Hello,
I run the full TBSS analysis and a goup comparison on my FA data. Now I want to repeat the T-Tests on the MD, RD and AD data.
I did this according to the UserGuide (http://fsl.fmrib.ox.ac.uk/fsl/fslwiki/TBSS/UserGuide):
1. Run the full TBSS analysis (see all steps above) on your FA data.
2. Create a new directory called L2 (or any other name) in your TBSS analysis directory (the one that contains the existing origdata, FA and stats directories from the FA analysis). Type: mkdir L2
Copy your L2 images into this new directory, making sure that they are named exactly the same as the original FA images were (look in origdata to check the original names - and keep them exactly the same, even if they include FA, which can be confusing; e.g. if there is an image origdata/subj005_FA.nii.gz then you need an image L2/subj005_FA.nii.gz and this file should contain the L2 data, even though it has FA in the name).
3. Now, making sure that you are in your top working TBSS directory (the one that now contains FA, stats and L2 subdirectories) and run the tbss_non_FA script, telling it that the alternate data is called L2. This will apply the original nonlinear registration to the L2 data, merge all subjects' warped L2 data into a 4D file stats/all_L2, project this onto the original mean FA skeleton (using the original FA data to find the projection vectors), resulting in the 4D projected data stats/all_L2_skeletonised. Run: tbss_non_FA L2
That's where the problem begins: for MD and RD I get 'all_MD/RD' files and the all_MD/RD_skeletonised but for AD I only get the 'all_AD' file. Do you know what the problem is here?
Then also, when I try to run a T-Test for MD I typed in this command but there popped up an error message:
"randomise -i all_MD_skeletonised -o tbss -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2''
I don't know much about tbss yet and the questions might be stupid but I hope s.o. will reply anyway.
Thanks a lot,
Vera
|