A Phenomenological Model of the Electrically Stimulated Auditory Nerve Fiber: Temporal and Biphasic Response Properties

Abstract

We present a phenomenological model of electrically stimulated auditory nerve fibers (ANFs). The model reproduces the probabilistic and temporal properties of the ANF response to both monophasic and biphasic stimuli, in isolation. The main contribution of the model lies in its ability to reproduce statistics of the ANF response (mean latency, jitter, and firing probability) under both monophasic and cathodic-anodic biphasic stimulation, without changing the model's parameters. The response statistics of the model depend on stimulus level and duration of the stimulating pulse, reproducing trends observed in the ANF. In the case of biphasic stimulation, the model reproduces the effects of pseudomonophasic pulse shapes and also the dependence on the interphase gap (IPG) of the stimulus pulse, an effect that is quantitatively reproduced. The model is fitted to ANF data using a procedure that uniquely determines each model parameter. It is thus possible to rapidly parameterize a large population of neurons to reproduce a given set of response statistic distributions. Our work extends the stochastic leaky integrate and fire (SLIF) neuron, a well-studied phenomenological model of the electrically stimulated neuron. We extend the SLIF neuron so as to produce a realistic latency distribution by delaying the moment of spiking. During this delay, spiking may be abolished by anodic current. By this means, the probability of the model neuron responding to a stimulus is reduced when a trailing phase of opposite polarity is introduced. By introducing a minimum wait period that must elapse before a spike may be emitted, the model is able to reproduce the differences in the threshold level observed in the ANF for monophasic and biphasic stimuli. Thus, the ANF response to a large variety of pulse shapes are reproduced correctly by this model.

Visual Speech Benefit in Clear and Degraded Speech Depends on the Auditory Intelligibility of the Talker and the Number of Background Talkers

“…However, limitations of the vocoding technique used here should be acknowledged, as the technique simply simulates the consequences of removing TFS from the signal and filtering it into a number of discrete frequency bands. Other difficulties faced by CI users, such as spiral ganglion excitability ( Horne, Sumner, & Seeber, 2016 ), among others, are not simulated. In addition, the tone vocoding used in the present study did not limit the range of modulations extracted from each channel, meaning that there will be F0 related modulations in the extracted envelopes.…”

The contribution of visual information to the perception of speech in noise with and without informative temporal fine structure

“…The acoustic simulation used here simulates only the consequences of removing TFS from the speech signal and filtering the speech into a discrete number of frequency bands. Many other factors, such as the spread of electrical current along and across the cochlea (Cohen et al, 2003), are not simulated, and the primary sources of stochasticity (normal hearing: inner haircell/auditory nerve synapse, Sumner et al, 2002; cochlear implant: spiral ganglion cell excitability, Horne et al, 2016) are very different. Thus, the encoding of speech on the auditory nerve is expected to be very different between electrical and tone-vocoded inputs.…”

Applications of Phenomenological Loudness Models to Cochlear Implants

“…This reduction in efficiency is well explained by the neural “leaky integrator” model ( Miller et al, 2001 ). The ability with which neurons integrate charge on their membranes depends on the site of activation (dendrite, cell body or axon) and physical attributes of the neurons such as size and health, for example presence or absence of myelin ( Parkins and Colombo, 1987 ; Horne et al, 2016 ). These neural properties lead to the amount of PD change versus current change for equivalent loudness change being different at different absolute current amplitudes and PDs ( McKay and McDermott, 1999 ; Carlyon et al, 2005 ), and between different people and different electrode positions in the same person ( Schvartz-Leyzac and Pfingst, 2016 ).…”