Computational modeling of the dependence of kindling rate on network properties

Abstract

The dependence of the rate of kindling on network properties, such as the number of neurons, number of stored memories, and the number of neurons used to store each memory, is studied through computer simulations of an appropriate neural network model for kindling of focal epilepsy. Simulations are performed for models of both chemical and electrical kindling. Larger and more complex networks are found to take longer time to kindle, as observed in experiments. The nature of the dependence of the kindling rate on network properties is somewhat different between the two types of kindling. A simple analysis of the process of chemical kindling is presented, which provides a semi-quantitative explanation of the behavior observed in our simulations. This analysis also shows that our main conclusions about the dependence of the kindling rate on the size and complexity of the network are independent of some of the assumptions made in our modeling.