Abstracts

Rationale: Recent reviews suggest simultaneously recorded MEG and EEG may maximize the advantages of each for presurgical epilepsy evaluations compared to either modality alone. EEG potential differences and MEG magnetic fields have similar linear functions of source strength and nonlinear functions of the source location. They are generated by the similar neurophysiological events (dendritic depolarizations). Both EEG and MEG have high temporal resolution but MEG can be limited by head movement and EEG signal can be distorted by skull impedances and other cranial tissues. At the source level, MEG is more sensitive to tangential sources whereas EEG is more sensitive to radial sources. The goal of this study was to determine the value of simultaneous MEG and dense array EEG (dEEG) for characterization of the epileptogenic zone in pediatric intractable epilepsy. Methods: We recorded simultaneous MEG with 275 sensors and dEEG with 256 or 128 electrodes, depending on the patient's head circumference, from 4 pediatric patients who were undergoing non-invasive epilepsy presurgical evaluation. MEG and dEEG recordings were synchronized within 1/1000 second level accuracy. The source estimations for both ictal onsets and interictal discharges were assessed with CURRY 7 software using 5 different algorithms including ECD, MUSIC, MNE, sLORETA and SWARM within the frequencies determined by wavelet time frequency analysis. The boundary elemental model (BEM) segmented from each patient's own MRI was the head model. The 2 patients who had seizures during MEG/dEEG simultaneous recording underwent respective surgery. One patient underwent invasive presurgical evaluation using intracranial EEG prior to respective surgery. The results from MEG/dEEG source estimations and intracranical EEG findings were compared. Results: Ictal events were captured from 2 patients, and interictal discharges were observed from all 4 patients. Ictal onset in MEG were detected approximately 150-500msec earlier than dEEG onset determined by the spectral power changes. For the patient who underwent intracranial EEG monitoring, MEG and dEEG ictal onsets were both localized to the left superior frontal gyrus concordant with ictal onset on intracranial EEG. Conclusions: We found that the combined use of whole-head MEG and dEEG in conjunction with advanced source modeling techniques promotes better noninvasive characterization of both the irritative regions and ictal onset zone in presurgical epilepsy evaluations with observation of the different directions of sources between MEG and dEEG. In addition, using frequency domain analysis judiciously, i.e., frequency bandwidths tailored to the signal properties rather than the fixed broad conventional bandwidths, better identifies the seizure onset zone as determined by intracranial EEG.