More Reliable EEG Electrode Digitizing Methods Can Reduce Source Estimation Uncertainty, but Current Methods Already...

Area of Science:

• Neuroscience
• Biomedical Engineering
• Signal Processing

Background:

• Electroencephalography (EEG) and source estimation are crucial for understanding cortical dynamics during cognitive tasks.
• Accurate source estimation necessitates precise knowledge of EEG electrode locations, which can be obtained via templates or digitization.
• Digitization methods vary in reliability, potentially impacting the accuracy of source localization and interpretation.

Purpose of the Study:

• To compare the reliability of five different EEG electrode digitization methods: ultrasound, structured-light 3D scan, infrared 3D scan, motion capture probe, and motion capture.
• To determine the relationship between the reliability of EEG electrode digitization and the resulting uncertainty in source estimation.
• To evaluate the practical implications of digitization reliability on Brodmann area accuracy for source localization.

Main Methods:

• Five digitization methods were used to record EEG electrode locations on a mannequin head five times each.
• Reliability and validity of each method were quantified.
• Five hundred sets of electrode locations were generated to simulate digitization variability, and source estimation was performed using the DIPFIT algorithm.

Main Results:

• The motion capture method, using markers directly on electrodes, demonstrated the highest reliability (0.001 cm average electrode variability).
• Reliability decreased in the order: motion capture probe, infrared 3D scan, structured-light 3D scan, and ultrasound.
• Source estimation uncertainty increased with greater electrode location variability; a 1 cm variability could shift a source by 2 cm.
• Average Brodmann area accuracy exceeded 80% for all digitization methods, significantly outperforming template-based localization (approx. 50%).

Conclusions:

• More reliable EEG electrode digitization methods reduce source estimation uncertainty.
• The significance of source estimation uncertainty is dependent on the required spatial resolution.
• All tested digitization methods provide sufficient accuracy for Brodmann area identification, outperforming traditional template methods.