Dual-process modeling of sequential decision making in the balloon analogue risk task.

Area of Science:

• Cognitive psychology
• Computational neuroscience
• Decision science

Background:

• Understanding sequential risk-taking is crucial as people frequently face uncertain decisions.
• Dual-process theory posits two cognitive systems: System 1 (fast, automatic) and System 2 (slow, conscious).
• The computational basis of decision-making in tasks like the Balloon Analogue Risk Task (BART) remains unclear.

Purpose of the Study:

• To computationally model the decision-making process in sequential risk-taking tasks.
• To evaluate different dual-process model architectures for explaining behavior in the BART.
• To identify how System 1 and System 2 interact during risky decision-making.

Main Methods:

• Developed a pool of computational dual-process models by combining BART models with various architectures of System 1 and System 2.
• Evaluated models using performance metrics including parameter identifiability, model recovery, and predictive accuracy.
• Utilized the Scaled Target Learning model of the BART for evaluation.

Main Results:

• The optimal model featured a competitive connection between System 1 and System 2.
• This best-performing model incorporated an assessment rate sensitive to trial number, pumping opportunity, and System 1 bias.
• Model performance was assessed through parameter identifiability, model recovery, and predictive accuracy.

Conclusions:

• A competitive interaction between fast (System 1) and slow (System 2) cognitive processes best explains sequential risk-taking behavior.
• Modeling choices and response times within a dual-process framework enhance understanding of risky decision-making.
• Findings provide insights into the cognitive mechanisms underlying uncertainty and risk in sequential tasks.