Graph models of brain diseases

Recent advances in network modeling of brain phenomena have opened up many opportunities for graph theorists and image analysis researchers to solve important clinical and scientific problems in neuroscience and neurology. Although initial studies were limited to applying conventional “social network” techniques to the brain connectomes, new studies are introducing bottom-up graph models specifically derived from the neural context of brain structure and function. In this paper we review the models of graph spread and explore their application in capturing the incidence, spread and prognostication of neurodegenerative diseases. We show that graph models can be used not only to characterize disease but also to predict future disease states from baseline imaging data.