Reinforcement learning and counterfactual reasoning explain adaptive behavior in a changing environment.

Area of Science:

• Cognitive Science
• Neuroscience
• Animal Behavior

Background:

• Animals adapt to environmental shifts for survival.
• Reinforcement learning is a known adaptation mechanism.
• Reinforcement learning may not fully explain rapid strategy changes.

Purpose of the Study:

• Investigate counterfactual reasoning as a mechanism for environmental change detection.
• Explore its role alongside reinforcement learning.
• Understand its contribution to rapid adaptation.

Main Methods:

• An experiment involving change detection in a dynamic task environment.
• Participants performed tasks with changing states and received monetary rewards.
• A cognitive model integrating reinforcement learning and counterfactual reasoning was developed.

Main Results:

• The cognitive model accurately explained human behavioral data.
• Counterfactual reasoning was identified as essential for explaining observed effects.
• The model demonstrated rapid adjustment of strategy utility.

Conclusions:

• Counterfactual reasoning is a key mechanism facilitating change detection.
• It complements reinforcement learning in adaptive behavior.
• This integration explains rapid behavioral adjustments to environmental changes.