Audiological Research Over Six Decades [1 ed.] 1635503701, 9781635503708

Table of contents :
Foreword
Introduction
Acknowledgments
Abbreviations
1. The Early Years
A Diagnostic Challenge
Professor Doctor Eberhart Lüscher
The Quantal Psychophysical Method
The SISI Test
Abnormal Auditory Adaptation
The Békésy Audiogram
Speech Audiometry
Synthetic Sentence Identification (SSI)
Brainstem Versus Temporal Lobe
Related Readings
2. Immittance Audiometry
The Tympanogram
The Stapedius Muscle Reflex
Reflex Averaging
Other Reflex Oddities
An Educational Adventure
Related Readings
3. Auditory Processing Disorder
The Birth of APD
The Italian Pioneers
Early Efforts in the U.S.
The Other Side of the Coin
An Account of an Interesting Patient
Medical Report
Neuropsychological Examination
Audiological Examination
Related Readings
4. A Once-in-a-Lifetime Opportunity
An Important Point
History and Medical Findings
First Admission
Second Admission
Basic Testing
Basic Audiometry
Some Psychoacoustic Measures
Auditory Localization in Space
Final Thoughts
Related Reading
5. Binaural Hearing Aids
An Early Study
But Is the Group Representative of Everyone in It?
Binaural Interference
A Very Intact Nonagenerian
Can We Explain AK’s Findings by Invoking a Relevant Cognitive Deficit?
A Possible Explanation
Related Readings
6. Cued Listening
Group Results
Three Illustrative Individual Patients
Last Thoughts
Related Readings
7. Aging and Gender Effects
Can Speech Understanding Problems in Elderly Persons Be Explained by the Audiogram?
Auditory Processing Disorder and Dichotic Listening
Auditory Processing Disorder Versus Cognitive Decline
A Longitudinal Case Study
Some Gender Differences
Another Gender Effect — The Shape of the Audiogram
Overview
Related Readings
8. Auditory Event-Related Potentials to Words
Waveforms
The Importance of Forcing a Decision
The Framework of an Auditory Event-Related Potential Procedure (AERP)
The Late Positive Component (LPC)
The Right Ear Advantage
But What About the Nontarget Words?
Processing Negativity (PN)
Repeating Words Back Versus Making Decisions About Them
More PN Examples
A Case of Multiple Sclerosis
A Clarification
Final Thoughts
Related Reading
9. A Twin Study
Basic Audiometry Behavioral Psychoacoustic Measures of Auditory
Processing
Standardized Cognitive/Linguistic Evaluations
Activation Patterns
Dichotic Listening
Diffusion Tensor Imaging
Related Readings
10. Odds and Ends
A Visit to Montreal
“Normal Audiometric Findings”
Simian Surgery
A Researcher’s Dream
A Voice From the Past
A Herculean Effort
Finis
Related Readings
Index

Citation preview

Audiological Research Over Six Decades James Jerger

AUDIOLOGICAL RESEARCH OVER SIX DECADES

Editor-in-Chief for Audiology Brad A. Stach, PhD

Also Available From James Jerger Binaural Interference:  A Guide for Audiologists James Jerger, Carol A. Silverman James Jerger:  A Life in Audiology James Jerger Audiology in the USA James Jerger

Order at https://www.pluralpublishing.com

AUDIOLOGICAL RESEARCH OVER SIX DECADES James Jerger, PhD

5521 Ruffin Road San Diego, CA 92123 e-mail: [email protected] Website: https://www.pluralpublishing.com Copyright ©2022 by Plural Publishing, Inc. Typeset in 11/14 Palatino by Flanagan’s Publishing Services, Inc. Printed in the United States of America by Integrated Books International All rights, including that of translation, reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means, electronic, mechanical, recording, or otherwise, including photocopying, recording, taping, Web distribution, or information storage and retrieval systems without the prior written consent of the publisher. For permission to use material from this text, contact us by Telephone: (866) 758-7251 Fax: (888) 758-7255 e-mail: [email protected] Every attempt has been made to contact the copyright holders for material originally printed in another source. If any have been inadvertently overlooked, the publisher will gladly make the necessary arrangements at the first opportunity. Disclaimer: Please note that ancillary content (such as documents, audio, and video, etc.) may not be included as published in the original print version of this book. Library of Congress Cataloging-in-Publication Data Names: Jerger, James, author. Title: Audiological research over six decades / James Jerger. Description: San Diego, CA : Plural Publishing, Inc., 2021. | Includes bibliographical references and index. Identifiers: LCCN 2021017668 (print) | LCCN 2021017669 (ebook) | ISBN 9781635503708 (paperback) | ISBN 9781635503692 (ebook) Subjects: MESH: Jerger, James. | Audiology — history | Hearing Disorders — history | Biomedical Research — history | History, 20th Century | History, 21st Century | United States | Case Reports Classification: LCC RF290 (print) | LCC RF290 (ebook) | NLM WV 11 AA1 | DDC 617.8 — dc23 LC record available at https://lccn.loc.gov/2021017668 LC ebook record available at https://lccn.loc.gov/2021017669

Contents

Foreword by Brad A. Stach ix Introduction xv Acknowledgments xvii Abbreviations xix



1

The Early Years A Diagnostic Challenge Professor Doctor Eberhart Lüscher The Quantal Psychophysical Method The SISI Test Abnormal Auditory Adaptation The Békésy Audiogram Speech Audiometry Synthetic Sentence Identification (SSI) Brainstem Versus Temporal Lobe Related Readings

1 2 3 7 11 13 14 22 26 28 36



2

Immittance Audiometry The Tympanogram The Stapedius Muscle Reflex Reflex Averaging Other Reflex Oddities An Educational Adventure Related Readings

39 42 43 45 49 50 51



3

Auditory Processing Disorder The Birth of APD The Italian Pioneers Early Efforts in the U.S.

53 54 55 57

v

vi  AUDIOLOGICAL RESEARCH OVER SIX DECADES

The Other Side of the Coin An Account of an Interesting Patient Medical Report Neuropsychological Examination Audiological Examination Related Readings

63 63 64 64 65 70



4

A Once-in-a-Lifetime Opportunity An Important Point History and Medical Findings First Admission Second Admission Basic Testing Basic Audiometry Some Psychoacoustic Measures Auditory Localization in Space Final Thoughts Related Reading

73 75 76 76 77 79 79 81 82 87 88



5

Binaural Hearing Aids An Early Study But Is the Group Representative of Everyone in It? Binaural Interference A Very Intact Nonagenerian Can We Explain AK’s Findings by Invoking a Relevant Cognitive Deficit? A Possible Explanation Related Readings

89 90 93 94 95 99 99 102

Cued Listening Group Results Three Illustrative Individual Patients Last Thoughts Related Readings

105 108 110 113 113



6

CONTENTS   vii



7

Aging and Gender Effects 115 Can Speech Understanding Problems in Elderly 117 Persons Be Explained by the Audiogram? 119 Auditory Processing Disorder and Dichotic Listening 122 Auditory Processing Disorder Versus Cognitive Decline A Longitudinal Case Study 125 Some Gender Differences 126 Another Gender Effect — The Shape of the 128 Audiogram Overview 133 133 Related Readings



8

Auditory Event-Related Potentials to Words 135 Waveforms 136 The Importance of Forcing a Decision 138 The Framework of an Auditory Event-Related 141 Potential Procedure (AERP) The Late Positive Component (LPC) 141 The Right Ear Advantage 144 But What About the Nontarget Words? 147 Processing Negativity (PN) 149 Repeating Words Back Versus Making Decisions 154 About Them More PN Examples 155 A Case of Multiple Sclerosis 157 A Clarification 159 Final Thoughts 159 Related Reading 160



9

A Twin Study 161 Basic Audiometry 162 Behavioral Psychoacoustic Measures of Auditory Processing 163

viii  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Standardized Cognitive/Linguistic Evaluations Activation Patterns Dichotic Listening Diffusion Tensor Imaging Related Readings

163 164 169 174 174

10 Odds and Ends 177 A Visit to Montreal 178 “Normal Audiometric Findings” 179 Simian Surgery 181 A Researcher’s Dream 181 A Voice From the Past 183 A Herculean Effort 185 Finis 188 Related Readings 188 Index 189

Foreword

This book is an adventure. It tells the story of the evolution of diagnostic audiology through the voice of one of its greatest contributors, Dr. James Jerger. His story begins in the late 1950s during his formative years as a student of Raymond Carhart and other notables, continues through his prolific years at the Baylor College of Medicine, and ends with his final act at the University of Texas at Dallas. Jerger’s lively narrative describes, in his wise and witty way, what he was thinking throughout six productive decades of game-changing audiological research. He provides us with a view, through case studies of his own work, of an unparalleled perspective, from the room where it happened. I first met Dr. Jerger when he was lecturing at a symposium in Nashville in the late 1970s. I was a young master’s degree student at Vanderbilt University. In those days, that was the clinical degree for entry into the audiology profession. I have this vague memory of that lecture, and it went something like this: He seemed to be presenting case after case after case, and I kept wondering, where are the data? In the classroom and from their treatment in the literature, of course, we were taught that a disorder is some sort of homogeneous entity and that test outcomes should be reflective of that. We were taught to wonder, how do groups of people with the disorder perform on various diagnostic measures? If they vary, then there must be something wrong with the test or, perhaps, with the disorder. It was not until I studied with Jerger that I began to understand. In 1981, I joined the group at the Baylor College of Medicine as a PhD student under Dr. Jerger’s tutelage. My first research assignment was in the lab of Dr. Makoto Igarashi and Dr. Glenn Thompson, where I was studying the effects of cortical control and the location of motoneurons of the stapedius reflex in squirrel

ix

x  AUDIOLOGICAL RESEARCH OVER SIX DECADES

monkeys. But at the end of every day, I would wander down to the clinic to join the clinical staff in their daily case staffings with Dr. Jerger. (I say daily, but during sailing season, we might be found sneaking away on Wednesday afternoons to Galveston Bay for an afternoon on Dr. Jerger’s sailboat Ixchel.) It was at these staffings that I began to understand the tremendous value of peer review, the power of the test battery approach to clinical data collection, and the important lessons that we could learn from individual patients. After I graduated, I stayed on with the team at the Baylor College of Medicine and the Methodist Hospital of Houston for a number of years and never missed a case staffing. During my 11 years there, I came to know Jim Jerger as an excellent scientist with the strongest theoretical background, but one who had unmatched clinical credibility. I tell you the story about case staffings because, as I was reading this book, I was struck by how often he uses case studies to help explain to you as a reader the point he is trying to make about the topic. In his Chapter 3 on auditory processing disorder (APD), he provides some historic background and some theoretical constructs, but then he teaches you about the nature of the disorder by showing you a well-studied and well-tested patient. Chapter 4 is all about a patient and Chapter 9 about twins. And he makes his most important teaching points by illustrating them in patient outcomes in his Chapter 5 on binaural hearing and in Chapter 6 on a very clever test of binaural listening. Finally, in Chapter 8, on auditory event-related potentials (AERPs), he says it best: “That is why we have focused so much of our work in AERPs on individual listeners, whether normal or abnormal: because as clinicians, we all want methods and techniques that we can use to investigate people, not groups of people.” In these days of big data and meta-analysis, this is a refreshing reminder of the importance of the individual and what we can learn from the variability of individual clinical outcomes. I left the Baylor College of Medicine in 1992. In those days, our only real mode of communication was the telephone, and

FOREWORD   xi

back then, they had cords. So, communication was not as simple as today, and I remember how much I missed understanding Dr. Jerger’s perspective on things. One thing I could count on, though, was that every month, he would write an editorial as Editor of the Journal of the American Academy of Audiology. I remember looking forward to understanding what interested him and what he saw as important in the work of others. I had the same anticipation as I was reading this book. I was there for a part of it, and I still found it interesting to see it through his eyes. The book begins with a chapter on the early years of diagnostic audiology. It is interesting on any number of levels, but I tried to imagine what it must have been like back in the days before computers, signal averaging, or radiologic imaging. His description of the approaches to the diagnostic questions is fascinating and, in particular, why he pursued loudness measures. It seems quite likely that it has been a while since the last publication of a Békésy audiogram, but you will see one in this chapter. Perhaps I never quite understood, or I had just forgotten, how rapidly a disordered auditory system can undergo adaptation. It made me wonder if some of the difficulty we have now in pinning down pure-tone thresholds on patients with auditory neuropathy might be related to how rapid this is in an asynchronous system. The chapter also includes a discussion of the Synthetic Sentence Identification (SSI) test. Students of speech audiometry should go back to the early work of Speaks and Jerger to learn why they chose the targets and competition for this test. The test might seem quirky, but it had the strikingly real advantage in that it actually worked as an effective clinical tool for identifying retrocochlear disorder and APD. The second chapter is on immittance audiometry and, in particular, the contribution of the acoustic reflex measurement and reflex patterns to auditory disorder diagnosis. It is a testament to Jerger’s clinical observation that his tympanogram typing has proven to be universally applicable 50 years later. I am impressed that, even today, the combination of vector tympanometry and

xii  AUDIOLOGICAL RESEARCH OVER SIX DECADES

acoustic reflex thresholds remains the best way to identify stiffness disorders such as otosclerosis. Heroic efforts to harness other tympanometric measures have consistently led to overly sensitive measures from a clinical usefulness perspective. And today, acoustic reflex patterns have become an integral component of the early diagnosis of third-window disorders. Each of the chapters in this book provides an insider’s view to Dr. Jerger’s thinking about the topic, a smattering of the data gathered, and a case or two to drive home the point. Other chapters include the topics of auditory processing disorder, binaural hearing aids, and the complexity of auditory aging. Two fascinating chapters cover his work on auditory event-related potentials. If you are not up to speed on any of these areas, you will be after you read Jerger’s eloquent summary of each. I want to tell another story about the patient described in Chapter 3 on APD and the cued-listening measurement described in Chapter 6. I had the privilege of working in the lab at the time this patient was being evaluated and this measurement technique was being developed. I’ll give you a brief preview. The patient with APD has an isolated, well-described dichotic deficit. In the cued-listening measurement, continuous discourse is played simultaneously from a speaker on both the right and left sides of the patient. The task is a very simple one that you will learn more about in Chapter 6. This particular patient performed well on the task in quiet. When noise was introduced, however, she had the perception that the sound coming from the left speaker was attenuated. We were astonished at what appeared to be a sound-field reflection of that dichotic disorder. That is how Jim Jerger learned — from patients. And that is how he tried to teach us to learn. As I reached the final chapter of this book, I was bemused by a story about one of my favorite Jerger articles, entitled “Normal Audiometric Findings.” I cannot remember if I ever knew the story behind the article or if I had just forgotten. Recently, I have resurrected a talk about the importance of the test battery

FOREWORD   xiii

approach to diagnostic testing as the result of identifying a patient with a sizable eighth nerve tumor based on acoustic reflexes and screening for speech-recognition rollover. I remember someone in the audience saying that “we never see patients with tumors.” And I remember thinking to myself, “and you never will, because you choose not to look for them.” In today’s world, we call that kind of thinking and testing confirmation bias. The story behind “Normal Audiometric Findings” was in response to that sort of confirmation bias. What Dr. Jerger shows is that you will see all kinds of interesting outcomes if you just look for them. Among my favorite articles written by Dr. Jerger is one not referenced in this book. He wrote it in 1962, and it was entitled “Scientific Writing Can Be Readable” (ASHA, April 1962, pp. 101– 104). If that sounds in any way pretentious, you can only imagine what the academics of the day must have thought of this precocious young whippersnapper. Regardless, in it he says, “You cannot communicate your research findings to other people unless you write about them in a way that allows other people to understand what you are talking about.” In this book, Jim Jerger lives up to his own lofty expectations and provides a treasure trove of insight for anyone who is a student of diagnostic audiology. — Brad A. Stach

Introduction

History is philosophy from examples. Ars Rhetorica — Dionysius of Halicarnasius Over a period of some 64 years, my colleagues, students, and I have completed research projects in a number of areas impacting audiology practice. We studied the diagnostic evaluation of persons with auditory disorders, ranging from the middle ear to the auditory centers in the brain. We reported these studies to our colleagues in 11 books and 335 articles, appearing in both audiology and otolaryngology publications. From these many sources, I have chosen to comment on articles and books representing 10 distinct areas of investigation. I have tried to avoid excessive technical detail. Readers will find no soporific statistical test results, no p values, no details of instrumentation, and few references to anyone else’s work. Readers are expected, however, to have a basic understanding of auditory evoked potentials and auditory event-related potentials. And, a sense of humor is always welcome. My hope is that this historical overview will convey to students new to the profession something of the satisfaction associated with an audiology research project. Another goal is to remind those few who can still remember the 1950s and 1960s how research in those years impacted and shaped the early formative development of the profession. This is, in many ways, the story of the evolution of diagnostic audiological assessment and a record of how thinking has evolved over the years. The book is divided into three major sections: (1) the early years, (2) the Baylor College of Medicine years, and (3) the University of Texas years. The early years included 9 years at Northwestern, shuttling between the School of Speech in Evanston and

xv

xvi  AUDIOLOGICAL RESEARCH OVER SIX DECADES

the Medical School in downtown Chicago. We left Chicago in 1961 to move to Washington, D.C. Here I spent one year, both at Gallaudet College and at the VA outpatient clinic. At Gallaudet, I learned about deafness and the incredible accomplishments of those wonderful young students. At the VA, I was one-third of a triumvirate, including Laszlo Stein and Stan Zerlin. We set the VA record for length of lunch break, which remains unbroken to this day. In 1962, we moved again, this time to Houston, where I served as head of research for the next 6 years at the Houston Speech and Hearing Center. In Houston, with Chuck Speaks and my wife, Susan Jerger, we spent much of our time on the development and evaluation of the Synthetic Sentence Identification (SSI) materials. Then, Chuck went up to the University of Minnesota and crafted a distinguished career in speech science. In 1968, Susan and I moved from the Houston Speech and Hearing Center across the street to the Baylor College of Medicine, within the Texas Medical Center, where we spent the next 29 years. Chapters 2 to 7 describe only a small fraction of our many interests during those 29 productive years, from 1968 to 1997. Finally, Chapters 8 and 9 cover some of the auditory event-related potentials studies we accomplished during the next 17 years at the University of Texas at Dallas. Chapter 10 fleshes out some odds and ends to complete the book. If you are interested in the right ear advantage, be sure to read the section, “A Visit to Montreal.” Even if you’re not, read it anyway. It is very interesting. Now Susan and I have been retired in Lake Oswego, Oregon, since 2014, and are loving it. But we miss many old friends.

Acknowledgments

If these studies have been useful to the profession, major credit must go to the many colleagues and students who made the individual studies possible: first, my colleague, wife, and helpmate for the past 57 years, Susan Wood Jerger. Her influence can be found throughout the publications cited in this book. At Northwestern, I welcomed the friendship and assistance of my mentor, Raymond Carhart, and of faculty members John Gaeth, Helmer Myklebust, Charles Lightfoot, and my dear friend and golfing buddy, Tom Tillman. At the Northwestern University Medical School, otolaryngologists George Shambaugh Jr., Gene Derlacki, and George Allen provided welcome support for our research in the audiology clinic. At the Houston Speech and Hearing Center, Charles Speaks brought his expertise in speech science to our collaboration in the development of the SSI test materials. We were ably assisted by Carolyn Malmquist, Jane Trammel, and James Thelin. At the Baylor College of Medicine in Houston, 29 fruitful years of research were facilitated by the active direction and cooperation of Bobby R. Alford, chair of the Department of Otolaryngology and Communicative Sciences. His dedication to the success of our laboratory over almost three decades never faltered. Memorable graduate students during those years included John Allen, Denise Brown, Robert Fifer, James Hall III, Maureen Hanley, Deborah Hayes, Susan Jerger, Karen Johnson, Craig Jordan, William Keith, Henry Lew, Brad Stach, Lois Sutton, and Ann Thompson. Otolaryngologic collaborators included Newton Coker, Mickey Stewart, and Makato Igarashi. We were ably assisted by a support staff including Larry Mauldin, Rose Chmiel Hardcastle, Sharon Smith, Terry Oliver, Connie Jordan, Emily Murphy, and Louise Loiselle. International visitors included Neil

xvii

xviii  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Lewis from Australia, Joan Grant from New Zealand, Ali.A. Ali from Egypt, and Anestis Psifidis from Greece. I will always be grateful to neuropsychologist Fran Pirozzolo for his invaluable contributions to our studies of auditory aging and to Charles Berlin and his staff at LSU Medical School for their development of the Dichotic Sentence Identification (DSI) test, based on our SSI materials. In our work on binaural interference, we have collaborated closely with Shlomo Silman at Brooklyn College and Carol Silverman at CUNY. Their earlier work on auditory deprivation was related to subsequent concepts underlying theoretical explanations of binaural interference. Finally, at the University of Texas at Dallas, we were indebted to Bert Moore, dean of the School of Behavioral and Brain Sciences, for his generous support of our AERP research efforts. We were ably assisted by graduate students Tara Davis, Rebecca Estes, Ralf Greenwald, Jeffrey Martin, Jyutika Mehta, Deborah Moncrief, Mary Reagor, Gail Tillman, and Ilse Wambacq. I would be remiss if I did not reiterate the critical role that colleague Brad Stach played in the founding of the American Academy of Audiology, in addition to his many other accomplishments during the Baylor years. Finally, the work on AERPs at UT Dallas would not have been as successful without the important contribution of colleague Jeffrey Martin.

Abbreviations

ABLB

Alternate binaural loudness balance test

ABR

Auditory brainstem response

AERP

Auditory event-related potential

AFSC

Aerospace Medical Division

AKA

Also known as

ALS

Advanced life support

AMLR

Auditory middle latency response

ANOVA

Analysis of variance

APD

Auditory processing disorder

BCM

Baylor College of Medicine

Bell Labs

Bell Telephone Laboratories

CELF

Clinical Evaluation of Language Function

CLT

Cued Listening Test

CNS

Central nervous system

CP3

Left centro-parietal electrode

CVC Consonant-vowel-consonant dB Decibel dBA

Decibels on the A scale of a sound level meter

d′

Detectability index

E800

Grason-Stadler self-recording audiometer

DSI

Dichotic Sentence Identification test

EEG Electroencephalographic EMS

Emergency medical service

HTL

Hearing threshold level

xix

xx  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Hz

Hertz (frequency in)

k 1,000 JND

Just noticeable difference

LE or L

Left

LEA

Left ear advantage

LPC

Late positive component

Max Maximum MCR

Message-to-competition ratio

Min Minimum MLR

Middle latency auditory-evoked response

ms Millisecond MNI

Montreal Neurological Institute

N1

Negative peak at about 100 ms on the evoked potential waveform

P2

Positive peak at about 200 ms on the evoked potential waveform

PZ

Midline parietal electrode

PAL

Harvard Psychoacoustic Laboratory

PB

Phonemically balanced

PBK

Phonemically balanced kindergarten list of words

PB50

Phonemically balanced 50-word lists

PI

Performance versus intensity

PN

Processing negativity

PTA

Pure-tone average

RE or R

Right ear

SCAN

Screening test for auditory processing disorder

SISI

Short Increment Sensitivity Index

SL

Sensation level

SPIN

Speech Perception in Noise Test

ABBREVIATIONS   xxi

SPL

Sound pressure level

TTC

Token Test for Children

USAF

United States Air Force

UTDallas

University of Texas at Dallas

USPHS

United States Public Health Service

W-1

List of Spondee words

WISC

Wechsler Intelligence Test for Children

WWII

World War II

This book is dedicated to the memory of Sadanand Singh, speech scientist, whose influence on our profession will never be forgotten.

1 The Early Years

1

2  AUDIOLOGICAL RESEARCH OVER SIX DECADES

A DIAGNOSTIC CHALLENGE Before the advent of brain imaging techniques, unilateral hearing loss in adults, either sudden or gradual, presented a diagnostic challenge. Meniere’s disease and acoustic tumor are two examples of the problem. Meniere’s disease results from abnormally high endolymphatic fluid pressure within the cochlea. It produces the distressing physical symptoms of unilateral hearing loss, dizziness, tinnitus, and nausea. Acoustic tumor (tumor affecting the eighth cranial nerve) derives from a schwannoma initially growing on the vestibular portion of the eighth nerve. As the tumor grows, it eventually invades the auditory portion of the nerve. The two etiologies initially produce roughly similar unilateral pure-tone audiometric patterns — relatively flatter in Meniere’s disease, usually more sloping from low to high frequencies in acoustic tumor — but the audiogram difference, by itself, is not sufficiently predictable to be useful diagnostically. Although Meniere’s disease is not a pleasant experience, it is usually not life threatening, but a growing tumor in the auditory canal can, eventually, affect nearby brainstem systems responsible for basic life functions such as breathing. It is understandable, therefore, that in the 1950s, before the advent of sophisticated brain imaging, it would have been desirable to know with some certainty which of the two possibilities was, in fact, the more likely case. The first suggestion that audiometric testing might provide an answer came from three British investigators, M. R. Dix, C. S. Hallpike, and J. D. Hood. Their 1948 paper in the Proceedings of the Royal Society of Medicine had a profound effect on the fledgling field of audiology. Dix et al. administered a loudness recruitment test, the alternate binaural loudness balance (ABLB) test, to 30 patients with unilateral Meniere’s disease and 20 patients with unilateral acoustic tumor. The ABLB test is a procedure in which the listener equates the loudness of a pure tone in each ear by adjusting the loudness

1.  THE EARLY YEARS   3

of the tone on one ear until it equals the loudness of the same tone on the other ear, as the tone alternates between the two ears. In the case of Meniere’s disease, Dix et al. found that, as loudness increased in the good ear, there was a concomitant increase in loudness in the impaired ear, but the equivalent increase in loudness on the impaired ear appeared to be compressed into a much smaller span of intensities. They attributed this compression of loudness to the phenomenon of “loudness recruitment,” long recognized as common in persons with sensorineural hearing loss, especially in the high-frequency region above 1 kHz. But in the case of patients with acoustic tumor, there was no compression of the loudness range on the impaired ear. As intensity increased, loudness appeared to grow at the same rate on both the better and the poorer ears. In other words, patients with acoustic tumor showed little or no recruitment of loudness. Dix et al. concluded that loudness recruitment, as measured by the ABLB test, was characteristically present in their patients with Meniere’s disease but characteristically absent in their patients with acoustic tumors. The implication seemed clear. You could distinguish between inner ear and auditory nerve sites of disorder by testing for recruitment. Its presence suggested an inner ear site while its absence pointed toward a site involving the auditory nerve.

Professor Doctor Eberhart Lüscher There was one major problem, however. The ABLB test was only feasible if the hearing loss was unilateral. The unaffected ear had to be relatively normal. But a Swiss otologist, Professor Doctor Eberhart Lüscher, immediately grasped the long-range significance of their findings. If you could devise a loudness recruitment test that did not require one normally hearing ear, reasoned Lüscher, you could expand the diagnostic value of the loudness recruitment phenomenon to include persons with bilateral sensorineural hearing loss as well as unilateral loss, a potentially major

4  AUDIOLOGICAL RESEARCH OVER SIX DECADES

advance in the audiological evaluation of virtually all sensorineural hearing disorders. Influenced by the 19th-century psychophysical research of German scientists Ernst Weber and Gustave Fechner, on the measurement of just-noticeable differences (JNDs) in the loudness of tones, Lüscher reasoned that, if loudness grew incrementally by the accumulation of JNDs, then each JND must be smaller in the ear with loudness recruitment since so many of them seemed to be packed into a much smaller range of intensities. The answer, Lüscher reasoned, was to measure a patient’s JNDs for loudness, which could be done on one ear independently of the recruitment status on the opposite ear: The ABLB requirement of unilateral loss would no longer be necessary. This reasoning did not hold up under subsequent research. Loudness in a listener with sensorineural hearing loss is much more complicated than a simple sum of JNDs. But, despite its faulty theoretical basis, the concept worked in practice. Lüscher next devised a method for measuring the loudness JND in the clinic. As a student of psychoacoustics, he turned to the classical method of constants. He instructed his lab assistant, a young Polish engineering student named Jozef Zwislocki, to fabricate a device that would allow the operator to vary the degree of amplitude modulation of a pure tone generated by an audiometer. Zwislocki, a Polish immigrant to Switzerland, subsequently immigrated to the U.S., studied at the Harvard Psychoacoustic Laboratory, and then moved on to Syracuse University, where he eventually became a distinguished auditory neuroscientist. Zwislocki died in 2016 at the age of 96. Lüscher reasoned that the smallest degree of amplitude modulation that the patient could just notice was a measure of the JND for loudness. In Lüscher’s methodology, modulation was varied randomly over a range of 0 to 5 dB in a series of trials separated by silent intervals. The JND was defined as the modulation level corresponding to the 50% correct point on the psychophysical function relating percent correct performance to modulation

1.  THE EARLY YEARS   5

level. You can imagine that after this procedure had been administered across three or four frequencies, both the tester and the listener were exhausted, but results were encouraging. After testing his method in the clinic, Lüscher concluded that his technique worked very well and could be viewed as a substitute for the ABLB as a test for loudness recruitment in patients with bilateral sensorineural losses. His results were published in the Journal of Laryngology and Otology in 1951. But word of his work had already spread among otology circles in Europe. An electronics company in London, Amplivox, was impressed with Lüscher’s paper and thought there might be a market for a device that one could attach to an audiometer to produce the amplitude modulation that formed the basis for the test. My mentor, Raymond Carhart, ordered one of the units for use in the Northwestern audiology clinics. Since the unit was small, light, and portable, it could be carried between the School of Speech in Evanston and the Department of Otolaryngology at the NU Medical School in downtown Chicago. At a meeting of the graduate students, Carhart asked whether any one of us was interested in trying it out. I immediately volunteered and took my first halting step into a career in audiological research. I was just beginning work on my master’s degree at Northwestern when Lüscher’s paper appeared. I took the Amplivox device to the Northwestern University Medical School in Chicago’s near northside, hooked it up to an audiometer, and began to examine patients with hearing loss. After testing 89 individuals with various auditory disorders, I analyzed the data, wrote a paper based on the findings, and submitted it to the audiology faculty as a master’s thesis. I then prepared a shortened version and submitted it for publication in The Laryngoscope. My findings were in good agreement with Lüscher’s. We had both demonstrated that patients with cochlear site of disorder appeared to have smaller JNDs for sound intensity than patients with either eighth nerve or more central sites. In addition, in my study, persons with more central sites appeared to

6  AUDIOLOGICAL RESEARCH OVER SIX DECADES

have abnormally large JNDs. It was difficult to reconcile this latter observation with the idea that a given loudness level was the result of a stack of JNDs for loudness, but as we have noted above, loudness and loudness discrimination are not so simply related. Nevertheless, for whatever reason, performing the “measurement of the JND” became a useful diagnostic procedure. At the time, I was so focused on differentiating between cochlear and eighth-nerve sites that I failed to pursue the interesting observation that persons with more central sites of disorder consistently showed larger than normal JNDs. Decades passed before we began to appreciate the many complex factors influencing auditory judgments. Professor Lüscher was so pleased with my article that he invited me to present at the biennial meeting of the International Congress of Audiology in Bonn, Germany, in 1960. I had the pleasure of meeting him and “talking shop” about JNDs. It happened that Jozef Zwislocki was also attending this meeting. It was now almost 10 years since Zwislocki had left Switzerland and Professor Lüscher’s lab to pursue an advanced degree at the Harvard Psychoacoustic Lab in Boston. He was now an established auditory scientist at Syracuse University and was viewed as one of the most promising younger investigators at work in the U.S. At the Congress, he told me that he had received a note from Professor Lüscher, requesting his presence at his hotel room at lunchtime. Zwislocki supposed that, over a hearty lunch, he would be congratulated by his former mentor for his many achievements in America since leaving the professor’s lab in Basle. Later in the day, I noticed him sitting alone in the hotel lobby. He looked depressed. I asked him how his meeting with his former employer, the Herr Professor, had gone. “Badly,” he muttered. “He only wanted me to carry a large painting his wife had just acquired, from her car to their hotel room.” With the MA degree behind me in 1952, I began work toward the PhD degree. My experience with the Amplivox device had not been altogether satisfactory. Surely there must be a better way

1.  THE EARLY YEARS   7

to approach the problem that would be less time-consuming than Lüscher’s approach. I attacked the problem with a view toward my PhD thesis. I had been reading about the quantal psychophysical method, and I thought it might be the basis for a different approach to JND measurement.

