Unraveling the synaptic and circuit mechanisms underlying a plasticity-driving instructive signal

项目总结 学习是认知的基础，需要以灵活的神经激活模式存储信息 以及神经元团内突触重量的变化(即可塑性)。这些陈述是 在几秒到几分钟甚至生命周期的时间尺度上，根据经验进行了修改。虽然是最近的 Pivotal的工作提供了对群体活动如何驱动记忆引导行为的洞察，许多 基本的问题仍然是关于神经可塑性机制，这些机制是这些疾病形成的基础 回应新体验的陈述。标准的突触可塑性规则(即，棘波计时- 依赖可塑性，STDP)需要精确定时和重复的突触前和突触后激活， 这与清醒行为的动物的网络看似混乱的活动是不一致的。在……里面 相比之下，行为时标突触可塑性(BTSP)，这是我最近与人共同发现的一条学习规则 海马CA1区经验依赖型空间表征的发展需要 只有一次诱导试验，并在认知相关的秒级时间尺度上运作。因此，BTSP 提供了第一批生物学上看似合理的机制之一，说明了单一体验如何产生 大脑活动中与学习相关的变化。这项先前的研究使我的实验室能够解决 关于学习背后的回路和突触机制的基本问题。建立在 我发表的工作，这一提议将测试内侧内嗅皮层第三层(MEC3)的模型 作为讲师，通过CA1神经元的簇状树突向其提供特定于上下文的目标信号， 从而驱动BTSP并指导CA1网络如何形成学习相关表示。 具体地说，我们将确定MEC3如何产生这个目标信号。我们将首先使用细胞外 用神经像素探头记录来监测大量内侧人群的神经活动 在灵活的空间记忆范式中，清醒小鼠的内嗅皮层(MEC)神经元 控制学习时间课程。使用这种方法，我们将确定信息流 通过MEC网络。第二，我们将使用体内MEC3神经元的全细胞记录 同样的学习任务，以确定指导信号背后的单细胞计算。我们会 确定涉及的过程，可能包括兴奋性、突触输入整合的变化，以及 可塑性。第三，我们将结合活动记录技术和光遗传学来确定 指导性信号通过本地计算产生或从上游皮层继承 地区。这一提议将对细胞、系统和认知神经科学产生深远影响。 由于学习是几乎所有主要大脑功能的基本组成部分，理解神经 学习算法，从突触到种群水平的神经编码，将提供一个基础 了解大脑如何执行所有依赖于学习的复杂任务。

PROJECT SUMMARY Learning, fundamental to cognition, requires storing of information in flexible neural activation patterns and synaptic weight changes (i.e., plasticity) within neuronal ensembles. These representations are modified with experience on the timescale of seconds to minutes and even lifetimes. Although recent pivotal work has provided insights into how population activity drives memory-guided behaviors, many fundamental questions remain about the neural plasticity mechanisms that underlie the formation of these representations in response to new experiences. The standard synaptic plasticity rule (i.e., spike timing- dependent plasticity, STDP) requires precisely timed and repetitive pre- and postsynaptic activation, which is incongruent with the seemingly chaotic activity of networks in awake behaving animals. In contrast, behavioral timescale synaptic plasticity (BTSP), a learning rule I recently co-discovered to underlie the development of experience-dependent spatial representations in hippocampal CA1, requires only a single induction trial and operates on the cognitively relevant timescale of seconds. Thus, BTSP provides one of the first biologically plausible mechanisms for how a single experience can produce learning-related changes in brain activity. This previous research has positioned my laboratory to address fundamental questions regarding the circuit and synaptic mechanisms underlying learning. Building upon my published work, this proposal will test the model that the medial entorhinal cortex layer 3 (mEC3) serves as an instructor, providing a context-specific target signal to CA1 neurons via their tuft dendrites, thereby driving BTSP and directing the CA1 network in how to form a learning-related representation. Specifically, we will determine how the mEC3 produces this target signal. We will first use extracellular recordings with Neuropixels probes to monitor the neural activity from large populations of medial entorhinal cortex (mEC) neurons in awake mice during a flexible spatial memory paradigm that allows control over the learning time course. Using this approach, we will determine the flow of information through the mEC network. Second, we will use in vivo whole-cell recordings of mEC3 neurons during the same learning task to pinpoint the single-cell computations underlying the instructive signal. We will identify the processes involved, which may include changes in excitability, synaptic input integration, and plasticity. Third, we will combine activity recording techniques and optogenetics to determine the extent to which the instructive signal is produced by local computation or inherited from upstream cortical regions. This proposal will have a far-reaching influence on cellular, systems, and cognitive neuroscience. As learning is a fundamental component of virtually all major brain functions, understanding the neural algorithms of learning, from synaptic to population level neural coding, will provide a basis for understanding how the brain performs all complex tasks that depend upon learning.