Depression classification method based on self-supervised learning and transfer learning

Abstract

A depression classification method based on self-supervised learning and transfer learning is characterized by tracking whole-brain white matter fiber bundles based on the diffusion tensor imaging data, constructing a whole-brain white matter fiber bundle network, and automatically learning the brain network node representation and the network representation irrelevant to downstream tasks through the comparative learning by using a self-supervised learning strategy; constructing a group network based on the brain network characterization and the non-image phenotype information, converting a depression classification problem into a network node classification problem, and classifying the depression patients and normal contrast by using a graph convolution neural network model based on spectrogram convolution. According to the method, the self-supervised learning and the transfer learning are utilized, the problem that depression samples are fewer is partially solved, the brain network level features related to depression are effectively mined, and the depression classification precision is improved.

Description

technical field [0001] This patent relates to the field of medical image processing and machine learning, especially a depression classification method based on self-supervised learning and transfer learning. Background technique [0002] Depression is a common mental illness worldwide, mainly manifested as low mood, loss of interest, impaired cognitive function, sleep and appetite disorders, and has become a major public health problem of global concern. Early diagnosis and treatment of depression are crucial to the recovery of depressed patients. The diagnosis of depression mainly relies on the clinical interview of the patient by a professional doctor to evaluate the patient's symptoms. Because the clinical symptoms of different mental diseases overlap, and depression is a heterogeneous disease, the clinical diagnosis of depression is very complicated and depends on the experience of doctors, which has a certain degree of subjectivity. Neuroimaging technology provides o...

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problems of insufficient feature mining capabilities of existing brain network analysis methods and fewer brain network labeled samples, inspired by comparative learning in the field of computer vision, the present invention proposes a depression classification method based on self-supervised learning and transfer learning. Multimodal brain network features can be extracted sufficiently quickly and efficiently for depression classification

Images