Hi Doug & others,

 

I am trying to use reg-feat2anat as it is described on the FSL web page:

 

However, I always seem t get a very bad registration. I tried using different values for  --maxangle --cost mutualinfo --bins, but maintained --dof always at 6 for rigid body registration of the T1 and EPI, but didn't gt any good results.

Initially I thought it might be because reg-fea2anat does not use a skull-stripped T1, but  when I run Feat using the T1.mgz (converted to .nii.gz) as the main structural image for Registration (and the MNI brain as standard space), both with 'Full Search' option and dof = 6, the registration (with FEAT)  works well.

Is there anything else I can try to make it work with reg-feat2anat. Why wouldn't the registration work on my data with reg-feat2anat given that it works with Feat?

 

Any comments or suggestions would be very appreciated.

 

best,

Sandy