Brain network classification method combining node attributes and multi-level topology

Abstract

The invention discloses a brain network classification method combining node attributes and multi-level topology, and the method comprises the following steps: S1, obtaining functional magnetic resonance brain image data, and carrying out the preprocessing; S2, based on the preprocessed data, generating a whole-brain network function connection matrix by using an automatic anatomical marking template, and constructing an unbiased brain network by using a Kruskal algorithm and taking DMN as a region of interest; S3, extracting brain region node betweenness from the unbiased brain network to serve as local attribute features, and extracting brain region features with inter-group significant differences by using a double-sample t test method; S4, extracting multi-level topological features on the unpartial brain network by using sub-network kernels, generating a sub-network kernel matrix, and extracting optimal topological features by using a kernel principal component analysis method. According to the method, the classification performance is remarkably improved, an abnormal brain area can be found, multi-level topological characteristics of brain area nodes can be captured, and the method has important significance in clinical auxiliary diagnosis of schizophrenia.

Description

technical field [0001] The invention belongs to the technical field, and in particular relates to a brain network classification method combining node attributes and multi-level topology. Background technique [0002] Schizophrenia (SchiZophrenia, SZ) is a severe chronic brain-damaged mental disease. Brain network classification has become a research hotspot among scholars in the fields of brain science research and brain disease diagnosis. The classification of brain networks in SZ can effectively improve the diagnostic accuracy of patients, which is of great significance for medical auxiliary diagnosis. [0003] There are many classification methods for the brain network of SZ. The traditional classification method is to extract a single local attribute feature from the brain network, such as betweenness, characteristic path length, clustering coefficient, etc., to form a long feature vector, and to train Support Vector Machines (SVM). , SVM) classifier for classificatio...

AI Technical Summary

Problems solved by technology

[0004] The above-mentioned measurement methods are all defined according to the theory of conventional graphs, ignoring that each node in the brain network represents a specific brain area, and each node in the brain network is unique and fixed, which cannot be fully realized. Considering the multi-level topological characteristics of nodes in brain networks limits the further improvement of classification performance

Images