by Zhang Nannan, Chinese Academy of Sciences

The simulation results demonstrate the superiority of the Hybrid-WM model over the classical Hybrid-PRE model. Credit: Yang Lizhuang

A study led by Prof. Li Hai from the Hefei Institutes of Physical Science of the Chinese Academy of Sciences has revealed that the balance between habitual and goal-directed decision-making strategies is influenced by the availability of working memory resources.

The findings, published in the Journal of Cognitive Neuroscience, provide a new framework for understanding how sequential decisions are made.

Everyday decisions often involve a series of choices aimed at reaching a goal-whether selecting a restaurant or deciding on the route. People vary in how they make decisions: some rely on habits, while others adjust based on new information and changing goals.

The key question, however, is what determines the balance between these two strategies. This challenge, known as meta-control, has long puzzled cognitive neuroscience.

To address this, the researchers combined behavioral experiments with computational modeling, demonstrating that working memory limitations play a crucial role in decision-making. They introduced the Hybrid-WM reinforcement learning model, which integrates working memory constraints into the decision-making process.

"Unlike older models, this new framework takes into account the brain's memory limits," said Prof. Li Hai. "By accounting for the brain's memory limits, it offers a more accurate representation of how we make decisions under various conditions—such as stress, distractions, or mental fatigue."

This research quantifies the role of working memory constraints in meta-control and offers new insights into sequential decision-making processes. "The findings hold promise for applications in human-computer interaction," said Prof. Li.

More information: Zhaoyu Zuo et al, Working Memory Guides Action Valuation in Model-based Decision-making Strategy, Journal of Cognitive Neuroscience (2024). DOI: 10.1162/jocn_a_02237

Journal information: Journal of Cognitive Neuroscience 

Provided by Chinese Academy of Sciences