Contribution of ultralow frequency LFPs to functional MRI

静息态功能磁共振成像（rs-fMRI）包含了丰富的信息， 大脑中神经活动的大规模结构，这是一个相对未被探索的区域。RS-fMRI具有 通过识别功能网络， 在不同的实验对象中是可重复的。函数网络通常被解释为代表时间- 不同类型的区域之间的相互作用，预计将出现从认知过程，但 同样的网络结构可以在认知受到抑制或缺乏的情况下（睡眠，昏迷， 麻醉）。这种持久的网络结构是rs-fMRI中挥之不去的难题之一。我们之前的工作 表明，大规模时空准周期模式（QPP）的电活动可以隔离， BOLD信号，使我们能够将缓慢的半周期调制与更局部化的非周期调制分开 由认知和信息处理引起的活动。这让我们假设， QPP解释了神经调节的持续背景模式，在该背景模式中， 来自认知和信息处理的贡献是叠加的。我们的初步数据显示 QPP产生于与信息处理相关的神经活动不同的大脑活动 和认知，但仍然占大脑功能连接的很大一部分。目标1： 扩展我们以前的工作，研究在QPP产生中发挥作用的神经生理学来源， 多模态成像在大鼠。我们的工作模型是，QPP产生于皮层下的局部输入 然后通过神经元和星形胶质细胞的协调作用在皮层传播。 在人类中，QPP在默认模式网络（DMN）中占主导地位，DMN是一种关键结构 参与许多功能并在许多疾病中改变。我们的初步数据显示， 占DMN中的连接性的相当大的部分。在目标2中，我们将比较函数网络 通过回归去除QPP之前和之后的整个大脑的度量，以确定 次声调制的存在影响标准分析。 我们的最终目标是直接检验QPP解释背景活动的假设，在此期间- 叠加了与认知更相关的各种活动。我们将计算QPP的相对贡献 BOLD信号作为大鼠麻醉深度的函数，我们预计它们的贡献会增加， 麻醉深度增加，在人类不同难度的任务中，我们预计他们的相对 作为认知需求增加的函数，总的来说，这方面的工作 这项提案将改变我们解释rs-fMRI的方式，允许对两种不同的功能进行单独检查。 大脑活动的组成部分，可能都是临床感兴趣的。

Resting state functional magnetic resonance imaging (rs-fMRI) contains a wealth of information about the large-scale structure of neural activity in the brain, an area that has been relatively unexplored. Rs-fMRI has provided some insight into the macroscopic organization of brain activity by identifying functional networks that are reproducible across subjects. The functional networks are often interpreted as if they represent time- varying interactions between areas of the type that would be expected to arise from cognitive processes, but the same network structure can be found in conditions where cognition is suppressed or absent (sleep, coma, and anesthesia). This persistent network structure is one of the lingering puzzles in rs-fMRI. Our previous work has shown that large-scale spatiotemporal quasi-periodic patterns (QPPs) of electrical activity can be isolated from the BOLD signal, allowing us to separate slow, semi-periodic modulations from the more localized aperiodic activity that is expected to arise from cognition and information processing. This led us to hypothesize that the QPPs account for a persistent background pattern of neuromodulation, over which time-varying contributions from cognition and information processing are superimposed. Our preliminary data indicates that QPPs arise from a different type of brain activity than the neural activity linked to information processing and cognition but still account for a substantial portion of the functional connectivity in the brain. In Aim 1 , we extend our previous work to investigate the neurophysiological sources that play a role in QPP generation using multimodal imaging in the rat. Our working model is that the QPPs arise from localized input from subcortical nuclei that then propagates across the cortex through the coordinated actions of neurons and astrocytes. In humans, the QPPs are most dominant in the default mode network (DMN), a critical structure implicated in numerous functions and altered in many disorders. Our preliminary data shows that QPPs account for a substantial portion of the connectivity in the DMN. In Aim 2, we will compare functional network metrics throughout the brain before and after the QPPs are removed by regression to determine how the presence of the infraslow modulation impacts standard analysis. Our final aim directly examines the hypothesis that QPPs account for background activity over which time- varying activity more relevant to cognition is superimposed. We will calculate the relative contribution of QPPs to the BOLD signal as a function of anesthetic depth in rats, where we expect their contribution to increase as anesthetic depth increases, and during tasks with varying difficulty in humans, where we expect their relative contribution to decrease as a function of increasing cognitive demand. Taken together, the work in this proposal will change the way we interpret rs-fMRI by allowing separate examination of two distinct components of brain activity that may both be of clinical interest.