Development of an in-silico framework for the engineering and evaluation of future optogenetic cochlear implants

by Lakshay Khurana
Doctoral thesis
Date of Examination:2024-05-15
Date of issue:2024-09-27
Advisor:Prof. Dr. Tobias Moser
Referee:Prof. Dr. Hansjörg Scherberger
Referee:Prof. Dr. Tim Gollisch
Persistent Address: http://dx.doi.org/10.53846/goediss-10750

 

 

Files in this item

Name:Dissertation_Khurana.pdf
Size:15.0Mb
Format:PDF

This file will be freely accessible after 2025-05-14.

Abstract

English

Hearing loss is a prevalent global challenge with profound cognitive, psychological, social, and economic implications. Traditional cochlear implants (CIs)—neuroprosthetic devices to partially restore hearing—stimulate the spiral-ganglion neurons (SGNs) with electrical pulses via an electrode array at a high stimulation rate. The wide spread of current and channel interactions lead to a poor understanding of speech and limited appreciation of music. Optogenetic cochlear implants (oCIs) represent a promising alternative, leveraging light-sensitive proteins (opsins) to activate the SGNs with spatially-confined light pulses. While several pre-clinical studies have shown promising results for improved spectral resolution, the question of potential improvements in speech perception remains elusive. This dissertation presents an exploration of the engineering and evaluation of the oCIs through a computational framework development, integrating models of sound processing, spatial spread in a human cochlea, and neuron stimulation. The modular structure of the framework allowed to simulate electrical and optical stimulation for an audio dataset in a fairly comparable manner. The spike patterns were compared to the sound spectrograph using a similarity measure based on the structural similarity index. The results demonstrate the superior spectral resolution of oCIs and offer an approach to evaluate the impact of lower stimulation rates that could arise from opsin kinetics and/or energy budget. The model indicates that the increased number of spectral channels (64 instead of 20) can compensate for a lower optical stimulation rate. Further, the oCIs are predicted to provide better hearing than the electrical CIs in noisy environments. By employing computational modelling approaches, this dissertation attempts to guide and accelerate the development and translation of oCIs.
Keywords: optogenetics; cochlear implants; hearing loss; spectral selectivity; computational model; neuron model; sound coding; ray tracing; spiral ganglion neurons; optical cochlear implants