> The difficulty with your suggestion of carrying out only steps 1-3 (below), without any clustering
> as step 4, is that you may encounter a scenario in which there is 1 voxel of overlap for tests 1
> and 2. Is this 1 voxel meaningful?

Yes, even one voxel out of the conjunction analysis would still be valid as long as each individual cluster was corrected prior to the conjunction analysis. The approach is labeled as minimum statistic compared to the conjunction null (MS/CN) in the Nichols et al. paper, and the conclusion is the 2nd paragraph in the Discusion section (and the rigorous proof is in Appendix B).

Gang



Edited 2 time(s). Last edit at 03/23/2015 05:09PM by Gang.