Diagnosis of epileptic seizure from EEG signals by least squares method

Epilepsy is characterized by temporary and unexpected electrical deterioration in brain. EEG is preferred in diagnosis. There are many studies in the literature on EEG signals to differentiate between groups in epileptic and non-epileptic individuals. In this study, EEG signals were examined for predicting seizures for the three pre-, during, and post-seizures. The EEG was filtered by Singular Spectrum Analysis. Then the maximum amplitude wave in the EEG signal is fitted to the exponential curve by the nonlinear Least Squares method. The slope of the exponential curve is obtained as a feature. The obtained feature was examined statistically. As a result, there was a significant difference between the during seizure and postseizure, pre-seziure and post-seizure. There was no difference between pre-seizure and during seizure.