Extended homogeneous field correction method based on oblique projection in OPM-MEG

Area of Science:

• Neuroimaging
• Biophysics
• Signal Processing

Background:

• Optically pumped magnetometer-based magnetoencephalography (OPM-MEG) is a flexible, wearable neuroimaging technique.
• Environmental noise suppression is critical for OPM-MEG, especially with limited channels.
• Existing methods like subspace projection and homogeneous field correction (HFC) have limitations in complex noise environments.

Purpose of the Study:

• To develop a novel noise suppression method for OPM-MEG systems.
• To address the limitations of existing methods in handling complex, non-homogeneous noise.
• To enhance signal quality at both sensor and source levels in OPM-MEG.

Main Methods:

• Proposed an extended homogeneous field correction method based on oblique projection (opHFC).
• opHFC constructs an oblique projection operator to divide signals into internal and external components.
• Employed temporal extension to eliminate complex interferences and addressed non-orthogonality between signal and noise subspaces.

Main Results:

• opHFC demonstrated superior noise suppression compared to four benchmark methods via simulations and experiments.
• The method achieved minimal signal distortion, enhancing signal quality at sensor and source levels.
• Validated performance using auditory and somatosensory evoked OPM-MEG data.

Conclusions:

• opHFC offers a novel and effective approach for reducing interference in OPM-MEG.
• The method expands the application scenarios for OPM-MEG systems.
• Provides high-quality signals crucial for scientific research and clinical applications.