Dear FreeSurfer experts,
I conducted a study with 3 groups of children (2 trainings, 1 control) and I want to analyze them longitudinally with LMEs. While preparing my data 2 topics came up that I would like to know more about.
A. controlling for ICV/mean thickness/surface area
1) I want to look at changes in cortical thickness and surface area at the whole brain level. Would I then have to control for mean thickness/ area, respectively, per parcellation? Or rather ICV? 2) Is it the same for looking at changes in thickness and area in different ROIs? 3) Since I have (up to) 3 scans per kid, do I include mean thickness/area/ICV for each scan or should I use it from baseline or a mean of all scans?
B. demeaning covariates
1) Is it necessary to demean the covariates? I have not seen any hints on the FS wiki pages about that yet. 2) Would I also have to demean the dependent variable (in my case whole brain cortical thickness/surface area)?
Thank you! Clara