Dear Fsl experts
I am running a second level regression analysis in Feat with pain scores. I obtained significant clusters at Z> 3.1 and p < 0.05. However when I extract the individual COPES to make a scatter plot the correlation is very weak ~ zero!?
When I try the same analysis using randomize instead of Feat I do not obtain any significant results. I believe the randomize results ; however why is feat showing such a significant result when the actual correlation is very weak ?! I wonder if this is a bug ?
Thanks
Paul
|