50% of fMRI papers in which journals show incorrect statistical analyses in the last 3 years? Which of those do people take seriously? Which journals operate below 10%? What about compared to 5 years ago? I'd say the general trend has been in the right direction for years now, across many fields. Where you can make a big contribution is where the rubber meets the road, which is when you and your co-authors are reviewing articles.

Why not submit this as a Letter to the Editor, or something like the "Comments and Controversies" section of Neuroimage? I'd say that is the appropriate venue to accomplish your stated goal of educating the uniformed. You wouldn't get a research article out of it though. :) Maybe that's what you're doing, in which case my comments don't apply - but I got the impression you're trying to get this published as original research.

For venicebeach - compared with how Vul et al. handled a similar topic, this is a party with clowns and flowers. I didn't think it was condemning of fMRI at all.

These conclusions were indeed made many years ago. As someone else has mentioned, most fMRI researchers are perfectly well aware of the issues at a deeper level than your poster goes, and do a careful job. The work described in the poster has no methodological deficiencies, but also no scientific novelty or significance; its merit is its humor. I'm not surprised you're having trouble publishing it. If I reviewed this for a journal, I would read it, enjoy, and recommend a polite but firm rejection. Does Neuroimage have a cartoon page? Incidentally, you didn't correct for autocorrelation in your study, which would also inflate the significance of your voxel-wise tests - but you didn't cover that in the poster. :)

We all got a good chuckle out it.... it will go down in history with classics like 0 Tesla MRI at ISMRM a few years back. The spatial extent-based correction isn't exactly "unknown" - I gather you mean using arbitrary cluster size thresholds? I agree that arbitrary thresholds with no attempt to control FWE should not be used. And any map once thresholded is a rather limited representation of the data. At least always report a correct p-value for each voxel or cluster... I'm not sure everyone does that either. Sadly, good analysis and reporting is challenging to do well with some common software packages. If you really want to get into it, why are we reporting voxel-wise hypothesis tests at all? I would rather see confidence intervals or posterior probability maps, perhaps. You are pointing to a more general problem that the data are just not easy to understand or report and well-known mistakes are still made. For all the neuroimaging skeptics around, any biostatistician will tell you it's not limited to fMRI. :)

I was there, I saw the poster, it was a humorous joke meant to remind fMRI newbies to control their type I error. It was in no way publishable research and was not intended to be. Most people who do fMRI research already make the effort to do the stats correctly. Multiple comparisons correction for fMRI is old news - the authors' most recent fMRI stats citation was from 15 years ago. And no there wasn't any activity or signal change or anything else in the brain of a dead salmon. It has nothing to do with the "double-dipping voodoo" of Vul and company's recent temper tantrum, which is a completely different statistical error.