Related Experiment Videos
Learning and decision making in monkeys during a rock-paper-scissors game
Daeyeol Lee1, Benjamin P McGreevy, Dominic J Barraclough
1Department of Brain and Cognitive Sciences, Center for Visual Science, University of Rochester, Rochester, NY 14627, USA. dlee@cvs.rochester.edu
Brain Research. Cognitive Brain Research
|August 13, 2005
Summary
Monkeys playing a game adjust strategies based on experience, not just rewards. Their learning combines reinforcement and belief updates, reflecting both actual and hypothetical payoffs in social decisions.
Area of Science:
- Neuroscience
- Behavioral Economics
- Game Theory
Background:
- Human decision-making deviates from game theory's optimal strategies, incorporating experience-based adjustments.
- Reinforcement learning (RL) and belief learning (BL) models describe strategy updates, differing in how value functions are modified based on outcomes or others' choices.
Purpose of the Study:
- To investigate the learning processes in primates during competitive social interactions.
- To determine how monkeys update their decision-making strategies in a game context.
Main Methods:
- Monkeys played a competitive ternary choice game (rock-paper-scissors) against a computer.
- Computer strategies varied, exploiting random targets, choice sequences, or both choice and reward histories.
Main Results:
- Monkeys exhibited target biases during random computer play.
- Computer exploitation of choice sequences led to systematic monkey biases, which decreased when both choice and reward histories were exploited.
- A combined RL-BL model better explained the monkeys' adaptive learning than individual RL or BL models.
Conclusions:
- Primate decision-making in social contexts is adaptive, integrating both direct outcomes and inferred payoffs from others' actions.
- Stochastic strategies in primates are dynamically adjusted based on a blend of reinforcement and belief-based learning mechanisms.