Analysis of epileptic seizure predictions based on intracranial EEG records

Abstract

Epilepsy affects to more than 50 million people in the world. This disease reduces the quality of life of patients and their families due to the constant danger of sudden convulsions or loss of consciousness. Thus, it is important to have automated seizure prediction systems that alert to patients about this risk. Many methods and techniques have been proposed in the last years to address this problem. However, further research is needed to improve the efficiency and adaptability of current systems. Hence, this work analyzes several configurations for a seizure prediction system based on spectral wavelet decomposition of electroencephalogram signals. The evaluation of this systems shows that a few electrodes can predict seizures with an accuracy of 99.9% and a sensitivity of 99.8%. We are convinced that studies like this will definitively help to improve the quality of live of people suffering from epileptic seizures.