Title:
PREDICTION SYSTEM, METHOD AND APPARATUS FOR ABNORMAL BRAIN CONNECTIVITY, AND READABLE STORAGE MEDIUM
Document Type and Number:
WIPO Patent Application WO/2023/077603
Kind Code:
A1
Abstract:
Disclosed in the present invention are a prediction system, method and apparatus for abnormal brain connectivity, and a readable storage medium. High-order correlated features in different modals and high-order complementary features between different modals are automatically extracted by means of a deep learning method, and the analysis of abnormal connectivity of a multi-modal brain network and the prediction of different cognitive diseases are realized by means of an adversarial training method. By means of the present invention, the problem of it being impossible to accurately evaluate change rules of the structure, morphology and function connectivity of a brain by using an existing method is solved. In the present invention, representations with interpretability are learned by using a prior knowledge guide model, and the consistency of different modal representation distributions is constrained by means of paired collaborative discriminators; then, brain graph data is reconstructed from feature codes by means of a reverse generator and a decoder; and finally, inter-modal and intra-modal high-order correlated features are extracted by means of a hypergraph perception fusion module, and an adversarial loss function, a reconstruction loss function and a classification loss function are provided to guide model learning, such that the aim of mining abnormal brain connectivity of Alzheimer's disease is achieved.
Inventors:
WANG SHUQIANG (CN)
ZUO QIANKUN (CN)
SHEN YANYAN (CN)
ZUO QIANKUN (CN)
SHEN YANYAN (CN)
Application Number:
PCT/CN2021/135844
Publication Date:
May 11, 2023
Filing Date:
December 06, 2021
Export Citation:
Assignee:
SHENZHEN INST ADV TECH (CN)
SHENZHEN INST OF ADV TECH CAS (CN)
SHENZHEN INST OF ADV TECH CAS (CN)
International Classes:
G16H50/50; G06N3/04; G06N3/08; G16H50/20
Foreign References:
| CN113724880A | 2021-11-30 | |||
| CN110580695A | 2019-12-17 | |||
| CN112735570A | 2021-04-30 | |||
| CN113094566A | 2021-07-09 | |||
| CN112863664A | 2021-05-28 | |||
| AU2020102977A4 | 2020-12-24 |
Other References:
QIANKUN ZUO; BAIYING LEI; SHUQIANG WANG; YONG LIU; BINGCHUAN WANG; YANYAN SHEN: "A Prior Guided Adversarial Representation Learning and Hypergraph Perceptual Network for Predicting Abnormal Connections of Alzheimer's Disease", ARXIV.ORG, 12 October 2021 (2021-10-12), XP091078549
Attorney, Agent or Firm:
BEIJING ZHONG XUN TONG DA INTELLECTUAL PROPERTY AGENCY CO., LTD. (CN)
Download PDF: