Quantitative Studies of Cortical Visual Processing

描述（由申请人提供）：本提案的目的是了解正在进行的皮层活动的性质，它代表什么，以及它如何与外部刺激相互作用，以在初级视觉皮层中产生“实时”反应。我们建议使用微加工电极阵列，同时记录局部场电位和单单位活动，结合新的数学分析方法，进行以下研究：（a）测量和数学模型的动态皮质活动，在自发和刺激诱发制度。我们将测试的假设，皮质活动的结构，可以在一个低维空间表示。将开发新的数学模型，在试验到试验的基础上预测群体反应。(b)确定神经元群体如何表示视野局部区域内的物理属性（如方向）的分布。我们将测试的假设，皮质状态之间的动态切换涉及的方向分布的代表。(c)通过单个或多组电极测量和数学建模皮质群体对电刺激的反应。由此产生的模型将被用于设计一个实时控制器的皮层活动。给定一个皮质活动模式作为“目标”，我们将测试我们通过电刺激将群体反应带入目标模式的能力。我们将通过限制跨电极阵列的电刺激的时空模式来开发安全的刺激方案。这项工作的重要性在于它有助于理解大脑皮层作为一个“实时”处理设备。研究结果将揭示正在进行的活动是如何结构化的，以及它如何与传入的视觉刺激相结合，以逐个试验的方式产生反应。我们将通过研究神经群体如何代表视野局部区域内物理属性的分布来推进我们对皮层功能的基本知识。最后，这里开发的技术和方法将有助于设计用于恢复感觉功能的皮质假体，这需要使用有限数量的刺激电极来控制大量神经元的活动。

DESCRIPTION (provided by applicant): The objective of this proposal is to understand the nature of ongoing cortical activity, what it represents, and how it interacts with external stimuli to generate a "real-time" response in primary visual cortex. We propose to use micro-machined electrode arrays to simultaneously record local-field potentials and single-unit activity, in combination with novel methods of mathematical analysis, to perform the following studies: (a) to measure and mathematically model the dynamics of cortical activity, in spontaneous and stimulus evoked regimes. We will test the hypothesis that cortical activity is structured and can be represented in a low-dimensional space. Novel mathematical models will be developed that predict population responses on a trial-to-trial basis. (b) To determine how populations of neurons represent distributions of physical attributes (such as orientation) within a local area of the visual field. We will test the hypothesis that dynamic switching among cortical states is involved in the representation of orientation distributions. (c) To measure, and mathematically model, the response of the cortical population to electrical stimulation through individual or groups of electrodes. The resulting model will be used in the design of a real-time controller of cortical activity. Given a pattern of cortical activity as a "target", we will test our ability to bring the population response into the target pattern via electrical stimulation. We will develop safe stimulation protocols by constraining the spatio-temporal pattern of electrical stimulation across the electrode array. The significance of the proposed work is its contribution to understanding the cortex as a "real-time" processing device. The findings will reveal how ongoing activity is structured, and how it combines with incoming visual stimulation to generate a response on a trial-by-trial basis. We will advance our basic knowledge of cortical function by investigating how neural populations represent distributions of physical attributes within a local region of the visual field. Finally, the techniques and methods developed here will be instrumental in the design of cortical prostheses for the restoration of sensory function, which require the activity of a large population of neurons to be controlled using a limited number of stimulating electrodes.