Psychosis automatic discrimination method based on multi-level feature fusion of functional connection networks

Abstract

The invention proposes a psychosis automatic discrimination method based on multi-level feature fusion of functional connection networks, and the method comprises the steps: constructing the functional connection network by using resting-state functional nuclear magnetic (Rs-fMRI), calculating features of two levels: network attribute features and functional connection features, wherein the network attribute features include six network local attributes and six network global attributes; stacking all functional connection networks all functions to calculate an average network, reserving a certain proportion of edges, and taking the correlation of the reserved positions as the features of the connection hierarchy; simplifying the features of two levels through the group Lasso with the consideration to the independence of brain regions and the correlation between features, and respectively constructing a support vector machine (SVM) classifier, and obtaining a final classification resultin a weighted voting mode. The method realizes automatic discriminant analysis of whether or not suffering from mental illness, and improves the accuracy of diagnosis of psychosis, and the method canbe applied to actual clinical diagnosis.

Description

technical field [0001] The invention relates to an automatic identification method for mental illness based on multi-level feature fusion of a functional connection network, and belongs to the technical field of disease classification and diagnosis. Background technique [0002] Schizophrenia is the most common severe mental illness, clinically manifested as a syndrome with various symptoms, involving obstacles in perception, thinking, emotion and behavior, as well as incoordination of mental activities. The diagnosis of schizophrenia in traditional medicine is mostly based on American DSM-IV, international ICD-10 and domestic mental disorder classification and diagnostic criteria. With the rapid development of medical imaging technology, designing an objective and convenient automatic discrimination and classification method for patients with schizophrenia will have a good application prospect in the differential diagnosis of psychosis. Functional Magnetic Resonance Imagin...

AI Technical Summary

Problems solved by technology

[0003] The use of functional magnetic resonance imaging for psychiatric diagnosis is a relatively objective and effective method. The common functional magnetic resonance imaging constructs a functional connection network to extract network features, but does not consider the commonality between the network and the average network. When extracting network attributes, it does not consider Grouping to Network Attributes

Images