Abstracts

Rationale: Intracranial depth macroelectrode recordings from patients with focal seizures demonstrate interictal and ictal high frequency oscillations (HFO, 100-450 Hz) that are mostly located within seizure onset zones. We report that HFOs do not consistently change in seizure onset channels immediately prior to focal seizures. Methods: Seven consecutive patients with intractable epilepsy and well-defined seizure onset zones were evaluated; 4 had mesial temporal epilepsy and 3 neocortical epilepsy. EEG was obtained from depth macroelectrodes, filtered at 500Hz and sampled at 2000Hz. One seizure was analyzed from each of the patients, and visually-identifiable HFOs were marked in seizure onset as well as control channels during 15 minute immediate pre-ictal periods. Spectral trends of the same pre-ictal segments were obtained for high frequency (HF, 100-200Hz) and very high frequency (VHF, 200-450Hz) bands. A linear regression was obtained for the power trends corresponding to pre-ictal durations of 1, 5 and 15min. We evaluated if the correlation was significantly different from 0 (p<0.01) to assess if the power showed a consistent increase or decrease before the seizure.Results: Using spectral analysis, 6/21 channels located in seizure onset zones of mesial temporal patients had significant positive pre-ictal HF or VHF regressions at one of the time intervals, compared to 6/26 control channels (outside the seizure onset zone). For neocortical patients, 3/8 seizure onset channels and 6/10 control channels had significant positive pre-ictal correlations. There were therefore no important differences between seizure onset channels and control channels for the HF nor for the VHF band. In addition there were no consistent negative pre-ictal correlations. Using the visual technique, marked HFOs were more frequently identified in seizure onset channels but did not show any tendency to increase or decrease over any of the pre-ictal periods; control channels did not show any tendency either.Conclusions: High frequencies in the 100-450Hz range are present in the pre-ictal period but do not appear to increase or decrease in seizure onset channels during the minutes before ictus regardless of the type of focal epilepsy. HFOs seem to rather increase abruptly during seizures. They may reflect an epileptogenic region but not an impending seizure and therefore may be of limited benefit for seizure prediction at least for the time scales evaluated for this study. (Supported by Canadian Institute of Health Research grant MOP-10189.)