Coefficient of Variation Based Analysis of Coherence Resonance in Hodgkin-Huxley Neuron Model

We study the nonlinear response of the Hodgkin- Huxley model without external periodic signal to Gaussian white noise (GWN) and Ornstein-Uhlenbeck noise (OUN) as synaptic current respectively near the saddle-node bifurcation of limit cycles. The coherence of the system, estimated from the Coefficient of Variation of interspike interval of membrane potentials and spike trains, is minimal at certain noise intensity, so that the coherence resonance occurs. When GWN is served as synaptic input, coherence resonance occurs when noise intensity is greater than a certain threshold; when OUN is served as synaptic input, coherence resonance only occurs only with a moderate diffusion coefficient: too large diffusion coefficient will not generate coherence resonance. These findings further our understanding of how neurons respond to different types of noises.