Dear Lorenzo,
> DFWE-corr is significant (0.013), while Duncorr is not (0.143)
Did you obtain these two p values with the same initial voxel threshold on whole brain level? If so, can you post a screenshot? I guess it's rather due to looking at one p value from a whole brain analysis and another obtained via small volume correction, in which case a corrected p value might indeed be smaller than an uncorrected one due to different initial voxel thresholds and/or search space (or possibly, p values resulting from different initial voxel thresholds).
Best
Helmut
|