诱发电位 (EP)信号的检测与分析技术是临床医学诊断神经系统损伤及病变的重要手段之一。但是 ,从人体体表所得到的EP信号含有大量的噪声 ,最典型的噪声是人体自发产生的脑电图信号 (EEG)。因此 ,为利用EP信号诊断神经系统的损伤和病变 ,需要从混合信号中去除EEG等噪声。独立分量分析 (ICA)是一种新近发展起来的统计信号处理方法。本文把ICA方法应用于EP信号的噪声消除 ,并与传统的自适应滤波方法进行了比较。计算机模拟表明 ,采用ICA方法进行信号噪声分离的结果明显优于自适应滤波方法 Evoked potential (EP) signal detection and analysis of clinical medicine is one of the important means of diagnosis of nervous system damage and lesions. However, the EP signal obtained from the human body surface contains a lot of noise. The most typical noise is the electroencephalogram signal (EEG) spontaneously generated by the human body. Therefore, it is necessary to remove noise such as EEG from the mixed signal in order to diagnose nervous system damage and lesion using EP signal. Independent Component Analysis (ICA) is a newly developed statistical signal processing method. In this paper, the ICA method is applied to the noise elimination of EP signals and compared with the traditional adaptive filtering method. Computer simulation shows that the ICA method for signal noise separation results significantly better than the adaptive filtering method