search

mne-tools / mne-hcp

MNE-HCP for processing the Human Connectome Project MEG data using MNE-Python
mne-tools.github.io/mne-hcp
BSD 3-Clause "New" or "Revised" License
22 stars 25 forks source link

more info on how to remove artefacts for epochs objects #18

Closed JohnGriffiths closed 7 years ago

JohnGriffiths commented 7 years ago

Hi.

Is it the case that arterfact rejection with removal of bad channels / epochs and ICA components needs to be done before source localization on the epochs data?

I know you have some pointers in the readme.md, but would it be possible to get a little bit more info on exactly how to use the annotations to reject segments / epochs in the epochs object, and likewise with the ICA components?

Reason I'm stuck is that it seems all of the MNE documentation on this starts with Raw objects; not with epochs objects. Not sure how to proceed.

Ta,

john

dengemann commented 7 years ago

Hi John,

if you start form the epochs you're already set, there's nothing to do for the MEG data. The epochs are preprocessed (bad ICA components subtracted, etc.); You will need to compute the inverse solution and what else you need. Hope this helps.

Denis

On Thu, Aug 25, 2016 at 11:21 PM John Griffiths notifications@github.com wrote:

Hi.

Is it the case that arterfact rejection with removal of bad channels / epochs and ICA components needs to be done before source localization on the epochs data?

I know you have some pointers in the readme.md, but would it be possible to get a little bit more info on exactly how to use the annotations to reject segments / epochs in the epochs object, and likewise with the ICA components?

Reason I'm stuck is that it seems all of the MNE documentation on this starts with Raw objects; not with epochs objects. Not sure how to proceed.

Ta,

john

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/mne-tools/mne-hcp/issues/18, or mute the thread https://github.com/notifications/unsubscribe-auth/AB0fioFxwUwEAeNzfmbVqs3sOmdGi-FNks5qjgdXgaJpZM4JthDB .

JohnGriffiths commented 7 years ago

Ok great. This was what I had thought, but then was a bit thrown by some comments saying 'bad segments and channels have to be read and used' and 'ICA components related to eye blinks and heart beats need to be removed from the data'. I guess those bits are referring specifically to replicating HCP sensor space analyses (?).

Thanks for speed reply.

dengemann commented 7 years ago

Yes, this is related to reprocessing from scratch. Hope it's clearer now. On Fri, 26 Aug 2016 at 00:48, John Griffiths notifications@github.com wrote:

Ok great. This was what I had thought, but then was a bit thrown by some comments saying 'bad segments and channels have to be read and used' and 'ICA components related to eye blinks and heart beats need to be removed from the data'. I guess those bits are referring specifically to replicating HCP sensor space analyses (?).

Thanks for speed reply.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/mne-tools/mne-hcp/issues/18#issuecomment-242569043, or mute the thread https://github.com/notifications/unsubscribe-auth/AB0fipMuhJyJ7xMdrq_f7MYE0skouP4Hks5qjhu6gaJpZM4JthDB .

dengemann commented 7 years ago

@JohnGriffiths have a look now at the repo, we had an unofficial release, the website is up and tests are running on continuous integration, travisCI so far. I'll close this for now. Please feel free to comment, open issues or contribute. We now have a fully fledged dev setup.