Automated seizure onset detection for accurate onset time determination in intracranial EEG.

Objective: A novel algorithm for automated seizure onset detection is presented. The method allows for precise identification of electrographic seizure onset times within large databases of electrographic data.

Methods: The patient-specific algorithm extracts salient spectral and temporal features in five frequency bands within a sliding window of an electrographic recording.