Abstracts

Rationale: High-frequency oscillations (HFOs) recorded during intracranial EEG (iEEG) have been proposed as a biomarker of epileptic brain tissue (1). Gamma frequency activity (GA: 30 ?" 100 Hz) may also be a biomarker of epileptic brain. Some studies considered 'event-related' HFOs as normal and spontaneous HFOs associated with interictal epileptiform spikes as pathologic events (2,3). Yet, this dichotomy may not fully address the localization problem, and whether 'event related' activity in epileptic brain and non-epileptic brain are distinguishable remains unclear. Previous findings(4,5) suggest the possibility that physiologically induced HFOs in epileptic brain have spectral properties similar to pathologic HFOs. Here we studied GA recorded during cognitive task stimulation within seizure onset zone (SOZ) and all regions other than those (NSOZ), to build a better understanding of how GA within epileptic tissue differs from non-epileptic brain. Methods: We used iEEG recorded during encoding of words in 11 epilepsy patients to analyze GA events within SOZ and NSOZ. Patients were implanted with subdural and depth electrodes for seizure monitoring as part of clinical treatment for drug-resistant epilepsy. A list of twelve words were presented on a laptop (each for 1600ms) for subsequent recall. The EEG signals were sampled at 500 Hz. Individual bursts of GA were detected trial-by-trial (4). Spectral and time domain properties of each GA detection were determined. Detections in low gamma (30-55 Hz) and high gamma (65-100 Hz) frequency ranges were included in further analysis. Based on prior studies (3), the detections were first thresholded by the upper adjacent value (UAV) of maximum amplitude of detections of each electrode in each band to narrow them to outliers with highest maximum amplitude. The properties of thresholded detections during the word presentation interval were compared within SOZ and NSOZ electrodes in each subject, across all subjects, and in different brain structures. Results: We observed significantly higher maximum and mean amplitude (in 8/11 patients for the low GA and 7/11 patients in the high GA (p < 0.05)), and longer duration (in 7/11 patients for the low GA, and 8/11 patients in high GA, (p < 0.05)) in SOZ compared to NSOZ. When considering all electrodes from all patients the GA detections in SOZ had higher maximum amplitude (low GA, p < 0.001, high GA, p < 0.001), mean amplitude (low GA, p < 0.001, high GA, p < 0.001) and longer duration (low GA, p < 0.001, high GA, p < 0.001) than those in NSOZ. Conclusions: GA events recorded during cognitive task stimulation within the SOZ had higher maximum and mean amplitude, and longer duration than GA events within NSOZ.The results suggest that tasks induced GA may be useful for mapping both normal and pathological brain regions. References Biomark Med 2011; 5(5); 537-544 Epilepsia 2013; 54(2); 370-376 (2013) J Neurophysiol 2013; 110(8); 1958-1964 Brain 2014; 137(8); 2231-2244 Neurology 2015; 84(6); 602-608 Funding: This work was supported by funding from the National Institutes of Health (NIH: R01-NS092882 and R01-NS063039), Czech Republic Grant agency (P103/11/0933), and European Regional Development Fund - Project FNUSA - ICRC (CZ.1.05/ 1.1.00/02.0123).