Cortical visual processing for navigation

项目摘要 视觉在我们的环境导航能力中起着关键作用，从识别地标和障碍物 来确定位置和航向虽然对视觉皮层的研究提供了对特性的理解， 例如方向选择性和物体识别，但对皮层回路如何提取信息知之甚少 并处理来自视觉场景的特征以支持导航。具体而言，有两个挑战。第一、 视觉刺激的性质在导航中是显著不同的，其中受试者的运动通过 世界创造了一个复杂的和动态的视觉输入，与标准的合成刺激相比， 静止的物体。其次，必须执行的视觉特征和计算的类型不同 比标准的探测或辨别模式更有优势。我们在这个提案中的目标是确定 大脑如何从丰富的、动态的视觉输入中提取相关的视觉特征，这些视觉输入是积极探索的典型特征， 并研究这些特征的神经表征如何支持视觉导航。 我们将通过三个平行的目标来研究这一点，这些目标是从视觉场景的表现中建立起来的 在V1中，在自由移动导航期间，计算导航所需的特定变量。在我们的第一个 目的，我们将使用微型头戴式摄像机测量自由移动小鼠的视觉输入， V1中的神经活动，以确定神经动力学如何在自然导航期间表示视觉场景。在 我们的第二个目标是，我们将使用多个皮层区域的大视野双光子成像，而小鼠则在导航。 在自然主义开放世界虚拟现实系统中，为了确定视觉特征如何在视觉上被表示， 皮质区在我们的第三个目标中，我们将在小鼠的旋转竞技场中使用双光子成像来确定视觉 输入用于动态更新关键导航变量：航向方向。总之，这个项目 通过机械回路研究对自由活动的动物进行基础测量， 视觉系统功能的一个重要方面

Project summary Vision plays a key role in our ability to navigate through the environment, from identifying landmarks and obstacles to determining location and heading. While studies of visual cortex have provided an understanding of properties such as orientation selectivity and object recognition, much less is known about how cortical circuitry extracts and processes features from the visual scene to support navigation. In particular, there are two challenges. First, the nature of the visual stimulus is dramatically different in navigation, where the subject's movement through the world creates a complex and dynamic visual input, in contrast to standard synthetic stimuli presented to stationary subjects. Second, the types of visual features and computations that must be performed are different in navigation than in standard detection or discrimination paradigms. Our goal in this proposal is to determine how the brain extracts relevant visual features from the rich, dynamic visual input that typifies active exploration, and investigate how the neural representation of these features can support visual navigation. We will investigate this through three parallel aims, that build up from the representation of the visual scene in V1 during freely moving navigation, to the computation of specific variables needed for navigation. In our first aim, we will measure the visual input in freely moving mice using miniature head-mounted cameras, together with neural activity in V1, to determine how neural dynamics represent the visual scene during natural navigation. In our second aim, we will use large field-of-view two-photon imaging of multiple cortical areas, while mice navigate in a naturalistic open-world virtual reality system, to determine how visual features are represented across visual cortical areas. In our third aim, we will use 2-photon imaging in mice in a rotational arena to determine how visual input is used to dynamically update a key navigational variable: heading direction. Together, this project bridges foundational measurements in freely moving animals with mechanistic circuit investigations, to provide insights into an important aspect of visual system function.