Then why do casinos make money?

This Nature article fits perfectly with my current neurology kick. Everybody's heard of Pavlovian experiments, where an animal is trained to produce a physiological response to a normally unrelated stimulus by repetition. In Pavlov's case, he simply rang a bell and fed steak, and over time was able to induce salivation by ringing the bell only.

That's all well and good, but this paper asks what happens when ringing the bell is only assigned a probability of bringing steak. They showed four of ten shapes to monkeys, and then allowed to look at either a green or red target. Each shape carried a probability with it of a reward being given for a red or green choice. They were then able to train the monkeys to select red or green based on the most likely outcome, given the shapes presented.

The paper focuses heavily on the neurons in the parietal lobe involved in probabilistic reasoning. That's where the statistics get a little too dense for me, but I find the concept interesting; they were able to demonstrate the region of the brain associated with logic based on probabilities, rather than proven results.