Motion Artifact Suppression Method Based on Dynamic Background Field Adaptive Filtering for Wearable OPM-MEG

Magnetoencephalography (MEG) is a noninvasive brain imaging technique that investigates dynamic brain changes by measuring magnetic fields generated by neural activities. Recently, significant advancements have been made in optically pumped magnetometer-based MEG (OPM-MEG). Compared with conventional superconducting quantum interference device (SQUID)-based MEG, OPM-MEGs operate at room temperature, reducing costs and allowing more flexible sensor configurations. Most importantly, OPM-MEG allows for measurements during natural head movements, facilitating more naturalistic experiments and broader applications, including studies involving motor tasks and populations, such as young children or patients with movement disorders. However, motion-compatible OPM-MEGs face significant challenges because of the severe impact of motion artifacts on data quality. This study introduced a dynamic background field adaptive filtering (DBFAF) method to suppress motion artifacts in OPM-MEG. The OPM-MEG data were processed by adaptively filtering out the motion artifacts measured using multiple orthogonally oriented OPM reference sensors fixed to a helmet. The simulation results demonstrate that the DBFAF method significantly improves data quality by effectively mitigating the impact of motion artifacts, surpassing the widely used moving-window regression methods. The efficacy and reliability of this method were further verified using median nerve stimulation in OPM-MEG experiments. This approach offers a novel solution for applying OPM-MEG in contexts involving participants’ head movements.