The Quantal Psychophysical Method A group at the Harvard Psychoacoustic Laboratory, S. Stevens, C. Morgan, and J. Volkmann, had developed the concept that loudness increased by the cumulative addition of “neural quanta,” or minimal changes in neural activity. Central to the concept was the requirement that, in any attempt to relate sound intensity with neural quanta, one must introduce the change in the intensity level of the signal immediately, without a silent interval between any two serially presented stimuli. This was a deliberate departure from Fechner’s method-of-constants in which the two stimuli to be compared are usually separated by a silent interval. But that introduces a “time error” that is difficult to deal with. In the new “quantal” psychophysical method, in contrast, relatively short-intensity increments were added to an ongoing steady tone, with literally no time between the two levels to be analyzed. The listener’s task was to respond to each increment heard. A single trial included 10 to 20 increments, all of equal size. Increment size was randomly varied over several trials. When a listener’s percent-correct responses were plotted against increment size, the data defined a psychometric function. It was traditional to regard such a function as a “phi function of gamma,” that is, a sequential accumulation of the area under the normal curve — in effect, a sigmoidal or “S”-shaped function. But Stevens et al. noted that their functions for individual listeners seemed to be fitted best by a simple rectilinear function as defined by the method of least squares. In the Stevens et al. model, the distance between the increment size corresponding

8  AUDIOLOGICAL RESEARCH OVER SIX DECADES

to 0% correct and the increment size corresponding to 100% correct defined a single neural quantum. The concept is illustrated graphically in Figure  1–1 with hypothetical data. Several data points are gathered as increment size is randomly varied over blocks of 20 increments each. These data are then fitted by a least squares rectilinear function. The neural quantum is defined as the difference in increment size between the value of the increment at 0% correct and the value of the increment at 100% correct (in this hypothetical example from 0.5 dB at 0% to 3.0 dB at 100%, a distance of 2.5 dB). Although the concept of the neural quantum is interesting, especially in view of previous concepts of the stacking of JNDs to produce total loudness, my interest was in the methodology per se. It seemed to me to be uniquely suited to the measurement of

Figure 1–1.  Example of a least squares linear fit to hypothetical data relating percent-correct performance to increment size in the quantal psychophysical method. The size of the “neural quantum” is the increment size at 100% correct performance minus the increment size at the 0% correct level, in this case 2.5 dB (3.0−0.5 = 2.5).

1.  THE EARLY YEARS   9

differential sensitivity because of the immediate change from one loudness level to another. For my doctoral dissertation, I followed the exact quantal procedure outlined by Stevens et al. I tested 20 participants: (1) a control group of adults with normal hearing threshold levels (HTLs) at all audiometric test frequencies to 8000 Hz and (2) an experimental group of listeners with high-frequency hearing loss and a history of excessive noise exposure, presumably due to cochlear damage. HTLs in this group were within normal limits out to 1000 Hz, but all had high-frequency sensorineural loss above 1000 Hz. All participants were tested at two frequencies, 1000 and 4000 Hz, at a constant sensation level of 10 dB. Figure 1–2 shows group results at the 1000-Hz test frequency, where threshold hearing level was within normal limits in both groups. As increment size varied from 1 to 5 dB, the averaged

Figure 1–2.  Actual quantal psychophysical functions for both experimental and control groups at 1000 Hz and SL = 10 dB. HTLs were within normal limits in both groups. Modified from Figure 4 in Jerger (1955).

10  AUDIOLOGICAL RESEARCH OVER SIX DECADES

rectilinear functions increased from 0% at an increment size slightly greater than 1 dB to 100% at an increment size of approximately 3.5 dB. Not unexpectedly there was little difference between the averaged functions for the two groups. But at the test frequency of 4000 Hz (Figure 1–3), the averaged rectilinear functions for the two groups diverged markedly. At 0% correct, the two functions converged at an increment size of about 0.5 dB. But as percent-correct performance increased from 0% to 100%, the two functions climbed at substantially different rates. For the normal/control group, the 100% correct level required, on average, an increment size of 3 dB. For the hearing-loss group, however, the 100% correct performance level required an increment size of less than 1.5 dB. As increment size increased, performance climbed more rapidly in the experimental group.

Figure 1–3.  Actual quantal psychophysical functions for both experimental and control groups at 4000 Hz and SL = 10 dB. HTLs were within normal limits in the control group, but there was significant sensorineural loss at 4000 Hz in the experimental group. Modified from Figure 4 in Jerger (1955).

1.  THE EARLY YEARS   11

It had been conventional to define the JND at the performance level of 50%. Here the difference between the two groups was about 0.8 dB. But these data suggested that, for purposes of differentiating the two groups, it might be more useful to contrast them at the 100% level, where the group difference was 1.5 dB. One can regard these results in at least two ways. One can say that the JND for loudness at a 10-dB sensation level is smaller in the ear with cochlear damage, or one can say that this has nothing to do with JNDs. The ear with cochlear damage is simply showing the same performance level at the 10-dB sensation level as the normal ear would show at the same absolute intensity level. Both statements are loosely true. But how you interpret the results depends on two factors: (1) whether you are interested in the question of whether JNDs are either normal or abnormal in people with hearing loss or (2) whether you are more interested in how to differentiate between cochlear versus neural sites of the hearing disorder. If you are aware that listeners with neural damage, as opposed to inner ear disorder, do not show exceedingly sensitive performance at low sensation levels, no matter the degree of sensorineural loss, then, quite aside from theoretical speculation about JNDs, you are looking at the basis for an interesting clinical diagnostic test. Here, I thought, might be a potential procedure to differentiate nicely between Meniere’s disease and acoustic tumors.

The SISI Test As I studied Figure 1–3, it struck me that one might devise a simple clinical test to make this distinction. Simply present to the listener a small increment size (1 dB) 20 times at a low sensation level, count the number of correct responses, and the resultant percentage should be higher in persons with a cochlear disorder like Meniere’s disease than in persons with acoustic tumors. This was the basis for the Short Increment Sensitivity Index (SISI) test. My collaborator at Northwestern, Earl Harford, and I made a few

12  AUDIOLOGICAL RESEARCH OVER SIX DECADES

modifications from the dissertation study. Preliminary exploration had suggested that, in actual clinical use, it would be better to raise the sensation level slightly, to 20 dB rather than 10 dB, and to present 20 successive 200-ms increments, at 5-s intervals. The choice of 1 dB as the single increment size had been based on subjective experience. It “seemed” to be a level that normals, conductives, and retrocochlears found quite difficult but cochlears heard very well. However, had we simply consulted the data of Figure 1–3 as, after reflection, we should have done and projected the actual difference between the two group functions, we would have noted (Figure 1–4) that group differences (vertical dashed black lines) were much greater for a 1.5-dB increment than for a 1-dB increment: Nonetheless, the 1-dB estimate actually turned out to work reasonably well in clinical practice. The SISI score was the percent of those 20 increments to which the listener made a correctly timed response. In subse-

Figure 1–4.  Modification of Figure 1–3, contrasting the difference between the two groups at increment size = 1 dB (30% difference between groups) and at increment size = 1.5 dB (62.5% difference between groups).

1.  THE EARLY YEARS   13

quent clinical testing, we found that persons with normal HTLs seldom scored better than 20%, more likely 0%. Persons with presumably cochlear disorders typically scored between 80% and 100% in the region of the sensorineural loss. Listeners with retrocochlear sites, such as acoustic neuroma, typically scored 0% consistently no matter the degree of loss. Well, that is the story of our brief early efforts to devise techniques for differentiating cochlear from retrocochlear disorders. They worked reasonably well until they were replaced by diagnostic imaging, but along the way, they provided unique observations regarding an even more interesting phenomenon, abnormal auditory adaptation, a discovery with deeper implications for diagnostic evaluation.

ABNORMAL AUDITORY ADAPTATION The auditory world is ordinarily perceived as stable. The loudness of the drip-drip-drip of a leaky faucet stays at the same irritating level over time, and the sound of a foghorn never alters the loudness of its mournful message. Even continuous low-frequency pure tones maintain their loudness levels over time. But in 1955, at the University of Iowa, Dean Lierle and Scott Reger reported a surprising finding. When a patient with an acoustic tumor tracked the threshold for a pure tone over time by means of a self-recording device, the recording pen gradually dropped over time. As seconds turned into minutes, the patient required more and more intensity to keep the continuous tone just audible in the ear with the tumor. This was perhaps the first convincing demonstration of what has come to be called “abnormal auditory adaptation” to an auditory stimulus. The concept of a self-recording audiometer of the sort employed by the Iowan investigators was not new. Georg von Békésy (1899– 1972) built a prototype self-recording threshold device while at the Karolinska Hospital in Stockholm, Sweden, during World War II.

14  AUDIOLOGICAL RESEARCH OVER SIX DECADES

His 1947 paper described it in some detail. An electric motor drove the frequency of a continuous pure tone across the frequency range from the low to the high end of the conventional audiometric range. At the same time, a second motor drove an intensity control from minimum to maximum level. The listener held a control button that reversed the direction of the motor driving the intensity control in the high-to-low direction. The listener was instructed to press the control button when the tone was just audible but to release the button when the tone was no longer audible. A pen mechanism attached to the intensity-level control button wrote out its position on a moving table to which an audiogram form had been attached. In this way, the movement of the pen appeared on the audiogram form as a tracking of the listener’s threshold across the entire frequency range. Alternatively, the frequency could be fixed at a single value, producing a threshold tracking at that fixed frequency as a function of time. In either event, the expected suggestion soon resurfaced that the distance between just-heard and just-not-heard must represent the “first JND.” As we shall see, however, subsequent research findings rendered this reasoning, too, an oversimplification of a very complex situation. Many years ago, I once shared a limousine with Békésy on a trip from the Denver airport to a conference at a resort hotel in Estes Park, Colorado. He was by no means a talkative person, but he did enjoy sharing his enthusiasm for Pre-Columbian art. He confided that he would not have accepted the invitation to speak at this conference had it not been for the excellent Pre-Columbian exhibit at the Denver Art Museum.

The Békésy Audiogram In the late 1950s, the Grason-Stadler Company produced a commercial version of Békésy’s audiometer. It happened that Rufus Grason, cofounder of the company with Steve Stadler, knew Békésy quite well from their days together at the Harvard Psy-

1.  THE EARLY YEARS   15

choacoustic Lab in the early 1950s. The instrument, the GrasonStadler E-800, followed Békésy’s overall concept but added a few additional features. First, the size of each intensity step, 2 dB in Bèkèsy’s original audiometer, was reduced to about 0.25 dB in the model E-800. Second, the tone could be either continuous or periodically interrupted by means of an electronic switch and timer. This latter feature greatly enhanced the versatility of the E-800. After reading the paper by Lierle and Reger and their fixedfrequency trackings, I convinced my mentor, Ray Carhart, that we desperately needed an E-800 for the clinic at the Northwestern Medical School. He agreed, and within a few months, we were in business. We set up a protocol that included: 1. A sweep frequency tracking of a continuous tone from 250 through 8000 Hz 2. A sweep frequency tracking of a periodically interrupted tone from 250 through 8000 Hz 3. Fixed-frequency tracking of both continuous and interrupted tones at frequencies suggested by the audiometric levels of the sweep frequency interrupted tracking data to examine the emerging pattern in greater detail if desired. We called these results, collectively, a “Békésy audiogram.” As a first step, we gathered data on 434 patients with various types of loss. We reported these results in 1960 in a paper entitled “Békésy Audiometry in Analysis of Auditory Disorders” in the Journal of Speech and Hearing Research. As we analyzed the clinical data, it soon became apparent that the diagnostic value of the technique lay in the relation between the tracking results for the continuous versus the periodically interrupted tones. This conclusion held for both the sweep-frequency and the fixedfrequency test modes. The data seemed to fall into four distinct patterns depending on the relation between the threshold tracking for the continuous tone and the threshold tracking for the corresponding interrupted tone. We labeled them, unimaginatively,

16  AUDIOLOGICAL RESEARCH OVER SIX DECADES

as Type I, Type II, Type III, and Type IV. In the subsequent discussion, interrupted tracings are always in black and continuous tracings always in red. In the Type I tracings, the results for the interrupted-tone threshold and for the continuous-tone threshold were essentially superimposed. The two tracings were virtually indistinguishable. This pattern characterized the data of persons with normal hearing and persons with purely conductive losses. In the Type II pattern, the two threshold tracings were superimposed over the low-frequency test range (e.g., below 1000 Hz), but above this frequency, the tracing of the continuous tone fell clearly below the tracing of the interrupted tone by 15 to 20 dB. Over this same range, the width of the tracing of the continuous tone tended to be substantially smaller than the width of the tracing of the interrupted tone. This pattern was consistently observed in patients whose hearing loss was due to cochlear disorder or whose etiology was unknown. In the Type III result, the tracing for the continuous tone overlapped the tracing for the interrupted tone only briefly before descending precipitously to the upper limit of the audiogram form. This extreme form of what can only be termed “abnormal auditory adaptation” characterized the losses of patients with eighth-nerve or brainstem disorders. In the Type IV result, the tracing of the continuous tone ran well below the tracing of the interrupted tone, usually across the entire test frequency range. But the tracing of the continuous tone ran parallel to the tracing of the interrupted tone rather than falling rapidly to the equipment limit, nor was the width of the continuous-tone tracing especially narrow as in the Type II pattern. We observed Type IV along with Type III tracings only in retrocochlear sites of loss. Figure 1–5 illustrates these four types of Békésy audiograms in the sweep-frequency mode. The Type III result was, of course, the most dramatic. It truly epitomized the concept of abnormal adaptation to continuous auditory stimulation. Figure 1–6 shows analogous results in the fixed-frequency mode. Here the advantage is the ability to see the abnormality change over time.

17 B

Figure 1–5. The four Békésy audiogram types as they appear in the sweep-frequency mode (A–D). In the Type I audiogram (A), the tracings for the continuous (red tracings) and interrupted tone (black tracings) conditions are essentially superimposed and not remarkably narrow. This type is largely limited to persons with normal hearing threshold levels and to persons with purely conductive losses. In the Type II audiogram (B), the continuous and interrupted tracings are usually superimposed at low frequencies, but somewhere in the region of 500 to 1000 Hz, the continuous tracings fall below the interrupted, by 15 to 25 dB; run below and parallel to the interrupted tracing to the upper frequency limit of the audiometer, usually 8000 Hz; and become quite narrow in amplitude. This type is typically observed in persons with cochlear loss.  continues

A

18 D

Figure 1–5.  continued  In the Type III Békésy audiogram (C), the continuous-tone tracing overlaps the interrupted tracing only in the low-frequency range but ultimately breaks away from the interrupted tracing and falls rapidly to the upper intensity limit of the audiometer and remains there until the operator terminates the test and turns off the equipment. The point at which the continuous-tone tracing begins to break away from the interrupted tracing can be anywhere along the frequency scale. The distinguishing feature is the rapid descent of the tracing to inaudibility at the maximum sound output of the test instrument. This type is virtually diagnostic of an eighth-nerve site of disorder. In the Type IV audiogram (D), the continuous tracing never descends to the equipment limit but falls immediately below the interrupted tracing as the test frequency sweeps upward from the lower frequency limit of the audiometer. Thereafter, it maintains a constant difference in the range of 10 to 40 dB. This type is also observed in eighth-nerve and low brainstem sites. Modified from Figure 1 in Jerger (1960).

C

19 B

Figure 1–6.  The four Békésy audiogram types as they appear in the fixed-frequency mode (A–D). Here the sweepfrequency mode of the audiometer is disarmed in order to view the tracings at any selected test frequency as a function of time. The four figures reflect how the four Békésy audiogram types appear in this mode of operation. Here the illustrative test frequencies are 250, 500, and 4000 Hz, but since the frequency source for the E-800 audiometer was a continuously variable oscillator, any desired test frequency was available. Modified from Figure 2 in Jerger (1960).  continues

A

20

Figure 1–6.  continued

C

D

1.  THE EARLY YEARS   21

We had occasion to study the progression of the Type III pattern over a 10-month period in a 19-year-old woman with right-sided acoustic neurinoma. Over that time period, the highfrequency loss at 4000 Hz declined from 40 dB to 100 dB. Interrupted and continuous sweep-frequency tracings over that same 10-month period declined substantially. As the high-frequency loss reflected by the tracing of the interrupted tone progressed, the frequency at which the tracing of the continuous tone broke away from the tracing of the interrupted tone progressed as well, from about 2000 Hz down to about 750 Hz. In another case of retrocochlear site, we studied the fixedfrequency threshold data after surgical removal of a meningioma in the right internal auditory canal. The fixed-frequency threshold tracings from the unaffected left ear showed no progression over time. In the case of the affected right ear, however, the tracing of the continuous tone broke away from the tracing of the interrupted tone at about 1500 Hz and descended rapidly to the upper equipment limit. The Type III pattern repeated in the fixedfrequency tracings at 4000 Hz. The tracing of the continuous tone threshold fell 50 dB in slightly over 2 minutes of tracing. In a case like this, the question arises, “Is there a critical ‘offtime’ of the interrupted-tone threshold, below which it behaves like a continuous-tone threshold?” In other words, is there an interrupted off-time between interrupted tones so short that it behaves like a continuous tone in the Type III situation. We addressed this question by systematically shortening the off-time between successive interrupted tones. The fixed-frequency tracing did not change as the off-time was shortened from 300 to 20 ms but changed immediately when the tone was switched to the continuous mode, returning to the precipitous drop in the tracing. The same overall pattern was repeated as the off-time returned to 20 ms and again to 50 ms, followed yet again by rapid fall in the continuous mode and again recouping at the 50-ms offtime. If there were a critical off-time for this patient, it was less than 20 ms.

22  AUDIOLOGICAL RESEARCH OVER SIX DECADES

These Békésy audiometric studies have highlighted one key difference between the consequences of cochlear versus retrocochlear sites of disorder — the presence of mild adaptation to continuous stimulation at cochlear sites, and confined to high frequencies, versus moderate to severe adaptation across the entire frequency range in the case of retrocochlear sites. We learned a good deal about hearing losses by running these Békésy audiograms routinely. It is regrettable that they are not done much anymore.

SPEECH AUDIOMETRY Scientists who study speech perception, especially with word stimuli, often plot correct responses as a function of speech intensity level. This has been traditionally termed the “articulation function.” In this way, performance is charted across the entire speech intensity range. More recently, these plots have been termed “performance versus intensity” or “PI” functions. When Raymond Carhart set out, in 1943, to adopt the PB word lists for clinical evaluation of hearing-impaired persons, he noted that for both hearing-impaired and normally hearing persons, these PI functions seemed to peak at about a speech sensation level of 25 dB. “Well,” he might have said to himself, “we could save a lot of time in the clinic if we just measured the PB score at sensation level (SL) = 25 dB and skipped testing at lower levels, since 25 dB is where everyone’s maximum score seems to be located anyway.” That was his first mistake. Subsequent studies showed convincingly that the maximum PB score is reached more nearly in the region of 40 dB SL. Carhart’s next idea was to abandon PB testing at SLs higher than 25 dB since they never really changed above 25 dB anyway. That was his second mistake. As a result, it was nearly three decades before the diagnostic value of a full PI function for PB words was recognized. It turns out that not everyone’s PI function stays up at levels above the PB

1.  THE EARLY YEARS   23

max. In some patients, it falls, often dramatically, as speech level increases beyond the maximum score. We have termed this the “rollover” phenomenon. Figure 1–7, for example, plots the PB scores, as a function of speech sensation level, for both ears of a woman after surgical removal of a meningioma from the right auditory canal. As expected, performance increased as speech level increased on the left ear. The score rose from 44% at the 10-dB sensation level to a maximum of 98% at the 40-dB sensation level. As intensity continued to increase, the percent-correct score remained relatively

Figure 1–7.  PI-PB functions of a patient with a meningioma in the right internal auditory canal. On the left ear, the PI function reached a maximum score of 98% at the 40-dB sensation level, then continued at that high level to the 80-dB sensation level. The right ear PB scores were substantially poorer than the left ear PB scores at every speech sensation level. After the right ear PI function reached its peak at the 40-dB sensation level, it declined systematically as speech level continued to increase. At the 80-dB level, it had fallen to only 20%. This is a good example of the “rollover” phenomenon characteristic of eighthnerve disorders. Modified from Figure 2 in Jerger and Bucy (1960).

24  AUDIOLOGICAL RESEARCH OVER SIX DECADES

stable up to the maximum speech sensation level of 80 dB. There is nothing unusual here. That is the way PI functions are supposed to look in a normal ear. In the right ear, however, the PB score rose from 18% at SL = 10 dB to a maximum of 94% at SL = 40 dB but then rapidly declined to only 20% at SL = 80 dB. This is a dramatic example of “rollover” of the PI function for PB words. It was published in 1960 in a paper coauthored by me and by my collaborators, Dr. Susan Jerger, at the Baylor College of Medicine, and Dr. Paul Bucy, then at the Department of Neurosurgery of the Northwestern University Medical School. We now recognize rollover of the PI function for words as another symptom of an auditory nerve, or more central, disorder. This was certainly not the first report of the rollover phenomenon. It had been noted and described repeatedly during the 1950s and 1960s by a number of European investigators, less so by U.S. investigators because of Carhart’s influence on the concept of measuring the PB Max at a single suprathreshold level. Also, investigators in both Europe and the U.S. often confused it with the minor declines in PB performance scores that can be observed at high speech levels in cochlear disorders (named by Huizing & Reyntjes, le Chapeau de Gendarme, after the helmet shape of the French police officer). Our lab’s principal contribution here was to demonstrate that it is possible to derive a quantitative measure of the rollover effect that does, in fact, differentiate between cochlear and eighthnerve sites. We calculated the rollover index, PB Max − PB Min / PB Max, where PB Max is the highest score on the PI function, and PB Min is the lowest score of the PI function in the range of speech intensity levels above the level defining the PB Max. We applied this computation to the data on the affected ear of 41 patients with presumably cochlear loss (supported by other tests of cochlear site) and 10 patients with surgically or radiographically confirmed auditory nerve tumors. Figure 1–8 shows each individual rollover index. It is clear that there is no overlap between the data of the two groups.

1.  THE EARLY YEARS   25

Figure 1–8.  Showing how well a quantification of the rollover effect (PB Max − PB Min / PB Max) differentiated cochlear from eighth-nerve sites. There was no overlap between the two groups. Modified from Figure 5 in Jerger and Jerger (1971).

How can we explain rollover of the PI function for speech? At one time, I thought that the high intensity levels of the speech input were, perhaps, overloading the auditory nerve with too much information for it to successfully code speech. I think now, however, that a more parsimonious explanation is simply abnormal auditory adaptation. At these very intense speech levels, the loudness adapts rapidly and interferes with audibility, and the higher the intensity, the more rapidly this abnormal adaptation occurs. At the very highest levels, few of the words are even audible. You may find this difficult to accept. I was skeptical myself until I had observed a number of patients with Type III Békésy audiograms. At some point along the frequency scale, the threshold tracing suddenly departed from the level of the interrupted tone and descended to the bottom of the audiogram form in a

26  AUDIOLOGICAL RESEARCH OVER SIX DECADES

matter of seconds. The patient appeared to be hearing nothing at an SPL of 100 dB. After about the third time that this happened in the Northwestern clinic, I opened the door of the test chamber and was greeted by the continuous, ear-splitting tone from the test earphone. In one of my less rational moments, I shouted, “Don’t you hear that?” The patient just stared at me quizzically, as if to ask, “Is the test over?” Well, why worry about these relics of the past? You don’t need measures of abnormal adaptation or rollover for diagnostic purposes any more. There are better procedures available. But I would argue that you ought to study them in order to fully understand the nature of the hearing losses that you must deal with in clinical audiological practice. We need to find out what patients, especially elderly patients, are actually hearing rather than assuming it is the same thing that we are hearing. We also need to think about how our hearing tests, themselves, may be affected by undiscovered abnormalities in the hearing of the listener.

Synthetic Sentence Identification (SSI) If I had it to do over again, I probably would not have spent so much time on the Synthetic Sentence Identification (SSI) test. Clearly, the project was a steep uphill climb from the outset. The original idea was fourfold. First, change the speech audiometric task from repetition (open set) to identification (closed set); second, control the informational content of the test items; third, change each test item from a single word to a seven-word sentence; and fourth, always test in the presence of continuous, unrelated, and competing speech. A unique feature, I thought, was the construction of the sentences as sequences of third-order approximations to actual English sentences. The idea was to control the informational content of each sentence by essentially remov-

1.  THE EARLY YEARS   27

ing all meaning in favor of meaningless seven-word sentences. Chuck Speaks, now retired from the University of Minnesota, developed the test sentences and carried out the initial evaluation on listeners with normal hearing. I chose, as the competing speech message, a historical account of the life of Davy Crocket from a Sunday supplement of the now-defunct Houston Post. The importance of our competing message can hardly be overstated. The task of hearing a sentence, finding it on a list of 10 possible sentences, and reporting the number on the list is not challenging. In order to fashion a useful test, we added a continuous passage of ongoing competing speech, at a sentence-to-competition ratio of 0 dB. Whenever you see an SSI score in this book, you may assume that the test procedure included the continuous speech competition at a message-to-competition ratio (MCR) of 0 dB. In the instances where there is an exception to this rule, the exception is duly noted. The SSI test met with some initial success, but clearly it was too much to swallow in one sitting. People who had grown up with the PB lists were not ready for meaningless sentences or for competing messages. It took years for the field as a whole to accept a level of change far too complicated to digest in one sitting. Second, we were also criticized for having only 10 items rather than the 25- or 50-item PB lists that people were used to. We never made it sufficiently clear that we always constructed SSI-Performance versus intensity (PI) functions involving at least 30, but more likely 40 to 50, items. Less valid (or so I thought) was the suggestion that the listener could identify the sentence heard from the scraps of information available during the silent intervals in the continuous competing message. Missing from this criticism was the understanding that the point of the competing speech message was not to achieve total peripheral masking but to cognitively engage the information-processing system. In any event, the SSI test paradigm provided some interesting investigations of retrocochlear disorders.

28  AUDIOLOGICAL RESEARCH OVER SIX DECADES

BRAINSTEM VERSUS TEMPORAL LOBE In the 1960s, as the distinction between cochlear and auditory nerve sites unfolded, our interest turned toward the remainder of the retrocochlear components of the auditory system, the crossed pathways in the brainstem, and the auditory areas of the cerebral cortex. In Europe, during the 1950s, there had been considerable research on binaural fusion tests in the evaluation of brainstem disorders with generally positive results. But there had been little systematic study of the auditory cortex in the temporal lobes. Well into the 1950s, it was still widely held that there were no obvious auditory deficits associated with temporal lobe disorders. But all of this changed abruptly when a group of Italian investigators, led by Ettore Bocca, published a series of papers showing that, if speech audiometric measures, especially sentence tests, could be made more difficult for the listener (their term was “sensitized”), then substantial auditory deficits could be readily demonstrated on the ear opposite the affected temporal lobe. It seemed to me that our SSI sentences were ideally suited to pursue the kinds of studies of “sensitized” speech audiometry pioneered by the Italians. They were keen, for example, on time compression of sentences as a means of sensitization. Well, you can’t compress a PB word very much. They only last for half a second. But a sevenword sentence is ideal for this purpose. An unusual opportunity presented itself when, over a short space of time, two patients, one with a well-documented brainstem disorder and the other with an equally well-documented temporal lobe lesion, became available for intensive study. The patient with the brainstem disorder was a 33-year-old man. Exploratory surgery had revealed an intra-axial, eccentrically placed pontine glioma located on the right side of the brainstem at the level of cranial nerves IV and V. The patient with temporal lobe disorder was a 43-year-old woman who had undergone two right osteoplastic craniotomies for the removal of a large infiltrating glioma focused in the parietal region of the right cerebral

1.  THE EARLY YEARS   29

hemisphere but extending over a large area, including Heschl’s gyrus, and the posterior two-thirds of the right temporal lobe. As a control patient, we recruited a 37-year-old woman. She had undergone a right occipital craniotomy for removal of a glioblastoma in the parietal-occipital region. Despite this substantial surgical intervention, Heschl’s gyrus and the entire left temporal lobe had been spared. This was an ideal control for the possibility that extensive brain surgery per se, not involving the auditory areas, might explain poor performance on any of the SSI measures we employed in the study of the two experimental patients. We tested all three patients at some length, generating PI functions first with the simple SSI sentences without ipsilateral competition. As noted above, we ordinarily carried out SSI testing in the presence of ipsilateral competing continuous speech. In these three patients, however, the effect of ipsilateral speech competition was one of the experimental conditions. Thus, these data are an exception to the previously stated rule that we always tested SSI with ipsilateral competing speech. In effect, we were essentially comparing various methods of rendering the listening task difficult, in particular, comparing ipsilateral with contralateral competition of the same competing speech passage. Our aim was to seek similarities and differences among the two experimental patients and their control, as we systematically altered the difficulty of the listening task, by low-pass filtering the sentences, time-compressing the duration of each sentence, mixing a competing speech message in the same ear with the sentences, or presenting the competing speech message to the opposite ear. Adding speech competition to the other conditions would have driven performance very nearly to zero level, especially in the case of the brainstem patient. Figure 1–9 shows PI functions for both ears in each of the three patients for SSI-sentences-in-quiet (i.e., without the presence of the competing speech message). The sentences were always presented at an intensity level of 50 dB SPL. This is an easy listening task. It is not surprising, therefore, that PI functions rose

30  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 1–9.  PI functions for unaltered SSI sentences in three patients, one with a well-documented disorder affecting the auditory pathways in the brainstem, a second with a well-documented disorder affecting the auditory area in the temporal lobe, and a third who had undergone occipital craniotomy in the parietal/occipital area, sparing any auditory areas. She was considered to be a control for the other two. All results appear to be equivalent. In particular, there were no interaural asymmetries in any of the three listeners. These unaltered sentence results serve as a baseline for comparison with various degradations of the signals in subsequent figures.

quickly to 100% correct scores in both ears of all three listeners as sentence intensity increased. We had, therefore, three equivalent baselines against which to evaluate the effects of subsequent changes in the listening task across the three study patients. Figure 1–10 shows the effect of low-pass filtering the sentences. It is evident that setting the low-pass frequency as low as 250 Hz had little effect on either the control or the temporal lobe listener but a profound effect on the brainstem listener. Here, performance on the right ear was only 60% at the 1000-Hz cutoff frequency and 0% at the 500-Hz frequency. Performance was even poorer on the left ear: only 40% at the 1000-Hz cutoff frequency

1.  THE EARLY YEARS   31

Figure 1–10.  The effect of low-pass filtering the SSI sentences. Speech energy above the low-pass cutoff frequency has been removed from the sentences over a filter cutoff frequency ranging from 200 to 2000 Hz.The control listener scored 70% to 80% at the 250-Hz cutoff.The temporal lobe patient did even better, scoring 90% to 100% at the 250Hz low-pass cutoff frequency. But the brainstem patient scored poorly on both ears and significantly worse on the left ear. Here scores at the 1kHz cutoff were 60% on the right ear and only 40% on the left ear. At the 2-kHz cutoff, SSI scores climbed to 80% on the left ear but fell to 30% on the right ear. Clearly, even mild low-pass filtering had a devastating effect on the brainstem patient. The contrasts with the performances of the temporal lobe patient and the control patient were striking.

and only 30% at the 2000-Hz cutoff. The fact that performance was substantially poorer on the left ear although the brainstem glioma was eccentric to the right side of the brainstem is consistent with the fact that the glioma was at the level of the fourth and fifth cranial nerves, well above the olivary complex where the inputs from the two ears enter the brainstem and cross and ascend on opposite sides of the brainstem. What we see here on the left ear is pretty much just a continuation of the expected picture for a patient with a right auditory nerve lesion, but a large

32  AUDIOLOGICAL RESEARCH OVER SIX DECADES

glioma at this level in the brainstem can affect the right ear/left brainstem pathway as well. In Figure 1–11, we see the effect of time compression. Again, the SSI sentences were presented at 50 dB SPL. Here the duration of each sentence had been compressed by 50% by discarding every other 25-ms time interval. Neither the control nor the temporal lobe patient was particularly affected by the shorter duration of each sentence. In the brainstem patient, however, a pattern similar to the effect of low-pass filtering was repeated: poorer performance on both ears but much worse on the left ear. And we revisit our old friend — rollover: severe on the PI functions of both ears. Figure 1–12 shows the effect of adding an ipsilateral competing speech condition (target and competition in the same ear) to

Figure 1–11. The effect of time-compressing the SSI sentences. Fifty percent time compression was achieved by discarding alternate 25-ms segments of the speech waveform and reforming the sentence continuity. Again, only the brainstem patient was affected. PI function scores were degraded on both ears but significantly poorer on the left ear. The temporal lobe patient’s scores were unaffected.

