Hardware Verification: Determining the Parameters of the Modified Izhikevich Neuron Model with Genetic Algorithm

The nonlinear function in Izhikevich neuron model (IzNM) makes difficult the digital hardware realizations of the model, so this parabolic function has been transformed to piecewise linear (PWL) functions in the literature. Some coefficients have been identified in the PWL functions by utilizing the classical step size method, but the values of these coefficients depend on the sensitivity of the step size considerably. In this study, the coefficients of the PWL functions in the modified IzNM are determined by using Genetic Algorithm (GA). After the parameter determination, the modified IzNM is simulated with the parameters, which are determined by both classical step size and GA. Also, the original and modified IzNMs exhibiting “tonic spiking” and “tonic bursting” behaviors are realized with digital programmable device, namely FPGA. Thus, it is tested the utility of the intelligent search algorithms in the neuronal structures and verified the adaptability of their results to the hardware implementations.