Lewandowski Achim, Rosipal Roman, Dorffner Georg
Section for Artificial Intelligence, Center for Medical Statistics, Informatics and Intelligent Systems, Medical University of Vienna, Austria.
J Psychophysiol. 2013 Jul 22;27(3):105-112. doi: 10.1027/0269-8803/a000092.
Research in recent years has supported the hypothesis that many properties of the electroencephalogram (EEG) are specific to an individual. In this study, the intra- and inter-individual variations of sleep EEG signals were investigated. This was carried out by analyzing the stability of the average EEG spectra individually computed for the Rechtschaffen and Kales (RK) sleep stages. Six EEG channels were used to account for the topographical aspect of the analysis. Validity of the results was supported by considering a wide dataset of 174 subjects with normal sleep. Subjects spent two consecutive nights in the sleep laboratory during which EEG recordings were obtained. High similarity between average spectra of two consecutive nights was found considering an individual. More than 89% of the second night recordings were correctly assigned to their counterparts of the first night. The average spectra of sleep EEG computed for each RK sleep stage have shown a high degree of individuality.
近年来的研究支持了这样一种假设，即脑电图（EEG）的许多特性是个体特有的。在本研究中，对睡眠EEG信号的个体内和个体间差异进行了研究。这是通过分析为 Rechtschaffen 和 Kales（RK）睡眠阶段分别计算的平均EEG频谱的稳定性来实现的。使用六个EEG通道来考虑分析的地形方面。通过考虑174名睡眠正常受试者的广泛数据集，支持了结果的有效性。受试者在睡眠实验室连续度过两个晚上，在此期间进行了EEG记录。考虑到个体，发现两个连续晚上的平均频谱之间具有高度相似性。超过89%的第二晚记录被正确地分配到了第一晚的对应记录中。为每个RK睡眠阶段计算的睡眠EEG平均频谱显示出高度的个体性。
