A realistic multimodal modeling approach for the evaluation of distributed source analysis: application to sLORETA

Area of Science:

• Neuroscience
• Biomedical Engineering
• Epileptology

Background:

• Electrical Source Localization (ESL) is crucial for epilepsy surgery planning.
• Both scalp EEG and intracranial EEG (iEEG) are used for ESL.
• Validation of ESL accuracy, especially spatial extent, is needed.

Purpose of the Study:

• To validate distributed ESL from scalp EEG and iEEG using realistic epileptic spike simulations.
• To assess the accuracy of ESL in determining the spatial extent of epileptic sources.
• To compare the performance of scalp EEG-based ESL versus iEEG-based ESL.

Main Methods:

• Applied ESL to averaged scalp EEG and iEEG spikes from two epilepsy patients.
• Developed a spatiotemporal source model based on patient-specific ESL results.
• Generated simulated EEG signals across various anatomic locations and spatial extents.
• Performed ESL using sLORETA on simulated data and quantitatively assessed accuracy for iEEG and scalp EEG.

Main Results:

• sLORETA overestimated source volume for both scalp EEG and iEEG, with increasing error for larger sources.
• ESL accuracy decreased significantly for larger sources.
• Reconstruction volumes shifted centrally for iEEG but remained stable for scalp EEG.
• iEEG-derived ESL showed more pronounced source mislocalization compared to scalp EEG.

Conclusions:

• A novel multiscale framework using realistic EEG simulations was developed for ESL evaluation.
• Scalp EEG-derived ESL generally yields more accurate results than iEEG-derived ESL due to better head coverage.
• iEEG-derived reconstruction results for widespread epileptic generators require cautious interpretation.