 331. Everaert, J., & Koster, E. H. W. (2020). The interplay among attention, interpretation, and memory ­biases in depression: Revisiting the combined cognitive bias hypothesis. In T. Aue & H. Okon-Singer (Eds.), Cognitive biases in health and psychiatric disorders: Neurophysiological ­foundations (pp. 193–213). Amsterdam, Netherlands: Elsevier. Everaert, J., Koster, E. H., & Derakshan, N. (2012). The combined cognitive bias hypothesis in depression. Clinical Psychology Review, 32, 413–424.

188  Cognitive biases in health and psychiatric disorders Everaert, J., Tierens, M., Uzieblo, K., & Koster, E. H. (2013). The indirect effect of attention bias on memory via interpretation bias: Evidence for the combined cognitive bias hypothesis in subclinical depression. Cognition & Emotion, 27(8), 1450–1459. Florea, V. G., & Cohn, J. N. (2014). The autonomic nervous system and heart failure. Circulation Research, 114, 1815–1826. Friedman, B. H., Thayer, J. F., & Borkovec, T. D. (2000). Explicit memory bias for threat words in generalized anxiety disorder. Behavior Therapy, 31, 745–756. Garcia, R. G., Valenza, G., Tomaz, C., & Barbieri, R. (2016). Relationship between cardiac vagal activity and mood congruent memory bias in major depression. Journal of Affective Disorders, 190, 19–25. Garcia, R. G., Zarruk, J. G., Guzman, J. C., Barrera, C., Pinzon, A., Trillos, E., et al. (2012). Sex differences in cardiac autonomic function of depressed young adults. Biological Psychology, 90(3), 179–185. Gawronski, B., Balas, R., & Hu, X. (2016). What do I know about it versus how do I feel about it? A theoretical analysis of response processes in evaluative conditioning. Social Cognition, 34, 357–381. Gotlib, I. H., & Hamilton, J. P. (2008). Neuroimaging and depression: Current status and unresolved issues. Current Directions in Psychological Science, 17, 159–163. Gotlib, I. H., & Joormann, J. (2010). Cognition and depression: Current status and future directions. Annual Review of Clinical Psychology, 6, 285–312. Gotlib, I. H., Traill, S. K., Montoya, R. L., Joormann, J., & Chang, K. (2005). Attention and memory biases in the offspring of parents with bipolar disorder: Indications from a pilot study. Journal of Child Psychology and Psychiatry, 46(1), 84–93. Hackmann, A., Clark, D. M., & McManus, F. (2000). Recurrent images and early memories in social phobia. Behavior Research & Therapy, 38, 601–610. Hamann, S. (2001). Cognitive and neural mechanisms of emotional memory. Trends in Cognitive Sciences, 5, 394–400. Hammen, C., & Zupan, B. A. (1984). Self-schemas, depression, and the processing of personal information in children. Journal of Experimental Child Psychology, 37(3), 598–608. Hankin, B. L., Wetter, E., Cheely, C., & Oppenheimer, C. W. (2008). Beck’s cognitive theory of depression in adolescence: Specific prediction of depressive symptoms and reciprocal influences in a multi-wave prospective study. International Journal of Cognitive Therapy, 1(4), 313–332. Hertel, P. T., & El-Messidi, L. (2006). Am I blue? Depressed mood and the consequences of selffocus for the interpretation and recall of ambiguous words. Behavior Therapy, 37, 259–268. Hirsch, C. R., Clark, D. M., & Mathews, A. (2006). Imagery and interpretations in social phobia: Support for the combined cognitive biases hypothesis. Behavior Therapy, 37, 223–236. Insel, T. R., Cuthbert, B. N., Garvey, M. A., Heinssen, R. K., Pine, D. S., Quinn, K. J., et al. (2010). Research domain criteria (RDoC): Toward a new classification framework for research on mental disorders. American Journal of Psychiatry, 167, 748–751. Johnson, J. E., Petzel, T. P., Hartney, L. M., & Morgan, L. M. (1983). Recall and importance ratings of completed and uncompleted tasks as a function of depression. Cognitive Therapy and Research, 7, 51–56. Joormann, J., Talbot, L., & Gotlib, I. H. (2007). Biased processing of emotional information in girls at risk for depression. Journal of Abnormal Psychology, 116, 135–143. Jovanovic, T., Kazama, A., Bachevalier, J., & Davis, M. (2012). Impaired safety signal learning may be a biomarker of PTSD. Neuropharmacology, 62, 695–704. Kensinger, E. A. (2007). Negative emotion enhances memory accuracy: Behavioral and neuroimaging evidence. Current Directions in Psychological Science, 16(4), 213–218.

Negative memory biases in health and psychiatric disorders  Chapter | 8  189 Kéri, S., & Kelemen, O. (2009). The role of attention and immediate memory in vulnerability to interpersonal criticism during family transactions in schizophrenia. British Journal of Clinical Psychology, 48(1), 21–29. Klein, S. B., & Loftus, J. (1988). The nature of self-referent encoding: The contributions of elaborative and organizational processes. Journal of Personality and Social Psychology, 55, 5–11. Koster, E. H., Crombez, G., Verschuere, B., & De Houwer, J. (2004). Selective attention to threat in the dot probe paradigm: Differentiating vigilance and difficulty to disengage. Behaviour Research and Therapy, 42, 1183–1192. Koster, E. H., De Raedt, R., Leyman, L., & De Lissnyder, E. (2010). Mood-congruent attention and memory bias in dysphoria: Exploring the coherence among information-processing biases. Behaviour Research and Therapy, 48, 219–225. Lepage, M., Sergerie, K., Pelletier, M., & Harvey, P. O. (2007). Episodic memory bias and the symptoms of schizophrenia. The Canadian Journal of Psychiatry, 52, 702–709. Long, N. M., Danoff, M. S., & Kahana, M. J. (2015). Recall dynamics reveal the retrieval of emotional context. Psychonomic Bulletin & Review, 22(5), 1328–1333. Lovibond, P. F., Liu, J. C., Weidemann, G., & Mitchell, C. J. (2011). Awareness is necessary for differential trace and delay eyeblink conditioning in humans. Biological Psychology, 87, 393–400. Lupien, S. J., & McEwen, B. S. (1997). The acute effects of corticosteroids on cognition: Integration of animal and human model studies. Brain Research Reviews, 24, 1–27. Markovic, J., Anderson, A. K., & Todd, R. M. (2014). Tuning to the significant: Neural and genetic processes underlying affective enhancement of visual perception and memory. Behavioural Brain Research, 259, 229–241. Mathews, A., & MacLeod, C. (1986). Discrimination of threat cues without awareness in anxiety states. Journal of Abnormal Psychology, 95(2), 131. Mathews, A., & MacLeod, C. (2005). Cognitive vulnerability to emotional disorders. Annual Reviews in Clinical Psychology, 1, 167–195. Matt, G. E., Vázquez, C., & Campbell, W. K. (1992). Mood-congruent recall of affectively toned stimuli: A meta-analytic review. Clinical Psychology Review, 12, 227–255. McVicar, A., Ravalier, J. M., & Greenwood, C. (2014). Biology of stress revisited: Intracellular mechanisms and the conceptualization of stress. Stress and Health, 30, 272–279. Michalak, J., Rohde, K., & Troje, N. F. (2015). How we walk affects what we remember: Gait modifications through biofeedback change negative affective memory bias. Journal of Behavior Therapy and Experimental Psychiatry, 46, 121–125. Mitte, K. (2008). Memory bias for threatening information in anxiety and anxiety disorders: A meta-analytic review. Psychological Bulletin, 134, 886–911. Mogg, K., Bradley, B. P., & Hallowell, N. (1994). Attentional bias to threat: Roles of trait anxiety, stressful events, and awareness. The Quarterly Journal of Experimental Psychology Section A, 47(4), 841–864. Mogg, K., Mathews, A., & Weinman, J. (1987). Memory bias in clinical anxiety. Journal of Abnormal Psychology, 96, 94–98. Moritz, S., Woodward, T. S., Jelinek, L., & Klinge, R. (2008). Memory and metamemory in schizophrenia: A liberal acceptance account of psychosis. Psychological Medicine, 38, 825–832. Neumann, A., Blairy, S., Lecompte, D., & Philippot, P. (2007). Specificity deficit in the recollection of emotional memories in schizophrenia. Consciousness and Cognition, 16, 469–484. Nielsen, S. E., Barber, S. J., Chai, A., Clewett, D. V., & Mather, M. (2015). Sympathetic arousal increases a negative memory bias in young women with low sex hormone levels. Psychoneuroendocrinology, 62, 96–106.

190  Cognitive biases in health and psychiatric disorders Ottaviani, C., Thayer, J. F., Verkuil, B., Lonigro, A., Medea, B., Couyoumdjian, A., et al. (2016). Physiological concomitants of perseverative cognition: A systematic review and meta-analysis. Psychological Bulletin, 142, 231–259. Packard, M. G., & Teather, L. A. (1998). Amygdala modulation of multiple memory systems: Hippocampus and caudate-putamen. Neurobiology of Learning and Memory, 69(2), 163–203. Pergamin-Hight, L., Naim, R., Bakermans-Kranenburg, M. J., van IJzendoorn, M. H., & Bar-Haim, Y. (2015). Content specificity of attention bias to threat in anxiety disorders: A meta-analysis. Clinical Psychology Review, 35, 10–18. Phillips, W. J., Hine, D. W., & Thorsteinsson, E. B. (2010). Implicit cognition and depression: A meta-analysis. Clinical Psychology Review, 30, 691–709. Prigerson, H. G., Maciejewski, P. K., & Rosenheck, R. A. (2001). Combat trauma: Trauma with highest risk of delayed onset and unresolved posttraumatic stress disorder symptoms, unemployment, and abuse among men. The Journal of Nervous and Mental Disease, 189(2), 99–108. Pyszczynski, T., Hamilton, J. C., Herring, F. H., & Greenberg, J. (1989). Depression, self-focused attention, and the negative memory bias. Journal of Personality and Social Psychology, 57(2), 351. Raes, F., Hermans, D., Williams, J. M. G., Demyttenaere, K., Sabbe, B., Pieters, G., et al. (2005). Reduced specificity of autobiographical memory: A mediator between rumination and ineffective social problem-solving in major depression? Journal of Affective Disorders, 87, 331–335. Ray, R. D., Ochsner, K. N., Cooper, J. C., Robertson, E. R., Gabrieli, J. D., & Gross, J. J. (2005). Individual differences in trait rumination and the neural systems supporting cognitive reappraisal. Cognitive, Affective, & Behavioral Neuroscience, 5, 156–168. Reisberg, D., & Heuer, F. (1995). Emotion’s multiple effects on memory. In J. L.  McGaugh, N. M. Weinberger, & G. Lynch (Eds.), Brain and memory: Modulation and mediation of neuroplasticity (pp. 84–92). New York, NY: Oxford University Press. Romero, N., Sanchez, A., & Vazquez, C. (2014). Memory biases in remitted depression: The role of negative cognitions at explicit and automatic processing levels. Journal of Behavior Therapy and Experimental Psychiatry, 45(1), 128–135. Romero, N., Sanchez, A., Vázquez, C., & Valiente, C. (2016). Explicit self-esteem mediates the relationship between implicit self-esteem and memory biases in major depression. Psychiatry Research, 242, 336–344. Sanchez, A., Duque, A., Romero, N., & Vazquez, C. (2017). Disentangling the interplay among cognitive biases: Evidence of combined effects of attention, interpretation and autobiographical memory in depression. Cognitive Therapy and Research, 41(6), 829–841. Sanz, J. (1996). Memory bias in social anxiety and depression. Cognition and Emotion, 10, 87–105. Sara, S. J., & Bouret, S. (2012). Orienting and reorienting: The locus coeruleus mediates cognition through arousal. Neuron, 76(1), 130–141. Schönfeld, S., & Ehlers, A. (2006). Overgeneral memory extends to pictorial retrieval cues and correlates with cognitive features in posttraumatic stress disorder. Emotion, 6(4), 611–621. Smith, R., Thayer, J. F., Khalsa, S. S., & Lane, R. D. (2017). The hierarchical basis of neurovisceral integration. Neuroscience & Biobehavioral Reviews, 75, 274–296. Stein, M. B., Yehuda, R., Koverola, C., & Hanna, C. (1997). Enhanced dexamethasone suppression of plasma cortisol in adult women traumatized by childhood sexual abuse. Biological Psychiatry, 42, 680–686. Tas, A. C., Luck, S. J., & Hollingworth, A. (2016). The relationship between visual attention and visual working memory encoding: A dissociation between covert and overt orienting. Journal of Experimental Psychology: Human Perception and Performance, 42, 1121–1138.

Negative memory biases in health and psychiatric disorders  Chapter | 8  191 Teasdale, J. D. (1983). Negative thinking in depression: Cause, effect, or reciprocal relationship? Advances in Behaviour Research and Therapy, 5, 3–25. Thayer, J. F., & Friedman, B. H. (2002). Stop that! Inhibition, sensitization, and their neurovisceral concomitants. Scandinavian Journal of Psychology, 43, 123–130. Thayer, J. F., Hansen, A. L., Saus-Rose, E., & Johnsen, B. H. (2009). Heart rate variability, prefrontal neural function, and cognitive performance: The neurovisceral integration perspective on self-regulation, adaptation, and health. Annals of Behavioral Medicine, 37, 141–153. Tran, T. B., Hertel, P. T., & Joormann, J. (2011). Cognitive bias modification: Induced interpretive biases affect memory. Emotion, 11, 145–152. Tryon, W. W. (1999). A bidirectional associative memory explanation of posttraumatic stress ­disorder. Clinical Psychology Review, 19(7), 789–818. van Praag, H. V., Asnis, G. M., Kahn, R. S., Brown, S. L., Korn, M., Friedman, J. H., et al. (1990). Nosological tunnel vision in biological psychiatry. Annals of the New York Academy of ­Sciences, 600(1), 501–510. Wald, I., Bitton, S., Levi, O., Zusmanovich, S., Fruchter, E., Ginat, K., et al. (2017). Acute delivery of attention bias modification training (ABMT) moderates the association between combat exposure and posttraumatic symptoms: A feasibility study. Biological Psychology, 122, 93–97. Williams, J. M. G. (1996). Depression and the specificity of autobiographical memory. In D. C. Rubin (Ed.), Remembering our past: Studies in autobiographical memory (pp. 244–267). Cambridge: Cambridge University Press. Williams, J. M. G., Barnhofer, T., Crane, C., Herman, D., Raes, F., Watkins, E., et al. (2007). Autobiographical memory specificity and emotional disorder. Psychological Bulletin, 133(1), 122. Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1988). Cognitive psychology and emotional disorders. Oxford, England: John Wiley & Sons. Williams, J. M. G., Watts, F. N., MacLeod, C., & Mathews, A. (1997). Cognitive psychology emotional disorders (2nd ed.). Chichester, United Kingdom: Wiley.

Further reading Joormann, J. (2018). Is the glass half empty or half full and does it even matter? Cognition, emotion, and psychopathology. Cognition and Emotion, 33(1), 133–138. Kellough, J. L., Beevers, C. G., Ellis, A. J., & Wells, T. T. (2008). Time course of selective attention in clinically depressed young adults: An eye tracking study. Behaviour Research and Therapy, 46(11), 1238–1243. MacLeod, C., & McLaughlin, K. (1995). Implicit and explicit memory bias in anxiety: A conceptual replication. Behaviour Research and Therapy, 33(1), 1–14.

Chapter 9

The interplay among attention, interpretation, and memory biases in depression: Revisiting the combined cognitive bias hypothesis Jonas Everaert, Ernst H.W. Koster Department of Experimental-Clinical and Health Psychology, Ghent University, Ghent, Belgium

Introduction Depression is a prevalent and recurrent mental disorder causing a severe personal and societal burden (Kessler & Bromet, 2013). Identifying the mechanisms involved in depression is an integral part of efforts to improve prevention and treatment strategies for this burdensome disorder. Cognitive theories posit that depression symptoms are caused in part by negative biases in the processing of emotional information (Clark, Beck, & Alford, 1999; Ingram, 1984; Williams et al., 1997). Consistent with this hypothesis, extensive research has linked depression to negative biases in cognitive processes such as attention, interpretation, and memory. Specifically, empirical studies have found that (sub) clinically depressed individuals may exhibit an attention bias toward negative self-relevant information (Armstrong & Olatunji, 2012; Peckham, McHugh, & Otto, 2010; Winer & Salem, 2016; but see also Rodebaugh et  al., 2016), an interpretation bias favoring negative explanations for ambiguous situations (Everaert, Podina, & Koster, 2017), and a memory bias featuring improved recollection of negative self-referential information (Gaddy & Ingram, 2014; Matt, Vázquez, & Campbell, 1992). Importantly, research suggests that biases of attention, interpretation, and memory may influence symptoms of depression (Hallion & Ruscio, 2011; Menne-Lothmann et al., 2014; Mogoaşe, David, & Koster, 2014; Vrijsen, Hertel, & Becker, 2016) and predict their longitudinal course (Johnson, Joormann, & Gotlib, 2007; Price et al., 2016; Rude, DurhamFowler, Baum, Rooney, & Maestas, 2010). Together, current research findings Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00009-X © 2020 Elsevier Inc. All rights reserved.

193

194  Cognitive biases in health and psychiatric disorders

indicate that attention, interpretation, and memory biases are not merely a byproduct of depressed mood but may confer risk to experiencing depression. While attention, interpretation, and memory biases have been investigated extensively in isolation, the interplay among these presumed risk factors has received only modest consideration. However, uncovering how these cognitive biases work together seems crucial to gain a comprehensive understanding of the cognitive mechanisms of maladaptation involved in depression. Indeed, researchers have repeatedly argued that it is unlikely that the heterogeneous symptoms of depression (e.g., sustained negative mood and anhedonia) would be caused or maintained by cognitive biases that operate in isolation (e.g., Hankin, 2012; Kraemer, Stice, Kazdin, Offord, & Kupfer, 2001; Wittenborn, Rahmandad, Rick, & Hosseinichimeh, 2016). Instead, theorists have proposed that it is much more likely that depression involves multiple causal chains involving cognitive biases that reinforce each other through mutual influences. This notion has been referred to as the combined cognitive bias hypothesis (CCBH; Hirsch, Clark, & Mathews, 2006). The CCBH, as originally formulated by Hirsch and colleagues, specifically states: “cognitive biases do not operate in isolation, but rather can influence each other and/or can interact so that the impact of each on another variable is influenced by the other. Via both these mechanisms we argue that combinations of biases have a greater impact on disorders than if individual cognitive processes acted in isolation” (p. 224; Hirsch et al., 2006). Although the CCBH was formulated in the context of social anxiety disorder, the hypothesis can be applied to other forms of psychopathology. Elaborating on the CCBH in the context of depression in a prior review (Everaert, Koster, & Derakshan, 2012), we outlined three broad categories of open research questions that required empirical consideration to further our understanding of how cognitive biases operate in the etiology, maintenance, and relapse depression. Specifically, we distinguished association, causal, and predictive magnitude questions. With association questions, we referred to research questions focusing on how biases of attention, interpretation, and memory are correlated across different stages of information processing (e.g., during encoding or retrieval of emotional information). Example association questions may concern whether negative attention bias during encoding is associated with improved memory for previously presented negative material or whether negative memory bias is related to attention biases toward matching emotional material. The second category of CCBH questions, the causal questions, focuses on the direction of the hypothesized influence of one cognitive bias on another bias. These questions concern unidirectional and bidirectional influences that may exist between two given cognitive biases. An example of a causal question is whether attention biases cause a congruent bias in interpretation and whether an interpretation bias in turn influences attention allocation toward stimuli that are congruent with the emotional interpretations. Finally, the third category of questions, the predictive magnitude questions, focuses on how multiple cognitive biases in concert influence the symptom course of depression over longer periods of time. Extending association and causal

Revisiting the combined cognitive bias hypothesis  Chapter | 9  195

­ uestions, predictive magnitude questions focus on the utility of a single cognitive q bias vs multiple cognitive biases in combination in predicting prospective changes in depression. For example, predictive magnitude questions may address whether cognitive biases have additive effects on depressive symptoms that extend beyond the isolated effect of each bias (see in the succeeding text). Recent years have witnessed an important upsurge of empirical studies addressing different aspects of the different CCBH questions in various forms of psychopathology. In light of the advances in research on the CCBH in depression, the purpose of this chapter is to review recent findings and both theoretical and methodological innovations and update our review article (Everaert et al., 2012). We think that an updated review of theory and research on the CCBH in depression is both timely and necessary given the increasingly complex picture that is arising from the empirical research examining interactions among cognitive biases in depression. Later, we start by describing theoretical contributions that can inform upon the interplay among cognitive biases in depression. Then, we discuss the major methods that have been used to investigate the association, causal, and predictive magnitude questions that originate from the CCBH. Next, we review findings from recent empirical work with respect to the different CCBH questions. Finally, we discuss limitations of current research on this topic and propose several ways in which this exciting area of research can be taken forward.

Major theories in the field Influential theoretical models of depression such as Beck’s schema theory (Beck & Haigh, 2014; Clark et  al., 1999), enhanced elaboration accounts (Ingram, 1984; Williams et al., 1997), and cognitive control accounts (Hertel, 1997; Joormann, Yoon, & Zetsche, 2007) have attributed a crucial role to cognitive biases in the etiology and maintenance of depressive symptoms. These dominant theoretical models have guided seminal research and led to important discoveries regarding the role of cognitive biases as (causal) risk factors for depression (for reviews, see Gotlib & Joormann, 2010; Mathews & Macleod, 2005). However, contemporary theoretical accounts often propose a number of cognitive biases without providing a detailed account of how these processes may influence one another and in concert influence the course of depression (for a review, see Everaert et al., 2012). It is only recently that specific ideas and hypotheses regarding the interplay between cognitive biases have begun to emerge in clinical research (Aue & Okon-Singer, 2015; Everaert et al., 2012; Hertel, 2004; Hertel & Brozovich, 2010; Hirsch et al., 2006; Wittenborn et al., 2016). In this section, we discuss novel conceptual contributions that have attempted to describe the interplay among cognitive biases in depression in a comprehensive manner. A discussion of the shared and unique predictions by traditional cognitive models with respect to the different CCBH questions can be found elsewhere (see Everaert et al., 2012).

