Authors :
Presenting Author: Kaya Scheman, BS – National Institutes of Health
Elena Hayday, BS – National Institutes of Health; Price Withers, BS – National Institutes of Health; Jeff Stout, PhD – National Institutes of Health; Antonio Triggiani, PhD – National Institutes of Health; Sara Inati, MD – National Institutes of Health
Rationale:
Magnetoencephalography (MEG) is a useful adjunctive tool in the presurgical evaluation of patients with drug resistant focal epilepsy (DRE). The equivalent current dipole (ECD) method is a validated gold standard for interictal epileptiform discharge (IED) source localization. Although effective, the ECD method models IED sources as highly focal, giving no information about either local or distant propagation of activity. Here, we present a novel method to investigate IED propagation patterns in interictal MEG recordings using dynamic statistical parametric mapping (dSPM) source mapping. We identify IED-related spatial clusters of significant gray matter activation in individual patients, showing time delays within each cluster suggestive of local propagation of activity.
Methods:
We examined the resting-state magnetoencephalography recordings of eight patients (four male, age 28 ± 15 years) who underwent presurgical evaluation for DRE at the National Institutes of Health Clinical Center. Similar IEDs were identified through an in-house automated detector and were manually verified. T1-weighted 3T MRI images were used to reconstruct cortical surfaces using Freesurfer. MNE software was used to create boundary element models. MEG data was filtered at 5-50 Hz, smoothed, and 2s epochs were created, centered around the marked IEDs. dSPM was used to reconstruct virtual source activity for the averaged epoch. We summed IED power at each source across a 100ms time window around the peak to quantify overall IED activity at each source. Spatial clustering was used to identify independent brain regions with significant IED-related co-activation. Within each identified cluster, time delays between the source points with the earliest and latest peaks were computed to evaluate whether activity in the cluster more likely represents volume conduction expected to have no or minimal time delays, or local propagation. Results:
We identified significant clusters of IED-related co-activation in 7/8 patients. More than one spatially independent cluster was identified in 6/7 (86%) with an average of 5.4 ± 5.5 clusters (Figures 1 and 2). Intra-cluster time delays were computed for 4/7 patients. Within clusters, the average time delay between the first and last peak was 14.6 ms. Conclusions:
Here, we present a novel magnetic source imaging approach to investigate propagation patterns of IEDs in patients with DRE undergoing presurgical evaluation. We were able to localize independent significant clusters in most patients, with multiple clusters of activation identified in most patients. Within the clusters, there were average delays of 14.6 ms from the first to last peak in the included virtual sensors, suggesting local propagation instead of volume conduction. This preliminary work is significant as this approach may allow for future investigation of common local and distant propagation patterns across IEDs within patients, but also across patients, which may be useful for understanding IEDs and for presurgical planning.Funding: National Institutes of Health Intramural Research Program