Hi FSL experts,
I had three groups of subjects who underwent the same scanning protocol.
However, I processed (tbss_1_preproc *.nii.gz, tbss_2_reg -T) these groups
in TBSS toolbox separately.
Recently, I need to include some of the controls from group A into group B,
so after I copied the related files in the FA folder from group A to group
B's FA folder, I rerun [tbss_3_postreg -S] and [tbss_4_prestats 0.2] scripts
so that I can run randomise for statistical analysis. However, I found that
for those subjects that I included later from group A, their MD and
RD[(L2+L3)/2] value is lower throughout the whole skeleton compared to
subjects in group B, which caused randomise give false significance.
Therefore, I am wondering:
1. Is it ok to combine subjects from different groups together after I ran
[tbss_1_preproc *.nii.gz] & [tbss_2_reg -T] on each group separately, based
on the fact that I am using the same version of TBSS?
2. I remember there was a scaling factor in the previous version of TBSS,
where did it go now? Does TBSS automatically scale the FA maps? If yes,
which step did it happen?
Thank you in advance.
Yi-Shin
|