Gomez Carlos, Poza Jesus, Gomez-Pilar Javier, Bachiller Alejandro, Juan-Cruz Celia, Tola-Arribas Miguel A, Carreres Alicia, Cano Monica, Hornero Roberto
Annu Int Conf IEEE Eng Med Biol Soc. 2016 Aug;2016:2830-2833. doi: 10.1109/EMBC.2016.7591319.
The aim of this pilot study was to analyze spontaneous electroencephalography (EEG) activity in Alzheimer's disease (AD) by means of Cross-Sample Entropy (Cross-SampEn) and two local measures derived from graph theory: clustering coefficient (CC) and characteristic path length (PL). Five minutes of EEG activity were recorded from 37 patients with dementia due to AD and 29 elderly controls. Our results showed that Cross-SampEn values were lower in the AD group than in the control one for all the interactions among EEG channels. This finding indicates that EEG activity in AD is characterized by a lower statistical dissimilarity among channels. Significant differences were found mainly for fronto-central interactions (p <; 0.01, permutation test). Additionally, the application of graph theory measures revealed diverse neural network changes, i.e. lower CC and higher PL values in AD group, leading to a less efficient brain organization. This study suggests the usefulness of our approach to provide further insights into the underlying brain dynamics associated with AD.
这项初步研究的目的是通过交叉样本熵（Cross-SampEn）以及从图论衍生出的两种局部测量方法：聚类系数（CC）和特征路径长度（PL），来分析阿尔茨海默病（AD）患者的自发脑电图（EEG）活动。记录了37例AD所致痴呆患者和29名老年对照者5分钟的EEG活动。我们的结果表明，对于EEG通道之间的所有相互作用，AD组的Cross-SampEn值均低于对照组。这一发现表明，AD患者的EEG活动特征是通道间的统计差异较低。主要在前额中央相互作用方面发现了显著差异（p < 0.01，置换检验）。此外，图论测量方法的应用揭示了不同的神经网络变化，即AD组的CC值较低而PL值较高，导致大脑组织效率较低。本研究表明我们的方法有助于进一步深入了解与AD相关的潜在脑动力学。
