Spatial sampling of MEG and EEG based on generalized spatial-frequency analysis and optimal design

Area of Science:

• Neuroscience
• Biophysics
• Medical Imaging

Background:

• Electroencephalography (EEG) and magnetoencephalography (MEG) measure brain activity using sensors on the scalp.
• Spatial sampling density is crucial for accurately capturing the spatial-frequency content of these brain signals.
• Current sensor placement strategies may not be optimized for maximizing information retrieval.

Purpose of the Study:

• To analyze the optimal spatial sampling density for EEG and MEG.
• To investigate the spatial-frequency content of signals acquired with different sampling strategies.
• To propose and evaluate a novel approach for determining optimal sensor locations.

Main Methods:

• Simulated realistic head models to analyze spatial sampling.
• Calculated spatial-frequency content for on-scalp MEG, off-scalp MEG, and EEG.
• Developed and tested a model-informed approach for sensor location optimization.

Main Results:

• On-scalp MEG, off-scalp MEG, and EEG benefit from up to 280, 90, and 110 spatial samples, respectively.
• A novel method for optimizing sensor locations based on prior assumptions and uniformity was proposed.
• Model-informed non-uniform sampling is beneficial for low sample numbers, while uniform grids suffice for high sample numbers.

Conclusions:

• The study provides evidence-based recommendations for spatial sampling densities in EEG and MEG.
• The proposed sensor placement approach allows for tailored optimization based on specific research needs.
• Both uniform and model-informed sampling strategies have roles depending on the number of spatial samples used.