Medical Research Council Brain Network Dynamics Unit, Nuffield Deptartment of Clinical Neurosciences, University of Oxford, Oxford OX1 3TH, United Kingdom.
Universite de Bordeaux, Institut des Maladies Neurodégénératives, 33076 Bordeaux, France; CNRS UMR 5293, Institut des Maladies Neurodégénératives, 33076 Bordeaux, France.
Neurobiol Dis. 2024 Oct 15;201:106652. doi: 10.1016/j.nbd.2024.106652. Epub 2024 Aug 28.
Defining spatial synchronisation of pathological beta oscillations is important, given that many theories linking them to parkinsonian symptoms propose a reduction in the dimensionality of the coding space within and/or across cortico-basal ganglia structures. Such spatial synchronisation could arise from a single process, with widespread entrainment of neurons to the same oscillation. Alternatively, the partially segregated structure of cortico-basal ganglia loops could provide a substrate for multiple ensembles that are independently synchronized at beta frequencies. Addressing this question requires an analytical approach that identifies groups of signals with a statistical tendency for beta synchronisation, which is unachievable using standard pairwise measures. Here, we utilized such an approach on multichannel recordings of background unit activity (BUA) in the external globus pallidus (GP) and subthalamic nucleus (STN) in parkinsonian rats. We employed an adapted version of a principle and independent component analysis-based method commonly used to define assemblies of single neurons (i.e., neurons that are synchronized over short timescales). This analysis enabled us to define whether changes in the power of beta oscillations in local ensembles of neurons (i.e., the BUA recorded from single contacts) consistently covaried over time, forming a "beta ensemble". Multiple beta ensembles were often present in single recordings and could span brain structures. Membership of a beta ensemble predicted significantly higher levels of short latency (<5 ms) synchrony in the raw BUA signal and phase synchronisation with cortical beta oscillations, suggesting that they comprised clusters of neurons that are functionally connected at multiple levels, despite sometimes being non-contiguous in space. Overall, these findings suggest that beta oscillations do not comprise of a single synchronisation process, but rather multiple independent activities that can bind both spatially contiguous and non-contiguous pools of neurons within and across structures. As previously proposed, such ensembles provide a substrate for beta oscillations to constrain the coding space of cortico-basal ganglia circuits.
定义病理性β 振荡的空间同步很重要,因为许多将其与帕金森病症状联系起来的理论都提出了皮质-基底神经节结构内和/或跨皮质-基底神经节结构的编码空间的维数降低。这种空间同步可能源于单个过程,即广泛地使神经元同步到相同的振荡。或者,皮质-基底神经节环路的部分分离结构可以为多个集合提供基础,这些集合可以在β 频率下独立同步。解决这个问题需要一种分析方法,该方法可以识别具有β 同步统计趋势的信号组,而这是使用标准的成对测量方法无法实现的。在这里,我们在帕金森病大鼠的外苍白球 (GP) 和丘脑底核 (STN) 的背景单位活动 (BUA) 的多通道记录上使用了这种方法。我们采用了一种基于原理和独立成分分析的方法的改编版本,该方法通常用于定义单个神经元的集合(即短时间尺度上同步的神经元)。该分析使我们能够确定局部神经元集合中β 振荡的功率变化是否随着时间的推移而一致变化,形成“β 集合”。单个记录中经常存在多个β 集合,并且可以跨越脑结构。β 集合的成员身份预测原始 BUA 信号中的短潜伏期 (<5 ms) 同步性和与皮质β 振荡的相位同步性显著更高,这表明它们包含了在多个水平上功能连接的神经元簇,尽管有时在空间上不连续。总体而言,这些发现表明β 振荡不是由单个同步过程组成,而是由多个独立的活动组成,这些活动可以绑定皮质-基底神经节回路内和跨结构的空间上连续和非连续的神经元池。如前所述,这样的集合为β 振荡提供了一个基础,以限制皮质-基底神经节回路的编码空间。