Hi,
Glad things are working smoothly.
I'm a little surprised at your problems with FNIRT if you find that the
FIRST registration looks better. Note that we don't have any plans
at this stage to produce a registration based on the FIRST segmentation,
as we find that FNIRT does just as well in general.
If you are finding consistent problems with FNIRT then please do let
us know.
All the best,
Mark
On 22 Aug 2010, at 23:01, Carlos Faraco wrote:
> Mark,
>
> I just rechecked my FEAT FNIRT setup to make sure I had it setup up
> properly when I had originally tried it, and yes, I did. I chose the
> _brain image and made sure the original non-BETed image was in the
> directory. For this particular subject, FNIRT seems to extend the
> frontal lobes a bit and cause some "shearing" (not sure if that is
> the correct word to use) on the surface of the cortex. Maybe I just
> need to run it on a few other subjects to make sure it just wasn't
> an anomaly with the subject I tried it on.
>
> In regards to FIRST, thanks for clearing that up, that is very
> helpful. It would be real nice though if the registration was later
> optimized based on the segmentation. =]
>
> In regards to timing, yes, it seems like it was going to take that
> long. Since the FIRST results are registered to the 1mm brain, for
> some reason I thought I should also implement this into the
> registration results for the use of FEAT. This was an obvious
> mistake on my part as there is no need to use the 1mm brain, just
> because the transformation was made into 1mm space.
>
> I redid the registrations with the 2mm brain and the transforms from
> FIRST, and everything seems to be going along smoothly.
>
> Thanks for your continued help!
>
> -Carlos
>
|