Post-Attentive Vision

我们无法完全感知眼睛一次接收到的所有视觉信息。如果可以的话，初学打字的人就不需要找C键了；他们可以简单地看一下键盘，就知道C是下一排左起第三个键。这是不会发生的。即使C很显眼，新手还是得一个键一个键地搜索，直到找到他们需要的那个。在找到C之前，打字员会在C的位置看到一些东西；他们只是没有完全理解它是字母c。这个例子说明，视觉可以大致分为两个阶段。首先，像颜色、运动和大小这样的基本特征是在整个视野中同时被感知的。（这种“预先注意”的基本特征可能有十几个。）然后，当注意力被引导到视野的特定区域时，该区域的特征被绑定在一起，形成物体的表征。因此，当我们第一次看到一个场景时，我们看到的是一系列预先注意的特征。只有当注意力集中在一个物品上时，这些特征才能在感知上结合在一起，形成一个可识别的物体。在这个项目中，我们正在研究“后注意力视觉”，即当注意力从一个物体上移开后会发生什么。这些特性是否仍然绑定在一起？传统的答案是肯定的。“然而，我们的初步数据表明，相反，当注意力离开一个物体时，视觉表征会恢复到其未受约束的注意前状态。”例如，当人们通过单个字母的视觉显示反复搜索目标字母时（例如，是否有E?，是否有G?），人们可能会认为，随着注意力被引导到显示中越来越多的字母上，搜索将变得越来越有效。然而，这并没有发生；显然，第100次试验的搜索效率并不比第1次试验高。基于这些和其他最近的发现，我们将对人们不会随着时间的推移而建立视觉感知的假设进行进一步的测试，我们也将探讨从这个假设中产生的几个问题。研究结果将回答关于视觉感知本质的基本问题，也将对我们理解涉及视觉搜索的现实世界任务（例如，驾驶、解读x光片、阅读）具有实际意义。

We cannot fully perceive all of the visual information our eyes can take in at once. If we could, beginning typists would not need to hunt for the C key; they could simply look at the keyboard and know that the C is the third key from the left in the bottom row. That doesn't happen. Even though the C is in plain sight, novice typists have to search from key to key until they find the one they need. Before finding the C, the typists see something at the location of the C; they just have not fully perceived it as the letter C. This example illustrates that vision can be broadly divided into two stages. First, basic features like color, motion, and size are perceived across the entire visual field at the same time. (There are, perhaps, a dozen of these `pre-attentive` basic features.) Then, as attention is directed to a particular region of the visual field, the features in that region are bound together into representations of objects. Thus, when we first look at a scene, what we see is a collection of pre-attentive features. Only when attention is deployed to an item are those features perceptually bound together into a recognizable object. In this project we are studying `post-attentive vision,` asking what happens after attention moves away from an object. Do the features remain bound together? The traditional answer has been `yes.` However, our preliminary data suggest that, to the contrary, when attention leaves an object, the visual representation reverts to its pre-attentive state of unbound features. For example, when people search repeatedly through a visual display of individual letters for a target letter (e.g., Is there an E?; Is there a G?) one might think the search would become increasingly efficient as attention was directed to more and more of the letters in the display. However, this does not happen; the search is apparently no more efficient on trial 100 than on trial 1. Building upon these and other recent findings, we will conduct further tests of the hypothesis that people do not build up a visual percept over time, and we will also explore several issues arising from this hypothesis. The results will speak to basic questions about the nature of visual perception, and will also have practical implications for our understanding of real world tasks involving visual search (e.g., driving, interpreting X-rays, reading).