ICA fit on raw data with projections

External Email - Use Caution

Dear all,

I create noise projections from meg raw data to remove environmental
artifacts (i.e., street car effects). I usually apply the projections to
the data and then perform a ICA fit to remove other biological artifacts
(i.e., ecg, eog). It seems like noise projections that are applied to the
data are temporarily removed while fitting the ICA object to it. I can see
that since the topographies of the noise projections are being replicated
in ICA topographies. Is it true that even if the projections are applied to
raw data before, they are kept separated while performing ICA fiting and
added afterwards again? Can anyone tell me how it works in MNE?

Best,
Dip
-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.nmr.mgh.harvard.edu/pipermail/mne_analysis/attachments/20191213/f8d9c011/attachment.html

External Email - Use Caution

hi Dip,

you have a discussion in the notes section of ICA doc:

https://mne.tools/stable/generated/mne.preprocessing.ICA.html

does it answer your question?

Alex

External Email - Use Caution

Yes, it does. Thanks!