Binding Across Perception and Action

每天，我们都会与周围环境中的许多物体互动，而我们在日常生活中取得的成功，很大程度上取决于我们能否成功地处理所遇到的物体。正因为如此，认知和视觉科学中一个经久不衰的问题是人类如何在视野中观察和感知物体。换句话说，物体在人类的思想和大脑中是如何表现的？早期的答案是“对象文件”，其中各种感知特征（即颜色，形状，大小）被绑定在一起以创建连贯的对象（例如，这个对象是一个黄色的小盒子）。这个概念演变成“事件文件”，其中动作（例如，左手响应）被添加到对象表示中。我的研究通过将认知概念视为特征（“认知特征”）并与感知和运动特征结合起来形成一种新型的对象表征——“情境文件”，继续了客体表征的进化。一个概念要被认为是一种认知特征，必须通过严格的实验测试；这个概念必须能够产生自动的注意力转移，在视觉处理中优先考虑这个概念。这些可能的认知特征包括所有权（谁拥有这个物体？）、奖励（这个物体有价值吗？）、animacy（这个物体有生命吗？）和亲和性（我能用这个物体做什么？）这些认知特征在大脑中都没有直接的检测器，并且只能在有限的时间内存在（例如，情境），但在定义对象方面发挥着重要作用。如果这些认知特征可以绑定到情境文件中，那么将对它们进行测试。一旦确定了可以绑定的认知特征，本研究的下一阶段将确定知觉、运动和认知特征绑定到情境文件的速度有多快，以及物体特征之间的绑定有多强。换句话说，这部分研究将检查绑定过程如何在情境文件中工作。在为情境文件理论建立了实验基础之后，拟建研究的最后阶段将应用情境文件在运动控制领域产生新的见解，特别是指导肢体运动的规划和产生（我们通常在与物体互动时做出的运动）。具体来说，情境文件的解释力将应用于与肢体运动规划相关的两个重要问题；准备移动某些手指对其他手指后续运动的影响，以及最终目标如何违反以速度换取准确性的法则。总的来说，提出的新的情境文件理论将把物体表征的概念扩展到更多的人类大脑的认知过程中，允许对我们对物体的认知、感知和运动行为进行更有力的解释。

Every day we interact with a multitude of objects in our environment, and a large part of the success we have in our daily lives depends on the success we have in dealing with the objects we encounter. Because of this, one of the enduring issues in cognition and vision science is how humans attend and perceive objects in the visual field. In other words, how are objects represented in human minds and brains? An early answer was “object files”, where various perceptual features (i.e., colour, shape, size) were bound together to create coherent objects (e.g., this object is a small yellow box). This notion evolved into “event files”, where actions (e.g., left hand response) were added to object representations. My research continues this evolution of object representations by theorizing that cognitive concepts can be treated as features (“cognitive features”) and as such can be bound with perceptual and motor features to form a new type of object representation – a “situation file”. To be considered a cognitive feature, a concept must pass a stringent experimental test; the concept must be capable of generating automatic shifts of attention that prioritizes the concept in visual processing. Such possible cognitive features that will be examined include ownership (who owns the object?), reward (is the object valued?), animacy (is the object is alive?), and affordance (what I can use this object for?). None of these cognitive features have direct detectors in the brain, and can exist for only limited periods of time (i.e., situations), but play important roles in defining objects. These cognitive features will then be tested if they can be bound into situation files. Once the cognitive features that can be bound are determined, the next phase of this research will determine how fast perceptual, motor, and cognitive features are bound into situation files, and how strong the binding is between features in objects. In other words, this portion of the research will examine how the binding process works in situation files. Having created the experimental foundation for the theory of situation files, the last phase of the proposed research will apply situation files to generate new insights in the field of motor control, particularly the planning and production of guided limb movements (the movements we typically make when interacting with objects). Specifically, the explanatory power of situation files will be applied to two important issues related to the planning of limb movements; the aftereffect of preparing to move certain fingers on the subsequent movement of other fingers, and how final targets violate a law of trading off speed for accuracy. Overall, the proposed new theory of situation files will extend the concept of object representations to significantly more of the human brain’s cognitive processes, allowing for much more powerful explanations of our cognitive, perceptual, and motor behaviours to objects.