Dear FSL friends,
We are trying to do the following. We have FA images of many different
subjects and we want to generate probabilistic tractography from seed 1
to seed 2. However we don't want to do it manually but rather
co-register all the subjects together using tbss, define the 2 masks for
the seeds and then go back to native space and do the tractography
independently for each subject. In that way we hope that we will
reduce the time of specifying the seeds separately for each subject.
Unfortunately, we are getting some unexpected results when we deproject
the masks to their native space.
Here is what we are doing:
1. Create FA images for each subject (using FDT).
2. TBSS steps:
i) tbss_1_preproc *.nii.gz
ii) tbss_2_reg -T
iii) tbss_3_postreg -S
iv) tbss_4_prestats 0.2
v) defined a mask on mean FA (and mean_FA_skeleton as well)
vi) tbss_deproject mask 2
Although we don't get any error from FSL we get results were the mask
have been moved very far away in the empty space of the image.
All the subjects that we are using are healthy subjects. However, we are
getting similar problematic results to most of our subjects. The version
of FSL we are using is 4.1.Are we doing any mistake? Any suggestions? Is
there an unresolved problem with tbss? Or is there a problem with our
methodology?
Thank you for your time.
Best regards,
Eleftherios Garyfallidis
|