Influence of head models on EEG simulations and inverse source localizations

Area of Science:

• Neuroscience
• Biomedical Engineering
• Computational Modeling

Background:

• Anatomical surface structures like gray matter, white matter, and cerebrospinal fluid (CSF) influence volume current flow in head models.
• This influence affects scalp potentials and the accuracy of inverse source localization techniques.

Purpose of the Study:

• To investigate how the complexity of finite element head models affects electroencephalography (EEG) source localization.
• To determine the impact of specific tissue layers, particularly CSF, on the accuracy of scalp potential simulations and source localization.

Main Methods:

• Four finite element head models of varying complexity were created from segmented MRI data.
• Models ranged from a detailed eleven-tissue model to a simplified five-tissue model, with variations in gray matter, white matter, and CSF conductivity.
• Lead fields and scalp potentials were computed for dipolar sources in the motor cortex, and inverse source localizations were performed across a range of signal-to-noise ratios (SNRs).

Main Results:

• The most complex model (Model 1) served as the reference and performed best.
• Model 3, lacking the CSF layer, exhibited the largest mean source localization errors (MLEs) and most significantly altered scalp potentials.
• Model 4 also showed higher MLEs than Models 1 and 2, with performance comparable to Model 3 at low SNRs but better at higher SNRs.

Conclusions:

• Head model complexity is a critical factor for accurate scalp potential simulations and inverse source localization.
• More heterogeneous models with detailed tissue surfaces, especially the inclusion of the CSF layer, lead to improved localization accuracy.
• Accurate simulations necessitate highly detailed head models that reflect complex anatomical structures.