Graph representations for imaging genetics: meso- and mixed-spatial scale modelling of brains and hearts

Graphs are particularly well-suited for modelling relationships between parts of a system. This includes organs such as the brain, where graph representations are ubiquitous and offer an expressive language to model spatial, structural, and functional relationship from data obtained in medical imaging. Graphs enable substantial compression and smoothing of imaging data, which helps build predictive models from high-dimensional volumetric time series such as found in functional magnetic resonance imaging (brain) or CINE imaging (heart).

In the introductory part of this talk, I will first briefly give an overview of the field of imaging genetics, a field at the intersection of medical image processing and bioinformatics. I will then discuss the notion of spatial scale in imaging genetics studies, illustrated by several results in brain and heart imaging. Further, I will illustrate how graphs can be computed from medical imaging data, and how they can be used in imaging genetics as an endophenotype representation and target for association studies, including genetic variation and transcriptomics. Finally, I will dicuss the development and use of graph neural networks and their latent spaces for multi-omics data analysis and heart imaging data analysis.