I am trying some FODs reconstruction and the dataset is peculiar in a way: The acquisition was not based on a coil gradient scheme, but the gradient was stable and directions changed by re-orienting the object inside the MF (with stable center), 60 times, on angles pre-set by electrostatic repulsion on sphere surface.
Each direction transformation was obtained by registration of the re-orientations to the reference and converted to a vector.
Could these vectors be directly used in -fslgrad bvecs selection or you spot any issue on the described process/difference with the regular bvecs from gradient schemes?
As long as the reorientation of the gradients is correct, I reckon that should work as expected. Personally, I’d be more worried about matching all the different geometric distortions, etc across volumes, but I assume you’ve got all that under control. I’d love to see what you’re cooking up here…
Hello,
So… the “cooking” in that case doesn’t have to do only with diffusion actually…!
The aim is to study the possibility of deriving FODs in white matter, based on high-angular local phase data (susceptibility induced phase) which regionally exhibit highly anisotropic characteristics, and subsequent comparison with the same analysis on DTI data.
The experiment acquisitions include multi-orientational DTI, STI data on post-mortem brains, and gradient scheme-based DTI acquisitions for reference. The processing is pretty long, including of course calculation of the FODs using MRTRIX.
On a first analysis (we are on the way for a new acquisition), the preliminary results look promising on the main fibers, like CST, despite I have some issues with the coloring of the FODs in mrview as the scanner vs image coordinate system seems to be confused in a non-clear way (for me).
I would be happy to share more details with you if you are interested, and of course any suggestion is more than welcome!
