Physiologically-based predictors of speech intelligibility

Accurate models of normal and impaired neural representations of sound are useful tools in understanding how acoustic stimuli are encoded in the brain, predicting speech intelligibility, and developing and testing speech processing schemes for hearing aids. In this talk I will review recent developments in modeling the effects of hair cell impairment on neural responses to speech stimuli in the auditory nerve and primary auditory cortex. Several important cochlear nonlinearities, such as compression and suppression, the shift in tuning with sound pressure level, and the component-1/component-2 transition at very high sound pressure levels, have been incorporated into the latest models of the auditory periphery. These properties of cochlear processing prove to be important not only in forming the normal neural representation of sound but also in determining the degradation of the neural representation in cases of hair cell impairment. We have evaluated these models by using them to predict the effects of presentation level, hearing loss and amplification on speech intelligibility. The models are able to predict both the effects of audibility on speech intelligibility and the "roll over" in speech intelligibility at high presentation levels for normal hearing listeners and for hearing impaired listeners using hearing aids. I will also discuss recent work in my lab aimed at understanding the contributions of temporal fine structure and envelope cues to speech perception.

Speaker: Ian C. Bruce, Mcmaster University

Room 3105