Spatiotemporal extended homogeneous field correction method for reducing complex interference in OPM-MEG

Abstract

Novel magnetoencephalography (MEG) systems based on optically pumped magnetometers (OPMs) have undergone rapid development in recent years. However, environmental interference significantly degrades data quality. When the number of sensors in the OPM-MEG system is small, the traditional subspace projection denoising algorithms reliant on sensor space oversampling will be difficult to apply. Although the recently proposed homogeneous field correction (HFC) method resolves this problem by constructing a low-rank spatial model, it lacks the ability to suppress complex environmental interference such as nonhomogeneous fields. Therefore, this paper proposes a novel OPM-MEG environmental interference suppression method based on HFC. We first use a projection operator constructed from a sensor orientation matrix to project original data and empty-room noise data onto the null space of the homogeneous field; this enables dimensionality reduction to eliminate homogeneous field interference. The remaining interference is then suppressed through subspace projection in the space and time domains. We compare our method to four benchmark algorithms based on simulations and somatosensory-evoked experiments. The experimental results demonstrate that the proposed method has better interference suppression performance than the benchmark algorithms. Therefore, our method can provide high signal-to-noise ratio data for subsequent clinical applications and brain scientific research.