A Pose Calibration Method for Sensors in OPM-MEG Systems with Experimental Validation

Optically pumped magnetometer-magnetoencephalography (OPM-MEG) is a noninvasive brain functional imaging technique that requires solving the source localization problem by inferring internal brain current sources from scalp magnetic field measurements. The position and orientation of OPMs relative to the subject's head are part of MEG forward modeling, and failure to accurately obtain their pose parameters will severely affect the accuracy of MEG source localization. Current optical scanning methods cannot obtain the equivalent measurement pose information of OPMs. Therefore, this article proposes a large coil calibration method suitable for OPM-MEG to directly determine the true measurement pose of OPM arrays. First, the large coils generate a magnetic field with a large uniformity volume to calibrate the pose and gain of OPMs, and then the iterative closest point (ICP) algorithm is employed to project the coil coordinate system onto the subject's head magnetic resonance imaging (MRI) coordinate system. This method overcomes the disadvantage that optical scanners and digitization devices cannot obtain accurate pose information of sensors. Subsequently, four equivalent current dipoles (ECDs) in a dry phantom were localized using the OPM array before and after calibration. The results show that the average localization error was reduced by a maximum of 90.1 % and a minimum of 35.7 %. The method was then applied to human somatosensory cortex experiments, where calibrated signals exhibited stronger postmovement beta rebound, demonstrating the importance of OPM calibration. This calibration method provides technical support for further improving the accuracy of MEG source localization.