In running a contrast on PET data, I set the threshold to be p<0.001 for 100
contiguous voxels, and I find in looking at the volume or cluster Pvalue
results that there are no clusters above that threshold. No until I lower
the thresold to p<0.005 (100 voxels again) do I see results. However, the P
uncorrected values listed in both the voxel-level as well as cluster-level
are less than 0.001. Why did these voxel-level statistics not appear when
p<0.001 was used?
Thanks in advance,
Suzanne Baker
|