Abstracts

Rationale: Spatiotemporal source localization of interictal spikes recorded in magnetoencephalography (MEG) and electroencephalography (EEG) may be useful for understanding origins and propagation of epileptic activities. Minimum norm estimates (MNE) is a new method for calculating spatiotemporal source distribution, and is believed to characterize the time-course of brain activities. Several case studies have suggested the usefulness of MNE for analyzing widespread spikes in presurgical evaluation of epilepsy. However, it has been still unclear whether MNE can represent possible spike propagation as recorded on intracranial EEG (IEEG). In the present study, we reconstructed source waveforms of MEG and EEG from MNE-based source distribution of interictal spikes, and compared them with IEEG in patients with intractable temporal lobe epilepsy. Methods: Simultaneous 306-channel MEG and 74-channel EEG were obtained from four patients with intractable temporal lobe epilepsy in Athinoula A. Martinos Center for Biomedical Imaging. High-resolution 3T MR images were acquired after MEG/EEG recordings. Segments containing epileptic spikes on both MEG and EEG were selected for analysis. In each spike, MNE-based source distribution was calculated from MEG and EEG, independently. A 3-layer boundary elemental model derived from each individual’s MRI was used for incorporating anatomical information into the source analysis. All patients underwent long-term IEEG recordings with an ipsilateral fronto-temporal coverage at Children’s Hospital after MEG/EEG recordings. We defined the locations of intracranial electrodes on the cortical surface obtained from each patient’s MRI, by using CT images including these electrodes. MEG and EEG waveforms were independently calculated from the MNE-based source distribution at each electrode site. Thus, reconstructed spikes on both MEG and EEG were obtained at “virtual” sensors on the cortical surface, corresponding to the locations of intracranial electrodes. These source waveforms were compared with the IEEG spikes at the corresponding sites. Results: In all patients, IEEG showed ipsilateral fronto-temporal involvement of interictal spikes. The temporal peaks of these spikes preceded the frontal peaks by approximately 10-100 ms. MEG/EEG spikes were distributed in the ipsilateral fronto-temporal area, and 10-16 MEG/EEG spikes were analyzed in each patient. In 69-100 % of reconstructed MEG spikes, the temporal peaks preceded the frontal peaks. The preceding temporal peaks were seen in 50-100 % of reconstructed EEG spikes. Other MEG/EEG spikes showed no time difference or preceding frontal peaks. The mean value of time delay in the frontal peaks of each individual patient ranged from 26.1-48.2 ms on MEG, and 10.4-58.7 ms on EEG. Conclusions: MNE-based source analysis of MEG and EEG spikes may represent the spike propagation appropriately as observed on IEEG in patients with temporal lobe epilepsy.