A Robust eLORETA Technique for Localization of Brain Sources in the Presence of Forward Model Uncertainties

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Electrophysiology

Background:

• Exact low-resolution electromagnetic tomography (eLORETA) is used for brain source localization.
• eLORETA's sensitivity to forward model uncertainties limits its clinical application, particularly in identifying the epileptogenic zone.
• Developing robust methods to handle these uncertainties is crucial for real-world use.

Purpose of the Study:

• To introduce ReLORETA, a robust version of eLORETA, designed to improve brain source localization accuracy despite forward model uncertainties.
• To address limitations in eLORETA stemming from unknown lead field matrices and model discrepancies.

Main Methods:

• An iterative approach was developed to estimate transformations correcting for lead field matrix uncertainties.
• Simulations incorporated major sources of uncertainty: geometry, conductivity, source space resolution, and electrode misalignment.
• ReLORETA and eLORETA were compared using simulated focal brain sources and real patient data.

Main Results:

• ReLORETA demonstrated significantly greater robustness and accuracy compared to eLORETA across various simulated noise levels and source locations.
• Performance improvements were consistent in both simulated data and real-world epilepsy patient data.
• The method effectively handled forward model uncertainties, including geometric and conductivity variations.

Conclusions:

• ReLORETA offers a promising advancement for clinical brain source localization by effectively managing forward model uncertainties.
• The developed technique shows potential for real-world applications, including epilepsy diagnosis.
• Robustness against model inaccuracies makes ReLORETA a valuable tool for neuroimaging.