Reward-Based Decision Making and Decision Making under Risk

Many of our decisions are influenced by the potential outcomes associated with different choice options. For instance, consumers consider positive and negative product attributes prior to purchase, or people use past experience to decide which means of transportation is the best to commute to work. Our research on this type of decision making examines how people use reward- and risk-related information to achieve desired outcomes. To examine these kinds of decisions, we abstract basic features from real-life decisions, such as the type of information and feedback available, and implement them in simpler tasks. These tasks are amenable to manipulation in a functional magnetic resonance imaging (fMRI) environment and to precise modeling.

Conducting fMRI experiments allows us to test models and theories by examining decision variables in brain activity that cannot be measured directly in behavioral experiments. Such variables are the prediction error in reinforcement learning models, which represents the deviation between expected and actual outcomes, or the decision threshold in sequential sampling models, which determines how much information needs to be collected before a decision is made. Furthermore, neuroimaging techniques allow us to develop theories that describe how the brain implements decision-making mechanisms.

Thus, to further our understanding of reward-based decision making and decision making under risk, and to develop integrative theories that explain reward-based decision making on different phenomenological levels, we are developing simple mathematical models that are derived from adaptive models of decision making and learning. These models are a central tool of our research allowing us to derive predictions for behavioral and neuroimaging data.