A Refractory Period-Incorporated LIF Neuron Circuit for Scalable Large-Scale Integration

Over the past few years, Spiking Neural Networks (SNNs), acknowledged as the third generation of neural networks, have garnered substantial research attention owing to their distinctive temporal dynamics, diverse encoding mechanisms, and event-driven nature, which fundamentally differentiate them from conventional artificial neural networks. In this paper, we improve the RLIF neuron model, propose a circuit structure that emphasizes the refractory period characteristics of neurons, and complete the circuit design based on 180nmCMOS technology. The circuit can be used to study the dynamic behavior of SNN neural networks, and can be used as the basic unit of integrated analog pulse neural network circuits and systems.