A study of dipole localization accuracy for MEG and EEG using a human skull phantom

Area of Science:

• Neuroscience
• Biophysics
• Biomedical Engineering

Background:

• Electroencephalography (EEG) and Magnetoencephalography (MEG) are crucial for non-invasive brain activity mapping.
• Accurate source localization of neural activity is essential for understanding brain function and dysfunction.
• Forward and inverse modeling techniques are used to estimate neural source locations from scalp recordings.

Purpose of the Study:

• To evaluate the accuracy of forward and inverse methods for dipole localization using EEG and MEG.
• To compare the performance of different head models (spherical vs. realistic) in dipole localization.
• To identify factors influencing the accuracy of EEG and MEG source localization.

Main Methods:

• A human skull phantom with realistic tissue conductivities was used.
• Thirty-two current dipoles were implanted, and EEG/MEG data were recorded.
• Dipole locations were estimated using the R-MUSIC inverse method with spherical and Boundary Element Method (BEM) forward models.
• X-ray CT data provided true dipole locations and head model anatomy.

Main Results:

• The average localization error was 7-8 mm for EEG and 3 mm for MEG.
• Realistic head models and simpler spherical models yielded comparable localization errors.
• MEG demonstrated superior localization accuracy compared to EEG.

Conclusions:

• MEG exhibits higher accuracy in dipole localization than EEG.
• EEG's reduced accuracy is attributed to its sensitivity to forward model errors, particularly concerning skull and scalp conductivity assumptions.
• Accurate head conductivity modeling is critical for precise EEG source localization.