Abstract: The effects of both regional topological structures and anatomical structures on human brain functional networks modeling were investigated， and generative models of human brain functional networks based on local communities were proposed. The local community topologies of the models are measured by not only the common neighbors between the two functional regions but also the connections among the neighbors. And the anatomical structures are represented by the anatomical distance between the two brain regions. In addition， the similarity energy index was proposed to evaluate the similarity between the generated network and the real data network based on functional magnetic resonance images(fMRI). The results show that the generative models based on local communities provide a good fit to the real data network in terms of network efficiency， clustering coefficient， modularity and degree distribution compared with traditional generated models.

Key words: brain functional network, generative model, local community, anatomical distance, network similarity