Classifying the Epilepsy Based on the Phase Space Sorted With the Radial Poincaré Sections in Electroencephalography

Background: Epilepsy is a brain disorder that changes the basin geometry of the oscillation of trajectories in the phase space. Nevertheless, recent studies on epilepsy often used the statistical characteristics of this space to diagnose epileptic seizures. Objectives: We evaluated changes caused by the seizures on the mentioned basin by focusing on phase space sorted by Poincaré sections. Materials & Methods: In this non-interventional clinical study (observational), ۱۹ patients with generalized epilepsy were referred to the Epilepsy Department of Razavi Hospital (Mashhad, Iran) between ۲۰۱۸ and ۲۰۲۰, which their disease had been controlled after diagnosis and surgery. In evaluating the effects of this disorder on the oscillation basin of the EEG trajectories, we used the MATLAB@ R۲۰۱۹ software. In this computational method, we sorted the phase space reconstructed from the trajectories by using the radial Poincaré sections and then extracted a set of the geometric features. Finally, we detected the normal, pre-ictal, and ictal modes using a decision tree based on the Support Vector Machine (SVM) developed by features selected by a genetic algorithm. Results: The proposed method provided an accuracy of ۹۴.۹۶% for the three classes, which confirms the change in the oscillation basin of the trajectories. Analyzing the features by using t test also showed a significant difference between the three modes. Conclusion: The findings prove that epilepsy increases the oscillations basin of brain activity, but classification based on the segment cannot be applicable in clinical settings.

Keywords:

Electroencephalography , Epilepsy , Decision trees