Hi Charlie,
Did you check the FA to T1 registration?
Cheers,
Manu
> On 15 Dec 2016, at 11:17, Charlie Lou <[log in to unmask]> wrote:
>
> Dear All,
>
> I'm new to FNIRT. Part of my work is to registrate AAL template to FA native space to construct white matter networks. And my FSL commands are as follows:
>
> /usr/share/fsl/4.1/bin/flirt -ref /home/charile/data/FNIRT_test/MNI152_T1_2mm_brain.nii.gz -in /home/charile/data/FNIRT_test/ab100358_T1_brain.nii -omat /home/charile/data/FNIRT_test/transfer_affine.mat -bins 256 -cost corratio -searchrx -180 180 -searchry -180 180 -searchrz -180 180 -dof 12 -interp nearestneighbour
>
> /usr/share/fsl/4.1/bin/fnirt --config=/usr/share/fsl/4.1/etc/flirtsch/T1_2_MNI152_2mm.cnf --ref=/home/charile/data/FNIRT_test/MNI152_T1_2mm_brain.nii.gz --in=/home/charile/data/FNIRT_test/ab100358_T1_brain.nii --aff=/home/charile/data/FNIRT_test/transfer_affine.mat --cout=/home/charile/data/FNIRT_test/nonlinear_transfer
>
> /usr/share/fsl/4.1/bin/convert_xfm -omat /home/charile/data/FNIRT_test/T1_to_FA.mat -inverse /home/charile/data/FNIRT_test/ab100358_FA_to_T1.mat
>
> /usr/share/fsl/4.1/bin/invwarp --ref=/home/charile/data/FNIRT_test/ab100358_T1_brain.nii.gz --warp=/home/charile/data/FNIRT_test/nonlinear_transfer.nii.gz --out=/home/charile/data/FNIRT_test/MNI_to_T1
>
> /usr/share/fsl/4.1/bin/applywarp --ref=/home/charile/data/FNIRT_test/ab100358_Concatenated_DTIs_MD_C_E_native_FA.nii --in=/home/charile/data/FNIRT_test/AAL_Template/AAL_Labels.nii --warp=/home/charile/data/FNIRT_test/MNI_to_T1 --postmat=/home/charile/data/FNIRT_test/T1_to_FA.mat --out=/home/charile/data/FNIRT_test/AAL_to_FA --interp=nn
>
> #The ab100358_T1_brain.nii file is the T1 image after the BET calculation with the original T1 image.
>
> Unfortunately, the result was like garbage. So I checked the result image step by step. The fnirt result, registered T1 image to MNI152 template, were still good(I used --iout instead of --cout for checking), but the final image was totally destoryed after the applywarp step. I guess it might due to the invwarp function but I can't take it for sure.
>
> Hope you experts could help me with this problem. Any suggestion would be appreciated.
>
> Best wishes,
> Charlie
> <screenshot_for_two_images.png>
|