Dear colleagues,
We have found a problem while running regression analysis on a result mask previously created with SPM12. This is how we have proceed:
First, we performed a regression analysis using a cortical mask obtained with the WFU Pickatlas toolbox, and saved the result mask (p<0.001 uncorrected). Finally, we performed the same regression analysis within this new SPM mask. Results are contradictory: while we found no significant results before applying FWE correction, clusters were even larger after FWE. We did similar analyses with other datasets and always got a decreased T threshold after FWE. This problem is limited to regression analysis.
Anyone knows why this is happening?
Many thanks in advance.
Best,
Jose
|