The Neural Basis of Functional MRI Responses

为了研究功能磁共振成像和电生理学之间的联系，必须能够对同一动物进行侵入性记录和功能磁共振成像扫描。 国家精神卫生研究所的校内研究项目是世界上唯一可以进行这种测量的地点之一。 在过去的几年里，我们已经改进了我们的方法，允许从清醒的、行为正常的非人类灵长类动物的同一组织中连续和同时测量电生理和功能磁共振成像信号。 这项工作的大部分是与神经生理学成像设施直接合作完成的，这是一个致力于结构和功能脑成像以及非人类灵长类动物的共享核心设施。 在以前的研究中，我们发现，在知觉抑制期间，视觉刺激是物理存在的，但暂时逃避知觉，功能磁共振成像信号和电生理信号在初级视觉皮层变得异常分歧。 具体来说，fMRI信号反映了猴子的感知状态，而大脑同一区域的单个神经元的尖峰反映了物理刺激的存在，但并没有根据报告的感知改变它们的活动。 这种分歧没有观察到在传统的视觉刺激的介绍。 我们实验室最近的研究表明，接受来自高级视觉中心反馈输入的视觉皮层上层的突触活动显示出反映猴子接受的活动变化。 我们目前正在进行实验，以了解这种上层活动如何导致初级视觉皮层的功能磁共振成像变化，但同时不影响同一区域神经元之间的动作电位放电速率。 去年，我们发表了一项研究，调查了自发性功能磁共振成像活动的基础，这被认为是整个神经成像界所谓的功能连接的基础。 当大脑没有遇到任何感官刺激时，它的活动水平仍然很高，观察到的缓慢的内源性波动的起源知之甚少。 以前的工作已经证明，这些波动的时空相关性表现出相似的已知功能网络，这种方法目前被数百个实验室用来研究人脑。 为了研究自发的fMRI波动是否有明确的电生理基础，我们在清醒的猴子同时进行了电生理和fMRI测量。我们发现，电活动与大脑皮层大片区域的活动变化相关，这表明这种广泛的活动（通常被视为血液动力学噪声）确实具有神经学基础。事实上，大脑活动的这种全局性可能特别重要，因为它占代谢消耗的很大一部分。 这些发现已经开始影响人类神经成像界对静息时测量的整体fMRI噪声的考虑。 最后，作为“视觉感知的神经生理学”项目中提到的盲视研究的一个分支，我们一直在研究对盲视视野视觉刺激的残余fMRI反应在多大程度上反映在同一皮层区域神经元的放电中。 从皮层V4区的初步研究结果表明，尖峰反应在这个领域是非常弱的，尽管适度的功能磁共振成像反应。 相比之下，局部场电位的反应是强的，这表明，在V1，突触输入可能会引起fMRI反应的神经尖峰的情况下。 总之，这三个项目开始阐明神经和功能磁共振成像信号之间的联系，在感官刺激和自发活动在休息。

To study the link between fMRI and electrophysiology, one must be able to conduct both invasive recordings and fMRI scanning on the same animal subjects. The Intramural Research Program at the National Institute of Mental Health is one of the only sites in the world at which such measurements can be conducted. In the last years, we have refined our methods to allow for both sequential and simultaneous measurement of electrophysiological and fMRI signals from the same tissue in awake, behaving nonhuman primates. Much of this work has been done in direct collaboration with the Neurophysiology Imaging Facility, a shared core facility dedicated to structural and functional brain imaging and nonhuman primates. In a previous study, we showed that during perceptual suppression, where a visual stimulus is physically present but temporarily escapes perception, fMRI signals and electrophysiological signals in the primary visual cortex become uncharacteristically divergent. Specifically, the fMRI signal reflected the monkeys perceptual state, whereas the spiking of individual neurons in the same region of the brain reflected the presence of the physical stimulus, but did not change their activity according to the reported perception. This divergence was not observed during the conventional presentation of visual stimuli. More recent work from our lab has suggested that synaptic activity in the upper layers of the visual cortex, which receive feedback input from higher visual centers, show activity changes that reflect the monkeys percept. We are presently conducting experiments to understand how the such upper-layer activity might lead to fMRI changes in the primary visual cortex, but at the same time not affect the rate of action potential firing among neurons in the same region. Last year, we published a study investigating the basis of spontaneous fMRI activity, which is taken as the basis for so-called functional connectivity throughout the neuroimaging community. When the brain is not confronted with any sensory stimuli, its activity level remains high, with the origin of the observed slow, endogenous fluctuations poorly understood. Previous work has demonstrated that the spatiotemporal correlations of these fluctuations exhibit a similarity known functional networks, and this approach is presently employed by hundreds of labs to study the human brain. To study whether the spontaneous fMRI fluctuations have a clear electrophysiological basis, we performed simultaneous electrophysiological and fMRI measurements in awake monkeys. We found that the electrical activity correlated with activity changes over large swathes of the cerebral cortex, indicating that such widespread activity, which is typically discarded as hemodynamic noise, does indeed have a neurological basis. In fact, this global aspect of brain activity may be particularly important because it accounts for a high fraction of metabolic consumption. These findings have already begun to influence the human neuroimaging communitys consideration of global fMRI noise measured during rest. Finally, as an offshoot of the blindsight study mentioned in the project The Neurophysiology of Visual Perception, we have been examining the extent to which residual fMRI responses to visual stimuli presented to a blind visual field are reflected in the firing of neurons in the same cortical area. Initial findings from cortical area V4 suggest that spiking responses in this area are very weak despite the moderate fMRI responses. By contrast, the local field potential responses are strong, suggesting that, as in V1, synaptic inputs may give rise to fMRI responses in the absence of neural spiking. Together, these three projects are beginning to elucidate the link between neural and fMRI signals during both sensory stimulation and spontaneous activity during rest.