Source analysis was conducted using a linearly constrained minimum variance beamformer (LCMV; Van Veen et al., 1997), which applies a spatial filter to the MEG data at each vertex of the 3D cortical mesh, in order to maximize signal from that location whilst attenuating signals elsewhere. Beamforming weights were calculated by combining the covariance matrix of the sensor data with leadfield information. Due to rank reduction following data cleaning with Maxfilter, the covariance matrix was kept at a rank below 64 components, which explained 99% of the variance. For all analyses, a common filter was used across baseline and grating periods, and a regularization parameter of lambda 5% was applied.
