I'm doing a cross-sectional analysis investigating differences in diffusion tensor scalars between controls and three different levels of disease groups along the skeletons of white matter fiber tracts. Instead of using tbss_skeleton the traditional way to create a whole brain white matter skeleton, I used it separately on the probability maps of 3 individual white matter tracts that were derived from a whole brain tractography separately. This method is described in this article: http://link.springer.com/chapter/10.1007/978-3-642-15705-9_77.
I want to use randomise to find where the scalar values are different along my white matter skeletons. Initially, I thought randomise should analyze the different tract skeletons separately. However, now I'm wondering if I need to combine all 3 tract skeletons into a single image for each subject and create my 4D volume for randomise to do a proper multiple comparisons correction across all ROIs.
I was hoping to get some feedback on this matter. Thank you for your time.