Julien Cohen-Adad

OK, first of all, I apologize but I misunderstood what `moco_params.tsv` was showing in the first place. I edited my comment https://github.com/spinalcordtoolbox/spinalcordtoolbox/issues/4344#issue-2086237735. Now, I have a bit of trouble understanding...

Thank you for the clarification @CarolineLndl, additional questions: > motion_magn is the mean magnitude across slices (average of magnitude across all slices for each time point) By 'magnitude' do you...

right-- i guess i was thinking more of an approach where we analyse the image data directly (as opposed to training a model from the CSA).

> I added the full SI axis I'm afraid in other situation this will be sub-optimal. Eg: for brain/cspine scans, where half of the image will be occupied by the...

> 20% of the width of the bounding box surrounding the entire segmentation Sounds reasonable, but I would not make it relative to the box. Instead, I would suggest a...

Thank you so much for the thorough analysis @joshuacwnewton ! This is very helpful, and it somewhat comforts me about going with the 'starting fresh' alternative 😅

We can probably close this issue

Funny, the CHANGES.md does not fully reflect the description in the release log: In any case, we need to find a way to systematically report PAM50's changelogs into SCT's changelog...

> Maybe there are other major dependencies where this also makes sense? I can also think of the deep learning models

Results from https://github.com/spinalcordtoolbox/spinalcordtoolbox/pull/4302/commits/3f301c20aa7337b4377aad09b1fe890516c0980a using `CRL_1501A_3T_20231102_123128062_csf_flow_fl2d_tra_6through-plane-C2-3_20231102093852_36.nii` (private data) Voxel in the spinal cord (expect minimum motion: [-0.2, 0.4] cm/s)  Voxel in the CSF (expect...