Effort and Substance Use: Differentiating Tobacco Use Through Reinforcement Learning of Effort Based Decision Making

Area of Science:

• Neuroscience
• Computational Psychiatry
• Behavioral Economics

Background:

• Effort-based decision-making is crucial for goal-directed behavior.
• Computational models offer insights into choice mechanisms and alterations in disorders like substance use disorders.
• Understanding these processes is key to characterizing motivation and behavior.

Purpose of the Study:

• To apply computational models to effort-based choice behavior.
• To characterize underlying decision processes.
• To determine if these mechanisms differ by substance use status.

Main Methods:

• Participants (n=100) completed the Effort Expenditure for Rewards Task.
• Computational models (Subjective Value, Reinforcement Learning) were fit to choice data.
• Multivariate analyses (PCA, LDA) examined model parameters across groups (no use, tobacco use disorder, former tobacco use disorder, tobacco and opioid use disorder).

Main Results:

• A temporal difference reinforcement learning model provided the best fit to effort-based choice behavior.
• Multivariate analyses revealed distinct patterns of learning rate, future discounting, and choice temperature across substance use groups.
• Linear discriminant analysis achieved 76% classification accuracy for group separation.

Conclusions:

• Reinforcement learning frameworks effectively explain effort-based decision-making.
• Substance use status is associated with dynamic behavioral changes in learning and decision parameters.
• These findings highlight the utility of computational modeling in understanding substance use disorders.