Mapping the genetic architecture of cognitive functions is a fundamental milestone to understanding the biological processes underlying human cognition. Decades of research have revealed the complexity of genetic composition of general intelligence, g. However, research on genetic contribution to more specific cognitive functions has been less successful.
Through a combined approach employing Exploratory Factor (EF) and Structural Equation Modelling (SEM) Analyses, we have constructed multifactorial models of cognitive testing data from the UK Biobank and ABCD study. We have identified a three-factor model of cognitive functioning, capturing fluid (Gf) and crystallised (Gc) aspects of cognition across both healthy adult (UK Biobank) (Ciobanu et al 2023) and adolescent (ABCD study) populations.
To explore the genetic determinants of these specific cognitive abilities, we applied an advanced technique for network-based polygenic scores calculation – a weighted single nucleotide polymorphism (SNP) correlation network analysis (WSCNA). Through comparison across the UK Biobank and ABCD study datasets, our findings indicate that network based PGS can provide more targeted polygenic scores, explaining higher genetic variance in cognitive phenotypes compared to traditional linear PGS. Moreover, we conducted functional characterisation of the SNP clusters predictive of specific cognitive functions in both datasets, revealing the biological significance of genetic networks implicated in cognition.
Our study introduces a novel framework for investigating the genetic underpinnings of cognitive functions beyond g, with potential implications for understanding other complex phenotype traits.