Phase contrast imaging acquisition for CSF flow measurements - How to label?

[cmadjar]

Hello,

I have a project that acquires CSF flow using phase contrast imaging acquisition. I am not entirely sure how to label those images so that they are following the BIDS spec.

Here are the three set of images I get:

• CSF flow
• CSF flow magnitude
• CSF flow phase

Does anyone know how I should label those? Thank you!

[tsalo]

@cmedjar, I'm not very familiar with phase-contrast imaging, but this may require a new suffix.

For example, we have an angio suffix for maps that generally come from phase-contrast imaging acquisitions, and there is an open BIDS extension proposal for susceptibility-weighted imaging, which proposes an swi suffix.

Can you explain what the difference between "CSF flow" and "CSF flow magnitude" is? At least for BOLD imaging, the primary data are generally the magnitude. BIDS has the part entity for when you have both magnitude and phase components from the same complex signal, but if there's a third file that has something that's neither magnitude, nor phase, that would complicate things, I think.

[cmadjar]

@tsalo thank you for getting back to me. I honestly don't know much about this type of acquisition. I don't even know what would be the scan type or category it should go within. The protocol name is for these acquisitions is: fl_pc_prospective

Looking at the JSON files, here are the differences I can see between the three files:

• CSF flow
  • SeriesDescription: fl_pc_prospective
  • SequenceName: *pc2d1
  • ImageType: [ORIGINAL, PRIMARY, M, ND]
• CSF flow mag
  • SeriesDescription: fl_pc_prospective_MAG
  • SequenceName: pc2d1_v020in
  • ImageType: [ORIGINAL, PRIMARY, MAG, ND]
• CSF flow phase
  • SeriesDescription: fl_pc_prospective_P
  • SequenceName: *pc2d1_v020in
  • ImageType: [DERIVED, PRIMARY, P, ND]

Not sure how much helpful that information is. Let me know. Thank you again!

[effigies]

Since these images are not currently defined in BIDS, the best bet is to do something BIDSy and then shoot to get it formalized.

I'm going to work off of case 1 described here, which seems to describe your situation.

I might start with something like:

sub-01/
  anat/
    sub-01_part-mag_CSFfl.nii.gz
    sub-01_part-phase_CSFfl.nii.gz
    sub-01_part-rephased_CSFfl.nii.gz

I use anat because the intent seems to be anatomical, to a first approximation. A proposal to standardize might end up using a different data type.

Note also that part-rephased would violate our current part- entity (which must be phase, mag, real or imag). rephased or combined might be worth adding, since this seems to be common to both SWI and CSF flow. This abuse of part is intended to keep a single suffix, to show the images are related. An alternative could be:

sub-01/
  anat/
    sub-01_part-mag_CSFdiff.nii.gz
    sub-01_part-phase_CSFdiff.nii.gz
    sub-01_CSFflow.nii.gz

[cmadjar]

Thank you @effigies for the suggestions! :-) I will pass the message along.

[RayStick]

Hi all, I was linked to this issue from a thread I had on the brainhack bids_general channel.

I collected phase contrast data in the feeding arteries to the brain (left and right ICAs; left and right VAs) to get blood velocity estimates across the cardiac cycle. I also get three outputs from the scanner:
scan_name
scan_name_MAG (magnitude)
scan_name_P (phase)

This data is pulse gated so I get 8 volumes of data in my output file, corresponding to 8 bins of the cardiac cycle. I will consider putting a BIDS proposal in sometime, as I don't think this is supported yet.

My sequence seems similar to the one described above, but perhaps just looking in a different location (CSF vs feeding arteries) therefore this data would be used for different purposes. Either way, the main reason I am commenting is that I think the scan_name without a magnitude or phase flag could be a MIP image (maximum intensity projection) that the scanner makes and outputs - so that's something to check if you're unsure about what that third output is.