Phenotyping and Computational Modeling of Diverse Forms of Genetic Hearing Loss
* Presenting author
Computational models of auditory processing that can incorporate pathology may be helpful in understanding the effects of hearing impairment and in the development of improved devices for those with hearing loss, such as hearing aids and cochlear implants. However, incorporating pathology into computational models to explain a human subject’s experimental data is typically problematic due to a lack of definitive knowledge about the pathology that they have.A number of diverse forms of genetic hearing impairment have been molecularly characterized from the genetically isolated human population of Newfoundland, Canada. Affected members of these families form a useful pool of subjects to target for computational modeling, as within each extended family the specific genetic mutation leads to a similar pattern of pathology and resulting pathophysiology and perceptual deficits, but the types of hearing loss are very different between families due to different gene mutations.We are developing computational models incorporating the suspected cochlear pathology for each subject, based on psychophysical tuning curves and DPOAE growth functions, in addition to routine audiometric measures. The model for each subject will be validated via quantitative predictions of advanced electrophysiological recordings (ABR and ECochG) and word perception in quiet and noise for that subject.