The role of dynamical criticality in human perception

项目摘要 大脑活动从未停止。当我们睡着了，注意力不集中，甚至在全身麻醉下， 人脑中相互连接的神经元继续自发地产生复杂的活动模式。 感官刺激扰乱了这种持续的自发神经元活动。为了被有意识地检测到， 这种扰动的影响需要足够大，以吸引数千个神经元并持续至少 至少几百毫秒。当我们清醒和专注时，最小的刺激就足以 引起大的扰动。然而，在全身麻醉下，即使是最有害的刺激也无法到达大脑。 有意识感知的阈值。在这里，我们解决一个基本的问题：为什么感官刺激能够 在某些状态下扰乱神经元活动，而在其他状态下不会？我们假设感官刺激的能力 干扰神经元活动与称为稳定性的动力系统的性质有关。如果是神经元 动力学是不稳定的，任何扰动的影响都会随着时间的推移而无限增长， 增加神经元的数量。相反，如果动力学太稳定，那么所有的扰动将很快 抑制下来，无法达到感知的阈值。因此，我们假设有意识的感知是 当神经元动力学恰好处于稳定和不稳定之间时， 政权。我们认为这一点是关键的。为了验证临界假设，我们开发了一种新的数学模型， 技术，并将其应用于人类和非人类灵长类动物的神经生理学记录。这些 初步研究结果有力地支持了这一假设。在拟议项目中，我们将严格测试 在植入电极的人类受试者中使用皮层电图（ECoG）进行临界假设， 癫痫定位我们将确定自发活动的稳定性如何作为睡眠的函数而变化 清醒、注意力集中和嗜睡，以及镇静和全身麻醉。我们将验证 临界假设和我们通过预测对电刺激的反应来估计神经元活动稳定性的能力。 脑刺激使用听觉掩蔽语音检测任务，我们还将确定 神经元动力学可用于预测在感知阈值处呈现的自然刺激是否将被 有意识地发现。虽然许多其他神经元活动的测量方法以前都与 唤醒和感知的变化，目前，不可能将现有措施应用于 明确区分有意识和无意识的大脑活动。因此，验证此 临界假设将是一个重大进步。除了解决神经科学中的一个基本问题外， 找到一个客观和可量化的感觉反应性测量方法， 神经学和麻醉学，其中在麻醉下或脑损伤后诊断隐蔽意识 不能用现有技术可靠地制造。

PROJECT SUMMARY Brain activity never ceases. When we are asleep, inattentive, or even under general anesthesia, networks of interconnected neurons in the human brain continue to spontaneously generate complex activity patterns. Sensory stimuli perturb this ongoing spontaneous neuronal activity. In order to be consciously detected, the effect of this perturbation needs to be large enough so as to engage thousands of neurons and persist for at least several hundred milliseconds. When we are awake and attentive, the smallest stimuli are sufficient to elicit a large perturbation. Under general anesthesia, however, even the most noxious stimuli do not reach the threshold for conscious perception. Here we address a fundamental question: why are sensory stimuli able to perturb neuronal activity in some states but not in others? We hypothesize that the ability of the sensory stimuli to perturb neuronal activity is related to the property of dynamical systems termed stability. If neuronal dynamics were unstable, the effect of any perturbation would grow over time without bounds and engage ever increasing number of neurons. Conversely, if the dynamics were too stable, then all perturbations will quickly dampen down and fail to reach threshold of perception. Thus, we hypothesize that conscious perception is most likely to occur when the neuronal dynamics are poised precisely between the stable and unstable regimes. We refer to this point as critical. To test the criticality hypothesis, we developed novel mathematical techniques and applied them to neurophysiological recordings in humans and in nonhuman primates. These preliminary findings strongly support the hypothesis. In the proposed project, we will rigorously test the criticality hypothesis using electrocorticography (ECoG) in human subjects implanted with electrodes for epilepsy localization. We will determine how the stability of spontaneous activity varies as a function of sleep and wake, attentiveness and drowsiness, as well as sedation and general anesthesia. We will validate the criticality hypothesis and our ability to estimate stability of neuronal activity by predicting responses to electrical brain stimulation. Using an auditory masked speech detection task, we will also determine whether stability of neuronal dynamics can be used to predict whether a natural stimulus presented at perceptual threshold will be consciously detected. While many other measures of neuronal activity have been previously associated with changes in arousal and perception, at present, it is not possible to apply the existing measures to unequivocally distinguish between activity in the conscious and unconscious brain. Hence, validating this criticality hypothesis would be a major advance. In addition to addressing a fundamental issue in neuroscience, finding an objective and quantifiable measure of sensory responsiveness has profound clinical significance in neurology and in anesthesiology where diagnoses of covert awareness under anesthesia or after brain injury cannot be made reliably with existing technology.