Keywords: Generative Modelling, GAN, GCN, Graph Convolutional Networks, Cortical Surface, MRI
TL;DR: We develop a GCN that can perform continuous age progression/regression of the neonatal cortical surface (myelination and sulcal depth maps). Generated images retain subject identity, and are accurate to the target age by 1.02 weeks
Abstract: Structural MRI-derived features of the cortical surface are known to correlate to phenotypes such as age, sex and cognitive outcomes. Deep generative modelling of cortical neurodevelopment can lead to clinically interpretable models of disease or identify atypical cases for clinical intervention, but deep modelling of non-Euclidean domains, such as surfaces, poses additional challenges. In this work, we adapt a graph convolutional network (GCN) to model the neonatal cortical surface, and synthesise realistic, age-conditioned images of myelination and sulcal depth cortical surface maps. We train our models without longitudinal data, using randomised aging cycles of varying length, which we validate by ablation and with comparison to a CycleGAN. An independently trained deep regression model evaluates the accuracy of the generated images as the difference between their apparent post-menstrual age (PMA) and their respective target ages, obtaining a mean absolute error (MAE) of 1.02 ± 0.28 weeks (baseline accuracy 0.6 weeks).