Electric and Magnetic Source Imaging in Epilepsy: Anatomical Insights from Realistic Simulations
- 1Cook Children's Health Care System
Abstract
Neurosurgery is an effective treatment for drug-resistant epilepsy (DRE), but its success depends on the accurate delineation of the epileptogenic focus, approximated noninvasively by electric and magnetic source imaging (ESI/MSI) of high-density electroencephalography (HD-EEG) and magnetoencephalography (MEG). These modalities have distinct sensitivity profiles; consequently, epileptiform activity detected by one modality may remain undetected by the other. However, the implications of these differences for spike detectability and localization accuracy, and the added value of combined electromagnetic source imaging (EMSI), remain insufficiently characterized under clinically realistic conditions. Here, we characterize HD-EEG and MEG sensitivity profiles and compare ESI, MSI, and EMSI, using realistic simulations with known ground truth. We generated HD-EEG and MEG datasets from cortical, insular, and hippocampal sources across multiple orientations, activated using patient-derived intracranial EEG spike waveforms. Spike sensitivity and localization accuracy were evaluated using dipole scanning, dynamic statistical parametric mapping (dSPM), and linearly constrained minimum variance (LCMV). EEG demonstrated higher spike detection rates and sensitivity for radially cortical sources at gyral crowns and sulcal bottoms (p<0.001), while MEG detected more spikes from tangential cortical sources along sulcal walls (p<0.001). MSI localized tangential sources more accurately than ESI across all methods (p<0.001), while dipole scanning outperformed dSPM and LCMV for cortical and insular sources (p<0.001). EMSI outperformed ESI (p<0.001) but did not improve localization over MSI. These findings suggest that complementary anatomical sensitivities of HD-EEG and MEG may improve epileptogenic source characterization, supporting simultaneous acquisition during the presurgical evaluation of patients with DRE.
Related
- Physics-informed neural operators for rapid electric field prediction in transcranial magnetic stimulation of healthy and tumor subjects
- Quantifying Reproducibility of Common Laboratory Tests in a Tertiary Care Hospital for More Realistic Synthetic Data in Machine Learning
- The collective turing test: large language models can generate realistic multi-user discussions
- Epilepsy Surgery vs Medical Management for Pediatric Drug-Resistant Focal Epilepsy
- Factors associated with drug-resistant epilepsy in children with epilepsy: a retrospective cohort study
BibTeX
@misc{sdoukopoulou2026electric,
title = {{Electric and Magnetic Source Imaging in Epilepsy: Anatomical Insights from Realistic Simulations}},
author = {G. Sdoukopoulou and S. Jahromi and L. Corona and H. Partamian and J. Lim and J. Afnan and C. Grova and C. Papadelis},
howpublished = {medRxiv},
year = {2026},
month = nov,
doi = {10.64898/2026.11.06.26367214},
url = {https://www.medrxiv.org/content/10.64898/2026.11.06.26367214v1},
note = {CC BY-NC-ND},
}
RIS
TY - UNPB
TI - Electric and Magnetic Source Imaging in Epilepsy: Anatomical Insights from Realistic Simulations
AU - G. Sdoukopoulou
AU - S. Jahromi
AU - L. Corona
AU - H. Partamian
AU - J. Lim
AU - J. Afnan
AU - C. Grova
AU - C. Papadelis
PY - 2026
DA - 2026/11/06
PB - medRxiv
DO - 10.64898/2026.11.06.26367214
UR - https://www.medrxiv.org/content/10.64898/2026.11.06.26367214v1
AB - Neurosurgery is an effective treatment for drug-resistant epilepsy (DRE), but its success depends on the accurate delineation of the epileptogenic focus, approximated noninvasively by electric and magnetic source imaging (ESI/MSI) of high-density electroencephalography (HD-EEG) and magnetoencephalography (MEG). These modalities have distinct sensitivity profiles; consequently, epileptiform activity detected by one modality may remain undetected by the other. However, the implications of these differences for spike detectability and localization accuracy, and the added value of combined electromagnetic source imaging (EMSI), remain insufficiently characterized under clinically realistic conditions. Here, we characterize HD-EEG and MEG sensitivity profiles and compare ESI, MSI, and EMSI, using realistic simulations with known ground truth. We generated HD-EEG and MEG datasets from cortical, insular, and hippocampal sources across multiple orientations, activated using patient-derived intracranial EEG spike waveforms. Spike sensitivity and localization accuracy were evaluated using dipole scanning, dynamic statistical parametric mapping (dSPM), and linearly constrained minimum variance (LCMV). EEG demonstrated higher spike detection rates and sensitivity for radially cortical sources at gyral crowns and sulcal bottoms (p<0.001), while MEG detected more spikes from tangential cortical sources along sulcal walls (p<0.001). MSI localized tangential sources more accurately than ESI across all methods (p<0.001), while dipole scanning outperformed dSPM and LCMV for cortical and insular sources (p<0.001). EMSI outperformed ESI (p<0.001) but did not improve localization over MSI. These findings suggest that complementary anatomical sensitivities of HD-EEG and MEG may improve epileptogenic source characterization, supporting simultaneous acquisition during the presurgical evaluation of patients with DRE.
ER -