Modeling auditory-nerve responses to electrical stimulation

Cochlear implants (CI) directly stimulate the auditory nerve (AN), bypassing the mechano-electricaltransduction in the inner ear. Trains of biphasic, charge-balanced pulses (anodic and cathodic) areused as stimuli to avoid damage of the tissue. The pulses of either polarity are capable of producing action potentials (AP) whereby the sites of initiation of the AP differ for the two polarities. A cathodicpulse triggers an AP in the peripheral axon, whereas an anodic pulse triggers an AP in the centralaxon. The latency difference between the APs initiated at the different sites is about 200μs, whichis large enough to affect the temporal coding of sounds and hence, potentially, the communication abilities of the CI listener. In the present study, two recently proposed models of electric stimulationof the AN [1, 2, 3] were considered in terms of their efficacy to predict the spike timing for anodic andcathodic stimulation of the AN of cat [4]. The models' responses to the electrical pulses of variousshapes [5] were also analyzed. It was found that, while the models can account for the ring rates inresponse to various biphasic pulse shapes, they fail to correctly describe the timing of AP in response to monophasic pulses. Strategies for improving the model performance with respect to correct AP timing are discussed. Eventually, a model that is able to account for correct spike timing in electrichearing will be useful for objective evaluation and improvement of CI stimulation strategies.