Stochastic Resonance and Coincidence Detection in Single Neurons

We demonstrate that a realistic neuron model expressed by the Hodgkin-Huxley equations shows a stochastic resonance phenomenon, by computing cross-correlation between input and output spike timing when the neuron receives both aperiodic signal input of spike packets and background random noise of both excitatory and inhibitory spikes. We consider that such a signal detection is realized because the neuron with active properties is sensitive to fluctuation caused by a sharp increase just after a sudden dip of excitatory noise spikes and a gradual decrease of inhibitory noise spikes. We also show that the model generates highly irregular firing of output spikes on the basis of the modulation detecting property.