RBF神经网络在诱发脑电信号分类中的应用研究

Application Research of Evoked Potential Signal Classification Based on RBF Neural Networks

采用RBF神经网络方法对麻醉状态下诱发脑电仿真信号进行分析,提出了一种通过基于RBF神经网络权值构造特征向量的特征提取方法,以达到提取信号特征的目的,并用BP神经网络分类器对提取的特征进行聚类,从而实现深浅麻醉状态下诱发脑电信号的分类,实验仿真结果表明了该方法的有效性.

A feature extraction method to evoked potential signal based on RBF networks'weights is presented in this paper, through the analysis of evoked potential signal.The features extracted from signals are classified with a BP neural network for classifying evoked potential signals. Experimental simulation results show that the method is effective.