MENTAL REPRESENTATIONS OF SPATIAL RELATIONS

描述(申请人摘要)：本提案中描述的研究 研究人类的空间记忆。该项目的目标是确定 环境中对象之间的空间关系如何在中表示 并从内存中检索。这项研究的具体目的是分离 空间记忆在什么条件下表示为 方向相关的方式，这样熟悉的视图更容易 被召回和承认的观点，而不是不熟悉的观点， 其中空间存储器以与方向无关的方式表示， 使得熟悉的和不熟悉的视图同样可访问。几个 将使用收敛的方法。在一项研究中，观察者将 了解物体在大空间中的位置，然后做出判断 使用他们对空间的记忆的相对方向(例如，“想象你 正对着鞋子站在台灯前。指着时钟“)这些 研究将操纵观察者被允许查看的数量 具有空间，可选空间参照系的一致性， 以及如何允许观察者在他或她学习到 它。第二系列实验使用场景识别来检查 空间记忆。人类观察者将看到一个或多个不同的视图 显示在三维空间中的一组点的旋转 图形监控，或将从一个或多个角度提交给内存 放置在曲面上的几个对象的位置。然后观察者们将 接受新旧识别测试，在测试中他们必须辨别熟悉的 和新近体验的空间布局的新奇观点 其他空间布局。熟悉的和新奇的观点将在 定位。这项研究的一个主要目标是确定现场 识别性能依赖于方向，并且当它是方向时 独立自主。将新的视图与存储进行匹配的机制 还将检查内存中的视图。场景的计算模型 将实施和测试识别。这一基础科学提供了一种 为理解个体在空间能力上的差异奠定基础。

DESCRIPTION (Applicant's Abstract): The research described in this proposal investigates human spatial memory. The goal of the project is to determine how spatial relations among objects in the environment are represented in and retrieved from memory. The specific aim of the research is to isolate the conditions under which spatial memories are represented in an orientation dependent manner, such that familiar views are more easily recalled and recognized than are unfamiliar views, and the conditions under which spatial memories are represented in an orientation independent manner, such that familiar and unfamiliar views are equally accessible. Several convergent methods will be used. In one line of research, observers will learn the locations of objects in large spaces and then make judgements of relative direction using their memories of the spaces (e.g., "Imagine you are standing at the lamp facing the shoe. Point to the clock".) These studies will manipulate the number of views that observers are allowed to have of the space, the consistency of alternative spatial reference frames, and how the observer is allowed to move in the space when he or she learns it. A second series of experiments uses scene recognition to examine spatial memory. Human observers will see one or more views, differing by rotation in three-dimensional space, of a collection of dots displayed on a graphics monitor, or will commit to memory from one or more perspectives the locations of several objects resting on a surface. The observers will then be given old-new recognition tests in which they must discriminate familiar and novel views of the recently experienced spatial layout from views of other spatial layouts. The familiar and novel views will differ in orientation. A principal goal of this research is to determine when scene recognition performance is orientation dependent and when it is orientation independent. The mechanisms by which novel views are matched to stored views in memory will also be examined. Computation models of scene recognition will be implemented and tested. This basic science provides a foundation for understanding individual differences in spatial ability.