A Silicon Resonate-and-Fire Neuron Based on the Volterra System

Abstract—We propose an analog integrated circuit that implements a resonate-and-fire neuron (RFN) model based on the Volterra system. The RFN model is a simple spiking neuron model that exhibits dynamic behavior observed in biological neurons, such as fast subthreshold oscillation, post-inhibitory rebound, and frequency preference. The RFN circuit was derived from the Volterra system to mimic such behaviors of the RFN model. Through circuit simulations, we will show that our circuit is expected to be useful for largescale integrated circuit implementation of functional spiking neural networks since it acts as a coincidence detector and a band-pass filter at a single unit level.