196  Cognitive biases in health and psychiatric disorders

The causal loop diagram of depression dynamics Wittenborn and colleagues recently proposed a causal loop diagram integrating cognitive, social, and environmental factors that may explain the etiology of depression (Wittenborn et al., 2016). Of particular relevance to the CCBH, this model specifies a reinforcing feedback loop involving attention, interpretation, and memory biases in the consolidation of negative cognitions. The model proposes that negative cognitive representations that are stored in long-term memory direct attention toward relevant information. Specifically, negative memory representations are hypothesized to both orient and maintain attention on negative material in the environment that matches the content of the memory representations. The resulting negative bias of attention is expected to increase one’s perceived stress level and produce negatively biased interpretations of the situation. This enhanced processing of negative material through biases of attention and interpretation is in turn expected to set the stage for increased negative affect and improved encoding of negative material into memory. This further consolidates the initial negative memory representations, which may in turn guide attention toward congruent information, etc. By defining a pathway with memory bias causing biases of attention and interpretation that in turn fuel memory bias, this model advocates the view that cognitive biases are highly interactive and interdependent processes that cannot be fully understood when studied in isolation.

The attention-memory bias-interaction research framework In an attempt to organize existing research and guide future empirical inquiry, we recently formulated a conceptual framework of interactions between attention and memory biases in psychopathology (Everaert, Bernstein, Joormann, & Koster, 2018). Grounded in basic cognitive research on interactions between attention and memory (for excellent review articles, see Awh, Vogel, & Oh, 2006; Chun, Golomb, & Turk-Browne, 2011; Hutchinson & Turk-Browne, 2012; Todd & Manaligod, 2017), the framework proposes several theoretical predictions about causal pathways between attention and memory biases at different stages of information processing. Specifically, the framework proposes that attention bias improves memory for negative material by influencing both encoding and retrieval of emotional material. That is, attentional bias may skew the processing of emotional material in favor of negative information, which increases the probability that negative material is encoded into memory and alters what is available for later recollection. In addition, attentional biases may also enhance memory biases after encoding by altering the retrieval of stored emotional items. In remembering emotional experiences, attention bias may influence which cues are used to guide memory search to retrieve details of a past event. This attention bias during memory search is expected to increase the likelihood of remembering negative memories. In addition to the role of attention bias in influencing emotional memory, the framework proposes that memory biases may guide attention biases toward matching emotional material.

Revisiting the combined cognitive bias hypothesis  Chapter | 9  197

Such a memory-guided attention, bias may occur because someone’s emotional learning history alters the attentional priority of certain cues in the environment. Alternatively, emotional memories can also be retrieved consciously when searching for relevant emotional information. Thus, the framework hypothesizes that attention bias influences and is influenced by memory bias. One important prediction of the framework is that the interactions between attention and memory biases unfold dynamically over time. Indeed, the model considers attention and memory biases as dynamic processes with mutually reinforcing influences. That is, memory biases may shape biases of attention, which may in turn influence what is processed and stored in memory, which again guides attention toward matching emotional information, and so on. Through mutual influences, attention and memory biases are hypothesized to instigate pathogenic cycles of increasingly negative information processing in depression. Critically, the framework argues that cognitive biases may not operate in a stable manner but are better conceptualized as processes that fluctuate over time and across contexts (e.g., before, during, and after stressful episodes). This implies that also the mutual influences between cognitive biases may change over time (e.g., stronger connections during stress). Therefore, the framework advocates the view that quantifying the dynamic nature of cognitive biases is key to understanding the intricate interplay between attention and memory biases.

Conceptualizing the combined influence of cognitive biases on depression over time While recent conceptual contributions predominantly focus on how cognitive biases may interact (i.e., association and causal questions), they do not elaborate extensively on how cognitive biases may combine to influence the course of depression symptoms (i.e., predictive magnitude questions). In this respect, frameworks from research on risk factors in psychiatry (Kraemer et al., 2001) and cognitive content factors in depression (Abela & Hankin, 2008) may help to understand the combined impact of cognitive biases on depression. In their seminal article, Kraemer et al. (2001) elaborated on five conceptually different ways in which risk factors may work together to influence an outcome. Putative risk factors may operate as proxy, overlapping, or independent risk factors as well as mediators or moderators. When applied to cognitive biases, several hypotheses can be formulated regarding how attention, interpretation, and memory biases may work together to influence symptoms of depression. A first possibility is that attention, interpretation, and memory biases are independent risk factors. This means that each cognitive bias has a unique relationship with depression symptoms and exerts an influence on depression independent from other cognitive biases. To qualify as independent risk factors, there should be no temporal precedence of different cognitive biases, and cognitive biases should not be correlated. A second hypothesis is that a cognitive bias may operate as a

198  Cognitive biases in health and psychiatric disorders

proxy risk factor for another cognitive bias. This may occur when one cognitive bias turns out to be a relatively stronger risk factor for depressive symptoms, and any of its correlating cognitive biases also appear to be a risk factor for depression (through the strong bias that is correlated with both). For example, through its correlation with memory bias, attention bias may be a proxy for memory bias as a risk factor for depression. The third hypothesis concerns a mediation model in which a cognitive bias intervenes in the relation between another cognitive bias and depressive symptoms. This means that the mediating cognitive bias explains how another cognitive bias influences depression. For example, it is possible that attention bias impacts depression symptoms such as sadness through its influence on interpretation bias. A fourth hypothesis about the combined impact of cognitive biases concerns a moderation model. It is possible that one cognitive bias may moderate effects of another cognitive bias on depression. A moderating cognitive bias determines under what conditions another cognitive bias operates to produce depression symptoms. For example, attention bias may affect the relationship between memory bias and depression, such that greater negative attention bias increases the potency of memory bias to generate depressive symptoms. A final hypothesis is that multiple cognitive biases could be overlapping risk factors in predicting depression. This possibility is likely when cognitive biases are correlated without temporal precedence and have unique relationships with depression symptoms. For example, interpretation and memory bias may be overlapping risk factors through the shared ­content of negative cognitions resulting from the bias. Given the mutual relations among attention, interpretation, and memory biases (see in the succeeding text for an overview of empirical research), several of these hypothetical interactions (e.g., overlapping risk factor and mediation effect) could be appropriate to characterize how cognitive biases work together to influence depression. In addition to examining how cognitive biases work together in influencing depressive symptoms, it is of importance to determine which (combination of) biases yield the greatest potency in predicting the symptom course over time. In this respect, research examining the predictive power of multiple depression-linked distortions in cognitive content factors may inform how to conceptualize and quantify the combined impact of cognitive biases. This research has frequently used additive and weakest link models to integrate multiple indices of risk factors (Abela & Hankin, 2008; Reilly, Ciesla, Felton, Weitlauf, & Anderson, 2012). The additive approach assumes that the severity of distorted cognitive factors has a cumulative effect, such that risk to develop depressive symptoms increases with each additional factor. Applied to cognitive biases, the model predicts that individuals with more severe negative biases in multiple processes are at greater risk to develop depressive symptoms compared with individuals with fewer negative biases. Alternatively, the weakest link approach predicts that the course of depressive symptoms depends on the most pathogenic cognitive factor and not on the number of factors. The best marker of future increases in depressive symptoms would then be the cognitive process that is dominantly biased toward negative

Revisiting the combined cognitive bias hypothesis  Chapter | 9  199

­ aterial. While these models may not specify how biases work together, they m provide useful integrative indices to conceptualize their combined predictive magnitude.

Methods used to investigate the CCBH Association questions In studies examining association questions, investigators have utilized two different methodological approaches. A first approach consists of combining experimental paradigms to measure multiple cognitive biases in a single study. For example, in a study by Gotlib et al. (2004), participants started with an encoding task of emotional and neutral words which was followed by a free recall task to measure memory bias for emotional words. Next, to measure attention biases, participants completed a dot-probe task in which emotional-neutral face pairs were presented. Specific to this first approach to study CCBH association questions is that the utilized experimental tasks are unrelated in that they present their own unique stimulus materials. Given this independence of the experimental tasks, this type of studies can address questions regarding the correlations among cognitive biases. For example, these studies may shed light on whether negative biases occur in one or multiple cognitive processes independently (Vrijsen, Van Oostrom, Isaac, Becker, & Speckens, 2014). Yet, a limitation of this type of CCBH studies is that the study design does not provide insight into specific pathways of how cognitive biases may work together. While hypothesized relations among attention, interpretation, and memory biases can be modeled statistically (e.g., see Sanchez, Duque, Romero, & Vazquez, 2017), it cannot be tested formally how emotional material is modulated by the different biases across different stages of information processing. To answer such association questions, researchers have adopted a second methodological approach. The second approach also combines multiple experimental tasks or measures of cognitive biases in a single study but modifies the tasks so that they use the same stimulus materials. These studies typically present the experimental tasks in a fixed temporal order to examine how different cognitive biases during encoding and/or retrieval of the emotional material are related to later processing of this stimulus material. For example, by presenting an attentional bias task followed by a memory task probing recall of the presented stimuli, it is possible to examine whether attention bias skews encoding in favor of negative information and how this is associated with a negative memory bias. By using the same stimulus materials across tasks, these studies enable the investigation of specific aspects of the potential interplay among cognitive biases in a controlled experimental context. For example, we recently examined whether attention bias modulates the interpretation of emotional material and subsequently alters emotional memory (Everaert, Duyck, & Koster, 2014). To this end, participants unscrambled emotional sentences in either a positive or

200  Cognitive biases in health and psychiatric disorders

FIG. 1  Example design to study the interplay among attention, interpretation, and memory biases.

a negative manner (e.g., “born winner am loser a I” into either “I am a born loser” or “I am a born winner”) to measure interpretation bias. When unscrambling the sentences, biases of attention toward negative (e.g., “loser”) vs positive (e.g., “winner”) words were measured using eye tracking. A subsequent free recall task prompted participants to recollect the constructed interpretations and served as a measure of memory bias. Fig. 1 depicts the task procedure. Indeed, the designed dependency among the different cognitive bias and measures allowed this study to track how the same emotional material was treated by the different biases. It is then possible to gain insight into potential pathways from attention bias to memory bias. At present, several promising paradigms have been developed to study the dependency among cognitive biases at different stages of information processing (see Everaert & Koster, 2015; Salem, Winer, & Nadorff, 2017; Wells, Beevers, Robison, & Ellis, 2010). However, we think that this field of research could benefit from the development of novel research paradigms to address open research questions. Indeed, many aspects of the interplay among cognitive biases have yet to be investigated such as the role of memory bias in guiding attention toward congruent information. In this respect, paradigms from basic cognitive science may provide ways forward (see Everaert et al., 2018).

Causal questions Though certain features of some cross-sectional study designs optimize conditions to examine the effect of one cognitive bias on another process (e.g., temporal order of tasks and similar stimulus materials across tasks), third variables could account for the observed relations. Direct proof of causality

Revisiting the combined cognitive bias hypothesis  Chapter | 9  201

requires experimental manipulation of one cognitive bias to track effects on other process. Cognitive bias modification methods provide the tools to test the causal influences among cognitive biases. These procedures encourage acquisition or attenuation of an emotional bias via exposure to experimentally established contingencies within specific task settings (Koster, Fox, & MacLeod, 2009). Studies that have tested the assumed influences among cognitive biases have typically employed cognitive bias modification procedures in combination with a transfer cognitive task. For example, in a study examining the influence of attention bias on emotional memory (Blaut et al., 2013), participants were trained to orient attention away from negative words. The attention manipulation involved a training variant of the dot-probe task in which a probe (the target) consistently replaced the neutral stimulus, instead of the equal replacement probability in the standard task design. The impact on memory bias was measured by presenting participants with a memory task in which novel neutral, negative, and positive words were presented. In a study testing whether interpretation bias influences memory (Tran, Hertel, & Joormann, 2011), interpretation bias was first modified by presenting participants with a series of ambiguous stories each ending with a word fragment for participants to complete. This word fragment imposed either a positive meaning (in the positive training condition) or negative meaning (in the negative training condition) on the ambiguous story. After the positive or negative interpretation bias training, participants completed an ambiguous stories interpretation task and a recall task testing memory for the ambiguous scenarios from the previous task. Such study protocols clearly enable the investigation of the effects of the manipulated cognitive bias on other cognitive processes. Though several studies have examined causal relations among cognitive biases, a potential threat to CCBH research is the availability of effective procedures to modify cognitive biases. While various novel approaches have been developed to modify attention and interpretation biases (Bernstein & Zvielli, 2014; Sanchez, Everaert, & Koster, 2016), there is some doubt about whether current approaches to modify cognitive biases such as attention bias using traditional dot-probe training tasks are successful (Koster & Bernstein, 2015; Okon-Singer, 2018). Moreover, cognitive bias modification procedures to target biases of memory are largely lacking (but see Vrijsen et al., 2014). Given the scarcity of appropriate methods, there is currently no research testing how memory bias may causally contribute to biases in attention or interpretation. The absence of reliable cognitive bias tasks and modification procedures may jeopardize progress in research on individual cognitive biases and integrative research testing the CCBH. Therefore, we think that the development of novel cognitive bias modification methods represents an important challenge for future research examining causal relations among cognitive biases.

202  Cognitive biases in health and psychiatric disorders

Predictive magnitude questions Longitudinal research designs are required to address predictive magnitude questions. To date, prospective studies exploring how multiple cognitive biases work together to influence depression are scarce. Existing research utilized a longitudinal design measuring cognitive biases at baseline and depression symptom levels at a later time point (Everaert, Duyck, & Koster, 2015). Though valuable as a first step, such study designs have important limitations that need to be addressed in future research. First, longitudinal designs with only two time points have a low temporal resolution. This restricts data analysis to tests of linear change models. While linear models might be appropriate to describe the change in depression symptoms and/or cognitive biases, it is plausible that the rate of change is not constant over time and may take a nonlinear shape. For example, the shape of change might be related to contextual factors such as the occurrence and chronicity of stressful events. By conducting at least four waves of data collection, studies can explore richer models of nonlinear change (e.g., quadratic growth) in depressive symptoms and how cognitive biases are related to different change parameters. Second, previous research has measured cognitive biases only at baseline. As such, these studies cannot examine whether there are changes in multiple cognitive biases over time and if such changes covary with changes in depressive symptoms or other contextual factors (e.g., perceived stress). Therefore, it is recommended that future longitudinal studies administer the same test battery of cognitive biases and symptom measures at each wave of data collection. Indeed, this again requires reliable cognitive tasks that also accurately conceptualize cognitive biases (Rodebaugh et al., 2016).

Empirical research on the CCBH Association questions Since the seminal study by Gotlib et  al. (2004), several studies have investigated correlations among cognitive biases using unrelated cognitive tasks. One study presented formerly depressed patients and never-disordered control individuals with a dot-probe task followed by a self-referential encoding and free recall task after a mood induction (Vrijsen, Van Oostrom, et  al., 2014). Similar to Gotlib et al.’s study in a clinical sample, no statistically significant correlations were found between attention and memory biases in formerly depressed individuals. By contrast, a recent study employing a different battery of experimental tasks reported some evidence for correlations among different cognitive biases (Sanchez et al., 2017). This study presented participants with varying depressive symptom levels with an engagement-disengagement task (measure of attention bias), scrambled sentences task (measure of interpretation bias), and a free recall autobiographical memory task (measure of memory bias). The results of that study showed that attentional biases for sad faces were

Revisiting the combined cognitive bias hypothesis  Chapter | 9  203

positively correlated with negative interpretation biases and negative memory biases. Taken together, research utilizing unrelated cognitive bias tasks has produced mixed findings, rendering conclusions regarding correlations among cognitive biases difficult. Important advances have been made by research that has attempted to model specific aspects of the interplay between cognitive biases. Building on initial work suggesting that depression-linked attention biases during encoding enhance later memory for negative material and impair memory for positive material (Ellis, Beevers, & Wells, 2011; Koster, De Raedt, Leyman, & De Lissnyder, 2010; LeMoult & Joormann, 2012; Wells et al., 2010), researchers studied the role of interpretation bias in the relation between attention and memory bias. In one recent study (Everaert et al., 2014), participants with varying depression levels completed an interpretation task requiring them to unscramble emotional sentences in either a positive or negative manner (e.g., “born winner am loser a I” into either “I am a born loser” or “I am a born winner”). When participants were unscrambling the sentences, biases in attention toward negative (e.g., “loser”) vs positive (e.g., “winner”) words were measured using eye tracking. A subsequent free recall task prompted participants to recollect the constructed self-referent meanings. This procedure is also depicted by Fig. 1. The results suggested that a bias in attentional selection toward negative words was related to a higher proportion of negative interpretations, which was in turn related to better memory for negative interpretations. Thus, interpretation bias may mediate the relation between attention selection bias and memory bias. Moreover, sustained attention toward negative words was directly related to a negative memory bias, without interpretation bias intervening in this relationship. This pattern of findings suggests that attention bias is both directly and indirectly related to memory bias via interpretation bias. Interestingly, the mediation model in which interpretation intervenes in the relation between attention bias and memory bias was also found by another study using an exogenous cueing task in combination with a scrambled sentences test and subsequent free recall task (Everaert, Tierens, Uzieblo, & Koster, 2013). Extending this work on explicit memory (i.e., storage systems that represent knowledge in a consciously accessible manner), one recent study examined whether symptoms of anhedonia are related to negative biases in attention and implicit memory (Salem et al., 2017). In that study, participants completed the attentional dot-probe task followed by a two-alternative forced-choice recognition task to measure implicit memory for stimuli that were presented during the dot-probe task. In the two-alternative forced-choice recognition task, participants were briefly presented with a stimulus word from the dot-probe task, which was then masked and replaced by two response choices (including the target word and a foil). Participants were instructed to select the word that was presented during the attention task. The results showed that implicit memory bias moderated the relation between attention bias and symptoms of anhedonia. Negative attention bias was associated with anhedonia only at high levels of

204  Cognitive biases in health and psychiatric disorders

implicit memory bias. However, the effect sizes were relatively small in this study. This suggests that attention biases do not influence implicit memory as they may influence explicit memory bias. Finally, there is some evidence suggesting that emotional memory may guide attention bias. Studies testing this hypothesis involve a learning phase followed by an attention task. The learning phase typically involves repeated pairings of neutral stimuli and emotional outcomes, while participants execute a cover task. The attention task subsequently presents only the neutral stimuli to examine whether the learned associations with emotional content may capture attention. Studies on depression have shown that individuals with higher depression levels do not orient attention toward stimuli associated with reward in an attentional cueing task (Brailean, Koster, Hoorelbeke, & De Raedt, 2014) or visual search task (Anderson, Leal, Hall, Yassa, & Yantis, 2014). In sum, research modeling specific hypothesized interactions among cognitive biases by using related cognitive tasks seems to produce more consistent findings compared with studies testing correlations among cognitive biases with independent tasks. The available research evidence suggests that attentional biases may be related to a subsequent congruent bias in memory and that emotional memory may modulate attention allocation.

Causal questions In researching causal relationships among cognitive biases, a fruitful line of studies has attempted to understand the causes of memory bias. Initial research has focused on the role of interpretation bias in explaining emotional biases in memory. Findings from studies in nonclinical samples suggest that memory recall is affected by interpretation biases operating during encoding of emotional material and by interpretation biases acquired after emotional material has been encoded (Salemink, Hertel, & Mackintosh, 2010; Tran et al., 2011). Recently, a study sought to extend this work by recruiting a sample of clinically depressed individuals (Joormann, Waugh, & Gotlib, 2015). In this study, participants completed an interpretation bias training inducing a positive interpretation bias. Next, they completed an ambiguous story interpretation task followed by a recall task testing memory for the ambiguous scenarios from the previous task. The results showed that only induced positive interpretation bias had trainingcongruent effects on the recall for endings (i.e., interpretations) of ambiguous scenarios. This finding substantiates the role of (positive) interpretation bias in influencing memory bias. In addition to interpretation biases, attention biases have also been hypothesized to influence emotional memory. In a recent study by Blaut et al. (2013), participants allocated to the training condition were trained to orient attention away from negative words to test the influence of attention bias on memory. Participants in the control condition were not trained to orient attention to

Revisiting the combined cognitive bias hypothesis  Chapter | 9  205

s­ pecific emotional stimuli. The attention manipulation involved a training variant of the dot-probe task in which a probe (the target) consistently replaces the neutral stimulus, instead of the equal replacement probability in the standard dot-probe task. It was found that individuals with higher depression levels did not exhibit a memory bias for negative words when they were trained to orient attention away from negative words. A typical memory bias occurred in the notraining control group. These findings suggest that attention bias may causally influence memory bias. Another line of research addressing causal CCBH questions examined causal links between biases of attention and interpretation in samples of individuals reporting varying depressive symptom levels. To test the hypothesized influence of attention bias on interpretation, two recent studies have trained participants to orient attention toward either positive or negative words using a dot-probe training task. Transfer of attention bias training to interpretation bias was examined by the scrambled sentence test following the training procedure (Everaert, Mogoaşe, David, & Koster, 2015). Across both studies, the training procedure was not successful at effectively inducing emotional biases in attention allocation in the training condition. No differences in attention bias were observed between the training and control condition. Instead, there was considerable variability in the extent to which participants acquired an attention bias in the training condition. However, the individual differences in attention bias acquisition were not related to the performance on the interpretation tasks. To date, there is one study examining whether interpretation bias influences attention bias in a sample of adolescents with major depression (LeMoult et al., 2017). In this study, participants received either six sessions of positive interpretation bias training or neutral training followed by an interpretation and attention bias task. The training was effective in that adolescents r­ eceiving positive training also interpreted ambiguous scenarios more positively than did p­ articipants who received the neutral training. However, there was no transfer of the training to the dot-probe task as a measure of attention bias. Consequently, no empirical support was found for an influence of interpretation bias on attention bias. Taken together, several studies have provided consistent evidence for the role of interpretation bias in influencing memory bias. Yet, current research has not been able to provide empirical support for mutual influences between attention and interpretation biases related to depression. This is remarkable in light of the empirical support for such mutual influences between attention and interpretation biases coming from studies on anxiety (cf. Amir, Bomyea, & Beard, 2010; White, Suway, Pine, Bar-Haim, & Fox, 2011). In examining the potential causal linkages between attention and interpretation bias in depression, studies could take advantage from novel and promising methods to induce or reduce biases of attention and interpretation (cf. Bernstein & Zvielli, 2014; Sanchez et al., 2016).

