Kubie John L, Fenton André A
Department of Anatomy and Cell Biology, The Robert F. Furchgott Center for Neural and Behavioral Science, S.U.N.Y. Downstate Medical Center, Brooklyn, NY 11203, United States.
Hippocampus. 2009 May;19(5):456-79. doi: 10.1002/hipo.20532.
Insect navigation is guided by heading vectors that are computed by path integration. Mammalian navigation models, on the other hand, are typically based on map-like place representations provided by hippocampal place cells. Such models compute optimal routes as a continuous series of locations that connect the current location to a goal. We propose a "heading-vector" model in which head-direction cells or their derivatives serve both as key elements in constructing the optimal route and as the straight-line guidance during route execution. The model is based on a memory structure termed the "shortcut matrix," which is constructed during the initial exploration of an environment when a set of shortcut vectors between sequential pairs of visited waypoint locations is stored. A mechanism is proposed for calculating and storing these vectors that relies on a hypothesized cell type termed an "accumulating head-direction cell." Following exploration, shortcut vectors connecting all pairs of waypoint locations are computed by vector arithmetic and stored in the shortcut matrix. On re-entry, when local view or place representations query the shortcut matrix with a current waypoint and goal, a shortcut trajectory is retrieved. Since the trajectory direction is in head-direction compass coordinates, navigation is accomplished by tracking the firing of head-direction cells that are tuned to the heading angle. Section 1 of the manuscript describes the properties of accumulating head-direction cells. It then shows how accumulating head-direction cells can store local vectors and perform vector arithmetic to perform path-integration-based homing. Section 2 describes the construction and use of the shortcut matrix for computing direct paths between any pair of locations that have been registered in the shortcut matrix. In the discussion, we analyze the advantages of heading-based navigation over map-based navigation. Finally, we survey behavioral evidence that nonhippocampal, heading-based navigation is used in small mammals and humans.
昆虫导航由通过路径积分计算得出的航向向量引导。另一方面,哺乳动物的导航模型通常基于海马体位置细胞提供的类似地图的位置表征。此类模型将最优路线计算为连接当前位置与目标的一系列连续位置。我们提出一种“航向向量”模型,其中头部方向细胞或其衍生物既作为构建最优路线的关键要素,又作为路线执行过程中的直线引导。该模型基于一种称为“捷径矩阵”的记忆结构,它是在对环境进行初始探索期间构建的,此时会存储一系列连续访问的路点位置对之间的一组捷径向量。我们提出了一种计算和存储这些向量的机制,该机制依赖于一种假设的细胞类型,即“累加头部方向细胞”。探索之后,通过向量运算计算连接所有路点位置对的捷径向量,并将其存储在捷径矩阵中。重新进入时,当局部视图或位置表征使用当前路点和目标查询捷径矩阵时,会检索到一条捷径轨迹。由于轨迹方向处于头部方向罗盘坐标中,因此通过跟踪被调整到航向角的头部方向细胞的放电来完成导航。手稿的第1节描述了累加头部方向细胞的特性。然后展示了累加头部方向细胞如何存储局部向量并执行向量运算以进行基于路径积分的归巢。第2节描述了捷径矩阵的构建和使用,用于计算已在捷径矩阵中注册的任意两个位置之间的直接路径。在讨论中,我们分析了基于航向的导航相对于基于地图的导航的优势。最后,我们综述了行为证据,表明小型哺乳动物和人类使用非海马体的、基于航向的导航。