基于前额联络皮层脑电信号的海洛因戒断大鼠个体觅药行为状态识别  

Identification of the status of individual drug-administration behavior in heroin-withdrawal rats based on FrA EEG

作　　者：黄磊[1] 潘群皖[2] 胡慧娴 徐晓燕[1] HUANG Lei;PAN Qun-wan;HU Hui-xian;XU Xiao-yan(Department of Medical Engineering,Wannan Medical College,Wuhu 241002;Department of Physiology,Wannan Medical College,Wuhu 241002,China)

机构地区：[1]皖南医学院医学工程学教研室,芜湖241002 [2]皖南医学院生理教研室,芜湖241002

出　　处：《中国应用生理学杂志》2020年第5期517-523,共7页Chinese Journal of Applied Physiology

基　　金：2017年度安徽高校自然科学研究重点项目(KJ2017A259);2017年皖南医学院第二批学术和技术带头人及后备人选项目资助。

摘　　要：目的:探究海洛因诱导位置偏爱大鼠不同行为状态下额叶联络皮层(FrA)无线遥测脑电新的分析方法,以期能精确实时识别海洛因诱导大鼠个体的强迫性觅药行为。方法:清洁级Wistar大鼠30只,于前额皮层行电极埋藏术后,随机分为手术对照组(n=10)和海洛因诱导组(n=20);海洛因诱导组皮下注射海洛因0.5 mg/(kg·d),之后每日递增0.25 mg/kg,连续注射7 d,对照组同时间注射等量生理盐水;利用CPP视频系统结合脑电无线遥测技术,同步记录白-黑箱穿梭、黑箱停留、黑-白箱穿梭、白箱停留四种行为状态下已成瘾大鼠FrA区脑电信号,辨识原始脑电中含肌电等噪声信号区域,针对性的给以小波分解及振幅阈值消噪预处理,提取不同行为状态下脑电数据的样本熵值及与4个节律频率对应的小波系数标准差值,利用支持向量机算法(SVM)实现对海洛因成瘾大鼠个体不同行为状态的实时识别。结果:SVM对20只经海洛因诱导产生明显药物依赖的大鼠个体白-黑穿梭、黑箱停留、黑-白穿梭、白箱停留四种不同行为状态的实时分类识别率均值都达到80%左右,其中与觅药行为关联最紧密的黑-白穿梭状态实时识别率均值达到83.88%。结论:本文建立的对海洛因诱导大鼠个体强迫性觅药行为的实时识别方法,可作为一种检测海洛因大鼠觅药行为发动、发生的有效手段,可用于海洛因依赖患者的临床观察和觅药行为的预防。Objective:To identify the compulsive drug-seeking behavior of the individual in the heroin-addicted rat,a novel analysis method of telemetering electroencephalogram(EEG)in the frontal association cortex(FrA)induced by heroin-dependent position preference in rats.Methods:Thirty clean-grade Wistar rats after implantation of prefrontal cortex electrodes,were randomly divided into the surgical control group(n=10)and heroin-inducing group(n=20).The heroin-induced group was subcutaneously injected with heroin 0.5 mg/(kg.d),and then increased daily by 0.25 mg/kg for seven days.The control group was injected with the same amount of normal saline at the same time.Using the CPP video system combined with electroencephalogram(EEG)wireless telemetry technology,EEG signals in FrA areas of the addicted rats were recorded simultaneously in four behaviors:white-black shuttle,black-white shuttle,black-chamber stay and white-chamber stay.The areas with EMG and other noisy signals in the original EEG were identified,and wavelet decomposition and amplitude threshold denoising pre-processing were used.The sample entropy values of EEG data and wavelet coefficients corresponding to 4 rhythm frequencies under different behavioral states standard deviation were extracted,and support vector machine algorithm(SVM)was used to achieve real-time identification of different behavioral states of heroin-addicted rats.Results:SVM real-time classification recognition rate of 20 heroin abstinence rats,which are staying in black or white chamber of video box,shuttling between black-white chambers or between white-black chambers,was about 80%.Among them,the real-time recognition rate of black-white shuttle,which is closely related to drug-seeking behavior,reached 83.88%.Conclusion:In this paper,the real-time identification method of heroin-induced obsessive-compulsive drug-seeking behavior in rats can be used as an effective method to detect the initiation and occurrence of heroin-seeking drug-seeking behavior in rats.It can be used for the clinical

关 键 词：海洛因 条件性位置偏爱 大鼠 额叶联络皮层 小波阈值消噪 支持向量机 

分 类 号：R338.8[医药卫生—人体生理学]