Typical Absence Epilepsy Identification on EEG

This paper describes the methodology and results obtained from the classification of EEG signals in two groups: 1) Patients with typical absence seizure; 2) Patients with other kind of epilepsy or healthy. Three main techniques were applied to identify the morphological features from EEG signals in order to evaluate recordings without having to train a model using a database: Continuous Wavelet Transform, Competitive Neural Networks and Correlation. An interface was developed to include clinical information in order to create an auxiliary system for the identification of absence epilepsy. Data from 24 patients, with different types of epilepsy and non-epileptic, were analyzed, and all of them were correctly classified. The system can be used as auxiliary in the identification of typical absence epilepsy either in clinic or in education.