Various comments on the process_data.sh script and README

[jcohenadad]

README:

• describe the type of data, the contrasts available and the file names, example spine-generic
• mention the native resolution of the image (relevant for the resampling part)

process_data:

• replace: https://github.com/sandrinebedard/Projet3/blob/2f35c88ef66e8db8640b3c386b9a60930a8f87bc/process_data.sh#L46 with:

  sct_label_utils -i ${file_seg}_labeled_discs.nii.gz -keep 3 -o ${FILELABEL}.nii.gz

• replace: https://github.com/sandrinebedard/Projet3/blob/2f35c88ef66e8db8640b3c386b9a60930a8f87bc/process_data.sh#L114 with:

  sct_register_to_template -i ${file_t1}.nii.gz -s ${file_t1_seg}.nii.gz -ldisc ${file_label}.nii.gz -c t1 -param step=1,type=seg,algo=centermassrot:step=2,type=seg,algo=syn,slicewise=1,smooth=0,iter=5:step=3,type=im,algo=syn,slicewise=1,smooth=0,iter=3 -qc ${PATH_QC} -qc-subject ${SUBJECT}

• we might want to double check the native resolution: https://github.com/sandrinebedard/Projet3/blob/2f35c88ef66e8db8640b3c386b9a60930a8f87bc/process_data.sh#L132

• Add QC after this step, in case patient moved (or MR tech repositioned) between T1 and T2_FLAIR: https://github.com/sandrinebedard/Projet3/blob/2f35c88ef66e8db8640b3c386b9a60930a8f87bc/process_data.sh#L145

[sandrinebedard]

@jcohenadad The native resolution of T1w is 2.4x2.4x2.4 mm and for T2w is resolution: 1.05x1x1 mm. For resampling, I wasn't sure if I have to put the native resolution in the parameter -m , in this line for T1w :

https://github.com/sandrinebedard/Projet3/blob/4ef8afeeb8a1081a269373d59cc458bee41c59c5/process_data.sh#L98

and in this line in T2w: https://github.com/sandrinebedard/Projet3/blob/4ef8afeeb8a1081a269373d59cc458bee41c59c5/process_data.sh#L132

[jcohenadad]

The native resolution of T1w is 2.4x2.4x2.4 mm

where did you get that information? it seems wrong. E.g.:

julien-macbook:~ $ fslhd /Users/julien/data/ukbank/1000032/T1/T1_orig_defaced.nii.gz 
filename    /Users/julien/data/ukbank/1000032/T1/T1_orig_defaced.nii.gz
size of header  348
data_type   FLOAT32
dim0        3
dim1        208
dim2        256
dim3        256
dim4        1
dim5        1
dim6        1
dim7        1
vox_units   mm

suggesting 1mm iso for the T1w. Maybe you looked at the DWI scan?

For resampling, I wasn't sure if I have to put the native resolution in the parameter -m , in this line for T1w :

it's the parameter -mm (not -m), and yes, you could put a fixed resolution close to the native, so in this case 1mm iso

[sandrinebedard]

Yes! You are right, I checked the wrong image for T1w, I also verified with UK biobank brain MRI documentation and for T1w, the resolution is 1x1x1mm.

https://biobank.ctsu.ox.ac.uk/crystal/crystal/docs/brain_mri.pdf (section 2.4) Thank you!