Electronic Model of FitzHugh-Nagumo Neuron

摘要

Nowadays neural networks are widely used to solve the tasks of signal processing, control and identification, pattern recognition and etc. [1, 2]. Traditional way to investigate the neural network is based on a mathematical model of the network. Despite the powerful computers the modeling of the entire neural network using numerical simulation requires a lot of calculation resources and time. Due to parallel modeling of processes in each of neurons, the neural network model, which consists of electronic models of neurons, should work much faster. The most famous mathematical model of the real neuron, whose adequacy was validated experimentally, is Hodgkin-Huxley model [1, 3]. But the model of Hodgkin-Huxley is described using four nonlinear differential equations therefore it would be difficult to implement this model into hardware. There are few variants of simplified Hodgkin-Huxley model. The FitzHugh-Nagumo model is one of them [1, 3]. This model is often used for investigation into neurodynamical systems [4-8]. The implementation of the FitzHugh-Nagumo model into hardware is also possible [3-8], so using of this model as a part of the neural network model can be used to increase the speed of modeling.