Abstracts

Rationale: Children with benign Rolandic epilepsy (BRE) present with a characteristic electroencephalogram showing high-voltage spikes or spikes and waves in the centrotemporal region. The term “benign” was initially adopted since the typical variant is not accompanied by developmental deficits. In recent reports, a “malignant” variant of Rolandic epilepsy with similar electroencephalographic features and atypical clinical evolution has been described. In this study we used MEG recordings to address whether the pattern of propagation of interictal perirolandic discharges differs in patients with the benign and the atypical variant of Rolandic epilepsy. Methods: Interictal MEG recordings from 9 patients were retrospectively reviewed after the clinical profile of these patients was classified as either typical/benign (6 cases) or atypical/malignant (3 cases). A minimum of 20 individual spikes, simultaneously present in EEG and MEG, were selected for review from each patients’ recording. A 400-ms MEG epoch centered at the peak of each spike was then analyzed using the equivalent current dipole (ECD) model. The resulting intracranial sources were estimated at 3 phases of the discharges: rising period, peak, and propagation. The topography and pattern of spike propagation (regional, intralobar, intrahemispheric or interhemispheric) was then described for the 2 groups. Results: Out of the 3 segments analyzed for each discharge, only the last one (which characterizes the pattern of propagation) differed between the 3 cases of “malignant” and the 6 patients with “benign” Rolandic epilepsy. Early segments of the discharges (rising period and peak) were localized in the perirolandic structures for all patients. In the benign group, the estimated sources corresponding to the last segment of the discharges were always localized within the same region defined by the sources that accounted for the previous 2 segments of the spike (i.e., intralobar or perirolandic propagation). For the 3 cases with the atypical variant, activity sources corresponding to the last phase of the spike discharge were found in the insula (2 cases) and in the superior temporal gyrus (1 case). Conclusions: Our results indicate that the pattern of propagation of the interictal perirolandic discharges is more focal in the benign form of Rolandic epilepsy (i.e., affecting only perirolandic structures) and is more distributed in cases of malignant Rolandic epilepsy. The spatiotemporal resolution of MEG is especially well suited to characterize the subtle differences in the topography of spike propagation in these two variants of Rolandic epilepsy.