Firing statistics of inhibitory neuron with delayed feedback. II. Non-Markovian behavior

TL;DR

This paper demonstrates that the sequence of inter-spike intervals in a simple inhibitory neuron with delayed feedback exhibits non-Markovian behavior, meaning future firing depends on an indefinite history of past impulses.

Contribution

It provides an exact analytical proof that the inter-spike interval sequence cannot be modeled as a finite-order Markov process in this neural setup.

Findings

Output ISI distribution depends on the entire past history.

The process cannot be simplified to a finite-order Markov chain.

Analytic expressions for conditional probabilities are derived.

Abstract

The instantaneous state of a neural network consists of both the degree of excitation of each neuron the network is composed of and positions of impulses in communication lines between the neurons. In neurophysiological experiments, the neuronal firing moments are registered, but not the state of communication lines. But future spiking moments depend essentially on the past positions of impulses in the lines. This suggests, that the sequence of intervals between firing moments (inter-spike intervals, ISIs) in the network could be non-Markovian. In this paper, we address this question for a simplest possible neural "net", namely, a single inhibitory neuron with delayed feedback. The neuron receives excitatory input from the driving Poisson stream and inhibitory impulses from its own output through the feedback line. We obtain analytic expressions for conditional probability density P(t_{n+1}| t_n,...,t_1,t_0), which gives the probability to get an output ISI of duration t_{n+1} provided the previous (n+1) output ISIs had durations t_n,...,t_1,t_0. It is proven exactly, that P(t_{n+1}| t_n,...,t_1,t_0) does not reduce to P(t_{n+1}| t_n,...,t_1) for any n>=0. This means that the output ISIs stream cannot be represented as a Markov chain of any finite order.