This looks fine to me.
Sometimes I would say that we get better results for step 1
by using the skull-stripped image for T1High, but normally
it doesn't matter much as long as you stick to 6 DOF.
Everything else looks great, so should work fine.
All the best,
On 1 Aug 2009, at 02:16, Nicole Strang wrote:
> I've primarily used afni in the past, but want to take advantage of
> great non-linear registration in fsl. Because I've already run the
> first-level stats is afni, I'd like to use fsl for registration
> alone, if
> possible. I couldn't find an answer to my the question online, so I
> really appreciate it if someone could let me know if there are any
> with my approach. If there is a problem, alternative would be very
> Convert all relevant files to NIFTI
> 1) Register run1 epi time series (there are 4 runs but all frames are
> aligned to first frame of run1) to skull on T1High (with flirt)
> 2) Register skull stripped T1High brain to MNI (with flirt)
> 3) Register skull on T1High to MNI (with fnirt)
> 4) Register output of glm which has been converted to percent signal
> with output of (3) as warp and output of (1) as premat.