`b0` files are mistakenly filed

[kousu]

A bunch of files are tagged acq-b0 The files affected are all DWI images:

u108545@joplin:~/data-multi-subject$ find . -name '*acq-b0*'
./sub-beijingVerio03/dwi/sub-beijingVerio03_acq-b0_dwi.nii.gz
./sub-beijingVerio03/dwi/sub-beijingVerio03_acq-b0_dwi.json
./sub-vallHebron02/dwi/sub-vallHebron02_acq-b0_dwi.nii.gz
./sub-vallHebron02/dwi/sub-vallHebron02_acq-b0_dwi.json
./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.nii.gz
./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.json
./sub-amu02/dwi/sub-amu02_acq-b0_dwi.nii.gz
./sub-amu02/dwi/sub-amu02_acq-b0_dwi.json
./sub-vallHebron01/dwi/sub-vallHebron01_acq-b0_dwi.json
./sub-vallHebron01/dwi/sub-vallHebron01_acq-b0_dwi.nii.gz
./sub-strasbourg04/dwi/sub-strasbourg04_acq-b0_dwi.json
./sub-strasbourg04/dwi/sub-strasbourg04_acq-b0_dwi.nii.gz
./sub-vallHebron04/dwi/sub-vallHebron04_acq-b0_dwi.nii.gz
./sub-vallHebron04/dwi/sub-vallHebron04_acq-b0_dwi.json
./sub-vallHebron05/dwi/sub-vallHebron05_acq-b0_dwi.nii.gz
./sub-vallHebron05/dwi/sub-vallHebron05_acq-b0_dwi.json
./sub-beijingVerio04/dwi/sub-beijingVerio04_acq-b0_dwi.nii.gz
./sub-beijingVerio04/dwi/sub-beijingVerio04_acq-b0_dwi.json
./sub-beijingVerio02/dwi/sub-beijingVerio02_acq-b0_dwi.nii.gz
./sub-beijingVerio02/dwi/sub-beijingVerio02_acq-b0_dwi.json
./sub-vallHebron07/dwi/sub-vallHebron07_acq-b0_dwi.json
./sub-vallHebron07/dwi/sub-vallHebron07_acq-b0_dwi.nii.gz
./sub-beijingVerio01/dwi/sub-beijingVerio01_acq-b0_dwi.json
./sub-beijingVerio01/dwi/sub-beijingVerio01_acq-b0_dwi.nii.gz
./sub-amu05/dwi/sub-amu05_acq-b0_dwi.json
./sub-amu05/dwi/sub-amu05_acq-b0_dwi.nii.gz
./sub-amu03/dwi/sub-amu03_acq-b0_dwi.json
./sub-amu03/dwi/sub-amu03_acq-b0_dwi.nii.gz
./sub-strasbourg05/dwi/sub-strasbourg05_acq-b0_dwi.nii.gz
./sub-strasbourg05/dwi/sub-strasbourg05_acq-b0_dwi.json
./sub-strasbourg02/dwi/sub-strasbourg02_acq-b0_dwi.nii.gz
./sub-strasbourg02/dwi/sub-strasbourg02_acq-b0_dwi.json
./sub-strasbourg06/dwi/sub-strasbourg06_acq-b0_dwi.json
./sub-strasbourg06/dwi/sub-strasbourg06_acq-b0_dwi.nii.gz
./sub-strasbourg01/dwi/sub-strasbourg01_acq-b0_dwi.json
./sub-strasbourg01/dwi/sub-strasbourg01_acq-b0_dwi.nii.gz
./sub-strasbourg03/dwi/sub-strasbourg03_acq-b0_dwi.json
./sub-strasbourg03/dwi/sub-strasbourg03_acq-b0_dwi.nii.gz
./sub-vallHebron06/dwi/sub-vallHebron06_acq-b0_dwi.json
./sub-vallHebron06/dwi/sub-vallHebron06_acq-b0_dwi.nii.gz
./sub-amu01/dwi/sub-amu01_acq-b0_dwi.nii.gz
./sub-amu01/dwi/sub-amu01_acq-b0_dwi.json
./sub-amu04/dwi/sub-amu04_acq-b0_dwi.json
./sub-amu04/dwi/sub-amu04_acq-b0_dwi.nii.gz

The only mention I can find of b0 in BIDS is about using the fmap/ datatype/subfolder, and it recommends four schemas depending on how many files you have:

• sub-*/{ses-*/,}anat/*_{magnitude{1,2}, phasediff}.nii.gz
• sub-*/{ses-*/,}anat/_{magnitude{1,2},phase{1,2}}.nii.gz
• sub-*/{ses-*/,}anat/_{magnitude,fieldmap}.nii.gz
• sub-*/{ses-*/,}anat/_epi.nii.gz

If our acq-b0 files are really fieldmaps, they only slot they fit into is the _epi case. But again, they're under dwi/ right now so I don't know what to make of the situation.

