基于rs-fMRI动态特征融合的阿尔茨海默病诊断

打开文本图片集

中图分类号：TP391.4 文献标识码：A 文章编号：2096-4706（2025）16-0076-06

Alzheimer's Disease Diagnosis Based on Dynamic Feature Fusion of rs-fMRl

LINKai，WANGi （School of Informationand Artificial Intelligence，Anhui BusinessCollege，Wuhu 241oo2，China）

Abstract：ThedynamicFunctional Connectivity（dFC）networksbasedonresting-state functional MagneticResonance Imaging （rs-fMRI） provide an important approach for decipheringthe pathological mechanisms of Alzheimer's Disease （AD） and Mild Cognitive Impairment （MCI）.Aiming atthe problems of insuffcient modelingof interactionrelationships between continuous time-series features andthe lack ofmulti-scale spatio-temporal feature fusion mechanisms in existing Dp Learning methods，a DeepLearning framework basedonDynamicFeature Fusion （DFF）is proposed fortheautomatic diagnosisof braindiseasesusingrs-fMRIdata.Experimentalvalidationbasedonrs-fMRIdataof174subjectsfromtheAlzheimer'sDisease Neuroimaging Initiativedatabase shows thatthe proposed method demonstrates emarkable diagnostic performance n both binary and multi-classification tasks.

KeyWords：dynamic Functional Connectivity;dynamic feature fusion;Alzheimer's disease;rs-fMRI;braindisease classification

0 引言

阿尔茨海默病（Alzheimer'sDisease，AD）以进行性认知障碍为特征，是老年人常见的神经退行性疾病，也是导致痴呆最常见的病因[1]。（剩余9288字）