Ghost stochastic resonance for a neuron with a pair of periodic inputs

Abstract

A small network like that proposed in [8] and [9] is studied when a pair of periodic signals are carried by the two different input neurons to verify the arising of ghost stochastic resonance phenomena in the third processing neuron. Suitable modifications of the stochastic leaky integrate-and-fire model are employed to describe the membrane potential of the input neurons while the processing neuron is modeled by means of a jump-diffusion process. A stochastic resonance behavior is detected for the processing neuron in correspondence with the "ghost" frequencies both in the harmonic and in the anharmonic case. The range of parameter values under which this behavior occurs is specified and an interpretation of the coincidence detection mechanism involved is provided. © 2007 Springer-Verlag Berlin Heidelberg.