Incentives Boost Model-Based Control Across a Range of Severity on Several Psychiatric Constructs

Area of Science:

• Cognitive Neuroscience
• Computational Psychiatry
• Behavioral Economics

Background:

• Human decision-making involves a balance between model-based and model-free control strategies.
• Metacontrol arbitrates between these strategies based on their computational costs and benefits.
• Model-based control offers accuracy but demands significant resources, making it sensitive to reward levels.

Purpose of the Study:

• To investigate how psychopathology symptoms and traits influence the use of model-based control in response to varying incentives.
• To determine if incentives can modulate model-based control irrespective of psychological construct severity.

Main Methods:

• A sequential decision task was employed with 839 online participants.
• Performance incentives were manipulated (small vs. large rewards).
• A dual-controller reinforcement learning model was used to quantify model-based control.

Main Results:

• No constructs were linked to a failure of incentives to boost model-based control.
• Sensation seeking and anxious-depression traits positively correlated with increased model-based control under high incentives.
• Most constructs showed decreased model-based control with severity, but social withdrawal was positively correlated.

Conclusions:

• Model-based control can be reliably enhanced by incentives across diverse symptom and trait severities.
• Incentives show potential as an intervention to improve model-based control in various populations.