206  Cognitive biases in health and psychiatric disorders

Predictive magnitude questions There is currently only one study that examined how biases in attention, interpretation, and memory combine to predict future depressive symptom levels (Everaert, Duyck, & Koster, 2015). This study tested the predictive value of two integrative approaches to model the combined impact of multiple cognitive biases, namely, the additive (i.e., cognitive biases have a cumulative effect) vs the weakest link (i.e., the dominant cognitive bias is important) model. To this end, this study organized a 1-year follow-up on the cross-sectional study quantifying relations among biases of attention, interpretation, and memory (Everaert et al., 2014). At the follow-up moment, participants’ depressive symptom levels were reassessed. The results revealed that the weakest link model had incremental validity over the additive model in predicting prospective changes in depressive symptoms, though both models explained a significant proportion of variance in the change in depressive symptoms. These initial findings suggest that the best cognitive marker of the evolution in depressive symptoms is the cognitive process that is dominantly biased toward negative material. This finding highlights the importance of considering idiographic cognitive profiles with multiple cognitive processes for understanding cognitive biases in depression. While this study aimed to conceptualize which (combination of) biases yields the greatest potency in predicting the symptom course over time, it does not cast light on how multiple cognitive biases work together in predicting future depression levels. As noted, addressing such questions requires longitudinal designs involving multiple waves of data collection to measure cognitive biases and depression symptoms at multiple time points. Further research is needed to address this open question.

Limitations and future directions Though research has made considerable progress in the past 6 years, several limitations remain and represent important directions for future research. First, various aspects of the interplay among cognitive biases have received only limited empirical attention or have yet to be investigated. To date, most studies have examined the role of attention bias during encoding of emotional material to explain emotional biases in memory. Indeed, much has yet to be discovered about whether and how attention bias modulates memory retrieval and how memory bias may shape attention allocation in depression. As outlined, several challenges also remain for research examining the causal direction of the observed associations between cognitive biases. Also, much remains to be understood about the combined influence of multiple cognitive biases on the course of depression. Advances with respect to these issues require targeted research that can be guided by the theoretical models discussed earlier. Second, much remains to be learned about the neural basis of interactions among cognitive biases in depression. This is because current research on the

Revisiting the combined cognitive bias hypothesis  Chapter | 9  207

CCBH has solely relied on behavioral tasks. Fortunately, research has identified neural mechanisms that are involved in cognitive biases in depression (Belzung, Willner, & Philippot, 2015; De Raedt & Koster, 2010; Disner, Beevers, Haigh, & Beck, 2011). For example, biased information processing has been linked to greater and more sustained amygdala reactivity, hypoactivity in the left dorsolateral prefrontal cortex and hyperactivity in the right dorsolateral prefrontal cortex. Moreover, memory bias has been associated with amygdala hyperactivity, which is positively correlated with activity in the hippocampus, caudate, and putamen (Disner et al., 2011). To gain insight into the interactions among cognitive biases at the neural level, future research may be guided by existing cognitive neuroscience research examining attention—memory interactions (Chun & Turk-Browne, 2007) and pathophysiological models of depression (Belzung et al., 2015). Third, prior research has generally studied cognitive biases at the disorder level and ignored the heterogeneous nature of depression. The dominant focus on the disorder level may be problematic because it overlooks critical differences in the importance of individual symptoms, differential relations between symptoms, and differential relations between symptoms and risk factors such as cognitive biases (Fried, 2015; Fried, Nesse, Zivin, Guille, & Sen, 2014). From both a theoretical and clinical stance, knowledge of whether (clusters of) symptoms are more closely related to (combinations of) cognitive biases is urgently required. Therefore, future research on the CCBH needs to adopt a symptom level approach to gain insight into how multiple cognitive biases may maintain clusters of depression symptoms. Fourth, while most studies testing the CCBH in depression focus on attention, interpretation, and/or memory biases, various relations with other critical cognitive factors await systematic empirical scrutiny. For example, studies have just started to examine the role of different executive control difficulties in the interplay between biases of attention and interpretation during encoding stages (e.g., Everaert, Grahek, & Koster, 2017). Also, research has started to document the role of expectancy biases in guiding attention in the context of anxiety (Aue & Okon-Singer, 2015). This research may inform advances in studies examining interactions among biases of attention, interpretation, and memory in depression. Indeed, broadening the scope of cognitive factors that are studied within a CCBH framework may provide an interesting avenue for future research and holds potential to facilitate an integrated understanding of the cognitive foundations of depression. Finally, currently little is known about factors that may modulate the interplay among cognitive biases. Indeed, the interplay among multiple cognitive biases may be influenced by factors that impact the expression of individual cognitive biases, such as stimulus-driven salience (but also see Niu, Todd, & Anderson, 2012), self-reference (Everaert, Podina, & Koster, 2017; Gaddy & Ingram, 2014), and counterregulation processes (Schwager & Rothermund, 2013). Of interest to understanding how c­ ognitive biases ­interact within a ­specific

208  Cognitive biases in health and psychiatric disorders

context are an individual’s current goals. Goals have been defined as motivational representations of a desired end state (Dickson, Johnson, Huntley, Peckham, & Taylor, 2017). The nature of an individual’s goals and the specific formulation (e.g., approach vs avoidance-oriented goals) may shape one’s cognitive set. This cognitive set may then alter which information will become more salient in the external environment and/or is more accessible for retrieval from memory. Thus, goals may influence memory retrieval and attention allocation toward cues in a congruent manner and may override initial negative biases to attain the goal (Vogt, Koster, & De Houwer, 2017). Exploring the role of goals in the interplay among cognitive biases in depression provides an important avenue for future research.

Summary The past several years have seen important theoretical, methodological, and empirical advances in discovering the interactions among attention, interpretation, and memory biases in depression. This review provides an overview of recent theories, methods, and research on the combined cognitive bias hypothesis in depression. The accumulated research findings provide evidence for interrelations among cognitive biases, suggesting that these processes should be studied in an integrative manner to understand their role in depression. Yet, there is much that remains to be understood about their complex interplay. To this end, guiding frameworks and methodological approaches are discussed to stimulate targeted research with the aim of gaining a comprehensive understanding of how multiple cognitive biases are involved in depression.

References Abela, J. R. Z., & Hankin, B. L. (2008). Cognitive vulnerability to depression in children and adolescents: A developmental psychopathology perspective. Handbook of Depression in Children and Adolescents, 35–78. Amir, N., Bomyea, J., & Beard, C. (2010). The effect of single-session interpretation modification on attention bias in socially anxious individuals. Journal of Anxiety Disorders, 24(2), 178–182. https://doi.org/10.1016/j.janxdis.2009.10.005. Anderson, B. A., Leal, S. L., Hall, M. G., Yassa, M. A., & Yantis, S. (2014). The attribution of value-based attentional priority in individuals with depressive symptoms. Cognitive, Affective, & Behavioral Neuroscience, 14(4), 1221–1227. https://doi.org/10.3758/s13415-014-0301-z. Armstrong, T., & Olatunji, B. O. (2012). Eye tracking of attention in the affective disorders: A meta-analytic review and synthesis. Clinical Psychology Review, 32(8), 704–723. https://doi. org/10.1016/j.cpr.2012.09.004. Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. https://doi.org/10.1016/j.cpr.2015.08.005. Awh, E., Vogel, E. K., & Oh, S. H. (2006). Interactions between attention and working memory. Neuroscience, 139(1), 201–208. https://doi.org/10.1016/j.neuroscience.2005.08.023. Beck, A. T., & Haigh, E. A. P. (2014). Advances in cognitive theory and therapy: The generic cognitive model. Annual Review of Clinical Psychology, 10(1), 1–24. https://doi.org/10.1146/ annurev-clinpsy-032813-153734.

Revisiting the combined cognitive bias hypothesis  Chapter | 9  209 Belzung, C., Willner, P., & Philippot, P. (2015). Depression: From psychopathology to pathophysiology. Current Opinion in Neurobiology, 30, 24–30. https://doi.org/10.1016/j.conb.2014.08.013. Bernstein, A., & Zvielli, A. (2014). Attention feedback awareness and control training (A-FACT): Experimental test of a novel intervention paradigm targeting attentional bias. Behaviour Research and Therapy, 55(1), 18–26. https://doi.org/10.1016/j.brat.2014.01.003. Blaut, A., Paulewicz, B., Szastok, M., Prochwicz, K., Koster, E. H. W., Paulewicz, B., et al. (2013). Are attentional bias and memory bias for negative words causally related? Journal of Behavior Therapy and Experimental Psychiatry, 44(3), 293–299. https://doi.org/10.1016/j.jbtep.2013.01.002. Brailean, A. M., Koster, E. H. W., Hoorelbeke, K., & De Raedt, R. (2014). Attentional modulation by reward and punishment cues in relation to depressive symptoms. Journal of Behavior Therapy and Experimental Psychiatry, 45(3), 351–359. https://doi.org/10.1016/j.jbtep.2014.03.003. Chun, M. M., Golomb, J. D., & Turk-Browne, N. B. (2011). A taxonomy of external and internal attention. Annual Review of Psychology, 62, 73–101. https://doi.org/10.1146/annurev. psych.093008.100427. Chun, M. M., & Turk-Browne, N. B. (2007). Interactions between attention and memory. Current Opinion in Neurobiology, 17(2), 177–184. https://doi.org/10.1016/j.conb.2007.03.005. Clark, D. A., Beck, A. T., & Alford, B. A. (1999). Scientific foundations of cognitive theory and therapy of depression. New York: John Wiley & Sons. De Raedt, R., & Koster, E. H. W. (2010). Understanding vulnerability for depression from a cognitive neuroscience perspective: A reappraisal of attentional factors and a new conceptual framework. Cognitive, Affective, & Behavioral Neuroscience, 10(1), 50–70. https://doi.org/10.3758/CABN.10.1.50. Dickson, J., Johnson, S. L., Huntley, C. D., Peckham, A., & Taylor, P. J. (2017). An integrative study of motivation and goal regulation processes in subclinical anxiety, depression and hypomania. Psychiatry Research, 256, 6–12. https://doi.org/10.1016/J.PSYCHRES.2017.06.002. Disner, S. G., Beevers, C. G., Haigh, E. A. P., & Beck, A. T. (2011). Neural mechanisms of the cognitive model of depression. Nature Reviews Neuroscience, 12(August), https://doi.org/10.1038/nrn3027. Ellis, A. J., Beevers, C. G., & Wells, T. T. (2011). Attention allocation and incidental recognition of emotional information in dysphoria. Cognitive Therapy and Research, 35(5), 425–433. https:// doi.org/10.1007/s10608-010-9305-3. Everaert, J., Bernstein, A., Joormann, J., & Koster, E. H. W. (2018). Toward understanding the dynamic interplay between cognitive biases in psychopathology: The Attention-Memory-BiasInteraction framework, submitted. Everaert, J., Duyck, W., & Koster, E. H. W. (2014). Attention, interpretation, and memory biases in subclinical depression: A proof-of-principle test of the combined cognitive biases hypothesis. Emotion, 14(2), 331–340. https://doi.org/10.1037/a0035250. Everaert, J., Duyck, W., & Koster, E. H. W. (2015). Emotionally biased cognitive processes: The weakest link predicts prospective changes in depressive symptom severity. PLoS One, 10(5), e0124457. https://doi.org/10.1371/journal.pone.0124457. Everaert, J., Grahek, I., & Koster, E. H. W. (2017). Individual differences in cognitive control modulate cognitive biases linked to depressive. Cognition & Emotion, 31(4), 736–746. https://doi.or g/10.1080/02699931.2016.1144562. Everaert, J., & Koster, E. H. W. (2015). Interactions among emotional attention, encoding, and retrieval of ambiguous information: An eye-tracking study. Emotion, 15(5), 539–543. https:// doi.org/10.1037/emo0000063. Everaert, J., Koster, E. H. W., & Derakshan, N. (2012). The combined cognitive bias hypothesis in depression. Clinical Psychology Review, 32(5), 413–424. https://doi.org/10.1016/j.cpr.2012.04.003. Everaert, J., Mogoaşe, C., David, D., & Koster, E. H. W. (2015). Attention bias modification via single-session dot-probe training: Failures to replicate. Journal of Behavior Therapy and Experimental Psychiatry, 49, 5–12. https://doi.org/10.1016/j.jbtep.2014.10.011.

210  Cognitive biases in health and psychiatric disorders Everaert, J., Podina, I. R., & Koster, E. H. W. (2017). A comprehensive meta-analysis of interpretation biases in depression. Clinical Psychology Review, 58, 33–48. https://doi. org/10.1016/j.cpr.2017.09.005. Everaert, J., Tierens, M., Uzieblo, K., & Koster, E.H.W.W. (2013). The indirect effect of attention bias on memory via interpretation bias: Evidence for the combined cognitive bias hypothesis in subclinical depression. Cognition & Emotion, 27(8), 1450–1459. https://doi.org/10.1080/02 699931.2013.787972. Fried, E. I. (2015). Problematic assumptions have slowed down depression research: Why symptoms, not syndromes are the way forward. Frontiers in Psychology, 6(March), 1–11. https://doi. org/10.3389/fpsyg.2015.00309. Fried, E. I., Nesse, R. M., Zivin, K., Guille, C., & Sen, S. (2014). Depression is more than the sum score of its parts: Individual DSM symptoms have different risk factors. Psychological Medicine, 44(10), 2067–2076. https://doi.org/10.1017/S0033291713002900. Gaddy, M. A., & Ingram, R. E. (2014). A meta-analytic review of mood-congruent implicit memory in depressed mood. Clinical Psychology Review, 34(5), 402–416. https://doi. org/10.1016/j.cpr.2014.06.001. Gotlib, I. H., & Joormann, J. (2010). Cognition and depression: Current status and future directions. Annual Review of Clinical Psychology, 6(1), 285–312. https://doi.org/10.1146/annurev. clinpsy.121208.131305. Gotlib, I. H., Kasch, K. L., Traill, S., Joormann, J., Arnow, B. A., & Johnson, S. L. (2004). Coherence and specificity of information-processing biases in depression and social phobia. Journal of Abnormal Psychology, 113(3), 386–398. https://doi.org/10.1037/0021-843X.113.3.386. Hallion, L. S. S., & Ruscio, A. M. M. (2011). A meta-analysis of the effect of cognitive bias modification on anxiety and depression. Psychological Bulletin, 137(6), 940–958. https://doi. org/10.1037/a0024355 2011-13412-001 [pii]. Hankin, B. L. (2012). Future directions in vulnerability to depression among youth: Integrating risk factors and processes across multiple levels of analysis. Journal of Clinical Child & Adolescent Psychology, 41(5), 695–718. https://doi.org/10.1080/15374416.2012.711708. Hertel, P. T. (1997). On the contributions of deficient cognitive control to memory impairments in depression. Cognition & Emotion, 11(5/6), 569–583. https://doi.org/10.1080/026999397379890a. Hertel, P. T. (2004). Memory for emotional and nonemotional events in depression: A question of habit? In D. Reisberg & P. Hertel (Eds.), Memory and emotion. New York: Oxford University Press. Hertel, P. T., & Brozovich, F. (2010). Cognitive habits and memory distortions in anxiety and depression. Current Directions in Psychological Science, 19(3), 155–160. https://doi. org/10.1177/0963721410370137. Hirsch, C. R., Clark, D. M., & Mathews, A. (2006). Imagery and interpretations in social phobia: Support for the combined cognitive biases hypothesis. Behavior Therapy, 37(3), 223–236. https://doi.org/10.1016/j.beth.2006.02.001. Hutchinson, J. B., & Turk-Browne, N. B. (2012). Memory-guided attention: Control from multiple memory systems. Trends in Cognitive Sciences, 16(12), 576–579. https://doi. org/10.1016/j.tics.2012.10.003. Ingram, R. E. (1984). Toward an information-processing analysis of depression. Cognitive Therapy and Research, 8(5), 443–477. https://doi.org/10.1007/BF01173284. Johnson, S. L., Joormann, J., & Gotlib, I. H. (2007). Does processing of emotional stimuli predict symptomatic improvement and diagnostic recovery from major depression? Emotion, 7(1), 201–206. https://doi.org/10.1037/1528-3542.7.1.201. Joormann, J., Waugh, C. E., & Gotlib, I. H. (2015). Cognitive bias modification for interpretation in major depression: Effects on memory and stress reactivity. Clinical Psychological Science, https://doi.org/10.1177/2167702614560748.

Revisiting the combined cognitive bias hypothesis  Chapter | 9  211 Joormann, J., Yoon, K. L., & Zetsche, U. (2007). Cognitive inhibition in depression. Applied and Preventive Psychology, 12(3), 128–139. https://doi.org/10.1016/j.appsy.2007.09.002. Kessler, R. C., & Bromet, E. J. (2013). The epidemiology of depression across cultures. Annual Review of Public Health, 34(1), 119–138. https://doi.org/10.1146/­annurev-publhealth-031912-114409. Koster, E. H. W., & Bernstein, A. (2015). Introduction to the special issue on cognitive bias modification: Taking a step back to move forward? Journal of Behavior Therapy and Experimental Psychiatry, 49, 1–4. https://doi.org/10.1016/j.jbtep.2015.05.006. Koster, E. H. W., De Raedt, R., Leyman, L., & De Lissnyder, E. (2010). Mood-congruent attention and memory bias in dysphoria: Exploring the coherence among information-processing biases. Behaviour Research and Therapy, 48(3), 219–225. https://doi.org/10.1016/j.brat.2009.11.004. Koster, E. H. W., Fox, E., & MacLeod, C. (2009). Introduction to the special section on cognitive bias modification in emotional disorders. Journal of Abnormal Psychology, 118(1), 1–4. https:// doi.org/10.1037/a0014379. Kraemer, H. C., Stice, E., Kazdin, A., Offord, D., & Kupfer, D. (2001). How do risk factors work together? Mediators, moderators, and independent, overlapping, and proxy risk factors. American Journal of Psychiatry, 158(6), 848–856. https://doi.org/10.1176/appi.ajp.158.6.848. LeMoult, J., Colich, N., Joormann, J., Singh, M. K., Eggleston, C., & Gotlib, I. H. (2017). Interpretation bias training in depressed adolescents: Near- and far-transfer effects. Journal of Abnormal Child Psychology, 1–9. https://doi.org/10.1007/s10802-017-0285-6. LeMoult, J., & Joormann, J. (2012). Attention and memory biases in social anxiety disorder: The role of comorbid depression. Cognitive Therapy and Research, 36(1), 47–57. https://doi. org/10.1007/s10608-010-9322-2. Mathews, A., & Macleod, C. M. (2005). Cognitive vulnerability to emotional disorders. Annual Review of Clinical Psychology, 1(1), 167–195. https://doi.org/10.1146/annurev.clinpsy.1.102803.143916. Matt, G. E., Vázquez, C., & Campbell, W. K. (1992). Mood-congruent recall of affectively toned stimuli: A meta-analytic review. Clinical Psychology Review, 12(2), 227–255. https://doi. org/10.1016/0272-7358(92)90116-P. Menne-Lothmann, C., Viechtbauer, W., Höhn, P., Kasanova, Z., Haller, S. P., Drukker, M., et al. (2014). How to boost positive interpretations? A meta-analysis of the effectiveness of cognitive bias modification for interpretation. PLoS One, 9(6), e100925. https://doi.org/10.1371/journal.pone.0100925. Mogoaşe, C., David, D., & Koster, E. H. W. (2014). Clinical efficacy of attentional bias modification procedures: An updated meta-analysis. Journal of Clinical Psychology, 70(12), 1133–1157. https://doi.org/10.1002/jclp.22081. Niu, Y., Todd, R. M., & Anderson, A. K. (2012). Affective salience can reverse the effects of ­stimulus-driven salience on eye movements in complex scenes. Frontiers in Psychology, 3, 336. https://doi.org/10.3389/fpsyg.2012.00336. Okon-Singer, H. (2018). The role of attention bias to threat in anxiety: Mechanisms, modulators and open questions. Current Opinion in Behavioral Sciences, 19, 26–30. https://doi. org/10.1016/j.cobeha.2017.09.008. Peckham, A. D., McHugh, R. K., & Otto, M. W. (2010). A meta-analysis of the magnitude of biased attention in depression. Depression and Anxiety, 27(12), 1135–1142. https://doi.org/10.1002/ da.20755. Price, R. B., Rosen, D., Siegle, G. J., Ladouceur, C. D., Tang, K., Allen, K. B., et al. (2016). From anxious youth to depressed adolescents: Prospective prediction of 2-year depression symptoms via attentional bias measures. Journal of Abnormal Psychology, 125(2), 267–278. https://doi. org/10.1037/abn0000127. Reilly, L. C., Ciesla, J. A., Felton, J. W., Weitlauf, A. S., & Anderson, N. L. (2012). Cognitive vulnerability to depression: A comparison of the weakest link, keystone and additive models. Cognition & Emotion, 26(3), 521–533. https://doi.org/10.1080/02699931.2011.595776.

