I have a VBM design in which I have 4 groups (factors), one continuous variable of interest (score on a psychology test) and a three nuisance covariates (total intracranial volume, disease severity and age). I initially wanted to see if the relationship between psychology score and brain volume is significantly different between disease groups (i.e. is there an interaction)? I therefore set up a full factorial design with group as a factor and my psychology score as a covariate with an interaction with this factor 1. To investigated differences between groups I set up t contrasts (e.g. 1 -1 0 0) to compare group 1 vs group 2. My question is, can I perform a t-contrast of 1, 1, 1, 1 to look at brain volume that correlates positively with psychology score across all groups - i.e. is this effectively the same as a multiple regression model in which group would be a covariate. I've attached a picture of my design with the 1, 1, 1, 1 contrast I'm referring to.

Thanks in advance!