Research On Brain Network Based On The Weighted K-order Propagation Number Algorithm

The brain's work often requires cooperation between various areas to complete complex cognitive tasks by forming brain Network.The topological structure and the importance of nodes of the brain Network are the research emphases.The analysis of the brain network topology is helpful to analyze the changes of brain network topology under different states,and the node importance contributes to the localization of disease and the recognition of brain functional areas.For the purpose of overcoming the shortcomings of traditional node importance methods,this study proposes an evaluating algorithm for the importance of the weighted network nodes,namely Weighted K-Order Propagation Number Algorithm.This method is abstracted from the spread of disease,combining the local and global characteristics of the network,and finally the importance of nodes can be obtained.To verify whether the Weighted K-Order Propagation Number Algorithm can effectively evaluate the importance of network nodes,this study conducts simulation experiments on the generated typical networks and public networks with true physical significance.This research will evaluate the node importance of the generated 10-node symmetric network.Since the algorithm presented in this research can well recognize the importance of bridge nodes,the obtained node importance will be more reasonable.Meanwhile,based on the method of intentional node attack,the Facebook forum network,the US 500 busiest commercial airports network,the non-US airport routing network and the Science Museum visitor network are analyzed.The algorithm in this paper only need remove a small number of nodes to achieve complete damage to the network structure.After the validity verification of the Weighted K-Order Propagation Number Algorithm,this research generates a kind of weighted brain network based on the phase-locking value for the positive and negative emotional EEG and analyzes the network topology and the differences of node importance.By studying node importance,it is found that there are obvious differences in the topology and node importance of the brain network in both positive and negative emotions under each rhythm.The network topology under Beta rhythm is far cry from that under Delta rhythm.In both Alpha and Theta,the cerebral limbic nodes have the characteristic of importance reversal.Finally,this paper combines the network topology and node importance features to design an emotion recognition algorithm model of multimodal fusion.The experimental results indicate that the multimodal fusion algorithm based on network topology and node importance is in possession of obvious advantages in recognition and accuracy compared with single modal.It is worth noting that the Weighted K-Order Propagation Number Algorithm proposed in this paper not only can be used in brain network,but also can explore the importance of other weighted networks.