Distinct cognitive trajectories in the early course of psychosis are associated with clinical and functional outcomes longitudinally

Cognitive dysfunction is a core dimension in psychotic disorders and among the strongest predictors of disability and poor quality of life. Cognitive impairments are highly heterogeneous, and cross-sectional studies have consistently found evidence of distinct cognitive profiles both within diagnoses and transdiagnostically. Findings regarding the course of cognitive impairments over time have been mixed. We hypothesized that subgroups of patients in the early course of psychosis show distinct cognitive trajectories that can be identified using data-driven methods, and that these subgroups differ on clinical and functional outcomes over time. Persons with schizophrenia-spectrum disorders or mood disorders with psychosis in the early course of illness (N=127) were assessed using clinical, functional and cognitive measures at three timepoints: baseline, 8 and 16 months. Group-based trajectory modeling was used to identify cognitive subgroups, which were then compared on clinical and functional measures using multilevel models. We identified three distinct cognitive subgroups: an Impaired group, an Average group, and a High-Functioning group. Cognition was stable over the follow-up period in the Impaired and High-Functioning groups, whereas the Average group showed cognitive improvement. Groups did not differ in terms of diagnostic distribution, baseline clinical symptoms, and most baseline functional and demographic measures. However, over the follow-up, group membership predicted changes in negative symptoms, social functioning, and patient-reported outcomes, with the Impaired group showing the most severe illness course. We conclude that patients in the early course of psychosis show distinct cognitive trajectories that predict future symptoms and social functioning, despite presenting no clinical differences at baseline. These findings have implications for understanding biology-cognition associations, which may be related to heterogeneity; developing predictive models for clinical and functional outcomes; and personalizing treatment to support patients' cognitive, clinical and functional needs towards improving illness course.