Abstracts

Rationale: This study aims to determine the association between surface and intracerebral EEG seizure-onset patterns in patients with lesional focal epilepsy, to know if surface patterns can predict pathology and intracerebral patterns. Methods: From 2003 to 2013, 60 consecutive patients with refractory focal epilepsy and a MRI-visible lesion underwent first a surface and then an intracerebral electrodes EEG recording (SEEG) for presurgical evaluation. We identified 38 patients in whom the seizures started in contacts of depth electrodes placed in lesional/perilesional areas. Two observers independently reviewed the seizure-onset morphology on surface and SEEG. Seizure-onset was defined as the first change in the EEG that led to seizure discharges without returning to background. Seizures were separated into clinical or electrographic (without clinical manifestations). Seizure-onset patterns were divided into 5 types on the surface and 7 on the SEEG. Surface EEG: rhythmic activity between 1-13 Hz, sinusoidal fast activity, repetitive epileptiform discharges, background attenuation and artifacts; SEEG: low-voltage fast activity (LVFA), low frequency high-amplitude periodic spikes, sharp activity ≤13 Hz, 2-4 Hz spike-and-wave activity, burst of high-amplitude polyspikes, burst suppression and delta brush. Only the predominant seizure-onset patterns of surface EEG and SEEG were compared. Clinical seizures were matched based on semiology and EEG seizures were matched based on topography. Results: Among the 38 patients, 16 were diagnosed with focal cortical dysplasia, 13 with mesial temporal sclerosis/atrophy, 3 with periventricular nodular heterotopia, 3 with polymicrogyria, 2 with cortical atrophy and one with tuberous sclerosis complex. Five of the patients had two matched seizure patterns, yielding a total of 43 matched seizures. Each intracerebral seizure-onset pattern corresponded to multiple surface patterns (Table 1). LVFA and sharp activity ≤13 Hz were associated with almost all surface patterns, while low frequency high-amplitude periodic spikes appeared on the surface as rhythmic activity or repetitive epileptiform discharges but were never associated with sinusoidal fast activity, attenuation or artifacts. Spike-and-wave and delta brush SEEG patterns appeared on the surface only as repetitive epileptiform discharges. In relation to pathology (Table 2), LVFA corresponded with sinusoidal fast activity and artifacts predominantly in FCD, and low frequency high-amplitude periodic spikes with rhythmic activity predominantly in mesial temporal sclerosis/atrophy. Conclusions: This study shows that some surface seizure-onset patterns exclude the presence of some intracerebral patterns, e.g. sinusoidal fast and artifact patterns exclude low frequency high amplitude periodic spikes, typical of mesial temporal sclerosis. Surprisingly, LVFA appears on the scalp with different patterns but not with attenuation. Funding: This project was supported by grant FDN 143208 of the Canadian Institute of Health Research.