Realistic and spherical head modeling for EEG forward problem solution: a comparative cortex-based analysis

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Modeling

Background:

• The accuracy of electroencephalography (EEG) forward models is crucial for reliable source reconstruction.
• Head tissue geometry significantly influences EEG forward model accuracy.
• The impact of different head model geometries on specific brain regions remains unclear.

Purpose of the Study:

• To compare spherical and realistic head modeling techniques for EEG forward solutions.
• To evaluate the sensitivity of different brain regions to head model geometry choices.
• To assess the accuracy of EEG source reconstruction using various head models.

Main Methods:

• Utilized three four-shell head models: two realistic (Boundary Element Method and Finite Difference Method) and one spherical.
• Simulated EEG forward solutions from dipole sources on a standard cortical space (Montreal Neurological Institute MRI data).
• Performed Point Spread Function and Lead Field cross-correlation analyses for 26 dipole sources.

Main Results:

• Realistic head geometry significantly improves EEG forward model accuracy compared to spherical models.
• The choice of head model geometry has a greater impact on sources located in the temporal and occipital cortex.
• Quantitative analyses confirmed the superiority of realistic models for source reconstruction accuracy.

Conclusions:

• Realistic head models are essential for improving the reliability of EEG source reconstruction.
• Source localization accuracy is particularly sensitive to geometric modeling in the temporal and occipital regions.
• This study provides quantitative evidence for the benefits of incorporating realistic head geometry in EEG analysis.