Computational Modeling of Sensorineural Hearing Loss

Abstract

Models of auditory signal processing have been used for many years to provide parsimonious explanations for how the normal auditory system functions at the physiological and perceptual levels, as well as to provide useful insight into the physiological bases for perception. Such models have also been used in many applications, such as automatic speech recognition, audio coding, and signal-processing strategies for hearing aids and cochlear implants. Here, an application framework is considered that is motivated by the long-term goal of using computational models to maximize the ability to apply physiological knowledge to overcome auditory dysfunction in individual patients. This long-term goal motivates the present ­chapter’s focus, which is on physiologically based computational models of auditory signal processing that have been used to explore issues related to sensorineural hearing loss (SNHL). Specifically, this chapter considers phenomenological signal processing models (rather than biophysical models) that predict normal and impaired peripheral responses to complex stimuli. These types of models are likely to be most useful in the specific applications needed to achieve the stated long-term goal, such as explaining the physiological bases for perceptual effects of SNHL, diagnosing the underlying physiological cochlear status of individual patients, and fitting and designing hearing-aid algorithms in a quantitative physiological framework.