search

bids-standard / bids-specification

Brain Imaging Data Structure (BIDS) Specification
https://bids-specification.readthedocs.io/
Creative Commons Attribution 4.0 International
277 stars 162 forks source link

filenames for T1w/T2w ratio images #462

Open loj opened 4 years ago

loj commented 4 years ago

I've been reading BEP003 (#265), and I'm uncertain how to handle the naming of T1w/T2w ratio images (ex: the HCP dataset <sub>/T1w/T1wDividedByT2w.nii.gz).

Perhaps <source_keywords>_desc-dividedByT2w_T1w.nii.gz?

Any thoughts, opinions, or insights are most welcome. :-)

nicholst commented 4 years ago

What about T1wT2wratio for the image type? Concise while hopefully being unambiguous, and it seems like you'd never want to mistakenly put it to use if you're looking for a T1w image.

effigies commented 4 years ago

The way I would interpret *_desc-dividedByT2w_T1w.nii.gz would be that the image can be used as a T1w map, but I should be aware that "dividedByT2w" happened to it. If that's reasonable, then I think that's fine, and it's certainly valid.

If this measure is not interpretable as a T1w (as it seems unlikely to be), then it should have its own suffix. I would be fine with @nicholst's suggestion of T1wT2wratio. Or is there another (briefer or more meaningful) name you can think of?

yarikoptic commented 4 years ago

FWIW, I do like T1wT2wratio -- it is somewhat inline with T1w as "T1 weighted" - so the additional notion ("weighted") is attached at the end, here ratio similarly follows. (Un)fortunately it would never become T1wT2w-ratio and T1wT2w+ratio kinda would make no sense, so T1wT2wratio might be the best (could not come up with shorter and not more confusing)

effigies commented 4 years ago

I see no reason not to add this to the derivatives PR. Sorry things are moving so slowly there... I'm aiming to finish up #301 on Monday, which should unblock a lot.

yarikoptic commented 4 years ago

so I guess the proper place to introduce it would be a new section within https://github.com/bids-standard/bids-specification/pull/207/files#diff-4ff87c6142082e27a0f84319b01b85ffR89 ?

effigies commented 4 years ago

I would say yes to that file, but that section specifically focuses on surface-mapped anatomical scalar derivatives. I think we have two options:

1) Change to "Anatomical scalar derivatives", and include patterns for volumetric NIfTIs, surface-mapped GIFTIs and volume/surface-mapped CIFTI-2s. 2) Add a "Volumetric anatomical scalar derivatives" section.

I would be inclined toward option 1, but we should check in with @edickie and @ahoopes, who wrote this section.

edickie commented 4 years ago

Yeah, I think I agree with option 1 - "Anatomical scalar derivatives" that, in this case, are volumetric NifTIs. Actually, "myelinmap" is already a proposed suffix in this section - but I agree that "T1wT2wratio" might be a better choice of suffix...