Abstract: Background: Schizophrenia (SCZ) is a severe neuropsychiatric disorder characterized by significant cognitive dysfunction, with 70-75% of patients performing below general population standards on cognitive tasks. Cognitive deficits, pronounced during first-episode psychosis (FEP) and worsening over time, are linked to brain abnormalities and deterioration. There is a genetic predisposition for cognitive performance and brain structural traits in psychotic and healthy subjects. This study tests various polygenic risk scores (PRS) for their ability to predict cognitive performance and brain structural metrics in FEP and healthy controls. Methods: Peripheral blood samples from 739 individuals (187 healthy controls and 552 FEP patients) were genotyped using the Infinium Global Screening Array-24 Kit (Illumina). PRS were estimated for schizophrenia (SCZ), neurostructural traits, cognitive traits, and dementia using available GWAS summary statistics. Correlation analyses identified non-correlated PRS within clusters. A subset of 139 FEP patients and 119 controls underwent MRI scans at baseline and at 1-, 3-, 5-, 10-, and 15-years follow-ups. Cognitive functions were assessed at baseline and 1, 3, and 10 years across seven domains, a global cognitive function (GCF) metric, and a global deficit score (GDS). Linear regression and mixed-effects models assessed the predictive ability of selected PRS for cognition and atypical trajectories of cortical thickness and subcortical volumetric maturation (quantified using centiles) at baseline and longitudinal conditions. Causal mediation analyses evaluated the mediation of neurostructural traits in PRS-cognition associations. FDR correction was applied to results. Results: PRS were estimated for 507 FEP patients and 162 controls. Educational Attainment (EA), brain volume, SCZ, and dementia PRS were selected for regression analyses. EA PRS predicted baseline and longitudinal processing speed (PS) in controls (Beta= 0.22/ 0.14). Dementia PRS predicted visual memory (ViM) in controls (Beta= -0.24); and ViM, motor dexterity (MD), attention (Att), GCF, and GDS in FEP at baseline (Beta= -0.15/ -0.37/ -0.75/ -0.19/ 0.17); and GCF and GDS in FEP at longitudinal condition (Beta= -0.13/ 0.10). SCZ PRS predicted all cognitive domains, GCF and GDS in controls at baseline (Beta= -0.63 (GCF)/ 0.60 (GDS)/-1.04 (PS)/ -0.44 (ViM)/ -0.86 (MD)/ -1.47 (Att)/ -0.73 (executive functioning, EF)/ -0.70 (verbal memory, VeM)/ -0.33 (working memory, WM)) and longitudinal conditions (Beta= -0.60 (GCF)/ 0.54 (GDS)/ -0.72 (PS)/ -0.54 (ViM)/ -0.42 (MD)/ -0.35 (Att)/ -0.32 (EF)/ -0.52 (VeM)/ (WM)). SCZ PRS predicted baseline and longitudinal global cortical thickness centiles in controls (Beta= -0.50/-0.38). No association was found for any regional brain centile. A mediation role for global cortical thickness centiles was found in the association between SCZ PRS and GCF, GDS, VeM, ViM, and PS at baseline (ACME= -0.10/ 0.10/ -0.10/ -0.07/ -0.07). Disease status did not moderate the mediation. Discussion: This study leverages the unique advantage of a well-characterized, longitudinal dataset spanning 15 years, including imaging and cognitive assessments in FEP and healthy individuals. Further research in this area is warranted to better understand and address cognitive deficits in psychosis, highlighting targets for therapeutic intervention.

Autoría

Descargar referencia