Hi,
This is normal. FEAT looks for the pre-BET images in the same directory
and uses them with FNIRT, as it is better for the non-linear
registration.
The linear registration with FLIRT is still better with the post-BET
images.
So you have done everything correctly and this is just how FEAT works.
It is detailed in the bubble help and the documentation for FEAT.
All the best,
Mark
On 1 Jun 2009, at 16:22, Emma Liang wrote:
> Dear FSLers,
>
> I use all the same imagings and setting to do linear and nonlinear
> registration respectively. But the nonlinear
> highres2standard(attached) shows the skull that had been removed by
> BET. Could anyone tell me why? Thanks a lot!
>
> Best regards,
> Emma
|