Hello,
I am doing a DTI study and in my study I have T1, T2 and DW images. I want to have the diffusion tensors on individual T1 space and not on the DTI space. I guess there are two ways to do this:
1) a) Use "eddy_correct" on DW image.
b) Use FLIRT to register T2 to T1 to obtain T2toT1 transformation matrix.
c) Use FLIRT to register b0 to T2 to obtain b0toT2 transformation matrix.
d) Cascade T2toT1 and b0toT2 to get b0toT1 transformation matrix.
e) Use this b0toT1 in "applyxfm" to register b0 and directional images to T1 space.
f) Change the gradient directions (bvec) according to the rotational part of the b0toT1 transfer matrix.
e) Apply "dtifit"
2) a) Use eddy_correct, bet and dtifit to DW image.
b) Use FLIRT to register T2 to T1 to obtain T2toT1 transformation matrix.
c) Use FLIRT to register b0 to T2 to obtain b0toT2 transformation matrix.
d) Cascade T2toT1 and b0toT2 to get b0toT1 transformation matrix.
e) Use "applyxfm" with b0toT1 transformation matrix for L1, L2, L3
f) Use "vecreg" with b0toT1 transformation matrix for V1, V2, V3 or diffusion tensor output
Is there any problem with those pipelines and which of those pipelines is more elegant? Thanks.
Best,
Ümit Aydin
|