基于人工神经网络优化脑脉通治疗缺血性脑中风组分配伍研究  

Optimization of Component Compatibility of Naomaitong for Anti-cerebral Ischemia Treatment Based on Artificial Neural Network

作　　者：吴纯伟 郭嘉雯[1] 陈超[1,2,3] 梁生旺[1,2,3] 王淑美[1,2,3] 

机构地区：[1]广东药学院中药学院,广州510006 [2]广东高校中药质量工程技术研究中心,广州510006 [3]国家中医药管理局中药数字化质量评价技术重点研究室,广州510006

出　　处：《中国药学杂志》2016年第6期454-458,共5页Chinese Pharmaceutical Journal

基　　金：国家自然科学基金资助项目(81274060;81274059;81473413)

摘　　要：目的采用人工神经网络对脑脉通有效部位进行组分配伍优化。方法采用均匀设计法将脑脉通中的五大有效部位10个不同水平的配比组合进行药效考察,除假手术组外其他各组制备大鼠大脑中动脉闭塞(MCAO)模型,以脑梗死面积和坏死区体密度作为药效指标,将获得的实验结果作为训练集,建立人工神经网络药效预测模型。结果基于建立的人工神经网络模型进行预测可知脑脉通组分最佳配伍为:大黄总蒽醌180 mg·kg^(-1),人参总皂苷70 mg·kg^(-1),葛根总黄酮450 mg·kg^(-1),川芎总酚酸27 mg·kg^(-1),川芎油110μL·kg^(-1)。并经实验验证优化的脑脉通组分配伍对各药效指标改善作用显著。结论脑脉通各组分配比均能不同程度的改善缺血性脑中风的症状,基于均匀设计结合人工神经网络优化组分配伍的方法是可行的。OBJECTIVE To optimize the component compatibility of Naomaitong for anti-cerebral ischemia treatment based on artificial neural network. METHODS The five effective parts in Naomaitong were divided into ten groups by uniform design. Except the rats in the sham groups, all rats were subject to right middle cerebral artery occlusion （MCAO） with the suture-occluded method by Longa. The effect of Naomaitong was evaluated based on the bulk density of necrotic zone and infarction area percentage. The artificial neural network model was established for pharmacodynamic prediction. RESULTS The model established in this study could predict the actions of different drug combinations. The best effect was obtained by the following formula： total anthraquinones of rhubarb （TAR） 180 mg·kg^-1, total saponins of Ginseng （TSG） 70 mg·kg^-1, total flavonoids of Pueraria （TFP） 450 mg·kg^-1, total phenolic acid of Ligustieum walliehii （TPLW） 27 mg·kg^-1, rhizome oil （RO） 110 μL·kg^-1. CONCLUSION Different eompatibili- ties of Naomaitong are all effective for the treatment of isehemia in rats. Uniform design combined with ANN can be a more efficient method to realize dose optimization for Naomaitong prescription.

关 键 词：脑脉通 组分配伍优化 缺血性脑中风 均匀设计 人工神经网络 

分 类 号：R284[医药卫生—中药学]