Implementation of FitzHugh-Nagumo Neurons using Nanoscale VO₂ Devices

Abstract

This study focuses on exploiting phase-transition properties in VO2 to develop a brain-inspired FitzHugh-Nagumo neuron. A key feature of our implementation is achieved through a self-coupling synapse that creates a neuron capable of generating complex firing patterns, such as mixed-mode oscillations. These oscillations hold promise for solving time-dependent computing problems. We study the advantages of coupling FitzHugh-Nagumo oscillators in terms of stability and circuit simplicity compared to Kuramoto-based relaxation oscillators. We show how they can excel in solving computational tasks, specifically in pattern recognition applications.