Dipole versus distributed EEG source localization for single versus averaged spikes in focal epilepsy

Area of Science:

• Neuroscience
• Medical Imaging
• Epileptology

Background:

• Focal epilepsy diagnosis relies on accurate localization of interictal epileptiform discharges (IEDs).
• Electroencephalography Source Localization (ESL) offers potential for non-invasive IED source identification.
• Comparing dipole and distributed ESL models is crucial for clinical application.

Purpose of the Study:

• To characterize and compare dipole and distributed ESL of IEDs in focal epilepsy.
• To evaluate the performance of different inverse models and source configurations.
• To assess the impact of independent component analysis (ICA) on ESL interpretation.

Main Methods:

• Applied independent component analysis (ICA) to single and averaged IEDs in benign focal epilepsy of childhood with centrotemporal spikes (BFEC) and mesial temporal lobe epilepsy (MTLE) patients.
• Utilized boundary element method forward modeling with four inverse models (two dipole, two distributed).
• Analyzed ESL solutions at various time points (onset, midupswing, peak) and fit criteria (strength maxima, residual deviation minima).

Main Results:

• Best-fit dipole and distributed ESL solutions localized to the IED midupswing, not the peak, in both BFEC and MTLE.
• Localization in BFEC was the lower Rolandic sulcus; in MTLE, it was basolateral or anterior temporal cortex.
• High variability was observed in single-to-averaged ESL in MTLE, indicating potential discordance.

Conclusions:

• Dipole and distributed ESL are complementary techniques for focal epilepsy.
• Optimal ESL solutions are found at the IED midupswing, aided by ICA for fit quality assessment.
• High single-to-averaged IED localization discordance requires further investigation for routine clinical use of ESL.