Effects of modeling errors and EEG measurement montage on source localization accuracy

Area of Science:

• Neuroscience
• Biophysics
• Medical Imaging

Background:

• Electroencephalography (EEG) measures brain electrical activity.
• Source localization aims to pinpoint the origin of EEG signals within the brain.
• Discrepancies between actual head geometry and spherical models introduce localization errors.

Purpose of the Study:

• To investigate localization errors caused by head shape and brain pan modeling inaccuracies.
• To assess the impact of different EEG montages on localization accuracy under these conditions.

Main Methods:

• Comparison of EEG source localization in realistic head models versus spherical models.
• Analysis of localization errors attributed to nonspherical head shape and brain pan geometry.
• Evaluation of various EEG electrode montages (referential, average reference, bipolar).

Main Results:

• Nonspherical head shape resulted in maximum localization errors of ~1 cm, largely independent of source parameters or montage.
• Brain pan geometry caused larger errors (up to ~2 cm), highly dependent on source location, orientation, and montage.
• Referential montages with a bottom reference showed the largest errors; other montages performed better.

Conclusions:

• Head geometry, particularly the brain pan, significantly impacts EEG source localization accuracy.
• EEG montage selection is crucial for mitigating localization errors, especially with complex head anatomy.
• Optimized montages can improve the reliability of brain source localization from EEG data.