Decomposition of Reinforcement Learning Deficits in Disordered Gambling via Drift Diffusion Modeling and Functional...

Area of Science:

• Neuroscience
• Computational Psychiatry
• Behavioral Economics

Background:

• Gambling disorder is linked to reward learning deficits.
• The computational mechanisms underlying these deficits are not well understood.

Purpose of the Study:

• To investigate the computational mechanisms of reinforcement learning deficits in gambling disorder.
• To compare individuals with gambling disorder and controls using computational modeling and fMRI.

Main Methods:

• A stationary reinforcement learning task was administered to individuals with gambling disorder (n=23) and matched controls (n=23).
• Reinforcement learning drift diffusion models (RLDDMs) were used for computational modeling.
• Functional resonance imaging (fMRI) was employed to examine neural correlates.

Main Results:

• The gambling group showed reduced accuracy but similar response times compared to controls.
• RLDDMs indicated that gambling disorder was associated with a more rapid reduction in decision thresholds and a reduced impact of value differences on drift rate.
• Shorter non-decision times were observed in the gambling group.
• fMRI revealed no significant group differences in prediction error or value coding in key brain regions.

Conclusions:

• Reinforcement learning impairments in gambling disorder are characterized by maladaptive decision threshold adjustments.
• Reduced consideration of option values during decision-making contributes to gambling disorder.
• Computational modeling provides insights into the cognitive processes underlying gambling disorder.