Abstracts

Rationale: Localization of seizure onset is one of the aims of intracranial epilepsy monitoring. As recordings techniques allow observation of higher frequencies, the occurrence of these components during epileptic seizures have been observed. The link between those patterns and epileptic focus localization is still being investigated. In this study we assessed the observation of high frequency patterns with current method of recordings. Methods: Twenty consecutive patients with intracranial recordings arrays (subdural grids with strips) between April 2005 and June 2006 were analyzed for this study. Age ranged from 10 to 50 years (27±13 yo). Eleven patients were diagnosed with mesial temporal lobe epilepsy, seven with neocortical onset epilepsy and two were suspected of having bilateral onset. All partial seizures (simple and complex) were analyzed with short-time Fourier transform and with matching pursuit analysis with comparable parameters (4.096 sec window – 1.024 sec overlapping). We focused our interest on activities in the frequency range 40-150Hz.Results: There is no a priori knowledge of the pattern that high-frequency components can take during a seizure. Therefore the analysis of the time-frequency reconstruction was first done by visual comparison between the image reconstructed by both methods. Unlike MP decomposition, FFT reconstruction highly depends on the choice of parameters used. We examined 175 partial seizures. Patients experienced between 1 and 14 seizures during their stay with one outlier (N=86). Epileptic events were stereotyped within each patient. Two patients presented clear high-frequency content during the onset of their seizures, five patients showed moderate sign of high frequency activities with either very low amplitude or not strictly timed with the onset. Recordings from the other thirteen patients did not show specific high frequency content. Currently 4 out of 12 patients with positive outcomes (Engel I or II) had strong or moderate high frequency content around focus area. Only 2 out of 6 patients which had less favorable outcomes (Engel III or IV) had recorded high frequency patterns. One patient with high frequency components did not have resective surgery.Conclusions: The nature of high-frequency components in epileptic seizures is the subject of research, due to occurrence of those components early in the focal area of some epileptic events. Indeed as techniques allow us to now observe activity above 60Hz, the potential of those markers deserves attention. We show on a series of twenty patients that typical FFT analysis do not always offer a clear picture of those time-frequency events. Only a limited number of patients seem to exhibit specific EEG activities above 40Hz that can truly be distinguished from simple broad-spectrum components. No definitive conclusions can be drawn regarding the potential of high-frequency content for seizure localization in larger populations. Future studies might need to focus on sub-populations of epileptic patients. Supported by NIH grant NS 48222