论文部分内容阅读
目的研究在连续循环反卷积(continuous loop averaging deconvolution,CLAD)方法中,刺激序列噪声增益系数Cdec与暂态诱发电位的重建质量之间的对应关系。方法首先研究真实脑电中的噪声分布特性,然后比较了不同刺激序列及其对应的噪声增益系数在中潜伏期听觉诱发电位(auditory evoked potential,AEP)提取时的性能。结果由于真实脑电中噪声的频率分布具有1/f特性,而非白噪声,从而导致Cdec标定CLAD刺激序列的反卷积性能方面存在不足。结论 Cdec不能单独作为评估刺激序列优劣的充分条件,而应在使用时综合考虑噪声的频率分布性质。
Objective To study the correspondence between the noise gain coefficient Cdec and the reconstruction quality of transient evoked potentials in the continuous loop averaging deconvolution (CLAD) method. Methods Firstly, the noise distribution in real EEG was studied. Then the performance of different stimulation sequences and their corresponding noise gain coefficients in the extraction of auditory evoked potential (AEP) were compared. Results As the frequency distribution of noise in real EEG has 1 / f characteristic, but not white noise, which leads to the lack of deconvolution performance of Cdec calibration CLAD stimulation sequence. Conclusion Cdec can not be used alone as a sufficient condition to evaluate the pros and cons of stimulation sequences. Instead, Cdec should consider the frequency distribution nature of noise synthetically.