Reorient images to AC-PC for registration plots

[tsalo]

What would you like to see added in fMRIPrep?

This is a bit of an edge case, but for one of the studies I'm helping out with, they are positioning participants with a pretty extreme tilt in the scanner. The registration plots produced by fMRIPrep and ASLPrep are hard to interpret because of this. If possible, it would be nice to reorient the brain ahead of time.

One idea would be to do an AC-PC alignment as part of the anatomical processing, like how QSIPrep does it, so that the desc-preproc_T1w would be rigid-body aligned with a template.

Do you have any interest in helping implement the feature?

Yes

Additional information / screenshots

No response

[effigies]

You're wanting the image actually resampled to rough cardinal alignment, or the diagnostic images?

For the BOLD images, at least, the diagnostic images are specifically in the boldref space to avoid resampling artifacts in the images that do not appear in the data.

For structural images, where in the workflow are you imagining this? I guess right after constructing the template, before anything else? What should be passed to FreeSurfer, which does its own ACPC alignment?

For BOLD, we could add an option to realign the diagnostic images in reports with a note that there may be resampling artifacts that are not present in the data.

[tsalo]

I'm not sure what makes the most sense. I suppose I was leaning toward focusing on structural images, and making it an initial step before constructing the template, so that the image-to-T1wref (or whatever it's called) transform includes ACPC alignment.

I don't think tilted BOLD figures are as much an issue as tilted structural figures.