CALM: Interpretable Cross-Modal Alignment for Biomarker Discovery from Unpaired Data

2026-07-01

AI Summary

This study addresses the challenge of integrating completely unpaired and unimodal neuroimaging and genomic data by proposing a linear projection–based cross-modal alignment method. The approach jointly aligns class-conditional distributions while preserving inter-group separability within a shared latent space, enabling—for the first time—the discovery of stable and interpretable associations between brain regions and gene pathways from non-overlapping samples. Applied to autism spectrum disorder research, the method successfully identifies immune- and metabolism-related gene pathways linked to specific cortical areas. These findings not only significantly outperform those of current state-of-the-art methods but also demonstrate strong generalizability and biological interpretability.

Abstract

The interaction between brain structure and genetic influences is key to understanding neuropsychiatric disorders. However, most large-scale datasets are unimodal, providing either neuroimaging or genetics data. We propose CALM, a framework that learns interpretable associations between brain ROIs and genetic pathways from completely disjoint populations. CALM aligns the two modalities in a shared latent space via linear projections that simultaneously match the class-conditional latent distributions and ensure group separability. These projections provide interpretable pathway--ROI associations. When trained on unimodal imaging and genetics datasets, CALM generalizes to an unseen paired dataset, outperforming several state-of-the-art methods and ablation baselines. We also demonstrate stability of the learned associations against a paired baseline. Our experiments on autism spectrum disorder reveal immune and metabolic pathways linked to specific cortical regions and are consistent with established literature. Thus, CALM opens the door to leveraging large unimodal repositories for studying cross-modal interactions in brain disorders across disparate datasets.

Problem

• cross-modal alignment
• biomarker discovery
• unpaired data
• neuroimaging
• genetics

Innovation

• cross-modal alignment
• unpaired data
• interpretable associations
• latent space projection
• biomarker discovery