Stimulation location encoding on the spike train of neuron models with passive dendrite

Abstract

Spike is the basic unit in the neuron communication, and different selections of the stimulation locations on the neuron might cause different spike trains, which infers that the spike trains may determine the information of the stimulation locations. The research on this subject deserves intensive attention, whether by numerical experiments or by electrophysiological ones. In this article, to answer the question of how does the spike train encode the stimulus location, by combining the cable model with the leaky integral firing model, a new neuron model called leaky integral firing model with passive dendrite is reconstructed from two levels (i.e., space and time) and in three forms (i.e., the conceptual model, the circuit model, and the mathematical model). Two types of stimulation are performed on this new model, which contain the constant electrode current and the synaptic one, where the latter is also divided into the excitatory current and the inhibitory one. Four coding ways are employed to encode the spike train, among them, by numerical experiments and some theoretical verification, it is shown that the first-to-spike-time coding method is the best one, which could clearly reflect the information of the stimulus position. To be more specific, the closer the stimulation location is to the axon hillock, the shorter the first-to-spike-time is. The neuron model proposed in this paper and the relating encoding methods for the stimulus location could also be applied to the brain-computer interface or constructing new types of neural networks.