Neural Basis of Depth Perception

描述（由申请人提供）：我们生活在一个三维（3D）的环境中，准确感知视觉场景的3D结构对许多日常活动至关重要。虽然我们对人类从多种线索（如双眼视差、运动视差和纹理）感知深度的了解很多，但我们对深度感知的神经机制的理解仍然非常不完整。这一建议解决了关于深度感知的神经基础的三个基本问题。Aim #1研究基于双目视差平滑空间变化的三维表面结构的神经表征。我们将描述层外视觉皮层的神经元如何编码由视差梯度定义的深度波的方向和空间频率，我们将探索线性或非线性感受野机制是否参与这种选择性。此外，我们将可逆地使这些大脑区域失活，以探索神经活动与感知差异定义的表面结构之间的因果关系。Aim #2为运动视差深度感知的潜在神经基础提供了第一个直接测试。我们将训练猴子区分深度和运动视差，并将使用单单元记录和电微刺激来验证MT区域在这种形式的深度感知中起重要作用的假设。Aim #3解决了一个未被探索的主要问题：我们如何在自我运动期间检测3D空间中移动的物体？我们设计了一个新的行为任务来证明检测两个深度线索（视差和运动视差）之间的不一致性为检测物体运动提供了一个强大的机制，并且我们测试了MT区域中对视差和运动视差进行不一致深度调整的神经元在这一过程中发挥重要作用的假设。本研究解决了神经回路如何从视觉场景中提取对解决特定行为任务具有重要计算意义的专门信息的一般问题，因此在系统神经科学的许多问题中具有广泛的应用。这项拟议的研究直接关系到国家眼科研究所斜视、弱视和视觉处理项目的研究重点。

DESCRIPTION (provided by applicant): We live in a three-dimensional (3D) environment, and accurate perception of the 3D structure of the visual scene is crucial for many daily activities. Although much is known about human perception of depth from multiple cues (such as binocular disparity, motion parallax, and texture), our understanding of the neural mechanisms of depth perception remains very incomplete. This proposal addresses three fundamental issues regarding the neural basis of depth perception. Aim #1 examines the neural representation of 3D surface structure based on smooth spatial variations in binocular disparity. We will characterize how neurons in extrastriate visual cortex code the orientation and spatial frequency of depth corrugations defined by disparity gradients, and we will explore whether linear or nonlinear receptive field mechanisms are involved in such selectivity. In addition, we will reversibly inactivate these brain regions to probe for causal links between neural activity and perception of disparity-defined surface structure. Aim #2 provides the first direct tests of a potential neural substrate for depth perception from motion parallax. We will train monkeys to discriminate depth from motion parallax, and will use single-unit recordings and electrical microstimulation to test the hypothesis that area MT plays an important functional role in this form of depth perception. Aim #3 tackles a major unexplored question: how do we detect objects moving in 3D space during self-motion? We have devised a novel behavioral task to demonstrate that detecting inconsistencies between two depth cues--disparity and motion parallax--provides a robust mechanism for detecting object motion, and we test the hypothesis that neurons in area MT with incongruent depth tuning for disparity and motion parallax play an important role in this process. This research addresses the general problem of how neural circuits extract specialized information from the visual scene that is computationally important for solving specific behavioral tasks, and thus has broad application to many problems in systems neuroscience. The proposed research is directly relevant to the research priorities of the Strabismus, Amplyopia, and Visual Processing program at the National Eye Institute. PUBLIC HEALTH RELEVANCE: The health-related value of this work will follow from a deeper understanding of how cognitive functions can be explained in terms of neural activity, as this will ultimately elucidate causes of various mental disorders. This work will also likely have practical applications to the development and assessment of 3D virtual environments which have growing importance in both commercial and entertainment applications. Basic science will aid development of virtual environments that are compelling, safe, and ergonomic.