MT has 4 dimensions for Calgary

[valosekj]

acq-MT_MTS.nii.gz images for the Calgary site (GE Discovery scanner) are 4D, see:

$ sct_image -i sub-cal056_ses-M0_acq-MT_MTS.nii.gz -header fslhd 
...
dim1        256
dim2        256
dim3        18
dim4        3
...

Visually, it looks that the null 3D volume represents acq-MTon_MTS:

The first 3D volume represents acq-MToff_MTS:

And the second 3D volume represents acq-T1w_MTS:

4D images are not optimal for processing, e.g., for the co-registration to T2w (which does not work for 4D images). I suggest splitting the 4D image into three individual images.

Additionally, to be compliant with the proposed derivatives conventions, I would use the following naming for the MT images:

acq-MTon_MTS → flip-1_mt-on_MTS
acq-MToff_MTS → flip-1_mt-off_MTS
acq-T1w_MTS → flip-2_mt-off_MTS

(This would also include renaming the MT images for other sites and updating the git-annex dataset)

[jcohenadad]

Good catch! Yes, this is the way to go for these data.

[valosekj]

Fixed in https://github.com/neuropoly/data-management/issues/211 --> closing this issue