Hi,
1) If you are using skull-stripped images then make sure you
use the *_brain images. You can choose a resolution
which suits you best (2mm is fairly standard, but 1mm for using
FNIRT - non-linear registration - with a good structural is fine).
And use the non-linear templates, MNI152_T1_* as they are
much better quality than the linear templates (MNI152lin_*).
The avg152 image is an identical copy of the MNI152_T1* which
was put there for reasons of backwards compatibility.
2) Yes, use the same ref and input image in std2imgcoord as in FLIRT
when registering the input image to the standard image. Also
use the same transform generated by FLIRT (not the inverse).
The line you've included looks fine.
3) xfm stands for transform (it's borrowed from MEDx, years ago)
and should be a flirt matrix. The units of the output coordinates
can either be mm (which is the default) or voxels (using the -vox
flag) and both should correspond to the coordinates that FSLView
reports.
All the best,
Mark
On 14 May 2009, at 05:47, SUBSCRIBE FSL ila wrote:
> Hi,
>
> I'm new to FSL.
> Could you please help me on these 3 questions?
>
> 1) I'd like to register a participant skull-stripped brain to an MNI
> brain,
> which reference image do I use in FLIRT (and why) e.g.
> 152MNI_T1_2mm_brain.nii.gz ; avg152MNIT1_brain.nii.gz?
>
>
> 2)when I use std2imgcoord in order to convert MNI coordinates in the
> coordinates of the registered braindo I have to put the same
> reference image
> I used in FLIRT?
>
> e.g. if in FLIRT I used as reference image avg152MNIT1_brain.nii.gz
>
> do I have to do:
>
> -std2imgcoord -img skullstrippedimage_brain.nii.gz -std
> /usr/local/fsl/data/standard/avg152T1_brain.nii.gz -xfm
> native2MNI_FLIRTtransformationfile.mat ?
>
> 3)what does -xfm stands for and in which units are the coordinates
> that come
> out from std2imgcoord ?
>
> I thank you in advance for your help!
>
> ila
>
|