Brain cognitive state recognition method based on network entropy

Abstract

The invention relates to a brain cognitive state recognition method based on network entropy. The method comprises the following steps: S1, constructing a complex network based on an electroencephalogram time sequence; s2, the connectivity of the complex network is judged, and the network shortest distance of any child node in the complex network is calculated; s3, performing mathematical statistics on the calculation result of the network shortest distance of any child node; s4, calculating a network entropy parameter based on a multi-scale network space distance according to a mathematical statistics result; and S5, inputting the network entropy parameter as a feature vector into a mode classifier, and performing mode classification of the cognitive state of the brain complex network. According to the method, a new network entropy parameter is constructed, the network entropy parameter can effectively represent various network states, and on the basis of the network entropy parameter, the cognitive activity state of the human brain can be effectively recognized.

Description

technical field [0001] The invention relates to the field of complex network data analysis, in particular to a brain cognitive state recognition method based on network entropy. Background technique [0002] A network is a system of specific entities and the interactions between them. Most complex physical systems can be expressed as a complex network. In the network obtained by modeling complex physical systems, nodes (vertices) present complex interactive relationships through edges (connections or arcs). A complex network can be described as a graph G={V, E, W}, where V represents a set of vertices or nodes, E represents a set of edges, W is a weight matrix, and W is mainly used to describe some attributes related to edges. [0003] In recent years, with the rapid development of complex network theory and information science, network entropy has been increasingly used to understand network properties and describe different types of networks. In information theory, entr...

AI Technical Summary

Problems solved by technology

[0005] Traditional network entropy parameters have limited feature representation capabilities when calculating and identifying the cognitive state of the human brain. Usually, the network entropy parameters are only based on the establishment of a local or global network structure characteristic. At the same time, due to the human brain The cognitive state model is complex and dynamically changing, and calculation and identification are difficult. The existing network entropy parameters are computationally intensive and cannot meet real-time requirements, and the recognition accuracy is also low.

Images