Skip to main content

Science: Poker Study Sheds Light on Social Decision-Making

Humans show different patterns of brain activity when they’re deciding to bluff against either a human or computer poker opponent, a new study in the 6 July issue of Science reveals. The unique experiment demonstrates how the action of social decision-making differs from non-social decision-making in the brain.

In recent years, cognitive neuroscientists have discovered that people rarely make decisions in isolation. Humans are sensitive to what others think. A quick phone call with mom or a long chat with a friend can make people think twice before making a decision, or can give them the conviction to forge ahead with a plan. Yet researchers are still trying to understand how social interactions affect the actual brain.

Scott Huettel, a professor of psychology and neuroscience at Duke University, and his colleagues designed an experiment where participants—individuals with no poker expertise—played simple virtual poker games separately against human and computer opponents. The participants met and shook hands with the human opponents right before playing the games.

Some of the time, the participant players were dealt a losing hand and had to decide whether to bluff their opponent. The researchers scanned 55 regions in each of the participants’ brains and then used the resulting functional magnetic resonance imaging (fMRI) data to predict those bluffs right before the act.

The researchers discovered that neural activity in a small region of the brain called the temporal-parietal junction can be used to predict how people will play the poker game—but only against a real human opponent. The region was much more likely to be active when the participant was deciding whether to bluff against a human opponent, compared to bluffs played against the computer.

The results offer concrete evidence that the brain works differently in social and non-social situations, the researchers concluded. The temporal-parietal junction may play a unique role, they suggest, in representing information that will predict how a person behaves during a social interaction.

Social decision-making requires elements different from non-social decision-making, including the ability to understand another person’s thought processes and to predict how the decision will guide future social interactions. These requirements, Huettel and colleagues say, could explain why brain activity differs in social and non-social situations.

Said Huettel: “Understanding how the brain identifies important competitors and collaborators—those people who are most relevant for our future behavior—will lead to new insights into social phenomena like dehumanization and empathy.”


Read the abstract, “A Distinct Role of the Temporal-Parietal Junction in Predicting Socially Guided Decisions,” by Scott Huettel and colleagues.



This script tag will cause the Brightcove Players defined above it to be created as soon
as the line is read by the browser. If you wish to have the player instantiated only after
the rest of the HTML is processed and the page load is complete, remove the line.

Listen to a podcast interview with Science author Scott Huettel.