Spiking Neural Networks - Models and Implementations: Exploring spiking neural network models and implementations for simulating biological neural networks and brain-inspired computing

Abstract

Spiking Neural Networks (SNNs) represent a class of artificial neural networks that mimic the behavior of biological neurons, offering a promising avenue for brain-inspired computing. Unlike traditional neural networks, which use continuous-valued signals, SNNs communicate through discrete, asynchronous spikes, enabling more efficient and bio-plausible computation. This paper provides a comprehensive review of SNN models and implementations, covering key concepts, architectures, learning mechanisms, and applications. We discuss various SNN models, including the spike response model, integrate-and-fire model, and the more biologically detailed Hodgkin-Huxley model. Additionally, we examine spike-based learning algorithms such as Spike-Timing-Dependent Plasticity (STDP) and its variants, which enable SNNs to learn and adapt to stimuli. Furthermore, we review hardware and software implementations of SNNs, highlighting neuromorphic hardware platforms and simulation tools. Finally, we discuss current challenges and future directions in SNN research, emphasizing the potential of SNNs in neuromorphic computing, cognitive modeling, and brain-machine interfaces.