RIKEN Brain Science Institute, 2-1 Hirosawa, Wako-shi, Saitama, 351-0198, Japan.
RIKEN Brain Science Institute, 2-1 Hirosawa, Wako-shi, Saitama, 351-0198, Japan; Graduate School of Education, The University of Tokyo, 7-3-1 Hongo, Bunkyo-ku, Tokyo, 113-0033, Japan; Advanced Center for Computing and Communication, RIKEN, 2-1 Hirosawa, Wako, Saitama, 351-0198, Japan; Faculty of Physics, The University of Warsaw, Pasteur 5, Warsaw, 02-093, Poland.
J Neurosci Methods. 2021 Jul 15;359:109219. doi: 10.1016/j.jneumeth.2021.109219. Epub 2021 May 23.
EEG of a resting state in Alzheimer disease (AD) patients and healthy controls (HC) are analyzed to identify the characteristics of EEG in AD.
A dynamic box plot approach to the theta/beta ratio with various window durations is proposed to analyze EEG.
Spectral results during a resting state in AD patients demonstrate the effect of relatively greater power in the low-frequency bands (i.e. 'slowing down' of the EEG). A significant difference is observed in the dynamic distribution of the theta/beta ratio in the AD and HC groups, which is related to the effect of 'slowing down'. There is a more obvious visual separation between the theta/beta ratio results for the AD and HC groups with increasing window durations. Variability of the theta/beta ratio can be observed with shorter window durations with a dynamic functional box plot. This provides a better classification accuracy by using the dynamic theta/beta ratio as a sensor to discriminate AD EEG from HC EEG by using the receiver operating characteristics (ROC) curve and the area under curve (AUC) with various window durations.
COMPARISON WITH EXISTING METHOD(S): EEG spectral analysis and theta/beta ratio used to evaluate EEG typically rely on long time averaging.
The dynamic box plot approach to the theta/beta ratio with various window durations provides the possibility of observing features of the EEG. The dynamic theta/beta ratio is a better sensor to discriminate AD EEG from HC EEG. Moreover, the reliability and accuracy of results can be increased by combining spectral analysis and the dynamic box plot approach to theta/beta ratio with various window durations.
对阿尔茨海默病（AD）患者和健康对照（HC）的静息状态脑电图进行分析，以确定 AD 中脑电图的特征。
提出了一种动态箱线图方法，用于分析不同窗长的θ/β比值。
AD 患者静息状态下的频谱结果表明，低频带的功率相对较大（即“脑电图减速”）。在 AD 和 HC 组中观察到θ/β比值的动态分布存在显著差异，这与“减速”效应有关。随着窗长的增加，AD 和 HC 组之间的θ/β比值结果之间存在更明显的视觉分离。使用动态功能箱线图可以观察到较短窗长下θ/β比值的可变性。通过使用动态θ/β比值作为传感器，使用不同窗长的接收器操作特性（ROC）曲线和曲线下面积（AUC），对 AD 脑电图和 HC 脑电图进行区分，该方法具有更好的分类准确性。
用于评估脑电图的脑电图频谱分析和θ/β比值通常依赖于长时间平均。
不同窗长的θ/β比值的动态箱线图方法提供了观察脑电图特征的可能性。动态θ/β比值是区分 AD 脑电图和 HC 脑电图的更好传感器。此外，通过结合频谱分析和不同窗长的动态箱线图方法对θ/β比值进行分析，可以提高结果的可靠性和准确性。
