Sensory Integration for Heading Perception

这项拟议中的研究旨在了解我们通过感知自我运动的神经回路。 空间，我们通过它来判断物体相对于我们的运动。视觉系统处理大量的- 视网膜图像运动（光流）的比例模式，以估计自运动（或航向）的方向。 然而，光流被我们自己的眼睛和头部的运动所污染，以及物体在视觉中的运动。 世界因此，视觉系统可能无法独自解决这些问题。我们寻求理解 猕猴联合收割机外纹皮层神经回路是如何将视觉运动与前庭信号结合起来的 非视觉输入源（例如，眼动的传出复制），以解决复杂交织的问题 来估计自我运动和物体运动。我们通过结合电生理学 和计算研究中的一种。在目标1中，我们提出了一个通用的计算框架 （多感觉正常化），解释了关于多感觉整合的不同观察结果。 单个神经元的输入。我们设计并测试了模型的关键预测，以确定是否 多感觉正常化发生在MSTd区域，该区域参与自我感觉的视觉前庭整合。 运动知觉在目标2中，我们研究了眼睛和头部旋转如何改变MSTd神经元的调谐， 由光流和前庭输入定义的航向。我们使用人口解码方法来量化 视网膜外信号和视觉深度信息对稳定 眼/头旋转时的神经元。在目标3中，我们研究了物体运动和自我之间的相互作用， 运动，并探讨前庭输入的贡献，解离自我运动和物体运动在两个 MST地区的细分。从广义上讲，这项研究解决了一个普遍的问题， 感兴趣的变量（例如，标题）是从受混杂内部干扰影响的神经活动中提取的 和外部变量，因此在系统神经科学中的许多问题中具有广泛的应用。的 拟议的研究与斜视、弱视和视觉的研究优先事项直接相关。 国家眼科研究所的处理程序

The proposed research seeks to understand the neural circuits by which we perceive our self-motion through space and by which we judge the motion of objects relative to ourselves. The visual system processes large- scale patterns of retinal image motion (optic flow) to estimate the direction of self-motion (or heading). However, optic flow is contaminated by our own eye and head movements, as well as motion of objects in the world. As a result, the visual system likely cannot solve these problems on its own. We seek to understand how neural circuits in extrastriate cortex of macaques combine visual motion with vestibular signals and other sources of non-visual input (e.g., efference copy of eye movements) to solve the complex intertwined problems of estimating self-motion and object motion. We approach this through a combination of electrophysiological and computational studies in trained animals. In Aim 1, we propose a general computational framework (multisensory normalization) that accounts for diverse observations regarding integration of multiple sensory inputs by single neurons. We devise and test a critical prediction of the model to establish whether multisensory normalization takes place in area MSTd, which is involved in visual-vestibular integration for self- motion perception. In Aim 2, we examine how eye and head rotations alter the tuning of MSTd neurons for heading as defined by optic flow and vestibular inputs. We use a population decoding approach to quantify the relative contributions of extra-retinal signals and visual depth information to stabilizing the heading tuning of neurons during eye/head rotation. In Aim 3, we examine the interactions between object motion and self- motion and explore the contributions of vestibular inputs to dissociating self-motion and object motion in two subdivisions of area MST. In broad terms, this research addresses the general problem of how critical variables of interest (e.g., heading) are extracted from neural activity that is influenced by confounding internal and external variables, 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