 |
 |
Dear SPM’ers, I have a question regarding the use of covariates in the statistical analysis with SPM. In an fMRI analysis, I’m trying a factorial design with 2 groups (patients and controls) to compare the disease effect. To reduce the effect of patient’s variability the idea is to add medication as nuisance variable. This covariate only has meaningful values for one group (patients) and zeros for the control group. I was trying the centering method according to the SPM manual (“Factor 1 mean” only for the patient’s group and “overall mean” for the controls group). It is correct? Additionally, due to the higher values of the covariates, I also tested normalizing the covariates between [0 and 1 + centering] and [-1 to 1 with no centering], obtaining in all cases different results for comparing patients vs controls. Whats the correct way to do that? Thank you very much. Regards, Francis