Experimental tests of EEG source localization accuracy in realistically shaped head models

Area of Science:

• Neuroscience
• Biomedical Engineering
• Medical Imaging

Background:

• Accurate localization of electrical sources in the human brain is essential for understanding neural activity.
• Traditional methods often employ simplified spherical head models, which may not fully represent individual anatomical complexities.
• Advancements in imaging techniques allow for the creation of more realistic head models.

Purpose of the Study:

• To assess the accuracy of brain electrical source localization using realistic head boundary element models.
• To compare the localization accuracy of realistic head models against conventional spherical head models.

Main Methods:

• Utilized electroencephalography (EEG) data previously recorded from human subjects with known electrical source locations.
• Employed computerized tomography (CT) scans to determine precise locations of implanted electrodes and individual head shapes.
• Calculated source locations using both realistically shaped head boundary element models and standard spherical head models.

Main Results:

• The average localization error was comparable between realistic models (10.5 mm) and spherical models (10.6 mm).
• For superior brain sources, realistic models showed an average error of 9.1 mm, very similar to spherical models (9.2 mm).
• For inferior brain sources, realistic models yielded an average error of 12.4 mm, closely matching spherical models (12.8 mm).

Conclusions:

• Realistically shaped head boundary element models offer localization accuracy comparable to that of spherical models.
• The findings suggest that complex anatomical details in head models do not significantly improve upon the localization accuracy achieved with simpler spherical models for EEG source analysis.