Transition from local to global phase synchrony in small world neural network and its possible implications for epilepsy

Abstract

Temporal correlations in the brain are thought to have very dichotomic roles. On one hand they are ubiquitously present in the healthy brain and are thought to underlie feature binding during information processing. On the other hand large scale synchronization is an underlying mechanism of epileptic seizures. In this paper we show a possible mechanism of transition to pathological coherence underlying seizure generation. We show that properties of phase synchronization in the 2-D lattice of non-identical coupled Hindmarsh-Rose neurons change radically depending on the connectivity structure. We modify the connectivity using the small world network paradigm and measure properties of phase synchronization using previously developed measure based on assessment of the distributions of relative interspike intervals. We show that the phase synchronization undergoes a dramatic change as a function of locality of network connections from local coherence strongly dependent on the distance between two neurons to global coherence exhibiting stronger phase locking and spanning the whole network.