Auditory evoked potentials are gaining popularity in assessing supra-threshold coding ability in the auditory brainstem. Because these aggregate population responses contain synchronized information converging from a large number of auditory-nerve fibers, they are particularly promising in diagnostics of noise-induced cochlear neuropathy, a newly discovered hearing deficit that affects the number and types of auditory nerve fibers available for a robust signaling of sound to the auditory brainstem.
In this project, we plan to bridge the gap between animal physiology and human applications via the development of a computational model of the auditory periphery that includes single-unit models of the auditory nerve, cochlear nucleus and inferior colliculus into a model for human auditory evoked potentials. Particularly, we plan on improving the adaptation properties of the single-unit model components and determine the weights of different contribution sources to match recorded population responses. After studying how different combinations of hearing deficits affect the source generators of auditory evoked potentials, we can use this knowledge to design auditory evoked potential metrics that are more precisely diagnosing subcomponents of hearing loss. Experimental validation of the model simulations is based on recordings of subcortical-steady state responses in normal and hearing-impaired listeners.
A Priority Project of