Detecting Ictal and Interictal Condition of EEG Signal using Higuchi Fractal Dimension and Support Vector Machine

Abstract

Ictal and interictal periods are the most important condition which needed to find for detecting and predicting seizure condition in epileptic patients. Neurologist spends hours to analyze electroencephalogram (EEG) signals in order to find a certain pattern for diagnosing epilepsy. This manual interpretation has a high chance of error and very time consuming. In order to minimize mistakes, various studies have proposed a computer-based detection system to support the detection of ictal and interictal conditions. In this study, we extract the EEG signal pattern by using the Higuchi fractal dimension to classify the ictal and interictal conditions of EEG signals. The features are extracted from five EEG sub-bands, delta, theta, alpha, beta, and gamma band. Those features are then fed to support vector machine as the classifier using 10-cross folds validation. The experiment shows that the use of HFD and the quadratic kernel is suitable for ictal detection. While the use of cubic kernel and HFD is suitable for detecting interictal conditions.