Effect of Polygenic Scores on the Relationship Between Psychosis and Cognitive Performance

Abstract

Background: Up to 80% of psychosis patients experience cognitive impairment. High heritability of both psychosis and cognition means cognitive performance could be an endophenotype for psychosis. Methods: Using samples of adults (N=4,506) and children (N=10,981), we investigated the effect of polygenic scores (PGSs) for schizophrenia and bipolar disorder on cognitive performance, and PGSs for intelligence and educational attainment on psychosis symptoms. Results: Schizophrenia PGS was negatively associated with visuospatial processing/problem-solving in the adult sample (beta: -0.0569; 95% confidence interval [CI]: -0.0926, -0.0212) and working memory (beta: -0.0432; 95% CI: -0.0697, -0.0168), processing speed (b: -0.0491; 95% CI: -0.0760, -0.0223), episodic memory (betas: -0.0581 to -0.0430; 95% CIs: -0.0847 to -0.0162), executive functioning (beta: -0.0423; 95% CI: -0.0692, -0.0155), fluid intelligence (beta: -0.0583; 95% CI: -0.0847, -0.0320), and total intelligence (beta: -0.0458; 95% CI: -0.0709, -0.0206) in the child sample. Bipolar disorder PGS was not associated with any cognitive endophenotypes studied. Lower values on the PGS for intelligence were associated with higher odds of psychosis in adults (odds ratio [OR]: 0.886; 95% CI: 0.811-0.968) and psychotic-like experiences in children (OR: 0.829; 95% CI: 0.777-0.884). In children, a lower polygenic score for educational attainment was associated with greater odds of psychotic-like experiences (OR: 0.771; 95% CI: 0.724-0.821). Conclusions: The relationship between psychosis and cognitive impairment can be demonstrated bidirectionally at the neurobiological level. The effect of schizophrenia PGS on cognitive performance differs across the lifespan and cognitive domains. Specific cognitive domains may therefore be better endophenotypes than overall cognition.