Modeling of the human skull in EEG source analysis.

Area of Science:

• Neuroscience
• Biophysics
• Computational Modeling

Background:

• Electroencephalography (EEG) is crucial for brain activity analysis.
• Accurate EEG source analysis relies on precise biophysical models of the human skull.
• Current models often simplify skull conductivity, potentially impacting source localization accuracy.

Purpose of the Study:

• To investigate the impact of skull modeling assumptions on EEG source analysis accuracy.
• To compare local, global, isotropic, and anisotropic skull conductivity models.
• To provide recommendations for improved skull modeling in EEG.

Main Methods:

• Computer simulations using finite element models of the human skull.
• Comparison of local and global skull models against a reference model with explicit bone layers.
• Inclusion of both isotropic and anisotropic conductivity assumptions.
• Evaluation of forward calculation and dipole position errors for sources across the entire brain.

Main Results:

• Accounting for local skull conductivity variations significantly improves accuracy.
• Isotropic versus anisotropic conductivity assumptions had minimal influence on results.
• Homogeneous, isotropic skull models require a lower skin/brain to skull conductivity ratio than the conventional 80:1.
• Specific recommendations for local and global skull modeling were derived.

Conclusions:

• Local skull variations are critical for accurate EEG source analysis.
• Standard homogeneous skull models may require adjusted conductivity parameters.
• Explicitly modeling bone layers (local models) is recommended when feasible.
• If local modeling is not possible, global models with adjusted conductivity values are advised.