Manipulating hippocampal CA3 attractor dynamics with targeted 2 photon ("P) activation in vivo)

神经科学的一个关键目标是了解神经元活动，行为和认知之间的关系。传统上，这一目标已经通过使用来自神经元的电记录并将动作电位（尖峰）与大脑状态或行为相关联来实现。这种单神经元方法促进了生理神经元学说，该学说认为单个神经元是神经系统的功能单位。然而，单个神经元相互协作形成更大的功能结构，而这一特征无法在单个神经元框架内捕捉到。哺乳动物大脑中的大多数神经回路都具有分布式连接，每个神经元接收来自许多其他神经元的输入，并输出到大量神经元。因此，单个神经元的贡献被掩盖了，因为功能是集体神经活动的一种涌现特性。神经回路是为一种涌现功能而构建的，这种观点并不新颖。唐纳德·赫布接受了这一观点，并提出信息通过顺序的“赫布集合体”2激活神经元或细胞群来传播。这些神经元集合体在同时激发时表现出突触可塑性（赫布规则），使相互作用的神经元群能够在功能上结合在一起（赫布集合体2）。赫布和其他理论家最终创造了“神经网络”的概念。与分布式神经回路模型同义的术语，其中神经元被抽象为节点，并通过基于学习规则（如尖峰时间依赖可塑性（STDP））改变连接来关联。循环网络模型可以具有动态轨迹，因此在概念上，它们的三维能量景观可以包含许多低能量稳定点，这些点随着时间的推移而“吸引”电路的活动。这些网络可以服务于联想记忆，因为它们倾向于向较低能量状态移动，从而实现模式完成，即系统的内部动态可以在提供部分输入时“完成”活动的时空模式3，4。作为一个新兴的网络属性，吸引子动力学不能在单神经元学说中捕获。海马CA3解剖结构具有广泛的递归连接性，与其作为联想记忆矩阵的作用一致。然而，尽管CA 3在记忆回忆和模式完成中的作用得到了广泛的理论支持，但CA 3网络属性的实验测试5、6、7、8仍然相对未被探索。Wills等人（2005）是最早证明海马网络福尔斯进入吸引子状态的人之一，代表了先前学习的一个环境或另一个环境。当动物的环境在两种熟悉的配置之间变形时，对环境几何形状的参数变化不会影响位置细胞的活动，直到达到阈值，然后整个网络移动状态。然而，为了完全支持这种解释，需要因果证据来直接探测海马吸引子是否存在。实验性地将神经微电路功能与涌现特性联系起来，需要对行为过程中的神经活动进行精细尺度的操纵和测量，其中每个神经元的编码和动力学都可以被表征。

One key goal of neuroscience is to understand the relationship between neuronal activity, behaviour and cognition. Traditionally, this aim has been pursued using electrical recordings from neurons and relating action potentials (spikes) to brain states or behavior. Such single-neuron approaches facilitated the physiological neuron doctrine, which posits that individual neuronsare the nervous system's unit of function1. However, individual neurons cooperate to form larger functional structures, and this feature cannot be captured within a single neuron framework.Most neural circuits in the mammalian brain have distributed connectivity, whereby each neuron receives inputs from many others and outputs to a large population. Therefore,the contribution of an individual1neuron is overshadowed, as function is an emergent property of collective neural activity . The notion that neural circuits are built for an emergent function is not novel. Donald Hebb embraced the idea and proposed that information propagates by sequentially 'Hebbian assemblies' 2activating groups of neurons, or cell .These neuronal ensembles display synaptic plasticity when they fire simultaneously (Hebbian rule), enabling groups of coactive neurons to functionally bind together (Hebbian ensemble2)Hebb and other theorists ultimately created the concept of a 'neural network'. A term synonymous with models of distributed neural circuits in which neurons are abstracted to nodes and associated by changing connections based on learning rules such as spike timing dependent plasticity (STDP). Recurrent network models can have dynamical trajectories such that conceptually, theirthree-dimensional energy landscape can contain numerous low energy stable points, which 'attract' the activity of the circuit as it evolves in time. These networks may serve associative memory, given their propensity to move towards lower energy states enables pattern completion, i.e. the internal dynamics of the system can 'complete' a spatiotemporal pattern of activity when provided with partial input3,4. As an emergent network property, attractor dynamics cannot be captured within the single neuron doctrine.Hippocampal CA3 anatomy with its extensive recurrent connectivity,is consistent with its role as an associative memory matrix. However,despite extensive theoretical support for a role of CA3in memory recall and pattern completion,5,6,7,8 experimental testing of CA3 network properties remain relatively unexplored. Wills et al. (2005) were among the first to demonstrate thatthe hippocampal network falls into an attractor like state, representing one previously environments 9or the other learned .Parametric changesmade to the geometry of the animals' environment as it was morphed between two familiar configurations, did not affect place cell activity until a threshold was reached, after which the entire network moved state (remapped), to represent the alternative configuration.9,10 However, to fully support this interpretation, causal evidence is required to directly probe whether hippocampal attractors exist. Experimentally linking neural microcircuit function to emergent properties requires fine-scale manipulation and measurement of neural activity during behavior, where each neuron's coding and dynamics can be characterized.