Hi,
I'm wanting to set up my design matrix before running randomise to see if the FA values correlate with patients' cognitive domain scores (5 domains). I only have a patient group (16 patients) without a control group. Because I'm not very familiar with setting up the glm for correlation analyses, I do not know exactly how to set up my EVs and contrasts.
1. For EV1, would it be all 1's?
2. EV2-EV6 are the different cognitive domains' scores? Do I have to demean them? If so, how do I demean?
3. How many contrasts would I need and what should they look like?
4. When running randomise: randomise -i all_FA_skeletonised -o tbss -m mean_FA_skeleton_mask -d design.mat -t design.con -n 500 --T2, do I need to use the "-d" option? Also how many permutations should I do?
Thanks very much,
Amy
|