The Neural Basis of Functional MRI Responses

认知神经生理学和成像单位(UCNI)正与邻近的神经生理学成像设备(NIF)密切合作，以阐明可用微电极测量的神经电信号与用功能磁共振成像(FMRI)测量的功能神经成像信号之间的关系。基于过去几十年的研究，神经生理学和成像研究普遍同意视觉刺激在大脑皮层中的处理方式。然而，许多研究发现，在注意力或主观知觉等认知变量与身体刺激分离的条件下，这些技术之间存在差异。例如，在某些视觉错觉中，对显著视觉模式的知觉抑制可以导致fMRI信号的显著减少，而完全不影响测量的电活动。我们对这种差异的基础很感兴趣，因为它可能提供对这两个信号之间关系的洞察。 我们一直在沿着几条线进行研究，以阐明在感知状态与物理刺激分离的条件下fMRI信号的神经基础。我们的第一种方法使用了被称为广义闪光抑制的新范式，在本年度报告的另一个项目中进行了描述。使用这一范式，我们测量了报告是否感觉到显著目标的猴子初级视觉皮质(V1)的fMRI信号和单神经元反应。正如人们所预期的那样，在靶子似乎可见的一半试验中，两个信号的反应都很高。然而，在靶子消失的另一半试验中，fMRI反应显示大幅下降，而神经反应保持稳定。这种知觉抑制过程中大脑信号的分离解决了文献中的一个主要谜团，表明解释观察到的差异的既不是物种差异(人与猴子)，也不是特定的范式差异。相反，在感知与物理刺激条件不匹配的情况下，功能磁共振成像和神经信号本质上是不同的。 我们使用多接触电极对这一初步观察进行了更详细的研究，允许同时监测不同层的皮质功能。哺乳动物的皮质是一个复杂的结构，其层叠是高度有组织的。在这项研究中，我们发现，在知觉抑制过程中，电活动的一个方面(具体地说，“局部场电位电流源密度”)受到调制，并且这种调制仅限于较高的视觉层。这些较浅的层很大程度上涉及不同皮质区域之间的连接，导致以下可能性：(1)皮质-皮质沟通对形成感知至关重要，(2)fMRI信号对皮层上层产生的电流特别敏感。 在我们目前的工作中，我们正在进一步研究神经活动的层状本质，因为它与功能磁共振信号有关。我们已经扩展了我们的实验曲目，现在包括磁孔内的神经生理学记录。这种方法使我们能够研究休息的大脑中的自发活动波动(事实上，从不活跃的意义上讲，大脑从来不会休息)。通过观察从睡眠到全神贯注的不同大脑激活状态下两种信号的耦合，我们希望更好地理解fMRI信号是如何从神经信号中衍生出来的，以及这两种信号之间的关系是否作为唤醒水平的函数是恒定的。我们预计，这些即将发表的集体发现将再次对大量依赖血液耦合功能磁共振信号来解释潜在大脑功能的人类研究产生重要影响。

The Unit on Cognitive Neurophysiology and Imaging (UCNI) is working closely together with the neighboring Neurophysiology Imaging Facility (NIF) to elaborate the relationship between electrical neural brain signals signals, measurable with microelectrodes, and functional neuroimaging signals, measured with functional magnetic resonance imaging (fMRI). Based on research in the last decades, neurophysiology and imaging studies in general agreement regarding the manner in which visual stimuli are processed in the cortex. A number of studies have, however, found differences between the techniques under conditions in which cognitive variables such as attention or subjective perception are dissociated from the physical stimulus. For example, during certain visual illusions, the perceptual suppression of a salient visual pattern can results in significant decrease in the fMRI signal, while not affecting the measured electrical activity at all. We are interested in the basis of this discrepancy, since it might provide insights into the relationship between these two signals. We have been researching along several lines to elucidate the neural underpinnings of the fMRI signal under conditions in the perceptual state is dissociated from the physical stimulation. Our first approach has used the novel paradigm called generalized flash suppression, described in another project of this annual report. Using this paradigm, we measured the fMRI signal and single-neuron response in the primary visual cortex (V1) of monkeys reporting whether they perceived a salient target. On half of the trials, where the target appeared visible, the responses were high in both signals, as one would expect. However, on the other half of the trials, where the target disappeared, the fMRI responses showed a major drop while the neural responses remained steady. This dissociation of brain signals during perceptual suppression solves a major mystery in the literature, indicating that it is neither the species difference (human vs. monkey) nor the specific paradigm differences that account for the observed discrepancy. Instead, the fMRI and neural signals are inherently different under condition in which perception does not match the physical stimulus condition. We investigated this initial observation in greater detail using multicontact electrodes, permitting the simultaneous monitoring of different layers of cortical function. The mammalian cortex is a complex structure that is highly organized in its lamination. In this investigation, we found that one aspect of the electrical activity (specifically, the "local field potential current source density") was modulated during perceptual suppression, and that this modulation was restricted to the upper visual layers. These more superficial layers are heavily involved in connections between different cortical areas, leading to the possibility that (1) cortico-cortico communication is critical for shaping perception and (2) the fMRI signal is particularly sensitive to currents generated in the upper cortical layers. In our present work, we are further investigating the laminar nature of neural activity as it relates to the functional MRI signal. We have expanded our experimental repertoire to now include neurophysiological recordings inside the magnetic bore. This approach permits us to investigate spontaneous activity fluctuations in the resting brain (which is, in fact, never at rest in the sense of being inactive). By looking at the coupling of the two types of signals under different states of brain activation ranging from sleep to full attention, we hope to better understand how the fMRI signal can be derived from the neural signal, and whether the relationship between these two signals is constant as a function of arousal level. We expect that these collective findings, to be published shortly, will again have important implications for a large number of human studies that rely upon the blood-coupled fMRI signal to interpret underlying brain function.