Hi Anderson,
Thank you very much! It is very helpful!
I am just not sure to understand why there is no need to demean Gender in this case as my design is quite similar to the example 3a (C=(1 -1 0 0) in http://mumford.fmripower.org/mean_centering/? Do you mean that in this case, mean-centering or not will not change the results?
Also I have used the group column to define 4 groups because I have patients and controls and some reasons to believe that the variance could be different between groups. Is it correct to do this when using Glm to produce the design.con and design.mat files that will be used in the Dualregression?
Thanks again!
Elisa
|