Hi Verena,
when transforming a ROI, you might also consider using nearest neighbour interpolation when transforming the ROI with flirt:
-interp nearestneighbour
This will keep the label values unchanged.
Cheers,
Jan
----- Ursprüngliche Mail -----
> Dear FSL-experts,
>
> please excuse me for coming up with an old topic, but I got stuck
> there.
> I tried to transform my ROI files from structural (T1) to diffusion
> (FA) space applying the following steps:
>
> FLIRT: -in: FA ref: T1 –out: diffusion2structural –omat:
> diffusion2structural.mat
>
> convert_xfm –omat: structure2diffusion.mat -inverse:
> diffusion2structural.mat
>
> FLIRT: -in: ROI_structural mask –ref: FA -applyxfm – init:
> diffusion2structural.mat –dof 6 -out: ROI_diffusion mask
>
> However, the ROIs (marked in red) that were originally placed in the
> subthalamic nucleus in the structural space (Fig1) look distorted and
> out of place in the diffusion space ( Fig.2).
>
> Where did I go wrong?
>
> Thanks a lot for your kind support,
> Best regards,
>
> Verena
|