混合遗传算法结合时空模型评估脑电多源定位问题.pdfVIP

24 3 Vol.24 No.3 2005 6 Chinese Journal of Biomedical Engineering June 2005 1 2 邹 凌 朱善安 1( , 213016) 2(, 310027) : , ,, , ,, , 2 , ; 3 , : ; ; ; Estimation of Multiple Dipole Source Locations from EEG Data by Using Hybrid Genetic Algorithm Combined Spatio_temporal Model 1 2 ZOU Ling Z U Shan_An 1(Department of Computer Science andTechnology ,Jiangsupolytechnic University , Changzhou 213016) 2 (Colloge of Electrical Engineering ,Zhej iang University ,Hangzhou 310027) Abstract: The problem of EEG source locationmay be summed up as optimization problem.It is generally accepted that nonlinear optimization methods are good tools to solve this problem.In thispaper,on the basis of spatio_temporal source model(STSM) ,a hybridgenetic algorithm( GA) combining genetic algorithmwith localalgorithm were applied onEEG to solve the location problem of multiple dipole sources.Simulation results indicated that GA and GA are superior to conventional nonlinear optimization methods on EEGmultiple source location.When the dipole number is two,the location resultsobtainedfrom GA arebetter than from GA;whenthe dipole number isthree,the performance of GA showed little differences from that of GA. Key ords:EEG source location; spatio_temporal sourcemodel(STSM) ; genetic algorithm; hybrid genetic algorithm R318119; A 0258-8021(2005)03-0307-06 , 引言