Population oscillations in a discrete model of neural networks of the brain

Abstract

A discrete model of biological neural networks is used to find out how synchronized firing of neurons emerges in a randomly connected neural population. The objective is to understand the mechanisms underlying brain waves and to find and characterize conditions which support spontaneous switching from disordered to rhythmic population activity as in case of an epileptic seizure. The model is kept as simple as possible to achieve on one hand a fast performance of computer simulations of networks with up to 10, 000 neurons and to keep on the other hand an overview of parameter dependences. Dynamics of the model can be classified into different regimes: random fluctuations, rhythmic oscillations and silence. When the ratio of the inhibitory/excitatory connectivity is raised the system crosses from the fluctuating regime through the rhythmic oscillating region to the silence regime. Close to the boundary between the fluctuating and the oscillating regimes the network shows spontaneous bursting of high amplitude rhythmic oscillations, which is characteristic of epileptiform behavior. The simulation results are in agreement with recent theories saying that focal epilepsy after injury of the brain could result from axonal sprouting of GABAergic neurons in the injured region.