基于电生理的大鼠DMN研究

发布时间：2018-06-27 05:49

本文选题：默认模式网络 + 脑电　；参考：《电子科技大学》2017年博士论文

【摘要】：静息态(resting state)下的大脑并非是“空载”的,而是表现出自发的、有组织的、持续的神经活动,这些神经活动形成了多种静息态网络(resting-state network,RSN)。默认模式网络(default mode network,DMN)是RSN的一种,因其在静息态下表现出独特的活动模式,并在众多神经和精神疾病中发生紊乱,受到越来越多的关注。新近的研究发现,DMN并非人类独有,在啮齿类动物中同样存在。这一发现为探索DMN的生理和病理机制提供了理想的临床前模型。然而,已有的研究大多基于代谢信号,并且绝大部分信号是在动物处于“非自然”状态下获得的,这些不足极大地限制了对啮齿类动物DMN的研究。在本论文中,我们采集了自由行为状态下大鼠DMN的电生理信号,并选取了觉醒静息(wakeful rest,WR)、慢波睡眠(slow wave sleep,SWS)和快速眼动睡眠(rapid eye movement sleep,REMS)三种不同警觉状态的数据,运用多种网络分析方法,从局部脑电(electroencephalogram,EEG)的振荡频段特征、功能网络特征(包括动态网络特征)和效应网络特征三个方面,对大鼠电生理的DMN进行了比较深入的研究。主要内容如下:第一,不同警觉状态下大鼠DMN的EEG频段的统计学划分。针对不同状态的振荡频段特征,我们采用因子分析(factor analysis)对EEG振荡的功率谱密度(power spectral density,PSD)的共变特性进行聚类,由此划分各DMN脑区的频段。我们发现这些划分出的频段在状态之间以及脑区之间都存在差异。特别是,REMS状态下的θ(theta)振荡可以被进一步细分为两个频段,分别对应于紧张型REMS(tonic REMS,tREMS)和相位型REMS(phasic REMS,pREMS)两个状态。同时,我们还在SWS的不同阶段提取到对应于不同频段的两种纺锤波,包括高电压纺锤波(high-voltage spindle,HVS)和低电压纺锤波(low-voltage spindle,LVS)。这些结果为大鼠DMN在不同警觉状态下的神经振荡活动提供了新的认识,并为后续研究的数据选取以及频段选择提供了依据。第二,探讨大鼠电生理DMN功能网络的连接特性。在此,我们选取WR、SWS(不含HVS)和tREMS状态的数据,使用PSD、相位锁时值(phase locking value,PLV)和模块化分析(modularity analysis),对大鼠DMN的局部活动和功能网络特征进行了研究。结果表明,在不同警觉状态下,大鼠DMN中的局部γ(gamma)频段能量的变化与人类同源脑区的代谢波动具有一致性;并且,基于大鼠电生理DMN的网络分析结果与功能磁共振(functional magnetic resonance imaging,fMRI)研究得到的结果具有广泛的相似性。这些结果为啮齿类动物大脑中DMN的存在提供了电生理的证据。第三,进一步探讨大鼠电生理DMN功能网络的动态连接特性。在此研究中,我们利用滑动窗分析(sliding window analysis)和因子分析,对不同警觉状态下DMN的网络动态特征进行了研究。我们发现大鼠电生理DMN是高度动态变化的,并且,动态的DMN可以被进一步提取为不同的空间模式(spatial pattern)。其中,部分空间模式仅存在于特定警觉状态,而其余的空间模式独立于警觉状态存在。进一步,我们发现空间模式的贡献随时间波动,并受到警觉状态的影响。这些波动的空间模式可能为神经信息的高效整合提供了一个构架,以维持灵活的认知和行为。这些结果为理解大鼠DMN的动态功能组织提供了新的视角。第四,探究DMN中子区域之间的效应连接(effective connectivity)。本研究中,我们采用定向相位转移熵(directed phase transfer entropy,dPTE),对不同警觉状态下DMN的效应连接进行了分析。我们发现,在δ(delta)频段,大鼠DMN内存在由前至后的信息流模式;而在θ频段,存在着相反的信息流模式。绝大多数在δ频段信息流出的脑区,在θ频段存在信息流入,反之亦然,形成了具有频段特异性的信息流环路。这一现象仅存在于WR和tREMS状态。上述发现可能揭示了再进入的(reentrant)神经信息整合机制,以及潜在的意识维持机理。综上所述,本文基于自由行为状态下的大鼠电生理数据,通过多种网络分析方法,从多个角度探讨了大鼠DMN的信息整合功能。我们的发现一方面佐证了啮齿类动物中DMN的存在,另一方面也为理解其信息整合及意识维持的生理机制提供了独特的视角。
[Abstract]:The brain under the resting state (resting state) is not "empty", but shows spontaneous, organized, persistent neural activity, which forms a variety of resting-state network (RSN). The default mode network (default mode network, DMN) is a kind of RSN, because it shows unique activity in the resting state. Recent studies have found that DMN is not unique to humans and exists in rodents. This discovery provides an ideal preclinical model for exploring the physiological and pathological mechanisms of DMN. However, most of the previous studies have been based on metabolic signals. And most of the signals are obtained in the "unnatural" state of animals, which greatly limit the study of rodent DMN. In this paper, we collected electrophysiological signals of DMN in rats under free behavior, and selected the awakening quiescent (wakeful rest, WR), slow wave sleep (slow wave sleep, SWS) and rapid development. Three different alerting states of rapid eye movement sleep (REMS), using a variety of network analysis methods, from the characteristics of the oscillation frequency of local electroencephalogram (electroencephalogram, EEG), functional network features (including dynamic network features) and effect network characteristics, three aspects, the electrophysiological DMN of rats is more in-depth. The main contents are as follows: first, the statistical division of the EEG frequency band of the rat DMN under different alert states. According to the characteristics of the oscillatory bands in different states, we use factor analysis (factor analysis) to cluster the co variation of the power spectral density of the EEG oscillation (power spectral density, PSD), and then divide the frequency bands of each DMN brain region. We found that these bands are