Why is stochastic resonance enhancement intrinsic in biological neurons?

Abstract

We claim that stochastic resonance enhancement is intrinsic in biological neurons and we try to back this claim by considering a single threshold detecting system. We consider this type of system configuration since a neuron, based on its empirical model, is essentially a single-threshold detector. We implemented two types of noise namely Gaussian white (GWN) and time-correlated noise (TCN). First, we show that stochastic resonance (SR) can be seen in this kind of system by looking at its performance (signal-to-noise ratio in this particular study) as you vary the noise's standard deviation and coherence. We then show that this SR can be enhanced at low frequency modulations by implementing refractory and excitatory regimes, similar to those seen in biological neuron firing, in the detection process and playing along with their combined effects. These results of improving SR in a system, parallel to a biological neuron, by incorporating a technique analogous to the empirical biological neuron firing model, suggests that indeed, SR enhancement is intrinsic in neurons.