Dear FSL Experts,
I am using easythresh_conj with a voxel threshold of 3.1 and a cluster threshold of 0.05. I am correcting over a binarized mask of all voxels above 3.1 for each of the individual contrasts. Three questions.
1. Many of the conjunction papers I am reading don't appear to apply a cluster correction, they just have a voxel level threshold but no cluster correction. Is it necessary to do a cluster correction?
2. For the cluster correction, is it reasonable to use the masked results of the individual contrasts to mask the cluster correction? (i.e. rather than the whole standard space brain)
3. When I do an analysis with a voxel threshold of 3.1 and cluster correction of 0.05 or even 0.01 I am getting "clusters" of 1 voxel that are passing the correction. That doesn't make any sense to me. Does that indicate some kind of a problem in the analysis?
|