Nipreps resampler

[effigies]

@oesteban @feilong @mgxd @HippocampusGirl Here is a single-shot resampler that accepts an original BOLD series and an fMRIPrep output directory.

Parts of this are cut-and-pasted from various places in sdcflows, nitransforms and feilong/process, but just dropped in a single place for rapid development. Some of it's new.

Todo:

• [ ] Add slice-timing correction step
• [x] Re-parallelize resampling
• [ ] Convert steps to interfaces for including in fMRIPrep/dMRIPrep
• [ ] Move as much of this as possible to nitransforms
• [ ] Add resampling targeting fsnative
  • [ ] Need to think about the Jacobian here.
• [ ] Consider Jacobians of the T1-MNI warp (we should be able to validate this with ANTs' CreateJacobianDeterminantImage)

[tsalo]

In this Mattermost conversation, @cookpa said:

Hi there, my thinking is that it's correct to modulate for the jacobian in the distortion correction stage, because the signal is being reduced or increased in approximate proportion to the expansion / contraction. I don't think that holds when going to a group template, you're deforming to account for anatomical variation, rather than a signal-modifying imaging artifact. If that makes sense.

Based on that, it doesn't seem like it's necessary to consider Jacobians for the T1-MNI warp for BOLD data.

[oesteban]

Correct

On Sat, May 4, 2024, 22:16 Taylor Salo @.***> wrote:

In this Mattermost conversation https://mattermost.brainhack.org/brainhack/pl/dkmfoby5bjr6fq3b4k9edymrzr, @cookpa https://github.com/cookpa said:

Hi there, my thinking is that it's correct to modulate for the jacobian in the distortion correction stage, because the signal is being reduced or increased in approximate proportion to the expansion / contraction. I don't think that holds when going to a group template, you're deforming to account for anatomical variation, rather than a signal-modifying imaging artifact. If that makes sense.

Based on that, it doesn't seem like it's necessary to consider Jacobians for the T1-MNI warp for BOLD data.

— Reply to this email directly, view it on GitHub https://github.com/nipreps/resampler/issues/1#issuecomment-2094217770, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAESDRRUDI54WECIVGCSKHLZATUTXAVCNFSM6AAAAAA5W7G3RWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDAOJUGIYTONZXGA . You are receiving this because you were mentioned.Message ID: @.***>