[jcohenadad]

If our acq-b0 files are really fieldmaps

no they are not. It is quite confusing. These are not B0 fieldmap (where B0 refers to the static magnetic field) but b=0 images (where b refers to the diffusion-encoding b-value).

this additional image is acquired in order to improve SNR when computing diffusion tensor metrics.

i'm not sure how we can make BIDS happy with this special image type...

[kousu]

Maybe it's

Single-Band Reference	sbref	Single-band reference for one or more multi-band dwi images.

so then

git mv ./sub-vallHebron06/dwi/sub-vallHebron06_acq-b0_dwi.nii.gz ./sub-vallHebron06/dwi/sub-vallHebron06_sbref.nii.gz

?

But maybe it's fine like it is. That spec page is pretty vague, only mentioning

The OPTIONAL acq-<label> key/value pair corresponds to a custom label the user may use to distinguish different sets of parameters.

which is also appropriate since the scanner set the different parameter set b=0!

Do you think the spec is implying or just suggesting that _dwi.nii.gz implies _dwi.bval and _dwi.bvec here?

sub-<label>/
   [ses-<label>/]
       dwi/
           sub-<label>[_ses-<label>][_acq-<label>][_dir-<label>][_run-<index>][_part-<label>]_dwi.bval
           sub-<label>[_ses-<label>][_acq-<label>][_dir-<label>][_run-<index>][_part-<label>]_dwi.bvec
           sub-<label>[_ses-<label>][_acq-<label>][_dir-<label>][_run-<index>][_part-<label>]_dwi.json
           sub-<label>[_ses-<label>][_acq-<label>][_dir-<label>][_run-<index>][_part-<label>]_dwi.nii[.gz]
           sub-<label>[_ses-<label>][_acq-<label>][_dir-<label>][_run-<index>][_part-<label>]_sbref.json
           sub-<label>[_ses-<label>][_acq-<label>][_dir-<label>][_run-<index>][_part-<label>]_sbref.nii[.gz]

[jcohenadad]

Single-Band Reference

no it's not that (for the record: SB reference refers to images acquired with SMS factor of >1)

But maybe it's fine like it is.

i can sleep at night with this status quo

Do you think the spec is implying or just suggesting that _dwi.nii.gz implies _dwi.bval and _dwi.bvec here?

dwi data definitely need a bval and a bvec in order to be able to compute DTI metrics

[kousu]

Do you think the spec is implying or just suggesting that _dwi.nii.gz implies _dwi.bval and _dwi.bvec here?

dwi data definitely need a bval and a bvec in order to be able to compute DTI metrics

Okay well, stupid question then: doesn't the BIDS naming scheme would seem to mean there's missing files here? The existence of

./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.json

should imply

./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.bval
./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.bvec

but if that's wrong then _dwi is the wrong suffix for these files isn't it?

[jcohenadad]

Do you think the spec is implying or just suggesting that _dwi.nii.gz implies _dwi.bval and _dwi.bvec here?

dwi data definitely need a bval and a bvec in order to be able to compute DTI metrics

Okay well, stupid question then: doesn't the BIDS naming scheme would seem to mean there's missing files here? The existence of

./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.json

should imply

./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.bval
./sub-vallHebron03/dwi/sub-vallHebron03_acq-b0_dwi.bvec

but if that's wrong then _dwi is the wrong suffix for these files isn't it?

ah! sorry i misunderstand your point above-- a b=0 image has (by definition) no bvec/bval associated with it.

that being said, we could also make the point that a very low b-value would be considered b=0 and labeled as is (and in this case a bvec/bval should be needed)-- so, i'm not sure what to advise here 🤷‍♂️

[kousu]

Maybe we can bring it up on the BIDS bug tracker when we get a free moment?