A very stochastic resonant neuron model

Abstract

We study the properties of a neuron model, the dynamical perceptron, under different stochastic inputs. This model is essentially a discrete time mapping, that can exhibit many neural features like activation threshold and pacemaker behavior, amongst others. Like real neurons, this model can show many different types of stochastic resonance. The addition of white noise causes autonomous stochastic resonance, oscillations with predominance of multiples of some fundamental period. The noise can also enhance subthreshold either periodic ( usual stochastic resonance) or aperiodic ( aperiodic stochastic resonance) external signals.