1.  THE EARLY YEARS   33

Figure 1–12.  Effect of ipsilateral speech competition. SSI sentences (message) were always presented at 50 dB SPL. The intensity of competing speech was varied to achieve a range of message-to-competition ratios (MCRs). The temporal lobe patient did almost as well as the control, but again, the brainstem patient did less well, especially at MCR = 0 dB.

an easily heard sentence (message) presented at 50 dB SPL, as the message-to-competition ratio (MCR) was varied from +10 dB to −20 dB. The +10-dB and 0-dB MCR conditions were relatively easy; both the control and the temporal lobe patients scored 90% to 100%, but somewhat less well at the more difficult MCR conditions of –10 dB and –20 dB (for the control patient, 60%–70%; for the temporal lobe patient, 40%–60%). But, again, the brainstem patient scored even less well (20% at MCR of −10 dB on the right ear and only 70% at +10 dB MCR on the left ear). Finally, Figure 1–13 shows the effect of introducing the same competing speech message not to the same ear but to the opposite ear (contralateral competing condition). This is an easier task. All three listeners did well on the right ear all the way down to an MCR of −60 dB. Moreover, for the first time, the brainstem patient did better than the temporal lobe patient, although

34  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 1–13.  Effect of contralateral speech competition. Interestingly, the right ear was unaffected by the contralateral speech competition, but both the brainstem and temporal lobe patients were affected on the left ear: And, for the first time, the effect was greater for the temporal lobe patient than for the brainstem patient. Here the wellknown primacy of the right ear to left side of the brain pathway for the processing of speech and language is the explanation for the left ear effect in both cases.

neither did well on the left ear. Here we see the defining pattern of a lesion in the auditory cortex, substantially poorer performance on a moderately difficult listening task on the ear contralateral to the affected side of the brain, but relatively normal on the ipsilateral side (in this case, the right ears of both the brainstem and temporal lobe patients). The SSI data in these last five figures suggest that the distinction between a brainstem disorder and a temporal lobe disorder is complicated. With the single exception of the contralateral competing condition, the brainstem disorder affected both ears but performance was uniformly poorer on the left ear. At first glance, it would seem that, since the brainstem tumor was eccentric to right side of the brainstem, the greatest degradation of per-

1.  THE EARLY YEARS   35

formance should have appeared on the right ear. But shortly after the auditory nerves from the two cochleae enter the brainstem, at the level of the olivary complex, they cross and rise up the brainstem, right ear path up the left side of the brainstem, left ear path up the right side of brainstem. Since the tumor was above the crossing point of the two pathways, at the level of cranial nerves IV and V, its rightward eccentricity affected not the right ear input but the left ear input. The key to understanding this reversal is to remember that the inputs from the two ears cross in the lower brainstem and ascend on opposite sides — right ear to left side of brain, left ear to right side of brain. The worst degradation of performance in the case of the brainstem patient occurred in the low-pass filtering condition, followed closely by the time-compression condition. In both conditions, percent-correct scores never exceeded 40% on the left ear. But in the contralateral speech competition condition, performance was at the 100% level on the right ear at all levels tested and climbed to 100% at the MCR of −20 dB on the left ear. In contrast, the scores for the temporal lobe patient were essentially equivalent to the performance of the control listener in all conditions except contralateral speech competition. Here, of course, we found the signature picture of a cortical auditory disorder, poor performance (70%) only on the ear contralateral to the affected hemisphere of the brain. As I look back over these five figures after half a century, I think that the SSI procedure did a good job of differentiating these two retrocochlear sites of auditory disorders. Time, tide, and new tests have filled the intervening years, but I do not think that the picture reflected in these SSI results has changed appreciably. This chapter pretty much covers the early years at Northwestern and at the Houston Speech and Hearing Center. They were, I think, quite productive years, but the next three decades at Baylor brought two important additions to our modest research program, graduate students, and access to a larger stream of clinical patients. I was now able to expand our research team and

36  AUDIOLOGICAL RESEARCH OVER SIX DECADES

to carry out studies involving very large groups of subjects. In my new position as Professor of Audiology in the Department of Otolaryngology and Communicative Sciences, at Baylor, and with the patients and resources provided by the Methodist Hospital in the Texas Medical Center, I was able to expand the audiology research program substantially. Chapters 2 to 7 cover some of the research carried out during that interesting period.

RELATED READINGS Jerger, J. (1952). A Difference Limen Recruitment Test and its diagnostic significance. Laryngoscope, 62, 1316–1322. Jerger, J. (1955). Differential intensity sensitivity in the ear with loudness recruitment. Journal of Speech and Hearing Disorders, 20, 183–191. Jerger, J. (1960). Békésy audiometry in the analysis of auditory disorders. Journal of Speech and Hearing Research, 3, 275–287. Jerger, J., & Bucy, P. (1960). Audiological findings in an unusual case of VIIIth nerve lesion. Journal of Auditory Research, 1, 26–35. Jerger, J., Carhart, R., & Lassman, J. (1958). Clinical observations on excessive threshold adaptation. Archives of Otolaryngology, 68, 617–623. Jerger, J., & Jerger, S. (1968). Progression of auditory symptoms in a patient with acoustic neurinoma. Annals of Otology, Rhinology, and Laryngology, 77, 230–242. Jerger, J., & Jerger, S. (1971). Diagnostic significance of PB word functions. Archives of Otolaryngology, 93, 573–580. Jerger, J., Shedd, J., & Harford, E. (1959). On the detection of extremely small changes in sound intensity. Archives of Otolaryngology, 69, 200–211. Jerger, J., Speaks, C., & Trammell, J. (1968). A new approach to speech audiometry. Journal of Speech and Hearing Disorders, 33, 318–328.

1.  THE EARLY YEARS   37

Speaks, C., Jerger, J., & Trammell, J. (1970a). Comparison of sentence identification and conventional speech discrimination score. Journal of Speech and Hearing Research, 13, 755–767. Speaks, C., Jerger, J., & Trammell, J. (1970b). Measurement of hearing handicap. Journal of Speech and Hearing Research, 13, 768–776.

2 Immittance Audiometry

39

40  AUDIOLOGICAL RESEARCH OVER SIX DECADES

In 1960, I toured a number of audiology facilities in Denmark and Sweden. In the laboratory of Dr. Harald Ewertsen, at the City Hospital in Copenhagen, I learned about the latest development in the measurement of what was then called “middle ear impedance”: the Madsen ZO60 instrument developed by Danes Knut Terkildsen and S. Scott-Nielsen. The ZO60, like earlier devices, measured the impedance of the middle ear in its natural state, sometimes called “static” or “absolute impedance,” but the Danes added a second dimension of measurement: They provided the additional capability of sealing the external ear canal to form a closed volume in which air pressure could be varied. This made it possible to plot “dynamic” impedance change as air pressure within the chamber was varied from +200 to −200 mm H2O. They called the resulting graph a “tympanogram” and immediately opened a new and powerful addition to the fundamental tools of audiological measurement. Dr. Ewertsen also showed me how the device could be used to detect a contraction of the stapedius muscle in the contralateral ear (the ability to detect the ipsilateral reflex came a little later). Back at Northwestern, Professor Carhart invited me to report to the audiology faculty and students on my impressions from the trip. I spent most of the talk on the Madsen device, suggesting that it was surely going to have a substantial impact on our clinical practice and that we all needed to learn more about impedance concepts. As the audience shuffled from the lecture room, I sensed a palpable lack of enthusiasm for any part of my talk. My mentor, too, was singularly unimpressed. Shortly thereafter, I moved from Northwestern to the VA in Washington, D.C., where I ordered a ZO60 unit from the Madsen Company. Our unit was the first device in the U.S., and I looked forward to exploring it. However, I soon left Washington for Houston and had to leave the unit behind. For the next 6 years, at the Houston Speech and Hearing Center (HSHC), my interest in impedance measurement lapsed as I became immersed in the

2.  IMMITTANCE AUDIOMETRY   41

SSI Sentences Project described in Chapter 1. At about that time, the Grason Stadler Company produced an ingenious mechanical bridge designed by Joe Zwislocki. I bought one of the Zwislocki bridges, only to learn that it was not useful in the clinic because it was cumbersome to operate and lacked the ability to create tympanograms. So, I was back to square one. In the meantime, the Madsen Company introduced the next version of the Madsen bridge, the model ZO70, in the U.S. I was anxious to acquire a model ZO70. But then my wife Susan and I moved to the Department of Otolaryngology of the Baylor College of Medicine (BCM), in the Texas Medical Center. BCM became our professional home for the next 29 years. Soon after moving to our new location, we were visited by an angel-indisguise, Mr. Jimmy Brown. Jimmy had the Madsen franchise for Texas, Oklahoma, and Louisiana. He came by to introduce himself and to see if I might be interested in buying a Madsen ZO70. I replied that I was indeed interested but was temporarily embarrassed for the necessary funds. “But,” said I (with my fingers crossed), “if you will just loan me one for a while, enough time for me to test some patients and write up the results, I think that might generate some sales in your territory.” Well, Mr. Brown agreed. That was in 1968. By the end of 1969, we had tested more than 400 patients and started writing up our results. The paper, entitled “Clinical Experience With Impedance Audiometry,” was published in 1970 in the Archives of Otolaryngology. Today, 50 years later, it continues to be cited. This was certainly not the first paper on impedance audiometry. Significant articles had already been published by Metz, Jepson, Ewertsen, and Terkildsen in Denmark; by Liden, Møller, Klockhoff, Anderson, Barr, and Wedenberg in Sweden; and by Brooks in the United Kingdom. All presented excellent work on one or another relatively isolated dimension of the problem from a research standpoint, but translating it all for practicing clinicians had not been effective.

42  AUDIOLOGICAL RESEARCH OVER SIX DECADES

I like to think that my 1970 paper brought it all together. I tried to make the important point that, of the three measures then popular, the absolute or static impedance, the tympanogram, and the acoustic reflex each had serious diagnostic limitations when viewed in isolation, but that, taken as a test battery, they yielded fairly consistent patterns of particular diagnostic value. This is a message that I have tried to convey in connection with all of our diagnostic audiological tests, trusting the concept of the test battery and its unique patterns, rather than searching for answers from individual test outcomes. In any event, as we considered the various patterns afforded by the three measures, it became clear that, of the three — static impedance, tympanogram, and stapedius muscle reflex — the least useful was the static impedance. The overlap among normal hearers and people with abnormalities was so great that this component added very little to the diagnostic challenge. Ironically, this was the dimension that had initially motivated the search for a satisfactory technique of middle ear measurement. It was not long before we had abandoned the static measure entirely and concentrated our attention on the tympanogram and the stapedius muscle reflex.

THE TYMPANOGRAM The tympanogram has come a long way since the days when it was measured at a single probe-tone frequency. Today, wideband acoustic immittance, based on tympanograms measured over a band of frequencies, has substantially broadened the clinical utility of tympanometry. In this area, we have only begun to reap the benefits of the new technology. I will say no more about tympanometry and focus further discussion of immittance audiometry on our diagnostic jewel, the reflexive contraction of the stapedius muscle.

2.  IMMITTANCE AUDIOMETRY   43

THE STAPEDIUS MUSCLE REFLEX There are, of course, two muscles in the middle ear, the tensor tympani and the stapedius. Most investigators have paid very little attention to the tensor tympani muscle because it does not ordinarily contract to sound (but see observation later recounted). Then what is it doing in there? Actually, it is part of the generalized reaction of the entire human organism to surprise or shock. Sound will elicit it, but only if the sound has some surprise or shock value (for example, a very loud sound that you are not expecting). The stapedius muscle, however, readily contracts to a lower range of sound intensities, usually in the range from 60 to 100 dB SPL. Moreover, the strength of the contraction is proportional to the sensation level of the eliciting sound. A number of different roles have been suggested for the stapedius reflex. At first, investigators thought that it evolved as a protective mechanism, preventing the footplate of the stapes from excessive inward motion against the fluid system of the cochlea. In recent years, however, alternative interpretations of its role seem more likely. One currently favored idea is that the contraction of the stapedius muscle evolved as a method for attenuating the low-frequency energy created by the action of the muscles of mastication, so that they do not mask the high-frequency sensitivity of the auditory system necessary for detecting the approach of potential enemies (or meals) in the wild. In our laboratory at the Baylor College of Medicine, in the 1970s and 1980s, we spent a good deal of time studying the effects of various auditory disorders on stapedius reflexes. Actually, there are four separate and distinct contractions, two crossed and two uncrossed. When sound is introduced to the right ear, both the stapedius muscle in the right middle ear (uncrossed/ipsilateral) and the stapedius muscle in the left middle ear (crossed/contralateral) contract. Similarly, when sound

44  AUDIOLOGICAL RESEARCH OVER SIX DECADES

is introduced to the left ear, both the stapedius muscle in the left middle ear (uncrossed) and the stapedius muscle in the right middle ear (crossed) contract. In the case of middle ear disorders and cochlear disorders, the patterns across the four reflex conditions are straightforward and predictable from the degree of loss involved. In persons with normal hearing, the reflex begins to contract when the sensation level (SL) of the sound stimulus reaches 60 to 70 dB. Further increases in stimulus level produce progressively stronger contractions up to an SL of 100 to 105 SL. In the case of cochlear hearing loss, the reflex contraction will occur at a much lower SL, apparently responding to the physical intensity of the stimulus rather than to its SL (but see later discussion of eighth-nerve tumors). Suppose, for example, that the threshold HTL is 50 dB in a cochlear ear. Then, a stimulus at 60 dB HTL might produce a reflex response equivalent to the response observed in the normal ear at 60 dB HTL. It is as if the reflex amplitude has, over the narrow range of only 10 dB, “caught up” with what the normal ear would look like at that physical intensity of stimulation. This is beginning to sound like “loudness recruitment”: And, indeed, the stapedius reflex at a lower-than-normal SL has been proposed by many investigators as a test of loudness recruitment, ergo a test for cochlear site of disorder. But there has been diminishing interest in loudness recruitment, or making the distinction between cochlear and retrocochlear sites of disorder in recent decades. As we studied the stapedius muscle responses in persons with retrocochlear sites of disorder, however, we noted some very interesting results. The patterns observed among the two crossed reflex waveforms and the two uncrossed reflex waveforms varied uniquely with the site of the retrocochlear lesion. Especially interesting were the results in patients with eighth-nerve and brainstem sites of disorder. Here results could vary from loss of only one reflex to loss of all four reflexes.

2.  IMMITTANCE AUDIOMETRY   45

Reflex Averaging As we pursued stapedial reflexes in patients with the available commercial testing devices, we began to wonder whether we were missing something by not studying the morphology of the reflex responses more closely. Commercial devices were limited in their capabilities. You could note whether a reflex was present or absent, observe the onset and offset of the reflex if it were strong enough, and deduce something about reflex decay; that was about it. But this was the era of the auditory brainstem response (ABR), an elegant example of the wonders of signal averaging. “Why not,” we thought, “apply the principle of signal averaging to the recording of the stapedius reflex?” Approaches to the manufacturers of impedance devices met cold stares. Admittedly, the profit potential was minimal. So, we decided that we would have to build our own stapedius-reflex-averaging system. I will not elaborate the details of the operation of this necessarily complex gadgetry or how much it cost in both dollars and effort. Suffice to say that the reflex-eliciting signal alternated between ears, the probe tone was 220 Hz, and each reflex response was averaged over eight sweeps of the computer. In the following pages, we show and discuss reflex findings in three listeners. All had normal pure-tone audiograms in both ears. No interaural differences can be attributed, therefore, to sensitivity differences between the two ears. Figure 2–1 shows all four reflexes recorded from the ears of a young adult with normal hearing and no auditory complaints. All uncrossed waveforms in this series (red) have been smoothed to eliminate the acoustic artifact due to the interaction of the probe tone and the reflex eliciting tone in the same cavity. The shape of the reflex response was two-phased, rising rapidly over the first 250 to 300 ms, then leveling off and continuing to rise more gradually over the next 200 to 250 ms. The relaxation of the normal response was more gradual than the onset, lasting another 400 to

46  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 2–1. The four signal-averaged stapedius muscle reflexes in response to a 2-kHz tone burst at 100 dB SPL for 500 ms. The two uncrossed waveforms are shown in red, the two crossed waveforms in black. All four are on the same time and amplitude scales. The listener was a young adult man with a normal audiogram and no auditory complaints. Note the steep early onset response, followed by a more gradual response to the peak amplitude, then gradual decline to baseline. Each uncrossed response is substantially larger than its crossed counterpart. Modified from Figure 1 in Hayes and Jerger (1982).

500 ms. Note as well that the maximum amplitudes of the waveforms were substantially greater for the uncrossed responses. Figure 2–2 shows the same four-waveform display in the case of a woman with an acoustic tumor on the right eighth nerve. Despite equivalent audiometric thresholds on the two ears, the reflex waveforms from the affected right ear were substantially reduced in amplitude. This is consistent with loss in suprathreshold loudness, characteristically observed in cases of unilateral acoustic tumor. In both the crossed and uncrossed waveforms for sound in the right ear, the initial rapid rise demonstrated in

2.  IMMITTANCE AUDIOMETRY   47

Figure 2–2. The four signal-averaged stapedius muscle reflexes in response to a 2-kHz tone burst at 100 dB SPL for 500 ms. The listener was a patient with an acoustic tumor on the right eighth nerve. The normal waveforms from the left ear showed the initial rapid-onset phase, followed by the more gradual rise to maximum. Because of the loudness loss in the right ear, due to the tumor, the two waveforms from the right ear stimulation are greatly attenuated in amplitude. In addition, there is absence of the initial steep phase of the normal waveform. Only the gradual phase remains. And, again, in both ears, the amplitude of the uncrossed response is larger than its crossed counterpart. Modified from Figure 3 in Hayes and Jerger (1982).

Figure 2–1 was clearly missing in comparison with the two waveforms from left ear stimulation. We had observed this pattern in several other cases of unilateral acoustic tumor. It speaks to what was at one time a spirited controversy over the basis for the contraction of the stapedius muscle. One school of thought asserted that the reflex was triggered by the physical intensity of the sound stimulus. This was based principally on the notion of protection against excessively intense acoustic events. The counterargument was that perceived loudness, not physical intensity, was, indeed,

48  AUDIOLOGICAL RESEARCH OVER SIX DECADES

the sufficient stimulus. These findings with unilateral acoustic tumor patients argue strongly that loudness is the key to strength of contraction. As illustrated in the case described above, when the two ears were stimulated by the same physical intensity, the loss of loudness due to the tumor significantly altered the reflex waveform in the affected ear. In Figure 2–3, the picture was dramatically different. The patient was a woman with multiple sclerosis involving the brainstem. Uncrossed reflexes in both ears were present and robust, reflecting a steep onset followed by a more gradual increase up to the peak of the waveform, but both crossed reflexes were almost eliminated. Here the known brainstem disorder suggests involve-

Figure 2–3. The four signal-averaged stapedius muscle reflexes in response to a 2-kHz tone burst at 100 dB SPL for 500 ms. The listener was a patient with multiple sclerosis (MS) involving the lower brainstem. Because the MS affected the crossed innervation pathways across the brainstem, the crossed reflexes were almost eliminated while the uncrossed reflexes were unaffected. Modified from Figure 4 in Hayes and Jerger (1982).

2.  IMMITTANCE AUDIOMETRY   49

ment of the crossed pathways through the brainstem innervating the stapedius muscle on the opposite side of the head. These figures appeared in print almost 40 years ago. They have always suggested to me that the dynamics of the stapedius muscle reflex are more complicated than one might suppose from illustrations in current textbooks. We showed that, in the normal case, there are three components: (1) a rapid-onset phase, followed by (2) a more gradual increase in strength, followed by (3) a gradual relaxation. The maximum amplitude of the uncrossed reflex is always greater than the maximum amplitude of the crossed reflex with sound in the same ear, a difference almost certainly due to some loss in strength of the innervation of the crossed reflex through the brainstem to the opposite side of the head. This suggestion is supported by the findings illustrated in Figure 2–3 for the patient with multiple sclerosis. Here the presence of brainstem disorder almost eliminated the crossed reflexes but had no effect on the uncrossed reflexes. In the case of eighth-nerve disorder, both the crossed and uncrossed reflex amplitudes were reduced in proportion to the loss in effective loudness of the reflex-eliciting stimulus. In addition, there appeared to be a loss in the early, fast-onset component of the normal reflex. The reflex characteristics of the opposite, normal ear were apparently not altered by the loudness loss on the affected ear. In the case of the intra-axial brainstem disorder due to multiple sclerosis, both uncrossed reflexes were unaffected, but both crossed reflexes were almost abolished.

Other Reflex Oddities We had many other interesting findings with our homemade reflex signal-averaging system. One colleague, Brad Stach, encountered an unusual facial nerve problem involving what looked like an acoustically elicited contraction of the tensor tympani muscle, certainly a rare event. On another occasion, we set

50  AUDIOLOGICAL RESEARCH OVER SIX DECADES

up the reflex recording apparatus in the animal lab of colleague Makoto Igarashi and recorded stapedius reflexes in a squirrel monkey (see Chapter 10 of this volume). We even succeeded in capturing the waveform of an electrically elicited stapedius contraction in a cochlear implant patient. It was a sad day in the lab when we finally had to disassemble the apparatus to make room for a multichannel brain-mapping system. But that story will appear in Chapter 8. I have always found these averaged reflex data intriguing. They suggest that there is much interesting information still to be mined from this vital feature of the auditory system in the ear and in the brain.

AN EDUCATIONAL ADVENTURE Shortly after the publication of the 1970 article “Clinical Experience With Impedance Audiometry,” I was visited in Houston by Mr. Paul Madsen, owner of the Danish company, Madsen Electronics. His company manufactured and distributed the Madsen ZO70 device, which, thanks to Jimmy Brown, I had used to gather all of the data for the 1970 article. Madsen was accompanied by Mr. Irwin Klar, a New York businessman whom Madsen had engaged to promote the sale of the Madsen ZO70 unit in the U.S. Madsen had the idea to send out a team of lecturers across the country to give 1-day courses on basic impedance concepts and use of the ZO70 in the audiology clinic. His idea was that Mr. Klar would organize all of this for the entire country. Irwin was not initially enthusiastic about the idea but agreed to take it on. Later, he became its strongest supporter. He recruited Jerry Northern as the key speaker if I would be there as well for backup. Thanks to Jerry’s enthusiasm and lecturing skills, the project enjoyed considerable success. We saw a number of cities from a new perspective and enjoyed meeting many fellow audiologists. By the time it all ended Jerry, Irwin, and I were a satisfied trio.

2.  IMMITTANCE AUDIOMETRY   51

RELATED READINGS Hayes, D., & Jerger, J. (1981). Patterns of acoustic reflex and auditory brainstem response abnormality. Acta Otolaryngology, 92, 199–209. Hayes, D., & Jerger, J. (1982). Signal averaging of the acoustic reflex: Diagnostic applications of amplitude characteristics. Scandinavian Audiology Supplement, 17, 31–36. Jerger, J. (1970). Clinical experience with impedance audiometry. Archives of Otolaryngology, 92, 311–324. Jerger, J., Fifer, R., Jenkins, H., & Mecklenburg, D. (1986). Stapedius reflex to electrical stimulation in a patient with cochlear implant. Annals of Otology, Rhinology, and Laryngology, 95, 151–157. Jerger, J., Harford, E., Clemis, J., & Alford, B. (1974). The acoustic reflex in eighth nerve disorders. Archives of Otolaryngology, 99, 409–413. Jerger, J., & Hayes, D. (1983). Latency of the acoustic reflex in eighth nerve tumor. Archives of Otolaryngology, 109, 1–5. Jerger, J., Oliver, T., & Jenkins, H. (1987). Suprathreshold abnormalities of the stapedius reflex in acoustic tumor: A series of case reports. Ear and Hearing, 8, 131–139. Jerger, J., Oliver, T., Rivera, V., & Stach, B. (1986). Abnormalities of the acoustic reflex in multiple sclerosis. American Journal of Otolaryngology, 7, 163–176. Jerger, J., Oliver, T., & Stach, B. (1986). Problems in the clinical measurement of the acoustic reflex latency. Scandinavian Audiology, 15, 31–40. Jerger, S., & Jerger, J. (1977). Diagnostic value of crossed vs uncrossed acoustic reflexes: Eighth nerve and brain stem disorders. Archives of Otolaryngology, 103, 445–453. Stach, B., & Jerger, J. (1984). Acoustic reflex averaging. Ear and Hearing, 5, 289–296. Stach, B., Jerger, J., & Jenkins, H. (1984). The human acoustic tensor tympani reflex: A case report. Scandinavian Audiology, 13, 93–102.

3 Auditory Processing Disorder

53

54  AUDIOLOGICAL RESEARCH OVER SIX DECADES

THE BIRTH OF APD In the early 1950s, as a graduate student at Northwestern, I was assigned, during one quarter, to the clinic of Dr. Helmer Myklebust. Myklebust held a weekly hearing clinic for the evaluation of children who had not developed adequate speech and language by the age of 3 to 6 years. The problem, as Myklebust saw it, was to differentiate among severe hearing loss, motor problems, and emotional issues. He was certainly not the first to face these challenging problems, but he was one of the first to attempt to apply hearing testing to children at such a young age. It was an era in which pediatricians and other medical professionals believed that you could not carry out successful audiometry until the child was at least 6 years old. But Myklebust had had considerable experience with testing children who were severely hearing impaired. There was deafness in his own family. Moreover, a brother headed a School for the Deaf. Myklebust chose to join the faculty of the audiology program at Northwestern precisely because of the university’s reputation as a leader in the young field of audiology and especially audiometry. As one who assisted in the attempted auditory testing of very young children in his clinic, I can attest to his dedication to the idea that children with deafness, and/or other problems, can and must be identified at the earliest possible age. Most of the children his clinic examined turned out to have severe hearing loss, but a significant number had relatively normal hearing levels for tones, noises, mother’s voices, and so forth. Yet these children, with apparently normal hearing sensitivity, were still deficient in language development. They did not seem to process auditory information appropriately. William Hardy, who conducted a similar clinic at The Johns Hopkins University, observed the same issues in Baltimore. He suggested that these children could hear well enough, but they did not seem to “aud” well. Why had such children not been described before? Probably because, in the days before sophisticated hearing testing

3.  AUDITORY PROCESSING DISORDER   55

was widely available, people assumed that most children with delayed language were either deaf or retarded, and they were routed to the appropriate residential school. Myklebust puzzled over these children for some time. In the end, he concluded that they might have a problem at a central level, perhaps analogous to the issues sometimes observed in children who showed visual perceptual problems as a result of central nervous system disorders. Myklebust entertained the idea of a specific auditory perceptual disorder, perhaps a very mild form of the brain-related phenomenon of agnosia, but insisted on keeping an open mind on the subject. Yet he was impressed by the work of Sir Henry Head, an English neurologist who had studied the effects of war-related head injuries on the speech and language of British soldiers wounded in World War I. In his two-volume set, Aphasia and Kindred Disorders of Speech, most of the patient material concerned the various forms of aphasia, but Head devoted considerable space in one volume to the subject of agnosia as well. Myklebust insisted that each student in his classes acquire a copy of Head’s two-volume work and study his description of agnosia. It seemed to our class that Myklebust had in mind some milder form of the brain injury that produced agnosia. In any event, it was never in doubt that Myklebust had some form of brain deficiency in mind. Myklebust’s observations, early in the 1950s, influenced a number of audiologists who had long been concerned about such children themselves. A few began to think about how one might develop tests to assist in the differential diagnosis of children who did not “aud” well.

The Italian Pioneers Within that same time frame, the early 1950s, a group of Italian investigators in Milan, led by Dr. Ettore Bocca, published

56  AUDIOLOGICAL RESEARCH OVER SIX DECADES

landmark papers on patients with temporal lobe tumors. Up to this point, most investigators believed that disorders of the brain, like temporal lobe tumors, had little effect on the pure-tone audiogram or on word recognition tests. This conclusion was largely based on animal studies in which large sections of the auditory areas of the animal’s brain were totally removed without obvious effect on responses to auditory stimuli. But Bocca’s lab demonstrated that if you heightened the difficulty of the listening task, the speech recognition score of the ear opposite the affected side of the brain was substantially poorer than the speech recognition score from the ipsilateral ear. They experimented with various techniques for “sensitizing” the tests by several means, including time compression, low-pass filtering, presenting the words at low sensation level, presenting low-pass filtered words in one ear and the same words at low sensation level in the other ear, and finally low-pass filtering the word to one ear while high-passing it to the other ear. All worked reasonably well. The key idea was simply to make the listening task more difficult by degrading the quality of the stimulus in some way. An example from the extensive research of Bocca and his associates is illustrated in Figure 3–1. It shows the effect of time compression on sentence understanding in the two ears of a patient after right temporal lobectomy. The PI function for the right ear was normal, but in the case of the left ear, the maximum never exceeded 64%. As speech sensation level continued to climb, there was marked rollover, down to less than 10% at SL = 70 dB. Note, also, that the deficit appeared not on the right ear but on the left ear, the ear opposite the affected side of the brain. By the 1950s, the fact that the auditory pathways from the ears to the brain crossed in the brainstem and traveled to the auditory areas of the temporal lobes on opposite sides of the head was well known. But Bocca’s data produced vivid illustrations of the consequences of the crossed wiring arrangement. Later we shall see how this fact accounts for the right ear advantage in dichotic listening to speech and the left ear advantage in dichotic listening to nonspeech.

