Abstract
Objective
To evaluate the ability of magnetoencephalography (MEG) to detect the activity of a deep brain structure, particularly the hippocampus, using simultaneous stereotactic electroencephalography (SEEG) and MEG recordings.
Methods
We included six patients with intractable temporal lobe epilepsy who underwent simultaneous SEEG and MEG. We recorded SEEG spikes and examined the MEG waveforms. We defined S-MEG spikes as MEG spikes synchronized with SEEG spikes in the deep brain structures. We examined the proportion of S-MEG spikes relative to SEEG spikes, the amplitude of S-MEG spikes, and the signal-to-noise ratio (SNR). We also examined the relationship between the spread of epileptic activity and the SNR of S-MEG spikes. We used SEEG spike-indexed averaging to improve the SNR and visualize the source of the S-MEG spikes.
Results
We recorded 1630 SEEG spikes, of which 575 (35.3%) were identified as S-MEG spikes. The mean amplitude of the S-MEG spikes derived from hippocampal spikes was 163 femtotesla per centimeter (fT/cm), significantly lower than the amplitude from spikes of cortical tissue (p < .001). The mean SNR of S-MEG spikes derived from hippocampal spikes was 0.96, significantly lower than that of cortical tissue spikes (p < .001). The mean SNR of S-MEG spikes derived from hippocampal interelectrode synchronized SEEG spikes was 1.09, significantly greater than the mean SNR of 0.92 without interelectrode synchronization (p = .0035). SEEG spike-indexed averaging could improve the poor SNR and depict dipoles in the hippocampi.
Significance
We detected MEG spikes localized in the hippocampus, the amplitude and SNR of which were significantly lower than those of spikes originating from the cortex. As it spreads, hippocampal epileptic activity becomes easier to detect with MEG, but it remains challenging for MEG to improve the poor SNR to detect the limited hippocampal activity without SEEG.
JUN