【摘要】 对偏头痛患者脑电信号进行特征分析研究,不仅可以提高对偏头痛疾病的确诊率,而且为偏头痛疾病的辅助诊断提供可靠依据。目前对偏头痛的脑电信号特征分析研究较少,使用优化的小波变换模极大值和神经网络技术结合的新方法来分析偏头痛患者脑电信号特征。改进小波变换模极大值法的阈值计算函数,将优化后的方法对脑电噪声信号进行去噪处理,再通过BP神经网络算法对消噪后的信号进行分类训练,设置神经网络各层的传递函数,统计神经网络分类准确率,从而分析偏头痛患者脑电信号的重要特征。仿真结果表明,改进方法可以有效地完成偏头痛脑电信号的特征提取。更多还原

【Abstract】 The Characteristic analysis of EEG signals in migraine patients can not only improve the diagnosis rate of migraine diseases, but also provide a reliable basis for the auxiliary diagnosis of migraine diseases. At present, there are few studies on the characteristics of EEG signals in migraine. This paper proposes a new method combining wavelet transform modulus maxima and neural network technology to analyze the characteristics of EEG in migraine patients. Threshold function for wavelet transform modulus maxima was improved, and then the EEG noise signal was denoised through optimized method. Finally, the denoised signals were trained and classified using the BP neural network algorithm. the transfer function of each layer of the neural network was set, and the accuracy of the classification of the neural network was calculated, thereby analyzing the important characteristics of the EEG in migraine patients. The simulation results show that the proposed method can effectively extract the features of migraine EEG signals.更多还原