An evolutionary optimization methodology of a low-power programmable Cochlear Implant

Hearing loss is a common problem of the modern society, the compensation of which requires auditory prosthesis. Cochlear Implants (CI) are implantable electronic devices which create auditory sensations by directly stimulating the auditory neurons. Low-power operation and wide-range parameter programmability are key features of CIs. This article proposes an evolutionary approach to optimizing a CI sound processor to enhance speech understanding and auditory experience. The novelty of this work consists in having the auditory perception incorporated into the objective function of the optimization algorithm. Auditory perception is objectively estimated over a Matlab-Eldo simulation platform, which performs transistor-level simulation of a low-power CI sound processor to generate stimulus signals, followed by perceptive signal reconstruction using a first-order extracellular stimulation model. The optimization objective is to maximize the performance similarity measure between the original and the reconstructed auditory signals.