Spontaneous Transitions in Deterministic Networks

The neural assemblies undergo spontaneous changes between various dynamical states characterized usually by spiking or bursting at a single neuron level. These microscopic states contribute to a global neural dynamics that may be measured in a form of electric signal referred to as a local field potential. Here, we present a model neural network composed with nodes exhibiting autonomous spiking dynamics. We show that under a particular coupling configuration and slight mismatches between the nodes, the neural network exhibits deterministic transitions between two possible configurations of clusters. The clusters, composed of two neurons each, differ in internal (always chaotic) dynamics as well as in synchronization properties. Such clusters features may contribute to a temporal increase or decrease of local field potential in the neural network, and thus give an insight into the possible mechanisms of the spontaneous brain transitions. We consider two different models for nodes, namely, forced FitzHugh-Nagumo equations and Rulkov map, and show that the presented results are node-type independent. Finally, we propose a mechanism explaining the origin of these transitions.