3.  AUDITORY PROCESSING DISORDER   57

Figure 3–1.  Effect of time compression on the PI function for Italian words spoken at approximately 350 words per minute, in a patient after a right temporal lobectomy. On the left ear, the ear opposite the affected temporal lobe, the PI function for words showed a lowered maximum score and extreme rollover. Time compression was achieved by recording a speaker while he talked very fast (E. Bocca, personal communication, circa 1964). Modified from Figure 15 in Bocca and Calearo (1963).

I once asked Professor Bocca how he produced time compression in his laboratory in those precomputer times. He smiled sheepishly, then whispered that they simply instructed the talker to talk faster. The talker had to practice this until he could reach 140, 250, and a staggering 350 words per minute. In those days, one had to be imaginative.

Early Efforts in the U.S. These findings of the Italian pioneers set off a burst of activity, especially in the U.S. The result was a catalogue of new tests for

58  AUDIOLOGICAL RESEARCH OVER SIX DECADES

“auditory processing disorder” or APD. There is scant evidence, however, that brain injury or deficit has been demonstrated in the bulk of children identified as having APD on the basis of APD testing. The evolving rationale for APD tests was that, if people with known brain injury do poorly on a particular APD test, then persons who do poorly on that test must be brain injured. But even if B follows A, it is not necessarily the case that B is due to A. You learned that in Logic 101. My own foray into this arena was a paper published in The Laryngoscope in 1960, entitled “Audiological Manifestations of Lesions in the Auditory Nervous System.” In it I presented what we then called “Special Test” data gathered on five patients with unilateral lesions along the pathway from the middle ear to the auditory cortex. I contrasted two auditory measures: (1) low-pass filtered PB words and (2) the alternate binaural balance test in patients with unilateral lesions ranging from the middle ear to the temporal lobe. My goal was to show that there was a distinctly different pattern of test results depending on whether the site of the auditory disorder was in the middle ear, the cochlea, the eighth nerve, the brainstem, or the temporal lobe. My test battery at the time included the SISI test, the Békésy audiogram, the PB score at SL = 25 dB, the PI PB function, low-pass filtered PB words, very faint PB words (SL = 5 dB), and the ABLB test. In retrospect, the only reason for including ABLB was the fact that it had played such a critical role in the 1948 paper by Dix, Hallpike, and Hood in making the distinction between Meniere’s disease and acoustic tumor. We were all so caught up in those days with the findings of Bocca and his team using sensitized speech tests that I didn’t pay as much attention to the ABLB results as I should have. But now, as I look back over those papers from 50 years ago, I have noted a startling consistency in virtually every patient with unilateral brainstem or temporal lobe site of disorder that we evaluated. Whereas patients with unilateral cochlear site all showed a “gain” in loudness as the tone intensity

3.  AUDITORY PROCESSING DISORDER   59

increased above the impaired threshold on the affected ear (i.e., loudness recruitment), in the case of unilateral brainstem and temporal lobe sites, there was a loudness “loss” as tone intensity increased above threshold on the affected (contralateral) ear. This difference is illustrated in Figure 3–2 in two patients, one with left Meniere’s disease, the other with left temporal lobe epilepsy. In each case, the familiar laddergram is shown for binaural balances at 1000 Hz. In the case of the patient with left Meniere’s disease, there was a 50-dB difference between the thresholds at the two ears, but that large difference almost “disappeared” at SL = 80 dB. The left ear had “caught up” with the right ear. It had “gained” or “recruited” loudness. But in the case of the patient with left temporal epilepsy, it was just exactly the opposite. A tone at an SL of 65 dB on the right ear (the ear opposite the brain lesion)

Figure 3–2.  Alternate binaural balance test results at 1000 Hz in two patients: one with Meniere’s disease in the left ear, the other with left temporal epilepsy. Modified from Figures 2 and 5 in Jerger (1960).

60  AUDIOLOGICAL RESEARCH OVER SIX DECADES

sounded as loud, to the patient, as a tone at only 45 dB SL on the left ear. Loudness had been “lost” on the right ear. Remarkably, we saw this configuration on every brainstem or temporal lobe patient we tested. Why was loudness lost on that right ear? Could it be some form of abnormal loudness adaptation unique to the brainstem and temporal lobe auditory pathways? In any event, I  wish that I had been paying more attention to those ABLB findings instead of focusing so keenly on the speech perception data. Audiologists who have attempted to obtain stable and repeatable audiograms from listeners with temporal lobe disorders often recite a common mantra. “You can never get the same thing twice! They are all over the map!” One possible explanation for this troubling moment-to-moment variability may be the unpredictable effects of interaural asymmetric adaptation on parameters of signal presentation. This can be explored by probing the effects of varying on-time and off-time of a train of tone bursts. Figure 3–3, for example, plots the effect of varying the “on-time” of a train of tonal pulses on thresholds for the two ears of a patient with extensive but asymmetric brain damage. There had been an infarction of the middle cerebral arteries of both temporal lobes. The damage was greater over the left temporal lobe. Off-time was always 2,000 ms. On the left ear, results were within normal limits. As on-time decreased from 1,000 ms to 20 ms, threshold increased from 20 to 35 dB SPL, but on the right ear, the ear opposite the side of greater temporal lobe damage, the threshold varied from 30 dB SPL to 85 dB SPL as on-time varied from 1,000 to 20 ms. On the left ear, as the duration of the sound decreased from 1,000 ms to 20 ms, the threshold SPL increased by about 15 dB, which is a normal threshold-versusduration function. Now try to imagine, in the opposite ear, a sound that adapts so quickly that you cannot hear it at all until the SPL reaches 85 dB. The difference is striking. One can hardly conceive of such rapid decline in response to shorter and shorter stimulus durations.

3.  AUDITORY PROCESSING DISORDER   61

Figure 3–3.  Effect of varying “on-time” of a train of pulses in a patient with asymmetric brain damage, greatest over the left temporal lobe. Results were normal in the left ear but abnormal on the right ear. Modified from Figure 29 in Jerger et al. (1969).

Now consider what happens if you shorten the silent interval between easily heard tone bursts. Figure 3–4 shows data from a patient with a cerebellar tumor pressing on the right auditory nerve. Here the off-time between successive tone bursts of 200-ms duration was systematically shortened from 500 to 20 ms. On the left ear, there was little effect of the temporal manipulation. But on the right ear, the threshold response required systematically more and more intensity as the off-time decreased below 200 ms. At an off-time of 40 ms, there was no response at the maximum output of the system, 110 dB SPL. This can only be interpreted as extraordinarily dramatic adaptation in which successive 200-ms tone bursts meld into a single continuous tone that adapts over

62  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 3–4.  Effect of shortening “off-time” in a train of pulses in a patient with a cerebellar tumor affecting the right eighth nerve. Results were normal on the left ear but abnormal on the right ear. Modified from Figure 5 in Jerger et al. (1966).

time to an astonishingly rapid degree. It is the psychophysical equivalent of a Type III Békésy audiogram. All of this effectively highlights why threshold responses in persons with retrocochlear disorders may be unstable and the extent to which such thresholds can vary unpredictably due to variation of minor, and seemingly unimportant, parameters of the test signal. So far, we have been examining auditory abnormalities in persons with known, well-documented, disorders of the eighth nerve, brainstem, and auditory temporal lobes. I hope you have noticed that there is nothing subtle about them. They have been large and reasonably obvious effects: grossly abnormal PI func-

3.  AUDITORY PROCESSING DISORDER   63

tions for sensitized speech materials, substantial loudness loss, and extremely abnormal on-time and off-time results.

The Other Side of the Coin What if there is no “hard” evidence of brain damage? Is there a role for “soft” evidence in clinical evaluation leading to the diagnosis of auditory processing disorder (APD)? Years ago, when I was working at Baylor, students, residents, fellow departmental faculty, and people from other departments sometimes asked, “What is APD?” I would reply (paraphrasing Bill Hardy of Johns Hopkins, circa 1955), “They can Hear, but they do not Aud well.” After 60 years of wrestling with the concept, I think that is still the best, most inclusive definition of APD. But, of course, it is never enough detail to satisfy the questioner. So, then I say, “Let me tell you about a patient I once had. I am confident that will answer your question fully.”

AN ACCOUNT OF AN INTERESTING PATIENT She was 18 years old, had just finished high school, and was entering the college of music in a nearby university in the fall. She came to see us because she said that she needed a hearing aid. She had always had trouble hearing her high school teachers and did not want to have the same thing happen with her college teachers. In particular, if there was any other noise in the classroom while the teacher was giving an assignment, she had missed important information. She also complained of hearing difficulties in group social situations, while watching TV, and in using the telephone. The telephone was satisfactory on the right ear, but “not as good” on the left ear. In brief, she was absolutely convinced that she was in desperate need of a hearing aid.

64  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Before exploring her auditory status, we thought it essential to rule out nonauditory problems that might explain her complaints. So we organized a comprehensive medical examination and an equally thorough neuropsychological examination.

Medical Report The following was excerpted from the report of her medical studies: “Physical examination revealed an alert, well developed young adult. Both external auditory canals and tympanic membranes were normal. There was no evidence of pathology in the middle ear clefts. Extraocular muscle movements were normal. Pupils were equally round and reactive to light and accommodation. Nasal cavity, oral cavity, pharynx and larynx were normal. Neck findings were negative. Cranial nerves III to XII were intact. Routine laboratory studies included a normal complete blood count with normal sedimentary rate. Serum T4 and T3 levels as well as T3 uptake level were normal. Urinalysis was negative. Five-hour glucose tolerance test was normal. A magnetic resonance image (MRI) of the brain and posterior fossa was performed with the administration of gadolinium. No abnormalities were observed in the brain or in the region of the VIII nerve complex.” In other words, no obvious extra-auditory physical explanation.

Neuropsychological Examination We enlisted the services of a neuropsychology colleague, Dr. Fran Pirozzolo, to carry out the neuropsychological exam. His summary was as follows: 1. Supranormal global intelligence 2. Supranormal speed of mental processing

3.  AUDITORY PROCESSING DISORDER   65

3. Supranormal visual-spatial organizational abilities 4. Supranormal cognitive abilities, including vocabulary and verbal similarities 5. Discrepancy between global intelligence and fund of knowledge (information) 6. The subject reports “missing” auditory information in school tasks, a condition that may account for the relative weakness in fund of information despite high global intelligence.

Audiological Examination With these medical and neuropsychological summaries in hand, we proceeded to administer a number of standard auditory measures. They included pure-tone audiometric levels, crossed and uncrossed stapedius muscle reflexes, PB-50 scores, and PI functions for SSI sentences in simultaneous competition. All of these results were within normal limits on both ears. The following 18 measures were also within normal limits on both ears: 1. Masking level difference 2. Tympanograms 3. Acoustic reflex thresholds 4. Perception of temporal order 5. Perception of voice onset time 6. Interaural intensity difference detection 7. Interaural temporal difference detection 8. Distortion-product otoacoustic emissions 9. Transiently-evoked otoacoustic emissions

66  AUDIOLOGICAL RESEARCH OVER SIX DECADES

10. Serial recall 11. Staggered spondee word test 12. Competing Environmental Sounds test 13. Absolute azimuth localization 14. Relative azimuth localization 15. Auditory brainstem response (ABR) 16. Auditory middle latency response (MLR) 17. Auditory long-latency vertex potential 18. Auditory long-latency event-related potential But the following five measures yielded clearly abnormal results on the left ear and on the left ear only. They were as follows: 1. Sound-field thresholds for 500-Hz tone, monosyllabic words, and SSI sentences, in ipsilateral versus contralateral speech competition 2. Sound-field speech understanding with contralateral speech competition 3. Cued listening task in sound field 4. Acoustic reflex amplitudes 5. Dichotic Sentence Identification (DSI) test By far, the most interesting results here are the DSI data summarized in Figure 3–5. Recognizing the many complicating issues relating to dichotic listening scores, we attempted to design a procedure that would yield unequivocal results if APD were actually this young lady’s problem. They are addressed from left to right

67

Figure 3–5.  Dichotic test results in a young woman with an auditory complaint. Note essentially equivalent left ear deficit across all dichotic conditions despite variation in cognitive demand across dichotic conditions. Modified from Figure 7 in Jerger et al. (1991).

68  AUDIOLOGICAL RESEARCH OVER SIX DECADES

across the figure. All bars represent the score for that ear and that condition based on the presentation of 20 sentences. Overall, we sought to control for possible explanations due to memory or attention deficits. The first issue was memory. Did she simply have substantial difficulty remembering things heard on either ear? At the far left of the figure are four bars representing scores in two monotic conditions, 20 items per ear and 40 items per ear (double sequential). In each condition, her task was simply to identify the sentence heard in the test ear. We began with 20 items per ear. Since she scored 100% on each ear in that condition, we doubled the number of items to 40 per ear. The two monotic scores were still 100%. Concluding that there was no obvious memory problem if sentences were presented in only one ear at a time, we moved on to dichotic testing. In each of the four dichotic conditions, scores were based on the presentation of 20 pairs of sentences. In the free report (sometimes called “divided report”) condition or mode, the listener was instructed to repeat everything heard after each pair. This “free report” condition is thought by some to be biased in favor of the right ear. Here it appeared as a 30% difference in favor of the right ear, the ear spontaneously recalled first. To counter the possibility of a right ear bias, we added a “directed report” condition in which the listener reported only the sentence heard in the right ear on 20 trials and heard only in the left ear on the other 20 trials. Instead of reducing the right ear advantage, the ear difference actually increased to 45%. Next we sought to evaluate how well our listener could rapidly switch attention from one ear to the other. We informed her that the to-be-reported ear might switch randomly from one trial to the next but that the to-be-attended pair would be precued before each trial. Results were essentially unchanged. The ear difference was 35%. In the postcued condition, the ear to be reported was signaled immediately following delivery of each sentence. Here she was required to attend to and remember both sentences until the designated ear-to-be reported was cued. We had

3.  AUDITORY PROCESSING DISORDER   69

expected that our various manipulations might produce substantial effects depending on whether we were making the listening task easier (precued) or more difficult (postcued), but there was little variation over the four dichotic conditions. This relatively constant ear asymmetry, despite variation in cognitive demands from one condition to the next, suggested an auditory-specific basis for the various performance deficits on the left ear, perhaps due to a problem in appropriately localizing test sentences to the left side of her auditory space. Well, I said at the outset of this chapter that I would give you an example of what I think APD is. In formal terms, it is when three conditions are met. First, the patient, parent, or teacher complains of a hearing problem; second, you can find no extraauditory explanation for the complaint; and third, there is objective evidence of a specifically auditory disorder. I believe that this young lady illustrates this series of findings: First, she complained of a hearing problem. She was convinced that she needed a hearing aid. Second, there was no obvious medical or neuropsychological explanation for her complaint. Third, there was ample evidence of a specific central auditory disorder, especially the consistent left ear disadvantage in dichotic listening tests despite manipulation of cognitive demand. I think that the three-step process breaks down when we do not spend enough time systematically investigating both the second and third factors. There is, I think, a danger in relying too heavily on one or two specifically auditory tests involving speech perception or gap detection. It has been said by some that dichotic listening is the key to understanding APD. We repeat it here. Tests go in and out of fashion, but dichotic listening tests, when properly controlled

70  AUDIOLOGICAL RESEARCH OVER SIX DECADES

and interpreted, will probably become our most useful tools in the evaluation of APD. Why this is so must remain speculative, but consider the following: The biological system for processing auditory input and converting it to speech output is based on an asymmetry in the brain. In most right-handed persons, specific areas in the left temporal lobe are specialized for this complicated task. Because of the way that the two ears are “wired” to the two hemispheres, the right ear input goes directly to the left hemisphere. It arrives at the language-processing center first. The left ear is always at a slight disadvantage. It must travel to the right hemisphere first, then find its way to the left hemisphere. The left ear input would be at an even greater disadvantage were it not for the corpus callosum, a bridge between the two hemispheres, that relays the left ear input from the right hemisphere to the language-processing center in the left hemisphere. But there is always a loss in time and efficiency. Dichotic listening tasks are so sensitive to this asymmetry in binaural processing that they readily demonstrate it in the case of otherwise normally hearing persons. Some call this the “right ear advantage.” But if you want to understand what is going on in the brain, it is better to think of it as a “left ear disadvantage.” Dichotic tests are, therefore, uniquely positioned to reveal any brain disorder upon which the mechanism for minimizing the built-in, left ear disadvantage depends (the corpus callosum). An example would be loss of the myelin insulation of neurons traversing the corpus callosum. Does this account for many cases of “APD”? I wonder.

RELATED READINGS Bocca, E., & Calearo, C. (1963). Central hearing processes. In J. Jerger (Ed.), Modern developments in audiology (pp. 337–370). Academic Press.

3.  AUDITORY PROCESSING DISORDER   71

Chmiel, R., & Jerger, J. (1996). Hearing aid use, central auditory disorder, and hearing handicap in elderly persons. Journal of the American Academy of Audiology, 7, 190–202. Fifer, R., Jerger, J., Berlin, C., Tobey, E., & Campbell, J. (1983). Development of a dichotic sentence identification test for hearing-impaired adults. Ear and Hearing, 4, 300–305. Head, H. (1926). Aphasia and kindred disorders of speech (Vol. I). Cambridge University Press. Jerger, J. (1960). Audiological manifestations of lesions in the auditory nervous system. Laryngoscope, 70, 417–425. Jerger, J., Jerger, S., Ainsworth, J., & Caram, P. (1966). Recovery of auditory function after surgical removal of cerebellar tumor. Journal of Speech and Hearing Disorders, 31, 377–382. Jerger, J., Johnson, K., Jerger, S., Coker, N., Pirozzolo, F., & Grey, L. (1991). Central auditory processing disorder: A case study. Journal of the American Academy of Audiology, 2, 36–54. Jerger, J., & Martin, J. (2006). Dichotic listening tests in the assessment of auditory processing disorders. Journal of Audiological Medicine, 21, 15–23. Jerger, J., Thibodeau, L., Martin, J., Mehta, J., Tillman, G., Greenwald, R., . . . Overson, G. (2002). Behavioral and electrophysiological evidence of auditory processing disorder: A twin study. Journal of the American Academy of Audiology, 13, 438–460. Jerger, J., Weikers, N., Sharbrough, F., & Jerger, S. (1969). Bilateral lesions of the temporal lobe: A case study. Acta Otolaryngology Supplement, 258, 1–51.

4 A Once-in-a-Lifetime Opportunity

73

74  AUDIOLOGICAL RESEARCH OVER SIX DECADES

In June 1968, Jim Endicott, an old friend from Northwestern days, called me from San Antonio. He was now Major James Endicott, chief of audiology at the Wilford Hall USAF Hospital, Lackland Air Force Base in San Antonio, Texas. Major Endicott said that he and his colleague, Alan Rost, had been testing an unusual patient. He was calling to ask whether I might be interested in doing further evaluation. He could arrange, he said, to have the patient, a United States Air Force (USAF) airman, transported to Houston and billeted there for 1 week, during which time he would be available for whatever testing I might want to carry out. At first, I wondered what could possibly be so interesting that it would warrant a week of testing time. But when he told me why he thought I would want to see this particular airman, I very nearly dropped the phone. A 21-year-old man had suffered two infarcts of his middle cerebral arteries, first on the left side of the brain and then, 7 months later, on the right side. The incidents affected first the auditory areas of the left temporal lobe, then the same areas on the right temporal lobe. The term “central deafness” immediately sprung to mind. There had been previous reports of bilateral temporal lobe damage in which severe loss of hearing, often described as “cortical deafness” or “central deafness,” was reported, but adequate documentation beyond anecdotal descriptions was usually sadly lacking. Indeed, Dr. Robert Goldstein, a prominent American audiologist in the latter half of the 20th century, who had studied such matters, asserted that there was not a single convincing case study of “central deafness” in which there was (1) evidence of premorbid normal hearing, (2) an unequivocally measured reduction in sensitivity with (3) a clearly demonstrable central nervous system (CNS) lesion, and (4) no demonstrable peripheral lesion that could conceivably account for the auditory findings. It seemed to me that this case might just be the first to meet all of Goldstein’s criteria. Even more interesting, moreover, would be the opportunity to study, for the first time, the psychoacoustic profile of lesions of both temporal lobes. Ours was

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   75

not the first patient to suffer bilateral temporal lobe damage, but I believe it was the first case to have the effects on auditory function studied in such detail.

AN IMPORTANT POINT Before proceeding, I need to make an important point. Dr. Goldstein was quite correct that, in the case of bilateral temporal lobe damage, the premorbid condition of the patient’s hearing had been almost always unknown, beyond vague statements that the patient had never reported a hearing problem, but there could have been an existing hearing loss before hospital admission. Without premorbid audiometry, there was no way of knowing whether a postmorbid loss had already existed before the brain injury. Thus, the fact of a measured hearing loss after the incident or incidents was usually ambiguous. That is just one of the reasons why this patient was so important. Thanks to the policies and procedures of the USAF, pure-tone audiometry is routinely carried out on every new recruit. Figure 4–1 shows the preinduction audiometric levels of our airman in 1965, three years before the infarcts. It shows that, between 500 and 4000 Hz, hearing sensitivity was well within the normal range, satisfying Goldstein’s

Figure 4–1.  USAF preinduction air conduction audiometric levels in 1965. Between 500 and 4000 Hz, audiometric thresholds for pure tones were well within normal limits on both ears. There was no evidence of preexisting hearing loss. Modified from Figure 4 in Jerger et al. (1969).

76  AUDIOLOGICAL RESEARCH OVER SIX DECADES

requirement that there was no preexisting hearing loss. This is a crucial point in relating hearing loss occurring after the brain incidents to the notion of central deafness. Here we have documented that no preexisting hearing problem could account for the postmorbid problems.

HISTORY AND MEDICAL FINDINGS The present history and medical findings were gathered by coauthors, neurologists Norbert J. Weikers and Frank W. Sharbrough III, Neurology Service, Department of Medicine, Wilford Hall USAF Hospital, Aerospace Medical Division (AFSC), Lackland Air Force Base, Texas. They collaborated on the preparation of the Acta Otolaryngologica Supplement in 1969 on which this account is based. I will always be grateful for their excellent contribution to this work. Our 21-year-old airman was left-eyed, right-footed, and ambidextrous. He used the left hand for writing and the right hand for eating and bowling. Parsing this information to reach a conclusion relative to hemispheric dominance for speech and language can only be described as challenging.

FIRST ADMISSION On the day of first admission to Wilford Hall USAF Hospital, our airman awoke, had breakfast, and then noticed the abrupt onset of a headache, numbness in the right hand, and difficulty producing and understanding speech. He was thought to have a transient mild expressive and receptive aphasia. Hearing testing, including The Whispered Voice test, and three tuning fork tests, Schwabach, Weber, and Rinne, were administered by an otolaryngologist at Wilford Hall Hospital. They were reported as indicating normal hearing, although a

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   77

pure-tone audiogram was, (sigh), not carried out at that time. Thirteen days after admission, a percutaneous left carotid arteriogram showed nonfilling of the angular and posterior parietal branches of the left middle cerebral artery during the arterial phase and retrograde filling of these branches during the late capillary and early venous phases. Branches of this artery overlie the superior temporal gyrus of the temporal lobe on each hemisphere. Two months after admission, the airman was discharged. He felt that he was completely back to normal. There was no evidence of any hearing difficulty.

SECOND ADMISSION The airman was readmitted to the hospital approximately 5 months later. On the night prior to admission, he had set his alarm clock and retired. The following morning, he awoke 30 minutes after his alarm clock had buzzed. He was immediately aware that the world around him was strangely silent. He could not hear the sound of running water from the faucet, the buzz of his electric shaver, or the sound of traffic on the street. He could see another airman outside his window, pushing a powered lawn mower, but he reported that he could hear nothing. When others spoke to him, he could understand nothing they said. Six weeks after admission, a percutaneous right carotid arteriogram exhibited nonfilling of the distal portions of the right middle cerebral artery with retrograde filling of these vessels during the capillary and early venous phases. In short, both the right and left temporal lobes were damaged. Hearing gradually improved over the second admission. During the next 6 weeks, Endicott and Rost completed three audiograms that they felt were reasonably reliable. The three audiograms obtained during this critical 6-week period suggested remarkable return of function. There was substantial recovery at every test frequency. The data are summarized in Figure 4–2.

78  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 4–2.  Recovery of pure-tone thresholds after second admission audiograms made at the audiology service, Wilford Hall USAF Hospital, Aerospace Medical Division, Lackland Air Force Base, Texas. Modified from Figure 4 in Jerger et al. (1969).

At the first test date, on the day of the second admission, the pure-tone average loss (PTA, average of HTLs at 500, 1000, and 2000 Hz) was 70 dB on both ears. Less than 3 weeks later, PTAs were 20 dB on the right ear and 22 dB on the left ear. Six weeks still later, PTAs were 3 dB on the right ear and 10 dB on the left ear. We can only describe this as remarkable recovery from a world “strangely silent” only 3 months earlier. It is likely, however, that the mild low-frequency losses at 250 and 500 Hz are permanent. Does all this satisfy Goldstein’s four criteria for central deafness? We think so. First, the normal result on the airman’s preinduction hearing test satisfies the requirement of premorbid normal hearing. Second, the measured 70- to 80-dB loss across the frequency range from 250 to 4000 Hz, plus the graphic description that he could hear nothing, is, we think, convincing evidence of a temporarily severe loss in sensitivity. Third, arteriograms provided ample evidence of damage to the auditory areas of the brain in both hemispheres. Fourth, although there was apparent

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   79

residual high-frequency loss at 4000 Hz on both ears, sensorineural loss of 70 dB across the range from 250 to 2000 Hz due to apparent cochlear damage is not likely to return to normal levels in the 500- to 2000-Hz range, as occurred in this airman.

Basic Testing All of our testing in Houston was carried out in a single week early in June 1968 at the Houston Speech and Hearing Center, then an affiliate of the Baylor College of Medicine, and located in the Texas Medical Center, in Houston, Texas. We had at our disposal sound-treated test rooms, specialized electronic test instrumentation, and an anechoic chamber for localization studies in a free sound field. We are grateful, also, to Dr. A. C. Coats, also at the Baylor College of Medicine, for carrying out and interpreting a complete electronystagmographic examination. These results were all within normal limits. It is important to note that every test administered to our airman was also administered to a control listener, an age-matched young adult college student who was paid for his participation in the project. These results are available in the original monograph. We never interpreted our airman’s data as abnormal until we had compared them to the performance of this control listener.

Basic Audiometry Our air conduction audiogram in Houston was in close agreement with the final audiogram summarized in Figure 4–2 (“Two months later”) from San Antonio. It showed the same lowfrequency loss at 250 Hz, the same PTAs across the 500- to 2000Hz range, and the same high-frequency losses on the two ears at 4000 Hz. The audiograms now appeared to have stabilized.

80  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Speech audiometry testing included Phonemically Balanced Words (PB-50) and Synthetic Sentence Identification (SSI) test materials, but without the customary competing speech message of the traditional SSI test. In this case, we thought that the competing speech message was superfluous. The actual SSI Max scores without competition, 36% in the right ear and 100% in the left ear, show the wisdom of our decision. In both the PB and SSI tests, complete performance-versus-intensity (PI) functions were generated in order to define the Max score for each function. Results were as follows: Right ear:  PB Max = 0%; SSI Max = 36% Left ear:  PB Max = 38%; SSI Max = 100% The two SSI Max scores, 36% on the right ear, and 100% on the left ear, reflect the consistently poorer performance on the right ear. The SSI task, without simultaneous speech competition, is relatively easy. You have a list of 10 sentences in front of you. You hear one of these sentences and you must indicate which sentence you heard. It is difficult to score less than 100% at an easily audible level in quiet unless you have a really serious problem processing ongoing speech. Thus, the 36% SSI Max score on the right ear reflects just that — a serious problem processing speech presented to the right ear. I believe that in the 60 years I was involved in supervising the testing of auditory function in patients, I never saw an SSI score as low as 36% when there was no competing speech and the sentences were presented at an easily audible level, unless there was a severe loss in hearing sensitivity. The word “central deafness” is, I think, inappropriate here. To be sure, there was evidence of relatively severe temporary “deafness,” but that resolved in short order. I suggest that “central speech processing disorder” would be a more accurate description of the long-term consequences of these two vascular events. Another puzzling aspect of these speech scores is the substantial difference between ears: 38% for PB words and 64% for SSI

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   81

sentences. If there were a difference in the amount of residual loss at 4000 Hz (40 dB), it certainly would not be sufficient to account for the 64% difference in SSI Max scores, where low-frequency sensitivity is as important as high-frequency sensitivity.

Some Psychoacoustic Measures We measured differential sensitivity to intensity change using the same quantal psychophysical method described in Chapter 1. At a comfortable listening level, the difference limen, measured at 50% correct on the psychometric function, was 2 dB on the left ear but 6 dB on the right ear. This interaural performance difference was a recurring pattern across all of the auditory measures we administered, except one. On one measure alone, there was no difference between ears. That was the judgment of temporal order. To measure temporal order, we adapted a procedure previously used by Hirsh and Sherrick in their pioneering study of the phenomenon. We presented short-tone bursts of two different frequencies simultaneously to the test ear. The two tone bursts were widely separated in frequency, 500 Hz and 2000 Hz. There was always a difference in the onset of the two tones, but they ended simultaneously. We instructed the airman to judge which tone, the low-pitched or the high-pitched, came on first. The duration of the longer tone was always 1,000 ms; the duration of the shorter tone was varied by a method of limits technique to arrive at a temporal order threshold. The result for the airman was 250 ms on both ears. Our control subject required only a 20-ms difference on each ear, a value consistent with the temporal order literature. We were not surprised by the much longer onset difference required by our airman, but we were surprised, at first, that there was no ear difference, in view of the relatively large interaural asymmetries observed for the speech audiometric measures and the differential intensity sensitivity scores. This prompted us to consult

82  AUDIOLOGICAL RESEARCH OVER SIX DECADES

the literature again. Here we discovered that, in the case of brain damage, interaural symmetry was usually the case in temporal order judgments, despite asymmetric damage to the auditory areas in the temporal lobes. Some have speculated that temporal order judgments are mediated at an extra-auditory site at the cortical level. And it was certainly the case that our subject did not lack for brain damage outside the classical auditory areas.

