The scaled target learning model: Revisiting learning in the balloon analogue risk task

Area of Science:

• Cognitive Psychology
• Computational Neuroscience
• Behavioral Economics

Background:

• The Balloon Analogue Risk Task (BART) is a key paradigm for assessing risk-taking behavior.
• Existing computational models for BART often struggle to capture the learning dynamics influenced by trial outcomes.
• Understanding adaptive learning is crucial for a comprehensive view of risky decision-making.

Purpose of the Study:

• To introduce and validate the Scaled Target Learning (STL) model for analyzing risk-taking in the BART.
• To model how individuals adjust their strategies based on experienced wins and losses.
• To enhance the characterization of psychological processes in sequential risk-taking.

Main Methods:

• Development of the Scaled Target Learning (STL) model, which quantifies strategy adjustments based on task outcomes.
• Assessing model sensitivity to experimental manipulations designed to elicit learning.
• Comparing STL performance against established computational models using metrics like parameter recovery and predictive accuracy.

Main Results:

• The STL model effectively captures learning dynamics in the BART, reflecting sensitivity to wins and losses.
• STL demonstrates robustness in parameter recovery and predictive accuracy.
• The model's performance equals or surpasses that of three competing computational models.

Conclusions:

• The STL model provides a more complete depiction of the psychological processes underlying sequential risk-taking behavior.
• STL's ability to characterize adaptive strategy adjustments offers new insights into risk-taking.
• This work advances computational approaches to understanding decision-making under uncertainty.