# Polygenic scores for executive functioning as predictors of performance improvements after repeated testing in major psychiatric disorders

## TL;DR

This study explores how genetic factors influence improvements in executive function performance over time in individuals with psychiatric disorders.

## Contribution

The study is the first to analyze how polygenic scores for executive functioning predict performance improvements in psychiatric patients.

## Key findings

- Polygenic scores for executive function (PGS-cEF) predicted improvements in latent executive function scores over 18 months.
- The prediction was only observable at the latent level, not in individual test scores.
- Psychotic disorders had lower PGS-cEF and higher PGS-PF compared to affective disorders and controls.

## Abstract

Executive Functions (EF) control goal-oriented behaviors and are disrupted in patients with psychiatric disorders (PPD). Repeated EF testing in clinical practice and research aims to evaluate disease progression and treatment efficacy. Performance improvements (or “practice effects”) are expected, while deficiencies had been proposed as markers of conversion to dementia. Given the high conversion risk in PPD, understanding the determinants of practice effects is of high relevance. However, the role of genetic predisposition remains unknown. We aim to test the influence of polygenic scores on practice effects across PPD. Data were drawn from the longitudinal deep-phenotyping PsyCourse Study (N = 1,558). Latent factor analysis revealed a common Executive Function factor (cEF), which explained shared variability of 5 different EF tests. We tested the association of polygenic scores for the cEF (PGS-cEF) – based on the largest GWAS on cEF to date (N = 427,037) - with cEF scores improvements in PsyCourse. The PGS for general psychopathology “p-Factor” (PGS-PF) were used as a psychiatric liability comparison. Volunteer adults (> 18 y.o.), with and without diagnoses within the Affective-Psychotic spectrum were recruited across Germany and Austria, followed-up four times at six months intervals. The mean PGS-cEF was lower in patients with psychotic vs. affective disorders (diff: -0.01, 95% CI -0.02 to -0.002) and controls (diff: -0.02, 95% CI -0.03 - -0.006); while the mean PGS-PF was higher in psychotic vs. affective disorders (diff: 0.02, 95% CI 0.01 to 0.04) and controls (diff: 0.03, 95% CI 0.02–0.05). Overall, the PGS-cEF x Visit interaction was associated with the phenotypic latent cEF scores (eta2 = 0.03; p = 1.32e-15), but not with the scores of individual EF tests. There was no PGS-PF x Visit interaction associated with any cognitive score. To the best of our knowledge, this is the first analysis of genetic factors related to EF improvements in PPD. The PGS-cEF, but not the PGS-PF predicted practice effects in cEF scores over 18 months. This prediction was only observable using the latent‑cEF level, and not with the individual scores. Genetic testing and deep phenotyping might hint further information about EF vulnerability in PPD.

The online version contains supplementary material available at 10.1038/s41598-026-41345-1.