Hi Anderson,
Thanks so much for the helpful response. You're exactly right about the set-up of our analysis (we are looking at location of voxels within our input masks related to clinical change measured by a quantitative EV), except that we are performing the correlation within a sample of only patients and no controls. In addition to including an intercept or using -D, should we include an EV coding them all as +1? Thanks again, we really appreciate the assistance.
Best,
Pranav
|