Dear all,
I am looking for some advice with some questionable results I have generated using randomise for TBSS. I have three groups and I wanted to look at differences in white matter between them with two covariates of no interest (age and gender). I ran an F-test using randomise with the following F-test contrasts (I also demeaned my covariates):
1 1 -1 0 0
1 -1 1 0 0
-1 1 1 0 0
I then used the following command to run the F-test:
randomise -i all_MD_skeletonised -o tbss_MD -m mean_FA_skeleton_mask -d design_CNNM2.mat -f design_CNNM2.fts -n 5000 -x --T2 -V
Three _tfce_corrp_fstat* files were then generated and I thresholded each image at .05. However, my results are quite unusual as for each F-test I have over 250 significant clusters with the largest clusters consisting of over 120,000 voxels. I expected quite subtle differences between my groups for this analysis. Have I made a mistake with this F-test?
I would really appreciate any advice. Thank you!
Sinead
|