I saw this article on Scientific American’s website about how our expectations relate to the effects of placebos. They did both a study with a placebo painkiller and a study using a gambling game. Subjects who had higher dopamine activity (and thus higher expectations of reward) during the gambling game were the same subjects who received more of an effect from the placebo. The conclusion is that your expectations of rewards are closely linked to the effectiveness of the placebo.

This is pretty interesting for a number of reasons. One is that it shows that an important aspect of having placebos work is simply being convinced that they will work. Looking at it from a reinforcement learning perspective, the dopamine activity can be viewed as the prediction of the reward signal. And then somehow expecting a higher reward signal can reduce the pain? Is the pain signal just a negatively valued feedback? What is the relationship between the expectation of reward and the actual reward received? Can simply expecting more or less reward affect the actual amount of reward received?