Eye Movements, Gaze Correction, and Visual Short-Term Memory

描述(由申请者提供)：人们每天进行数以万计的眼球跳动，其中很大一部分无法落在预期的物体上。为了提高效率，视觉系统必须快速纠正这些错误，并对目标物体进行矫正扫视。先前的研究表明，在简单的实验室条件下，凝视校正是快速和自动的，在这种条件下，只有一个物体存在。然而，在自然世界中，存在许多可能的凝视目标，并且确定哪个对象是原始扫视目标通常需要使用目标对象的存储器表示。然而，之前没有研究试图证明记忆被用来纠正眼跳错误。在这项拟议的研究中，我们计划证明对眼跳目标的记忆确实被用来纠正眼跳错误，并且这些纠正是快速、准确、自动的，甚至可能是无意识的。我们的中心假设是，视觉短时记忆(VSTM)存储了整个扫视过程中的目标信息，以便将目标对象与其他可见对象区分开来，并有效地纠正凝视。为了测试这一想法，我们开发了一种新的范式来模拟眼跳错误，在这种模式下，刺激阵列在眼跳过程中略有移动，从而使眼睛落在目标物体和分心物体之间。我们的初步结果表明，在这些条件下，基于记忆的凝视校正是准确、快速和自动的。这项拟议的工作将探索用于这些凝视纠正的记忆表征的性质，包括(A)测试记忆表征是否存储在VSTM中，(B)检查这些表征中存储了什么信息，(C)探索这些表征是如何自下而上地组合感觉信息的，以及(D)确定表征是否包括关于其他附近物体的信息。这些研究将促进我们对凝视矫正的基本机制的理解，这些机制支持我们与世界上的物体和代理人进行有效的互动。大多数复杂的视觉任务(准备食物、驾驶)依赖于注视一系列物体，而凝视校正在确保眼睛有效地引导到与目标相关的物体方面发挥着重要作用。因此，这项拟议的研究将为理解涉及凝视控制缺陷的条件提供必要的信息。

DESCRIPTION (provided by applicant): People make tens of thousands of saccadic eye movements each day, and a substantial proportion of these fail to land on the intended object. To be efficient, the visual system must rapidly correct these errors and make a corrective saccade to the intended object. Previous research has shown that gaze correction is fast and automatic under simple laboratory conditions, in which only a single object is present. In the natural world, however, there are many possible gaze targets, and determining which object was the original saccade target will often require the use of a memory representation of the target object. However, no previous studies have attempted to demonstrate that memory is used to correct saccade errors. In the proposed research, we plan to show that memory for the saccade target is indeed used to correct saccade errors, and that these corrections are fast, accurate, automatic, and perhaps even unconscious. Our central hypothesis is that visual short-term memory (VSTM) stores target information across the saccade so that the target object can be discriminated from other visible objects and gaze can by efficiently corrected. To test this idea, we developed a new paradigm to simulate saccade error, in which a stimulus array is shifted slightly during a saccade so that the eyes land between the target object and a distractor object. Our preliminary results suggest that memory-based gaze correction under these conditions is accurate, fast, and automatic. The proposed work will explore the nature of the memory representations used for these gaze corrections, including (a) testing whether the memory representations are stored in VSTM, (b) examining what information is stored in these representations, (c) exploring how these representations are combined bottom-up sensory information, and (d) determining whether the representations include information about other nearby objects. These studies will advance our understanding of basic mechanisms of gaze correction that support efficient interaction with objects and agents in the world. Most complex visual tasks (preparing food, driving) depend on fixating a series of objects, and gaze correction plays an important role in ensuring that the eyes are directed efficiently to goal-relevant objects. Thus, the proposed research will provide essential information for understanding conditions that involve deficits in the control of gaze.