Hello,
I exported the raw images (after DIFF_PREP) using the FSL sorted option (first button). I'm trying to run some tests in FSL, and was wondering if the gradient table is rotated to account for the registration to the T2? When I fit the tensor in FSL with the exported bvecs, the eigen vectors look a bit strange (even if I rotate X/Y).
In other words, does the tensor need to be calculated in the space the image is currently in (i.e., target space), or should I first ‘back transform’ the image to its original, native space. If there is a way to access the transformation matrix from raw diffusion space to target space (or preferably the inverse of that), I could apply it to the data and see if the tensors look more reasonable then.
Thanks,
Gustavo