I may be trying to do the impossible, or at least the unlikely, by attempting TBSS with noisy 7T DTI data. I only have 25 1.7 mmm axial slices. After registration to FMRIB 58_FA_1mm, I can see that the FA images are grossly distorted by the nonlinear registration, which leads to problems down the line, of course. Here are some questions related to my problems with this dataset:
1. how important is a good brain extraction to the TBSS process? BET does not work well with my dataset
2. can I modify the TBSS registration to make it linear, or less-strenuous nonlinear?
Thanks,
-Jeff
|