Surface-volume self-supervised representation learning of brain MRI for genetic discovery

2026-10-01

AI Summary

Existing brain imaging GWAS phenotypes typically rely on single modalities such as volume or surface area, limiting their capacity to comprehensively characterize the genetic architecture of brain anatomy. This work proposes MEVA, a framework that introduces a self-supervised autoencoder integrating MRI voxel intensities with 3D mesh geometric features, including cortical curvature and thickness. By fusing these multimodal data into a unified representation space for joint learning, MEVA overcomes the inherent limitations of single-modality phenotyping. In UK Biobank GWAS experiments, the proposed framework identifies a greater number of genome-wide significant loci compared to conventional approaches. Furthermore, it achieves superior predictive performance for age and sex relative to single-modality baselines. These findings demonstrate that MEVA effectively enhances the analytical power of brain imaging genetics association studies by leveraging complementary structural information across modalities.

Abstract

Existing genome-wide association studies (GWAS) of brain imaging provide predefined or deep-learning-derived imaging phenotypes, yet these phenotypes come from either volumetric scans or cortical surface meshes, so each captures only part of the heritable variation in brain anatomy. Here we introduce MEVA (Mesh-Enhanced Volumetric Autoencoder), a self-supervised framework that encodes voxel-level image intensity together with cortical mesh geometry, including curvature and cortical thickness at each surface vertex, into one shared set of imaging features. Combining the mesh and volumetric inputs in MEVA yields modest performance gains in age and sex prediction over models that use either input alone. When these features serve as phenotypes for GWAS in the UK Biobank, they reveal more genome-wide significant loci than features learned from volumes alone or from meshes alone. These results suggest that adding cortical surface geometry to volumetric self-supervised learning captures additional heritable variation and so increases the number of loci detected.

Problem

• GWAS
• brain MRI
• imaging phenotypes
• heritable variation
• self-supervised representation learning

Innovation

• Self-supervised learning
• Mesh-Enhanced Volumetric Autoencoder
• Brain MRI
• Genome-wide association study
• Surface-volume representation