In an article published in J Neurosci, the author established a parametric modulation GLM, with three parametric modulators (A, B, and their interaction AxB) were entered. Then the authors said:
“In a follow-up analysis, we controlled for a trial-by-trial X by adding it as a covariate in the model described above. Furthermore, to ensure that the effect in regions found to express the interaction between A and B were not confounded by another factor Y, we included choice as a covariate in our parametric modulation analysis by adding Y as a regressor orthogonalized to the interaction regressor in that order. Thus, in this stringent analysis, we modeled all trials with a parametric modulator of A and B, and modeled trials in the critical condition with the interaction regressor while controlling for X and Y.”
Here are my questions:
(1). My understanding is that “parametric modulator” is entered the first-level GLM, whereas “covariate” is entered in group-level model. Am I correct? So I’m confused by the terminology here.
(2). If X and Y are entered as parametric modulators and to be regressed out (because the authors are interested in the interaction effect AxB), the order of the 5 parametric modulators should be X, Y, A, B, and finally AxB, right?
Thanks.
Mike
|