Brain imaging-derived phenotypes influence risk of bipolar disorder: comprehensive exposome-wide association and wide-angled genetic analyses
Despite many reported imaging markers for bipolar disorder (BD), systematic assessment across diverse brain imaging-derived phenotypes (IDPs) is lacking. We used epidemiological and genetic analyses to progress from association to causality and identify candidate biomarkers and potential pathways of BD. We conducted an Exposome-Wide Association Study (ExWAS) in 35,680 UK Biobank participants to assess associations of 686 brain IDPs with BD. Genetic correlations were estimated via linkage disequilibrium score regression (LDSC). Bidirectional two-sample Mendelian randomization (MR) and meta-analysis were performed using large-scale GWAS data to infer causality. Additional analyses, including Functional Mapping and Annotation (FUMA), protein-protein interaction (PPI) network analysis, and spatiotemporal gene expression profiling during human brain development, were performed to investigate potential biological mechanisms of BD and IDPs. In our ExWAS analyses, we identified 10 significant exposome-level IDPs, and LDSC further revealed 67 IDPs with significant genetic correlations with BD. The MR analyses provided robust causal evidence for 12 IDPs, 6 of which were further validated by meta-analysis. Integrative analyses identified two associations supported by both epidemiological and genetic evidence (IDP.1510: fractional anisotropy in the forceps major, and IDP.0387: volume of the right lateral orbital frontal cortex). Additionally, we identified pleiotropic genes between IDPs and BD. Gene enrichment analyses implicated shared biological pathways, such as nucleosome assembly and neutrophil extracellular trap formation . PPI network analysis suggested that histone family members occupied central hubs in the interaction network. Spatiotemporal expression profiling identified region-specific developmental trajectories of pleiotropic genes in brain regions, including the cerebellum and thalamus, with the late fetal to early postnatal period potentially representing a sensitive developmental period for neurodevelopmental effects of BD susceptibility genes. This study establishes an atlas of brain IDPs associated with BD, by integrating epidemiological and genetic evidence. The findings highlight candidate IDPs, including right lateral orbital frontal cortex volume and fractional anisotropy in the forceps major, that were consistently associated with BD across complementary analyses. Our results provide genetically supported targets for early intervention and precision psychiatry.