Investigating neural substrates of generalization from past experience

项目摘要/摘要： 在新的环境中，我们有非凡的能力从先前的学习中总结出帮助我们实现目标的能力。这 能力被认为取决于我们表现相关经历中存在的共同关系的能力， 例如，在各种超市中的任何一家购买杂货所涉及的一系列步骤。这些背景- 任务的广义表示被称为图式。在大脑中，图式被认为是 由神经集成代表，它在不同的上下文中以类似的方式对任务进行编码。最近的研究发现 前额叶皮质(PFC)的图式表征。这些模式表示如何形成和支持 然而，在新的环境中快速学习并没有得到很好的理解。 一种被称为系统巩固理论的长期记忆形成的成熟理论提供了一种 图式表示如何在大脑中出现的模型。根据这一理论，海马体(HPC) 在新的背景下对特定情节的记忆进行快速编码。在接下来的休息期间，这些HPC 记忆被重新激活，这被认为会导致相应记忆的持久加强 大脑皮层内和大脑皮层之间的痕迹。由此产生的新皮质记忆被认为强调共同的 跨上下文的功能，为模式表示提供了基础。我们的总体假设是基于 系统整合理论认为，HPC记忆再激活促进了图式的形成 PFC中的表示(目标1和3)，并且这些表示在新的上下文中的表达依次实现 快速学习(目标2和3)。为了验证这一假设，我们将在新的环境中配对学习的大鼠模型，其中 根据先前的经验，大鼠表现出快速学习，同时记录PFC和HPC神经 集合(目标1-3)和因果干预研究(目标3)。 这些目标的完成有可能产生对神经基础的基本见解 从过去的经验中进行概括。我们的发现还将为支持或反对提供重要证据 对系统整合理论的长期预测，到目前为止，在没有 分布式同时录制。这项研究将在我的研究赞助人罗伦博士的实验室进行 弗兰克，在加州大学旧金山分校(UCSF)工作。Frank实验室位于桑德勒神经科学公司 中心，这是一个高度创新和协作的教职员工和学生社区的所在地 神经科学的问题。我在此奖学金下的培训计划将为我作为一名独立的 学术内科科学家，长期目标是揭示认知基础上的神经计算 流程。除了拟议的研究外，这一准备工作将通过严格的量化来实现 课程作业，与充满活力的知识分子社区的有计划的接触，以及与临床相关的活动。

PROJECT SUMMARY/ABSTRACT: In new settings, we have a remarkable ability to generalize from prior learning to help us achieve our goals. This capacity is thought to depend on our ability to represent the common relationships present in related experiences, such as the series of steps involved in grocery shopping at any one of a variety of supermarkets. These context- generalized representations of a task have been termed schemas. In the brain, schemas are thought to be represented by neural ensembles that encode a task similarly across contexts. Recent work has identified schema representations in the prefrontal cortex (PFC). How these schema representations form and support rapid learning in new contexts, however, is not well understood. A well-established theory of long-term memory formation known as systems consolidation theory provides a model for how schema representations may arise in the brain. According to this theory, the hippocampus (HPc) rapidly encodes memories of specific episodes in new contexts. In periods of rest that follow, these HPc memories are reactivated, and this is thought to result in the long-lasting strengthening of corresponding memory traces within and across neocortex. The resulting neocortical memories are thought to emphasize common features across contexts, providing a basis for schema representation. Our overarching hypothesis based on the theory of systems consolidation is that HPc memory reactivation promotes the formation of schema representations in PFC (Aims 1 and 3), and expression of these representations in new contexts in turn enables rapid learning (Aims 2 and 3). To test this hypothesis, we will pair a rat model of learning in new contexts wherein rats exhibit rapid learning given prior experience, with simultaneous recordings of PFC and HPc neural ensembles (Aims 1–3) and causal intervention studies (Aim 3). Completion of these Aims has the potential to yield fundamental insights into the neural substrates of generalization from past experience. Our findings will additionally provide important evidence for or against longstanding predictions of systems consolidation theory that have so far been difficult to test in the absence of distributed simultaneous recordings. This study will be carried out in the lab of my research sponsor, Dr. Loren Frank, at the University of California, San Francisco (UCSF). The Frank lab is located in Sandler Neurosciences Center, which is home to a highly innovative and collaborative community of faculty and students pursuing neuroscience questions. My training plan under this fellowship will prepare me for an independent career as an academic physician scientist with the long-term goal of revealing neural computations underlying cognitive processes. In addition to the proposed research, this preparation will be achieved through rigorous quantitative coursework, planned engagement with vibrant intellectual communities, and clinically geared activities.