Disrupted state transition learning as a computational marker of compulsivity.

Area of Science:

• Neuroscience
• Computational Psychiatry
• Cognitive Psychology

Background:

• Disorders characterized by compulsivity, fear, and anxiety are associated with a diminished perception of environmental predictability.
• The underlying mechanisms driving these beliefs and their impact on learning remain poorly understood.
• This research investigates a potential mechanistic link between compulsivity and impaired probabilistic learning.

Purpose of the Study:

• To test the hypothesis that individuals with compulsivity exhibit deficits in learning probabilistic action-state contingencies.
• To elucidate the nature of this learning impairment, specifically whether it involves learning that is too fast or too slow.
• To identify potential therapeutic targets for conditions involving compulsivity.

Main Methods:

• A novel online task was developed to isolate state transition learning.
• Computational models were employed to estimate state transition learning rates.
• Learning rates were analyzed across three studies with varying environmental stability (stable and changing state transitions).

Main Results:

• Individuals with higher compulsivity levels demonstrated impaired state transition learning.
• Compulsivity was associated with learning that was too fast in stable environments and too slow in changing environments.
• Preliminary findings suggested a link between this impairment and a shared factor of compulsivity and fear.

Conclusions:

• Compulsivity is characterized by dysregulated state transition learning, where learning rates are maladaptive to environmental dynamics.
• This dysregulation represents a potential mechanism contributing to the beliefs associated with compulsivity, fear, and anxiety.
• Targeting dysregulated state transition learning may offer a novel therapeutic avenue for compulsivity.