Neural basis of flexible decisions in naturalistic environments

项目总结 大脑不会简单地被动地处理关于世界上物体的视觉信息。它是收费的 通过将关于物体的这些信息放在相对于我们的环境中，以一种有助于 我们与他们互动。为了在野外觅食，一只动物认出遥远的水果，权衡获得 根据每一个水果的成熟度，制定并执行路线计划，并从经验中学习。 这是可能的，因为我们的视力具有两种非凡的能力--过滤掉不相关的信息，比如 不能食用的枝叶要专注于果实，并制定灵活的计划，以达到可以达到的目标 在飞行中改变了。为了研究这些能力的神经基础，我将使用新的概念和技术 在认知和自然交互的背景下研究物体视觉的进展，如导航和 寻找一个关于视觉如何实现灵活行为的统一理论。 大脑区域的一些计算，如前额叶皮质、内嗅觉皮质等，一直被认为是 调节各种认知环境中的灵活行为，如注意力、工作记忆和导航。在这 建议，我将检验中央假设，即背外侧前额叶皮质(DlPFC)中的信号与 这些不同形式的灵活性反映了指导如何从视觉上读出对象属性的计算 V4区。鼓励将多种视觉信息来源和与对象的交互结合在一起 在本课题中，我设计了一个虚拟现实觅食平台，以验证以下关于对象如何 视觉使人的行为更加灵活。在目标1中，我将测试dlPFC(以前)中的计算是否以及如何进行 在注意的上下文中识别的)调解对具有窄范围的学习的对象的搜寻 通过影响V4中相关和不相关属性的读数来获得所需的属性。在目标2中，我将学习 对象熟悉度如何以视点不变的方式编码，以及V4和dlPFC之间的交互如何 引导觅食行为。在目标3中，我将探索视觉处理如何使获得性关联 通过比较V4和dlPFC中导致 顺序选择的执行。 这些研究的结果将对视觉感知模型具有广泛的科学意义 解释对象交互、关联和动作执行的行为和临床含义 中风和阿尔茨海默病患者受损，而视觉处理则幸免于难。在培训过程中获得的技能 指导部分-开发视觉觅食VR任务，测量和处理类似注意力的信号 以及我在生成视觉场景以分析神经编码方面的经验，将为 未来对形成健康和健康中视觉引导的体验和行动的认知过程的研究 疾病。

PROJECT SUMMARY The brain does not simply passively process visual information about the objects in the world. It is charged with placing this information about objects in the context of their environment relative to us in a way that helps us interact with them. To forage for food in the wild, an animal recognizes distant fruits, weighs the effort to get to each fruit against its perceived ripeness, makes and executes a route plan, and learns from the experience. This is possible because our vision enables two remarkable abilities – filtering out irrelevant information like inedible branches and leaves to focus on the fruit and creating flexible plans to reach the goal that can be altered on the fly. To study the neural basis of these abilities, I will use new conceptual and technological advances to study object vision in the context of cognition and naturalistic interactions like navigation and foraging to formulate a unified theory of how vision enables flexible behavior. Several computations in brain areas, like the prefrontal cortex, entorhinal cortex, etc., have been thought to mediate flexible behavior in various cognitive contexts like attention, working memory, and navigation. In this proposal, I will test the central hypothesis that signals in the dorsolateral prefrontal cortex (dlPFC) related to those different forms of flexibility reflect computations that guide how object properties are read out of visual area V4. To encourage the combination of several sources of visual information and interaction with objects in the task, I have designed a virtual reality foraging platform to test the following hypotheses about how object vision enables flexible behavior. In aim 1, I will test whether and how computations in dlPFC (previously identified in the context of attention) mediate the foraging of objects that have a narrow range of learned desirable properties by affecting the readout of relevant and irrelevant properties from V4. In aim 2, I will study how object familiarity is encoded in a viewpoint-invariant fashion and how interactions between V4 and dlPFC guide foraging behavior. In aim 3, I will explore how visual processing enables learned associations between objects and sequential actions by comparing the changes in neural activity in V4 and dlPFC that lead to the execution of sequential choices. The results of these studies will have broad scientific implications for models of visual perception that explain behavior and clinical implications for how object interaction, association, and action execution can be impaired in stroke and Alzheimer’s disease patients while visual processing is spared. Skills gained in the mentored portion – developing a visual foraging VR task and measuring and manipulating attention-like signals in dlPFC – along with my experience in generating visual scenes to analyze neural coding, will set the stage for future studies of the cognitive processes that shape visually guided experiences and actions in health and disease.