Reward-based decision making under social interaction Damon Tomlin MURI Kick-Off meeting September 13, 2007
The decision task A B
The underlying structure % A Reward Reward A Reward B Average
A more interesting case % A Reward Reward A Reward B Average
Adding social interaction... Feedback –None –Choice history –Individual rewards
Adding social interaction Feedback Different games NEO data
Conditions in the experiment: “Alone”
Conditions in the experiment: “Rewards”
Conditions in the experiment: “Choices”
Logistics Group size Subject payment Behavioral cohort Imaging cohort
Game elements Crossing points Optimal reward Short term vs. long term gains
Games within the experiment "Simple" Rising Optimum % A Reward A Reward B Average Reward How frequently do subjects find the optimum? Once found, do they stay?
Games within the experiment "Simple" Rising Optimum % A Reward A Reward B Average Reward Are subjects naturally biased toward A or B?
Games within the experiment “Complex" Rising Optimum % A Reward A Reward B Average Reward Can subjects find a more subtle strategy? How do social partners affect adherence to it?
Individual behavior
Games within the experiment Converging Gaussians % A Reward A Reward B Average Reward How much exploration occurs in a simple task?
Individual behavior
Games within the experiment Diverging Gaussians % A Reward A Reward B Average Reward How does social information produce herd behavior?
Individual behavior
Summary Binary choice decision paradigm Social conditions: –Alone –Reward Information –Choice Information Games examining: –Exploratory behavior –Herd behavior –Strategy maintenance