【摘要】 目的探讨社区老年性痴呆(AD)筛选诊断的人工神经网络模型。方法通过现场流行病学整群抽样调查获取研究对象及有关信息;AD确诊根据中国精神疾病分类(CCMD-3)诊断标准,并参照国际疾病分类第10版(ICD-10)相关内容。采用原子吸收法检测研究对象全血中宏、微量元素含量,放射免疫分析法检测相关神经递质含量;数据库建立采用SPSS13.0软件,反向传播算法-人工神经网络模型(BP-ANN)的建立使用Clementine12.0软件。结果南昌、吉安、宜春三市6个社区实查≥60岁者4 350人,确诊AD患者214人,患病率为4.92%。从被研究者中抽取AD患者和非AD者各60名为建模对象,将其血中铁(Fe)、铜(Cu)、铁(Zn)、五羟色胺(5-HT)、多巴胺(DA)等17个变量作为网络的输入层,进行BP-ANN的拟合。所建模型的预测精度为75.00%,灵敏度为76.67%,特异度为83.33%;能够正确预测建模对象中88.33%的非AD者,78.79%的中型AD患者和47.37%的重型AD患者;输入变量敏感性系数排在前四位的依次为Al(0.156 3)、Cr(0.120 6)、5-HT(0.1090)和年龄(0.1010)。结论BP-ANN在社区老年性痴呆筛选(预测)中精度较高,具有一定的开发应用前景。更多还原

【Abstract】 Objective To explore an artificial neural network model for Alzheimer’s disease(AD) screening in the community.Methods The subjects and related information were obtained by Field Epidemiology cluster sample;AD was diagnosed by Chinese Classification of Mental Disorders(CCMD-3) diagnostic criteria and International Classification of Diseases 10th edition(ICD-10);the detection of the whole blood macro and trace element was tested by atomic absorption method,then the neurotransmitters were detected by radioimmunoassay;SPSS13.0 software was adopted to establish the database,the back-propagation algorithm-Artificial Neural Network(BP-ANN) was established by Clementine12.0 software.Results 4 350 people older than 60 years were investigated from 6 communities of Nanchang,Ji’an,Yichun,and 214 patients were diagnosed as AD,the prevalence was 4.92%.60 AD and 60 non AD cases were extracted as model objects to build the BP-ANN,the input layer of which were serum levels of Fe,Cu,Zn,5-HT,DA and other 17 variables.The model yielded the prediction accuracy of 75.00%,the sensibility and specificity was 76.67% and 88.33% respectively;the probabilities of judging non-AD,medium and heavy AD were 88.33%,78.79% and 47.37% respectively.According to the sensitivity coefficient,rowing in the top four were aluminum(Al,0.156 3),chromium(Cr,0.120 6),5-hydroxytryptamine(5-HT,0.109 0) and age(0.101 0).Conclusions BP-ANN has a high accuracy in AD screening(prediction) in the community,which has wide application prospect.更多还原