自发脑电的频谱不规则 ,与有效信号ERP频谱相重叠 ,传统滤波方法难以奏效。但由于随机噪声奇异性指数与有效信号奇异性指数大小不一样 ,小波变换模极大值在不同尺度下传播行为也不一样 ,据此可将有效信号从随机噪声中提取出来。我们发现利用小波变换模极大值恢复算法可以提取出单次ERP ,并将此方法用于 11名被试听觉ERP的实时提取研究 The spectrum of spontaneous EEG is irregular, overlapping with the effective signal ERP spectrum, the traditional filtering method is hard to work. However, since the singularity index of random noise is not the same as the singularity index of effective signal, the maximum of wavelet transform modulus is different at different scales, so the effective signal can be extracted from the random noise. We find that the single ERP can be extracted by using wavelet transform modulus maxima recovery algorithm, and this method is applied to the real-time extraction of auditory ERP in 11 subjects