Prediction of EEG Dynamics under Epilepsy using Machine Learning Methods

Epilepsy is a neurological disorder that leads to disturbances in normal brain activity with characteristic recurring seizures. Not all types of diseases are amenable to medical treatment and surgery is associated with minor risks. In order to make life easier for patients with epilepsy, there is a need to predict the onset of epileptic seizures. In this paper, prediction of seizures based on electroencephalography (EEG) data is performed using machine learning methods. The task is to make classification of EEG segments with interictal and preictal activity. Several classification algorithms were tested, such as logistic regression, support vector machine, gradient boosting and CatBoost. As a result, for some patients, it was possible to solve the problem using a smaller number of EEG recording channels, and the number of predicted epilepsies was 70% on the test sample.