Dear Jo,
Partial volume registration when the anatomy differs is extremely challenging and sometimes it just wont work. The best thing would be if you had a full brain image of the same subject that you could register your partial T2-weighted image to. Then you can register the full brain to the MNI and concatenate these transformations to get the partial T2-weighted to MNI transformation.
Your use of nudge and flirt seem fine, but I would strongly recommend that you try the above suggestion instead, since it is likely to be the change of anatomy that is causing you the biggest problems.
All the best,
Mark
On 23 Oct 2013, at 18:05, Jo Hale <[log in to unmask]> wrote:
> Hi,
>
> I'm trying to coregister a partial volume T2-weighted image to the MNI brain. With previous subjects I've been able to use FLIRT with 3DOF to do this; however I am struggling to do so with this participant.
>
> I found a post from Mark Jenkinson from 2007, which seemed to address this issue and suggested using the Nudge GUI. I have generated an overlay using NUdge which looks sensible and then went ahead as the post recommended to attempt to generate a 'cost-function weighting volume', creating a mask of 1's where the brain is. I then ran the following in the command window (again as suggested in the earlier post)
>
> flirt -in partial_brain -ref anatomical -dof 6 -nosearch -init nudge_matrix.mat -inweight partial_brain_weighting_volume -out partial2anat -omat partial2anat.mat
>
> However, the outputted image doesn't look like the overlay I had when using Nudge and doesn't line up with the MNI. Could anyone suggest where I may be going wrong Or an alternative method for doing this
>
> Thanks and Best Wishes,
>
> Jo
|