afni.nimh.nih.gov

I have used 3dttest++ for a group analysis and would like to get p-values for each cluster. For the first level analysis, the smoothness is first estimated from the residuals, and 3dClustSim is then used with this estimate. What should I use to estimate the smoothness for the group analysis (3dFWHMx states that it should not be used on the statistical results) ?