Hi everyone,
I am trying to do something that I thought would have been pretty easy, and indeed is not.
I want to have some T1 images and some FA images in (exactly) the same space. To do that, I thought that I could do as follow:
1) BET the T1
2) flirt the FA onto the (brain extracted) T1 (default options), output the affine matrix
3) flirt the T1 (brain extracted) onto the MNI 2mm template, output the affine matrix
4) fnirt the T1 (non brain extracted) onto the MNI 2mm tamplete using the affine calculated in the previous step and the relative config file
5) apply the warp obtained in 4 to the FA image, using the FA -> T1 affine matrix as premat.
However, although the registration of FA on T1 is good, and the T1 -> MNI is also good, the result of step 5 is pretty poor. The white matter tract are not well aligned with the template and the boundary of the normalized FA looks spiky and not well aligned with the boundary of the normalized T1.
I am doing something wrong ?
Alain
|