Abstracts

Rationale:
Studies comparing short-time MEG recordings with long-term intracranial electroencephalography (iEEG) have reported that MEG identifies signals consistent with iEEG spikes and that a significant portion of the epileptic network analyzed from iEEG can also be detected by MEG.




However, since these studies were not performed simultaneously, it is difficult to compare and interpret the interactions in different epileptic waves, such as spikes and high-frequency oscillations, in different sessions. By simultaneously recording MEG and stereotactic intracranial EEG (SEEG), it is possible to determine whether the abnormal waves detected by MEG and SEEG correspond to the same epileptic activity, and to examine the consistency between modalities in the exact same condition.




Methods:
Eight patients with refractory epilepsy who underwent SEEG (frontal lobe epilepsy: 4 cases, temporal lobe epilepsy: 4 cases).






MEG measurements were focused on electrodes capable of detecting SEEG anomalous waves; MEG was evaluated for detection of anomalous signals at the sensor level, dipole analysis was performed, and further spatial filter analysis was performed at the source level.




Results:
SEEG electrodes at a distance of 10 cm from the sensor could confirm the MEG signal if the signal strength was above approximately several hundred µV, but even if a signal was detected, current dipole (ECD) analysis was difficult to perform. It was difficult for MEG to detect abnormal signals localized to the hippocampal electrodes. Spatial filter analysis showed anomalies at the source level for cases where dipole analysis was difficult, but it was difficult to capture a distinct localized response.




Conclusions: 　Deep magnetic field sources are accessible by MEG despite their low amplitude at the sensor level, consistent with previous reports. Abnormal waves at the epileptic focus at a depth of a few centimeters from the scalp (about 10 cm from the sensor) can be visually detected with amplitudes larger than several hundred microV and clearly different from the background magnetic field. On the other hand, localized signals from the hippocampus are difficult to detect with MEG, even when the amplitude is greater than 2 mV. This indicates the peculiarity of the neural network in the hippocampus. The distance between the location of the MEG Dipole and the electrode of the maximum signal of SEEG is reported to be 20.7mm in the shallow area and 20.7mm in the deep area, and our analysis results show the same thing. Although there are reports that spatial filter analysis and independent component analysis (ICA) can reconstruct the current source with excellent sensitivity and specificity, our data did not point out the superiority of spatial filter analysis.

Funding: none