Identifying cognitive remediation change through computational modelling--effects on reinforcement learning in...

Area of Science:

• Neuroscience
• Cognitive Psychology
• Psychiatry

Background:

• Individuals with schizophrenia exhibit significant deficits in reinforcement learning, particularly in adaptive and flexible task performance.
• These learning impairments are often linked to dysfunctions within the brain's dopamine reward systems.

Purpose of the Study:

• To characterize reinforcement learning impairments in schizophrenia using computational modeling.
• To investigate the efficacy of cognitive remediation therapy (CRT) in ameliorating these specific learning deficits.

Main Methods:

• Computational modeling estimated three reinforcement learning parameters (reward sensitivity R, punishment sensitivity P, choice consistency D) from Wisconsin Card Sorting Test (WCST) data.
• Study 1 compared these parameters between individuals with schizophrenia (n=100) and healthy controls (n=50).
• Study 2 assessed CRT's impact on these parameters in individuals with schizophrenia, comparing a CRT group (n=37) with a treatment as usual (TAU) group (n=34).

Main Results:

• Individuals with schizophrenia demonstrated significant impairments in reward (R) and punishment (P) sensitivity compared to controls.
• CRT significantly improved both reward (R) and punishment (P) sensitivity in the schizophrenia group compared to TAU.
• Improvements in R and P correlated with WCST performance and were associated with gains in working memory and reduced negative symptoms, respectively.

Conclusions:

• Reinforcement learning deficits, specifically in reward and punishment sensitivity, are a key feature of schizophrenia impacting cognitive flexibility.
• Cognitive remediation therapy shows promise in enhancing these crucial learning parameters.
• The identified parameters (R and P) may serve as valuable biomarkers for assessing treatment efficacy in experimental medicine studies for schizophrenia.