Functional Properties of Resonate-and-Fire Neuron Circuits for Bio-Inspired Chemical Sensor Array

Kazuki Nakada, Katsumi Tateno, Hatsuo Hayashi, Kiyonori Yoshii

Abstract: The purpose of our project is to develop brain-inspired chemical sensor arrays from physiological, theoretical, and engineering points of view. In the previous work, a computational model for chemical sensor arrays has been proposed based on physiological properties of mouse taste bud cells (TBCs). The model consists of three functional parts: the chemical sensor, the random pulse generator, and the stochastic synchronizer. The chemical sensor array based on the computational model detects the concentration of chemical substances as the degree of stochastic synchronization. For practical implementation, we compare the probabilistic firing properties of an array of resonate-and-fire neuron (RFN) models and those of an array of leaky integrate-and-fire neuron models in order to consider their feasibility as the second part of a chemical sensor array. We further investigate stochastic synchronization in uncoupled integrated circuits implementing the RFN model in order to confirm if they act as the third part of a chemical sensor array on a practical hardware platform.