基于改进的遗传算法求解脑电逆问题  

Application of modified genetic algorithm in inverse problem of electroencephalogram

作　　者：邢大地[1] 吴效明[1] 

机构地区：[1]华南理工大学生物科学与工程学院,广东省广州市510640

出　　处：《中国组织工程研究与临床康复》2009年第17期3268-3271,共4页Journal of Clinical Rehabilitative Tissue Engineering Research

摘　　要：由大脑头皮电压推断大脑内活动源的信息,称之为脑电逆问题,即通过脑电图数据去反演可以反映脑电活动等效偶极子源的参数信息。其中,遗传算法由于其良好的鲁棒性、自适应性和全局优化性,是解决这类非线性问题的有力工具,但仍存在一些问题。文章分别采用了改进的自适应遗传算法和混合遗传算法,分别对单偶极子和双偶极子模型进行仿真。仿真结果表明,相对于基本遗传算法,混合遗传算法在搜索的早期效率较高,而自适应遗传算法在中后期防止局部最优的性能方面有较大提高。Inverse problem of electroencephalogram （EEG） means using EEG data to get the information of equivalent dipole sources that can reflect the activity of EEG. Among them, the genetic algorithm methods are available to this estimation due to its good quality of robust, adaptability and overall optimization. However, there still have some problems. This paper applies the hybrid genetic algorithm （HGA） and the adaptive genetic algorithm （AGA） to simulate to the single and twin dipoles model. Computer simulation demonstrates that in contrary to SGA, HGA exhibits better efficiency during the early stage. But the AGA has an advantage about the convergent speed and avoiding the local optimization during the later stage.

关 键 词：脑电逆问题 偶极子 混合遗传算法 自适应遗传算法 

分 类 号：R319.1[医药卫生—基础医学]