Liang Zhenhu, Liang Shujuan, Wang Yinghua, Ouyang Gaoxiang, Li Xiaoli
Institute of Electric Engineering, Yanshan University, Qinhuangdao 066004, China.
Department of Anesthesia, No. 1 Hospital of Qinhuangdao, Qinhuangdao 066004, China.
Clin Neurophysiol. 2015 Feb;126(2):412-22. doi: 10.1016/j.clinph.2014.05.012. Epub 2014 May 28.
Coupling in multiple electroencephalogram (EEG) signals provides a perspective tool to understand the mechanism of brain communication. In this study, we propose a method based on permutation cross-mutual information (PCMI) to investigate whether or not the coupling between EEG series can be used to quantify the effect of specific anesthetic drugs (isoflurane and remifentanil) on brain activities.
A Rössler-Lorenz system and surrogate analysis was first employed to compare histogram-based mutual information (HMI) and PCMI for estimating the coupling of two nonlinear systems. Then, the HMI and the PCMI indices of EEG recordings from two sides of the forehead of 12 patients undergoing combined remifentanil and isoflurane anesthesia were demonstrated for tracking the effect of drug on the coupling of brain activities. Performance of all indices was assessed by the correlation coefficients (Rij) and relative coefficient of variation (CV).
The PCMI can track the coupling strength of two nonlinear systems, and it is sensitive to the phase change of the coupling systems. Compared to the HMI, the PCMI has a better correlation with the coupling strength in nonlinear systems. The PCMI could track the effect of anesthesia and distinguish the consciousness state from the unconsciousness state. Moreover, at the embedding dimension m=4 and lag τ=1, the PCMI had a better performance than HMI in tracking the effect of anesthesia drugs on brain activities.
As a measure of coupling, the PCMI was able to reflect the state of consciousness from two EEG recordings.
The PCMI is a promising new coupling measure for estimating the effect of isoflurane and remifentanil anesthetic drugs on the brain activity.
多个脑电图（EEG）信号中的耦合为理解大脑通信机制提供了一种透视工具。在本研究中，我们提出一种基于排列交叉互信息（PCMI）的方法，以研究EEG序列之间的耦合是否可用于量化特定麻醉药物（异氟烷和瑞芬太尼）对大脑活动的影响。
首先采用罗塞尔 - 洛伦兹系统和替代分析，比较基于直方图的互信息（HMI）和PCMI对两个非线性系统耦合的估计。然后，展示了12例接受瑞芬太尼和异氟烷联合麻醉患者前额两侧EEG记录的HMI和PCMI指数，以追踪药物对大脑活动耦合的影响。所有指数的性能通过相关系数（Rij）和相对变异系数（CV）进行评估。
PCMI可以追踪两个非线性系统的耦合强度，并且对耦合系统的相位变化敏感。与HMI相比，PCMI与非线性系统中的耦合强度具有更好的相关性。PCMI可以追踪麻醉效果并区分意识状态和无意识状态。此外，在嵌入维度m = 4和延迟τ = 1时，PCMI在追踪麻醉药物对大脑活动的影响方面比HMI具有更好的性能。
作为一种耦合度量，PCMI能够从两个EEG记录中反映意识状态。
PCMI是一种很有前景的新耦合度量，用于估计异氟烷和瑞芬太尼麻醉药物对大脑活动的影响。