212  Cognitive biases in health and psychiatric disorders Rodebaugh, T. L., Scullin, R. B., Langer, J. K., Dixon, D. J., Huppert, J. D., Bernstein, A., et al. (2016). Unreliability as a threat to understanding psychopathology: The cautionary tale of attentional bias. Journal of Abnormal Psychology, 125(6), 840–851. https://doi.org/10.1037/ abn0000184. Rude, S. S., Durham-Fowler, J. A., Baum, E. S., Rooney, S. B., & Maestas, K. L. (2010). Self-report and cognitive processing measures of depressive thinking predict subsequent major depressive disorder. Cognitive Therapy and Research, 34(2), 107–115. https://doi.org/10.1007/s10608009-9237-y. Salem, T., Winer, E. S., & Nadorff, M. R. (2017). Combined behavioural markers of cognitive biases are associated with anhedonia. Cognition & Emotion, 31, 1–9. https://doi.org/10.1080/ 02699931.2017.1307808. Salemink, E., Hertel, P. T., & Mackintosh, B. (2010). Interpretation training influences memory for prior interpretations. Emotion, 10(6), 903–907. https://doi.org/10.1037/a0020232. Sanchez, A., Duque, A., Romero, N., & Vazquez, C. (2017). Disentangling the interplay among cognitive biases: Evidence of combined effects of attention, interpretation and autobiographical memory in depression. Cognitive Therapy and Research, https://doi.org/10.1007/s10608-017-9858-5. Sanchez, A., Everaert, J., & Koster, E. H. W. (2016). Attention training through gaze-contingent feedback: Effects on reappraisal and negative emotions. Emotion, 16(7), 1074–1085. https:// doi.org/10.1037/emo0000198. Schwager, S., & Rothermund, K. (2013). Counter-regulation triggered by emotions: Positive/negative affective states elicit opposite valence biases in affective processing. Cognition & Emotion, 27(5), 839–855. https://doi.org/10.1080/02699931.2012.750599. Todd, R. M., & Manaligod, M. G. M. (2017). Implicit guidance of attention: The priority state space framework. Cortex, https://doi.org/10.1016/j.cortex.2017.08.001. Tran, T. B., Hertel, P. T., & Joormann, J. (2011). Cognitive bias modification: Induced interpretive biases affect memory. Emotion, 11(1), 145–152. https://doi.org/10.1037/a0021754. Vogt, J., Koster, E. H. W., & De Houwer, J. (2017). Safety first: Instrumentality for reaching safety determines attention allocation under threat. Emotion, 17(3), 528–537. https://doi.org/10.1037/ emo0000251. Vrijsen, J. N., Becker, E. S., Rinck, M., Van Oostrom, I., Speckens, A., Whitmer, A., et al. (2014). Can memory bias be modified? The effects of an explicit cued-recall training in two independent samples. Cognitive Therapy and Research, 38(2), 217–225. https://doi.org/10.1007/ s10608-013-9563-y. Vrijsen, J. N., Hertel, P. T., & Becker, E. S. (2016). Practicing emotionally biased retrieval affects mood and establishes biased recall a week later. Cognitive Therapy and Research, 40(6), 764–773. https://doi.org/10.1007/s10608-016-9789-6. Vrijsen, J. N., Van Oostrom, I., Isaac, L., Becker, E. S., & Speckens, A. (2014). Coherence between attentional and memory biases in sad and formerly depressed individuals. Cognitive Therapy and Research, 38(3), 334–342. https://doi.org/10.1007/s10608-013-9590-8. Wells, T. T., Beevers, C. G., Robison, A. E., & Ellis, A. J. (2010). Gaze behavior predicts memory bias for angry facial expressions in stable dysphoria. Emotion, 10(6), 894–902. https://doi. org/10.1037/a0020022. White, L. K., Suway, J. G., Pine, D. S., Bar-Haim, Y., & Fox, N. A. (2011). Cascading effects: The influence of attention bias to threat on the interpretation of ambiguous information. Behaviour Research and Therapy, 49(4), 244–251. https://doi.org/10.1016/j.brat.2011.01.004. Williams, J. M. G., Watts, C., MacLeod, C. M., Mathews, A., Williams, M. G., Watts, F. N., et al. (1997). Cognitive psychology and emotional disorders (2nd ed.). Chichester: Wiley.

Revisiting the combined cognitive bias hypothesis  Chapter | 9  213 Winer, E. S., & Salem, T. (2016). Reward devaluation: Dot-probe meta-analytic evidence of avoidance of positive information in depressed persons. Psychological Bulletin, 142(1), 18–78. https://doi.org/10.1037/bul0000022. Wittenborn, A. K., Rahmandad, H., Rick, J., & Hosseinichimeh, N. (2016). Depression as a systemic syndrome: Mapping the feedback loops of major depressive disorder. Psychological Medicine, 46(03), 551–562. https://doi.org/10.1017/S0033291715002044.

Further reading Marchetti, I., Everaert, J., Dainer-Best, J., Loeys, T., Beevers, C. G., & Koster, E. H. W. (2017). Specificity and overlap of attention and memory biases in depression. Journal of Affective Disorders, 225, 404–412. https://doi.org/10.1016/j.jad.2017.08.037.

Chapter 10

The impact of top-down factors on threat perception biases in health and anxiety Tamara J. Sussmana,b, Jingwen Jinc,d, Aprajita Mohantye a

Department of Psychiatry, Columbia University Irving Medical Center, New York, NY, United States, bDepartment of Psychiatry, New York State Psychiatric Institute, New York, NY, United States, cDepartment of Psychology, The University of Hong Kong, Pok Fu Lam, Hong Kong, dThe State Key Laboratory of Brain and Cognitive Sciences, The University of Hong Kong, Pok Fu Lam, Hong Kong, eDepartment of Psychology, Stony Brook University, Stony Brook, NY, United States

Introduction Organisms that can respond rapidly to a fearsome predator or a delicious treat are at an evolutionary advantage. Therefore, our perceptual system prioritizes emotional over neutral information, allowing us to more effectively avoid threats and obtain rewards. Threatening stimuli, such as snakes, spiders, and angry faces, are thought to belong to a special class of stimuli that are perceptually prioritized due to their importance for survival (Brosch, Pourtois, & Sander, 2010; Eimer & Holmes, 2007; New, Cosmides, & Tooby, 2007; Seligman, 1971). This hypothesis has been supported empirically by research showing that potential threats, including spiders, snakes, and angry faces, are detected more quickly than nonthreatening stimuli, such as mushrooms, flowers, and neutral faces (Hansen & Hansen, 1988; Horstmann, 2007; Ohman, Flykt, & Esteves, 2001). Eye movements orient more quickly to threatening compared with neutral stimuli (Bannerman, Milders, de Gelder, & Sahraie, 2009), and task performance is improved on attentional blink paradigms for threatening compared with neutral stimuli shown as a second target in a rapid stream of images (Anderson, 2005). Positive emotional stimuli are also perceptually favored compared with neutral stimuli; however, these improvements in detection tend to be smaller than those elicited by threats, particularly for schematic images of faces (Carretie, Mercado, Tapia, & Hinojosa, 2001; Dijksterhuis & Aarts, 2003; Nummenmaa & Calvo, 2015; Smith, Cacioppo, Larsen, & Chartrand, 2003; Stefanics, Csukly, Komlosi, Czobor, & Czigler, 2012; Sussman, Weinberg, Szekely, Hajcak, & Mohanty, 2017). Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00010-6 © 2020 Elsevier Inc. All rights reserved.

215

216  Cognitive biases in health and psychiatric disorders

Much of the current literature has attributed the perceptual prioritization of threatening stimuli to differences in bottom-up processing, driven by the physical characteristics or evolutionary significance of these stimuli (Bannerman et al., 2009; Ohman et al., 2001). Similarly, affective neuroscience research examining threat perception has focused on neural pathways that promote “automatic” perception of emotional stimuli (Mendez-Bertolo et  al., 2016; Vuilleumier & Pourtois, 2007). Researchers have theorized that threatening stimuli are prioritized due to a perceptual processing or attentional bias for threatening stimuli (Bar-Haim, Lamy, Pergamin, Bakermans-Kranenburg, & van Ijzendoorn, 2007; Cisler, Bacon, & Williams, 2009). However, perceptual and attentional biases can influence decision-making even before arrival of the threatening stimuli, for example, by prioritizing contexts and cues that indicate upcoming threats. While the research outlined earlier has focused on the relatively automatic aspects of stimulus detection, processes that facilitate and hinder perception have been shown to start before a stimulus is encountered. Empirical findings have demonstrated that perception can be guided by top-down factors such as goals and expectations (Bacon & Egeth, 1994; Itti & Koch, 2001), challenging the hypothesis that threatening stimuli are prioritized only via automatic, bottom-up processes. For example, prestimulus cues improve target detection whether they are implicit or explicit (Chen & Zelinsky, 2006; Stein & Peelen, 2015; Wolfe, Butcher, Lee, & Hyle, 2003). Emotional information has also been shown to guide perception in a top-down manner (Sussman et al., 2017; Sussman, Szekely, Hajcak, & Mohanty, 2016; Wormwood, Lynn, Feldman Barrett, & Quigley, 2016). For example, both positive and negative stimuli guide attention when they are task relevant, and do not impair performance when they are irrelevant to task goals (Hahn & Gronlund, 2007; Lichtenstein-Vidne et al., 2017; Lichtenstein-Vidne, Henik, & Safadi, 2012; Van Dessel & Vogt, 2012; Vogt, De Houwer, Crombez, & Van Damme, 2013; Williams, Moss, Bradshaw, & Mattingley, 2005). Motivational factors, such as reward and punishment can also alter attentional allocation (Della Libera & Chelazzi, 2006; Engelmann & Pessoa, 2007). For example, cues indicating potential rewards reduce the impact of threatening distractors on a perception task (Padmala, Sirbu, & Pessoa, 2017). Overall, there is increasing research showing that emotional information can bias perception endogenously in a top-down manner. However, it is also important to note that recent research has found that neutral prestimulus (compared with threat-related) cues improved the detection of neutral stimuli in a visual search paradigm, whereas threat-related (compared with neutral) cues did not aid in the subsequent perception of threatening stimuli (Aue, Chauvigne, Bristle, Okon-Singer, & Guex, 2016; Aue, Guex, Chauvigne, Okon-Singer, & Vuilleumier, 2019; Aue, Guex, Chauvigne, & Okon-Singer, 2013). Therefore, the type of task being performed may increase or decrease the influence the impact of top-down factors on the perception of emotional stimuli. The impact of endogenous biases may be particularly important to consider in anxiety disorders, as they are associated with inaccurate expectations about

Impact of top-down factors on threat perception biases  Chapter | 10  217

future negative events (Grupe & Nitschke, 2013). In addition to anxiety disorders, dispositional anxiety is associated with exaggerated estimates of the likelihood and impact of negative future events (Aue & Okon-Singer, 2015; Grupe & Nitschke, 2013). The influence of endogenous processes on perception in anxiety is also made evident by studies demonstrating that prestimulus, threatrelated cues have distinct impacts on perception in different types of anxiety (Sussman et al., 2016). Therefore, in this chapter, we explore the possibility that the perceptual prioritization of threatening stimuli, in normal function and in anxiety, may be attributed to prestimulus biases, in addition to biases that occur after the stimulus is presented. We first discuss how emotion-related endogenous factors (such as expectation of threat) can guide perception, highlighting major theories in the field. Next, we discuss conceptual and methodological drawbacks that result from neglecting emotional top-down factors in threat perception. We also discuss neurobiological factors related to endogenously guided threat perception and differences between top-down threat perception in healthy function and clinical anxiety. Finally, we discuss future directions for the field.

Major theories and evidence supporting top-down threat processing Threats can take us by surprise. However, we can often predict what kinds of threatening stimuli we are likely to encounter in a given context. Both explicit cues, such as a sign warning of bears on a mountain trail, and implicit cues, such as a densely forested area, provide us with information about what we are likely to encounter, improving subsequent perception of predictable stimuli. Top-down processes related to emotion or motivation (looking for an angry face or for a wallet full of money) can influence the process of perception. Positive and negative facial expressions are perceptually prioritized only when they are the goal of a search task, and not when they distract from the task at hand (Hahn & Gronlund, 2007; Williams et al., 2005). Prestimulus cues indicating a threat-related decision were found to improve perceptual sensitivity and to speed subsequent detection (Sussman et  al., 2016; Sussman et  al., 2017). Furthermore, compared with neutral stimuli, threatening stimuli were detected faster following the fear cue, but not following a neutral cue (Sussman et al., 2017). Therefore, in addition to the impact of bottom-up processes, top-down factors improve the efficiency of visual search for emotional stimuli and can improve target detection as well. Several theories examine the relationship between the perception of emotional stimuli and performance. We focus on how these theories posit that emotional information is used in a voluntary, top-down manner to guide attention and perception. According to the arousal-based competition (ABC) theory, emotional cues increase arousal, biasing attention toward the perception of stimuli that are salient to the task at hand and weakening perception of task-irrelevant stimuli (Mather & Sutherland, 2011). Arousal biases p­ erception

218  Cognitive biases in health and psychiatric disorders

both in instances when the target is unexpected (e.g., a snake suddenly appears in our path) and when it is goal-relevant, and top-down processes are involved in perception (e.g., snake appears in our path after viewing signs warning of rattlesnakes). The ABC model theorizes that emotional arousal biases perception in favor of the relevant stimulus while simultaneously suppressing task-irrelevant stimuli. This enhancement is hypothesized to be the result of the arousal-induced release of norepinephrine into the locus coeruleus, which increases levels of both norepinephrine and glutamate at the site of the goalrelevant representation, thereby boosting the representation of the goal-relevant stimulus (Mather, Clewett, Sakaki, & Harley, 2015). Consistent with this model, on a word-identification task, emotion-related prestimulus cues led to improved target detection compared with neutral cues, while emotional distractors did not impact performance, demonstrating that searching for emotional stimuli can improve processing of subsequent emotional targets (Zeelenberg, Wagenmakers, & Rotteveel, 2006). Similarly, on a visual search task, target detection was improved when prestimulus cues accurately predicted upcoming angry, compared with neutral faces, demonstrating that top-down processes related to emotional target features can improve performance (Mohanty, Egner, Monti, & Mesulam, 2009). The dual-competition framework proposes that affective significance (e.g., threats or rewards) influences competition at the level of perceptual and executive functioning to determine the behavioral outcome (Pessoa, 2009). Emotionladen stimuli are prioritized perceptually, due to their enhanced representation in sensory cortices (a stimulus-driven effect). They are also prioritized due to top-down processes, such as attention. Stronger sensory representations receive increased attention, likely via interactions between the amygdala and controlrelated brain regions, such as dorsal lateral prefrontal cortex. However, since subcomponents of executive control (e.g., inhibition or shifting) share limited resources, one subcomponent can be compromised if another is utilizing more resources. High-threat stimuli are theorized to also utilize these shared resources, reducing available resources. Thus, the dual-competition framework provides an explanation of how top-down processes simultaneously prioritize threat perception while also leading to decrements in performance, depending on the context and the task at hand. Motivation is also theorized to impact how executive processes are allocated, in an effort to maximize rewards or reduce punishment. This theory has been extended to include a neural model of how top-down processes impact threat perception (Pessoa & Adolphs, 2010). The multiple-wave model challenges the notion that there is an “automatic” subcortical pathway involving the superior colliculus, pulvinar, and amygdala that is specialized to process emotional stimuli. Instead, it posits that the amygdala receives input from frontal brain regions and allocates processing resources when salient, ambiguous, or unpredictable stimuli are sensed by modulating processing in networks involved in visual processing, via its broad connectivity with the visual cortex, prefrontal cortex, and other subcortical regions.

Impact of top-down factors on threat perception biases  Chapter | 10  219

Attentional control theory (Eysenck, Derakshan, Santos, & Calvo, 2007), in a similar vein, proposes that trait anxiety increases the attention allocated to threats and impairs attentional control. More specifically, trait anxiety is theorized to boost bottom-up processing of threats and impair top-down guidance of attention. This impairment is thought to occur as a result of reduced efficiency of the processing of nonthreatening stimuli, rather than anxiety simply leading to wide-ranging decrements in performance. This theory has been supported by empirical results demonstrating that accuracy on verbal and spatial memory tasks (Calvo, Eysenck, Ramos, & Jimenez, 1994; Ikeda, Iwanaga, & Seiwa, 1996; Markham & Darke, 1991) and reaction time on tasks involving visual attention, such as the emotional Stroop are not always negatively impacted by anxiety (Bishop, Duncan, & Lawrence, 2004; Compton et al., 2003; Whalen et al., 1998). Attentional control theory also posits that as demands on executive control increase, the influence of trait anxiety on performance also increases, leading to greater decrements in performance for nonthreatening stimuli (Ashcraft & Kirk, 2001; Eysenck, 1985). The aforementioned three theories can be applied to understand how functioning is altered in anxiety. While it is clear that anticipatory attention and expectation regarding threat influences perception, this is not well studied in anxiety disorders. Anticipation of negative future events is a cardinal feature of anxiety, including the overestimation of both the likelihood and negative consequences of future negative events (Aue & Okon-Singer, 2015; Grupe & Nitschke, 2013). Therefore, endogenous processes related to future expectations may play a crucial role in the development and maintenance of anxiety disorders. In a room full of cobwebs, a person with anxiety or spider-related fear may overestimate the likelihood of confronting a spider and/or the danger involved compared with a nonanxious person in the same room. As a result, anxious individuals may scan the environment for spiders and are likely to detect them faster, if present. Focusing on the anticipatory phase in anxiety has been suggested as an effective strategy to determine psychological and neurobiological factors involved in anxiety development and maintenance (Davis, Walker, Miles, & Grillon, 2010). Similarly, the uncertainty and anticipation model of anxiety posits that hypervigilance, or increased attention to expected threatening information can lead to both faster detection of threatening stimuli and a misinterpretation of neutral stimuli (Grupe & Nitschke, 2013). For example, anxiety caused by the threat of shock leads to (a) faster detection of negative stimuli (Robinson, Letkiewicz, Overstreet, Ernst, & Grillon, 2011), (b) interpretation of neutral faces as being negative in socially anxious individuals (Yoon & Zinbarg, 2008), and (c) interpretation of ambiguous interoceptive experiences as being negative in people high in anxiety sensitivity (Richards, Austin, & Alvarenga, 2001). This model also posits that inflated estimates of threat probability and cost can lead to improved performance (Paulus & Yu, 2012), via overweighting of low probability events, for example, inflating the likelihood of a rare negative outcome (Mukherjee, 2010).

220  Cognitive biases in health and psychiatric disorders

Theoretical considerations While not currently a part of a theory, it is important to note that prior knowledge regarding emotional stimuli is thought to influence subsequent perceptual decision-making via two documented mechanisms, (1) an attention-related mechanism, based on the salience of the upcoming stimulus, that increases sensitivity in perceptual discrimination and (2) an expectation-related mechanism that creates a prestimulus response bias, favoring the stimulus with a higher probability of occurrence (Summerfield et al., 2006; Summerfield & de Lange, 2014; Summerfield & Egner, 2016). The use of attention-related top-down processes, encouraged by tasks that involve feature-based, object-based, and spatial attention (Kanwisher & Wojciulik, 2000), may lead to more sensitive detection of subsequent, relevant stimuli. For example, information regarding upcoming target locations improved target detection at the attended location, even when saccadic eye movements did not overtly move to the target (Carrasco, Ling, & Read, 2004; Posner, Snyder, & Davidson, 1980). Similarly, knowledge regarding upcoming locations of threatening targets has been shown to improve their detection (Mohanty et al., 2009). Top-down spatial attention has also been shown to reduce the impact of salient distractors at noncued locations (Theeuwes, 1991; Yantis & Jonides, 1984). Stimuli presented in contexts in which they have a higher probability of occurring are also recognized more rapidly (Gold & Shadlen, 2007; Heekeren, Marrett, & Ungerleider, 2008) than objects in unexpected contexts (Bar, 2004; Brattico, Naatanen, & Tervaniemi, 2002; Enns & Lleras, 2008). For example, a classic study of contextual effects demonstrated that visual search in a kitchen scene led to faster identification of a loaf of bread than a drum (Palmer, 1975). Unexpected stimuli in complex scenes are typically detected more slowly and less accurately (Biederman, Mezzanotte, & Rabinowitz, 1982). It has thus been proposed that humans integrate incoming sensory information with prior knowledge regarding the likelihood of future events in a manner consistent with Bayes’ theorem (Bach & Dolan, 2012).

Methods used to investigate threat perception Research examining the psychological and neural mechanisms of threat perception has generally focused on bottom-up processing of threat, reinforcing the view that the perception of threat is automatic, and not subject to endogenous processes (Vuilleumier & Driver, 2007). This view has led to development of several paradigms in which emotional stimuli are unexpected, distracting, or irrelevant to the task at hand (Armony & Dolan, 2002; Holmes, Green, & Vuilleumier, 2005; Keil, Moratti, Sabatinelli, Bradley, & Lang, 2005; Mogg et al., 2000; Mogg & Bradley, 1999a; Stormark & Hugdahl, 1996, 1997). For example, the dot-probe task simultaneously presents an emotional and a neutral stimulus peripherally, and one of these two stimuli is followed by an attentional

Impact of top-down factors on threat perception biases  Chapter | 10  221

probe (Holmes et al., 2005; Mogg et al., 2000; Mogg & Bradley, 1999a). Faster responses to probes in threat-associated locations (valid trial) compared with neutral locations are interpreted as evidence of bias to threatening stimuli. However, as the importance of top-down processing has become more apparent, a number of studies examined the impact of prior knowledge and attention on threat perception. These studies have used paradigms in which emotional cues are presented centrally (rather than peripherally) and can be used endogenously or voluntarily to guide attention and perception (LoBue, 2014; Mohanty et al., 2009; Sussman et al., 2016; Sussman et al., 2017; Zeelenberg et al., 2006). To examine the anticipatory aspects of attentional bias, a cued visual probe task was recently developed and validated, demonstrating an anticipatory attentional bias to threat (Gladwin, Mobius, McLoughlin, & Tyndall, 2019). This new paradigm provides a measurement of the influence of top-down processes on attentional bias, therefore providing additional information compared with the classic version of the task. The classic dot-probe task was developed to examine attentional biases by measuring reaction time in response to a stimulus presented in an attended versus nonattended location. Each trial starts with the presentation of a pair of stimuli, typically one threatening and one neutral (Mogg & Bradley, 1999b). Subsequently, a target stimulus is presented in the same location as either the threatening or neutral stimulus. In the novel, cued dot probe, neutral symbols are paired with either neutral or angry faces (Gladwin et al., 2019). These symbols are subsequently presented on a screen and then followed by a probe and a nontarget stimulus. Probes in the same location as the cue that have been paired with the angry face are located more quickly. In addition to explicit cues, emotional contexts and states can guide attention and perception in a top-down manner. For example, spatial configurations or contexts predicting threat can be learned and used to improve the detection of embedded threats. For example, when target faces are shown in a field of neutral faces, they are detected more quickly when they have been shown in the same field of faces before, an effect that does not occur for neutral faces (Szekely, Rajaram, & Mohanty, 2017, 2019). Negative and angry mood inductions can lead to decreased perceptual sensitivity (Wormwood et  al., 2016) or increase the rate of identifying a neutral stimulus as threatening (Baumann & DeSteno, 2010). However, a fearful mood induction can speed up the perception of threatrelated, but not neutral, targets (LoBue, 2014). Overall, the view that emotional stimuli are processed in a bottom-up automatic manner has led to development of paradigms that do not examine the impact of goals or expectations on the perception of emotional stimuli. However, newer paradigms in which emotional information is provided in advance in the form of cues or contexts demonstrate that emotion-related top-down processes play a key role in perceptual processes (Aue et al., 2016; Aue et al., 2019; Aue, Guex, et al., 2013; Barrett & Simmons, 2015; Sussman et al., 2016; Szekely et al., 2019). Therefore, it is important for the field to continue developing paradigms in which emotional information is provided in both a top-down and

222  Cognitive biases in health and psychiatric disorders

bottom-up manner. It is necessary to process an emotional stimulus when it is relevant to one’s immediate well-being. However, if the same emotional stimulus does not indicate a salient threat, it is beneficial to be able to ignore it and focus on the task at hand. Therefore, truly adaptive behavior balances the interaction between endogenous and exogenous processing in real time.

