利用小波神经网络求解脑电等效偶极子源参数  

WAVELET NEURAL NETWORK FOR ELECTROENCEPHALOGRAM DIPOLE SOURCE LOCALIZATION

作　　者：吴清[1] 沈雪勤[1] 颜威利[1] 

机构地区：[1]河北工业大学计算机科学与软件学院,天津300130

出　　处：《中国生物医学工程学报》2004年第3期210-214,共5页Chinese Journal of Biomedical Engineering

基　　金：国家自然科学基金重点资助项目( 5 993 7160);河北省自然科学基金资助项目( 5 0 10 3 7)

摘　　要：提出了多维单尺度径向基小波神经网络的构造性算法 ,并将之应用到脑电等效偶极子源定位问题 ,从而避开对模型系统描述的困难和现有的迭代类求解方法计算时间较长的问题。通过对解空间的随机抽样由正向计算形成学习样本 ,使小波神经网络在训练过程中建立起自己的逆映射联想记忆 ,以正确推断出头皮观测电位与脑内源发生器之间的内在联系 ,使之对于新的脑电数据能实时地、准确地估算出等效偶极子源的参数 ,为脑电的动态分析提供一条途径。A multi-dimensional single-scaling radial-basis wavelet neural network (WNN) was constructed, and applied to the study of EEG source localization based on equivalent current dipole. The difficulty in the description of inverse model and the disadvantage of the high computational cost of iterative approaches could be avoided by using this network. A self-associative-memory was established in the training of WNN, which could determine the inner relation between the observation of EEG on the scalp and the electric current sources in the brain correctly. The parameters of the sources can be worked out rapidly for new EEG data that providing a novel approach to the dynamic analysis of EEG. The simulation results proved the effectiveness of the WNN method.

关 键 词：脑电图 等效电流偶极子 源定位 小波神经网络 

分 类 号：R318[医药卫生—生物医学工程]