Abstracts

Rationale: Although seizures are defined as sudden interruptions of normal brain function, growing evidence suggests that the process by which the brain transitions into a seizure is not abrupt. We investigated EEG changes occurring at the transition to seizure in pharmacoresistant focal epilepsy, and whether these changes differ across distinct underlying pathologies. Methods: Forty consecutive patients with pharmacoresistant lesional focal epilepsy undergoing intracranial EEG investigations were included in this study. For each patient, representative seizure types were analyzed. For each seizure, two 8-sec sections were selected from the unfiltered EEG: baseline, ending 30 sec prior to the first EEG change suggestive of seizure activity; immediate preictal, ending at the first EEG change suggestive of seizure activity. In each section, standard frequency bands (delta, theta, alpha, beta and gamma) and high-frequency oscillations (ripples and fast ripples) were assessed using spectral analysis and visual inspection, respectively. Frequency band data were compared between the baseline and immediate preictal sections, in the entire set of contacts and in four contact subsets: non-lesional/non-seizure-onset zone (SOZ); lesional only; SOZ only; lesional/SOZ. These analyses were performed for the entire patient sample and for three subgroups, defined according to underlying pathology: mesial temporal sclerosis, local/regional atrophy, and malformations of cortical development. Results: In the entire sample, there was a significant increase in each EEG frequency band activity from the baseline to the immediate preictal section. For standard frequency bands, the preictal increase in spectral power was 22.1% for delta, 18.0% for theta, 12.0% for alpha, 7.7% for beta, and 6.1% for gamma (all p<0.001). For high-frequency bands, the preictal increase in the percentage of time occupied by high-frequency oscillations per section was 35.7% for ripples (p<0.001) and 14.6% for fast ripples (p<0.05). Preictal increases in each EEG frequency band were widespread, with significant changes in lesional and/or SOZ contacts, as well as non-lesional/non-SOZ contacts. Widespread preictal changes were observed in all pathologies, although they were more pronounced in patients with mesial temporal sclerosis. Conclusions: Changes across the entire EEG bandwidth precede the onset of focal seizures by a few sec at least, independently of the underlying pathology. The changes include the focus as well as remote regions. Considering our conservative approach to define the seizure onset (first EEG change suggestive of seizure activity), these results are unlikely due to inclusion of ongoing seizure activity in the preictal section. Rather, they could represent a facilitating state of the brain, which enables a susceptible region to generate seizures. Elucidating mechanisms underlying these changes could lead to an improved understanding of seizures and epilepsy. This study was supported by the American Brain Foundation, the American Epilepsy Society, the Epilepsy Foundation, and the Canadian Institutes of Health Research.