Neural mechanisms involved in threat processing Neural mechanisms of bottom-up threat processing Bottom-up processing of emotional stimuli is thought to be mediated by the amygdala and its interactions with the sensory cortices (Cisler & Koster, 2010; Dolan, 2002; Ohman, 2002, 2005). Results from studies using backward masking paradigms suggest that fearful faces activate the amygdala in the absence of conscious awareness (Morris et al., 1998; Whalen et al., 1998). Vuilleumier and colleagues examined the possibility that emotion and voluntary attention are impacted by distinct influences. Using an event-related fMRI paradigm in which subjects fixated on a central cue and matched either two fearful or neutral faces or two houses presented eccentrically, they found that fearful face-related activity in the fusiform gyrus (FG)—a face-sensitive region—was modulated by voluntary attention, while amygdala responsiveness was not (Vuilleumier, Armony, Driver, & Dolan, 2001; Vuilleumier, Richardson, Armony, Driver, & Dolan, 2004). Along these lines, bottom-up appraisal of emotional stimuli has been associated with greater connectivity between the amygdala and limbic regions, such as the anterior cingulate cortex (Comte et al., 2014). The amygdala has been hypothesized to quickly detect threatening stimuli (Cunningham & Brosch, 2012; Sander, Grafman, & Zalla, 2003), via a subcortical pathway that (1) passes through the superior colliculi and the thalamus, (2) does not require cortical input (LeDoux, 2000), and (3) carries low-spatial frequency information (Vuilleumier, Armony, Driver, & Dolan, 2003). Supporting evidence has demonstrated that emotional faces can be processed without conscious awareness in a subject with a lesioned striate visual cortex (de Gelder, Vroomen, Pourtois, & Weiskrantz, 1999). Similarly, fearful faces presented in low-spatial frequency images, which are considered coarse information, produce more amygdala activation than the same images displayed with more detailed high-spatial frequency information (Vuilleumier et  al., 2003). These results lend support to the proposal that course information can be transmitted quickly to the amygdala for expedited processing. Recent supporting evidence using human intracranial electrophysiological data showed that the amygdala responds quickly, beginning as early as 74 ms after the onset of threatening facial expressions, much more rapidly than the FG (Mendez-Bertolo et al., 2016). This theory has been challenged by the multiple-wave model, which posits that the amygdala’s primary role is to allocate neural processing resources when a threat is present, and that the thalamic pulvinar nucleus plays a critical role in threat detection (Pessoa & Adolphs, 2010). This model has been

Impact of top-down factors on threat perception biases  Chapter | 10  223

supported by evidence that emotional faces produced greater amygdala activity only when they are actively attended, in other words, only when top-down processes are involved (Okon-Singer et al., 2014; Pessoa, Kastner, & Ungerleider, 2002). Evidence that rapid fear detection can rely on high-spatial frequency information, suggesting involvement of sensory cortices (Stein, Seymour, Hebart, & Sterzer, 2014) also supports the multiple waves model. Once the amygdala has identified information as threatening, it can increase neural processing in ­perception-related regions—for example, visual cortices—via re-entrant feedback (Ohman, 2005; Vuilleumier, 2005; Vuilleumier & Driver, 2007). The amygdala is a particularly well-positioned brain region to influence critical processes and behaviors, as it receives sensory information from all modalities and projects to a wide range of cortical and subcortical brain regions (Janak & Tye, 2015). The theory that the amygdala’s rich connectivity is employed during threat perception is supported by findings demonstrating that amygdala lesions led to less activation of face-sensitive brain regions during the viewing of threatening faces (Hadj-Bouziane et al., 2012; Vuilleumier et al., 2004). Corroborating evidence includes that greater connectivity between the amygdala and sensory cortices has been found during threat perception compared with the perception of neutral stimuli (Morris et al., 1998). Taken together, this body of research provides support for a top-down neural mechanism that perceptually prioritizes threatening stimuli via feedback from the amygdala. A ventral network of brain regions including the temporoparietal junction (TPJ) and the ventral frontal cortex (VFC) is another likely site for the integration of top-down and bottom-up signals. This circuit has been shown to aid in bottom-up threat perception: Salient task-irrelevant stimuli increase TPJ activity (Geng & Mangun, 2011), suggesting that this ventral circuit can shift focus to salient stimuli, even when this is in conflict with current task goals (Dolcos & McCarthy, 2006) or when the salient information source is beyond the current focus of sensory processing (Corbetta & Shulman, 2002). Moreover, interactions between top-down and bottom-up processes have also been observed in this network, suggesting that it is sensitive to both kinds of processing. For example, activity in the TJP can be suppressed by top-down attentional guidance (Shulman et  al., 2003; Shulman, Astafiev, McAvoy, d’Avossa, & Corbetta, 2007).

Neural mechanisms of top-down threat processing The predictive coding theory suggests a neurobiologically plausible mechanism by which top-down factors could aid the perception of emotional stimuli. According to this theory, rather than passively absorbing incoming sensory information, the brain actively predicts what will be perceived, aiding perception (Mumford, 1992; Rao & Ballard, 1999). In this framework, visual perception can be understood in terms of higher-order brain regions feeding back predictions regarding future perceptions to subsequent lower level brain regions where

224  Cognitive biases in health and psychiatric disorders

they are matched with sensory evidence, and prediction errors are then fed ­forward (Lee & Mumford, 2003; Mesulam, 2008; Mumford, 1992; Rao & Ballard, 1999). One possibility is that top-down information is utilized by forming a perceptual template of the anticipated stimulus in higher-order areas, to which subsequent bottom-up input can then be matched (Friston, 2005; Mumford, 1992; Rao & Ballard, 1999). Consistent with this view, studies comparing patterns of brain activity evoked by internal stimulus representations to patterns of activity in response to a directly viewed stimulus have found significant similarity during several different types of tasks, including paradigms that examined (1) preparatory attention (Peelen, Fei-Fei, & Kastner, 2009; Stokes, Thompson, Nobre, & Duncan, 2009; Zelano, Mohanty, & Gottfried, 2011), (2) expectation (Kok, Failing, & de Lange, 2014), and (3) mental imagery (Albers, Kok, Toni, Dijkerman, & de Lange, 2013; Lee, Kravitz, & Baker, 2012). Another possibility is that top-down information enhances perception by increasing baseline activity in visual sensory areas tuned to the expected stimulus, thereby biasing perceptual decision-making (Esterman & Yantis, 2010; Kok, Jehee, & de Lange, 2012; Puri, Wojciulik, & Ranganath, 2009). For example, expectation during a cued face/house discrimination task increases blood oxygen level-dependent (BOLD) signal in face- and house-specific visual cortical areas (Esterman & Yantis, 2010). Similarly, anticipation of forthcoming stimuli has been found to increase prestimulus neuronal activity in sensory regions associated with the perception of the anticipated stimulus and decision-related brain regions (Kastner, Pinsk, De Weerd, Desimone, & Ungerleider, 1999; Ress, Backus, & Heeger, 2000; Summerfield & de Lange, 2014). More specifically, neurons in the inferior temporal lobe that encode expected stimuli increase activation in response to a predictive cue (Erickson & Desimone, 1999). Similarly, neurons in regions sensitive to object motion become activated before the presentation of a predicted stimulus in motion (Albright, 2012; Sakai & Miyashita, 1991). Furthermore, greater prestimulus BOLD signal in fusiform face area predicts subjects’ report of seeing the Rubin’s vase illusion as a face rather than as a vase (Hesselmann, Kell, Eger, & Kleinschmidt, 2008), and face-related cues increase BOLD signal in face-related areas before face-stimulus onset (Bar et al., 2001; Esterman & Yantis, 2010; Puri et al., 2009). Cues predicting where a target will appear can bias the starting point of decision-related oscillatory activity in motor cortex before the accumulation of sensory evidence (de Lange, Rahnev, Donner, & Lau, 2013). Furthermore, attention to a target created by prestimulus cues activates target-specific representations in stimulus-specific brain regions, creating a preparatory bias for the attended stimulus (Zelano et al., 2011). Patterns of brain activity evoked by maintaining a stimulus in working memory are similar to patterns of brain activity when directly viewing the same stimulus (Harrison & Tong, 2009; Serences, Ester, Vogel, & Awh, 2009; Sreenivasan, Sambhara, & Jha, 2011). While these studies used relatively neutral stimuli, prestimulus activity prior to stimulus onset may also describe a neural mechanism by which emotional stimuli are perceptually prioritized.

Impact of top-down factors on threat perception biases  Chapter | 10  225

For example, one study used threat-related prestimulus cues to determine how perceptual biasing is instantiated in the brain. At the start of each trial, the letter “F,” indicating an upcoming decision regarding fearful faces, or the letter “N,” indicating a decision regarding neutral faces, was presented. Threatrelated cues increased activation in the superior temporal sulcus, a region sensitive to emotional faces, both when while viewing the cue, and when viewing the stimulus (Sussman et al., 2017). Furthermore, greater activation to threatrelated cues compared with neutral cues predicted improved task performance (Sussman et al., 2017). Threat-cues also led to stimulus-triggered increases in amygdala activity for threatening, but not neutral, faces. In addition to preparatory processing seen in relevant sensory cortices, prior emotional and motivational information may modulate perception via higher-order prefrontal and parietal brain regions. Endogenous biasing of spatial attention involves a frontoparietal network including the intraparietal sulcus in the posterior parietal cortex (PPC) and the frontal eye fields (FEF), the anterior cingulate cortex, and supplementary motor area (ACC/SMA; (Corbetta & Shulman, 2002; Gitelman et al., 1999; Kastner et al., 1999; Kastner, De Weerd, Desimone, & Ungerleider, 1998; Mesulam, 1981, 1999; Reynolds, Chelazzi, & Desimone, 1999). This spatial attention network is thought to produce a search template that integrates the salience of the expected stimulus and likely spatial coordinates, biasing sensory brain regions toward perception of the sought-after object (Egner, 2008; Gottlieb, 2007; Thompson & Bichot, 2005). One study manipulated the guidance of attention via predictive cues offering information about the location and emotional salience of a subsequent stimulus (Mohanty et al., 2009). Both cues led to improved target detection, indicating that, like cues indicating spatial locations of targets, emotional cues can lead to endogenously driven improvements in target detection. Brain imaging demonstrated that the presentation of spatially informative cues prior to stimulus presentation activated the frontoparietal spatial attention network including the intraparietal sulcus and frontal eye fields (FEF), as well as the FG. Cues related to emotional salience were associated with increased neural activity in regions associated with the processing of emotional stimuli, including the amygdala. Importantly, additive effects of spatial and emotional cues were observed in intraparietal sulcus, FEF and FG, suggesting a mechanism by which emotional and location information is integrated to direct attention in space. Furthermore, this study demonstrated that threat-related cues initiated input from the amygdala to the spatial attention network and inferotemporal visual areas, improving threat detection. Overall, these studies highlight neural mechanisms involving prefrontal, parietal, and sensory cortices, as well as their interactions with limbic regions, as the means by which emotional information can guide attention and perception in a top-down manner. However, other studies have found that threat-related cues do not improve subsequent detection of threating images, while neutral stimuli receive a perceptual boost from neutral predictive cues (Aue et al., 2019).

226  Cognitive biases in health and psychiatric disorders

Peripheral mechanisms of threat processing The relationship between top-down processes, threat perception, and somatovisceral responses has largely been examined in anxiety disorders, as anxiety is thought to be characterized both by exaggerated negative predictions about future events (Grupe & Nitschke, 2013) and by increased physiological arousal in response to threats (Clark & Watson, 1991). However, some studies have been performed in nonanxious populations. For example, several studies have found that expectancy biases for threatening stimuli are associated with increased skin conductance (Amin & Lovibond, 1997; Diamond, Matchett, & Davey, 1995; Honeybourne, Matchett, & Davey, 1993). Furthermore, in both individuals with and without specific phobia, top-down processes evoked by cues indicating the likelihood of seeing a bird vs a spider on the subsequent trial lead to changes in pupil diameter and heart rate when the target to-be-detected was a bird (Aue et  al., 2016; Aue, Guex, et  al., 2013). Comparable top-down processes were not visible in these studies when the target was a spider, however. This finding points to the necessity of identifying potential preconditions or moderators of top-down influences on threat perception. Studies examining somatovisceral responses in anxiety have demonstrated that anxiety is associated with physiological hyperarousal. Both elevated baseline arousal and exaggerated reactivity to stressors have been found in anxious adults (Harrison & Turpin, 2003; Sarlo, Palomba, Angrilli, & Stegagno, 2002; Thayer, Friedman, Borkovec, Johnsen, & Molina, 2000). Similarly, in children, increased heart rate (HR) and skin conductance reactivity (SCR) are associated with self-report of anxiety (Weems, Zakem, Costa, Cannon, & Watts, 2005). Furthermore, increased startle response to images of feared stimuli has been found in adults with specific phobia (Globisch, Hamm, Esteves, & Ohman, 1999; Hamm, Cuthbert, Globisch, & Vaitl, 1997; Sabatinelli, Bradley, & Lang, 2001; Wendt, Lotze, Weike, Hosten, & Hamm, 2008). Importantly, avoidance in individuals with spider phobia, indexed by shorter gaze duration, is associated with increased HR and SCR (Aue et al., 2016; Aue, Hoeppli, Piguet, Sterpenich, & Vuilleumier, 2013). These results shed light on how physiological arousal contributes to a cardinal symptom and key maintenance factor in anxiety disorders. Techniques used in the treatment of anxiety disorders have been found to have an impact on physiological arousal. For example, reappraisal of physiological arousal leads to cardiovascular outcomes that are associated with more positive outcomes (increased cardiac efficiency and reduced vascular resistance), as well as decreased attentional bias toward threat (Jamieson, Nock, & Mendes, 2012). Similarly, an attentional training paradigm that taught participants with social phobia to attend to nonthreatening, compared with threatening, cues reduced SCR (Heeren, Reese, McNally, & Philippot, 2012).

Threat processing in anxiety versus healthy populations Anxiety is distinguished from fear by the proximity of the threat causing distress. Whereas anxiety is conceptualized as a response to future threats, fear is

Impact of top-down factors on threat perception biases  Chapter | 10  227

thought of as an immediate response to a present stimulus (Davis et al., 2010; Grupe & Nitschke, 2013; LeDoux, 2015). Several studies show that anxiety is characterized by increased physiological arousal during anticipation of future threats (for a review, see Grillon, 2008). Despite the importance of anticipatory arousal in anxiety, threat perception in anxiety is typically examined by measuring response to the stimulus itself. Furthermore, these stimuli are often presented in paradigms designed to tap the bottom-up aspects of threat processing (Mathews & MacLeod, 1994; Ohman & Mineka, 2001). For example, in past studies of threat processing in anxiety, threatening stimuli “pop out” among nonemotional stimuli (Fox et al., 2000; Ohman et al., 2001), are peripheral to fixation (Mogg & Bradley, 1999a), appear in a rapid stream of images (Arend & Botella, 2002), or are irrelevant to task at hand (Williams, Mathews, & MacLeod, 1996). However, consistent with the conceptualization of anxiety as anticipatory in nature, empirical evidence demonstrates that anxiety can impact perception via top-down processes. One study showed that prestimulus threat-related cues improved perceptual sensitivity as a function of the level of trait anxiety and current levels of induced anxiety (Sussman et al., 2016). In this study, two groups of participants varying in levels of trait anxiety identified degraded emotional and neutral stimuli in a cued two-alternative forced-choice perceptual discrimination task. One group completed the task under threat of shock (induced anxiety), while the other group performed the task in the absence of shock. Higher levels of trait anxiety were associated with larger gains in perceptual sensitivity in the presence of shock but were associated with worse perceptual sensitivity in the absence of shock. Therefore, the utilization of threat-related information to guide threat perception in anxiety is most efficient at high levels of trait and situationally induced anxiety. Aue, Guex, et al. (2013), Aue, Hoeppli, et al. (2013), and Aue et al. (2016) manipulated the probability of the presentation of upcoming threatening and neutral targets to examine whether threat expectancy (inflated estimates of threat probability) impacts visual search for a spider or a bird in individuals with or without spider fear. In two studies, expectancy was manipulated via prestimulus cues indicating a 90% or 50% likelihood of seeing a spider or a bird. Cues indicating spider and bird influence the speed of the detection of bird targets, but likelihood did not have a significant impact on the speed of spider detection. (Aue et al., 2016; Aue, Guex, et al., 2013). Therefore, top-down manipulation of expectation had an impact on nonthreatening, but not on threatening targets. Notably, however, those participants who were characterized by high fear spiders were particularly fast in detecting the spiders (irrespective of the cues that had been presented). Taken together, the Sussman et al.’s (2016, 2017) and Aue et al.’s (2013, 2016) studies not only indicate that anxiety influences the impact of top-down processing in threat perception but also suggest that prestimulus cues may have varying impact on threat perception depending on the nature of the task.

228  Cognitive biases in health and psychiatric disorders

Maintaining task-relevant representations in working memory is required to use prior threat-related information (Sreenivasan et  al., 2011). Since representations of emotional stimuli are maintained with more vividness than neutral representations (Bywaters, Andrade, & Turpin, 2004), they may utilize more working memory resources. Empirical evidence suggests that trait anxiety has a negative impact on the efficiency on tasks requiring the inhibition function (Calvo & Eysenck, 1996; Yiend & Mathews, 2001), the shifting function (Gopher, Armony, & Greenshpan, 2000), and to a certain extent the updating function (Duff & Logie, 2001). Therefore, individuals high in trait anxiety may perform worse than nonanxious individuals on tasks that require the maintenance of threat-related information in working memory or that encourage the unnecessary entry of threat-related information into working memory, particularly when state anxiety is not induced, as some empirical evidence suggests that anxious individuals can perform better under stress (Stout, Shackman, & Larson, 2013; Sussman et al., 2016).

Neural mechanisms of threat processing in anxiety versus healthy populations In general, the amygdala is more active for people with anxiety disorders compared with healthy controls (Williams et  al., 2015). This finding extends to anticipation; cue-related activity in the amygdala is greater for individuals with generalized anxiety disorder than for healthy individuals, when cues signified both upcoming threatening and neutral images (Nitschke et  al., 2009). Hyperactivation has also been found in other brain regions in anxiety. When asked to imagine giving a public speech, individuals with social phobia had increased activity in limbic and paralimbic regions compared with healthy controls (Lorberbaum et al., 2004). Similarly, anticipating giving a public speech led to increased limbic activation and decreased striatal activation in socially anxious individuals (Boehme et al., 2014) and to reduced functional connectivity between limbic regions and cortical regions involved in emotion regulation (Cremers et al., 2015). Similar to limbic activity described earlier, very early sensory processing is increased in individuals with clinical anxiety (Knott et al., 1994) and in children with higher levels of dispositional anxiety (Woodward et al., 2001). Enhanced sensory-perceptual functions have been observed in both high trait anxiety and induced anxiety (Robinson, Vytal, Cornwell, & Grillon, 2013) and could explain improved perceptual sensitivity in threatening contexts in anxiety. For example, the threat of shock has been found to change neural processing to a sensory-vigilance mode that prioritizes threatening stimuli (Arnsten, 2009; Shackman, Maxwell, McMenamin, Greischar, & Davidson, 2011). Activation of sensory cortices has also been found when anticipating a feared stimulus in individuals with specific phobia. For example, expectancies related of encountering a phobogenic stimulus (e.g., a spider) was associated with a distinct pat-

Impact of top-down factors on threat perception biases  Chapter | 10  229

tern of brain activation in the precuneus, lateral prefrontal cortex, and primary visual cortex (Aue et al., 2015). Overall, these studies demonstrate that, in anxiety, prestimulus anticipation is associated with greater activity in limbic, sensory, and prefrontal regions and their connectivity prior to the threatening stimulus. However, the role of this anticipatory activity and connectivity in heightening perception and attention to subsequently presented threatening targets can be better characterized by future studies aimed at determining the neural mechanisms underlying top-down guided threat perception in anxiety. To better examine how top-down threat-related information is utilized in anxiety, it is important to consider the research showing that anxiety is associated with reduced recruitment of prefrontal brain regions involved in top-down control. In healthy populations, predictive representations are maintained in the dorsal and ventral medial prefrontal cortices (DMPFC and VMPFC; Summerfield et al., 2006). However, anxiety is associated with poorer recruitment of DMPFC (Shin et al., 2005) and dorsolateral prefrontal cortex (DLPFC; Bishop, 2009), which may lead to an impaired ability to maintain and utilize threat-related perceptual templates in the process of threat perception. For example, while low-trait-anxious individuals benefit from cues preceding a visual search task, individuals high in trait anxiety cannot use these cues as effectively (Berggren & Derakshan, 2013). Similarly, individuals higher in trait anxiety have decrements in perceptual sensitivity for threatening but not neutral cues (Sussman et al., 2016). Anxiety-related decrements in performance increase with greater task demands on the central executive (Eysenck et  al., 2007). Therefore, the maintenance of a perceptual set for threatening stimuli, which might be more complex than for neutral stimuli, could be more demanding. Overall, further research needs to be conducted to understand how greater anticipation of threat interacts with worse working memory capacity for maintaining these threat-related representations and its effect on threat perception.

