Gambling pigeons: Primary rewards are not all that matter

In nature, actions rarely lead to reward 100% of the time; most decisions involve an element of risk or uncertainty. The human decision-making literature is full of examples in which people show biased or irrational risky decisions, suggesting that choices are not merely based on expected value. In this talk, I describe two lines of research in which pigeons exhibit similar biases and irrationalities, suggesting that these behaviours may reflect a shared, evolutionarily ancient learning mechanism. In the first example, pigeons show systematic biases when choosing between fixed and risky options that have the same expected value (yield the same average food reward): pigeons gamble more when choosing between options that both yield more food rewards (relative wins) than when choosing between options that both yield less food rewards (relative losses). In the second example, pigeons show a striking example of irrational behaviour in which they choose an option that leads to a lower probability of food over one that provides that same food with greater certainty. This suboptimal choice occurs only when the food rewards for both choices are delayed and there is a signal during the delay indicating that the food is coming (i.e., a signal for good news). This suggests that signals for good news can be powerful reinforcers of behavior and serve to bias choices. Together these findings indicate that seemingly biased and maladaptive decision making is not unique to humans and may have an adaptive purpose with deep evolutionary roots.