查看文章

Comprehending the interplay between spatial and temporal characteristics of neural dynamics can improve our understanding of information processing in the human brain. Graph neural networks provide a novel possibility to interpret graph-structured signals as typically observed in complex brain networks. This thesis presents an application of spatio-temporal graph neural networks to model functional dynamics observed in magnetic resoance imaging data. It is shown that graph neural network models are able to scale to large brain networks, and can help us to derive directed functional dependecies based on the structural brain network.