Limitations While there is increasing attention to the importance of top-down processes in threat perception, in both normal function and anxiety, a deeper understanding of how endogenous mechanisms impact perception is required. Designing paradigms that manipulate top-down expectation and attention to threatening stimuli is a complex process. This may explain why these types of paradigms are used less often to study threat perception in anxiety than study designs that compare behavioral and neural response with threatening stimuli in anxious and healthy populations. Furthermore, the impact of top-down factors such as moods, motivational states, and nonspatial contexts on prestimulus processing can be challenging to quantify effectively. Preliminary research suggests that limbic, paralimbic, and prefrontal regions are involved with the guidance of threat perception by top-down control. However, the role of this activity in in-

230  Cognitive biases in health and psychiatric disorders

fluencing attention and perception in anxiety is not well characterized. It is also unclear whether increased anticipatory brain activity is associated with a more negative interpretation of cues regarding the salience and the likelihood of an upcoming threatening stimulus.

Future directions While empirical research has begun to characterize the influence of endogenous factors on threat perception, further elucidation of the differential impact of different types of threat-related top-down mechanisms on perception will propel the field forward. For example, future work could distinguish the impact of prestimulus attention, prioritizing perception based on the salience of the anticipated stimulus to current goals, from the impact of prestimulus expectation, prioritizing perception based on the estimated likelihood of the anticipated stimulus. This distinction could provide more fine-grained information about how top-down factors influence perception in both normal function and in anxiety. Computational modeling of both behavior and neural signals could allow us to examine the specific mechanisms that contribute to the perceptual prioritization. For example, computational modeling of behavioral and neural pre- and poststimulus responses could provide information regarding the differential contribution of pre- and poststimulus biases to top-down guided perceptual prioritization of threatening stimuli. Using model-based techniques to examine neural data such as multivariate pattern analysis or dynamic causal modeling could also add significantly to our understanding of how top-down processes contribute to threat perception. Future research should also aim to better isolate the impact of top-down from bottom-up factors on threat perception, as results could advance our understanding of how these factors interact to perceptually prioritize emotional stimuli. While several studies have manipulated top-down task-driven factors and bottom-up stimulus-driven factors in the same experiment, the top-down factors tend to be cognitive in nature, and the bottom-up variables tend to be emotional in nature. Hence, it remains unclear how top-down and bottom-up threat-related manipulations interact to influence threat perception. Distinctions between pre- and poststimulus factors in threat perception are of particular importance in anxiety disorders, since inaccurate predictions regarding the likelihood and costs of future negative events are thought to underlie these disorders (Grupe & Nitschke, 2013). Improved understanding how anxiety influences the interaction between top-down and bottom-up factors could link differences in basic perceptual processes with clinical symptoms and could foster the development of novel treatments and prevention strategies for clinical anxiety. For example, by leveraging top-down processes, a form of ABMT was designed to enhance vigilance for threat in soldiers before they were exposed to

Impact of top-down factors on threat perception biases  Chapter | 10  231

combat. This treatment was found to reduce posttraumatic stress and depression symptoms following combat (Wald et al., 2016).

Summary and conclusions The preponderance of past research examining the perceptual prioritization of threatening stimuli has focused on relatively automatic aspects of threat perception. However, an emerging body of literature in the field of visual perception demonstrates that top-down processes can play an important role in the perception of threatening stimuli. For example, prestimulus emotional cues have been shown to improve target detection (Sussman et al., 2016; Sussman et al., 2017; Wormwood et al., 2016). Furthermore, task-relevant emotional stimuli can guide attention when consistent with current goals without impairing performance when they are irrelevant to task goals (Hahn & Gronlund, 2007; LichtensteinVidne et al., 2012; Lichtenstein-Vidne et al., 2017; Van Dessel & Vogt, 2012; Vogt et al., 2013; Williams et al., 2005). Furthermore, motivational factors, such as rewards, also influence how attention is allocated (Padmala et al., 2017). In terms of peripheral mechanisms, biases for threatening stimuli are associated with increased skin conductance (Amin & Lovibond, 1997; Diamond et  al., 1995; Honeybourne et al., 1993). While the studies cited earlier have provided evidence of top-down guidance of threat perception, in some cases, prestimulus cues can aid more in the perceptual prioritization of neutral compared with threatening stimuli (Aue et al., 2016; Aue et al., 2019; Aue, Guex, et al., 2013). The perceptual prioritization of threatening stimuli may be instantiated in the brain via the formation of perceptual templates of the anticipated stimulus in higher-order brain regions, against which incoming sensory information may be matched (Friston, 2005; Mumford, 1992; Rao & Ballard, 1999). Alternatively, top-down processes could enhance perception by increasing baseline activity in relevant sensory brain areas and biasing perceptual decision toward the expected stimulus (Esterman & Yantis, 2010; Kok et al., 2012; Puri et al., 2009). Threat-related cues have been shown to increase both cue- and stimulus-related brain activation in sensory brain regions, and the level of cue-related activation in these regions predicts subsequent task performance (Sussman et al., 2017). Threat-related cues also lead to stimulus-related increases in amygdala activity for threatening, but not neutral, stimuli (Sussman et al., 2017). Anxiety influences the guidance of perception via top-down processes, and therefore may significantly impact how top-down processes aid in the perception of threat. For example, one study found that higher levels of trait anxiety were associated with larger gains in perceptual sensitivity when trials started with a threat-related cue in the presence of shock but were associated with worse perceptual sensitivity in the absence of shock (Sussman et  al., 2016). These results suggest that, for individuals high in trait anxiety, perception of visually degraded faces, guided via threat-related cues, is most efficient when situational anxiety is present. However, using a cued visual search paradigm,

232  Cognitive biases in health and psychiatric disorders

Aue, Guex, et  al. (2013), Aue, Hoeppli, et  al. (2013), and Aue et  al. (2016) did not find differences in top-down guided perception of threatening stimuli in anxious or nonanxious individuals. Therefore, the influence of top-down ­factors on threat perception in anxiety may depend on the type of task at hand. In individuals with specific phobia, the activation of sensory cortices has been found when anticipating a feared stimulus. For example, expectancies related to encountering a phobogenic stimulus (e.g., a spider) was associated with a distinct pattern of brain activation in the precuneus, lateral prefrontal cortex, and primary visual cortex (Aue et al., 2015). These results suggest that anticipatory amygdala activity may, in concert with other key areas names earlier, influence threat perception in anxiety. Overall, the current body of research demonstrates that top-down factors can influence threat perception and suggest that this process may be altered in anxious individuals. Therefore, it is important to include top-down factors in future empirical examinations and models of threat perception and to further characterize the influence that top-down processes exert on the perception of threat, both in anxiety and in healthy populations.

References Albers, A. M., Kok, P., Toni, I., Dijkerman, H. C., & de Lange, F. P. (2013). Shared representations for working memory and mental imagery in early visual cortex. Current Biology, 23(15), 1427–1431. Albright, T. D. (2012). On the perception of probable things: Neural substrates of associative memory, imagery, and perception. Neuron, 74(2), 227–245. Amin, J. M., & Lovibond, P. F. (1997). Dissociations between covariation bias and expectancy bias for fear-relevant stimuli. Cognition and Emotion, 11(3), 273–289. Anderson, A. K. (2005). Affective influences on the attentional dynamics supporting awareness. Journal of Experimental Psychology-General, 134(2), 258–281. Arend, I., & Botella, J. (2002). Emotional stimuli reduce the attentional blink in sub-clinical anxious subjects. Psicothema, 14(2), 209–214. Armony, J. L., & Dolan, R. J. (2002). Modulation of spatial attention by fear-conditioned stimuli: An event-related fMRI study. Neuropsychologia, 40(7), 817–826. Arnsten, A. F. (2009). Stress signalling pathways that impair prefrontal cortex structure and function. Nature Reviews Neuroscience, 10(6), 410–422. Ashcraft, M. H., & Kirk, E. P. (2001). The relationships among working memory, math anxiety, and performance. Journal of Experimental Psychology-General, 130(2), 224–237. Aue, T., Chauvigne, L. A., Bristle, M., Okon-Singer, H., & Guex, R. (2016). Expectancy influences on attention to threat are only weak and transient: Behavioral and physiological evidence. Biological Psychology, 121(Part B), 173–186. Aue, T., Guex, R., Chauvigne, L., & Okon-Singer, H. (2013). Varying expectancies and attention bias in phobic and non-phobic individuals. Frontiers in Human Neuroscience. Aue, T., Guex, R., Chauvigne, L. A. S., Okon-Singer, H., & Vuilleumier, P. (2019). Expectancies influence attention to neutral but not necessarily to threatening stimuli: An fMRI study. Emotion, 19(7), 1244–1258. Aue, T., Hoeppli, M. E., Piguet, C., Hofstetter, C., Rieger, S. W., & Vuilleumier, P. (2015). Brain systems underlying encounter expectancy bias in spider phobia. Cognitive, Affective, & Behavioral Neuroscience, 15(2), 335–348.

Impact of top-down factors on threat perception biases  Chapter | 10  233 Aue, T., Hoeppli, M. E., Piguet, C., Sterpenich, V., & Vuilleumier, P. (2013). Visual avoidance in phobia: Particularities in neural activity, autonomic responding, and cognitive risk evaluations. Frontiers in Human Neuroscience, 7, 194. Aue, T., & Okon-Singer, H. (2015). Expectancy biases in fear and anxiety and their link to biases in attention. Clinical Psychology Review, 42, 83–95. Bach, D. R., & Dolan, R. J. (2012). Knowing how much you don’t know: A neural organization of uncertainty estimates. Nature Reviews Neuroscience, 13(8), 572–586. Bacon, W. F., & Egeth, H. E. (1994). Overriding stimulus-driven attentional capture. Perception & Psychophysics, 55(5), 485–496. Bannerman, R. L., Milders, M., de Gelder, B., & Sahraie, A. (2009). Orienting to threat: Faster localization of fearful facial expressions and body postures revealed by saccadic eye movements. Proceedings of the Royal Society B-Biological Sciences, 276(1662), 1635–1641. Bar, M. (2004). Visual objects in context. Nature Reviews Neuroscience, 5(8), 617–629. Bar, M., Tootell, R. B., Schacter, D. L., Greve, D. N., Fischl, B., Mendola, J. D., et al. (2001). Cortical mechanisms specific to explicit visual object recognition. Neuron, 29(2), 529–535. Bar-Haim, Y., Lamy, D., Pergamin, L., Bakermans-Kranenburg, M. J., & van Ijzendoorn, M. H. (2007). Threat-related attentional bias in anxious and nonanxious individuals: A meta-analytic study. Psychological Bulletin, 133(1), 1–24. Barrett, L. F., & Simmons, W. K. (2015). Interoceptive predictions in the brain. Nature Reviews Neuroscience, 16(7), 419–429. Baumann, J., & DeSteno, D. (2010). Emotion guided threat detection: Expecting guns where there are none. Journal of Personality and Social Psychology, 99(4), 595–610. Berggren, N., & Derakshan, N. (2013). Attentional control deficits in trait anxiety: Why you see them and why you don’t. Biological Psychology, 92(3), 440–446. Biederman, I., Mezzanotte, R. J., & Rabinowitz, J. C. (1982). Scene perception: Detecting and judging objects undergoing relational violations. Cognitive Psychology, 14(2), 143–177. Bishop, S. J. (2009). Trait anxiety and impoverished prefrontal control of attention. Nature Neuroscience, 12(1), 92–98. Bishop, S. J., Duncan, J., & Lawrence, A. D. (2004). State anxiety modulation of the amygdala response to unattended threat-related stimuli. Journal of Neuroscience, 24(46), 10364– 10368. Boehme, S., Ritter, V., Tefikow, S., Stangier, U., Strauss, B., Miltner, W. H., et al. (2014). Brain activation during anticipatory anxiety in social anxiety disorder. Social Cognitive and Affective Neuroscience, 9(9), 1413–1418. Brattico, E., Naatanen, R., & Tervaniemi, M. (2002). Context effects on pitch perception in musicians and nonmusicians: Evidence from event-related-potential recordings. Music Perception, 19(2), 199–222. Brosch, T., Pourtois, G., & Sander, D. (2010). The perception and categorisation of emotional stimuli: A review. Cognition & Emotion, 24(3), 377–400. Bywaters, M., Andrade, J., & Turpin, G. (2004). Determinants of the vividness of visual imagery: The effects of delayed recall, stimulus affect and individual differences. Memory, 12(4), 479–488. Calvo, M. G., & Eysenck, M. W. (1996). Phonological working memory and reading in test anxiety. Memory, 4(3), 289–305. Calvo, M. G., Eysenck, M. W., Ramos, P. M., & Jimenez, A. (1994). Compensatory reading strategies in test anxiety. Anxiety Stress and Coping, 7(2), 99–116. Carrasco, M., Ling, S., & Read, S. (2004). Attention alters appearance. Nature Neuroscience, 7(3), 308–313.

234  Cognitive biases in health and psychiatric disorders Carretie, L., Mercado, F., Tapia, M., & Hinojosa, J. A. (2001). Emotion, attention, and the ‘negativity bias’, studied through event-related potentials. International Journal of Psychophysiology, 41(1), 75–85. Chen, X., & Zelinsky, G. J. (2006). Real-world visual search is dominated by top-down guidance. Vision Research, 46(24), 4118–4133. Cisler, J. M., Bacon, A. K., & Williams, N. L. (2009). Phenomenological characteristics of attentional biases towards threat: A critical review. Cognitive Therapy and Research, 33(2), 221–234. Cisler, J. M., & Koster, E. H. (2010). Mechanisms of attentional biases towards threat in anxiety disorders: An integrative review. Clinical Psychology Review, 30(2), 203–216. Clark, L. A., & Watson, D. (1991). Tripartite model of anxiety and depression: Psychometric evidence and taxonomic implications. Journal of Abnormal Psychology, 100(3), 316–336. Compton, R. J., Banich, M. T., Mohanty, A., Milham, M. P., Herrington, J., Miller, G. A., et  al. (2003). Paying attention to emotion: An fMRI investigation of cognitive and emotional Stroop tasks. Cognitive, Affective, & Behavioral Neuroscience, 3(2), 81–96. Comte, M., Schon, D., Coull, J. T., Reynaud, E., Khalfa, S., Belzeaux, R., et al. (2014). Dissociating bottom-up and top-down mechanisms in the cortico-limbic system during emotion processing. European Psychiatry, 29. Corbetta, M., & Shulman, G. L. (2002). Control of goal-directed and stimulus-driven attention in the brain. Nature Reviews Neuroscience, 3(3), 201–215. Cremers, H. R., Veer, I. M., Spinhoven, P., Rombouts, S. A., Yarkoni, T., Wager, T. D., et al. (2015). Altered cortical-amygdala coupling in social anxiety disorder during the anticipation of giving a public speech. Psychological Medicine, 45(7), 1521–1529. Cunningham, W. A., & Brosch, T. (2012). Motivational salience: Amygdala tuning from traits, needs, values, and goals. Current Directions in Psychological Science, 21(1), 54–59. Davis, M., Walker, D. L., Miles, L., & Grillon, C. (2010). Phasic vs sustained fear in rats and humans: Role of the extended amygdala in fear vs anxiety. Neuropsychopharmacology, 35(1), 105–135. de Gelder, B., Vroomen, J., Pourtois, G., & Weiskrantz, L. (1999). Non-conscious recognition of affect in the absence of striate cortex. Neuroreport, 10(18), 3759–3763. de Lange, F. P., Rahnev, D. A., Donner, T. H., & Lau, H. (2013). Prestimulus oscillatory activity over motor cortex reflects perceptual expectations. Journal of Neuroscience, 33(4), 1400–1410. Della Libera, C., & Chelazzi, L. (2006). Visual selective attention and the effects of monetary rewards. Psychological Science, 17(3), 222–227. Diamond, D., Matchett, G., & Davey, G. C. (1995). The effect of prior fear levels on UCS-­expectancy ratings to a fear-relevant stimulus. The Quarterly Journal of Experimental P ­ sychology. A, 48(1), 237–247. Dijksterhuis, A., & Aarts, H. (2003). On wildebeests and humans: The preferential detection of negative stimuli. Psychological Science, 14(1), 14–18. Dolan, R. J. (2002). Emotion, cognition, and behavior. Science, 298(5596), 1191–1194. Dolcos, F., & McCarthy, G. (2006). Brain systems mediating cognitive interference by emotional distraction. Journal of Neuroscience, 26(7), 2072–2079. Duff, S. C., & Logie, R. H. (2001). Processing and storage in working memory span. The Quarterly Journal of Experimental Psychology. A, 54(1), 31–48. Egner, T. (2008). Multiple conflict-driven control mechanisms in the human brain. Trends in Cognitive Sciences, 12(10), 374–380. Eimer, M., & Holmes, A. (2007). Event-related brain potential correlates of emotional face processing. Neuropsychologia, 45(1), 15–31. Engelmann, J. B., & Pessoa, L. (2007). Motivation sharpens exogenous spatial attention. Emotion, 7(3), 668–674.

Impact of top-down factors on threat perception biases  Chapter | 10  235 Enns, J. T., & Lleras, A. (2008). What’s next? New evidence for prediction in human vision. Trends in Cognitive Sciences, 12(9), 327–333. Erickson, C. A., & Desimone, R. (1999). Responses of macaque perirhinal neurons during and after visual stimulus association learning. Journal of Neuroscience, 19(23), 10404–10416. Esterman, M., & Yantis, S. (2010). Perceptual expectation evokes category-selective cortical activity. Cerebral Cortex, 20(5), 1245–1253. Eysenck, M. W. (1985). Anxiety and cognitive-task performance. Personality and Individual Differences, 6(5), 579–586. Eysenck, M. W., Derakshan, N., Santos, R., & Calvo, M. G. (2007). Anxiety and cognitive performance: Attentional control theory. Emotion, 7(2), 336–353. Fox, E., Lester, V., Russo, R., Bowles, R. J., Pichler, A., & Dutton, K. (2000). Facial expressions of emotion: Are angry faces detected more efficiently? Cognition & Emotion, 14(1), 61–92. Friston, K. (2005). A theory of cortical responses. Philosophical Transactions of the Royal Society B-Biological Sciences, 360(1456), 815–836. Geng, J. J., & Mangun, G. R. (2011). Right temporoparietal junction activation by a salient contextual cue facilitates target discrimination. NeuroImage, 54(1), 594–601. Gitelman, D. R., Nobre, A. C., Parrish, T. B., LaBar, K. S., Kim, Y. H., Meyer, J. R., et al. (1999). A large-scale distributed network for covert spatial attention—Further anatomical delineation based on stringent behavioural and cognitive controls. Brain, 122, 1093–1106. Gladwin, T. E., Mobius, M., McLoughlin, S., & Tyndall, I. (2019). Anticipatory versus reactive spatial attentional bias to threat. British Journal of Psychology, 110, 3–14. Globisch, J., Hamm, A. O., Esteves, F., & Ohman, A. (1999). Fear appears fast: Temporal course of startle reflex potentiation in animal fearful subjects. Psychophysiology, 36(1), 66–75. Gold, J. I., & Shadlen, M. N. (2007). The neural basis of decision making. Annual Review of Neuroscience, 30, 535–574. Gopher, D., Armony, L., & Greenshpan, Y. (2000). Switching tasks and attention policies. Journal of Experimental Psychology-General, 129(3), 308–339. Gottlieb, J. (2007). From thought to action: The parietal cortex as a bridge between perception, action, and cognition. Neuron, 53(1), 9–16. Grillon, C. (2008). Models and mechanisms of anxiety: Evidence from startle studies. Psychopharmacology, 199(3), 421–437. Grupe, D. W., & Nitschke, J. B. (2013). Uncertainty and anticipation in anxiety: An integrated neurobiological and psychological perspective. Nature Reviews Neuroscience, 14(7), 488–501. Hadj-Bouziane, F., Liu, N., Bell, A. H., Gothard, K. M., Luh, W. M., Tootell, R. B., et al. (2012). Amygdala lesions disrupt modulation of functional MRI activity evoked by facial expression in the monkey inferior temporal cortex. Proceedings of the National Academy of Sciences of the United States of America, 109(52), E3640–E3648. Hahn, S., & Gronlund, S. D. (2007). Top-down guidance in visual search for facial expressions. Psychonomic Bulletin & Review, 14(1), 159–165. Hamm, A. O., Cuthbert, B. N., Globisch, J., & Vaitl, D. (1997). Fear and the startle reflex: Blink modulation and autonomic response patterns in animal and mutilation fearful subjects. Psychophysiology, 34(1), 97–107. Hansen, C. H., & Hansen, R. D. (1988). Finding the face in the crowd—An anger superiority effect. Journal of Personality and Social Psychology, 54(6), 917–924. Harrison, L. K., & Turpin, G. (2003). Implicit memory bias and trait anxiety: A psychophysiological analysis. Biological Psychology, 62(2), 97–114.

