dead salmon categorizes human emotions
This nice study shows the need to beware of poorly controlled fMRI studies. (In particular: predefined significance thresholds = bad, multiple comparisons correction = good.)
The image next to the header is "Ghost" by Neil Grayson.

I'm *really* curious what y'all think about this -- is it just obvious and spurious?
Reply to this
The point of the paper is that many people have been using bogus analysis on fMRI data, and this is a reductio of said techniques.
Reply to this