Auditory Localization in Space Loss in the ability to localize sounds in space has often been cited in descriptions of auditory deficits due to temporal lobe damage. Suspecting that our airman might have similar difficulties, we devised a system to gather data on his localization abilities for several types of sounds, including pure-tone bursts at different frequencies, clicks, and noises (Figure 4–3). In the center of an anechoic chamber, we mounted a loudspeaker on a boom and pivoted it over the listener’s head so that it could travel half the perimeter of a circle 16 feet in diameter, that is, from 90 degrees to the listener’s right through 0 degrees (directly ahead of the listener) to 90 degrees to the listener’s left, marked off in 15-degree segments. We blindfolded the airman, told him to listen for sounds that lasted 10 seconds, provided him with a 3-foot pointer, and instructed him to point to each sound as it was presented from the loudspeaker. Then we presented 10-second bursts of pure tones, noises, or clicks in a position randomly chosen at one of 13 possible points on the 8-foot arc (i.e., one of 13 positions spaced 15 degrees apart). After each stimulus presentation, we translated the pointer position to degrees azimuth and recorded it for later analysis. Figure 4–4 compares accuracy scores for the two listeners. In the case of perfect performance, all data points would lie along the diagonal lines. The control listener did fairly well on the tasks, but the airman seldom even came close to the diagonal, except

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   83

Spe ake r

8 F t

0

90 L Bo om

90 R Pivot

Figure 4–3.  Apparatus for the measurement of sound localization. In an anechoic chamber, we pivoted a 16-foot boom over the listener’s head. A single loudspeaker, mounted at one end of the boom, could be moved over a semicircle, from 90 degrees to the right of 0 degrees azimuth to 90 degrees to the left of 0 degrees azimuth. Ten-second tone bursts of both tones and noises were presented at randomly chosen positions separated by 15 degrees of arc. Modified from Figure 19 in Jerger et al. (1969).

at 0 and 90 degrees azimuth, that is, straight ahead or directly to the right or to the left. We repeated the same procedure with pure tones of 400 Hz, 2000 Hz, 4000 Hz, and sawtooth noise. Results were pretty much the same except at 4000 Hz, where neither listener did well. We were not really surprised by any of this; impaired localization of sounds in space has been a common observation in temporal lobe damage for many years. We were

Figure 4–4.  Sound localization accuracy scores; comparison with an age- and sex-matched young adult with normal hearing. Errors are represented as deviations from the diagonal lines. Signals were 10-second, 1000-Hz tone bursts. Modified from Figure 20 in Jerger et al. (1969).



84

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   85

surprised, however, by a query that the airman posed at the end of the test session. “Do you want me to tell you where the sound starts or where it ends?” We were dumbfounded! Further questioning established that all of the sounds, except the clicks, moved in space over the 10-second sound duration, almost always from left to right. We were frankly astonished. We are always prepared to expect the unexpected in cases of brain injury, but the idea that a sound emanating from a fixed location in space could be heard as moving from its initial position to a different final position was certainly not expected. As it turned out, the answer was disarmingly simple. There is a mechanism in audition that might be termed “critical on time.” For very short sounds, there is a trade-off between threshold and duration. At 1000 Hz, for example, over the first 1,000 ms of sound duration, the threshold gradually improves by about 20 dB in a normal system. If you substitute “loudness” for “threshold,” you will begin to see where I am going with this. The answer to the apparent mystery of why a stationary sound can seem to move in space is illustrated in Figure 4–5. You will remember this figure from Chapter 3, where it appeared, for a different purpose, as an example of how critical on-time can be affected by temporal lobe disorder. In persons with normal auditory systems, the trade-off between threshold-versus-duration functions for the two ears will be similar. As a sound at 0 degrees azimuth progresses over the first 1,000 ms, loudness will rise equally in the two ears; the 10-second tone burst will be perceived as either directly ahead or directly behind the listener. This provides an important clue to localization. As the sound source moves away from the midline and toward the right sound field, a “head shadow” effect will attenuate slightly the loudness at the left ear, shifting the apparent location toward right auditory space. Moving the sound source away from the midline toward left auditory space will move the apparent source of the sound toward the left space as

86  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 4–5.  Effect of varying “on-time” of a train of tone bursts. Results were normal on the left ear but abnormal on the right ear. Modified from Figure 29B in Jerger et al. (1969).

well. In both cases, the head shadow effect creates a clue to the location of the sound. But suppose that loudness increases differently on the two ears over the first 1,000 ms. This is precisely what Figure 4–5 demonstrates in our airman. At a duration of 20 ms, the difference between thresholds (i.e., loudness) was an astonishing 50 dB. Then, as duration increased, the difference between thresholds gradually decreased to only about 6 dB. Is it any wonder that, for him, the sound seemed to move from left to right? From this experience, we learned, or I should say “relearned,” lessons from long-forgotten lectures. You are not finished testing until you have talked to the patient about his or her reaction to the procedures. In the military, they used to call it “debriefing.” Without it, we would have undoubtedly missed an extremely

4.  A ONCE-IN-A-LIFETIME OPPORTUNITY   87

important dimension of bilateral temporal lobe damage, one possible explanation for the extreme variability that many investigators have noted when attempting to carry out routine audiometry with these patients.

FINAL THOUGHTS We have retold the story of a 21-year-old airman in the USAF whose preinduction physical examination included audiograms showing normal hearing levels over the frequency range from 500 to 4000 Hz on both ears. Three years later, he sustained damage to the middle cerebral artery in both hemispheres of the brain, first on the left side and then 5 months later on the right side. The first event had little lasting effect with the exception of mild transient receptive aphasia, but the second event produced an immediate, incontestable, severe bilateral loss in auditory sensitivity. An audiogram at this point in time revealed audiometric levels at 70 dB across the frequency range from 250 to 4000 Hz on both ears. These results speak to a long-standing controversy. Is there such a thing as “central deafness” or at least severe hearing sensitivity loss associated with damage to one or more temporal lobes? Our results with this airman certainly confirm that, yes indeed, in the case of involvement of both temporal lobes, there was clear evidence of an associated severe loss of sensitivity to pure tones, in the region of 70 dB HTL. This loss gradually recovered to nearnormal levels from 1000 to 4000 Hz, with the exception of apparently permanent loss in the right ear at 4000 Hz. If you want to call this “temporary central deafness,” you certainly can. It meets our traditional notion of “deafness” for pure tones, but you would be missing two other dimensions of hearing that, in the case of our airman, may never recover to premorbid status. First, there was a continuing problem in processing the fine grain of speech. Two-way communication with the airman

88  AUDIOLOGICAL RESEARCH OVER SIX DECADES

was possible but labored. Sometimes he seemed to understand the gist of what we asked, but at other times, understanding was not so clear. Even on standard PB lists, he did not do well. Second, he will probably always have substantial difficulty with sound localization. This aspect of audition has been severely damaged. A revised concept of “central deafness” might read something like this: 1. Evidence of premorbid normal hearing sensitivity, normal speech understanding, and normal ability to localize sounds 2. Severe loss in pure-tone sensitivity, which may, or may not, return to near-normal levels 3. Severe, perhaps permanent, impairment in the ability to understand speech 4. Severe, perhaps permanent, difficulty in the ability to localize the sources of sounds in space 5. Do not expect to see this overall picture unless there has been damage to both temporal lobes. Well, that is the story of our “once-in-a-lifetime opportunity.” It needn’t take a week of testing to understand the problems that such a patient faces, but it does take more than an audiogram.

RELATED READING Jerger, J., Weikers, N., Sharbrough, F., III, & Jerger, S. (1969). Bilateral lesions of the temporal lobe: A case study. Acta Otolaryngologica Supplement, 258, 5–51.

5 Binaural Hearing Aids

89

90  AUDIOLOGICAL RESEARCH OVER SIX DECADES

The famous Italian scientist and astronomer, Galileo Galilei, was widely disliked by church authorities in 17th-century Florence because he refused to recant his support of the Copernican thesis that the Earth rotates around the sun. While my own situation hardly rivals Galileo’s in import, I have been widely disliked by many of my colleagues because I have refused to recant my position that not all binaurally hearing-impaired persons will benefit from binaural hearing aids. My experience in this arena began when binaural aids first became available in the 1950s. There was a good deal of hype at the time from various quarters that wearing two hearing aids would revolutionize the hearing aid experience. There were certainly sound theoretical reasons to expect this. Many colleagues repeatedly pointed them out, especially directionality and the head shadow effect. It was tacitly assumed that all hearing aid users would revel in the new world of amplified binaural sound.

AN EARLY STUDY It was within this zeitgeist that we approached our first serious attempt to study binaural aids at Northwestern. The result was a paper entitled “Binaural Hearing Aids and Speech Intelligibility,” published in the Journal of Speech and Hearing Research in 1961 with coauthors Raymond Carhart and Donald Dirks. We set up two loudspeakers in a sound room, one at a 45-degree angle to the left of the listener’s 0 degrees azimuth, the other to the right of 0 degrees azimuth. The two loudspeakers were separated by 90 degrees at a distance of about 4 feet from the listener. There were two speech understanding tests. In the first, Test 1, PB words were delivered from one loudspeaker, while simultaneously, Bell Lab sentences were presented from the other loudspeaker. In the second, Test 2, 30 questions and commands from PAL Auditory Test 8 (PAL-8), devised at the Harvard Psychoacoustic Lab dur-

5.  BINAURAL HEARING AIDS   91

ing World War II, were presented from one loudspeaker while, simultaneously, unrelated continuous discourse of two talkers was presented from the other loudspeaker. As the median scores will confirm, both tests were very difficult listening situations. We meant for them to be difficult. We purposely designed both listening tasks to require the listener to attend to the PB words, or the questions and commands from one direction while ignoring the competing speech intruding from the other direction. We reasoned that these listening tasks would be more likely to reflect the value of two-eared input than one-eared input. And we were right. We divided a total of 48 listeners into two groups on the basis of chronological age. Group I consisted of 24 persons below the age of 53 years; Group II included 24 persons above the age of 59 years. Thirty-two were long-time monaural users, 2 had worn binaural aids for 4 years, and the remaining 14 had never used aids at all. We analyzed the results as though they were two homogeneous groups, which, of course, they were not. In any event, we analyzed the data several different ways and concluded that, in view of excessive variability, there was no convincing statistical evidence that binaural fittings were any better for speech understanding than monaural fittings. And yet, informal examination of the median scores for monaural versus binaural aids, and for the young versus the elderly group, presented, in retrospect, an interesting picture. Figure 5–1 summarizes median scores for monaural and binaural conditions separately in the young and elderly groups. In the young group, repeating PB words (Test 1) was so difficult that the median monaural score was only 13% but improved to 27% for the binaural condition. In the elderly group, the median scores were 22% for the monaural condition and 27% for the binaural condition. Dealing with the questions and commands of the PAL-8 procedure (Test 2) proved to be an easier task in both groups, but the median score for the young group was still

92  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 5–1.  Median scores on two auditory tests in 24 young and 24 elderly hearing-impaired listeners. The comparison is between monaural and binaural fitting of the same aid. In both tests, two loudspeakers were arranged at a 45-degree angle from 0 degrees azimuth. In Test 1, one loudspeaker delivered a PB word while simultaneously the other loudspeaker delivered a sentence from the Bell Labs intelligibility lists. In Test 2, one loudspeaker delivered a question or command, taken from Harvard PAL-8 while simultaneously the other loudspeaker delivered the continuous discourse of two different talkers. Blue percentages are monaural scores. Red percentages are binaural scores. Modified from Tables 2 and 3 in Jerger et al. (1961).

only 60% in the monaural condition. It improved to 77% in the binaural condition. In the elderly group, however, there was no binaural improvement. Median scores were 65% in both amplification conditions. With only 48 subjects and considerable variability, we did not consider these results seriously at the time. But, in retrospect, they foreshadowed two conclusions that have since been more firmly demonstrated. First, when the listening task is difficult because of unrelated speech competition, there is an undoubted advantage to binaural aids in many, perhaps most,

5.  BINAURAL HEARING AIDS   93

hearing-impaired persons. Second, elderly people do not seem to get as much benefit as younger people. All of this happened 60 years ago with far less sophisticated hearing aids. Also, I think that we would have seen more dramatic results if the PB word test had been less difficult. In any event, many more carefully conducted group studies over the years have documented that the aided binaural score is, on average, better than the aided monaural score.

BUT IS THE GROUP REPRESENTATIVE OF EVERYONE IN IT? In those early days of binaural aids, we were sharply focused on group research. We did not consider that some persons might reap the benefits of the new technology, whereas others might not. Early research was dominated by designs in which large groups were tested with the ubiquitous PB lists on the assumption that if you compared results for one aid with results for two aids, the group with two aids would do better on average. Back then, I was as guilty as everyone else. Either there was wholesale improvement or there wasn’t. But in the early 1960s, when the golden age of binaural aids seemed to be collapsing, I began to have serious doubts that we were all drawing research subjects from the same normally distributed pool of listeners. We had been warned! As early as 1939, Vern Knudsen, a distinguished acoustician at UCLA, had observed that, in his experience, some hearing-impaired persons experienced the same sound as a different pitch in the two ears. He suggested that this might interfere with their successful use of binaural aids if binaural aids ever became available. It was many years in the future before any of us began to take seriously the idea that there might be two different groups of hearing aid users: those who did benefit from binaural aids and those who did not.

94  AUDIOLOGICAL RESEARCH OVER SIX DECADES

BINAURAL INTERFERENCE We were not the first to suggest that not all hearing aid users will necessarily enjoy a binaural advantage. But I think that we were among the first to present serious documentation of the phenomenon. Some people whom we had fitted in the Houston Methodist Hospital insisted that they heard better with just one aid. After we had encountered this complaint several times, we set out to research it. Meanwhile, the public relations person at our hospital had heard, from the chief of our service, about what we were doing. She came down to the lab to interview me. I told her that whereas we formerly fitted only one hearing aid per patient, we now had the capability to fit two aids, one to each ear. Many people, I emphasized, thrived on the change, reporting that this two-eared or “binaural” fitting was giving them a better result, but there were dissenters who insisted that they did better with one aid on just one ear, a “monaural fitting.” We were just gearing up to study why binaural fittings didn’t seem to help all hearing aid users. Our public relations person prepared a press release based on what I had told her and distributed the story to one of the national wire services. I promptly forgot all about this until I received a letter, weeks later, from a woman in North Carolina. I am working from a long-ago memory, but I think it went something like this: “God bless you, sir, please continue your important research. I too believe that I hear better with just one ear aided. But no one will believe me.” Well, that letter kept us on task even as some of our colleagues decried the idea that binaural aids were not the best solution for everyone. The slogan “Two ears — two aids” was often heard at conventions. Eventually, in collaboration with colleagues Shlomo Silman, of Brooklyn College, and Carol Silverman, of City University of New York, we were able to document the reality of what has

5.  BINAURAL HEARING AIDS   95

come to be called “binaural interference” in six different ways as follows: 1. Via tests of speech understanding 2. Via a cued monitoring task 3. Via the middle latency auditory evoked potential (MLR) 4. Via dichotic speech tests 5. Via a detectability index (d′) 6. Via the late positive component (LPC) of the P300 auditory event-related potential I will illustrate some of these with a brief account of just one patient (AK) who was tested fairly extensively in our laboratory at BCM in Houston in 1996.

A VERY INTACT NONAGENERIAN At the Baylor College of Medicine, Rose Chmiel and I collaborated on the evaluation of AK, a 90-year-old woman with a 13-year history of hearing loss following aspirin therapy for shoulder pain. She was active and in good general health. She showed no evidence of stroke, dementia, or neurologic or other systemic disease. She had been fitted binaurally by another center and came to us with the complaint that the aids were not satisfactory, especially in the presence of background noise. Pure-tone audiometry showed a bilateral low-frequency loss of about 45 to 50 dB, extending from 250 to 1000 Hz, then peaking at 2000 Hz at about 20 dB HTL, finally falling off to 40 to 50 dB HTL at 4000 Hz and 70 dB HTL at 8000 Hz. Distortion-product otoacoustic emissions were strong out to 10 kHz in both ears, arguing for reasonably healthy cochleae. Especially noteworthy,

96  AUDIOLOGICAL RESEARCH OVER SIX DECADES

in view of later interaural findings, was the fact that emissions were actually slightly stronger from the left ear over the frequency range from 4000 to 10,000 Hz. We began testing with routine aided SSI sentence recognition scores without competition. Results were as follows: Aided right:  100% Aided left:  30% Aided binaural:  60% Despite relatively similar pure-tone thresholds on the two ears, sentence recognition was decidedly poorer on the left ear. The effect on binaural performance was predictable: better than the left ear but poorer than the right ear, a classic demonstration of binaural interference. We pursued a plan designed to explore the basis for the interference. In elderly people, especially if they are in their 90s, the issue of cognitive decline inevitably arises. To what extent do flagging attention, immediate memory loss, and generalized mental decline play a role in the phenomenon of binaural interference? A useful initial approach is to explore dichotic listening. There are many dichotic listening tests, and they all work well. Within a framework in which two different auditory events are presented to the two ears simultaneously, one may explore the effects of various instructions designed to manipulate cognitive variables. In the case of listener AK, we used the Dichotic Sentence Identification (DSI) test. From a list of six synthetic sentences, two different sentences are randomly selected for presentation to the two ears. Information is gained by systematically changing the instructions to the listener. Figure 5–2 summarizes results for AK across the complete dichotic testing protocol. There were two initial monotic procedures, single and double sequential, designed to screen for a possible severe memory problem. In the monotic single condition (far-left side of figure),

97

Figure 5–2.  Monotic and dichotic SSI scores of patient AK, a 90-year-old woman, who had been fitted binaurally but insisted that she heard better with just one aid on the right ear. Monotic scores showed little difference between ears, but dichotic scores revealed a 70% to 90% ear difference. The extreme left ear disadvantage was unchanged as task difficulty was manipulated. Modified from Figure 3 in Chmiel et al. (1997).

98  AUDIOLOGICAL RESEARCH OVER SIX DECADES

a sentence was presented to only the right ear or to only the left ear. After the presentation of each sentence, AK was instructed to find the sentence on the list. She had no problem, scoring 100% correct for both ears. In the monotic double-sequential condition, two sentences were presented in sequence during each trial, in order to increase the memory load. The listener must now identify both sentences in the correct sequence after their presentation. Here the right ear score fell to 90% and the left ear score to 75%, not perfect but not unreasonable in a 90-year-old. In any event, a severe memory problem was not evident. In the first dichotic condition, free report, AK was instructed to find on her list both sentences and to identify them in any order, that is, the “free-report” condition. In this dichotic presentation, she correctly identified 90% of sentences presented to the right ear but only 15% of sentences presented to the left ear. In order to ensure that this was not a cognitive neglect or spatial localization issue involving left auditory space, we repeated the procedure in two separate blocks of sentences but instructed her to report on each trial only what was heard in one ear on each trial of the block. She reported only what was heard in the right ear throughout one block and the left ear throughout the other block. This was the directed-report condition. If the result of the earlier free recall condition were reflecting a cognitive issue with left auditory space, this new result would be an expected improvement in the left ear score, but it was actually slightly poorer, from 15% to 10%. In the next or “precued” condition, we instructed AK on which side to report before each trial. The side to be attended was then randomly varied from trial to trial. Here we anticipated that varying the ear-to-be attended trial-by-trial might make the task easier because of the immediacy of the instruction. But for AK, it made no difference. She continued to score 100% on the right ear and 10% on the left ear. Finally, we changed the instruction to postcueing, instructing her on which ear to report only after the two sentences had been presented. This, we thought, ought to make the

5.  BINAURAL HEARING AIDS   99

task more difficult, and indeed the right ear score dropped from 100% to 70%, but the left ear score remained at 10%.

Can We Explain AK’s Findings by Invoking a Relevant Cognitive Deficit? Not likely! There was a consistent performance difference between the two ears. First, there were the monaural PB scores (RE = 100%, LE = 30%); then there were the monotic double sequential scores of Figure 5–2 (RE = 90%, LE = 77%), the dichotic free-report condition (RE = 90%, LE = 15%), the dichotic directed-report condition (RE = 100%, LE =10%), the dichotic precued condition (RE = 100%, LE = 10%), and the very difficult postcued condition (RE = 70%, LE = 10%). Ordinarily, it would be difficult to imagine a cognitive deficit that would consistently affect only one of the two ears of a listener, especially an ear difference of such magnitude. Whenever we have the combination of reduced audiometric levels and poor speech understanding, the first explanation that comes to mind is damage to the exceedingly important end organ, the delicate cochlear mechanisms. And, indeed, a sensory loss is where you can expect to find the explanation most of the time but not all of the time. In the case of AK, otoacoustic emissions were not only robust in both ears but also actually stronger from the left ear. We must look more centrally in the auditory system to explain AK’s unique set of unilateral abnormalities.

A Possible Explanation This substantial, unwavering right ear advantage/left ear disadvantage is perhaps best explained by the structural model of Doreen Kimura. The basic idea is that the right ear input travels

100  AUDIOLOGICAL RESEARCH OVER SIX DECADES

directly to the language-processing center in the left hemisphere of the brain, but the left ear input is routed first to the right hemisphere, then to left hemisphere language-processing center via the corpus callosum. In normal young adults, this slight delay in left-sided input causes the well-known right ear advantage (and left ear disadvantage) in dichotic listening. Ordinarily, it is a small effect, hardly noticed in everyday life. But as the corpus callosum deteriorates with age, the left ear disadvantage becomes progressively greater. In the case of AK, the effect was striking. In single-eared listening, each ear worked reasonably well, but when both ears were stimulated dichotically, the left ear was at a severe disadvantage. But for successful binaural amplification, both ears must be working in synchrony. Is it any wonder that AK preferred the right ear–aided condition over the both ears–aided condition? This theoretical explanation holds only for verbal materials, analyzed thanks to specialization of the left hemisphere language-processing center. What about nonverbal materials like musical sounds and sounds whose meaning is related to pitch and spectrum? Here the evidence points toward an explanation favoring the right hemisphere as the key processor. In this situation, it is the right ear that suffers the disadvantage. Its input travels directly to the left hemisphere, then must cross over to the right hemisphere via the corpus callosum for further processing as a nonverbal event. If we were right that AK’s problem was in the corpus callosum, then comparing the right and left ears responding to nonverbal stimuli, in a dichotic paradigm, should produce a left ear advantage analogous to the right ear advantage for verbal events. We explored this idea by means of the late positive component (LPC) of the auditory event-related potential (AERP). (See Chapter 8 for more about the late positive component of the AERP.) The value of AERP comparisons here is that we could evoke an LPC response using the same speech materials to compare verbal

5.  BINAURAL HEARING AIDS   101

and nonverbal processing by simply changing the nature of the listening task. We fitted a 19-electrode array to AK’s scalp and proceeded to gather AERP waveforms. Materials were 300 trials of pairs of PB words presented dichotically. There were two conditions. In one condition, rare (oddball) targets were verbal, words that rhymed with the word “book.” The instruction to AK was to press a button whenever a target word was heard in either ear. All PB words had been digitally recorded by a single male talker. In the second condition, the same rare PB word targets were recorded by a female voice and substituted for the male target words in the original recording of all male target and nontarget words. The instruction was to press the button when a female talker was heard instead of a male talker in either ear. This altered the listening task from a phonemic linguistic target to a spectral, nonlinguistic target. These procedures yielded four LPC waveforms, both right ear and left ear LPC waveforms for the verbal task, as well as both right ear and left ear LPC waveforms for the nonverbal task. The LPC peaks were all in the latency region of 900 ms. Figure 5–3 shows the amplitude of the positive LPC peak of the waveform of the AERP for these four waveforms. In the case of a verbal target, the right ear peak was 3.2 microvolts, and the left ear peak was 0.8 microvolts, a strong right ear advantage. In the case of the nonverbal targets, the peak amplitude was 1.8 microvolts for the right ear but 3.4 microvolts for the left ear, a strong left ear advantage. These results are entirely predictable from the structural model of the central auditory system with (1) crossed inputs from ears to hemispheres, (2) specialization for linguistic processing in the left hemisphere, and (3) specialization for nonlinguistic processing in the right hemisphere. In summary, listener AK was typical of many elderly people: very old, still relatively intact cognitively, but the victim of an aging brain, a brain that was helped by compassionate counseling but not by binaural aids.

102  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 5–3.  Peak amplitudes of LPC auditory event-related potential waveforms for AK in two conditions: verbal targets and nonverbal targets. Verbal targets show a substantial left ear disadvantage, but nonverbal targets show just the opposite, an equally substantial right ear disadvantage.These results are predicted by the structural model of the central auditory system. Modified from Figure 8 in Chmiel et al. (1997).

RELATED READINGS Chmiel, R., Jerger, J., Murphy, E., Pirozzolo, F., & Tooley-Young, C. (1997). Unsuccessful use of binaural amplification by an elderly person. Journal of the American Academy of Audiology, 8, 1–10. Jerger, J., Carhart, R., & Dirks, D. (1961). Binaural hearing aids and speech intelligibility. Journal of Speech and Hearing Research, 4, 137–148. Jerger, J., Darling, R., & Florin, E. (1994). Efficacy of the cuedlistening task in the evaluation of binaural hearing aids. Journal of the American Academy of Audiology, 5, 279–285. Jerger, J., Silman, S., Lew, H., & Chmiel, R. (1993). Case studies in binaural interference: Converging evidence from behavioral

5.  BINAURAL HEARING AIDS   103

and electrophysiologic measures. Journal of the American Academy of Audiology, 4, 122–131. Jerger, J., & Silverman, C. (2018). Binaural interference: A guide for audiologists. Plural Publishing.

6 Cued Listening

105

106  AUDIOLOGICAL RESEARCH OVER SIX DECADES

If you want to demonstrate the advantage of a binaural hearing aid fitting, do not waste your time with PB 50 word lists. There is a better way to show not only the advantage when it exists but also the disadvantage when there is binaural interference. One of the principal auditory dimensions that binaural aids are expected to restore is directionality, the ability to localize the direction of a sound in space, a property of every two-eared system. A single hearing aid, whether worn on the head or on the body, cannot restore directionality. Does it not follow that we ought to be assessing directionality as well as speech understanding in any binaural system? In our lab at the Baylor College of Medicine in Houston, in the early 1990s, we thought so. Graduate student Craig Jordan and I devised a testing arrangement that we called “cued listening.” It was expressly designed to assess the ability of hearing aid users to appreciate the directionality of sounds in free space. This is how it worked. Figure 6–1 shows the physical arrangement of the 8-foot × 8-foot sound booth that we adapted for the study. A listener was seated in the exact center of the booth facing one corner. A loudspeaker faced the listener’s right ear (90 degrees azimuth) and another loudspeaker faced the left ear (270 degrees azimuth). A third (Babble) loudspeaker was mounted on the ceiling just behind the listener’s head. Immediately in front of and facing the listener was a pair of lights labeled L and R. These two lights cued the side to which the listener should attend. A recorded story about day-to-day events in the life of an ordinary person, written in the first person, was presented through both the right and left loudspeakers. Our story was a narrative about a private detective. (“I woke up early, took a shower and got ready to face the day. I looked for my gun but I could not find it. I worried about that, but breakfast called.”) The point of all this is that the story included many instances of the pronoun “I.” The listener’s task was to press a button whenever the word “I” was heard. As long as the red light was on, the listener had to monitor the speaker to the right side. When the

6.  CUED LISTENING   107

L SPKR

Sig 1.

L

na

l L

R

6 m

igh

ts

1. 6 m

Babble Speaker

R SPKR

Figure 6–1. Arrangement of loudspeakers and signal lights in the sound booth for the Cued Listening Test. See text for details of test procedure. Modified from Figure 1 in Jerger and Jordan (1992).

red light went off and the blue light came on, the listener had to switch attention to the left side. The same story was presented from both the right and left loudspeakers, but one recording had been purposely delayed by 60 seconds relative to the other recording, so that whenever the listener switched attention from one side to the other, he or she was always monitoring a different part of the story. The instruction was simple. “Listen to the story. Whenever you hear the word ‘I,’ press this button. When the red light is on only listen for ‘I’s from your right side. When the blue light comes on, only listen for ‘I’s from your left side.” For the duration of the testing, multitalker babble was played from the ceiling loudspeaker.

108  AUDIOLOGICAL RESEARCH OVER SIX DECADES

There were 10 listening periods during which targets, “I”s, embedded in the context of the narrative, were presented from the right side and 10 from the left side. Each listening period contained 10 “I”s. A response was scored as correct if the button push was within 1.5 seconds after an “I” occurred from the target side. If a button push occurred outside the 1.5-second response interval, that was scored as incorrect, even when the “I” was from the target side. If the response was to an “I” from the nontarget side, that was scored as a miss. Since there were 100 target “I”s from each side (10 intervals × 10 “I”s per interval), each correct response was worth 1% of the total number of “I”s for that ear. Listening periods were sequenced randomly with the constraints that there must be 10 intervals from each side and the same-sided interval may not appear in immediate sequence more than once per ear.

GROUP RESULTS As a first step in evaluating the procedure, we tested 10 experienced and successful binaural aid users. We reasoned that, if our test did not predict successful directionality of binaural aids in this group, we were wasting our time. The average age of the 10 listeners was 74.1 years, with a range from 48 to 88 years. Unaided PTAs averaged 51.1 dB for the right ear and 49.3 dB for the left ear. Months of binaural use averaged 32.6 with a range from 1 to 68 months. Figure 6–2 summarizes average Cued Listening Test (CLT) scores under four conditions: (1) unaided, (2) aided left, (3) aided right, and (4) aided binaural. It is interesting that, compared to the unaided condition, aiding the left ear failed to improve average group performance (40% in both conditions). Aiding the right ear improved performance by only 4%. But aiding binaurally raised average performance to an encouraging 64%.

6.  CUED LISTENING   109

Figure 6–2. Average Cued Listening Test (CLT) results for 10 experienced and successful users of binaural aids. Note that the binaural score is 20% better than the best monaural score (aided right ear). Across these 10 successful binaural users, individual advantages over the unaided condition ranged from 3% to 61%. All percentage scores are collapsed across side of target. Modified from Figure 5 in Jerger et al. (1994).

Across the 10 individual listeners, the improvement of the binaural condition over the best monaural condition ranged from 13% to 24%. Well, group results are always interesting, but clinicians want to know how any test works on the individual patient. In the following three case reports, we illustrate the range of performance profiles one is likely to encounter over a series of potential users of aids.

110  AUDIOLOGICAL RESEARCH OVER SIX DECADES

THREE ILLUSTRATIVE INDIVIDUAL PATIENTS Figure 6–3A shows CLT results in an extraordinarily successful user of binaural aids, a 70-year-old man with a moderate, bilateral sensorineural loss. The binaural score (71%) was 23% better than the best monaural score (right ear aided, 48%). Note also that the right ear aided score was, in turn, 21% better than the left ear aided score. This was the case throughout the group and individual data. The right ear aided score was usually better than the left ear aided score, often by a substantial margin. It seemed that binaural amplification accentuated the unaided right ear advantage.

Figure 6–3A.  Cued Listening Test results in a successful user of binaural aids, a 70-year-old man with a moderate bilateral sensorineural loss. Note the substantial 23% improvement in binaural aided score over right ear aided score and 41% improvement over the unaided condition. All percentage scores are collapsed across side of target. Modified from Figure 3 in Jerger et al. (1994).

6.  CUED LISTENING   111

In Figure 6–3B, you will see CLT results in another member of the group, a 78-year-old man with a moderate to severe bilateral sensorineural loss. Here results were not nearly so dramatic, but the binaural aided score (49%) was still 13% better than the best monaural score, 36%. And, in another bitter defeat for the left ear, the aided left ear score was 15% worse than no aid at all. Finally, Figure 6–3C summarizes CLT results in the case of a 68-year-old woman with a moderate bilateral sensorineural hearing loss above 500 Hz. She complained that hearing aids were not helpful in most listening situations. Little wonder! Every one of the three amplified scores was worse than the unaided score.

