Computational neuroscientists discover a rich interplay of emotions, experiences, and decision making.
Any stock exchange investor is regularly confronted with the choice of either selling their shares of a given company or keeping them, accepting a potential loss as a tradeoff for a possible profit. But how do people decide when to take a risk and when to be cautious? Do previous outcomes of their actions have an effect on the decision? And what role do emotions play?
Neuroscientists Vikki Neville and her collaborators at the University of Bristol together with Peter Dayan from the Max Planck Institute for Biological Cybernetics and the University of Tübingen designed an experiment to investigate these fundamental questions.
“We wanted to understand in detail the relationship between the risk-taking behavior of humans, their ongoing experiences, and their emotions,” says Dayan. “Therefore, we had to come up with a new decision-making experiment and a novel computational model to analyze its results. This is a paradigmatic example of basic research in the new field of computational psychiatry.”
Risky choices under time pressure
At the stock exchange, agents have to try and predict the share price development from ambiguous information, taking into account a myriad of economic and political events. The pay-off or penalties associated with these guesses can be small or substantial. In the experiment designed by the team from Bristol and Tübingen, dots in a short animated movie moved to either the right or the left, and subjects could make either a safe choice, releasing a key or ‘going’; or a risky choice, holding onto the key or ‘staying’. ‘Going’ led to neither loss nor gain; ‘staying’ accrued loss or gain depending on the true direction of the dots, with amounts that changed trial to trial.
Sometimes the information was rather unequivocal (when many dots moved in the same direction), and hence the best action to take was clear. In other trials, however, the movement of the dots was harder to make out and therefore participants had lower quality information about what might happen if they gambled. Participants could try waiting longer to gather more information, but that was itself risky, since they might find themselves to have ‘stayed’, and thus gambled, by default. This raises the question: what determines the sorts of risks that people are willing to take?
Gains and losses – and happiness
Participants were immediately informed of their losses and wins after each trial. From time to time, they were also asked about their emotional states – if they experienced positive or negative feelings, and if they felt rather sleepy or rather alert.
Perhaps not surprisingly, participants felt happier when they were more successful in the task. “What we hadn’t expected, however, was that being more successful didn’t make people take more risks – in fact, the opposite was the case,“ says Dayan: “success made participants more likely to choose the safe, ‘leave’ option.” In line with this, but also unexpected, was that negative emotions and states of high arousal made participants seek out more risks; and conversely, when participants reported to be feeling happy and calm, they played it safe.
Could it be that traders in the stock market tend to go for riskier transactions when they are making losses? And what would that imply for the stability of the markets as a whole? It is certainly not possible to extrapolate from a well-controlled laboratory study to this much richer (or poorer) scenario, so these questions must remain unanswered for now.
Next on the agenda for the researchers from Bristol and Tübingen is to extend beyond the healthy population and to use these methods to investigate decision making in people with mental health disorders such as depression.