I found the issue. Since I had retained only 1 subject from your dataset but all five empty room recordings, the script was not associating any of the five noise covariance matrices with the subject causing the failure of source localisation module.
Now that I am able to reproduce the result on data set published in the paper, I am using my own resting state MEG data but I do not have empty room recordings, nor head shape points. I could follow steps listed for GUI to process the data. But I get the following error while trying to project the psd values onto the cortical surface.
One surface has an atlas "Structures", the other does not. You need to use surfaces coming from the same software to project sources.
How do I resolve this?
Best,
Pradeep