Hi FSL experts,
I've run into some issues using the flirt bbr option when trying to register the aparc+aseg file from Freesurfer to the dMRI b0 file. I tried doing this without bbr, i.e.:
flirt -in T1 -ref dMRI -omat T1_to_dMRI.mat
flirt -in T1 -ref dMRI -o T1_to_dMRI -init T1_to_dMRI.mat -applyxfm
flirt -in aparc+aseg.nii.gz -ref dMRI -o aparc+aseg-new.nii.gz -init T1_to_dMRI.mat -applyxfm –interp nearestneighbour
this works fine.
The T1 image is a skull-stripped, intensity normalized image ('brain.mgz' from Freesurfer output converted to 'nii.gz' format).
However, when doing:
flirt -in T1 -ref dMRI -omat T1_to_dMRI.mat
flirt -in T1 -ref dMRI -dof 6 -cost bbr -wmseg T1_brain_pve_2.nii.gz -init T1_to_dMRI.mat -omat T1_to_dMRI-bbr.mat -schedule $FSLDIR/etc/flirtsch/bbr.sch
flirt -in aparc+aseg.nii.gz -ref dMRI -o aparc+aseg-new.nii.gz -init T1_to_dMRI-bbr.mat -applyxfm –interp nearestneighbour
the aparc+aseg registration turned out to be poorly aligned to the diffusion image. I'm not sure if this is because I am not specifying the correct WM segmentation for -wmseg? Or could it be that the commands I've used are inaccurate?
Would appreciate any advice!
M.
|