A review on the stochastic firing behaviour of real neurons and how it can be modelled

Abstract

The types of spike trains recorded in real neurons from different parts of the brain, can either be completely random or bursty. Certainly, at very, high firing rates regular spike trains are observed. This paper examines the neurobiological spike trains observed experimentally and analytically and presents how they can be modelled and accounted for by using the biologically inspired Temporal Noisy-Leak), Integrator (TNLI) neuron model, with partial reset. The complete randomness or high firing variability, can be achieved for certain input parameter values at high firing rates, which results from the dendritic temporal summation of postsynaptic responses and the use of random synaptic inputs. It is also demonstrated that bursting behaviour can indeed be achieved, using the TNLI, which is a result of the use of random synapses and distal inputs. The firing variability is demonstrated by calculating the Coefficient of Variation (Cv) of the interspike interval (ISI) distribution and by observing the corresponding ISI histograms. © 1995, Springer Verlag. All rights reserved.