A perceptual model of auditory deafferentation

Overexposure to intense sound produces temporary threshold shifts but permanent loss of afferent nerve terminals. Here, we present a vocoder designed to explore the perceptual consequences of this type of damage. The basic idea is that the spike train produced by an individual auditory afferent resembles a stochastically digitized binary version of the stimulus waveform and that the quality of the waveform representation in the whole nerve depends on the number of aggregated spike trains. Sounds were processed by filtering them into 10 adjacent frequency bands. For the signal in each band, multiple spike trains were then obtained in an attempt to mimic the different representations of that signal conveyed by different auditory afferents innervating a given cochlear region. The aggregated spike train was multiplied by the original signal to obtain an acoustic version of the simulated nerve waveform. Tone-in-noise and speech-in-noise perception tests were performed by young, normal-hearing listeners using different numbers of afferents per frequency band. Results support that deafferentation impairs perception in noise more than in quiet. The proposed vocoder may be extended to model other types of hearing damage and to guide the design of hearing aids and cochlear implants.