Dear Alessandra,
I cannot see any affine registration matrices in your directories,
and the misalignment looks like something that could easily be
due to a failed affine or missing affine registration. Did you run
the FLIRT command to get a 12 DOF (affine) registration before
running FNIRT? It is step 1 in the online instructions.
If you have done the affine registrations then let me know and please
provide the registration outputs so that we can debug the problem.
Also note that there were no README files in the directories on the
website.
All the best,
Mark
On 9 Feb 2012, at 14:57, Alessandra Carla' wrote:
> Hi all.
>
> We tried to perform a non-linear registration between a T1-image of a subject and the MNI152.
>
> We followed the instructions in http://www.fmrib.ox.ac.uk/fslcourse/lectures/practicals/reg/index.html.
>
> MRimages and relative output are in http://hpmagic5.unile.it/BRAIN_IMAGE/
>
>
> We performed two tests using the command list you can see in the README file.
>
> In test1 we used the original T1_2_MNI152_2mm.cnf.
> in test2 we modified configuration as in T1_2_MNI152_2mm_mod.cnf
>
> The output are not satisfactory in both tests.
>
> Please, could you suggest us which parameters we have to use in order to obtain a good non-linear registration?
>
> Many thanks and best wishes.
>
> Alessandra Carlà
>
|