Measuring functional connectivity in MEG: A multivariate approach insensitive to linear source leakage

► We demonstrate a method to image cortico-cortical interactions in MEG data. ► The method employs a multivariate statistical framework in MEG source space. ► The technique has reduced sensitivity to signal leakage across voxels. ► In simulation we show spatial accuracy and slightly conservative false positive rate. ► In real data we show cortico-cortical interaction between left and right motor cortex.