afni.nimh.nih.gov

Jessica-

To your original question, my quick test suggests that the implementation of 3dRegAna should give you almost identical results for Coef #3 and the F-stat, with the exception that the t-stat sub-brik will give you directionality, whereas the F-stat will display all differences as positive. I believe that 3dRegAna F-test evaluates if the Full Model is significantly different from the Restricted Model. In which case, the F-stat should be the squared value of the t-stat for a single variable that differs between the two models. As shown in the figure below, forgive the rounding.

I'm not entirely sure the best way to do your analysis in 3dttest++ or 3dMEMA, but I'm hoping that Gang can help translate your original code to 3dMEMA or 3dttest++. I tend to only use 3dttest++ and 3dMEMA for single covariates.





Edited 1 time(s). Last edit at 05/21/2013 04:02PM by Peter Molfese.