AAU Student Projects - visit Aalborg University's student projects portal
A master's thesis from Aalborg University
Book cover


Development and Objective Evaluation of an EAS Cochlear Implant Model

Author

Term

4. term

Publication year

2014

Submitted on

Pages

81

Abstract

Dette projekt udvikler en computermodel, der simulerer kombineret akustisk og elektrisk stimulering (EAS) hos brugere af cochleært implantat (CI). Den akustiske del bygger på MATLAB Auditory Periphery (MAP) modellen, som er tilpasset til at efterligne et høretab. Den elektriske del anvender en CI-model fra Neurelec/Oticon Medical. De to dele samles i Goldwyns punktprocesramme, som beskriver hørenervens spike-aktivitet. Vi kombinerer inputtene ved at antage superposition af neurale affyringsintensiteter – dvs. at de akustiske og elektriske bidrag lægges sammen. Den samlede model vurderes med neurogram similarity index measure (NSIM), som sammenligner mønstre af neural aktivitet, og vi undersøger EAS-relaterede problemer som tidsmæssig desynkronisering (responser, der ikke er justeret i tid) og place-mismatch (stimulering andre steder i cochlea end forventet ud fra frekvens).

This project develops a computer model that simulates combined acoustic and electric stimulation (EAS) for cochlear implant (CI) users. The acoustic pathway is modeled with the MATLAB Auditory Periphery (MAP) model, adapted to represent hearing loss. The electric pathway uses a CI model from Neurelec/Oticon Medical. Both parts are integrated in the Goldwyn point process framework, which represents auditory-nerve spike activity. We combine the two inputs by assuming superposition of neural firing intensities—that is, the acoustic and electric contributions add together. The complete model is evaluated with the neurogram similarity index measure (NSIM), which compares patterns of neural activity, and we examine EAS-related issues such as temporal desynchronization (responses not aligned in time) and place-mismatch (stimulation at different cochlear locations than expected from frequency).

[This abstract was generated with the help of AI]