EEG Signal Processing for Epileptic Seizure Detection

All patients presenting a possible epileptic disease should be evaluated with an EEG, because the presence of epileptiform activity during the evaluation of a patient with suspected epilepsy in numerous situations definitely supports the diagnosis. The proposed method involves using an open-source database of EEG signals, processing those signals and extracting specific features. Various classification techniques are used for a binary classification to distinguish between the classes. The results obtained are focused to identify the best type of signals and the best classifier based on the accuracy and sensitivity of the classification results. Once this discrimination is understood, the main goal in this field is the prediction of seizures.