世界科学技术-中医药现代化2025，Vol.27Issue(11)：3119-3125,7.DOI:10.11842/wst.20250720004
基于度中心性对亚健康脾气虚证与肾气虚证患者的机器学习分类研究
Research on Machine Learning Classification of Sub-health Patients with Spleen Qi Deficiency Syndrome and Kidney Qi Deficiency Syndrome Based on Degree Centrality
摘要
Abstract
Objective To build the machine learning classification model for sub-health patients with spleen qi deficiency syndrome and kidney qi deficiency syndrome based on degree centrality characteristics.Methods A total of 80 sub-healthy subjects were included,including 40 subhealth patients with spleen qi de ficiency syndrome and 40 subhealth patients with kidney qi deficiency syndrome were enrolled and underwent resting-state functional magnetic resonance imaging scans.The DPABI software was used to extract the degree centrality characteristics of all subjects,and a random forest model was used for classification.Results Between the sub-health spleen qi deficiency syndrome group and the sub-health kidney qi deficiency syndrome group,after feature selection,10 degree centrality features were finally obtained:The supplementary eye field,posterior cingulate gyrus,MT+area,pre-supramarginal sulcus,ventromedial visual area,secondary visual cortex,supramarginal sulcus,precuneus,ventral supramarginal gyrus complex and dorsolateral prefrontal cortex.Following hyperparameter optimization and leave-one-out cross-validation,a random forest classification model was obtained with an accuracy rate of 0.71.Conclusion The significant changes in the centrality of brain regions such as the supplementary eye field and posterior cingulate gyrus may be key brain regions underlying the neural mechanism differences between subhealth spleen qi deficiency syndrome and kidney qi deficiency syndrome,providing neuroimaging evidence for the cognitive neural basis differences in traditional Chinese medicine theories of"spleen in storing idea"and"kidney storing will".关键词
脾气虚证/肾气虚证/亚健康/度中心性/机器学习Key words
Spleen qi de ficiency syndrome/Kidney qi deficiency syndrome/Sub-health/Degree centrality/Machine learning分类
医药卫生引用本文复制引用
冯思同,吴子遥,董麟瑞,宁艳哲,贾竑晓..基于度中心性对亚健康脾气虚证与肾气虚证患者的机器学习分类研究[J].世界科学技术-中医药现代化,2025,27(11):3119-3125,7.基金项目
国家自然科学基金委员会面上项目(82174311):基于"脾藏意"的脾气虚证所致工作记忆异常的神经机制研究,负责人:贾竑晓 （82174311）
国家自然科学基金委员会面上项目(81873398):基于"肾藏志"理论的肾虚证所致注意网络异常及神经机制研究,负责人:贾竑晓 （81873398）
深圳市'医疗卫生三名工程'项目资助(SZZYSM202411001):首都医科大学附属北京安定医院贾竑晓教授中医睡眠情志病诊疗团队,负责人:贾竑晓. （SZZYSM202411001）
