Abstracts

Rationale: The purpose of this study is to compare the results of using an automated detection method of identifying interictal epileptiform activity against the standard method of human (manual) visual identification of this activity using dense-array EEG. An automated detection method should use information from all channels in order to identify this activity. During manual identification of interictal epileptiform activity it is likely that some EEG channel's information will not be observed and thus provide inconsistent results or lost information. Methods: We acquired EEG from five epileptic patients using a 256-channel sensor configuration. Portions of the file that were artifact-free were selected to be included in this analysis. The EEG recordings were then processed manually for interictal epileptiform activity and processed through an automated detection algorithm that identified interictal epileptiform activity. A neurologist then verified the detection rates of the manual and automated techniques for validity. Results: We found that the automated detection algorithm identified interictal epileptiform activity more consistently then the manual visual identification method. We also found that the time of processing was less with the automated detection method then with manual detection. Conclusions: Automated detection of interictal epileptiform activity can be a valuable tool for working with this patient population. Results of processing with this technique are more consistent and require less human processing time. This decrease in time of processing could be very valuable in the pre-surgical planning for this patient population and provide an avenue to quickly understand the epileptiform activity following an EEG exam.