The neurobiology of forgetting - spontaneous synaptic remodeling, cortical representation stability,memory retention and behavioral flexibility

为什么我们会忘记呢？令人着迷的遗忘现象最初是由实验心理学家研究的，后来由实验、计算和理论神经学家研究，一个多世纪以来一直是研究的焦点。虽然健忘通常被认为是有害的，尤其是与痴呆症等病理学相关的疾病，但我们现在知道，健忘对于行为灵活性--适应不断变化的环境或对已有知识的概括--以及减轻创伤事件至关重要。几十年来，遗忘一直被认为是持续不断的生活经历的结果，这些经历干扰了之前的记忆，覆盖了它们，或者混淆了它们的找回。最近，人们还发现，遗忘还涉及到主动、有选择地“抹去”不受欢迎的信息的生物机制。这方面的大部分工作都集中在突触上，突触是将神经细胞连接成巨大而复杂的网络的特殊连接。这种关注源于一个被广泛接受的理论，即突触的变化代表了神经系统学习新任务和存储新信息的主要机制。这一学说还与一种隐含的假设相关联，即突触属性在不受生理提示或“擦除”机制驱动时，将无限期地存在。然而，与大多数人造存储设备不同的是，信息通常会持续存在，直到主动擦除，突触属性的持久性不能被认为是理所当然的：突触是由动态的、短暂的(几天)组件组成的，这些组件在几分钟到几小时的时间尺度上不断地在突触之间移动。事实上，这些动力驱动突触特性的自发变化，这些变化与生理信号驱动的特性具有相同的幅度。虽然人们试图将突触连接的波动性与突触作为信息存储设备的规范观点进行协调，但一个直接的可能性是，突触连接的自发变化可能只是导致遗忘。令人惊讶的是，这种明显的可能性几乎没有被探索过。因此，这个项目的总体目标是探索自发突触重塑、遗忘和行为灵活性之间的关系。主要目的是(A)设计特异性和选择性地影响自发突触重建率的扰动，(B)将最有效的分子修饰引入小鼠模型，(C)验证它们对体内突触重塑动力学的影响，(D)检查这些扰动如何影响这些动物皮质表征的稳定性，以及(E)通过行为测试检查这些扰动如何影响记忆、遗忘和行为灵活性。为了实现这一目标，我们提出了一个深度整合实验、先进分析方法和理论的多层次(分子、突触、网络、行为)的协作方案。(...)

Why do we forget? The fascinating phenomenon of forgetting, originally studied by experimental psychologists and later by experimental, computational, and theoretical neuroscientists, has been a focus of research for over a century. Although forgetting is often regarded as detrimental, in particular in association with pathologies such as dementia, we now know that it is crucially important for behavioral flexibility – the adaptation to changing environments or the generalization of acquired knowledge – as well as for mitigating traumatic events. For decades, forgetting has been viewed as a consequence of ongoing life experiences, which interfere with prior memories, overwrite them or obfuscate their retrieval. More recently, forgetting has also been shown toinvolve biological mechanisms that actively and selectively ‘erase’ undesirable information. Much of this work has focused on synapses, the specialized junctions that interconnect nerve cells into vast and complex networks.This focus stems from the widely accepted doctrine that changes to synapses represent the main mechanism by which nervous systems learn new tasks and store new information. This doctrine is also associated with an implicit assumption that synaptic properties, when not driven to change by physiological cues or ‘erasure’ mechanisms, will persist indefinitely. Yet, unlike most human-built storage devices, in which information typically persists until actively erased, persistence of synaptic properties cannot be taken for granted: Synapses are made of dynamic, short-lived (days) components that continuously move in, out and between synapses over time scales of minutes and hours. Indeed, these dynamics drive spontaneous changes in synaptic properties which are of the same magnitude as those driven by physiological signals. While attempts have been made to reconcile this volatility of synaptic connections with canonical views of synapses as information storage devices, a straightforward possibility is that spontaneous changes in synaptic connection might simply drive forgetting.Surprisingly, this evident possibility has hardly been explored.The overall goal of this project is thus to explore relationships between spontaneous synaptic remodeling, forgetting and behavioral flexibility. The major thrust is to (a) devise perturbations that specifically and selectively affect spontaneous synapse remodeling rates, (b) introduce the most effective molecular modifications into mouse models, (c) validate their effects on synaptic remodeling dynamics in vivo, (d) examine how these perturbations affect the stability of cortical representations in these animals, and (e)examine by behavioral testing, how these perturbations affect memory, forgetting, and behavioral flexibility. To attain this goal, we propose a collaborative, multilevel (molecular, synaptic, network, behavior) program which deeply integrates experimentation, advanced analytical methods and theory. (....)