Automatic Identification of Epileptic Focus on High-Frequency Components in Interictal iEEG

Abstract

The localization of the seizure focus affected by epilepsy is crucial for epilepsy treatment due to observing long-term interictal intracranial electroencephalogram (iEEG) for categorizing the patterns of seizures by the neurological experts. Therefore, a computer-aided system based on machine learning method for automatic localization of focal patterns is promising future. In this study, we presents a filter-bank entropy-based feature-extraction approach in high-frequency components to detect epileptic focus, which consider a valid biomarkers to guide epilepsy surgery. The experimental results on real-world interictal iEEG recorded from eight patients demonstrate that our proposed method can achieve average AUC 0.79, which can reduce the workload of clinical experts for detection of epileptic focal.