Exploring function-structure covariant patterns in Alzheimer's disease and mild cognitive impairment based on multimodal magnetic resonance imaging
10.3969/j.issn.1005-202X.2025.10.006
- VernacularTitle:基于多模态磁共振成像探讨阿尔茨海默症和轻度认知障碍的功能-结构共变模式
- Author:
Yifan SHEN
1
;
Ruipeng NING
;
Renren LI
;
Chenxi PAN
;
Wei ZHANG
;
Zheyu LI
;
Zhihao XU
;
Qiurong YU
;
Dazhi YIN
;
Yunxia LI
;
Mingxia FAN
Author Information
1. 华东师范大学物理与电子科学学院上海市磁共振重点实验室,上海 200062;华东师范大学医学磁共振与分子影像技术研究院,上海 200062
- Publication Type:Journal Article
- Keywords:
Alzheimer's disease;
mild cognitive impairment;
covariant pattern;
multimodal fusion;
magnetic resonance imaging
- From:
Chinese Journal of Medical Physics
2025;42(10):1298-1305
- CountryChina
- Language:Chinese
-
Abstract:
Objective To explore function-structure covariant patterns in Alzheimer's disease(AD)and mild cognitive impairment(MCI),and to investigate their associations with cognitive function and activities of daily living.Methods three-way parallel group independent component analysis(three-way pGICA),was used to identify the covariant patterns of resting-state functional MRI temporal data,gray matter density maps,and fractional anisotropy(FA)maps,and the differences between different groups were compared.Furthermore,the associations of covariant patterns with the Montreal Cognitive Assessment-Basic(MoCA_B)Scale scores and Activities of Daily Living Scale scores were analyzed.Results The function-structure covariant patterns in AD and MCI were characterized by the enhanced negative functional connectivity between the left posterior salience network and the right default mode network,the decreased gray matter density in the bilateral dorsolateral prefrontal cortex,and the reduced FA values in the left superior corona radiata(correlations:P<0.001,FDR corrected).Compared with HC group,AD group showed significant abnormalities in all identified covariant patterns(P<0.01,FDR corrected),but MCI group only exhibited a significant decrease in gray matter density in the bilateral dorsolateral prefrontal cortex(P<0.05,FDR corrected).Additionally,AD group had significantly lower FA value in the left superior corona radiata than MCI group(P<0.05,FDR corrected).The loadings reflecting the degree of covariation were significantly correlated with the Activities of Daily Living Scale scores(P<0.05,FDR corrected)but not with MoCA_B Scale scores.Conclusion The function-structure covariant patterns in AD and MCI are consistent with the declines in activities of daily living.The multimodal fusion analysis(three-way pGICA)provides a novel approach to understand the brain damage mechanisms underlying the covariant evolution of MCI and AD.