Quantifying the effects of the cochlear amplifier on temporal and average-rate information in the auditory nerve by Michael Gregory Heinz Sc.B., Brown University (1992) M.S.E., Johns Hopkins University (1994) Submitted to the Division of Health Sciences and Technology in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Speech and Hearing Sciences at the MASSACHUSETTS INSTITUTE OF TECHNOLOGY June 2000 c Massachusetts Institute of Technology 2000. All rights reserved. Author.................................................................... Division of Health Sciences and Technology May 18, 2000 Certifiedby............................................................... Laurel H. Carney, Ph.D. Associate Professor of Biomedical Engineering, Boston University Thesis Supervisor Accepted by . ............................................................ Martha L. Gray, Ph.D. Edward Hood Taplin Professor of Medical and Electrical Engineering Co-director, Harvard-M.I.T. Division of Health Sciences and Technology Quantifying the effects of the cochlear amplifier on temporal and average-rate information in the auditory nerve by Michael Gregory Heinz Submitted to the Division of Health Sciences and Technology on May 18, 2000, in partial fulfillment of the requirements for the degree of Doctor of Philosophy in Speech and Hearing Sciences ABSTRACT An active mechanism referred to as the cochlear amplifier is believed to be responsible for sharp tuning and excellent sensitivity in the normal auditory system, and is impaired or absent in many common forms of sensorineural hearing loss. The cochlear amplifier is thought to benefit normal-hearing listeners, especially in complex listening environments in which hearing-impaired listeners have difficulty even with hearing aids. A modeling approach was developed to relate nonlinear physiological response properties associated with the cochlear amplifier to human psychophysical performance. Quantitative methods combined analytical and computational population models of the auditory-nerve (AN) with statistical decision theory to evaluate performance limits imposed by the random nature of AN discharges (modeled by a nonstationary Poisson process). A new theoretical ap- proach was developed to predict performance for psychophysical tasks that use random-noise stimuli to mask signal information. The ability of temporal and average-rate information in the AN to account for human performance was evaluated for several listening tasks for which the cochlear amplifier has been suggested to be significant. The benefit of the cochlear amplifier for extending the auditory system’s dynamic range was evaluated in terms of AN information available for encoding changes in stimulus level. An analytical model included nonlinear gain, level-dependent phase, and high-, medium- and low-spontaneous-rate AN fibers. Both nonlinear-phase and average-rate information were required to encode levels up to 120 dB SPL based on a narrow range of AN characteristic frequencies. A physiologically realistic mechanism to decode nonlinear gain and phase cues is monaural, across-frequency coincidence detection. Level-discrimination performance of a coincidence-counter population matched human performance across the entire dynamic range of hearing at both low and high frequencies. Results suggest that the cochlear amplifier is beneficial for encoding sound level within narrow frequency regions, and has only a small influence on simple listening tasks in quiet. The cochlear amplifier alters tuning within the normal auditory system based on the spectral and temporal configuration of the stimulus. The influences of compression and suppression on psychophysical measures of auditory frequency selectivity were evaluated. Implications for the interpretation of psychophysical methods for estimating auditory filters are discussed. Thesis Supervisor: Laurel H. Carney, Ph.D. Title: Associate Professor of Biomedical Engineering, Boston University 2 Acknowledgments There are many people to whom I owe thanks for their contributions to my dissertation work. I would like to thank the members of my thesis committee for extremely valuable and enjoyable interactions throughout the course of this dissertation. Their continual encourage- ment to focus this work on fundamental issues of broad interest has provided an important lesson to me, and has made it possible for me to finish in a reasonable amount of time. Don Eddington’s influence on my work began when I was a 4th-semester student in his research-methods class. He emphasized the benefits of formulating a good thesis topic and proposal, the importance of considering relevant clinical issues, and the often forgotten fact that working on a dissertation should be an enjoyable experience. Lou Braida, the chairman of the committee, also made an early and profound contribution to this work through his auditory-perception class, which introduced me to the method of combining statistical decision theory with auditory-nerve models to evaluate psychophysical performance limits. Lou has also helped to make me aware of specific conditions in which hearing-impaired listeners have the most difficulty, even with hearing aids. Andrew Oxenham provided valuable input to this work by continually reminding me of the important contributions that have been made to understanding the significance of cochlear nonlinearity through psychophysics, and of the many experiments for which hearing- impaired listeners demonstrate near-normal performance. Bertrand Delgutte made an extremely valuable contribution through his critical evalu- ations of this work, both in committee meetings as well as in conversations with me. He continually reminded me of the limitations of the auditory-nerve models used in this disserta- tion, as well as the new contributions of this work. I have greatly appreciated his willingness to critically read manuscripts, which were always too lengthy. I consider myself to be extremely fortunate to have had the opportunity to work so closely with both Steve Colburn and Laurel Carney during this dissertation. Steve’s willingness to meet with Laurel and me almost every week is a primary reason that this work has progressed to its current state. Steve is equally as kind as he is brilliant. His intuition about mathematics and the auditory systemhas strongly influenced myway of thinking, and the importance of his previous work to this dissertation is immeasurable. Finally, it is difficult to describe the profound influence Laurel Carney has had on this dissertation, as well as on me as a person. She has contributed invaluable guidance to every aspect of this dissertation, while giving me unrestricted freedom to work on issues that were of interest to me. Through her willingness to meet with me far more frequently than she had time for, she pushed me to continually move forward. Laurel’s encouragement to finish the dissertation quickly while maintaining rigor has provided valuable lessons on the practical issues of doing quality research. The ability to combine her enthusiasm for both learning and teaching with a genuine concern for her students has made her a great role model. Laurel’s advice and friendship in both academic and personal settings has made working with her an experience that I will never forget. In addition to the committee, I would like to thank Craig Formby and Moise Goldstein, my advisors at Johns Hopkins University, for introducing me to the field of hearing, and for giving me support and guidance at the beginning of my graduate-school experience. 3 This dissertation would not have been possible without financial support fromthe Na- tional Institute on Deafness and Other Communication Disorders in the form of a training grant to the Harvard-MIT Speech and Hearing Sciences Program. Additional support was provided by the Office of Naval Research and the National Science Foundation. The Speech and Hearing Sciences (SHS) Ph.D. Programhas provided the ideal multidis- ciplinary environment for me as a graduate student. I thank Susan Voss, a friend since our days in Brown Engineering, for introducing me to the SHS program. I am grateful to Nelson Kiang for his role in founding this program, and for our thought-provoking conversations. Bill Peake provided early encouragement and advice for my interest in carrying out my research with Laurel Carney at Boston University. He also maintained an interest in my work, which included his willingness to read a lengthy manuscript. John Rosowski, who also provided early encouragement for my working with Laurel, was a great role model during my time as a teaching assistant, continually demonstrating the importance of focusing on the students’ perspective. Larry Frishkopf provided much valuable advice on academic and career issues in his role as my academic advisor at MIT. I thank Bill Siebert for his willingness to talk with me about my work, which would clearly not have been possible without his pioneering work in this area. Jason Smith and Geoff Meltzner were instrumental in making the first two years of classes in the SHS program a fun, productive, and memorable experience. In addition, A.J. Aranyosi, Harold Cheyne, Brad Cranston, Greg Huang, John Iversen, Sridhar Kalluri, Courtney Lane, Chris Long, Martin McKinney, Janet Slifka, and Jason Sroka have helped to make graduate school an enjoyable experience beyond all the hard work. I have truly enjoyed working in the Boston University Hearing Research Center (HRC), which is where the day-to-day research for this dissertation