Depressive symptom-cognition associations in major depressive disorder and bipolar disorder: A diagnosis-stratified analysis

Background: Cognitive impairment is common in mood disorders. Most studies relate cognition to total depressive severity rather than individual symptoms. Aims: We examined associations between individual Montgomery-Åsberg Depression Rating Scale (MADRS) items and cognitive domains separately in major depressive disorder (MDD) and bipolar disorder (BD). We also explored whether associations differed by diagnosis. Method: This cross-sectional study included 225 participants with MDD and 231 with BD. Cognitive scores were standardized against 177 healthy controls. Age-, sex-, and education-adjusted linear models tested 60 item-domain associations in each diagnosis with false discovery rate (FDR) correction. Follow-up analyses included diagnosis-by-item interactions, rest-score adjustment, healthy-control norm bootstrapping, and reverse and continuous models. Results: In MDD, 10/60 associations survived FDR correction. Apparent sadness, reported sadness, and concentration difficulties were associated with global cognition, inhibition, working memory, or processing speed. In BD, only suicidal thoughts (MADRS item 10) were associated with worse inhibition (β = 0.439, FDR p = 0.002). In the reverse model, each 1-SD worse inhibition score was associated with higher odds of item 10 scores of at least two (OR = 1.419, 95% CI 1.079–1.866; p = 0.012). No diagnosis-by-item interaction survived FDR correction across the complete 60-test family. Conclusions: MADRS item-cognition associations showed different descriptive patterns within MDD and BD. The data did not confirm diagnosis-specific effects. The BD item 10-inhibition association was robust within this sample but requires independent and longitudinal replication.