Thanks for your responses. I tried 3dttest++ with one participant group and behavior scores as a covariate. When examining the results of the 3dttest++ in afni, do results produced by entering behavior t-stat as the threshold represent behavior varying with connectivity or mean connectivity for all participants covarying behavior? I’m interested in the relationship between connectivity and behavior. I'm trying to determine if it is appropriate to use 3dttest++ with the scores that I want to correlate provided in covariates file.