Simple tunable generator of neuron-like activity

Abstract

This study aims to develop a new electronic realization of a neuron. We based our work on a previously constructed circuit and modified it to reduce the number of elements. All modeling was done using ngSPICE circuit simulator. For the new developed circuit current–voltage characteristics were constructed, then dependence of signal amplitude on resistance of potentiometer was measured. In this way, we found that oscillations arose as a result of Andronov–Hopf bifurcation like in the FitzHugh–Nagumo neuron. Various implementations of nonlinearity were considered and the dependence of the shape of the generated signal on the asymmetry of the nonlinear element was investigated. It is shown that with increasing asymmetry this signal becomes more similar to nonlinear spikes, and with decreasing asymmetry it approaches the signal of a strongly nonlinear van der Pol oscillator. The oscillation frequency is mainly determined by the resonant frequency of the circuit and weakly depends on asymmetry: the period slightly decreases with increasing asymmetry. As a result, we developed a new simplified electronic neuron of FitzHugh–Nagumo type simple for hardware realization and modification.