different between States and between the brain regions. In particular, the theta oscillation in the REMS state can be further subdivided into two bands, corresponding to the two states of the tense REMS (tonic REMS, tREMS) and the phase REMS (phasic REMS, pREMS). At the same time, we also carry out the different stages of SWS. Two kinds of spindle waves corresponding to different frequency bands, including high voltage spindle wave (high-voltage spindle, HVS) and low voltage spindle wave (low-voltage spindle, LVS), are taken. These results provide a new understanding for the nervous oscillation activity of rat DMN in different alert states, and provide the basis for data selection and frequency selection for subsequent research. Second, discuss the connection characteristics of the rat electrophysiological DMN function network. Here, we select WR, SWS (not HVS) and tREMS state data, use PSD, phase lock time value (phase locking value, PLV) and modularized analysis (modularity analysis), and study the local activity and functional network characteristics of the rat. Under the alert state, the local gamma (gamma) band energy changes in the rat DMN are consistent with the metabolic fluctuations in the human homologous brain region; moreover, the results of the network analysis based on the electrophysiological DMN of the rat and the functional magnetic resonance (functional magnetic resonance imaging, fMRI) have extensive similarity. These results are the meshing results. The existence of DMN in the cerebrum of the teeth provides evidence of electrophysiology. Third, we further explore the dynamic connection characteristics of the DMN functional network of rats. In this study, we used sliding window analysis (sliding window analysis) and factor analysis to study the dynamic characteristics of DMN network under different alertness. The electrophysiological DMN of rats is highly dynamic, and dynamic DMN can be further extracted as a different spatial pattern (spatial pattern). In which some spatial patterns exist only in a specific alert state, while the rest of the spatial patterns exist independently of the alert state. The spatial patterns of these fluctuations may provide a framework for the efficient integration of neural information to maintain flexible cognition and behavior. These results provide a new perspective for understanding the dynamic functional organization of the rat DMN. Fourth, explore the effect connection between the DMN neutron region (effective connectivity). In the study, we use directed phase transfer entropy (dPTE) to analyze the effect connection of DMN under different alarm states. We find that in the Delta (delta) band, the rat DMN is in the information flow pattern from the front to the back, while there is the opposite information flow pattern in the theta band. The vast majority of the information flow in the delta band is in the delta band. In the brain region, there is information flow in the theta band and vice versa, forming a loop of frequency specific information flow. This phenomenon exists only in the state of WR and tREMS. The above discovery may reveal the mechanism of the re entry of (reentrant) neural information and the underlying mechanism of the maintenance of consciousness. In summary, this article is based on the state of free behavior. The electrophysiological data of the rats, through a variety of network analysis methods, explored the information integration function of the rat DMN from a variety of angles. Our discovery, on the one hand, supported the existence of DMN in rodents, and on the other hand, it provided a unique perspective for understanding the physiological mechanism of its information integration and the maintenance of consciousness.
【学位授予单位】：电子科技大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：R338

【参考文献】