Abstract
BACKGROUND: Schizophrenia spectrum disorders (SSDs) often feature social cognitive deficits. However, little work has focused on the factor structure of social cognition, and results have been inconsistent in schizophrenia. This study aimed to elucidate the factor structure of social cognition across people with SSDs and healthy controls. It was hypothesized that a 2-factor model, including lower-level "simulation" and higher-level "mentalizing" factors, would demonstrate the best fit across participants. METHODS: Participants with SSDs (N = 164) and healthy controls (N = 102) completed social cognitive tasks ranging from emotion recognition to complex mental state inference, as well as clinical and functional outcome, and neurocognitive measures. Structural equation modeling was used to test social cognitive models, models of social cognition and neurocognition, measurement invariance between cases and controls, and relationships with outcome measures. RESULTS: A 2-factor (simulation and mentalizing) model fit the social cognitive data best across participants and showed adequate measurement invariance in both SSD and control groups. Patients showed lower simulation and mentalizing scores than controls, but only mentalizing was significantly associated with negative symptoms and functional outcome. Social cognition also mediated the relationship between neurocognition and both negative symptoms and functional outcome. CONCLUSIONS: These results uniquely indicate that distinct lower- and higher-level aspects of social cognition exist across SSDs and healthy controls. Further, mentalizing may be particularly linked to negative symptoms and functional outcome. This informs future studies of the neural circuitry underlying social cognition and the development of targeted treatment options for improving functional outcome.