Dear FSL experts,
I saw a post on this from a few years ago, but just wanted to clarify. I would like to run randomise on DTI data to look for areas in which there is a progressive change in FA with my outcome variable, but not necessarily linear. In the old post, Anderson seemed to suggest that this is possible by using the appropriate contrasts, but I 'm not sure how this would be set up. I have 4 groups, which are the quartiles of my outcome variable, 50 subjects per group, with other demographic covariates as nuisance variables. Any suggestions for how I would set up the design matrix and the contrasts would be most appreciated. Thanks!
Mark Wagshul
|