基于距离测度法的中枢神经系统损伤检测

Distance Measurements Based Injury Detection of the Central Nervous System

基于ＡＲ模型的ＥＥＧ信号距离测度方法可以较好地表征中枢神经系统损伤的情 况．主要探讨在这种方法中利用不同的定阶准则建立相应的ＡＲ模型及其对距离测度结果 的影响．同时，使用自适应滤波器对原始信号进行滤波；将分离后的平稳和非平稳成分再 进行距离测度法的研究．对滤波器的参数分析表明，这种方法可以很好地消除ＥＥＧ信号 中可能混有的不同程度的噪声，得到的距离测度结果也能更好地表示损伤的变化情况．

AR model based distance measurements with EEG signals are well reliable for the injury detection of the central nervous system. This paper uses several criteria to estimate the order of the AR model and the influence for the distance measurement outcomes. Also, an adaptive filter is employed to separate steady and unsteady conponets from original EEG signals. Both components are used to detect the brain injury. The simulation results show that this method is helpful for the noise elimination in EEG signals and the distance measurement out- comes reflect the injury changes in EEG well.