Investigating disorder-specific and transdiagnostic alterations in model-based and model-free decision-making.

Knolle, Franziska; Sen, Pritha; Culbreth, Adam; Koch, Kathrin; Schmitz-Koep, Benita; Gürsel, Deniz A.; Wunderlich, Klaus; Avram, Mihai; Berberich, Götz; Sorg, Christian; Brandl, Felix

Background: Decision-making alterations are present in psychiatric illnesses like major depressive disorder (MDD), obsessive–compulsive disorder (OCD), and schizophrenia, linked to symptoms of the respective disorders. We sought to analyze unique and shared decision-making alterations in these disorders, which is crucial for early diagnosis and treatment, especially given potential comorbidities. Methods: Using 2 computational modelling approaches — logistic regression and hierarchical Bayesian modelling — we analyzed alterations in model-based and model-free decision-making in a transdiagnostic cohort of patients with MDD, OCD, or schizophrenia. Our aim was to identify disorder-specific and shared alterations and their associations with symptoms. Results: We included 23 patients with MDD, 25 patients with OCD, 27 patients with schizophrenia, and 25 controls. Overall, participants of all groups relied on model-free decision-making. Patients with schizophrenia had the lowest learning rate and highest switching rate, indicating low perseverance. Furthermore, patients with OCD were more random in both task stages than controls and patients with MDD. All patient groups exhibited more randomness in responses than controls, with the schizophrenia group showing the highest levels. Increased model-free behaviour correlated with elevated depressive symptoms, and more model-based decision-making was linked to lower anhedonia levels across all patient groups. Limitations: The sample size in each group was small. Conclusion: This study highlights disorder-specific and shared decision-making alterations among people with MDD, OCD, or schizophrenia. Our findings suggest that anhedonia and depressive symptoms, which are present in all 3 disorders, share underlying behavioural mechanisms. Improving model-based behaviour may be a target for intervention and treatment. Furthermore, completely random behaviour in the 2-step task appears to distinctly differentiate patients with schizophrenia in remission.

DIAGNOSIS of schizophrenia; DIAGNOSIS of mental depression; DIAGNOSIS of obsessive-compulsive disorder; SCHIZOPHRENIA treatment; STATISTICAL models; DECISION support systems; EARLY medical intervention; RESEARCH funding; LOGISTIC regression analysis; DECISION making; OBSESSIVE-compulsive disorder; LONGITUDINAL method; ANHEDONIA; EARLY diagnosis; COMPARATIVE studies; MENTAL depression; COMORBIDITY

Journal of Psychiatry & Neuroscience, 2024, Vol 49, Issue 6, pE389

1180-4882

Academic Journal

10.1503/jpn.240020