Spatial coding in the hippocampal formation: boundaries and grids

记忆是我们日常生活的核心，它决定了我们是谁。大脑中一个叫做海马体结构(HF)的区域对我们拥有的一些最有用的记忆至关重要(这些记忆告诉我们在哪里可以找到我们关心的人，我们想吃的食物，哪些地方可以让我们兴奋，哪里可以避免我们可能害怕的地方)；对我们最宝贵的记忆(那些关于我们生活中的事件的记忆，比如初吻、婚礼、出生和死亡的记忆，给了我们身份和人性)也是至关重要的。事实证明，高频对于一种被称为分心性空间记忆的空间记忆是至关重要的。同心空间指的是大尺度或类似地图的空间，它是根据环境来定义的，而不是自我中心空间，自我中心空间是根据你的视网膜和皮肤等感官以及你的四肢等运动效应器定义的。简而言之，以自我为中心的表征对于接球或从树上摘水果等行为至关重要；但记住水源位置、在自然地形上长途航行等行为也需要以分配为中心的、类似地图的表征。海马体结构对记忆也是至关重要的，因为记忆依赖于对背景的表征，例如情节记忆(对于我们生活中的事件)。这可能是海马体提供了一个时空背景，记忆的内容(新娘、母亲、他们的座位位置、鸡、恶作剧)可以与之绑定。我们知道，从许多类型的证据来看，HF对这类记忆至关重要，特别是海马区受损的患者找不到方向，不容易形成新的情节记忆。我们对位置和背景的感觉很可能是由高频“位置细胞”驱动的，也就是说，神经元在依赖于环境的特定位置放电。这种地方感来自：1)自我运动线索；2)外部环境线索。“网格单元”在所有环境中以非常几何的方式发出，就像瓷砖一样，代表着自我运动的信息。就像人类的A-Z地图集一样，在城镇街道和建筑上铺设等边正方形，哺乳动物的大脑在环境上铺设宽广等边的三角形。我们的研究团队最近对边界矢量细胞的发现。当环境边界(如墙、落差)位于与受试者的特定距离和偏心方向时，这些信号就会触发，这补充了理解环境线索贡献的关键缺失部分。因此，我们似乎知道神经元根据我们自己的运动建立空间位置，也知道神经元根据我们与周围环境的关系建立我们的空间位置。我们现在需要的是一个广泛的、系统级别的理解，了解环境边界的表示在高频中是如何组织的，以及边界单元如何与其他空间单元相互作用。我们通过记录大量单个神经元和局部场电位(向我们提供有关脑波的信息)来实现这一点。为了帮助我们进一步操纵神经元的特定子集，在我们记录的同时，我们向经过基因改造的神经元发射激光，这些神经元受到特定频率的激光的影响(激发、失活)。为了能够以受控的方式操纵外部环境线索，我们还使用了虚拟现实。我们想要探索HF的特殊解剖回路，重点放在称为下丘的区域，这可能是边界表征的组织中心，以及内嗅皮层和CA1，两者都与下丘紧密相连。我们想要测试我们的新想法，关于下丘是如何组织的，以及它如何向HF提供与边界相关的信息。我们的工作集中在空间认知和记忆上，但我们希望揭示情景记忆的神经表征和组织的一般原理。

Memory is central to our everyday lives, and make us who we are. An area of the brain called the hippocampal formation (HF) is crucial to some of the most useful memories we have (those that tell us where to find the people we care about, the food we want to eat, the places that promise excitement, and where to avoid places we might be scared of); and crucial to our most treasured memories (those memories of the events of our lives, such as first kisses, weddings, births, and deaths, that give us our identity and humanity). It turns out that the HF is crucial for a kind of spatial memory called allocentric spatial memory. Allocentric refers to large-scale or 'map-like' space, which is defined with reference to the environment as opposed to egocentric space, which is defined in relation to sensory organs like your retina and skin, and in relation to motor effectors like your limbs. Briefly, egocentric representations are crucial for behaviours such as catching a ball or picking fruit from a tree; but behaviours such as remembering the location of water-sources, and navigating long distances over natural terrain also require allocentric, map-like, representations. The hippocampal formation is also crucial for memory that relies on a representation of a context, such as episodic memory (for the events of our lives). It may be that the hippocampus provides a spatiotemporal context to which the contents of a memory (a bride, a mother, their seat locations, the chicken, bad jokes) can be bound. We know that the HF is crucial for these kinds of memories from many types of evidence, notably that patients with damaged hippocampi don't find their way about well, and can't easily form new episodic memories. Our sense of location & context is likely driven by HF 'place cells', that is, neurons which fire in context-dependent, specific locations. This sense-of-place derives from: 1) self-motion cues, and; 2) external environmental cues. 'Grid cells', which fire in a very geometric tile-like way throughout all environments, represent self-motion information. Like human A-Z atlases, tiling out equal-sided squares over town streets & buildings, the mammalian brain tiles out broadly-equilateral triangles over environments. The recent discovery by our research team of boundary vector cells. which fire when an environmental boundary (e.g. walls, drop-offs) is located at a specific distance & allocentric direction from the subject, supplied the crucial missing piece in understanding the environmental cue contribution. So we seem to know about neurons which establish spatial location based on our own movements, and about neurons which establish our spatial location based on where we are in relation to our surroundings. What we now need is a broad, systems-level understanding of how the representation of environmental boundaries is organised in the HF, and how boundary cells interact with other spatial cells. We go about this by recording a lot of individual neurons and also local field potentials (informing us about brain waves). To help us further manipulate particular subsets of neurons, while we record, we beam laser light at neurons that have been genetically modified to be affected (excited, inactivated) by laser light at particular frequencies. To be able to manipulate external environmental cues in a controlled way, we also employ Virtual Reality. We want to explore particular anatomical circuits in the HF, focusing on regions called the subiculum, which may be an organizing centre for boundary representations, and the entorhinal cortex and CA1, which are both strongly connected to the subiculum. We want to test our new ideas about how the subiculum is organised, and how it might provide boundary-related information to the HF. Our work is focused on spatial cognition and memory, but we hope to reveal general principles of the neural representation and organization of episodic memory.