A hierarchical reinforcement learning model explains individual differences in attentional set shifting.

Area of Science:

• Cognitive neuroscience
• Computational psychiatry
• Psychometrics

Background:

• Attentional set shifting is crucial for cognitive flexibility, showing significant variation in the general population.
• Impairments in set shifting are observed in psychiatric conditions, but underlying mechanisms are unclear.
• The Cambridge Neuropsychological Test Automated Battery (CANTAB) Intra-Extra Dimensional (IED) set-shift task is a standard measure.

Purpose of the Study:

• To investigate the computational mechanisms underlying attentional set shifting.
• To explore the relationship between set shifting performance and psychiatric symptoms, particularly compulsive symptoms.
• To develop and validate a novel modeling approach for analyzing CANTAB IED data.

Main Methods:

• Utilized a hierarchical computational modeling approach.
• Analyzed data from two large-scale, independent online general-population samples performing the CANTAB IED task.
• Included psychiatric symptom assessment in one of the samples.

Main Results:

• A hierarchical model learning both feature values and dimension attention best explained the observed data.
• Compulsive symptoms were significantly associated with slower learning rates.
• Individuals with higher compulsive symptoms exhibited greater attentional bias towards the initial relevant stimulus dimension.

Conclusions:

• The findings suggest a specific computational mechanism for attentional set shifting.
• The study provides a potential mechanistic explanation for individual differences in set shifting ability.
• The results highlight a link between compulsive symptoms and specific deficits in learning and attention within set shifting.