Hello,
I am trying run randomise on DTI data to see group differences. However, when i choose to run randomise using -n 500 permutations i see more significant results (after thresholding corrp_tstat map at 0.95) when compared to using -n 5000 or -n 10000.
1) How do i understand this behavior?
2) what is the optimum number of permutations to be used for a particular size of study group?
BW
Jay
|