基于时频分布的EEG中棘波的相关检测

本文提出了一种检测EEG中棘波的新方法。这种方法首先对仿真信号进行时频变换,得到模型棘波的时频特征。EEG信号经过时频变换后,计算EEG信号与模型棘波的时频相关系数,利用相关系数和时频分布中高频部分的能量,给出了一个确定EEG信号中是否含有棘波的指标,并定义这个指标为棘波概率。检测过程中,以棘波概率的大小检测棘波。实际应用表明,这种方法能够准确地检测EEG信号中的癫痫样瞬态特征

Correlation Detection of Spikes in EEG Based on Time-frequency Distribution

A new method for detecting spikes from EEG signals is presented in this paper. First, the characteristic of a simulated spike modeled signal in time frequency domain is determined by computing its discrete Choi Williams distribution(DCWD). Then, the correlation coefficients between the time frequency distribution of EEG signals and the one of the modeled spikes are computing to check their similarity. Based on the correlation coefficients and the powers of high frequency components of EEG signals in the time frequency distribution, a criterion for judging whether there are spikes in EEG signals is proposed, and the criterion is defined as the spike probability. In the process of the detection, the spike waves in EEG singnals are distinguished from the background EEG by the spike probabilities. The results show that the new method works well.