This study integrated brain eQTL–predicted gene-expression features with resting-state functional network connectivity (FNC) using an independent multimodal decomposition framework. Gene-expression and FNC features were independently decomposed into latent components. These latent components were then screened for associations with schizophrenia and for sex differences, and cross-modal associations between gene-expression components and FNC components were tested.
The approach applied publicly available templates and methods to produce comparable intrinsic functional networks and genetically regulated expression projections. The analysis focused on whether genetically regulated molecular variation couples to large-scale functional dysconnectivity in schizophrenia and whether that coupling is moderated by sex.
Primary analyses used 367 participants pooled from the FBIRN and COBRE datasets. Replication analyses used UK Biobank data in two sets: 2,273 individuals (cases and matched controls) to test the schizophrenia-associated gene-expression effect, and 4,206 participants with matched genetic and imaging data to test the multimodal association and its sex-dependent interaction.
All human data used were de-identified and accessible through controlled-access application procedures (FBIRN and COBRE via COINSTAC; UK Biobank via application). The NeuroMark 1.0 template of intrinsic functional networks was referenced for network definitions.
From the eQTL-predicted gene-expression features, the investigators identified a latent gene-expression component that associated with schizophrenia status. This component also exhibited a significant sex difference, indicating that genetically regulated expression variance relevant to schizophrenia differed between males and females in the discovery cohorts.
Component identification proceeded through multimodal decomposition and statistical screening; specific decomposition algorithmic parameters and statistical thresholds were described in the source methods (details not restated here beyond those reported in the source).
The schizophrenia-associated gene-expression component was enriched in cerebellar tissues. Functional enrichment analyses implicated biological processes related to vesicle and membrane organization and intracellular trafficking. These enrichments suggest that the genetically regulated expression signal tied to schizophrenia in this component relates to cellular processes important for synaptic and membrane dynamics, with a tissue preference toward the cerebellum.
The identified schizophrenia-associated gene-expression component was significantly associated with a schizophrenia-associated FNC component. The FNC component encompassed cortical, subcortical, and cerebellar systems, indicating that the molecular regulatory signal linked to schizophrenia covaried with distributed functional-network dysconnectivity across these brain systems.
This cross-modal coupling supports a convergence of genetically regulated molecular variation and macroscopic functional-network alterations in schizophrenia.
A significant sex-by-gene interaction characterized the multimodal association. Specifically, the association between the gene-expression component and the FNC component was stronger and negative in males, whereas in females the association was weaker and positive. In other words, the coupling pattern between genetically regulated expression and functional dysconnectivity differed by sex, with opposite directions and magnitudes reported in the source.
This sex modulation indicates that sex moderates how regulatory liability maps onto large-scale functional-network abnormalities in schizophrenia.
The schizophrenia-associated gene-expression effect observed in the discovery cohorts replicated in UK Biobank in an analysis of 2,273 individuals with schizophrenia and matched controls. The multimodal association between the gene-expression component and the FNC component, and its sex-dependent interaction, also replicated in a separate UK Biobank set of 4,206 participants with matched genetic and imaging data.
Replication across these larger, independent samples supports the robustness of both the gene-expression schizophrenia association and the sex-modulated multimodal coupling reported in the discovery cohorts.
The findings imply that genetically regulated gene-expression variation associated with schizophrenia may converge with distributed functional-network dysconnectivity, and that sex moderates this coupling across biological scales. Enrichment in cerebellar tissue and trafficking-related biological processes points to specific cellular and regional contexts for the regulatory signal.
Limitations reported in the source include reliance on openly available, de-identified datasets and the analytic choices inherent to eQTL-predicted expression and multimodal decomposition; further methodological specifics and potential confounds are described in the original preprint. The source did not report additional experimental details beyond those summarized here.
Overall, the study provides evidence for sex-modulated molecular-to-network coupling in schizophrenia that replicated in UK Biobank samples, suggesting a moderated pathophysiological pathway deserving continued investigation.