应用多分辨率小波变换提取脑电信号异常节律

脑电信号是非平稳的随机信号，其中包含了大量的生理和疾病信息，对于医生判断脑都是否有器质性的病变具有重要作用。因此对脑电信号的分析和处理一直是人们努力研究的领域。考虑到小波变换良好的时频局部化特性，利用多分辨率小波变换方法来实现脑电信号异常节律的提取，脑电信号经多分辨率小波变换后所得到的各个尺度的信号不仅反映了信号的频率信息，即尺度越大，对应信号的频率越低，同时也反映了信号的时间信息，即反映此时的EEG状态，实验结果表明，选择合适的小波基，可以有效地提取脑电信号中的异常节律。

Multiresolution Wavelet Transform for Detection of EEG Abnormal Rhythms

The signal of brain activity is a non-stationary random signal including lots of physiology and disease information, which is of important action for doctors to judge pathological changes in brain. So the analysis and process of the EEG signals are always attended. In this paper, the authors take account of the time-frequency localization of wavelet transform and use multiresolution wavelet transform to detect EEG abnormal rhythms. The signals of different scales after EEG signals are transformed by multiresolution wavelet transform not only reflect the frequency information of the signals, namely the more great scale is the lower of the frequency of the signals,but also reflect the time information of the signals, namely EEG state at that time. The test results indicate that the abnormal rhythms of the EEG signals can be detected effectively if right wavelet basis is selected.