236  Cognitive biases in health and psychiatric disorders Harrison, S. A., & Tong, F. (2009). Decoding reveals the contents of visual working memory in early visual areas. Nature, 458(7238), 632–635. Heekeren, H. R., Marrett, S., & Ungerleider, L. G. (2008). The neural systems that mediate human perceptual decision making. Nature Reviews Neuroscience, 9(6), 467–479. Heeren, A., Reese, H. E., McNally, R. J., & Philippot, P. (2012). Attention training toward and away from threat in social phobia: Effects on subjective, behavioral, and physiological measures of anxiety. Behaviour Research and Therapy, 50(1), 30–39. Hesselmann, G., Kell, C. A., Eger, E., & Kleinschmidt, A. (2008). Spontaneous local variations in ongoing neural activity bias perceptual decisions. Proceedings of the National Academy of Sciences of the United States of America, 105(31), 10984–10989. Holmes, A., Green, S., & Vuilleumier, P. (2005). The involvement of distinct visual channels in rapid attention towards fearful facial expressions. Cognition & Emotion, 19(6), 899–922. Honeybourne, C., Matchett, G., & Davey, G. C. L. (1993). Expectancy models of preparedness effects: A UCS–Expectancy bias in phylogenetic and ontogenetic fear-relevant stimuli. Behavior Therapy, 24, 253–264. Horstmann, G. (2007). Preattentive face processing: What do visual search experiments with schematic faces tell us? Visual Cognition, 15(7), 799–833. Ikeda, M., Iwanaga, M., & Seiwa, H. (1996). Test anxiety and working memory system. Perceptual and Motor Skills, 82(3), 1223–1231. Itti, L., & Koch, C. (2001). Computational modelling of visual attention. Nature Reviews Neuroscience, 2(3), 194–203. Jamieson, J. P., Nock, M. K., & Mendes, W. B. (2012). Mind over matter: Reappraising arousal improves cardiovascular and cognitive responses to stress. Journal of Experimental PsychologyGeneral, 141(3), 417–422. Janak, P. H., & Tye, K. M. (2015). From circuits to behaviour in the amygdala. Nature, 517(7534), 284–292. Kanwisher, N., & Wojciulik, E. (2000). Visual attention: Insights from brain imaging. Nature Reviews Neuroscience, 1(2), 91–100. Kastner, S., De Weerd, P., Desimone, R., & Ungerleider, L. C. (1998). Mechanisms of directed attention in the human extrastriate cortex as revealed by functional MRI. Science, 282(5386), 108–111. Kastner, S., Pinsk, M. A., De Weerd, P., Desimone, R., & Ungerleider, L. G. (1999). Increased activity in human visual cortex during directed attention in the absence of visual stimulation. Neuron, 22(4), 751–761. Keil, A., Moratti, S., Sabatinelli, D., Bradley, M. M., & Lang, P. J. (2005). Additive effects of emotional content and spatial selective attention on electrocortical facilitation. Cerebral Cortex, 15(8), 1187–1197. Knott, C., Beyer, J., Villarruel, A., Denyes, M., Erickson, V., & Willard, G. (1994). Using the Oucher developmental approach to pain assessment in children. MCN: American Journal of Maternal Child Nursing, 19(6), 314–320. Kok, P., Failing, M. F., & de Lange, F. P. (2014). Prior expectations evoke stimulus templates in the primary visual cortex. Journal of Cognitive Neuroscience, 26(7), 1546–1554. Kok, P., Jehee, J. F. M., & de Lange, F. P. (2012). Less is more: Expectation sharpens representations in the primary visual cortex. Neuron, 75(2), 265–270. LeDoux, J. E. (2000). Emotion circuits in the brain. Annual Review of Neuroscience, 23, 155–184. LeDoux, J. E. (2015). Anxious. London: Oneworld Publications. Lee, S. H., Kravitz, D. J., & Baker, C. I. (2012). Disentangling visual imagery and perception of real-world objects. NeuroImage, 59(4), 4064–4073.

Impact of top-down factors on threat perception biases  Chapter | 10  237 Lee, T. S., & Mumford, D. (2003). Hierarchical Bayesian inference in the visual cortex. Journal of the Optical Society of America A-Optics Image Science and Vision, 20(7), 1434–1448. Lichtenstein-Vidne, L., Henik, A., & Safadi, Z. (2012). Task relevance modulates processing of distracting emotional stimuli. Cognition & Emotion, 26(1), 42–52. Lichtenstein-Vidne, L., Okon-Singer, H., Cohen, N., Todder, D., Aue, T., Nemets, B., et al. (2017). Attentional bias in clinical depression and anxiety: The impact of emotional and non-emotional distracting information. Biological Psychology, 122, 4–12. LoBue, V. (2014). Deconstructing the snake: The relative roles of perception, cognition, and emotion on threat detection. Emotion, 14(4), 701–711. Lorberbaum, J. P., Kose, S., Johnson, M. R., Arana, G. W., Sullivan, L. K., Hamner, M. B., et al. (2004). Neural correlates of speech anticipatory anxiety in generalized social phobia. Neuroreport, 15(18), 2701–2705. Markham, R., & Darke, S. (1991). The effects of anxiety on verbal and spatial task-performance. Australian Journal of Psychology, 43(2), 107–111. Mather, M., Clewett, D., Sakaki, M., & Harley, C. W. (2015). Norepinephrine ignites local hot spots of neuronal excitation: How arousal amplifies selectivity in perception and memory. The Behavioral and Brain Sciences, 1–100. Mather, M., & Sutherland, M. R. (2011). Arousal-biased competition in perception and memory. Perspectives on Psychological Science, 6(2), 114–133. Mathews, A., & MacLeod, C. (1994). Cognitive approaches to emotion and emotional disorders. Annual Review of Psychology, 45, 25–50. Mendez-Bertolo, C., Moratti, S., Toledano, R., Lopez-Sosa, F., Martinez-Alvarez, R., Mah, Y. H., et  al. (2016). A fast pathway for fear in human amygdala. Nature Neuroscience, 19(8), 1041–1049. Mesulam, M. (2008). Representation, inference, and transcendent encoding in neurocognitive networks of the human brain. Annals of Neurology, 64(4), 367–378. Mesulam, M. M. (1981). A cortical network for directed attention and unilateral neglect. Annals of Neurology, 10(4), 309–325. Mesulam, M. M. (1999). Spatial attention and neglect: Parietal, frontal and cingulate contributions to the mental representation and attentional targeting of salient extrapersonal events. Philosophical Transactions of the Royal Society of London Series B-Biological Sciences, 354(1392), 2083. Mogg, K., & Bradley, B. P. (1999a). Orienting of attention to threatening facial expressions presented under conditions of restricted awareness. Cognition & Emotion, 13(6), 713–740. Mogg, K., & Bradley, B. P. (1999b). Some methodological issues in assessing attentional biases for threatening faces in anxiety: A replication study using a modified version of the probe detection task. Behaviour Research and Therapy, 37(6), 595–604. Mogg, K., McNamara, J., Powys, M., Rawlinson, H., Seiffer, A., & Bradley, B. P. (2000). Selective attention to threat: A test of two cognitive models of anxiety. Cognition & Emotion, 14(3), 375–399. Mohanty, A., Egner, T., Monti, J. M., & Mesulam, M. M. (2009). Search for a threatening target triggers limbic guidance of spatial attention. Journal of Neuroscience, 29(34), 10563–10572. Morris, J. S., Friston, K. J., Buchel, C., Frith, C. D., Young, A. W., Calder, A. J., et  al. (1998). A neuromodulatory role for the human amygdala in processing emotional facial expressions. Brain, 121, 47–57. Pt 1. Mukherjee, K. (2010). A dual system model of preferences under risk. Psychological Review, 117(1), 243–255.

238  Cognitive biases in health and psychiatric disorders Mumford, D. (1992). On the computational architecture of the neocortex. 2. The role of corticocortical loops. Biological Cybernetics, 66(3), 241–251. New, J., Cosmides, L., & Tooby, J. (2007). Category-specific attention for animals reflects ancestral priorities, not expertise. Proceedings of the National Academy of Sciences of the United States of America, 104(42), 16598–16603. Nitschke, J. B., Sarinopoulos, I., Oathes, D. J., Johnstone, T., Whalen, P. J., Davidson, R. J., et  al. (2009). Anticipatory activation in the amygdala and anterior cingulate in generalized anxiety disorder and prediction of treatment response. The American Journal of Psychiatry, 166(3), 302–310. Nummenmaa, L., & Calvo, M. G. (2015). Dissociation between recognition and detection advantage for facial expressions: A meta-analysis. Emotion, 15(2), 243–256. Ohman, A. (2002). Automaticity and the amygdala: Nonconscious responses to emotional faces. Current Directions in Psychological Science, 11(2), 62–66. Ohman, A. (2005). The role of the amygdala in human fear: Automatic detection of threat. Psychoneuroendocrinology, 30(10), 953–958. Ohman, A., Flykt, A., & Esteves, F. (2001). Emotion drives attention: Detecting the snake in the grass. Journal of Experimental Psychology-General, 130(3), 466–478. Ohman, A., & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module of fear and fear learning. Psychological Review, 108(3), 483–522. Okon-Singer, H., Mehnert, J., Hoyer, J., Hellrung, L., Schaare, H. L., Dukart, J., et  al. (2014). Neural control of vascular reactions: Impact of emotion and attention. Journal of Neuroscience, 34(12), 4251–4259. Padmala, S., Sirbu, M., & Pessoa, L. (2017). Potential reward reduces the adverse impact of negative distractor stimuli. Social Cognitive and Affective Neuroscience, 12(9), 1402–1413. Palmer, S. E. (1975). Effects of contextual scenes on identification of objects. Memory & Cognition, 3(5), 519–526. Paulus, M. P., & Yu, A. J. (2012). Emotion and decision-making: Affect-driven belief systems in anxiety and depression. Trends in Cognitive Sciences, 16(9), 476–483. Peelen, M. V., Fei-Fei, L., & Kastner, S. (2009). Neural mechanisms of rapid natural scene categorization in human visual cortex. Nature, 460(7251), 94–U105. Pessoa, L. (2009). How do emotion and motivation direct executive control? Trends in Cognitive Sciences, 13(4), 160–166. Pessoa, L., & Adolphs, R. (2010). Emotion processing and the amygdala: From a ‘low road’ to ‘many roads’ of evaluating biological significance. Nature Reviews Neuroscience, 11(11), 773–783. Pessoa, L., Kastner, S., & Ungerleider, L. G. (2002). Attentional control of the processing of neural and emotional stimuli. Brain Research. Cognitive Brain Research, 15(1), 31–45. Posner, M. I., Snyder, C. R., & Davidson, B. J. (1980). Attention and the detection of signals. Journal of Experimental Psychology, 109(2), 160–174. Puri, A. M., Wojciulik, E., & Ranganath, C. (2009). Category expectation modulates baseline and stimulus-evoked activity in human inferotemporal cortex. Brain Research, 1301, 89–99. Rao, R. P., & Ballard, D. H. (1999). Predictive coding in the visual cortex: A functional interpretation of some extra-classical receptive-field effects. Nature Neuroscience, 2(1), 79–87. Ress, D., Backus, B. T., & Heeger, D. J. (2000). Activity in primary visual cortex predicts performance in a visual detection task. Nature Neuroscience, 3(9), 940–945. Reynolds, J. H., Chelazzi, L., & Desimone, R. (1999). Competitive mechanisms subserve attention in macaque areas V2 and V4. Journal of Neuroscience, 19(5), 1736–1753. Richards, J. C., Austin, D. W., & Alvarenga, M. E. (2001). Interpretation of ambiguous interoceptive stimuli in panic disorder and nonclinical panic. Cognitive Therapy and Research, 25(3), 235–246.

Impact of top-down factors on threat perception biases  Chapter | 10  239 Robinson, O. J., Letkiewicz, A. M., Overstreet, C., Ernst, M., & Grillon, C. (2011). The effect of induced anxiety on cognition: Threat of shock enhances aversive processing in healthy individuals. Cognitive, Affective, & Behavioral Neuroscience, 11(2), 217–227. Robinson, O. J., Vytal, K., Cornwell, B. R., & Grillon, C. (2013). The impact of anxiety upon cognition: Perspectives from human threat of shock studies. Frontiers in Human Neuroscience, 7, 1–21. Article 203. Sabatinelli, D., Bradley, M. M., & Lang, P. J. (2001). Affective startle modulation in anticipation and perception. Psychophysiology, 38(4), 719–722. Sakai, K., & Miyashita, Y. (1991). Neural organization for the long-term memory of paired associates. Nature, 354(6349), 152–155. Sander, D., Grafman, J., & Zalla, T. (2003). The human amygdala: An evolved system for relevance detection. Reviews in the Neurosciences, 14(4), 303–316. Sarlo, M., Palomba, D., Angrilli, A., & Stegagno, L. (2002). Blood phobia and spider phobia: Two specific phobias with different autonomic cardiac modulations. Biological Psychology, 60(2– 3), 91–108. Seligman, M. E. (1971). Phobias and preparedness. Behavior Therapy, 2(3), 307–320. Serences, J. T., Ester, E. F., Vogel, E. K., & Awh, E. (2009). Stimulus-specific delay activity in human primary visual cortex. Psychological Science, 20(2), 207–214. Shackman, A. J., Maxwell, J. S., McMenamin, B. W., Greischar, L. L., & Davidson, R. J. (2011). Stress potentiates early and attenuates late stages of visual processing. Journal of Neuroscience, 31(3), 1156–1161. Shin, L. M., Wright, C. I., Cannistraro, P. A., Wedig, M. M., McMullin, K., Martis, B., et al. (2005). A functional magnetic resonance imaging study of amygdala and medial prefrontal cortex responses to overtly presented fearful faces in posttraumatic stress disorder. Archives of General Psychiatry, 62(3), 273–281. Shulman, G. L., Astafiev, S. V., McAvoy, M. P., d’Avossa, G., & Corbetta, M. (2007). Right TPJ deactivation during visual search: Functional significance and support for a filter hypothesis. Cerebral Cortex, 17(11), 2625–2633. Shulman, G. L., McAvoy, M. P., Cowan, M. C., Astafiev, S. V., Tansy, A. P., d’Avossa, G., et al. (2003). Quantitative analysis of attention and detection signals during visual search. Journal of Neurophysiology, 90(5), 3384–3397. Smith, N. K., Cacioppo, J. T., Larsen, J. T., & Chartrand, T. L. (2003). May I have your attention, please: Electrocortical responses to positive and negative stimuli. Neuropsychologia, 41(2), 171–183. Sreenivasan, K. K., Sambhara, D., & Jha, A. P. (2011). Working memory templates are maintained as feature-specific perceptual codes. Journal of Neurophysiology, 106(1), 115–121. Stefanics, G., Csukly, G., Komlosi, S., Czobor, P., & Czigler, I. (2012). Processing of unattended facial emotions: A visual mismatch negativity study. NeuroImage, 59(3), 3042–3049. Stein, T., & Peelen, M. V. (2015). Content-specific expectations enhance stimulus detectability by increasing perceptual sensitivity. Journal of Experimental Psychology-General, 144(6), 1089– 1104. Stein, T., Seymour, K., Hebart, M. N., & Sterzer, P. (2014). Rapid fear detection relies on high spatial frequencies. Psychological Science, 25(2), 566–574. Stokes, M., Thompson, R., Nobre, A. C., & Duncan, J. (2009). Shape-specific preparatory activity mediates attention to targets in human visual cortex. Proceedings of the National Academy of Sciences of the United States of America, 106(46), 19569–19574.

240  Cognitive biases in health and psychiatric disorders Stormark, K. M., & Hugdahl, K. (1996). Peripheral cling of covert spatial attention before and after emotional conditioning of the cue. International Journal of Neuroscience, 86(3–4), 225–240. Stormark, K. M., & Hugdahl, K. (1997). Conditioned emotional cueing of spatial attentional shifts in a go/no-go RT task. International Journal of Psychophysiology, 27(3), 241–248. Stout, D. M., Shackman, A. J., & Larson, C. L. (2013). Failure to filter: Anxious individuals show inefficient gating of threat from working memory of threat from working memory. Frontiers in Human Neuroscience, 7, 1–10. Article 58. Summerfield, C., & de Lange, F. P. (2014). Expectation in perceptual decision making: Neural and computational mechanisms. Nature Reviews Neuroscience, 15(11), 745–756. Summerfield, C., & Egner, T. (2016). Feature-based attention and feature-based expectation. Trends in Cognitive Sciences, 20(6), 401–404. Summerfield, C., Egner, T., Greene, M., Koechlin, E., Mangels, J., & Hirsch, J. (2006). Predictive codes for forthcoming perception in the frontal cortex. Science, 314(5803), 1311–1314. Sussman, T. J., Szekely, A., Hajcak, G., & Mohanty, A. (2016). It’s all in the anticipation: How perception of threat is enhanced in anxiety. Emotion, 16(3), 320–327. Sussman, T. J., Weinberg, A., Szekely, A., Hajcak, G., & Mohanty, A. (2017). Here comes trouble: Prestimulus brain activity predicts enhanced perception of threat. Cerebral Cortex, 27(4), 2695–2707. Szekely, A., Rajaram, S., & Mohanty, A. (2017). Context learning for threat detection. Cognition & Emotion, 31(8), 1525–1542. Szekely, A., Rajaram, S., & Mohanty, A. (2019). Memory for dangers past: Threat contexts produce more consistent learning than do non-threatening contexts. Cognition & Emotion, 33(5), 1031–1040. Thayer, J. F., Friedman, B. H., Borkovec, T. D., Johnsen, B. H., & Molina, S. (2000). Phasic heart period reactions to cued threat and nonthreat stimuli in generalized anxiety disorder. Psychophysiology, 37(3), 361–368. Theeuwes, J. (1991). Exogenous and endogenous control of attention: The effect of visual onsets and offsets. Perception & Psychophysics, 49(1), 83–90. Thompson, K. G., & Bichot, N. P. (2005). A visual salience map in the primate frontal eye field. Progress in Brain Research, 147, 251–262. Van Dessel, P., & Vogt, J. (2012). When does hearing laughter draw attention to happy faces? Task relevance determines the influence of a crossmodal affective context on emotional attention. Frontiers in Human Neuroscience, 6, 294. Vogt, J., De Houwer, J., Crombez, G., & Van Damme, S. (2013). Competing for attentional priority: Temporary goals versus threats. Emotion, 13(3), 587–598. Vuilleumier, P. (2005). How brains beware: Neural mechanisms of emotional attention. Trends in Cognitive Sciences, 9(12), 585–594. Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2001). Effects of attention and emotion on face processing in the human brain: An event-related fMRI study. Neuron, 30(3), 829–841. Vuilleumier, P., Armony, J. L., Driver, J., & Dolan, R. J. (2003). Distinct spatial frequency sensitivities for processing faces and emotional expressions. Nature Neuroscience, 6(6), 624–631. Vuilleumier, P., & Driver, J. (2007). Modulation of visual processing by attention and emotion: Windows on causal interactions between human brain regions. Philosophical Transactions of the Royal Society of London. Series B, Biological Sciences, 362(1481), 837–855. Vuilleumier, P., & Pourtois, G. (2007). Distributed and interactive brain mechanisms during emotion face perception: Evidence from functional neuroimaging. Neuropsychologia, 45(1), 174–194.

Impact of top-down factors on threat perception biases  Chapter | 10  241 Vuilleumier, P., Richardson, M. P., Armony, J. L., Driver, J., & Dolan, R. J. (2004). Distant influences of amygdala lesion on visual cortical activation during emotional face processing. Nature Neuroscience, 7(11), 1271–1278. Wald, I., Fruchter, E., Ginat, K., Stolin, E., Dagan, D., Bliese, P. D., et al. (2016). Selective prevention of combat-related post-traumatic stress disorder using attention bias modification training: A randomized controlled trial. Psychological Medicine, 46(12), 2627–2636. Weems, C. F., Zakem, A. H., Costa, N. M., Cannon, M. F., & Watts, S. E. (2005). Physiological response and childhood anxiety: Association with symptoms of anxiety disorders and cognitive bias. Journal of Clinical Child and Adolescent Psychology, 34(4), 712–723. Wendt, J., Lotze, M., Weike, A. I., Hosten, N., & Hamm, A. O. (2008). Brain activation and defensive response mobilization during sustained exposure to phobia-related and other affective pictures in spider phobia. Psychophysiology, 45(2), 205–215. Whalen, P. J., Bush, G., McNally, R. J., Wilhelm, S., McInerney, S. C., Jenike, M. A., et al. (1998). The emotional counting Stroop paradigm: A functional magnetic resonance imaging probe of the anterior cingulate affective division. Biological Psychiatry, 44(12), 1219–1228. Williams, J. M. G., Mathews, A., & MacLeod, C. (1996). The emotional stroop task and psychopathology. Psychological Bulletin, 120(1), 3–24. Williams, L. E., Oler, J. A., Fox, A. S., McFarlin, D. R., Rogers, G. M., Jesson, M. A., et al. (2015). Fear of the unknown: Uncertain anticipation reveals amygdala alterations in childhood anxiety disorders. Neuropsychopharmacology, 40(6), 1428–1435. Williams, M. A., Moss, S. A., Bradshaw, J. L., & Mattingley, J. B. (2005). Look at me, I’m smiling: Visual search for threatening and nonthreatening facial expressions. Visual Cognition, 12(1), 29–50. Wolfe, J. M., Butcher, S. J., Lee, C., & Hyle, M. (2003). Changing your mind: On the contributions of top-down and bottom-up guidance in visual search for feature singletons. Journal of Experimental Psychology-Human Perception and Performance, 29(2), 483–502. Woodward, S. A., McManis, M. H., Kagan, J., Deldin, P., Snidman, N., Lewis, M., et al. (2001). Infant temperament and the brainstem auditory evoked response in later childhood. Developmental Psychology, 37(4), 533–538. Wormwood, J. B., Lynn, S. K., Feldman Barrett, L., & Quigley, K. S. (2016). Threat perception after the Boston Marathon bombings: The effects of personal relevance and conceptual framing. Cognition & Emotion, 30(3), 539–549. Yantis, S., & Jonides, J. (1984). Abrupt visual onsets and selective attention: Evidence from visual search. Journal of Experimental Psychology. Human Perception and Performance, 10(5), 601–621. Yiend, J., & Mathews, A. (2001). Anxiety and attention to threatening pictures. The Quarterly Journal of Experimental Psychology. A, 54(3), 665–681. Yoon, K. L., & Zinbarg, R. E. (2008). Interpreting neutral faces as threatening is a default mode for socially anxious individuals. Journal of Abnormal Psychology, 117(3), 680–685. Zeelenberg, R., Wagenmakers, E. J., & Rotteveel, M. (2006). The impact of emotion on perception: Bias or enhanced processing? Psychological Science, 17(4), 287–291. Zelano, C., Mohanty, A., & Gottfried, J. A. (2011). Olfactory predictive codes and stimulus templates in piriform cortex. Neuron, 72(1), 178–187.

Chapter 11

Cognitive biases across development: A detailed examination of research in fear learning Rivkah Ginat-Frolich, Tomer Shechner Department of Psychology and the Integrated Brain and Behavior Research Center, University of Haifa, Haifa, Israel

Introduction The current anthology focuses on the myriad forms of cognitive biases that effect both normative and pathological behaviors. The book is comprehensive in its presentation of different cognitive biases related to various stages of information processing. While literature on the nature and etiology of these mechanisms has proliferated greatly in the past 20 years, the bulk of existing studies in the field have been conducted with adults. Examining cognitive biases through a developmental lens presents a unique opportunity to answer two specific questions. The first relates to which specific biases can be observed at different stages of development (e.g., infants, toddlers, children, and adolescents). The second question relates to how these biases affect a child’s present and future psychological functioning. Together, examining these two questions can inform our understanding of the mechanisms that contribute to typical and atypical development in youth. This in turn could lead to the establishment of additional research-informed treatment interventions in at-risk and clinical pediatric populations (Fig. 1). The limited research on cognitive biases in youth, as compared with adults, is likely due to methodological difficulties inherent to developmental ­studies. These include challenges related to measurement, such as greater age-­ related variability in reaction time, and children’s ability to provide reliable judgements of complex situations (Davidson, Amso, Anderson, & Diamond, 2006; Richardson, Anderson, Reid, & Fox, 2011; Shechner et  al., 2012;

Cognitive Biases in Health and Psychiatric Disorders: Neurophysiological Foundations https://doi.org/10.1016/B978-0-12-816660-4.00011-8 © 2020 Elsevier Inc. All rights reserved.

