[ENH] Clarify SEEG tutorial coordinate frames#9601
Conversation
|
I need to add a test for the new method, doing that now |
|
Not to toot my own horn but I really like the applying a I think this works really well for finding channel coordinates from just a CT; the fiducials could be found on the CT and a head to CT transform could be used to get the coordinates in CT space and be applied directly to the montage. |
|
you need to add it to |
drammock
left a comment
drammock
left a comment
There was a problem hiding this comment.
I'm having trouble following 10_ieeg_localize.html, although the places that I find difficult to follow (or where I find typos/awkward passages) appear to have not changed in this PR, so maybe this is not the best place to discuss fixes. @alexrockhill LMK if you would prefer to make those changes in this PR, and if so, we can schedule a time to talk it through.
Sure that's fine, can you talk this morning? Feel free to coordinate on Discord |
|
From discord with @alexrockhill and @drammock we decided to merge this and @alexrockhill will 1) make a PR to clarify the narrative after the GUI work is done (or as part of it) and 2) open a dipy issue to see if SDR provides a direct (non-discretized) mapping from points in the moving space to the static space (i.e., the equivalent of an affine |
|
Thanks @alexrockhill ! |
| # Freesurfer surfaces when they were assigned to the ``raw`` object, they were | ||
| # transformed to "head" so we need to have a transform to put them back | ||
| subj_trans = mne.read_trans(op.join(misc_path, 'seeg', | ||
| 'sample_seeg_trans.fif')) |
There was a problem hiding this comment.
sorry I saw this late. Is it kosher to have sEEG data as fif files and thus be transformed to "head" coordinates? Don't you normally get the coordinates in MRI coordinates as a separate text file or mat file?
There was a problem hiding this comment.
Yeah I agree it's not ideal but I think to be backward compatible with MEG/EEG, FIF requires it to be in the "head" frame. I've thought it a bit like how Freesurfer uses surface RAS as its internal representation frame, MNE uses "head". They are all arbitrary reference frames and at least there is a standard and you know how it's stored. Ideally the coordinate frame could be specified in the FIF but that seems like pretty large change.
* rework transform narrative * add more clarifying language * add test * Update doc/changes/latest.inc Co-authored-by: Eric Larson <larson.eric.d@gmail.com>
5 participants
This clarifies and I think simplifies what's going on with the coordinate transforms for the SEEG case.
I think this will be really helpful for the GUI as they will hopefully be able to follow the same steps.