Quantifying value-based determinants of drug and non-drug decision dynamics

Area of Science:

• Neuroscience
• Behavioral Economics
• Addiction Research

Background:

• Substance use disorder (SUD) is increasingly viewed as a decision-making impairment.
• Existing research often uses simplified lab conditions, potentially limiting understanding of drug value in complex environments.

Purpose of the Study:

• To investigate how dynamic reward conditions and predictive cues influence choices between food and remifentanil (REMI) in rats.
• To analyze choice data using quantitative matching and reinforcement learning (RL) models to understand determinants of drug value.

Main Methods:

• Rats made choices between identical food options (isomorphic) and between remifentanil (REMI) and food (allomorphic).
• Reward probabilities, magnitudes, and predictive cues were dynamically altered.
• Choice data were analyzed using quantitative matching and reinforcement learning models.

Main Results:

• Decreased reward probability or magnitude reduced preference for both food and REMI.
• Reward-predictive cues enhanced preference for both food and REMI rewards.
• Models indicated cues altered economic substitutability and supported a shared, relative reward value update in RL models.

Conclusions:

• Value-based choice models effectively capture decision-making in complex contexts, supporting drug choice as a multidimensional valuation process.
• Decision-making for drugs and non-drug rewards shares commonalities.
• Findings validate economic SUD therapies and highlight neurobehavioral processes for future treatment development.