Hi Robert,
I am not sure I understand your last sentence....
Here is some advise on trouble shooting this:
1. Make sure you provide MNI->Diff as --xfm, and Diff->MNI as --invxfm
2. Run probtrackx from the M1-MNI mask setting step size to 2 (no waypoints etc.), this will produce a fdt_paths that looks like your mask in MNI space
2.1. If it does, then registration is good, and you should look at orientations (V1, dyads1)
2.2. If it doesn't, then it's a registration problem: make sure (1) is correct, and test these transform using applywarp.
If these don't help, then feel free to upload your data on our server.
Cheers,
Saad.
On 5 Oct 2010, at 16:29, Robert Schulz wrote:
> Hello FSLers,
>
> I have 1mm³ structural T1 in MNI space in which seed mask are created. FLIRT/FNIRT registration to the diffusion nodif_brain resulted in the according str2diff and diff2str warp fields. I have checked the matching via applywarp to both nodif and t1 with excellent results. But now when I enter the masks into probtrack and give the warp files for nonlinear xfms, 300 voxels in M1-MNI to spinal give me 2 tracks. The resulting fdt-path file contains only a few red spots going down. Applying both the warp fields to the M1 masks shows me that voxels get lost, sometimes the whole mask is disappearing in the diff-space, e.g. the spinal thin 1 voxel layer as the target mask.
>
> Any comments? Where is the problem? Wrong approach?
>
> Thanks very much for your help. I am looking forward to your answers.
>
> Kind regards,
>
> Robert
>
--
Saad Jbabdi
University of Oxford, FMRIB Centre
JR Hospital, Headington, OX3 9DU, UK
(+44)1865-222466 (fax 717)
www.fmrib.ox.ac.uk/~saad
|