Forgetting phenomena in the Iowa Gambling Task: a new computational model among diverse participants

Area of Science:

• Cognitive Science
• Computational Neuroscience
• Psychology

Background:

• The Iowa Gambling Task (IGT) is crucial for assessing decision-making and executive functions.
• Current computational models often overlook the impact of forgetting on dynamic decision processes.

Purpose of the Study:

• To introduce the Exploitation and Exploration with Forgetting (EEF) model, integrating forgetting into a reinforcement-learning framework.
• To enhance the understanding of decision-making by accounting for information decay.

Main Methods:

• Developed the EEF model incorporating a dynamic forgetting parameter (λ) and first-choice priors.
• Fitted the EEF model to IGT data from 504 healthy individuals.
• Assessed model performance using goodness-of-fit metrics, simulations, and behavioral validation.

Main Results:

• The EEF model demonstrated superior fit compared to five established models across multiple cohorts.
• Introduced novel metrics (SED and FI) to quantify forgetting's influence on exploratory behavior.
• Observed age and gambling frequency effects on forgetting and decision strategies.

Conclusions:

• Forgetting plays a fundamental role in complex decision-making.
• The EEF framework provides novel insights into cognitive dynamics and decision strategies.
• The EEF model serves as a valuable tool for future computational and empirical research on decision-making.