Towards the Automatic Localization of the Irritative Zone Through Magnetic Source Imaging

Area of Science:

• Neuroscience
• Biophysics
• Medical Imaging

Background:

• Interictal epileptiform discharges (IEDs) are key indicators of epilepsy.
• Accurate localization of IED generators is crucial for epilepsy surgery planning.
• Current methods like dipole fitting can be subjective and time-consuming.

Purpose of the Study:

• To validate the Bayesian multi-dipole modeling algorithm (SESAME) for localizing IED generators.
• To compare SESAME's performance against established methods (dipole fitting, RAP-MUSIC, wMNE).
• To assess the relationship between source localization concordance and post-surgical outcomes.

Main Methods:

• Resting-state magnetoencephalographic (MEG) recordings were used.
• Bayesian multi-dipole modeling (SESAME) was applied to localize IED generators.
• Results were benchmarked against expert-performed Equivalent Current Dipole (ECD) fitting and compared with RAP-MUSIC and wMNE.

Main Results:

• SESAME and RAP-MUSIC showed agreement with ECD fitting in identifying affected cerebral lobes.
• SESAME's localization results were closer to ECDs than RAP-MUSIC.
• Concordance between surgical plans and SESAME-identified lobes predicted seizure freedom post-surgery.

Conclusions:

• Bayesian multi-dipole modeling (SESAME) provides a reliable and objective alternative to manual dipole fitting for epilepsy surgery.
• Automated source localization methods reduce subjectivity in clinical applications.
• Improved source localization accuracy correlates with better surgical outcomes.