noise regularization in group analysis

✉️ Mailing List Archive (read-only)
1

Hello,

I have a naive questions about noise covariance matrix regularization.

For the group analysis, should the same noise covariance regularization
(e.g 'shrunk') be applied to all the subjects or would it be better to
apply regularization that is 'optimal' ('auto' option in the mne.
compute_covariance) for each particular subject ?

With regards,
Elena
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20161024/4f54ac45/attachment.html

2

Hi,

I have a naive questions about noise covariance matrix regularization.

For the group analysis, should the same noise covariance regularization (e.g
'shrunk') be applied to all the subjects or would it be better to apply
regularization that is 'optimal' ('auto' option in the
mne.compute_covariance) for each particular subject ?

To me "optimal" is recommended but I am curious if you see any
impact of "optimal" vs. "all shrunk" on the group results.

Alex