Figure 6–3B.  Cued Listening Test results in a less, but still successful, user of binaural aids, a 78-year-old man with moderate to severe bilateral sensorineural loss. The aided binaural score was only 10% better than the unaided score and 13% better than the best monaural aided score (aided right ear). All percentage scores are collapsed across side of target. Modified from Figure 4 in Jerger et al. (1994).

112  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 6–3C.  Cued Listening Test results in an unsuccessful user of binaural aids, a 68-year-old woman with moderate bilateral sensorineural loss above 500 Hz. The binaural aided score (20%) was 23% worse than the unaided condition (43%) and 11% worse than in the best monaural aided condition (left ear aided 31%). All percentage scores are collapsed across side of target. Modified from Figure 6 in Jerger and Jordan (1992).

Adding insult to injury, the poorest performance was the aided binaural score. I think you can see why Craig Jordan and I were so excited by this project with the Cued Listening Test. It seemed to be particularly useful in separating the satisfied from the unsatisfied users of amplification. In particular, it seemed sensitive to both the advantages and disadvantages of binaural amplification. To be sure, the Cued Listening Test is weighted heavily toward directionality. Yet the listener must monitor actual running speech in order to detect when a target “I” occurs in the

6.  CUED LISTENING   113

context of the narrative. It is certainly the case, however, that the fine grain of phonological analysis is not probed in detail. For that, the clinician must rely on one of the excellent speech intelligibility tests-in-noise already available.

LAST THOUGHTS When Craig and I embarked on this study, we were mainly concerned with finding a better way to demonstrate the value of binaural aids by focusing on directionality. As we gained experience with patients, however, we sensed that broader applications for the technique might exist. It was evident early on, for example, that, in elderly people, the right ear advantage/left ear disadvantage seemed to be accentuated along with everything else. Here, we think, we may have a robust method for assessing abnormally large left ear disadvantages in elderly people. Dichotic testing under earphones certainly reveals the abnormal left ear deficits but has its own set of difficulties. In the small number of patients that we have tested with the CLT, the effect of aiding the left ear alone has often proven to be disastrous. This not only impacts the difference between monaural versus binaural amplification but also suggests a unique method for evaluating auditory processing problems in general. It may aid in everyone’s search for a better understanding of the auditory processing problems described in Chapters 3 and 4 of this volume.

RELATED READINGS Jerger, J., Darling, R., & Florin, E. (1994). Efficacy of the cuedlistening task in the evaluation of binaural hearing aids. Journal of the American Academy of Audiology, 5, 279–285. Jerger, J., & Jordan, C. (1992). Age-related asymmetry on a cuedlistening task. Ear and Hearing, 13, 272–277.

7 Aging and Gender Effects

115

116  AUDIOLOGICAL RESEARCH OVER SIX DECADES

My first foray into the study of auditory aging was motivated by the concept of “phonemic regression,” first advanced by one of my mentors at Northwestern, John Gaeth. In his 1949 doctoral dissertation, “A Study of Phonemic Regression Associated With Hearing Loss,” Gaeth advanced the concept of a syndrome we now know as sensorineural loss, associated with inappropriately poor speech understanding, and often, but not always, observed in elderly persons. I learned a good deal about this study from Gaeth himself because in the early 1950s, we both lived in Glenview, a Chicago suburb a few miles west of Evanston, and sometimes shared the commute to Northwestern’s Evanston campus. In those days, the idea of speech audiometry was fairly new. Gaeth was plowing fresh ground, laying the foundation for the idea that speech audiometry, especially the word lists, had much to offer the budding field of audiology. Based on his dissertation data, Gaeth concluded that, in some elderly persons, word scores were substantially worse than you would expect from the degree of high-frequency hearing loss. Almost immediately it became a controversial issue among audiologists. Some agreed with the concept, but others argued that the decline in word scores with advancing age was entirely explained by their audiometric losses, especially at 2000 Hz and above. Time has worked in Gaeth’s favor. Over the years, evidence has accumulated that some elderly people do, indeed, have more trouble with speech understanding than you can explain from their audiograms. In the following sections, I address some of our research carried out in an attempt to explain the speech understanding problems experienced by so many elderly persons. Once you get into this very complicated business, you must juggle three factors simultaneously: (1) age-related changes in the pure-tone audiogram, (2) age-related changes in the central auditory system (APD), and (3) age-related changes in cognition. In the following sections, we review research evidence bearing on each of these three issues.

7.  AGING AND GENDER EFFECTS   117

CAN SPEECH UNDERSTANDING PROBLEMS IN ELDERLY PERSONS BE EXPLAINED BY THE AUDIOGRAM? To seek an answer to this question, we selected 137 files from a pool of 200 individuals who had participated in a previous study of aging. We selected them on the basis that they could be divided into four age groups, all with similar group-averaged pure-tone audiograms. In addition, speech test scores for SSI lists in the presence of continuous speech competition were available on each listener. The four age groups were: 50–65 years 66–70 years 71–75 years 76–90 years The left panel of Figure 7–1 shows the average audiogram for each group over the frequency range from 250 to 4000 Hz. The right panel of Figure 7–1 shows average scores for SSI sentences for each age group. We reasoned that, if speech understanding scores are explained solely by the audiometric loss in high frequencies, and since high-frequency loss was constant across the four age groups, then there should have been no decline in SSI scores across the four age groups. But as you can see in Figure 7–1, the average SSI score declined by more than 30% across the four age groups. Clearly, this systematic age-related change in speech understanding cannot be attributed to changes in highfrequency hearing sensitivity loss alone. In the next section, we consider the issues of APD, and agerelated cognitive changes, as explanatory factors in the decline of speech understanding in elderly persons.

Figure 7–1.  Left panel — Average audiograms for the right ears of 137 elderly persons divided into four age groups such that average audiometric levels are similar across all four groups. Right panel — Decline in average SSI scores across the four age groups. Decline in SSI scores across the 51to 91-year age cannot be attributed to differences in high-frequency sensitivity loss.



118

7.  AGING AND GENDER EFFECTS   119

AUDITORY PROCESSING DISORDER AND DICHOTIC LISTENING As people age, both their higher-level auditory processing abilities (APD) and their cognitive abilities tend to decline. Certainly, each plays a role in the decline of speech understanding with age. The question is, how do you tell the difference between these two key factors? Changes in age-related cognitive skills are more complicated than age-related hearing loss. The cognitive abilities of particular relevance to hearing, especially speech understanding in difficult listening environments, include memory, attention, and speed of mental processing. Auditory processing disorder (APD) problems, while not as complex as cognitive decline, have their own problems in separating the relative contributions of peripheral and central issues to decline in speech understanding. To compare the relative contributions of auditory processing disorder versus cognitive decline, we recruited, through newspaper advertisements, 130 elderly persons in the age range from 51 to 91 years to participate in a study involving APD tests and cognitive measures. All were paid to participate in the project. Each participant was subjected to extensive auditory testing and equally intensive neuropsychological evaluation. The auditory test battery included PB-SSI difference scores, Dichotic Sentence Identification (DSI) scores, and Speech Perception in Noise (SPIN) test scores. The subject was designated as APD if any one of those three test scores was positive. The neuropsychological test battery included the Minnesota Multiphasic Personality inventory, the Wechsler Adult Intelligence Scale, the Wechsler Memory Scale, the Buschke Selective Reminding Test, the Boston Naming Test, the Spatial Orientation Memory Test, the Simple Auditory Reaction Time Test, the Simple Visual Reaction Time Test, and the Four-Choice Visual Reaction Time Test. Neuropsychological tests were chosen and administered by our neuropsychology collaborator in the

120  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Department of Neurology at Baylor, Dr. Fran Pirozzolo. He selected, administered, and scored the battery of neuropsychology tests, then categorized each subject as either normal or abnormal cognitively. The auditory testing and interpretation were carried out without knowledge of the cognitive battery results. Similarly, all of the neuropsychology testing and interpretation were carried out without knowledge of the auditory results. Each participant had two separate outcome measures: (1) either normal or abnormal APD results and (2) either normal or abnormal cognitive results. Figure 7–2 summarizes the results of the study. The largest group, 36% of the total cohort, was normal on both sets of tests. This was not an unexpected finding. It is consistent with much previous research emphasizing the variability in the rate at which symptoms of aging appear in the elderly population. Interestingly, 14% of the total group were normal on the APD tests but abnormal on the cognitive tests. The fact that some participants did well on the APD tests despite abnormal cognitive results supports the view that the two outcomes are not inextricably interwoven. It is clearly possible to do well on APD tests despite cognitive decline. Similarly, 23%, almost one-fourth of the total sample, showed signs of APD despite normal cognitive function. Again, this further supports the view that there is, in the aging population, a meaningful entity, APD, that can exist without concomitant cognitive decline. Finally, 27% of the total group were abnormal on both test batteries. In this group, the aged participants were deficient on both APD and cognitive measures. In answer to the question partially motivating this study, can APD signs in elderly persons be explained as the inevitable result of cognitive decline, the answer is clearly no! In 23% of the total group, almost one in four persons, poor performance on APD tests could not be explained by cognitive decline.

7.  AGING AND GENDER EFFECTS   121

Figure 7–2.  Distribution of APD and cognitive test outcomes among elderly persons in the age range from 51 to 91 years. In 36% of the total group, results were normal in both test domains. They were normal on the APD test battery but abnormal on the cognitive battery in 14%, abnormal on APD but normal on cognitive in 23%, and abnormal in both domains in the remaining 27% of the total group. The fact that APD results were abnormal but cognitive results normal on 23% of the total group supports the view that APD can exist in elderly people independently of cognitive status. Modified from Figure 1 in Jerger et al. (1989).

The importance of this research is the fact that it demonstrated the potential independence of age-related changes in APD and in cognitive decline in two ways. In 14% of the group, cognitive findings were abnormal while, in the same people, APD findings were normal, and in 23% of the total group, APD findings were abnormal but cognitive status was normal. Clearly, decline in auditory processing ability and decline in cognitive abilities are not necessarily two sides of the same coin. Each can exist without the concomitant involvement of the other. Would it not, therefore, be helpful if we could easily differentiate between the two? That is the subject of the next section.

122  AUDIOLOGICAL RESEARCH OVER SIX DECADES

AUDITORY PROCESSING DISORDER VERSUS COGNITIVE DECLINE In Chapter 3, we talked at length about the importance of dichotic listening tests in the evaluation of APD. I think it is safe to say that many investigators regard dichotic tests as benchmarks of APD, so long as performance is not influenced by a complicating factor like cognitive decline. A large interaural difference on a dichotic test administered in the free-report mode is consistent with an APD problem, but it is also consistent with a cognitive problem associated with the second ear reported. In the early days of dichotic testing, the listener was instructed to report what was heard in each ear after each trial. If two different words were presented dichotically, the listener was told to report both words in any order. This is, of course, the free-report mode. It is fair to say that the free-report mode is a difficult, but not impossible, task for many elderly listeners. It does, however, impose a cognitive load in the form of attention to, and memory of, the second word to be reported. This cognitive load can be greatly attenuated by changing the instruction from “report both words” to “report only the word you hear on the right/left ear.” This, of course, is the directed-report mode. In theory, it substantially attenuates the cognitive stress encountered when the listener must first store both words in memory, then retrieve them serially. Because of this cognitive confound in the free report mode, many investigators have switched to testing dichotically only in the directedreport mode. But my colleagues and I had a different idea. We reasoned that, if you tested a person in both dichotic modes, free report and directed report, you might have a way of isolating any cognitive component involved in the free-report data. Here was our thinking. If there is a substantial interaural difference in the free-report mode, this result, by itself, is ambiguous. It could be a diagnostic sign of APD or it could be a cognitive problem of memory,

7.  AGING AND GENDER EFFECTS   123

storage, and retrieval of both words. But if you also test dichotically in the directed-report mode, where the cognitive demand is greatly attenuated, if the interaural difference declines, then the cognitive explanation is supported. On the other hand, if the score is unchanged in the directed-report mode, then the diagnosis of APD is supported. To explore this issue, we examined dichotic listening in both modes in 172 elderly persons. The Dichotic Sentence Identification (DSI) test was administered in both the free-report and directed-report modes. In the free-report mode, the listener was required to identify the sentence heard in each ear after each trial. But in the directed-report mode, the listener reported only the sentence heard in one ear according to pretest instruction. In both modes, there was a score for each ear. We reasoned that if there were a substantial difference between ear scores in the free-report mode but not in the directed-report mode, this could only be attributed to a cognitive deficit due to the difference in cognitive demand between the two modes. On the other hand, if the ear score difference was the same in both modes, then the problem could be reasonably ascribed to an auditory processing problem. If you have been at all into the dichotic literature, you will not be surprised that the poorer ear throughout these results was almost always the left ear. In any event, Figure 7–3 summarizes the results of the study. Data on the 172 elderly persons fell into three patterns. In Pattern I (both conditions normal), all four ear scores (free report, RE and LE; directed report, RE and LE) were within normal limits. In Pattern II, one ear score was below normal in the free-report mode but not in the directed-report mode. In Pattern III, one ear score was below normal in both the free- and directed-report modes. We reasoned that Pattern II represented cognitive deficit because the poor score in the more difficult free-report mode disappeared in the less difficult directed-report mode. We further reasoned that Pattern III represented a genuine auditory processing problem since scores were low on the poorer ear in both modes.

124  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 7–3.  Distribution of DSI test outcomes in 172 elderly persons in the age range from 50 to 91 years. The DSI test was administered in two modes, free report and directed report. Both modes were normal in 19% of the group. If there were an abnormal ear difference in the free-report condition but there was no significant ear difference in the directed-report condition, this outcome was classified as “cognitive,” since the ear difference was resolved by the less cognitively demanding directed-report outcome. But if there were a significant ear difference in both the free-report and directed-report conditions, this outcome was classified as “auditory processing disorder,” since the abnormal ear difference was consistent across modes. More than half of the group reflected the cognitive outcome. The remaining 23% showed the APD outcome, a significant ear difference in both report conditions.

Figure 7–3 shows that Pattern I (all conditions normal) was present in 19% of the 172-member cohort. Pattern II (one ear worse in free-report mode but not in directed-report mode) was present in 58% of the group. Finally, Pattern III (one ear worse in both free and directed modes) was present in only 23% of the total cohort. This result has always fascinated me because it sug-

7.  AGING AND GENDER EFFECTS   125

gests that more than half of elderly people with hearing loss have an associated cognitive deficit sufficient to appear on a relatively simple dichotic test. Additionally, 23% reflected a possible auditory processing problem. It is noteworthy that almost one in four elderly persons show a pattern of results suggesting APD. An interesting follow-up question is the extent to which either abnormal outcome, cognitive or APD, impacts the successful use of amplification in elderly persons. In a study of the effect of DSI test results in 63 elderly listeners, we found that listeners with a Type III DSI test result (APD) reported significantly less benefit from amplification than those with a Type I (normal) DSI result. But this is a broad question requiring a good deal more study. Interestingly, the patients’ significant others (mostly wives) reported more improvement on the Hearing Handicap Scale for the Elderly (HHIE) than did the study patients. In summary, our data suggested that approximately one in four elderly persons with peripheral hearing loss has an associated auditory processing disorder. In contrast, more than half of such a cohort suffers some degree of cognitive decline. It is an open question whether either or both groups might enjoy more benefit from an assistive listening device than from a conventional hearing aid.

A LONGITUDINAL CASE STUDY My long-time friend and colleague, Dr. Brad Stach, is now chief of audiology at the Henry Ford Hospital in Detroit, Michigan, but in the early 1980s, he directed the audiology service of the Methodist Hospital in Houston, Texas. One of his clinicians brought to Dr. Stach’s attention a 79-year-old patient who complained that his hearing aid, fitted to the right ear, was no longer working. But the clinician could find no problem with the aid. Searching further into the case, Dr. Stach discovered that the same individual had been evaluated by his service on three previous dates

126  AUDIOLOGICAL RESEARCH OVER SIX DECADES

in 1975, 1979, and 1980. On all of those occasions, he complained about difficulty in understanding speech in noisy backgrounds but functioned reasonably well in subdued noise environments. Now, however, he could understand little through the aid. He was convinced that the battery needed to be replaced. Dr. Stach immediately sensed a unique opportunity to study, longitudinally, the progression of difficulty in speech understanding over the 9 years for which records were available. The patient’s audiograms reflected the typical, expected gradually falling contour on both ears. PTAs were approximately 40 to 45 dB HTL on all of the four evaluations in 1970, 1975, 1976, and 1979. Over that same time frame, PB Max scores fell slightly from the high 90% levels to the high 70% levels, all expected minor age-related declines. But there was nothing in these numbers to explain the sudden seemingly total loss in speech understanding that led the patient to insist that the aid needed a new battery because it wasn’t working at all. As Dr. Stach plotted the SSI sentence identification scores, however, the answer became clear. Figure 7–4 plots these PI functions, for the right ear, over the four test dates as speech intensity level increased. In 1975, the SSI Max score was 80% and the degree of rollover was 20%. Four years later, in 1979, the SSI Max was still high at 90%, but rollover had increased to 45%. In 1980, the rollover was still 45%, but the SSI Max score had dropped to 70%. Throughout this time span, the patient’s assessment of the aid was that while it was “OK in quiet environments it was hard to understand speech in noise.” Finally, by 1984, the SSI Max had declined to 10%, and the aid was “no longer working.” Presumably, speech of any kind was now lost to him.

SOME GENDER DIFFERENCES In the same group of elderly persons reported earlier, we analyzed the right ear advantage (REA) data by gender. We formed

7.  AGING AND GENDER EFFECTS   127

Figure 7–4.  SSI PI functions in an elderly man who used a hearing aid on his right ear moderately successfully from 1975 through 1980 but by 1984 was convinced that it didn’t work at all. The sequence of events mirrored the decline in SSI Max. Modified from Figures 1, 2, and 3 in Stach et al. (1985).

a new grouping of 78 participants who were about equally divided between men and women. We noted an interesting gender difference in the REAs of the two groups. If the data were first collapsed across age group, then separated into men (n = 44) and women (n = 34), the gender difference was striking. Figure 7–5 shows that, for both free-report and directed-report conditions, the REA was substantially smaller in women than in men. The difference between the average percent-correct score for the right ear and the average percent-correct score for the left ear was almost 30% in the men but only slightly more than 10% in the women. Why this should be is a mystery. Perhaps the smaller average diameter of the female head means that the average length of the corpus callosum is shorter in women than in men, but that difference seems insufficient to account for the size of the REA differences reflected in Figure 7–4. We can only suggest that perhaps females age at a different rate than males. Hence, loss in

128  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 7–5.  A group of 84 elderly listeners was collapsed across age groups but separated by gender. Right ear advantage (REA) was defined as the difference between the right ear percent-correct score and the left ear presentcorrect score. In both the free-report and directed-report modes, the REA of men averaged almost 30%, but women only about 10%. Do males age more rapidly than females?

the efficiency of interhemispheric transfer perhaps proceeds at a slower pace in females.

ANOTHER GENDER EFFECT —  THE SHAPE OF THE AUDIOGRAM After many years of looking at audiograms I developed the feeling that there was a gender difference in the shape of the audiometric contour among elderly listeners. So, I assigned two graduate students to go through our files and calculate, separately by gender, the average audiogram for the right ear of 687 former patients

7.  AGING AND GENDER EFFECTS   129

with sensorineural loss who were over the age of 50 years. Of the total group, there were 341 males and 346 females. Figure 7–6 shows the result of that effort. Note that females had greater losses at 250 and 500 Hz than males, and males had greater losses at 3000 and 4000 Hz than females. This comparison is ambiguous, however, for two reasons. First, the total group of 687 files examined included both patients who were referred to us because they had an auditory complaint and volunteers who had no auditory complaint but had agreed to participate in various studies of auditory aging in our Baylor lab. Anticipating

Figure 7–6.  Average audiograms of 341 males and 346 females from the files of our Baylor lab. Females showed more loss than males at frequencies below 1000 Hz, and males showed greater loss than females above 2000 Hz. But male high-frequency data were contaminated by possibly greater noise exposure in the male group. Modified with permission from Figures 3, 4, 6, and 7 in Jerger et al. (1993).

130  AUDIOLOGICAL RESEARCH OVER SIX DECADES

that the complaint group was likely to have more hearing loss, especially in the high-frequency region above 2000 Hz, we eliminated all data from persons referred for an auditory complaint. The remainder, “no-complaint group” numbered 198. But a second source of concern was the likelihood that the male hearing levels at 3000 and 4000 Hz were contaminated by excessive noise exposure. So, we removed the data of all persons (male and female) in the remaining group of 198 records who had reported a history of exposure to high levels of noise, either in connection with their work environment or in pursuit of sporting or hunting activities. Figure 7–7 shows both the result of eliminating listeners with an auditory complaint (left panel, n = 198) and the result of further eliminating those with a history of excessive noise exposure (right panel, n = 123). Removing those who complained of a hearing loss (left panel) did little to change the results noted for the original group of 687 listeners. Indeed, the loss at 4000 Hz was slightly greater in the no-complaint group, but the original pattern persisted. In the right panel, however, the result of removing the data of listeners with a history of excessive noise exposure had little effect on the female levels but had a substantial effect on the male data, especially at 2000, 3000, and 4000 Hz. One question remains. Is there something about the U.S. that provokes results like this? Will the same pattern recur in less heavily industrialized countries? We found, in the literature, publications of average audiograms as a function of gender in four somewhat more pastoral countries: Sudan, Jamaica, Finland, and Scotland. Figure 7–8 shows average audiograms from each country. We did not see, in any of the four countries, the difference between male and females so evident in our highfrequency data. But we did see the same consistent tendency for sensitivity at 250 and 500 Hz to be poorer in females than in males, across all four countries. This suggests a more genderspecific cause than an exogenous explanation like more or less noise exposure. What that gender-specific factor might be must await further research.

131

Non-Noise Exposed Subgroup

Figure 7–7.  Data of Figure 7–6 after removal of males and females with an auditory complaint (no-complaint group) or with a history of excessive noise exposure (non-noise-exposed subgroup). Difference between genders in the region above 2000 Hz has been attenuated, but the overall pattern remains: more loss in females below 1000 Hz and more loss in males above 2000 Hz. Modified with permission from Figures 3, 4, 6, and 7 in Jerger et al. (1993).

No Complaint Group

132

Figure 7–8. Average audiograms from four countries less industrialized than the U.S. The highfrequency regions of the audiograms in all four countries reflect less noise exposure, but the greater loss in females in the lowfrequency region is evident in all four countries. Modified with permission from Figures 3, 4, 6, and 7 in Jerger et al. (1993).

7.  AGING AND GENDER EFFECTS   133

OVERVIEW In this brief glimpse of aging and gender effects, I think we made some useful contributions. We showed, I hope convincingly, that speech understanding declines with age, especially in one’s 70s and 80s, even as the pure-tone audiogram remains unchanged. We went on to show that one cannot blame everything on agerelated cognitive changes. Auditory processing, indexed best by dichotic listening tests, can be compromised even in the presence of normal cognitive function. Enigmatic gender differences are always intriguing. One thing I have learned from a lifetime of audiological research is that a gender difference in how we hear is the rule rather than the exception. Well, let us move on to auditory event-related potentials.

RELATED READINGS Jerger, J., Chmiel, R., Allen, J., & Wilson, A. (1994). Effects of age and gender on dichotic sentence identification. Ear and Hearing, 5, 274–286. Jerger, J., Chmiel, R., Stach, B., & Spretnjak, M. (1993). Gender affects audiometric shape in presbyacusis. Journal of the American Academy of Audiology, 4, 42–49. Jerger, J., Jerger, S., Oliver, T., & Pirozzolo, F. (1989). Speech understanding in the elderly. Ear and Hearing, 10, 79–89. Stach, B., Jerger, J., & Fleming, K. (1985). Central presbyacusis: A longitudinal case study. Ear and Hearing, 6, 304–306.

8 Auditory Event-Related Potentials to Words

135

136  AUDIOLOGICAL RESEARCH OVER SIX DECADES

The concept of a word recognition score as a tool for understanding hearing loss has had a long history in audiology. The procedure is disarmingly simple. A single-syllable consonantvowel-consonant (CVC) word, usually from a PB-50 list, is presented to the listener, who must then repeat back what was heard. The response is scored as correct or incorrect and the next word is presented. After 50 words have been scored, the total percent-correct score is calculated. The procedure was devised 76 years ago. The problem with this scheme is that the amount of mental effort expended to achieve the score is largely unknown. But as any hearing aid user will tell you — and tell you — and tell you — listening effort expended is a critical issue in the successful use of hearing aids. As a hearing aid user myself, I can attest to the stress and exhaustion I feel trying to understand what someone is saying in another room while the television blares away. You can repeat words back ad infinitum but that will seldom capture the essence of active, purposeful listening. In this chapter, we introduce the reader to some of our research in auditory event-related potentials (AERPs). We turned to AERPs in an effort to understand how a listener actually actively processes words. The long-term research plan here was to search for better ways to assess the processes involved in active rather than passive listening to a list of words. Just as the auditory brainstem response gave us a much better way to assess difficult-to-test patients, especially babies, I believe that the AERP response, recorded from electrodes attached to the scalp, will help us gain a better understanding of the various factors contributing to listening effort. To aid the reader, we must first introduce some basic evoked-potential concepts.

WAVEFORMS If you are an audiologist you have already encountered the notion of a waveform from your exposure to the auditory brain-

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   137

stem response (ABR). You affix the necessary vertex electrode and associated reference electrodes, instruct the listener to go to sleep, turn on the ABR recording device, deliver a train of 2,000 clicks to the earphone, and watch the waveform of the ABR build up on the computer screen. What you see on the recording apparatus is simply a visual display of how the electrical activation of the brain under the vertex electrode, sensed indirectly through the skull bone, varies as a function of time after click onset. The unit of measure is the microvolt. In the case of the ABR, the time frame is usually 1 to 18 milliseconds (ms), and there are five “bumps” or “waves,” imaginatively labeled I, II, III, IV, and V. Parenthetically, I once asked Don Jewett why he chose Roman numerals to label the five peaks. He said that he wanted to reserve Arabic numbers for the animal ABR. It was on animals that he first discovered these early electrophysiological responses from the brain (four peaks from animals; five peaks from humans). These five classic waves simply show how electrical activity changes over time after the onset of each click. In this chapter, we look at waveforms like the ABR, but with these differences: 1. The auditory stimulus is a consonant-vowel-consonant (CVC) word, not a click. 2. The time scale of interest is usually the latency region from 500 to 1,800 ms, not the first 18 ms after stimulus onset. 3. There are generally no more than three to four readily identifiable bumps, waves, or peaks. 4. Amplitude peaks and valleys vary with time after word onset. 5. More than one recording electrode is applied, usually 32 in our lab at UTDallas. The electrodes are spread out uniformly across the surface of the listener’s scalp from ear to ear and from nasion to inion.

138  AUDIOLOGICAL RESEARCH OVER SIX DECADES

6. As each word is presented, the listener is urged to attend closely to the word, reach a decision about that word, and make an appropriate motor response. The goal is to generate not an oral response from the listener but an auditory event-related potential (AERP), that is, an electric potential evoked by requiring that the listener make a decision after the event of hearing each word. For example, if the target category is names of animals and the listener hears the word “cow,” then the listener must decide whether “cow” is the name of an animal and, if so, press the YES button. It is helpful to visualize auditory evoked potentials as ranging across a continuum from very simple (ABR) to very complex (AERP), with all the others — for example, auditory middle latency response (AMLR) and auditory late response (ALR) — between these two extremes. Figure 8–1 compares the ABR with the AERP across a number of dimensions. It may prove helpful to refer back to this figure from time to time as you proceed through the chapter.

THE IMPORTANCE OF FORCING A DECISION The key idea here is that, if you want to study active listening, you must structure the listening task so that the brain is put to work. Once you get into this arena, you will discover that the possibilities for crafting appropriate listening tasks are limitless. One of our favorites has been a simple category judgment. It goes like this. Make recordings of four or five dozen CVC words chosen at random. Now make recordings of another dozen CVC words that are all in the same category, for example, names of animals (dog, cat, mouse, duck, and so forth). Insert the category words, the “targets” randomly within the list of the previously recorded, randomly selected, words (nontargets). Now instruct the listener as follows: “You are going to hear a list of short words. Some of

Figure 8–1.  Comparison between auditory brainstem response (ABR) and auditory event-related potential (AERP). Each represents one end of the continuum of evoked auditory potentials.



139

140  AUDIOLOGICAL RESEARCH OVER SIX DECADES

the words are the names of animals, but most are not. You see before you two buttons. One is labeled YES; the other is labeled NO. If the word you heard was the name of an animal, press the YES button. Otherwise press the NO button.” Thus, a decision is forced after each word, target or nontarget, has been presented. If you want to make it more difficult, you can add another category, for example, names of articles of clothing (hat, glove, shirt, coat, and so forth). Now instruct the listener to press the YES button if either the name of an animal or an article of clothing was heard. You can begin to see the possibilities here. In a typical test run in our lab, we presented 60 CVC words. Of the total number, 9 words (15%) were targets. The remaining 51 words were nontargets. The 9 target words were inserted among the nontarget words in random fashion, but with the constraint that no more than 2 target words might occur sequentially. These are variations on the classic “oddball” paradigm first explored years ago by the late Sam Sutton of the Psychophysiology Laboratory at the New York State Psychiatric Institute. In one of his typical procedures, two different frequencies of tone bursts, either 1000 Hz or 2000 Hz, were presented over a series of trials. The key difference between them was that the 1000-Hz tone burst was presented on most trials (i.e., “frequently”; 70%–90% of the trials) while the 2000-Hz tone burst was presented only rarely (10%–30% of the trials). The listeners’ task was to ignore the frequent, lower-pitched tone bursts (nontargets) but to respond positively to the rare, higher-pitched tone bursts (targets). It turned out that responses to the 1000-Hz frequent tones showed little deviation from baseline, but the relatively rare 2000-Hz tones generated a waveform characterized by a robust positive peak at a latency of about 300 ms after tone-burst onset. Not unexpectedly, this came to be called the “P300” response and was thought to be indexing the decision-making process. But the distinction between two tone bursts an octave apart is not very challenging for most listeners. Therefore, the decision-making process is not taxing. Subsequent research with more difficult listening proce-

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   141

dures revealed that as the degree of difficulty in making the decision increased, the latency of the positivity evoked by the target stimuli increased. It was considered more prudent, therefore, to avoid attaching a single number like 300 to the response. It is now called the “late positive component,” or LPC. In our lab at UTDallas, we frequently observed LPC latencies in excess of 800 ms, in the case of even the simplest word task. Variations on the oddball paradigm are legion. The unique contribution of our lab was simply to apply the oddball paradigm to the analysis of CVC words.

THE FRAMEWORK OF AN AUDITORY EVENTRELATED POTENTIAL PROCEDURE (AERP) The presentation of each word is considered a “trial.” The basic idea is to divide trials into two categories, “targets” and “nontargets.” Targets are words to which the listener should respond “YES.” Nontargets are words to which the listener should respond “NO.” If you make the number of “target” trials small in comparison with the number of “nontarget” trials, the average of target trials will reflect a robust positive response in comparison with the average on nontarget trials. This average of the target trials defines the late positive component or LPC.

