Feed

Cognitive deficits across multiple domains are prevalent in patients with schizophrenia (PWS), and metabolic syndrome (MetS) may significantly contribute to this impairment. To clarify the complex relationships between individual MetS components and multidimensional cognitive dysfunction in PWS, we conducted a multicenter study involving 727 clinically stable patients recruited from ten psychiatric hospitals. Cognitive function was assessed using the Chinese Brief Cognitive Test (C-BCT). We employed network analysis and structural equation modeling (SEM) to explore these associations, with machine learning techniques applied for further validation. The results revealed statistically significant differences in several cognitive domains between patients with and without dyslipidemia (DL). Patients with hypertension (HT) also exhibited overall poorer cognitive performance. Network analysis indicated meaningful distinctions between patients presenting two or more MetS components (MetS-2+) and those without, showing a sparser network configuration in the MetS-2+ group. Across both groups, the Symbol Coding task demonstrated the highest strength centrality. SEM indicated that metabolic indicators, specifically DL and HT, mediated the relationship between clinical symptoms and cognitive function. Furthermore, a transformer-based machine learning model performed effectively in predicting cognitive dimensions, supporting the predictive utility of MetS components for multidimensional cognitive outcomes. In summary, specific MetS components, particularly DL and HT, show intricate associations with cognitive function in stable-phase PWS. Our findings suggest that management of HT in this population may represent a potential pathway for cognitive enhancement and improved social functioning. Trial registration: MR-11-23-007343.