Stochastic Resonance Of Neuronal Models With Threshold

Stochastic resonance(SR) is a phenomenon, where the response of a system to a weak periodic signal is enhanced at a nonzero level of noise. In neuronal information processing, SR of neuronal system with threshold driven by external period input and background noise are of great interest. In general, SR occurs during neuronal spiking driven by noise and period input which is illustrated by the signal-noise ratio(SNR), the mean interspike interval(ISI) and the amplitude gain versus the input frequency, the intensity of noise,the correlation time and so on are nonmonotomous.This article mainly discussed two aspects of SR of neuron models with threshold. On one hand, SR of an integrate-and-fire neuronal model with threshold subject to white background noise and input signal is investigated.The probability density function of ISI and the mean ISI are obtained by the method of mirror image. Numerical simulation shows that the evolution of the peak height of probability density function with regard to the noise intensity and the mean ISI versus the input frequency is nonmonotomous. Those show that SR occurs during neuronal spiking driven by background noise and periodic input. On the other hand, most researches only have taken into account the channel noise and they didn't consider the synaptic noise. In our work, stochastic resonance of an integrate-and-fire neuron model with threshold subject to synaptic noise and input signal is investigated. The amplitude gain of the output signal is obtained by the method of average. Numerical simulation shows that the evolution of the amplitude gain with regard to the input frequency is nonmonotonous in different intensity, correlation time and asymmetry of the synaptic noise. Those show that SR occurs during neuronal spiking driven by synaptic noise and period input.