High-order Brain Network Analysis of Depression Based on Dynamic Functional Connectivity

The realization process of brain cognitive and emotional functions involves the dynamic integration and coordination across multiple time scales. Simultaneously, anomalous topological structures in higher-order brain network have been founded as important characteristics for depression identification. To capture the high-order dynamic temporal characteristics of the brain for depression group across time scales, this study proposed a method for analyzing the high-order brain network in depression based on dynamic functional connectivity. Firstly, a dynamic functional connectivity matrix was constructed to obtain the temporal sequences of functional connectivity. Secondly, a hyper-network based on LASSO method was constructed using functional connectivity as nodes to explore the topological structure of the brain functional network. The results revealed the presence of anomalous long-range connections between brain regions in the depression group, particularly concentrated within the Dorsal Attention Network(DAN) and the Default Mode Network(DMN). These connections were mainly distributed in the frontal, temporal, and parietal regions. In terms of network metrics analysis, the clustering coefficient of the depression group was significantly lower than that of the normal group, especially in hyper-network clustering coefficients 2(HCC2). This suggested a weakened information connection between brain regions in the depression group. Furthermore, the experiment also found that the selection of hyper-network parameters had a significant impact on the network structure. In summary, this study demonstrated the feasibility of exploring the anomalous topological structure of the brain using the concept of the high-order network dynamics reconstruction for depression patients, and provided new insights into the construction of brain functional networks.