Computational approaches to modeling gambling behaviour: Opportunities for understanding disordered gambling

Area of Science:

• Computational neuroscience
• Psychiatry
• Cognitive science

Background:

• Computational modeling is crucial for understanding normal and pathological behaviors.
• Reinforcement learning (RL) and drift diffusion modeling (DDM) are established frameworks applied to gambling disorder.
• Existing models simplify or overlook key aspects of real-world gambling decision-making.

Purpose of the Study:

• To review the application of RL and DDM in studying cognitive components of gambling and gambling disorder.
• To explore the potential of bespoke modeling approaches for gambling disorder.
• To introduce Bayesian models as a tool for tailored computational approaches.

Main Methods:

• Review of studies utilizing RL and DDM frameworks in the context of gambling disorder.
• Discussion of cognitive components investigated by these models.
• Overview of Bayesian modeling methodologies.

Main Results:

• RL and DDM have provided insights into cognitive processes relevant to gambling disorder.
• Current models have limitations in capturing the full complexity of gambling behavior.
• Bayesian models offer potential for more nuanced and tailored analyses.

Conclusions:

• Computational modeling, particularly RL and DDM, is valuable for investigating gambling disorder.
• There is a need for more sophisticated, bespoke models to address the complexities of gambling behavior.
• Further research utilizing advanced modeling techniques like Bayesian approaches can advance the understanding of gambling disorder's cognitive mechanisms.