Epileptic activity terminates spontaneously, typically lasting only seconds or minutes. In addition, in many instances regional seizure activity stops simultaneously in all channels (Fig. 1). The actual mechanisms for this spontaneous regional termination are not well established. To study this phenomena under controlled conditions, we have studied burst termination in model neural networks of purely excitatory neurons.
Neuronal modeling to date has produced multi-compartment models of neurons with discrete dendritic, somatic and axonal properties. Some networks have been modeled based on either hippocampus or neocortex and studied to help explain burst generation and enhancement of excitation. To analyze common, non-specific properties of networks of different neocortical neurons, we have implemented reduced neuron models (Rinzel, 1985; Av-Ron, 1994). The neuronal compartments have been simplified and the interactions have been defined in terms of net excitation. This type of model allows for simulations of a large number of realistic neurons.
Our previous simulations (Kudela et al., 1999) have shown that propagation of synchronous epileptiform activity can be sustained in locally connected excitatory networks. In another earlier study (Kudela et al., 1997) we have shown that bursting activity in a loop of two excitatory neurons can be terminated by additional excitatory input.In this study we study epileptiform activity caused by the summation of excitation in small and large (up to 62,500 neurons) purely excitatory networks and examine the termination of this activity in these model networks.