Skip to content

DWI models: Define "bvec" orientation reference#98

Merged
Lestropie merged 3 commits intobep-016from
bvec_reference_axes
Jun 5, 2024
Merged

DWI models: Define "bvec" orientation reference#98
Lestropie merged 3 commits intobep-016from
bvec_reference_axes

Conversation

@Lestropie
Copy link
Copy Markdown
Collaborator

Addresses part of #94.

While I have not yet completed all of the relevant work, I have at least convinced myself of the fact that the eigenvector images produced by FSL's dtifit are not defined with respect to the image axes; rather, they use the same convention as the bvecs format. Therefore, to be able to write such data to a BIDS-compliant dataset without image intensity modification, that coordinate system needs to be included in the specification.

While I suspect the bedpostx outputs use the same convention, I've not yet confirmed that for #94, so I'm deferring that for now.

@Lestropie Lestropie self-assigned this May 15, 2024
- FSL bedpostx command fibre orientations stored as spherical coordinates are confirmed to use the "bvec" reference frame.
- Enforcing azimuth angle to obey the right hand rule about the zenith axis would necessitate direct modification of image intensities from bedpostx outputs to store. Therefore it would be preferable to instead define the sign of the azimuth angle based on the direction of the second reference axis.
Closes #95.
Closes #94.
@arokem
Copy link
Copy Markdown
Collaborator

arokem commented Jun 5, 2024

I think this is ready to go, right? Could you take a look/resolve the conflicts? I just updated the bep-16 branch with a merge from bids-standard/bids-specification:master

Conflicts:
	src/derivatives/05-diffusion-derivatives.md
@Lestropie
Copy link
Copy Markdown
Collaborator Author

I'm not 100% happy with the description of the azimuth angle. I wonder if it needs to be explicit about the trigonometric relationships involved in either direction. Another prospect to consider would be tailored figures showing the interpretations of coefficients relative to any arbitrary reference frame. But it's good enough, so I'll merge and if a better idea comes across anyone we can revisit.

@Lestropie Lestropie merged commit 1ce38c2 into bep-016 Jun 5, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants