In our paper we show, how one can obtain special class exitable systems with more complex dynamics. We discuss principal features of such models. For specific example, derived from complex nehpron model, we demonstrate new features of noisy dynamics, such as two maxima of regularity.
In order to treat the effect of subthreshold dynamics on noisy neuron behavior we focus on parameters region of FitzHugh-Nagumo
model close to the so called canard-explosion. Such parameter region corresponds to transition from excitable regime to continuous spiking. We observe the number of noise-induced effects, such as (i) noise-induced stabilization of firing frequency; (ii) noise-induced
suppression of spiking; (iii) noise-induced chaos. We show that for small ensemble of resonator-type neurons activated by noise there is the global maxima of firing frequency at some optimal noise intensity. The underlying mechanisms of such behavior are closely related to noise-activated subthreshold dynamics.