Optimizing Network Structure in a Model of Human Aging

Our network model of aging represents aspects of human health as nodes in a network. The nodes damage stochastically over time based on the health of their neighbours. Mortality occurs stochastically with a rate proportional to the overall damage in the network. Previous work has shown that human health and mortality data can be captured using a scale-free network and that the model is very sensitive to the choice of network structure. Does the network structure represent an organism’s robustness to damage in a meaningful way? We address this question from the bottom up using a network optimization approach. We optimize the network structure with respect to various health outcomes, such as longevity and health-related quality of life. We investigate which network motifs emerge depending on the optimization health outcome, and how these aspects of the network structure affect damage propagation. We add measures of network entropy to penalize “trivial” network structures.