Biological basis of hearing-aid design.

Abstract

We show that we can accurately model the auditory-nerve discharge patterns in response to sounds as complex as speech and ask how we may exploit this knowledge to test new strategies for hearing-aid signal processing. We describe the auditory-nerve representations of vowels in normal and noise-damaged ears. The normal representations are predicted well by a cochlear signal processing model originally developed by Carney (Carney, L. H. J. Acoust. Soc. Am. 93:401-417, 1993). Basilar-membrane tuning is represented by a time-varying narrow-band filter. Outer hair cell control of tuning is exerted by a nonlinear feedback path. We show that the effects of noise-induced outer hair cell damage can be modeled by scaling the feedback signal appropriately and use the model to test one strategy for hearing-aid speech processing. We conclude by discussing some aspects of future trends in biomedical engineering approaches to problems of hearing impairment.