Hi,
I have just tried different SNRs (--snr) in mne_make_movie with one data
set (simulated dipoles), and didn't see big differences for SNRs of 1
and 1000, for example (correlations of distributions ~0.98). I would
have expected huge differences in smoothness, if SNR is inversely
related to the regularization parameter lambda. Has anyone made similar
experiences, or does this point to a problem in my set-up?
Regards,
Olaf