A phenomenological model of the neural response to pulsatile electrical stimulation: a novel approach and theoretical considerations

Cochlear implants stimulate the spiral ganglion cells (SGCs) with trains of charge-balanced current pulses. How the SGCs respond has often been studied with the leaky integrate-and-fire (LIF) model. However, while the LIF model partially reproduces how SGCs respond to monophasic pulses, which are not charge balanced, it does not reproduce how the opposite-polarity phases of charge-balanced pulses interact to reduce their efficacy, and nor does it reproduce the temporal distributions of the evoked spikes, or how they depend on the strength of the stimulus. To address these limitations, I extended the LIF model by adding an initiation period to spiking, delaying spike emission by a stochastic, stimulus-dependent duration, so that the temporal distribution of the model reproduces that of the SGC. During its initiation, a spike may be cancelled by anodic current, thus allowing opposite-polarity phases to interact, reducing the stimulus’s efficacy in a way that reproduces how the thresholds of SGCs depend on the delay between the phases of cathodic-leading biphasic pulses. A spike may only be cancelled by anodic current after it has been initiated by cathodic current, and thus, cancellation only occurs in response to pulses in which a cathodic phase precedes an anodic phase. Reversing the phase order of a pulse therefore changes how the phases interact. To investigate whether phase order has a similar effect in the real neuron, I analysed the phase plane of a biophysical point-neuron model and found that the qualitative description of excitation depends on phase order in a way that is consistent with the theory of spike cancellation. That the phase order of a biphasic pulse affects how excitation occurs has direct consequences for cochlear implant coding strategies, as it affects whether the interactions between consecutive biphasic pulses are facilitatory or inhibitory