 |
 |
Matthew Brett wrote, in a thread that looks to be winding down: > I think this gets right at the key point. Let's consider two regions > A and B, and an activation map for task X. > > Thresholded map: A survives thresholding, B does not. Evidence that > A is more activated than B = very near zero. > > Unthresholded map: A is in area outlined as significant, B has zero > or negative signal on unthresholded map. Evidence that A is more > activated than B = moderate. > > Neither of course constitutes a statistical test of region A vs region > B. It is just that the unthresholded map offers considerably more > evidence than the thresholded map. > > Do we agree on that? For the most part, sure. How I feel about the words "moderate" and "considerably" depends on how the threshold was chosen, but that's fine. To me, the issue has little to do with more vs. less evidence. We're talking about studies in which the authors and reviewers agree that direct comparison of regions A and B is not relevant (otherwise either there would already be a direct comparison or the article would be rejected out of hand), so I don't think supporting this comparison better is a compelling argument. The space of irrelevant hypotheses is large, and we can't ask authors to support all of them in their reports. The alpha blended outline map is nice, because there's roughly no cost to it. But absent that option, I don't see why authors should be universally compelled to include a figure about which they have nothing to say. dan