The unbearable brightness of being a functional magnetic imaging brain scan

For more than several years, I've observed the following phenomenon. It needs a good name (and greater opprobrium).

• Before: through diligent experimentation and careful thought, scientists construct a theory of some complex human behavior. Example: a model of human memory (Atkinson-Shiffren memory model) highlighting the different time scales over which memories persist.
• After: we have the same set of behaviors and models, but also functional magnetic imaging scans of the brain. These scans show some correlation between an assigned task and a region of the brain. Therefore, it is concluded, the features of the theory must be true: we can see it in the brain.

This is mostly nonsense. What it does say is that some brain functions are highly localized. However, other functions are not localized and won't show up on scans. That some well-described function is not apparent in a scan does not imply anything about the veracity of the theory of the function. Somehow, the visual concreteness of the scan shortcircuits any critical thinking faculty in the reader. I'd wager that because scans in color are fancier, they are given greater evidentiary weight.

This has been studied. See http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2778755/