A finite-element reciprocity solution for EEG forward modeling with realistic individual head models

Area of Science:

• Computational neuroscience
• Biomedical engineering
• Medical imaging

Background:

• Accurate head modeling is crucial for precise source localization in electroencephalography (EEG).
• Existing methods for solving the EEG forward problem have computational limitations or may not fully capture tissue properties.

Purpose of the Study:

• To present a novel finite element modeling (FEM) implementation for the EEG forward problem.
• To leverage Helmholtz's principle of reciprocity for computational efficiency.
• To incorporate anisotropic conductivities in realistic head models.

Main Methods:

• Developed an FEM implementation based on Helmholtz's principle of reciprocity.
• Validated the method against established alternatives (OpenMEEG, SimBio) using a spherical model.
• Applied the FEM approach to human MRI data, creating a multi-tissue head model.
• Calculated conductivity tensors from diffusion-weighted MR images to model anisotropy.

Main Results:

• The FEM implementation demonstrated comparable accuracy to state-of-the-art methods.
• Incorporating white matter anisotropy reduced orientation-specific errors in the leadfield matrix.
• Ignoring white matter's directional conductivity led to significant forward model inaccuracies.

Conclusions:

• The developed FEM method offers a fast, accurate, and open-source solution for EEG forward problems.
• Accounting for white matter anisotropy is essential for precise individual-based source localization.
• This approach enhances the realism and reliability of EEG head models.