Abstracts

Magnetoencephalography (MEG) is gaining increasing acceptance as a valuable and valid pre-surgical clinical evaluation tool for epilepsy cases. The literature has demonstrated both a correlation between MEG epilepsy spike localizations and intra-operative electrocorticography, and, MEG sensory cortex localization with direct cortical stimulation of motor cortex. One modality that has not been confirmed is the correlation between MEG-identified sensory areas using somatosensory evoked fields (SEF) with intra-operative somatosensory evoked potentials (SEP). In this retrospective study, we compare the location of the MEG SEF with the location of the intra-operative SEP in a group of pediatric patients undergoing subdural grid placement. 9 patients (2-17yrs) with intractable epilepsy underwent a MEG. To obtain the SEF, median nerves were stimulated and the resultant waveforms submitted to dipole analyses where the magnetic correlates of the electrical N20 were localized. SEF dipole locations were then co-registered onto a structural MRI. For 4/9 patients, these data were transferred to an intra-operative neuronavigation system (Zeiss Systems); for the other 5/9, the MRI data underwent 3-dimensional reconstruction but were not taken directly into the operating room. During epilepsy surgery, prior to subdural grid placement, electrical SEPs were recorded from the cortex using a 2x4 electrode array. The contralateral median nerve was stimulated and the SEP N20 phase reversal was used to identify sensory and motor cortices. Identification of motor cortex was always confirmed by direct cortical stimulation. For the 4/9 patients whose SEFs were transferred to the neuronavigation system, the SEF-SEP difference could be compared directly. Of these 4 patients, 1 showed an exact overlay, 2 were adjacent within 1 cm, and 1 was adjacent within 2 cm; all localizations were on the same gyrus. For the other 5/9 patients without neuronavigation, the 3-dimensional reconstruction allowed the SEF dipole location to be projected to the cortical surface, and the distance between this and the intra-operative SEP could be measured. 4 of these 5 patients showed an SEF-SEP difference of within 2 cm, also on the same gyrus. The last patient did not demonstrate an SEF despite a technically clean MEG; interestingly, intra-cortical SEPs also could not be obtained despite stimulation at above motor threshold. We demonstrate a high correspondence between the MEG localization of sensory cortex using SEF and the gold standard, which is the recording of intra-cortical SEP phase reversals. This adds to the existing literature demonstrating the validity of MEG localizations, and points to the utility of MEG as an non-invasive tool for the accurate localization of function.