If you use the anatomical as the reference image and the BOLD as the input
image perhaps you will have some more success? You can always invert the
transformation matrix with convert_xfm -inverse later.
Peace,
Matt.
-----Original Message-----
From: FSL - FMRIB's Software Library [mailto:[log in to unmask]] On Behalf
Of James Sheehan
Sent: Tuesday, October 13, 2009 11:09 AM
To: [log in to unmask]
Subject: [FSL] Difficulty with Flirt -> High Res 2 bold
Hi all,
I know this should not be so difficult, but I don't understand where I'm
going wrong here.
I am trying to update a workflow that bounced between AFNI and FSL and used
ANALYZE.
There was a multipart registration that they preformed to register a high
res anatomical and bold data while keeping the anatomical in 1x1x1.
Here is what I think should work but does not work.
flirt -in mprage.nii -ref filt_func_data.nii -out mprage_dummy -omat
txfm1.mat -dof 6
flirt -in mprage.nii -ref mprage.nii -out mprage_converted -init txfm1.mat
-applyxfm
this produces images that are not even remotely aligned... Anyway...
Below is the old work flow.
This is the old work flow that I'm trying to update.
#declare -x FSLOUTPUTTYPE=ANALYZE
#fslroi filt_func_data epi7th 7 1
#3dcopy epi7th.hdr epi7th (AFNI COMMAND converts file format)
#3dresample -dxyz 1.0 1.0 1.0 -prefix resampled_epi7th -inset epi7th+orig
-master Aanat_bet+orig (AFNI COMMAND Resamples epi into 1x1x1)
#3dAFNItoANALYZE -orient RAI -4D resampled_epi7th resampled_epi7th+orig.BRIK
(AFNI COMMAND file conversion to ANALYZE)
#flirt -in resampled_epi7th -ref Aanat_bet -out epi7th_trans_DUMMY -omat
txfm1.mat -dof 6 -cost corratio -searchcost corratio
#convert_xfm -omat txfm_inverse.mat -inverse txfm1.mat
--tranfsorm the anatomical space to that of the 7th epi, use without ref
(otw resolution problem)
#flirt -in Aanat_bet -ref Aanat_bet -out Aanat_bet_f1 -init txfm_inverse.mat
-applyxfm
Any help to get this corrected would be a huge relief right now... Thank
You.
--J Sheehan
|