Different architectures and techniques havedeveloped in the neuromorphic field to mimic andinvestigate the activity of biological neural networks.This paper presents a set of piece-wise linear approximationsof a two-dimensional Hindmarsh–Rose neuronmodel for digital circuit implementation to achievehigher speeds and lower hardware costs in large-scaleimplementation of the biological neural networks. Theperformance of the model was evaluated with a timedomain signal error. Synthesis and hardware implementationon a field-programmable gate array, as aproof of concept, indicates that the proposed modelreproduces several neuronal behaviors similar to theoriginal model with higher performance and considerablylower implementation costs.
A modified Izhikevich model for circuit implementation of spiking neural networks