Abstracts

Rationale: Interictal epileptiform discharges may localize to the seizure onset zone, particularly if there is a single interictal focus. Even if there are multiple interictal foci, if the events are detected and clustered accurately, these clusters when localized separately provide prior hypotheses to guide analysis of seizure onset.Methods: We propose a clinical protocol for (1) automatic detection and clustering of interictal events, (2) physician review and source localization of those interictal events, (3) examination of seizure onset in both noninvasive dense array EEG and intracranial recordings to determine the utility of the interictal hypotheses.Results: In five patients examined in this preliminary study, interictal epileptic discharges were identified in each patient with the automatic algorithm, showing good agreement with clinician review. Clustering of spikes into typical groups was possible with the automatic k-means algorithm, allowing the clinician to form hypotheses from the interictal clusters for the likely zone(s) of seizure onset. Seizure onset was determined both from noninvasive dEEG source localization and from intracranial recording. Conclusions: Automated detection and source localization of interictal epileptiform activity can be a valuable tool when included in a physician s review. The results determined can be informative about the seizure onset and prepare a more informed hypothesis for the patient s future treatment plan.