Heart-brain connections: Phenotypic and genetic insights from magnetic resonance images

Abstract

Cardiovascular health interacts with cognitive and mental health in complex ways, yet little is known about the phenotypic and genetic links of heart-brain systems. We quantified heart-brain connections using multiorgan magnetic resonance imaging (MRI) data from more than 40,000 subjects. Heart MRI traits displayed numerous association patterns with brain gray matter morphometry, white matter microstructure, and functional networks. We identified 80 associated genomic loci (P < 6.09 × 10^-10) for heart MRI traits, which shared genetic influences with cardiovascular and brain diseases. Genetic correlations were observed between heart MRI traits and brain-related traits and disorders. Mendelian randomization suggests that heart conditions may causally contribute to brain disorders. Our results advance a multiorgan perspective on human health by revealing heart-brain connections and shared genetic influences.

Fig. 1.. Overview of the study design and analyses.

(A) Overview of the study. We used CMR and brain MRI traits as endophenotypes to explore the phenotypic and genetic connections between the heart and the brain. (B) Description of the overall workflow and the key analyses involved in each step.

Fig. 2.. Phenotypic heart-brain associations.

(A) The −log10 (P value) of phenotypic correlations between 82 CMR traits and five groups of brain MRI traits, including 101 regional brain volumes, 63 cortical thickness traits, 110 DTI parameters, 92 resting fMRI traits, and 92 task fMRI traits. The dashed line indicates the Bonferroni significance level (P < 1.33 × 10⁻⁶). Each CMR trait category is labeled with a different color. (B) Significant correlations (P < 1.33× 10⁻⁶) between fractional anisotropy values of white matter tracts and AAo minimum area. (C) Significant correlations (P < 1.33 × 10⁻⁶) between mean diffusivity values of white matter tracts and global myocardial wall thickness at end diastole (global wall thickness).

Fig. 3.. Genetics of CMR traits in the UKB.

(A) SNP heritability of 82CMR traits across the six categories. The x axis displays the short names of CMR traits; see table S1 for the full names of these traits. The average heritability of each category is labeled. (B) Ideogram of 80 genomic regions associated with CMR traits (P < 6.09× 10⁻¹⁰). Red and brown name labels denote genomic regions that have been replicated in the validation dataset after applying Bonferroni correction and at a nominal level, respectively. (C) LVESV was associated with the 22q11.23 region in both the UKB (index variant rs5760061) and BBJ (index variant rs5760054) studies. (D) LVESV was associated with the 8q24.13 region in both the UKB and BBJ studies (shared index variant rs34866937).

Fig. 4.. Selected genetic loci associated with both CMR trait and other complex traits and diseases.

(A) In 6p21.2, we observed colocalization between the global myocardial wall thickness (WT) at end-diastole (WT global, index variant rs4151702) and atrial fibrillation (index variant rs3176326). The posterior probability of Bayesian colocalization analysis for the shared causal variant hypothesis (PPH4) is 0.997. In this region, the WT global was also in LD (r² ≥ 0.6) with ischemic stroke. (B) In 7p21.1, we observed colocalization between the DAo minimum area (DAo min area, index variant rs2107595) and systolic blood pressure (index variant rs57301765, PPH4 = 0.998). In this region, the DAo min area was also in LD with stroke, intracranial aneurysm, coronary artery disease, and moyamoya disease. (C) In 15q25.2, we observed colocalization between the regional myocardial wall thickness at end-diastole (WT AHA 7, index variant rs11638445) and schizophrenia (index variant rs12902973, PPH4 = 0.922). In this region, the WT AHA 7 was also in LD with bipolar disorder. AHA 7, American Heart Association (AHA) region 7. (D) We illustrated the colocalization between the AAo maximum area (AAo max area) and functional connectivity between the default mode and orbito-affective networks (shared index variant rs1678983) in 15q21.1 (PPH4 = 0.964).

Fig. 5.. Genetic correlations between CMR traits and other complex traits and diseases.

(A) We illustrated selected genetic correlations between CMR traits (x axis) and complex traits and diseases (y axis). The asterisks highlight genetic correlations that have passed multiple testing adjustments using the Benjamini-Hochberg procedure to control the FDR at the 5% level. (B to E) Illustration of CMR traits that exhibited genetic correlations with stroke (B), schizophrenia (C), anorexia nervosa (D), and cognitive function (E).

Fig. 6.. Genetic causal effects of CMR traits on psychiatric disorders.

We illustrated selected significant (P < 1.68 × 10⁻⁴) causal genetic links from CMR traits (exposure) to psychiatric disorders (outcome) after adjusting for multiple testing using the Benjamini-Hochberg procedure to control the FDR at the 5% level. Category, the category of CMR traits; #IVs, the number of genetic variants used as instrumental variables. Different Mendelian randomization methods and their regression coefficients are labeled with different colors. See table S14 for data resources of psychiatric disorders.