Did a brain scan of a dead salmon really find thoughts?

A dead salmon once seemed to show brain activity, and that is the point. In a real but satirical fMRI study, a dead salmon was shown pictures of humans in different emotional states, and two commonly used statistical tests suggested meaningful activity in areas of its brain. The study was used to highlight the need for more careful statistical analyses.

Here is why. fMRI measures brain activity by detecting changes associated with blood flow. It relies on the coupling of cerebral blood flow and neuronal activation: when an area is in use, blood flow to it increases. A voxel, the scan's tiny 3D pixel, typically contains a few million neurons and tens of billions of synapses, and a typical scan has a large number of voxels. That is where the multiple comparisons problem bites. If 100 tests are each run at the 5% level and all null hypotheses are true, the expected number of false positives is 5. With that many tests, some false alarms are expected unless the analysis corrects for them.

Practice has changed since. Before the controversies were publicized in 2010, between 25 and 40% of published fMRI studies did not use corrected comparisons; by 2012 that had dropped to 10%. But fMRI has other limits too. It can localize activity to within millimeters but, using standard techniques, no better than within a window of a few seconds. And a colourful blob does not tell you what a person was thinking: most researchers agree that fMRI's ability to detect deception in a real life setting has not been established.