Model based prediction of post epilepsy surgery

Abstract

In about 30% of epileptic patients, epilepsy is not controlled by medication. For some

of these patients surgery is an option. However, the surgery requires accurate

determination of the seizure onset zone. So the prediction of reliable epileptic brain

area is of crucial importance.

Here, a dynamic model is developed to predict the seizure onset zone and outcome of

surgery in patients. The model uses the inter-ictal ECoG data as a connectivity to

predict the seizure onset zone in the brain.

The examination of betweenness centrality of the node gives the strong correlation

with the seizure onset zone. Betweenness centralization of the network depends on the

seizure onset zone, if seizure onset is removed the betweenness centralization of the

network decreases. Increase in betweenness centralization of the network is also

associated with the seizure onset zone.

It is shown that the probability of seizure occurrence is less with decrease in

betweenness centralization ofthe network. Hence, the model can be used to predict

the seizure onset zone and outcome of surgery.