基于ICA的消除ERP中伪迹干扰的方法

概述了消除事件相关脑电位(ERP)伪迹的一般方法和存在的问题，介绍了独立分量分析(ICA)的基本原理及其两个关键内容，将ICA算法用于消除ERP伪迹干扰，根据ERP伪迹的特点用ICA算法编写了一组用于盲信号分离和伪迹干扰消除的程序，利用该组程序，进行了仿真实验和消除ERP实测数据中伪迹干扰的处理实验，仿真实验完全分离出了4组人为随机混合的信号，实测信号实验较好地消除了伪迹干扰的影响，实验结果表明，使用ICA算法消除ERP伪迹干扰是一种较理想的方法。

Using ICA to Move Artifacts from Event-Related Potentials

The current methods to remove artifacts from ERP and the existent problems were summarized, while the fundamental of ICA and its two key facts were introduced. By applying the ICA to remove the artifacts from ERP, a set of programs used in blind signal separation and ERP artifact removal are programmed according to the character of ERP artifact and ICA. The set of programs are used to make a simulation and an experiment to remove artifact from ERP collected by NEROSCAN system. In the simulation the four signals being mixed at random are separated entirely, and in the experiment the artifacts are almost removed. Both of the results show that ICA is an ideal algorithm to move artifact from ERP.