径向基函数网络在神经元锋电位分类中的应用及改进  

The Application and Improvement of Radial Basis Function Network in Spike Sorting

作　　者：降惠[1] 王琳娟[1] 李杰 

机构地区：[1]山西农业大学文理学院 [2]山西长治县法院

出　　处：《电脑开发与应用》2010年第10期54-56,共3页Computer Development & Applications

摘　　要：提出了一种改进的基于径向基函数网络的锋电位分类方法。针对传统的径向基函数网络对叠加锋电位信号识别准确率不高的问题,将分段加权的思想引入了这种网络,同时实现了锋电位的分类和完全叠加信号的分离,并且有效提高了完全叠加波的识别准确率。最后用多组不同信噪比的实验数据验证了该方法。This paper proposes an improved spike sorting method based on radial basis function （RBF） network. The piecewise and weighted form is introduced into RBF network in view of the low recognition rate of traditional RBF network. It realizes the classification of single waveform and the decomposition of overlapped waveform simultaneously and improves the recognition accuracy of fully overlapped waveform. The last part of this paper verifies the proposed method using experiment including multigroup of data with different SNR.

关 键 词：神经元锋电位分类 径向基函数网络 分段加权 完全叠加锋电位 

分 类 号：TP183[自动化与计算机技术—控制理论与控制工程]