Signal Space Separation for Spin-Exchange Relaxation-Free magnetometer

Magnetoencephalography (MEG) provides a real-time, non-invasive investigation of brain activity, which is very important for deep understanding of neuroscience. But MEG signals are often contaminated by various artifacts. Signal space separation (SSS) is a technique based on quasi-static Maxwell equations and Laplace equations. It can be used as a spatial filter for MEG signals denoising. In this paper, the SSS spatial filtering is carried out with multi-channel Spin-Exchange Relaxation-Free (SERF) magnetometer equipment for the first time, and the signal fluctuation is significantly reduced after filtering. One automatic method has been developed in this study to find the best SSS parameters based on the Sequential Least Squares Programming. This paper used MNE-pyhton software to generate spatial noise and internal source signal, and compared the denoised signal with internal simulation signal to find the optimal SSS parameters. The experimental results showed that the auditory evoked response was more obvious and the signal-to-noise ratio of the signal is improved in the induced period after SSS filtering.