Seizure Onset Zone Localization in Focal Epilepsy using stereo-EEG: SCSA and Visibility Graph complementary features

Despite how limiting the symptoms of Epilepsy are in daily life, over a third of epileptic conditions are drug-resistant. These seizures stem from a specific region of the brain: the Seizure Onset Zone (SOZ). The available treatment options include surgical resection or neurostimulation of the SOZ and both options require its accurate localization. In this scope, we aim to analyse stereo-electroencephalograph (s-EEG) data from a feature engineering perspective to propose reliable computer-aided SOZ localization algorithms. This paper investigates Semi-Classical Signal Analysis (SCSA) and Visibility Graphs (VG) features and establishes their complementarity. The study considers brain -scale signals simulated through the Epileptor model using The Virtual Brain (TVB). Results have been undeniably satisfactory and a valid contribution to Seizure Onset Zone localization.Clinical relevance- The paper contributes to the design of frameworks for assisting medical experts in Seizure Onset Zone localization in focal epilepsy.