THE LATE POSITIVE COMPONENT (LPC) We studied the LPC to words in two kinds of listeners: (1) young adults with normal hearing and (2) listeners of all ages with various disorders. Sometimes we averaged results over small groups, but more often we focused on individual patients. Figure 8–2 shows two examples of such LPC waveforms. Before proceeding, however, I must point out that in all of the waveforms you will see in this chapter, there are two consistent features characterizing the onset of each waveform. They are a sharp negative peak

142  AUDIOLOGICAL RESEARCH OVER SIX DECADES

(Talker Gender)

P2

N1

Figure 8–2.  Late positive components (LPCs) in a young adult with normal hearing, under two conditions: noise target (black curve) and spectral-feature target (female talker—red curve). Note reduced peak amplitude and longer peak latency of the LPC as the listening task becomes more difficult. Modified with permission from Figure 3.11 in Jerger et al. (2014).

at a latency (time after onset of the word) of about 100 ms and an equally sharp positive peak at a latency of about 200 ms. These are called the N1 and P2 peaks. They are always there when you examine the waveforms in response to word stimulation. The N1 peak signals the onset of the word; the P2 peak is less well understood. It may be that it indexes something about what sort of sound is coming up (e.g., music, noise, speech, and so forth). These two peaks are important components of all AERP waveforms. For current purposes, however, they hold little interest

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   143

except as time markers. The positive going peak following the N1 and P2 peaks is the late positive component, the LPC. Figure 8–2 shows two examples of such LPC waveforms. They were recorded from the same young adult listener under two conditions. The black line shows the LPC generated by interposing nine noise bursts (targets) among 51 randomly selected CVC words (nontargets). The listener was instructed to press the YES button if a noise burst was heard and to press the NO button if a word was heard. The LPC waveform (black curve) peaked at a latency of about 400 ms, reflecting a relatively easy task. The red curve is the same individual’s LPC response to a spectral target (talker gender), nine random CVC words spoken by a female rather than a male talker. It differs from the black curve in two ways, both a consequence of increasing the difficulty of the task: (1) The peak latency is much longer, closer to 700 ms, and (2) the amplitude is much smaller, dropping from almost 8 microvolts for the black curve to about 5 microvolts for the red curve. This is a dramatic example of how the LPC is affected by changing the task difficulty; the difference between detecting a burst of noise among words and detecting words with a different kind of spectral feature (talker gender) among other words is not subtle. In Figure 8–3, we see data from the same young adult, but now the red curve shows the effect of changing the target to a common phonemic characteristic, words that rhyme with “jet.” Again, the relatively less difficult noise-target task (blue waveform) generated an LPC peak at 400 ms, but, in the case of the more difficult phonemic task, the peak latency was closer to 800 ms and the amplitude much attenuated. These two figures illustrate how increasing task difficulty alters the LPC waveform in two ways: (1) longer peak latency and (2) lower peak amplitude. Both are consequences of requiring that a decision be made after each word or noise burst. Note also that the N1 and P2 peaks at 100 and 200 ms, respectively, are clearly evident.

144  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 8–3.  Late positive components (LPCs) in the same young adult as in Figure 8.2. Here, comparison is between a noise target and a phonemic feature target (words that rhyme with “jet”). Note, again, how the increase in task difficulty reduced peak amplitude and increased peak latency of the LPC. Modified with permission from Figure 2.3 in Jerger et al. (2014).

THE RIGHT EAR ADVANTAGE Can the well-known right ear advantage (REA) in dichotic listening be reflected in the LPC? Actually, it was surprisingly easy to demonstrate. We first digitally recorded, sequentially, two children’s fairy tales, Cinderella and Snow White & the Seven Dwarfs. We needed both because we wanted speech materials that listeners were familiar with, but we also needed enough material in which to embed nine target words that were either

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   145

semantically or syntactically inappropriate to the ongoing story. Dichotic stimulation was achieved by delaying the story to one ear by 60 seconds relative to its position on the other ear. The total listening experience was divided into two blocks of 60 trials each. In one block, the listener was instructed (precued) to report only target words heard in the right ear and to disregard any target words heard in the left ear. In the other block, the listener was precued to report only target words heard in the left ear and to disregard target words heard in the right ear. Reports from the wrong side, whether correct or incorrect, were disregarded. In effect, the listener heard the same target words on the two ears, but at different points in time, and responded only to the targets on the side to be attended in that block. Figure 8–4 shows the averaged LPC waveforms generated by targets from the two sides in 20 young adults with normal hearing. Note that these waveforms are much smoother than the waveforms of Figures 8–2 and 8–3. That is because they are averages over 20 listeners rather than waveforms from a single listener. These right and left ear LPCs differ in two respects. First, the latency of the left ear LPC is delayed slightly compared to the right ear LPC. Second, the maximum amplitude of the left ear LPC is slightly less than the maximum amplitude of the right ear LPC. These are the two key factors indicating that one decision is more difficult than another. It is, of course, possible that a listener might respond to a target word coming from the wrong side. We did not keep track of these but we should have, because they could be treated as “false alarms.” Pairing these with the “hits” (correct responses to a target on the precued side) would have permitted computation of a detectability index (d′) that would have reflected a more accurate analysis of the interaural contrasts. In any event, the sensitivity of the LPC waveforms from the two sides of the head was sufficient to document the right ear advantage electrophysiologically. In summary, in these experiments, the LPC of the AERP was the average of the individual responses to 9 target words, out of a

146  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 8–4.  Late positive component (LPC) waveforms for targets to right and left ears in a dichotic listening procedure. Data are averaged across 20 young adult listeners. Targets were morpho-syntactically deviant words that had been inserted in an ongoing fairytale. Note smaller peak amplitude and slightly delayed latency of the left ear LPC compared to the right ear LPC. Modified with permission from Figure 4.1 in Jerger et al. (2014).

total of 60 words presented to the listener. In the jargon of eventrelated potentials (ERPs), we say that the a priori probability of a target was 0.15 or 15%. The positivity peaked over the latency range from 400 to 800 ms. That is why it is no longer called the P300 response. The LPC is all about decision making. When the decision “Yes, I heard a target word” is relatively easy for the listener, the peak latency is short and the peak amplitude is large. But as the decision becomes more difficult, the latency increases and the peak amplitude decreases.

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   147

Note also that all of these LPC waveforms were recorded at the midline parietal electrode, PZ (see Chapter 9, Figure 9–1A to view the placements of electrodes over the surface of the scalp). This is the region of the scalp where the LPC response is typically at its maximum amplitude. It is in the posterior region of the scalp, just forward of the occipital electrode array. In the next section, we will be looking in a different region of the scalp.

BUT WHAT ABOUT THE NONTARGET WORDS? A number of years ago, I had occasion to give a talk in Sweden on this very topic, the averaging of nontarget words in an AERP context. I spent at least 5 minutes explaining the rationale and showing several examples. During the posttalk question period, a member of the audience asked, impatiently, why in the world anyone would average the nontargets in this paradigm. I resisted temptation and simply tried to explain the idea in simpler language. I can understand why the idea is, at first, counterintuitive, but hear me out. Nontarget words constitute 70% to 90% of trials in our AERP studies. In the past, their waveform has usually been either (1) subtracted from the LPC target waveform, changing very little, or (2) simply ignored. We would have ignored them as well, except that something about AERP word data troubled us from the very beginning of our research. In previous work with clicks and tones, we did not expect to see much activity from nontarget data, but as we examined our nontarget waveforms in response to words, they all seemed to fall into negative voltage territory immediately after the N1, P2 peaks. Indeed, sometimes the negativity even seemed to begin from word onset and to be carrying the N1 and P2 peaks down with them. This was so foreign to our previous experience with simpler stimuli, like clicks and tone bursts, that we spent a lot of lab time discussing

148  AUDIOLOGICAL RESEARCH OVER SIX DECADES

it and wondering what was going on. Why were the waveforms of nontarget words going negative? There were of course, previous AERP studies showing negativity to linguistic entities (for example, the N400 response to incongruous words at the ends of sentences), but this response concerned target, rather than nontarget, words. So, we started to look at the waveforms generated by nontarget words, beginning with simple category judgment experiments. Gradually, we came to the conclusion that we were looking at something unique, the processing of words that are not targets. When the target is a tone burst of a certain pitch, and the nontarget is a tone burst of a different pitch, you don’t spend a lot of time analyzing nontargets. You are listening for something else. But the unique feature about word nontargets is that each nontarget is a distinctly different word. In all of the conditions that we have studied, the nontargets have always been a list of randomly selected unrelated CVC words. Each nontarget word must be analyzed in its own right. In effect, each nontarget word is a mini-experiment in word processing, leading to the decision, “this is not a word I am supposed to say yes to.” Such a decision seems to invoke different listening mechanisms than those involved in the YES decision, mechanisms that evoke a sustained negative potential rather than an abrupt positive potential, as in the LPC response. Furthermore, the negativity is maximal, not in the midline parietal region but in the left frontal region of the scalp. That is why you will not see much of it in the parietal region where the LPC dominates. Figure 8–5 shows a typical waveform generated by nontarget words. After the initial N1 and P2 peaks, the waveform drops quickly into a negative wave almost as deep as the N1 component. It peaks at about 500 ms, then returns slowly to baseline at 1,800 ms. Here the negativity was maximal at a latency of approximately 500 ms and was largest at the left fronto-central electrode (FC-3).

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   149

Figure 8–5. An example of processing negativity in the average waveform of nontarget words. Note peak negativity at a latency of about 500 ms and prolonged return to baseline.

PROCESSING NEGATIVITY (PN) We have called this phenomenon “processing negativity,” or simply PN, because it seems to reflect unique aspects of word processing. Note, in Figure 8–5, that the N1 and P2 peaks are better formed than in the LPC examples above. That is because the averaged waveforms are based on 51 trials rather than 9 trials. At a latency of about 300 ms, the waveform breaks away from the negativity following the P2 peak and continues to decline, reaching a peak negativity in the 500- to 600-ms range, then slowly returns to baseline at 1,800 ms. Within the area encompassed by that roughly 1,000-ms latency range after the P2 peak, a number

150  AUDIOLOGICAL RESEARCH OVER SIX DECADES

of processing activities seem to be going on. We believe that they include attention, phonological analysis, and semantic analysis of the heard word. We should also note that the PN component characterizing the analysis of each nontarget evokes activity in a different region of the scalp than the LPC potential. The waveform of Figure 8–5 was in fact recorded at electrode FC-3, the fronto-central region of the scalp, just left of midline. Recall that the LPC component is located over the midline Pz electrode. One useful feature of processing negativity (PN) is that you can manipulate the PN by simply changing the instruction to the listener, or changing the nature of the task. Let me just run that by you again with added emphasis. You can change the PN response to nontarget words by altering nothing more than the instructions to the listener on the nature of the task relative to the target words. This remarkable effect on nontarget words permits a wide variety of applications. Consider, for example, Figure 8–6. It shows the average effect on the PN component when 19 young-adult normal-hearing listeners were tested in three conditions: (1) visual decoy, (2) nonword targets, and (3) semantic targets. In the visual decoy condition (green waveform), both the target and nontarget CVC words were presented in the usual fashion, but the listener’s attention was directed to a computer monitor placed directly before him. On each successive visual decoy presentation, four geometric figures were briefly presented in the four quadrants of the computer screen. On nontarget trials, either four circles or three circles and an ellipse were presented. On target trials, two or more ellipses were presented. The viewer’s task was to press the response button whenever two or more ellipses were observed on a trial. The screen changed every second. We constructed this fairly demanding visual task in the hope that it would draw the listener’s attention away from the words being presented auditorily. Despite the visual distraction, one can see the expected automatic N1 and P2 waveforms at 100 and 200 ms,

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   151

Figure 8–6.  Processing negativity (PN) waveforms under three conditions: (1) visual decoy, (2) nonword targets, and (3) semantic category targets. Average of five young adult listeners. See text for further details. Modified with permission from Figure 6.1 in Jerger et al. (2014).

respectively, but, in addition, a small PN component beginning at about 300 ms, falling to a negative peak of 2.5 microvolts at a latency of about 550 ms, then slowly returning to baseline by 900 ms. We did not expect that we would see such an obvious PN response to the nontarget words in view of the complicated visual distraction. It suggests that you cannot-not-process words when they enter your auditory system, whether you are actively distracted from them or not. In the second condition, the blue curve, the same nontarget words were used, but the target words were actual CVC words but played backward. We called this the nonword condition. The auditory task was not difficult, but the negative peak of the PN component dropped to about 5.0 microvolts.

152  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Finally, in the third condition, the red curve, the same nontarget words were used, but the target words were in designated semantic categories, either the names of animals or articles of clothing. On this more demanding task, the negative peak of the PN component dropped further to negative 7.6 microvolts. The most important thing to remember about these three differently colored lines is that they are all waveforms in response to the exact same nontarget words. The differences among the three conditions are only in the nature of each target-related task. Actually, this should not be surprising. If processing negativity of the nontarget words is due to strategies involved in deciding that the heard word is not a target word, then the fact that the nature of the decision process involving target words changes substantially across the three conditions must necessarily be reflected in the processing of the nontarget words across those same three conditions. The important lesson is that the nature of the task imposed on the target words calls into play differences in the analysis of the nontarget words. The dichotic paradigm can be used to illustrate even quite subtle differences across conditions and how they affect the PN component. Figure 8–7, for example, shows the average waveforms in 20 young adults with normal hearing. It was extracted from a study of dichotic listening carried out by a graduate student as part of her doctoral dissertation at UTDallas. Here the dichotically presented pairs of words were always preceded by a reference word. Then two words were presented dichotically. On each trial, there were three response possibilities: (1) The word heard dichotically on the right ear matched the semantic category of the reference word (Match to the Right), (2) the word heard dichotically on the left ear matched the semantic category of the reference word (Match to the Left), or (3) the semantic category of the reference word was not heard on either ear dichotically (No Match). The reference word differed on every trial. There were nine semantic categories from which the various reference and target words were selected. They included (1) animals, (2) foods,

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   153

Figure 8–7.  Processing negativity (PN) waveforms in three conditions of a dichotic procedure. Greater negativity in the Match-Left condition compared to the Match-Right condition is another index of the right ear advantage (REA). See text for details. Data were averaged across 20 young adults with normal hearing. Modified with permission from Figure 4.3 in Jerger et al. (2014).

(3) body parts, (4) articles of clothing, (5) natural earth formations, (6) types of vehicles, (7) weather phenomena, (8) kitchen utensils, and (9) musical instruments. Figure 8–7 shows the three PN waveforms recorded at the left centro-parietal (CP3) electrode. Matching the reference word to a word of the same semantic category presented dichotically to the right ear (red curve) required least effort. The difference between Match to the Right Ear (red curve) and Match to the Left Ear (blue curve) waveforms is due, of course, to our old friend, the right ear advantage

154  AUDIOLOGICAL RESEARCH OVER SIX DECADES

(REA). In the Match-Left condition, the PN was slightly deeper at peak negative amplitude and took longer to return to baseline. Finally, the No Match condition (green curve) required the greatest mental effort, not returning to baseline until a latency slightly greater than 1,200 ms. This is a potentially quite interesting aspect of dichotic listening. In any event, I am obliged to remind you, again, that across all three conditions, there was no difference in the nontarget words being averaged. The differences were only in the instructions to the listener concerning target words.

Repeating Words Back Versus Making Decisions About Them In the typical audiology clinic, speech understanding ability is assessed with lists of CVC words. The listener hears a word, then must repeat it back to the tester for immediate scoring. This is a task requiring only a minimum of mental effort. We are reminded that parrots handle it amazingly well. In one AERP study, we compared the effort expended in this repetition paradigm with the effort required by a task in which the same 60 words were recast in a category judgment task. Nine words were the names of animals. Figure 8–8 shows the result in a 31-year-old man with normal hearing. In the “repeat word” condition (blue curve), there was a PN peak of about −3.5 microvolts at a latency of 375 ms, with recovery to baseline at about 800 ms. This was the response to simple repetition of each word. There was no distinction between target and nontarget words. They were all treated as words to be repeated back. In comparison, the PN for the category judgment condition showed a PN peak of almost −5.0 microvolts at a latency of slightly more than 550 ms. Moreover, recovery to baseline was not complete until about 1,100 ms. The difference between the red and blue waveforms reflects the difference in cognitive effort invoked under each condition.

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   155

Figure 8–8.  Processing negativity (PN) curves in an adult listener with normal hearing under two conditions: (1) simply repeating back the complete list of 60 words as they were presented and (2) making category judgments on the 9 target words. Both curves show a PN component, but it is more prolonged in the case of the category judgments.

More PN Examples In Figure 8–9, you will find four widely different examples of the PN component. In panel a (left upper quadrant), two different PN curves are shown. The black curve was the PN response to noiseburst targets. The red curve was generated by semantic category targets (words that belonged to the category — animals). Both waveforms show a processing negativity component peaking at about 400 ms, but the red curve is twice the depth of the black curve because more mental effort is required to decide whether

156  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 8–9.  Four different examples of processing negativity (PN) applications. a. Two different tasks: noise targets and semantic category targets. b. Repeat word versus make category judgment. c. Context priming. d. The N400 effect. Modified with permission from Figure 3.6 in Jerger et al. (2014).

a word belongs to a given semantic category than to decide that what you heard was not a word at all but a burst of noise. In both tasks, however, the waveform of the PN component was based on the same nontarget words. The difference is in the cognitive

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   157

resources that must be marshaled for each of the two tasks, not in the nontarget words themselves. They were the same in both conditions. We shall have occasion to remind the reader of this fact later in the chapter as well. It cannot be overstressed that what changes, in these comparisons of different conditions, is not the nontarget words. They are always the same, just presented in different orders across experimental conditions. In panel b (upper right quadrant), the black curve was the PN waveform generated by simple word repetition as in PB word testing. A list of words including both semantic category targets and nontargets was presented to the listener. The task was simply to repeat aloud each heard word. Then the red curve was generated by repeating the procedure but requiring a semantic category judgment instead of simple repetition. Note that the repetition task generated almost no processing negativity, but the category judgment task generated a substantial negativity peaking at about 500 ms. Again, both the black and the red curves are based on the exact same nontarget words. It is only the context in which they were presented that differed. Panel c, in the lower right quadrant, shows the substantial effect of context priming on the PN component. Here we see the difference between waveforms evoked by words that have been either primed or not primed by the previous word in a memory task. Panel d (lower left quadrant) shows the familiar N400 effect generated by presenting sentences in which the last word may or may not be congruous with the gist of the sentence. In the case of the incongruous words, there is a clear negativity peaking at about 700 ms.

A Case of Multiple Sclerosis Figure 8–10 shows dichotic test results in a 41-year-old woman diagnosed with multiple sclerosis. We also recruited a 39-year-old man (VG) as a control. The dichotic task was to judge whether

Figure 8–10.  Upper panel —Late positive component (LPC) and processing negativity (PN) components in a young normal adult. Lower panel —Same conditions in a patient with multiple sclerosis. Note absence of LPC but presence of PN. Modified with permission from Figure 4.10 in Jerger et al. (2014).

158

8.  AUDITORY EVENT-RELATED POTENTIALS TO WORDS   159

a word from the cued-side word belonged to the appropriate semantic category. First, study VG’s data. There are three curves, red for right-sided targets, blue for left-sided targets, and green for nontargets. The red and blue curves reflect, of course, the respective LPC components for the right ear and left ear targets. Note also the REA reflected by these two curves. The left ear LPC has a longer peak latency and a smaller peak amplitude than the right ear. The right ear advantage again? Turning now to listener JW, it is obvious that she had no LPCs. Apparently, one consequence of her malady was that she could not make the decisions required for the development of the necessary right and left target LPC waveforms. But all three PN curves, right target words, left target words, and nontarget words, show similar classical PN functions. Here is a good example of the difference between the LPC, requiring a conscious decision, and the PN, not apparently under strong conscious control.

A CLARIFICATION The thoughtful reader may wonder why we employed 32 active electrodes when we have only shown examples from one or two electrodes. There are at least two good reasons. First, the data from all 32 electrodes permit the construction of a topographic map showing how the distribution of electrical activity across the entire surface of the scalp changes with the experimental conditions. We have used these data in other publications (see, for example, the next chapter). Second, such maps are useful in selecting which individual electrodes yield the optimal waveforms for comparisons among conditions.

FINAL THOUGHTS My aim, in going through all of these examples of AERPs to words, is to try to convey, to my fellow audiologists, the extraordinary

160  AUDIOLOGICAL RESEARCH OVER SIX DECADES

versatility of the procedures as they apply to the issue of assessing the mental effort required for useful speech communication. We must get beyond repeating words back to the examiner. There are much better ways if we will only try them. The AERP paradigm, especially the processing negativity component, is totally open-ended. You often hear it said that auditory evoked potentials are fine with group data but not useful for assessing individual listeners. This is certainly not true of ABR (ask any baby who has been screened for hearing loss), and it is not true of AERPs either. That is why we have focused so much of our work in AERPs on individual listeners, whether normal or abnormal, because, as clinicians, we all want methods and techniques that we can use to investigate people, not groups of people.

RELATED READING Jerger, J., Martin, J., & Fitzharris, K. (2014). Auditory event-related potentials to words: Implications for audiologists. Create Space Independent Publishing Platform.

9 A Twin Study

161

162  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Twin studies are always interesting but most valuable if only one twin presents with a problem. Since both twins typically share a common environment, the other twin can serve as an ideal control. This chapter summarizes a study of two 10-yearold fraternal twin girls. One had developed normally, but the other had suffered a febrile seizure at age 17 months that lasted about 10 minutes. This twin — hereafter identified as “Twin E” (experimental) — suffered persistent academic problems. She had difficulty understanding concepts and remembering complex instructions. Her sister — hereafter referred to as “Twin C” (Control) — demonstrated none of these problems. On the Children’s Auditory Performance Scale (CHAPS), Twin C’s scores were all in the pass range, but Twin E was rated as at risk for (1) listening in noise, (2) listening in quiet, (3) memory, and (4) attention. Our task was to explore the possibility of auditory processing disorder (APD) in Twin E in greater detail. In the School of Behavioral and Brain Sciences at UTDallas, we were fortunate to have colleagues and graduate students with considerable skills in the behavioral evaluation of children with suspected APD. They included Linda Thibodeau, Jeffrey Martin, Jyutika Mehta, Gail Tillman, Ralf Greenwald, Lana Britt, Jack Scott, and Gary Overson. Together we developed a plan to administer a series of both behavioral and electrophysiological measures to both twins. Our goal was to search for evidence of any significant performance differences between Twin E and her sister, Twin C, consistent with APD.

BASIC AUDIOMETRY We administered the following basic test battery: Pure-tone audiogram Word recognition in quiet (PBK lists)

9.  A TWIN STUDY   163

Hearing in Noise Test for Children (HINT-C) Tympanograms Ipsilateral acoustic reflexes Auditory brainstem response (ABR) Transient evoked otoacoustic emissions

BEHAVIORAL PSYCHOACOUSTIC MEASURES OF AUDITORY PROCESSING We administered the following test battery: Screening Test for Auditory Processing Disorders (SCAN) Simultaneous and Backward Masking Frequency Sweep Discrimination Binaural Gap Detection for tones and clicks (based on the gap detection research of Robert Keith)

STANDARDIZED COGNITIVE/ LINGUISTIC EVALUATIONS Here the following tests were administered: Wechsler Intelligence Scale for Children III (WISC III) Clinical Evaluation of Language Function (CELF III) Token Test for Children (TTC) As it turned out, we found little evidence of any significant deficit in Twin E on any of these measures. Yet the behavioral

164  AUDIOLOGICAL RESEARCH OVER SIX DECADES

auditory gap detection test results were intriguing: One twin was better for tones and the other was better for clicks, although both were within the norms for both stimuli. We resolved to subject gap detection and dichotic listening to further examination, by means of auditory event-related potential (AERP) analyses.

ACTIVATION PATTERNS In Chapter 8, we concentrated on the waveforms of the late positive component (LPC) of the AERP for target words and the waveforms of the processing negativity (PN) of the AERP for nontarget words. We analyzed how voltage changed over time at an individual electrode. It is entirely appropriate to ask why, then, we needed to record from 32 active electrodes. Actually, for two good reasons: (1) in order to choose the electrode reflecting the maximum voltage of the LPC or PN and (2) in order to study how activation is spread over the surface of the skull — whether activation is greater in one hemisphere than the other and whether more active in the frontal region of the skull or in the more central and posterior regions. But there is a price to be paid for this additional spatial information. In order to ask questions about spatial distribution, you must give up information about time. There is a different spatial distribution at every moment of poststimulus activity. You can view the changing pattern over time in the form of a motion picture, but it is usually more helpful to concentrate analysis on the spatial pattern at the latency yielding the maximum voltage over poststimulus time. Let us consider some basic aspects of such activation patterns. Figure 9–1A shows the labeling of the 32 electrodes in our system. Moving from the front of the head to the back (from anterior to posterior), each row is labeled as follows: Prefrontal (FP) Frontal (F)





Figure 9–1A.  Configuration of recording electrodes in a 32-channel electrode array. Beginning at the very front of the head and moving toward the back of the head, successive rows of electrodes are described as prefrontal (FP), frontal (F), fronto-central (FC), frontotemporal (FT), temporal (T), central (C), temporo-parietal (TP), centroparietal (CP), parietal (P), and occipital (O). All midline electrodes end in Z, all electrodes to the right of midline end in an even number (e.g., C4), and all electrodes to the left of the midline end in an odd number (e.g., F3). Electrode CZ is also known as the Vertex electrode. Modified with permission from Figure 1.1 in Jerger et al. (2014).

165

166  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Fronto-central (FC) Fronto-temporal (FT) Central (C) Temporal (T) Temporo-parietal (TP) Centro-parietal (CP) Parietal (P) Occipital (O) All electrodes in the midline end in Z (e.g., FCZ). All electrodes to the right of midline end in even numbers (e.g., P8). All electrodes to the left of the midline end in an odd number (e.g., TP7). The CZ electrode is also known as the Vertex electrode. In this chapter, we are only concerned with LPC activation. The activation pattern for the LPC to words usually occupies the space bounded by electrodes C3 to CZ and P3 to PZ in width and C3 to CZ and CZ to PZ in height, an area that is posterior and eccentric to the left hemisphere. Well, if you have a spatial distribution, or activation of the LPC, based on 32 data points, how do you convey differences in voltage levels at each individual electrode? Actually, it is done with colors! Some systems assign different individual colors to the varying voltages across the surface of the scalp. Other systems employ gradual variation over shades of gray, or just two colors, but varying in intensity from a strong representation of one of the two colors, through white (no color), then gradually building up into a strong representation of the other color. Figure 9–1B shows, for illustrative purposes, an example of a topographic map from a different experiment, rendered in the two-color system (blue and red). The voltage scale is shown as a series of bars to the left of the map. It begins at dark blue for zero

9.  A TWIN STUDY   167

Figure 9–1B. Example of an activation pattern based on a two-color voltage display. Black dots are the electrode positions labeled in Figure 9–1a. Number below display is latency in milliseconds to peak voltage across the electrode array. Red area over right hemisphere indicates region of maximum activity. Voltage scale runs upward from 0 microvolts (deep blue) to positive 30 microvolts (strong red).

voltage, becomes lighter shades of blue, merges into white, and then climbs from weak to strong red, peaking at 30 microvolts. This topographic map was recorded at the peak latency of the red activity area — in this case, 1,000 ms (shown at the base of the topographic map). We had originally planned to study LPC activation in the twins, but only for words presented in the dichotic mode. However, the interesting results of the behavioral studies of the binaural random gap detection test (one twin better for tones, the other twin better for clicks) prompted us to set up AERP studies

168  AUDIOLOGICAL RESEARCH OVER SIX DECADES

of auditory gap detection as well. Each trial consisted of a 500-ms burst of white noise. In 25% of trials, a 10-ms gap was inserted in the temporal center of the burst. These were considered target trials. In the remaining 75% of trials, there was no gap. These were treated as nontargets. Our principal interest was the LPC component of the AERP, the response to targets, to each twin in each condition. Figure 9–2 shows the activation patterns of the LPC for each twin, derived from the gap detection procedure. In the case of Twin C, it was a strong, essentially bihemispheric response at a peak latency of 710 ms. In the case of Twin E, however, there was virtually no response — nothing but dark blue over most of the surface of the skull. This difference between the

Figure 9–2. Activation pattern for binaural auditory gap detection for each twin. The activation pattern for Twin C rose to a maximum of 18 microvolts at a peak latency of 710 ms. It extended over both hemispheres but was greater over the left hemisphere. In contrast, the pattern of activation for Twin E never rose above the zero baseline over the area covered by Twin C’s activation pattern. Modified with permission from Figure 4 in Jerger et al. (2002).

9.  A TWIN STUDY   169

twins was striking. At first, we could not believe such a strong effect, especially after the long list of negative findings for the various behavioral measures listed earlier. I thought it must be some undetected technical failure in the recording apparatus. I had to reject this explanation, however, when one of the graduate students pointed out that Twin E had been tested first — before Twin C. Another graduate student suggested that it might be interesting to look at visual gap detection in both twins. If visual gap detection were normal in Twin E, that would further confirm the absence of a technical problem with her auditory gap detection data. In addition, it might support an argument for modality specificity in Twin C’s data. So, we devised an apparatus for measuring visual gap detection. A trial occurred every 1.6 seconds. On each trial, the twin saw a white dot, 8 cm in diameter, on the black screen of a computer monitor for 900 ms. On 25% of trials (target trials), the dot was erased for 200 ms. On the remaining 75% of trials, there was no gap in the dot (nontarget trials). We chose the 200-ms duration on the basis that lab members informally judged it to be at about their thresholds for detecting the visual gap. Figure 9–3 shows the activation patterns for each twin in the visual gap detection study. Again, positivity was prominent in the 750- to 770-ms latency range. The twins’ data showed almost identical results. The LPC for Twin C peaked at 14 microvolts at 770 ms. The data for Twin E showed a similar result, an LPC peak of 15 microvolts at a latency of 750 ms. This result does argue for modality specificity, but firmer data are lacking.

DICHOTIC LISTENING To study dichotic listening, perhaps the most sensitive APD test in children, we devised a dichotic test appropriate to the twins’ ages. We presented two different CVC words simultaneously to the twins’ two ears in an “oddball” paradigm designed to

170  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 9–3.  Activation pattern for binocular visual gap detection for each twin. The activation pattern for Twin C rose to a maximum of 14 microvolts at a peak latency of 770 ms. The activation pattern for Twin E rose to a maximum of 15 microvolts and peaked at 750 ms. The activation patterns were similar for the two twins. This figure confirms that the deficit in auditory gap detection for Twin E was modality specific. In both twins, the activation pattern was symmetric over the two hemispheres. Modified with permission from Figure 3 in Jerger et al. (2002).

yield an LPC. On each trial, the targets were words that shared a common phonemic feature, specifically words “that rhymed with “jet,” for example, “bet,” “let,” “met,” “pet.” The nontargets were words chosen at random from a previously recorded list of CVC words. We presented target words on 30% of trials and nontarget words on the remaining 70% of trials. Target words were presented to the right ear on 15% of trials and to the left ear on 15% of trials. Each twin was instructed to listen carefully, on a precued ear, for words that rhymed with the word “jet” and to press a response button labeled “Yes” if such a rhyming word was heard in the precued ear and to press a response button labeled “No” if

9.  A TWIN STUDY   171

a rhyming word was not heard in the precued ear. We analyzed results separately for target words presented to the right ear only and to the left ear only (that is, directed report rather than free report). We then analyzed activation patterns for the LPC component of the AERP separately for the two conditions, targets to right ear and targets to left ear. Figure 9–4 compares LPC activation patterns for the two twins when the target word was always presented to the left ear only. The Twin C pattern peaked at 16 microvolts while the Twin E waveform peaked at only 6 microvolts, almost a 3:1 amplitude difference favoring Twin C.

