AI Summary: This paper introduces a Graph Variational Autoencoder that can reconstruct brain connectivity patterns while 'conditioning' them on specific traits like sex or intelligence. Using a massive dataset from the UK Biobank, the model identifies specific network axes that correlate with cognitive performance.
Uncovering the network mechanisms of cognitive assessments requires generative frameworks that model brain connectivity in conjunction with demographic variables. We present a conditional graph variational autoencoder (cGVAE) for modeling human brain functional connectivity features using over 20,000 subjects from the UK Biobank. Our model demonstrates high-fidelity reconstructions that preserve condition-specific network patterns associated with biological sex and fluid intelligence. By probing the latent dimensions, we identify interpretable connectivity patterns that act as functional signatures for these variables. This scalable framework offers potential applications in characterizing individual differences and functional signatures for mental health conditions.
Share your opinion to help other learners triage faster.
Write a reviewInvite someone by email to share an invited review for Conditioned Graph Reconstruction of Brain Functional Network Connectivity Reveals Interpretable Latent Axes of Sex and Fluid Intelligence.