首页|期刊导航|数字中医药(英文)|中医精-神-寐理论的数字化框架及其在哺乳动物寿命建模中的应用

中医精-神-寐理论的数字化框架及其在哺乳动物寿命建模中的应用OA

A theory-constrained digital framework of the traditional Chinese medicine Jing-Shen-Mei theory for mammalian lifespan modeling

中文摘要英文摘要

目的 构建传统中医精-神-寐的理论约束数学表达框架,并评估预设有理闭式模型在跨物种哺乳动物寿命预测中的适用性.方法 将精、神、寐分别操作化为有限生命预算储备、有效整合能力和睡眠介导的日常负担消减机制.整合 Human Ageing Genomic Resources(HAGR)中的 AnAge 数据库、波士顿大学哺乳动物睡眠数据库以及比较神经解剖学整理数据,构建包含最大寿命、睡眠时长和皮层/背侧皮层神经元数量 3个变量的完整跨物种数据集.主要分析集共纳入 53 个物种,其中 48个物种采用直接测量或整理的神经元数据,另有 5个物种采用明确标注的脑重外推神经元数据.模型参数 α=2/5、β=1/4和 η=1/5在经验分析前预先设定,仅对全局系数 C在训练集中进行估计.采用 100 次重复五折交叉验证评价模型性能,并结合 48个直接测量物种子集分析和留一交叉验证(LOOCV)进行稳健性检验.同时,以自由幂律模型和代表性机器学习模型作为对照,采用均方根误差(RMSE)、平均绝对误差(MAE)和决定系数(R2)进行评价,并通过 729组参数网格搜索分析参数可辨识性.结果 在 53个物种数据集上,预设有理闭式模型交叉验证结果为 RMSE=13.73年、MAE=9.84年、R2=0.660.在相同交叉验证条件下,与代表性机器学习模型及代数模型相比,该模型在保持良好可解释性的同时表现出具有竞争力的预测性能.剔除 5个脑重外推神经元数据后,其性能仍保持稳健:48个直接测量物种重复五折交叉验证结果为 RMSE=14.27 年、MAE=10.14年、R2=0.647;LOOCV结果为 RMSE=14.16 年、MAE=10.08 年、R2=0.652.参数网格搜索得到的最低RMSE为 13.03年,而预设有理闭式模型 RMSE 为 13.73年;共有 11、27 和 69 组参数组合分别落在最优结果 1%、5%和 10%的误差范围内,表明该有理闭式模型处于近最优参数区域,但并非唯一最优解.结论 本研究初步实现了传统中医精-神-寐理论的计算化形式表达,并表明在稀疏跨物种数据条件下,仅包含一个拟合系数的预设有理闭式模型仍具有良好的经验预测能力.

Objective To construct a theory-constrained mathematical framework for the traditional Chi-nese medicine(TCM)Jing(精,essence)-Shen(神,spirit)-Mei(寐,sleep)theory and to evalu-ate the applicability of a pre-specified rational closure to cross-species mammalian lifespan prediction. Methods Jing,Shen,and Mei were operationalized as a finite life-budget reservoir,effective homeostatic capacity,and sleep-mediated daily burden reduction,respectively.A harmo-nized complete-case dataset was constructed by integrating maximum lifespan,sleep dura-tion,and cortical/pallial neuronal abundance from three sources:the AnAge database within the Human Ageing Genomic Resources(HAGR),the Boston University mammalian sleep records,and curated comparative neuroanatomical compilations.The primary dataset in-cluded 53 species,comprising 48 directly measured or curated neuronal records and 5 explic-itly labelled records derived from brain-mass-estimated.The rational exponents(α=2/5,β=1/4,and η=1/5)were pre-specified prior to any empirical fitting,leaving only the global coef-ficient C to be estimated within training folds.Model performance was evaluated using re-peated five-fold cross-validation with 100 repeats,alongside sensitivity analyses on the 48-species direct-measurement subset and leave-one-out cross-validation(LOOCV).Compara-tors included an unconstrained power-law model and representative machine learning(ML)regressors.Root mean square error(RMSE),mean absolute error(MAE),and coefficient of determination(R2)served as primary metrics,and a 729-combination parameter-grid analy-sis was conducted to assess parameter identifiability. Results On the 53-species curated dataset,the theory-selected rational closure achieved cross-validated RMSE=13.73 years,MAE=9.84 years,and R2=0.660.Compared with repre-sentative ML and algebraic benchmark models under identical out-of-fold evaluation,the ra-tional closure showed competitive predictive performance while maintaining a lower-dimen-sional interpretable structure.Performance remained robust after excluding the five brain-mass-estimated neuronal records:repeated five-fold cross-validation on the 48 directly mea-sured species achieved RMSE=14.27 years,MAE=10.14 years,and R2=0.647,and LOOCV achieved RMSE=14.16 years,MAE=10.08 years,and R2=0.652.The grid-search analysis identified a minimum RMSE of 13.03 years,while the pre-specified rational closure delivered an RMSE of 13.73 years.Notably,11,27,and 69 parameter combinations fell within 1%,5%,and 10%of the minimum,respectively,indicating that the rational closure is near-optimal yet non-unique. Conclusion This study provides a preliminary,computationally grounded formalization of the TCM Jing-Shen-Mei framework and demonstrates that a pre-specified,one-coefficient ra-tional closure can maintain empirical utility in a sparse-data,cross-species predictive proto-col.

陈俊名;金霞

桂林航天工业学院机电工程学院,广西 桂林 541004,中国桂林航天工业学院机电工程学院,广西 桂林 541004,中国

数字中医精-神-寐框架哺乳动物寿命理论选择有理闭包整合能力稀疏数据协议机器学习基准测试

Digitalization of traditional Chinese medicineJing-Shen-Mei frameworkMammalian lifespanTheory-selected rational closureIntegrative capacitySparse-data protocolMachine learning benchmarking

《数字中医药(英文)》 2026 (3)

331-343,13

Scientific Research Start-up Fund under the Second Batch of the Guangxi Young Talent Universal Support Program(No.325,2025).

10.1016/j.dcmed.2026.08.003

评论