论文部分内容阅读
对人的一维脑电信号进行了m维的重构,计算其m阶KC复杂度与最大Lyapunov指数。实验结果表明:癫痫发作前期的特征曲线出现了变化,该特性可用于预测癫痫发作。通过分析5例实验样本与8例验证样本的24小时动态脑电信号,预测准确率为87.5%,平均预测时间为17.54s。
The one-dimensional EEG signals of human beings are reconstructed in m dimensions, and their m-order KC complexity and maximum Lyapunov exponent are calculated. The experimental results show that: the characteristic curve of the pre-seizure changes, this feature can be used to predict seizures. By analyzing 24 hours dynamic EEG signals from 5 experimental samples and 8 validated samples, the prediction accuracy was 87.5% and the average prediction time was 17.54s.