Hi,

It sounds like your registration failed.
I assume that after you swapping commands
that both images are correctly labeled in
FSLView - right?  Oh, and by the way, if
they start with correct labels, but are just in
an orientation that is different from the standard
then you can use the new tool, fslreorient2std
which is easier and safer (it will never swap
left and right).

So if your images are labeled correctly and
in the same orientation as nodif_brain then
you need to solve the registration problem.
I would check the following
  - have you run BET on T1out?
  - are you using T1out (after BET) in the call
	to FLIRT and not the original T1 (which
	is what you've written in the email, but
	may just be a typo)?
  - use 6 dof for an intra-subject registration
  - if you've done all of the above and it still
	isn't working then try using -nosearch
	if the images seem fairly well aligned
	(at least in rotation) to begin with

If you have done all of the above and still
have trouble then let us know.

All the best,
	Mark


On 21 Jul 2010, at 06:46, Kahori Kita wrote:

> Hi,
>
> I'm a beginner of fsl, so please give me an advise.
>
> I try to convert T1 to diffusion space, but there is a problem.
>
> First, I tried to reorient T1 image by following command,
> because T1 is sagittal slice, and DTI is axial slice.
>
> fslorient -deleteorient T1
> fslswapdim T1 -z -x y T1out
> fslorient -setsformcode 1 T1out
>
> I checked the sformcode and qformcode of T1out, and they are 1.
> it is same as "nodif_brain", I got it from DTI using BET.
>
> I also checked the images by fslview, and both look right.
> I mean upper left of fslview space is coronal, upper right is  
> sagittal,
> and lower left is axial.
>
> And then, I try to convert T1 to diffusion space.
> flirt -in T1 -ref nodif_brain -omat t12diff.mat
> flirt -in T1 -ref nodif_brain -applyxfm -init t12diff.mat -out  
> t1_brain_diffspace
>
> T1 is a file name of T1 image, and nodif_brain is acquired from BET.
>
> I checked the resultant t1_brain_diffspace by fslview, and it is  
> quite wrong.
> Upper left is axial and rotate (i.e. nose comes to "R" position),
> upper right is coronal and rotate (i.e. top of the head comes to "P"  
> position),
> and lower left is sagittal and also rotate (i.e. top of the head  
> comes to "P" position).
>
> Any ideas?
>
> Thank you for your help.
>
> Regards
> kahori
>