Addictive behaviors constitute a massive societal problem in terms of both quantifiable costs and an unquantifiable toll on intrapersonal relationships and personal happiness.
Formal models of reinforcement learning have already played an exciting role in understanding neural function in reward processing circuits in the brain and in leading to the development of improved treatments for addictive disorders. On the other hand, individual differences in susceptibility to addictive behaviors are widely acknowledged and have been characterized in terms of social and inheritable risk factors. Yet very little is known about how and to what extent differences in how individuals implement reinforcement learning algorithms in their behaviors confers risk. We thus plan to carefully characterize individual differences in reinforcement learning strategies across a large population of volunteers, and relate this information to the display of addictive behavior to gain fresh insight into risk factors that are currently unknown and to lay the groundwork for mechanism-based individualized treatments.
We will develop a small battery of simple cognitive and motor skill learning games that can characterize individual differences in reinforcement learning strategies that dictate individual performance. We will couple this with surveys that provide insight into addictive tendencies and behaviors. The convergence of this information will allow us to leverage the Lab-in-the-Wild platform to relate the diversity in the features of human reinforcement learning to a spectrum of addictive behaviors.
Team: Krzysztof Gajos, Maurice Smith, and one postdoctoral fellow