Noise Reduction and Localization Accuracy in a Mobile Magnetoencephalography System

Area of Science:

• Neuroscience
• Biophysics
• Biomedical Engineering

Background:

• Magnetoencephalography (MEG) offers non-invasive insights into human brain electrophysiology.
• Optically pumped magnetometers (OPMs) are emerging as an alternative to traditional cryogenic sensors in MEG systems.
• Current cryogenic MEG systems require large, shielded rooms, limiting accessibility.

Purpose of the Study:

• To evaluate the performance of OPM-based MEG recordings within a compact, cylindrical magnetic shield (1 × 2 m²).
• To assess the effectiveness of passive shielding and post hoc noise reduction algorithms.
• To quantify the localization accuracy of OPM sensors using simulations and phantom recordings.

Main Methods:

• Investigated passive shielding efficacy (field attenuation and isotropy) in a small-footprint cylindrical shield.
• Applied post hoc noise reduction algorithms to OPM recordings.
• Quantified localization accuracy using a bespoke current dipole phantom and 104 OPM sensors in a fixed helmet array.

Main Results:

• Passive shielding reduced magnetic fields significantly across DC, power line, and 10-200 Hz ranges.
• Post hoc noise reduction provided an additional 5-15 dB attenuation.
• Phantom signals were localized with 3 mm accuracy, demonstrating no significant localization bias.

Conclusions:

• A compact OPM-based MEG system validated with a phantom achieves high localization accuracy comparable to existing systems.
• The small footprint and validated performance pave the way for lower-cost, more accessible MEG installations.
• This advancement supports the global adoption of MEG, including mobile and resource-limited settings.