Human brain function relies on interactions between brain areas. Therefore, a complex network appears as a natural model for the brain, and network science indeed holds potential for understanding the brain. However, constructing a functional brain network is not trivial. I will address common problems in brain network construction, concentrating on a step that strongly affects obtained network properties: definition of network nodes. Using multiple fMRI datasets, I will show that the currently used node definition strategies produce nodes that suffer from low functional homogeneity, which may lead to spuriosity in observed network structure and data losses. Further, I will show that functional homogeneity changes in time, reflecting nodes&rsq
uo; roles in network topology. Finally, as a potential solution for node definition problem, I will introduce a flexible node definition strategy that accounts for the time-dependent nature of brain function.
Presential seminar, with parallel Zoom stream: https://zoom.us/j/98286706234?pwd=bm1JUFVYcTJkaVl1VU55L0FiWDRIUT09