Figure 9–4.  Dichotic listening to words. Phonemic target to left side. The activation pattern for Twin C rose to a maximum of 16 microvolts at a peak latency of 850 ms. It covered a large area centered over the left posterior region of the skull, extending from the left occipital electrode (O1) to the left central electrode (C3) and the vertex electrode (CZ). For Twin E, there was a small region of activity in the right parietal region roughly between electrodes P4 and PZ. The activation pattern for Twin E rose to a maximum of 6 microvolts and peaked at 880 ms. Modified with permission from Figure 7 in Jerger et al. (2002).

172  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Figure 9–5 shows the results when targets were presented to the right ear only. In both cases, LPC peaks were in the range from 850 to 880 ms. Here we have what is perhaps the worst-case scenario. Twin E was poorer than her sister not only on the left ear but on the right ear as well. In Figure 9–6, the activation patterns of Figures 9–4 and 9–5 have been rearranged in order to contrast results for the two ears of each twin. In Twin C, the activation pattern was almost identical for the two ears, but in Twin E, the activation pattern was

Figure 9–5.  Dichotic listening to words. Phonemic target to right side. The activation pattern for Twin C rose to a maximum of 16 microvolts at a peak latency of 850 ms. It covered a large area centered over the left posterior region of the skull, extending from the left occipital electrode (O1) to the left central electrode (C3) and the vertex electrode (CZ). For Twin E, a small region of activity in the right hemisphere covered an area in the right parietal region roughly bounded by electrodes P4, PZ, CPZ, and CP4. The activation pattern for Twin E rose to a maximum of 10 microvolts and peaked at 880 ms in the right parietal region. Modified with permission from Figure 8 in Jerger et al. (2002).

173

!"#$%'

Figure 9–6.  Same dichotic data as shown in Figures 9–4 and 9–5 but rearranged to facilitate comparison between target-right and target-left activation patterns. In the case of Twin C, responses to target-right and target-left conditions were strong, equivalent, and greater over the posterior left hemisphere. In the case of Twin E, however, the response to the target-left condition was weak, greater in the target-right condition, and located over the right hemisphere.The results for Twin E reflect a marked left ear disadvantage, a common sign of APD in children.

Target to Left Ear

Target to Right Ear

!"#$%&



174  AUDIOLOGICAL RESEARCH OVER SIX DECADES

substantially smaller for the left ear than for the right ear. These are abnormally large right ear advantages, or left ear disadvantages in Twin E. What does all of this have to do with auditory processing disorder? Actually, quite a bit. It has been suggested that such abnormally large left ear disadvantages can best be explained as a loss in the efficiency of transfer of information from the right hemisphere to the specialized language-processing area in the left hemisphere via the corpus callosum. Why, then, not take a look at the corpus callosum?

DIFFUSION TENSOR IMAGING We did just that. We actually examined this possibility in the twins. Our colleague Roderick McColl, at the Department of Radiology, University of Texas Southwestern Medical Center, studied diffusion tensor imaging on both twins. He found, in Twin E, reduced anisotropy over the length of the midline corpus callosum and adjacent lateral structures, indicating reduced myelin integrity (“anisotropy” means having unlike properties in different directions). Well, here was “hard evidence” of a firm link between observational evidence of an auditory processing problem with something actually known to be wrong in the brain. One of my mentors at Northwestern, Helmer Myklebust, who opened the door to all of this almost 70 years ago, would have been pleased indeed. He always bristled a bit when a medical colleague would say, “Yes, Helmer, but where is the hard evidence?”

RELATED READINGS Jerger, J., Martin, J., & Fitzharris, K. (2004). Auditory event-related potentials to words: Implications for audiologists. Create Space Independent Publishing Platform.

9.  A TWIN STUDY   175

Jerger, J., Martin, J., & McColl, R. (2004). Interaural cross correlation of ERP waveform reveals auditory processing disorder. Journal of the American Academy of Audiology, 15, 79–87. Jerger, J., Thibodeau, L., Martin, J., Mehta, J., Tillman, G., Greenwald, R., Britt, L., Scott, J., & Overson, G. (2002). Behavioral and electrophysiological evidence of auditory processing disorder: A twin study. Journal of the American Academy of Audiology, 13, 438–460.

10 Odds and Ends

177

178  AUDIOLOGICAL RESEARCH OVER SIX DECADES

A VISIT TO MONTREAL You will have noticed, in the previous chapters, that our labs at Baylor College of Medicine in Houston and at UTDallas did a good deal of research in the area of dichotic listening. I first encountered the concept in the late 1950s. I was on a consulting visit to Montreal for my friend and former colleague, Earl Harford. After finishing his PhD degree at Northwestern, Earl had taken a job as head of audiology at the Royal Victorian Hospital in Montreal. The hospital, on the campus of the McGill University Health Center, was near the famed Montreal Neurological Institute (MNI). Earl suggested that, while I was in Montreal, I ought to visit the laboratory of Brenda Milner at MNI. He thought she was doing some interesting work that might interest me. Earl and I had barely entered the MNI building when I was introduced, in the hall, to Herbert Jasper, famed for the international 10-20 EEG electrode placement system and for his collaboration with world-renowned neurosurgeon, Wilder Penfield. Our next stop in the building was an operating room where Penfield was then operating on a brain. I did not get to meet him but was grateful for the opportunity to see him at work. Dr. Milner could not have been more gracious. She welcomed me to her lab; introduced me to her associate, Doreen Kimura; and showed me the apparatus they were using to study dichotic listening. It was a modest, two-channel tape recorder. On it they had recorded pairs of numbers. The listener was instructed to repeat back both numbers on each trial. They were planning to take the recorder up to the neurology wards to test patients with various brain disorders. Before embarking on the project, however, they wisely tested a control group of normal young adults to establish a baseline for their test. When they analyzed their results, they were surprised to find that the young adults scored slightly better on the right ear than on the left ear. The effect was small but reliable. Thus was the right ear advantage, or REA, born. All the way back to Chicago on the plane, I mulled over

10.  ODDS AND ENDS   179

the thought that they were on to something important, but I had other interests at the time and didn’t get into it myself in a serious way until I moved to UTDallas in 1997. But I have treasured the memory of meeting Herbert Jasper, Brenda Milner, and Doreen Kimura and seeing the great Wilder Penfield at work.

“NORMAL AUDIOMETRIC FINDINGS” In the late 1970s, I participated in an educational seminar at a ski resort in Idaho. One of the medical lecturers was giving a paper on otological diagnosis. When it came to the question of hearing loss, he said that if the audiological findings were negative, there was no need for further evaluation. He amplified this by defining “audiometrically normal” as air and bone conduction audiograms within normal limits and PB Max scores within normal limits. When I returned to Houston, I asked one of my assistants to “go to our files and, working backward into the past, extract the folders of the first 20 patients with retrocochlear, or more central, problems who had normal audiograms and normal PB Max scores.” The assistant came back about 30 minutes later with 20 folders and placed them on my desk. In those days, we had just two criteria for a retrocochlear finding: (1) significant rollover of the performance-versus-intensity (PI) function for PB words and (2) SSI Max significantly poorer than PB Max. Figure 10–1 summarizes what her search revealed. Every one of these 20 individuals would have been declared audiometrically normal if the decision were based on the audiograms and the PB Max score. But all had serious retrocochlear findings based only on PB rollover and the PB-SSI discrepancy. Connie Jordan and I wrote a paper on this. It was published in the American Journal of Otology but triggered almost no response from any quarter. Connie and I published that paper in 1980, over 40 years ago. I am told by colleagues who know about such things that “Air, Bone, and a PB List at SL = 40 dB” is still considered the standard



Figure 10–1.  Audiometric findings in 20 patients with retrocochlear auditory dysfunction. In all 20 patients, pure-tone audiograms and PB Max scores were within normal limits, but all failed at least one of two indices of retrocochlear disorder. See text for more details. Modified from Table 1 in Jerger and Jordan (1980).

180

10.  ODDS AND ENDS   181

audiological test battery in medical environments. The first time I heard this, I had to consult a calendar to be sure that I had not accidentally fallen into a negative time warp.

SIMIAN SURGERY A colleague at Baylor, Makoto Igarashi, studied the auditory and vestibular systems of squirrel monkeys. In 1978, he came to our lab with a question. He knew that we had set up a system to signal average the stapedius muscle reflex (see Chapter 2 of this volume) and asked whether the system could be adapted to study the stapedius muscle reflex in his squirrel monkeys. A group in Sweden had just published a paper in which they had studied the effect of transection of the olivocochlear bundle on the stapedius muscle reflex in rabbits. Igarashi wanted to do a similar transection in his monkeys. Could we help him? It took a bit of doing, but we rigged up a system that would work if the monkey was wide awake but restrained from movement. It was our first and last foray into the animal kingdom, but it actually worked out pretty well. It turned out that midline resection of the olivocochlear bundle had no effect on the crossed reflexes of Igarashi’s squirrel monkeys, a result that did not agree with the Swedish results. But Igarashi suggested that the disagreement might be the result of a species difference (monkey versus rabbit). We returned to the study of humans who, thankfully, need less restraint to be tested via immittance audiometry.

A RESEARCHER’S DREAM It has been said that people’s personality characteristics are often influenced by their names. That was certainly true of my friend, Dr. Paul Pepe. He stormed into my office at Baylor, one day in 1982, brimming over with pep and energy. “Have you ever,” he

182  AUDIOLOGICAL RESEARCH OVER SIX DECADES

asked, “tested the hearing of firemen?” I ventured a cautious “no,” fearing what might come next. In addition to his teaching duties on the faculty of the Department of Medicine at Baylor, Paul was charged with monitoring the health and well-being of the Houston Fire Department’s Emergency Medical Service (EMS) personnel. They are the ones who rush off to the latest medical emergency in ambulances (called ALS or advanced life support vehicles) with the flashing lights and loud sirens. Paul was concerned about their hearing because the noise levels in the cabs of the ALS vehicles varied from 95 dBA with the windows shut to 110 dBA with the windows open. He also mentioned in passing that the department had recently hired a number of EMS personnel who were Black, and he wanted to be sure that their hearing was adequate to the stressing tasks they faced on a daily basis. When he mentioned Blacks, my mind began to spin. Data comparing hearing sensitivity and both temporary and permanent threshold shifts of Blacks and Whites after excessive noise exposure were not plentiful. There had been surveys, but relevant variables were difficult to control. Here was an unusual opportunity to compare the hearing of the two groups while the following critical variables were under automatic control: 1. Age 2. Gender 3. Educational level 4. Health status 5. Place of employment 6. Equivalent minimal employment entrance requirements 7. Equivalent income 8. Work-related noise exposure level 9. Nature and extent of non-work-related noise exposure

10.  ODDS AND ENDS   183

Finally, constant exposure to the high-level emergency siren virtually guaranteed significant loss of hearing over a short period of work experience in the Black and White groups. Here was a rare opportunity to compare the rate at which hearing loss developed and increased as noise exposure accumulated in both groups. From a pool of about 180 EMS personnel, we were able to form two well-matched groups as follows: 28 Black males, average age = 28.2 years, and 28 White males, average age = 27.9 years. Duration of employment ranged from 22 to 174 months in the Black group with a mean duration of 73.2 months. In the White group, duration of employment ranged from 17 to 173 months with a mean duration also of 73.2 months. Was this not a researcher’s dream? From their audiometric thresholds, we calculated, for each participant, the average loss across the frequency range from 3000 to 6000 Hz on both the better ear and the poorer ear. We wanted to concentrate analysis in this high-frequency range where any loss was most likely to appear. We then formed four subgroups within each group based on duration of employment. Average results for the better ears of both groups are plotted in Figure 10–2. Similar patterns were observed on the poorer ear as well. Hearing loss was greater in the White group, and the group difference increased as duration of employment increased. We speculated that the reason for the apparent racial difference might be the greater quantity of protective melanin in the striae vascularis of the cochlear duct in the Black participants.

A Voice From the Past Before writing a report on our audiometric findings and submitting it to Dr. Pepe, I wanted to be sure that I was using the proper terminology for the two study groups. I could not have found a better consultant than Dr. Willis Beasley — yes, the Willis Beasley

184  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Hearing Loss in dB

20

White Black

10

! ! !

0 17–50

51–65

66–95

96–107

Duration of Employment in Months Figure 10–2.  Hearing levels of Black and White Houston Fire Department Emergency Medical Service (EMS) personnel, as a function of length of employment. Note consistently poorer hearing levels in the White group and steeper rate of increasing loss with duration of employment in the White group. Modified from Figure 2 in Jerger et al. (1986).

himself — the author of the report on the 1935–1936 USPHS Survey of Human Hearing in the USA: the basis for the first American standard for the calibration of audiometers. As a young officer in the U.S. Public Health Service, he had been commissioned to carry out the survey so that all audiometers in the country could be calibrated to the same standard. For audiologists, he has always enjoyed an iconic status. Now, almost 50 years later, he was still active in his chosen field. At the moment, he was consulting at the University of Texas Medical School in Houston, which happened to be just a few buildings away from my office at the Baylor College of Medicine. We chatted for some time about his famous survey before I raised the reason for my visit.

10.  ODDS AND ENDS   185

“What is the proper wording,” I asked, “to describe racial groups in public health documents?” I girded myself for a long, technically complicated answer. “That’s easy!” he replied, “White, Black, and other!”

A HERCULEAN EFFORT I would like to take you back to the year 1957. The scene is the Auditory Research Laboratory on the Evanston campus of Northwestern University. It is lunch time. Tom Tillman and a few other graduate students are in the lunch room with me. We are playing a game of cards (Sheepshead) while eating our brown bag lunches. Dr. Raymond Carhart enters the room, stares disapprovingly at our unproductive activity, but then gets down to business. “Fellows,” he says, “I want you to tackle a very important problem. While the reference SPLs for normal pure-tone thresholds are finally standardized (the original ASA standard for diagnostic audiometers), the standard for speech audiometers is in a chaotic state. It is supposed to be standardized relative to the normal threshold of a 1,000 cps (now 1000 Hz) tone, but they [he never told us who ‘they’ were] have the difference all wrong. They are suggesting 6 dB (that is, the difference between the SPL at normal threshold of a 1,000-cps tone and the SPL of a 1,000-cps tone recorded at the average of frequent peaks of spondee words should be 6 dB), but I am certain that that number is much too small. Based on our experiences at Deshon Hospital during the war, I suspect that it is at least 10 dB and probably more.” There was a long pause and then, “I want you fellows to find the right number. And while you are at it, there are a lot of variables to consider; you need to start by making a list.” It was not difficult. We came up with a list of five variables: (1) gender, (2) ear, (3) test order, (4) prior knowledge of spondee vocabulary, and (5) method of testing (Clinical versus Adaptive).

186  AUDIOLOGICAL RESEARCH OVER SIX DECADES

We tested 96 young adults with normal hearing. There were two results for each listener: the threshold for a 1000-Hz tone and the threshold for spondee words. Prior knowledge of spondee words was imparted to 48 of the participants by reading them the entire W-1 list of spondees before testing. Similarly, 48 participants were tested using the Hughson-Westlake method of threshold seeking. The other half of the total group was tested by an adaptive psychophysical method called “Up and Down,” in which the intensity is lowered until the listener does not respond, and then the level is increased until a response returns, then decreased until a response again disappears, and so forth. It is basically a simple 50% adaptive rule. That testing, in the late 1950s, may well have been the first application of an adaptive testing technique with hearing-impaired individuals. Such elegant techniques are still not widely employed in audiological testing. In any event, making a list and testing the 96 young adult normal listeners was the comparatively easy part of the project. After all of the data had been collected, we were faced with the daunting task of carrying out a five-factor analysis of variance (ANOVA). Today’s young investigators have no real idea of what data crunching was like in those days. There were no ANOVA programs for computers because there were no computers. We had only books telling us what were the equations to be applied in order to gather all of the sums of squares necessary to calculate all the variances necessary to arrive at the mean squares and F ratios necessary to decide whether anything was significantly different. The only labor-saving device at our disposal, beyond a pencil and a pad of paper, was an aged Friden calculator that allowed one to enter 96 numbers, each number twice, on two different parts of the machine’s keyboard, in order to arrive at one of the squared numbers necessary to calculate the sums of squares necessary to proceed with the analysis. If you have access to one of the older statistics textbooks, take a look at the number of sums of squares involved in a five-factor ANOVA. If I were to remind

10.  ODDS AND ENDS   187

you further of the number of interactions among five factors, you would stop reading now. So I will move on. Coincidentally, our Friden calculator, acquired from an Army/Navy surplus store, was the same model as the Friden calculator I had used as a clerk in the Finance Office of the Army Separation Center at Fort Lawton, Seattle, in 1946. But I digress. As project leader, I asked Tom Tillman to carry out the various calculations required to complete the ANOVA. After 3 exhausting days, he complained that the tips of his fingers no longer felt like they were attached to the rest of his hands. I reminded him that difficult work was character building. He took a deep breath, then opined that he might suggest someone else who needed character building more than he did. As it turned out, there was only one significant difference. Of the five main effects, only one, familiarity with the spondee words, exceeded the 1% confidence level. The threshold SPLs for the 1000-Hz tone hovered around 9 dB, and the threshold SPLs for the spondee words hovered around 22 dB, so that 13 dB was a pretty good estimate of the difference that should be built into the two types of audiometers. The effect was even larger than what Carhart had expected. He had thought it might be at least 10 dB and possibly even more. In the case of prior familiarity with the spondee words, the difference was 14.4 dB in the subgroup without prior exposure to the W-1 list and 11 dB in the subgroup exposed to the list. This difference of 3.4 dB was, indeed, statistically significant. The take-home message was that, before attempting to measure the spondee threshold, one should read a list of spondee words to the patient and inform him or her that these are the words he or she will be asked to listen for. All in all, it was something of a tempest in a teapot. A few years later, the American standard for pure-tone audiometers was abandoned anyway, in order to be in agreement with the new international standard.

188  AUDIOLOGICAL RESEARCH OVER SIX DECADES

FINIS So, here we are at the end of our journey through six decades of research that I have been privileged to be a part of. I hope that you have enjoyed reading about it as much as I have enjoyed writing about it.

RELATED READINGS Igarashi, M., Mauldin, L., & Jerger, J. (1979). Impedance audiometry in the squirrel monkey. Effect of transection of crossed olivocochlear bundle. Archives of Otolaryngology, 105, 258–259. Jerger, J., Carhart, R., Tillman, T., & Peterson, J. (1959). Some relations between normal hearing for pure tones and for speech. Journal of Speech and Hearing Research, 2, 126–140. Jerger, J., Jerger, S., Pepe, P., & Miller, R. (1986). Race differences in susceptibility to noise-induced hearing loss. American Journal of Otology, 7, 425–429. Jerger, J., & Jordan, C. (1980). Normal audiometric findings. American Journal of Otology, 1, 157–159. Jerger, J., Malmquist, C., & Speaks, C. (1966). Comparison of some speech intelligibility tests in the evaluation of hearing aid performance. Journal of Speech and Hearing Research, 9, 253–258. Pepe, P., Jerger, J., Miller, R., & Jerger, S. (1985). Accelerated hearing loss in urban emergency medical services firefighters. Annals of Emergency Medicine, 14, 438–442.

Index

Note:  Page numbers in bold reference non-text material.

A ABLB test (alternate binaural loudness balance test), 2–3, 58 Abnormal auditory adaptation, 13–14 ABR (auditory brainstem response), 136–137, 138, 139 Absolute impedance, 40, 42 Acoustic reflex, 42, 43–50 Acoustic tumors, Meniere’s disease differential diagnosis, 2, 3, 11–13, 12, 58 Activation patterns, twin studies, 164, 165, 166–169, 167, 168 Active listening, 138, 140–141 AERP (auditory event-related potential), 100–101, 102 Aging, 101 auditory aging, 116–133 auditory processing abilities, 119–123 binaural interference, 96 binaural vs. monaural hearing aids, 89–102 gender differences and, 126–133 Agnosia, 55 Alternate binaural loudness balance (ABLB) test, 2–3, 58 Amplivox, 5, 6

Animal studies, simian surgery, 181 APD. See Auditory processing disorder Aphasia and Kindred Disorders of Speech (Head), 55 Articulation function, 22 Audiogram Békésy audiogram, 14–16, 17–20, 21–22, 25–26, 62 continuous-tone threshold, 21 in elderly populations, 117, 118 fixed-frequency threshold, 21 gender effects on shape, 128–130, 129, 131, 132 interrupted-tone threshold, 21 “Audiological Manifestations of Lesions in the Auditory Nervous System” (Jerger), 58 Audiometric testing, 2, 6 Audiometry binaural vs. monaural hearing aids, 89–102 hearing loss despite normal findings, 179, 180, 181 immittance audiometry, 40–50, 181 speech audiometry, 22–27 standards for, 185–187 tympanogram, xi–xii, 42 Auditory aging, 116–133

189

190  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Auditory brainstem response (ABR), 136–137, 138, 139 Auditory event-related potential (AERP) auditory brainstem response (ABR) and, 138, 139 late positive component (LPC), 100–101, 102, 141–143, 142, 144–147, 144, 166–168 multiple sclerosis case study, 157, 158, 159 nontarget words, 147–148, 149 processing negativity (PN), 148, 149–154, 149, 151, 153, 155–157, 156 32-channel electrode array, 164, 165, 166 twin studies, 164, 165, 166–169, 167, 168 word recognition score, 135–160 Auditory localization in space, 82–83, 83–84, 85–87, 86 Auditory processing, aging and, 119–123 Auditory processing disorder (APD), xii, 54–70 case study, 63–66, 67, 68–70 cognitive decline and, 119–123 in the elderly, 119–123, 121 history, 54–63 Auditory stimulus, abnormal auditory adaptation, 13–14

B Baylor College of Medicine (BCM), 35, 41, 43, 63, 79, 95, 106, 178, 182, 184 Beasley, Willis, 183–184 Békésy, George von, 13–14

Békésy audiogram, 14–16, 17–20, 21–22, 25–26, 62 Binaural balance test results, 58–60, 59 Binaural hearing aids directionality of sounds in, 106–113 monaural hearing aids vs., 89–102 “Binaural Hearing Aids and Speech Intelligibility” (Jerger, Carhart & Dirks), 90 Binaural interference, 95–102 Bocca, Ettore, 28, 55–57, 58 Brainstem disorders, auditory deficits and, 28–35, 58–60 Britt, Lana, 162 Brown, Jimmy, 41 Bucy, Paul, 24

C Carhart, Raymond, 5, 15, 22, 40, 90, 185 Case studies aging and hearing, 125–126 auditory processing disorder, 63–66, 67, 68–70 binaural vs. monaural hearing aids, 95–102 central deafness, 76–88, 78, 83–84, 86 cued listening, 110–113 dichotic testing in multiple sclerosis, 157, 158, 159 Central deafness (cortical deafness), 74–88 case study, 76–88, 78, 83–84, 86 Goldstein’s criteria, 78 revised concept of, 88 Chapeau de Gendarme, 24

INDEX   191

Chmiel, Rose, 95 “Clinical Experience with Impedance Audiometry” (Jerger), 41 CLT (Cued Listening Test), 108–109, 109, 112–113 Coats, A.C., 79 Cochlear disorders, 24 Cognitive decline, auditory processing disorder and, 119–123 Competing messages, 27–35 Consonant-vowel-consonant (CVC) word, 137, 138, 140, 143, 154, 169 Continuous-tone threshold, 21 Contralateral speech competition, 33–34, 34 Cortical deafness. See Central deafness Cued listening, 105–113 case studies, 110–113 measurement, xii testing, 108–109, 109 Cued Listening Test (CLT), 108–109, 109, 112–113 CVC words, 137, 138, 140, 143, 154, 169

D Danish audiology facilities, 40 Diagnostic audiology, xi, 2–36 Dichotic listening left ear advantage, 56 right ear advantage (REA), 56, 68, 126–128, 127, 128, 144–147, 146, 178–179 testing, 56, 67, 68, 69–70, 96, 178 twin studies, 169–172, 170–173, 174

Dichotic Sentence Identification (DSI) test, 96, 123–125, 124 Diffusion tensor imaging, 174 “Directed report” condition, 123 Dirks, Donald, 90 “Divided report” condition, 68 Dix, M.R., 2, 3, 58

E Elderly. See Aging Endicott, Jim, 74, 77 Ewertsen, Harald, 40

F Fechner, Gustave, 4, 7 Fixed-frequency threshold, 21 “Free report” condition, 68, 123

G Gaeth, John, 116 Galileo, 90 Gender aging and, 126–133 audiogram shape and, 128–130, 129, 131, 132 right ear advantage (REA), 126–128, 127, 128, 144–147, 146 Goldstein, Robert, 74, 85 Grason, Rufus, 14 Grason-Stadler Co., 14, 41 Grason-Stadler E–800 (instrument), 15 Greenwald, Ralf, 162

H Hallpike, C.S., 2, 58 Hardy, William, 54

192  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Harford, Earl, 11–12, 174 Head, Sir Henry, 55 Hearing critical variables, 182 simian surgery, 181 Hearing aids binaural aids, 89–102 directionality of sounds in, 106–113 Hearing loss auditory aging, 116–133 in Blacks and Whites, 182, 183, 184 with normal audiometric findings, 179, 180, 181 retrocochlear disorders, 62, 179, 180 sensorineural hearing loss, 116 speech understanding in the elderly, 125–126 Hirsh, I.J., 81 Hood, J.D., 2, 58 Houston Speech and Hearing Center (HSHC), 35, 40–41, 79

I Igarashi, Makoto, 50, 181 Immittance audiometry, 40–50, 181 Interrupted-tone threshold, 21 Ipsilateral competing speech, 29, 32–34, 33

J Jasper, Herbert, 178 Jerger, Susan, xvi, 24, 41 Jewett, Don, 137 Jordan, Connie, 179

Jordan, Craig, 106, 112, 113 Just-noticeable differences (JNDs), 4, 5–6, 8, 11

K Kimura, Doreen, 99, 178 Klar, Irwin, 50 Knudsen, Vern, 93

L Late positive component (LPC), 100, 101, 102, 141–143, 142, 144–147, 144, 166–168, 171, 171 Left ear advantage, dichotic listening, 56 Lierle, D.M., 15 Localization of sound in space, 82–83, 83–84, 85–87, 86 Loudness recruitment, 3, 44 Lüscher, Eberhart, 3–7

M Madsen, Paul, 50 Madsen ZO60 (instrument), 40 Madsen ZO70 (instrument), 40, 50 Martin, Jeffrey, 162 McColl, Roderick, 174 Mehta, Jyutika, 162 Meniere’s disease, acoustic tumor differential diagnosis, 2, 3, 11–13, 12, 58 Meningioma, PI-PB functions, 23, 23 Message-to-competition ratio (MCR), 27

INDEX   193

Method-of-constants, 7 Methodist Hospital (Texas Medical Center), 36 “Middle ear impedance,” 40 Milner, Brenda, 178 Monaural hearing aids, vs. binaural aids, 89–102 Monkeys, surgery on auditory system, 181 Montreal Neurological Institute, 178 Morgan, C., 7 Multiple sclerosis, dichotic testing in, 157, 158, 159 Myklebust, Helmer, 54–55, 174

Processing negativity (PN), 148, 149–154, 149, 151, 153, 155–157, 156 Pure tone audiometry, standards for, 185–187

Q Quantal psychophysical method, 7–11, 8–10

R

Neural quantum, 8 Noise, emergency siren and hearing loss, 182, 183 Nonword condition, 151 Northern, Jerry, 50 Northwestern University, 35, 40, 54, 90, 185

Reflex averaging, 45–49, 46–48 Reger, S.N., 15 Retrocochlear disorders, 62, 179, 180 Right ear advantage (REA) dichotic listening, 56, 68, 126–128, 127, 128, 144–147, 146, 178–179 LPC and, 144–147, 146 Rollover phenomenon, 24–25, 25 Rost, Alan, 74, 77

O

S

“Off-time” pulses, 60, 62 “On-time” pulses, 60, 61, 86, 86 Overson, Gary, 162

“Scientific Writing Can Be Readable” (Jerger), xiii Scott, Jack, 162 Scott-Nielsen, S., 40 Sensorineural hearing loss, 116 Sharbrough, Frank W., 76 Sherrick, C.E., Jr., 81 Short Increment Sensitivity Index (SISI) test, 11–13, 12 Silman, Shlomo, 94 Silverman, Carol, 94 Simian surgery, 181

N

P PB testing, 22–24 Penfield, Wilder, 178 Pepe, Paul, 181–182, 183 Phonemic regression, 116 PI functions, 22, 24, 56, 57 Pirozzolo, Fran, 64, 120

194  AUDIOLOGICAL RESEARCH OVER SIX DECADES

Sound directionality of sounds in hearing aids, 106–113 localization in space, 82–83, 83–84, 85–87, 86 Speaks, Chuck, xvi, 27 Speech audiometry, 22–27 articulation function, 22 contralateral speech competition, 33–34, 34 ipsilateral competing speech, 29, 32–34, 33 PB testing, 22–24 PI functions, 22, 24, 56, 57 SSI (synthetic sentence identification), 26, 26–35, 28, 29–35, 30–34 Speech recognition auditory event-related potential (AERP) and word recognition score, 135–160 in the elderly, 110–133 repeating word back, 154, 155 Squirrel monkeys, surgery on auditory system, 181 SSI (synthetic sentence identification), 26, 28, 29–35, 30–34, 80, 96, 117 Stach, Brad, 49, 125–126 Stadler, Steve, 14 Stapedius muscle, 43–44 Stapedius muscle reflex, 42, 43–50, 181 Static impedance, 40, 42 Stevens, S., 7 “A Study of Phonemic Regression Associated with Hearing Loss” (Gaeth), 116 Sutton, Sam, 140 Swedish audiology facilities, 40

Synthetic sentence identification (SSI), 26, 28, 29–35, 30–34, 80, 96, 117

T Temporal lobe disorders auditory deficits and, 28–35, 56, 58–60 central deafness and, 74 Tensor tympani muscle, 43 Terkildsen, Knut, 40 Texas Medical Center (Methodist Hospital), 36 Thibodeau, Linda, 163 Threshold-versus-duration functions, 85 Tillman, Gail, 162 Tillman, Tom, 185, 187 Twin studies, 162–174 activation patterns, 164, 165, 166–169, 167, 168 dichotic listening, 169–172, 170–173, 174 diffusion tensor imaging, 174 LPC, 166–168, 171, 171 Tympanogram, xi–xii, 42

U Unilateral hearing loss, diagnosis, xi, 2–36

V Volkmann, J., 7

W Waveforms, 136–138, 139 Weber, Ernst, 4

INDEX   195

Weikers, Norbert J., 76 Word recognition score, auditory event-related potential (AERP), 135–160

Z Zwislocki, Joe, 4, 6, 41