Noise in integrate-and-fire models of neuronal dynamics

Abstract

The sequence of action potentials produced by a neuron is best characterized in terms of a stochastic point process. In this contribution we will be primarily concerned with different variants of stochastic leaky-integrator models for the membrane potential. The point process representation is then achieved by the fast passage time transformation of the underlying membrane potential model. Different sources of the noise in the diffusion neuronal models resulting from the stochastic leaky-integrator model will be discussed.