Dynamic Functional Segregation and Integration in Human Brain Network During Complex Tasks.

The analysis of the topology and organization of brain networks is known to greatly benefit from network measures in graph theory. However, to evaluate dynamic changes of brain functional connectivity, more sophisticated quantitative metrics characterizing temporal evolution of brain topological features are required. To simplify conversion of time-varying brain connectivity to a static graph representation is straightforward but the procedure loses temporal information that could be critical in understanding the brain functions. To extend the understandings of functional segregation and integration to a dynamic fashion, we recommend dynamic graph metrics to characterise temporal changes of topological features of brain networks. This study investigated functional segregation and integration of brain networks over time by dynamic graph metrics derived from EEG signals during an experimental protocol: performance of complex flight simulation tasks with multiple levels of difficulty. We modelled time-varying brain functional connectivity as multi-layer networks, in which each layer models brain connectivity at time window t + Δt. Dynamic graph metrics were calculated to quantify temporal and topological properties of the network. Results show that brain networks under the performance of complex tasks reveal a dynamic small-world architecture with a number of frequently connected nodes or hubs, which supports the balance of information segregation and integration in brain over time. The results also show that greater cognitive workloads caused by more difficult tasks induced a more globally efficient but less clustered dynamic small-world functional network. Our study illustrates that task-related changes of functional brain network segregation and integration can be characterized by dynamic graph metrics.