243

244  Cognitive biases in health and psychiatric disorders

FIG. 1  The extended neural network underlying fear learning.

Shechner, Hong, Britton, Pine, & Fox, 2014). Further, in developmental research, tasks need to be tailored to specific age groups. For example, designing a task assessing cognitive appraisal of a potential threat will be done differently when targeting this ability in toddlers as compared with schoolaged children. Although this is true for all experimental research conducted with youth, it is particularly true when measuring cognitive biases, as cognitive skills evolve tremendously at distinct developmental stages. In addition, when dealing with emotion-related cognitive biases, attention must be paid to the behavioral, psychophysiological, and neural findings indicating that patterns of emotional processing change drastically throughout childhood and adolescence (Vink, Derks, Hoogendam, Hillegers, & Kahn, 2014). Undeterred by these difficulties, however, several specific cognitive biases have been examined in youth populations. These include explorations of various attentional biases, such as attention bias relating to threat (e.g., Dudeney, Sharpe, & Hunt, 2015; Shechner et al., 2012; Waters, Wharton, Zimmer-Gembeck, & Craske, 2008) or emotional faces (e.g., Kujawa et al., 2011; McAllister, Kelman, & Millard, 2015; Waters, Mogg, Bradley, & Pine, 2008) and negative interpretation and judgment biases (Cannon & Weems, 2010; Gonzalez et  al., 2017; Rozenman, Vreeland, & Piacentini, 2017). Of note, the bulk of work on cognitive biases in youth have been done through the lens of a specific disorder, such as depression (e.g., Platt, Waters, Schulte-Koerne, Engelmann, & Salemink, 2017), anxiety (e.g., Rheingold, Herbert, & Franklin, 2003; Shechner et  al., 2012), and chronic pain (e.g., Lau et al., 2018). In addition, interventions such as attention bias modification training and cognitive bias modification training have been used in clinical youth populations in an attempt to impact the effects of maladaptive cognitive biases (Cristea, Mogoașe, David, & Cuijpers, 2015; Krebs et al., 2018; Shechner, Hong, et al., 2014). Since a thorough overview of cognitive biases in youth is beyond the scope of the current chapter, we decided to provide a narrow and deep focus on neurocognitive aspects of fear learning. Fear is one of the first discrete emotions

Cognitive biases across development  Chapter | 11  245

to develop, with evidence of fear expression evident as early as infancy in both human and animal studies (Izard, 2011), making it an ideal candidate for developmental research. Indeed, fear learning is an early and rudimentary developmental process involving both cognitive mechanisms (e.g., threat appraisal) and associative learning vital to understanding safe and danger cues in one’s environment. Further, perturbations in fear learning have been implicated in the etiology and maintenance of anxiety disorders, the most prevalent form of pediatric psychopathology (Beesdo, Knappe, & Pine, 2009). An enhanced understanding of these perturbations will help isolate when and how cognitive biases can be measured in youth and their subsequent impact on normative and pathological developmental trajectories.

Cognitive biases related to fear learning across development Since the 19th century, classical fear acquisition and extinction processes have been examined extensively, primarily using animal models (Fanselow & Poulos, 2005; Pavlov, 2010). A century ago, fear conditioning was for the first time examined in humans in a study best known by its sole participant: a 9-month-old child named Little Albert. In this experiment, Albert learned to fear a white rat after it was repeatedly paired with a loud noise (Watson & Rayner, 1920). This experiment set the stage for the hundreds of studies that have been conducted using fear conditioning tasks in humans, almost exclusively with adult participants (Sehlmeyer et al., 2009). The rich experimental literature amassed over the last century, together with the development of technological advances, enables a more precise examination of neurocognitive mechanisms underpinning fear learning. More recently, fear learning studies in youth have received more empirical attention. In this section, we will provide an overview of cognitive mechanisms relating to different phases and aspects of fear learning. The “Fear acquisition and extinction” section will provide an overview of fear acquisition and fear extinction, the “Retrieval of fear extinction memories” section will detail processes related to retrieval of extinction memory (extinction recall), and the “Biases related to overgeneralization of fear” section will explore the cognitive and neural mechanisms underlying adaptive and maladaptive fear generalization. In each of these sections, findings from studies conducted with typically developing children will first be presented. This will be followed by an examination of anxious vs nonanxious child studies, with a specific focus on perturbations in cognitive and learning processes. Last, neural structures and functions related to each fear learning process in typically developing and, when available, anxious youth will be reviewed. The goal of each section will be to present how cognitive biases are reflected in each of these phases of learning and their subsequent influence on typical versus pathological developmental trajectories. Given the scarcity of youth studies, some of the findings presented will

246  Cognitive biases in health and psychiatric disorders

be reviewed from the adult literature with an attempt to focus on their developmental implications.

Fear acquisition and extinction Fear conditioning tasks offer a means to examine associative learning processes and are frequently used in laboratory settings to assess the behavioral, psychophysiological, and neural underpinnings of fear learning. In classical fear conditioning tasks, during acquisition, a neutral stimulus (conditioned stimulus [CS]; e.g., tone) is repeatedly paired with an aversive stimulus (unconditioned stimulus [US]; e.g., shock). Through this pairing, the formerly neutral stimulus comes to elicit a fear response (conditioned response [CR]; e.g., freezing) and in this way becomes a danger cue (CS+). In differential fear conditioning tasks, a second neutral stimulus is presented in the absence of the US and in this way serves as a safety cue (CS−). Thereafter, during extinction, both CSs are repeatedly presented in the absence of the US. This causes the conditioned response to the CS+ to gradually decline and rather than eradicating the memory of the CS-US pairing, a new learning takes place whereby the CS+ is no longer associated with the US (Quirk & Mueller, 2008). When successful extinction takes place, the CS+ no longer elicits a fear response. In human studies, fear acquisition and extinction are typically measured through self-reported fear, autonomic activity (e.g., skin conductance response [SCR], heart rate [HR], and fear-potentiated startle via electromyography [FPS-EMG]), and neural activity (Lonsdorf et al., 2017). Responses to each discrete CS and differential responding (i.e., the difference between an individual’s response to the CS+ and CS−) can then be examined. As compared with the adult literature, relatively few studies have used fear conditioning tasks to assess fear learning in child populations. This is likely a result of ethical and methodological challenges that are inherent to fear learning studies in youth, such as the selection of a US potent enough to elicit a fear response while still being developmentally appropriate (Shechner, Hong, et al., 2014). However, studies assessing developmental populations provide several important findings. First, during acquisition, youth typically exhibit differential fear learning with greater fear attributed to the CS+ as compared with the CS− as indexed by self-report and psychophysiological measures (Shechner, Hong, et al., 2014). Indeed, infants as young as 3 months old have been shown to be capable of acquiring fear in controlled fear learning studies (Gao, Raine, Venables, Dawson, & Mednick, 2010; Ingram & Fitzgerald, 1974). However, significant safety learning differences have been observed among children, with younger children showing higher responses to the CS− than older children (Glenn et  al., 2012). In line with this finding, older children show increased CR (e.g., CS+ > CS−) compared with younger children, which indicates that older children have a better cognitive ability to discriminate between threat and safety cues (Gao et al., 2010; Glenn et al., 2012; Jovanovic et al., 2014).

Cognitive biases across development  Chapter | 11  247

Second, within-session extinction effects have consistently been observed in developmental populations (Britton et al., 2013; Shechner et al., 2015; Waters, Theresiana, Neumann, & Craske, 2017). Interestingly, in one developmental study with adults, adolescents, and children, children and adults showed comparable levels of extinction to the CS+ as indexed by SCR. However, levels of attenuated extinction learning were observed in adolescents. This was hypothesized to be due to the heightened synaptic plasticity, particularly in the prefrontal cortex (Casey, Tottenham, Liston, & Durston, 2005), which is evident during this period along with additional development-related changes (e.g., sexual maturation and transition to independence from parents) (Pattwell et al., 2012). Impairments in fear acquisition and extinction processes have been implicated in the etiology and maintenance of pathological anxiety. To this end, dozens of studies have examined differences between anxious and nonanxious adults during these two stages of fear learning, with their combined effects examined in two metaanalyses. During fear acquisition, Lissek et  al. (2005) found that anxious adults exhibited greater fear to danger cues (CS+) as compared with nonanxious adults. In addition, Duits et al. (2015) found that anxious adults exhibited greater fear to safety cues (CS−) than did nonanxious adults. Further, during extinction, both metaanalyses found that, as compared with nonanxious adults, anxious adults displayed a stronger fear response to the CS+. Interestingly, neither metaanalysis found significant group differences in differential learning (Duits et al., 2015; Lissek et al., 2005). Fewer studies have examined differences between anxious and nonanxious children, yielding mixed findings. During fear acquisition, two studies found that anxious youth exhibited higher overall fear to the CS+ than did nonanxious youth (Britton et al., 2013; Lau et al., 2008). In another study, nonanxious youth exhibited differential learning following fear acquisition, while anxious youth did not exhibit differential learning during this phase (Liberman, Lipp, Spence, & March, 2006). Yet another study found that the CS+ was rated as more unpleasant than the CS− among anxious but not among nonanxious youth (Craske et  al., 2008). During extinction, similar equivocal findings emerge. Some studies reported that both anxious and nonanxious children exhibited extinction (Britton et al., 2013; Shechner et al., 2015), while other studies demonstrated impaired extinction in anxious youth (as indicated by high reported fear levels to the CS+ than to the CS−) as compared with nonanxious youth (Craske et al., 2008; Liberman et al., 2006). These differences could be due to differences in methodology between studies (i.e., US variability, age range, or number of CS presentations in each phase). However, based on findings from studies with both adults and developmental populations, it appears that individuals suffering from anxiety disorders exhibit biases in threat appraisal. Threat appraisal refers to an individual’s assessment of a stimulus’ perceived ability to inflict harm (see Britton, Lissek, Grillon, Norcross, & Pine, 2011). Though this is more clear in the adult literature, the difference in responding to the CSs during both fear conditioning and extinction in anxious samples indicates both

248  Cognitive biases in health and psychiatric disorders

increased threat appraisal of dangerous and safe stimuli and possible perturbations in safety learning processes in anxious populations. Further research is necessary in anxious youth so as to better understand how these biases manifest across development. Research on the neural circuitry underlying fear learning in adults has implicated neural activation in an extended “fear network,” with a specific focus on the amygdala, the anterior cingulate cortex (ACC), the anterior insular cortex (AIC), the dorsolateral, ventral and medial prefrontal cortices (dlPFC, vPFC, and mPFC), and the hippocampus. The involvement of the orbitofrontal cortex (OFC), anterior thalamus, ventral putamen and pallidum, and midbrain substantia nigra/ventral tegmentum has also been observed (Fullana et al., 2016; Mechias, Etkin, & Kalisch, 2010). Convergent lines of research have consistently highlighted the central role of the amygdala in the acquisition, storage, and expression of conditioned fear memory (Kim & Jung, 2006; Sergerie, Chochol, & Armony, 2008). In addition, extensive connectivity-related changes in amygdala-cortical regions and structural changes in the amygdala occur between childhood and adolescence (Gabard-Durnam et al., 2014; Wierenga et al., 2014). This is important as, over the course of development and on the basis of individual experiences, positive (e.g., safety) or negative (e.g., danger) associations are established with formerly neutral stimuli. Notably, the amygdala has been implicated as having a particularly significant role, while these associations are being formed (Tottenham, Hare, & Casey, 2009). Few fMRI studies have been published examining fear circuitry in adolescents during fear learning. In one such study, compared with adults, adolescents were found to be more likely to engage early-maturing subcortical structures (i.e., the amygdala and hippocampus) when discriminating between the CS+ and CS−. Further, only adults showed engagement of l­ate-maturing prefrontal cortex regions (i.e., dlPFC) that was positively correlated with CS− fear ratings during the task (Lau et  al., 2011). In this way, the mature dlPFC aids in the disambiguation of similar appearing stimuli, in this case enabling discrimination learning. In adolescents, by contrast, the dlPFC was negatively correlated with CS− fear ratings, implying that under conditions of increased ­ambiguity—and thus, higher cognitive load—the dlPFC is less able to aid in adaptive discrimination. Taken together, these findings suggest that the development of cortical and prefrontal regions may partially explain agerelated differences in CS+/CS− discrimination (Lau et  al., 2011; Shechner et al., 2014; Shechner, Hong, et al., 2014). One neuroimaging study examined the differences between anxious and nonanxious adolescents during fear conditioning (aged 11–17 years) (Haddad, Bilderbeck, James, & Lau, 2015). Age-related increases in neural activation in response to the CS+ (relative to control) in both groups were evident in several regions, particularly in the insula and dlPFC. Further, group differences emerged in response the CS−, with a higher correlation between age and

Cognitive biases across development  Chapter | 11  249

a­ ctivation in the dlPFC, bilateral anterior insula, and striatum in the nonanxious group as compared with the anxious group. This may reflect typical age-related changes in responding to safety cues (e.g., improved regulation of emotional response to safety cues, better threat/safety discrimination, and increased contingency awareness), which may be perturbed in anxious youth (Haddad et al., 2015). Last, in several regions, including the medial PFC, bilateral amygdala, left hippocampus, and left striatum, greater activation to the CS+ (as compared with a control cue) was displayed in nonanxious youth as compared with among anxious youth. Of note, this elevated activation is not in line with findings from other studies at later stages of fear learning (see “Retrieval of fear extinction memories” and “Biases related to overgeneralization of fear” sections), where elevated activation has been observed in anxious as compared with nonanxious adults and adolescents in response to threat cues as compared with safety cues (Britton et al., 2013; Greenberg, Carlson, Cha, Hajcak, & Mujica-Parodi, 2013b). This finding stresses the need for further research on fear-related neural structures among adolescents in general and anxious adolescents more specifically. To date, there are no neuroimaging studies looking specifically at extinction in developmental populations. However, in adult studies, the amygdala has also been highlighted as a key region in extinction learning, with increased amygdala activation observed in relation to the CS+–no US association (Knight, Smith, Cheng, Stein, & Helmstetter, 2004; Milad et al., 2007). In addition, the mPFC and the ACC have been implicated as being particularly important during extinction learning (Phelps, Delgado, Nearing, & Ledoux, 2004; Sotres-Bayon, Cain, & LeDoux, 2006). Taken together, as several of the brain regions involved during both fear acquisition and extinction exhibit significant structural and functional connectivity-related changes during development, neuroimaging research in youth populations is paramount. Further, such work could aid in informing a more mechanistic understanding of threat appraisal-related biases, such as the expression of fear in response to a neutral stimulus, among youth.

Retrieval of fear extinction memories Retrieval of fear extinction memories (often referred to as extinction recall) occurs when, following extinction, the extinguished CS+ is represented at a later time. Following successful extinction, the extinguished CS+ should no longer elicit a CR. Consequently, a strong level of fear expression (i.e., CR) to the CS+ during this phase indicates poor extinction recall (i.e., biased threat appraisal), whereas low levels of fear expression are indicative of successful extinction recall (Quirk & Mueller, 2008). In differential fear conditioning tasks, extinction recall is typically measured by showing multiple presentations of the extinguished CS+ and CS−. In addition, additional morphs are sometimes presented so as to better identify discrimination abilities between threat and safety cues.

250  Cognitive biases in health and psychiatric disorders

To date, very few studies have examined age-related differences in responding during extinction recall in developmental populations. One study where extinction recall was measured 3 weeks following extinction found age-related differences in self-reported contingency awareness (i.e., explicit recall of the CS-US pairing), with younger children (aged 5–6 and 7–8 years) exhibiting more difficulty discriminating between the presented stimuli than did older children (aged 9–10 years). Further, the youngest children in this study (aged 5–6 years) displayed increased fear responding to the CS− as compared with additional presented stimuli (Michalska et  al., 2016). A second study found that following successful extinction, children (aged 6–11 years) showed poor extinction recall based on higher self-reported fear and elevated SCRs to the CS+ as compared with the CS−. In addition, behaviorally, children maintained greater distance (i.e., showed increased avoidance) to the CS+ than to the CS− during this phase (Marusak, Peters, Hehr, Elrahal, & Rabinak, 2018). A third study that compared adolescents with adults found no age-related interactions in self-reported threat appraisal and explicit memory ratings, though post hoc analyses revealed that adults correctly remembered the CS+ to a greater extent than youth (Britton et al., 2013). Together, these results build on the finding that threat and safety cue discrimination ability improves with age, with biases in threat appraisal decreasing, and the explicit memory and maintenance of this discrimination continuing to develop, during childhood and adolescence. Neuroimaging studies have highlighted the role of the vmPFC and hippocampus during extinction recall. Indeed, increased vmPFC activation is typically observed during this phase (Kalisch et al., 2006), with a positive correlation emerging between the degree of extinction retention and level of signal change in both vmPFC and hippocampal activation (Milad et al., 2007). Further, structural imaging studies have found vmPFC thickness to be correlated with how well healthy adults retain their extinction memory (i.e., the degree of successful extinction recall) (Milad et al., 2005). At present, few studies have used neuroimaging methods to examine extinction recall in youth. In one study, no significant difference emerged in vmPFC activation among healthy adults or healthy adolescents (Britton et al., 2013). In contrast, another study observed reduced activity in adolescents in the vmPFC and the dlPFC relative to adults during this phase. However, unlike the negative vmPFC-amygdala functional connectivity observed in adults, negative functional connectivity was observed between the dlPFC and the amygdala in adolescents. Notably, although dlPFC-amygdala connectivity is not typically considered to be involved in extinction recall, it plays a role in emotional regulation strategies (Hartley & Phelps, 2010). Indeed, less negative dlPFC-amygdala connectivity was associated with higher state anxiety among adolescent participants (Ganella, Barendse, Kim, & Whittle, 2017). In addition, Ganella and colleagues observed negative functional connectivity between the dlPFC and the hippocampus in adolescents, which may underlie an attempt to cognitively suppress negative and intrusive memories in this group

Cognitive biases across development  Chapter | 11  251

(Ganella et al., 2017). In a third study conducted with preadolescent children (aged 6–11 years), increased behavioral avoidance of the CS+ during extinction was positively correlated with increased hippocampal-dACC coupling during extinction recall. Positive hippocampal-dACC coupling has been implicated in the expression of conditioned fear. Indeed, following successful extinction, one would expect decreased hippocampal-dACC coupling. Therefore, the earlier finding may suggest that avoidant behavioral patterns might interfere with the retention of extinction memory (Marusak et al., 2018). Further, activation of the hippocampal-dACC-­insula circuitry in the earlier study may suggest relatively immature hippocampal-vmPFC circuitry during preadolescence. Notably, this hypothesis is in line with findings from developmental neuroscience, which have observed decreases in gray matter volume in the vmPFC during childhood and adolescence (Keding & Herringa, 2015) and a peak in hippocampal volume occurring between ages 9–11 (Uematsu et al., 2012). In addition, there are agerelated increases in hippocampal connectivity in specific regions of the default mode network (DMN) across childhood (Blankenship, Redcay, Dougherty, & Riggins, 2017). As the DMN has been implicated as being important in relation to the contextualization of safety memories (Marstaller, Burianová, & Reutens, 2017), immaturity in these areas may contribute to difficulty in retrieving extinction memory in a safe context (Marusak et al., 2018).

Biases related to overgeneralization of fear Fear generalization describes the process whereby characteristics of a dangerous stimulus transfer onto a similar, neutral stimulus. The cognitive biases that will be examined in this section relate to when fear generalization occurs in excess (i.e., fear overgeneralization). This manifests as maladaptive fear responding to neutral stimuli that bear some resemblance to a danger cue but are in fact innocuous. For example, while it is appropriate to exhibit fear in response to a dog that had previously attacked you, overgeneralizing this fear to anything that looks like a dog would be maladaptive. In a laboratory context, fear overgeneralization is typically measured using generalization gradients. These gradients consist of generalization stimuli (GS) ranging in either perceptual (e.g., color and size; for review, see Dymond, Dunsmoor, Vervliet, Roche, & Hermans, 2015) or categorical (e.g., tools and mammals; for review, see Dunsmoor & Murphy, 2015) similarity on a continuum from a threat cue to a safety cue. In this way, the neurocognitive and psychophysiological correlates of fear generalization can systematically be assessed. Though it is difficult to determine the precise boundary whereby fear generalization becomes maladaptive, fear generalization gradients can help isolate perturbations in this mechanism. In contrast to adaptive fear generalization, where fear responses are typically only observed to GSs relatively similar to the CS+, fear overgeneralization can be indexed by heightened fear responses to GSs less similar to the CS+ (i.e., further down on the continuum).

252  Cognitive biases in health and psychiatric disorders

Though there has been a recent surge in research on fear generalization in humans, at present, only three studies have assessed fear generalization following fear acquisition in a laboratory context among typically developing youth. In one study, as compared with adults, children (aged 8–10 years) had higher self-reported arousal (on a scale from “very calm” to “very arousing”) and SCR to GSs. Of note, no age-related difference in valence (on a scale from “very unpleasant” to “very pleasant”) ratings was observed in this sample (Schiele et al., 2016). In a second study, children (aged 9–14 years) who did not receive cognitive training aimed at decreasing fear generalization exhibited heightened FPS-EMG responses to GSs perceptually similar to the CS−. Further, no difference was observed in the measure of FPS-EMG between the CS− and CS+ (Ginat-Frolich, Gendler, Marzan, Tsuk, & Shechner, 2019). Notably, a third study using a single GS (a 50% morph of the CS+ and CS−) found that, while younger children (aged 8–10 years) had higher FPS-EMG responses to the CS− than to the GS morph, older children (aged 11–13 years) displayed increasing FPS-EMG responses as stimuli increased in similarity to the CS+ (i.e., CS−