CutFEM forward modeling for EEG source analysis

Area of Science:

• Computational neuroscience
• Biomedical engineering
• Medical imaging

Background:

• Electroencephalography (EEG) source analysis relies on solving the forward problem, which models scalp potentials from brain activity.
• Finite Element Method (FEM) is crucial for accurate head modeling but faces challenges with mesh generation for complex geometries.
• Existing FEM approaches struggle to balance geometric flexibility with computational efficiency.

Purpose of the Study:

• Introduce CutFEM, an unfitted FEM, for advanced EEG forward simulations.
• Integrate the advantages of both hexahedral and tetrahedral meshes in EEG modeling.
• Improve the accuracy and efficiency of modeling volume conduction effects in the human head.

Main Methods:

• Developed and applied CutFEM, a type of unfitted finite element method, for EEG forward simulations.
• Decoupled mesh and geometry representation to handle complex head models.
• Validated CutFEM in controlled spherical models and real-world somatosensory-evoked potential reconstructions.

Main Results:

• CutFEM demonstrated superior numerical accuracy compared to traditional FEM approaches.
• Achieved significant reductions in memory consumption and computational time.
• Successfully meshed arbitrarily touching compartments, enabling more realistic head models.

Conclusions:

• CutFEM offers a balanced solution for EEG forward modeling, enhancing numerical accuracy and computational efficiency.
• Provides smooth approximation of complex geometries previously unattainable with standard FEM.
• Represents a significant advancement in FEM-based EEG forward modeling capabilities.