Optimal response vigor and choice under non-stationary outcome values

Area of Science:

• Decision neuroscience
• Computational psychiatry
• Behavioral economics

Background:

• Decision-making involves selecting actions and their vigor based on reward maximization.
• Previous models assumed constant outcome values, limiting applicability to dynamic environments.

Purpose of the Study:

• To develop a computational model for optimal action selection under non-stationary outcome values.
• To investigate the impact of changing outcome values on action vigor and strategy.

Main Methods:

• Derived a computational model from an optimal control framework.
• Analyzed action selection dimensions: alternative choice and vigor.
• Incorporated uncertainty about temporal horizons (e.g., session duration).

Main Results:

• Optimal response rate remains constant despite non-stationary outcome values.
• Changes in action rate are not solely due to outcome value fluctuations.
• Model explains both probability matching and maximization strategies.

Conclusions:

• The derived model offers a quantitative analysis of optimal decision-making under dynamic conditions.
• Uncertainty in temporal horizons explains constant optimal response rates.
• Provides testable predictions for future empirical research on action selection.