Dear Paolo,
Why do you want to run fnirt between these images?
Are they from the same subject?
If so, then you should run flirt instead (which can do mutual info).
If not, do you have a T1 image from each subject? In that case
you can combine within-subject between DTI and T1 and between-
subject (with fnirt) between T1's.
All the best,
Mark
On 3 Nov 2011, at 18:04, Paolo Zaffino wrote:
> Dear FSL users/developers,
> I am new of FSL and I am trying to register an MRI diffusion weigthed on a MRI T1 weighted.
> I am using fnirt but I read in the documentation that the only cost function avaible is the sum of squared differences.
> In my opinion the mutual information (in theory) should be the best choise in this case...Am I wrong?
> Is there a way to use a different cost function instead of the SSD?
>
> Thank you very much.
> Regards.
>
> Paolo
>
|