Noisy neural nets exhibiting epileptic features.

On the basis of our previous studies of noisy neural nets we propose a model for the explanation of epileptic phenomena. Our neural net model is capable of exhibiting epileptic features if the number of spontaneously firing neurons is periodically increased beyond a certain threshold. Some alternative epileptogenic mechanisms are also discussed. The epileptic behavior of the neural net is determined by a combination of certain parameters of its phase diagram. The general features of the model are consistent with several experimental observations and explain some poorly understood clinical phenomena. The differences between normal and epileptic neural nets are explained in terms of the structural properties of the model.