Yes, just using flirt would work fine. Just remember to save the
transformation matrix.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of Live Eikenes
Sent: Thursday, November 15, 2007 8:53 AM
To: [log in to unmask]
Subject: Re: [FSL] nodif_brain in registration
Thank you very much for your answer, but I need to know one more thing
about the registration. If I am to use the dti_FA for registering the
structural image to the diffusion image, and the registration in Fdt, I
have to call it nodif_brain (in the bedpostX directory), otherwise it
wont work. Should I instead of the registration in Fdt use FLIRT?
Thank you very much!
Live.
Sitat Matt Glasser <[log in to unmask]>:
> As long as your bvecs and bvals take into account all of your b0
> volumes,
> they will be included in the DTIFIT and BEDPOSTX calculations. You
> should
> use the dti_FA for registering the image to the structural to
> generate a
> transformation matrix, but not as the nodif_brain image.
>
> Regarding the different acquisition plane and non-isotropic voxel
> size, so
> long as you don't have a gap between the slices it should be fine.
> -----Original Message-----
> From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On
> Behalf
> Of Live Eikenes
> Sent: Thursday, November 15, 2007 8:08 AM
> To: [log in to unmask]
> Subject: [FSL] nodif_brain in registration
>
> Dear FSL users,
>
> I have a simple question about the nodif_brain file needed for the
> registration between the structural image and the diffusion image
> after
> BedpostX and before probtrackX in Fdt:
>
> Should the nodif_brain file only be one volume, and should it be just
> one average, or all the b0 averages that I have?
>
> I also read that the registration between the structural image and
> the
> diffusion image becomes better when using the dti_FA image instead of
> the b0 image (nodif_brain image). Should I therefore use the dti_FA
> as
> my nodif_brain image in the BedpostX directory when applying the
> registration?
>
> Another question about the structural image is: does it matter if the
> structural image is a sagital image (not isotropic, 1mm x 1mm x
> 1.3mm)), and that the diffusion image is an axial image?
>
> Thank you very much!
>
> Best regards, Live Eikenes.
> --
>
--
|