Knowledge-based control of cognition and action

熟练的认知行为，如阅读单词、识别物体或使用工具，都得到了经常相互冲突的多种心理表征的支持。例如，咖啡杯提供了许多不同的、相互排斥的抓手(例如，对手柄的闭合抓取、对马克杯桶的打开抓取、甚至对杯子边缘的捏抓)，而与咖啡杯的交互需要从其他抓取中选择一个抓取来执行。这种选择是如何发生的，对未被选择的候选人会产生什么后果？我的研究计划的目标是揭示基于知识的影响，这些影响唤起候选人的行动，并决定选择过程的结果。我的方法是由两个可行的假设指导的。首先，包括常识在内的相关先前经验的记忆是在任务执行过程中无意识地被招募的，形成了熟练执行的基础。第二，任务情境影响知识的招募和选择。在通常涉及单词阅读、对象识别或句子理解的一系列任务域上，将应用四种测量熟练动作的组成部分的方法：(1)快速识别或分类响应(例如，尽可能快地识别对象)，(2)使用适应一系列关节手势的特殊构造的响应设备执行的手部动作(例如，在相关或不相关对象的上下文中执行提示的手部动作)，(3)当观察者观看视觉显示时跟踪眼部运动(例如，搜索自然场景的图片以寻找目标对象)，以及(4)使用连接在手指和手上的传感器来记录手的形状和位置的细节，当观察者做出快速伸手和抓取反应时。这项研究将提供有关概念知识和行动表征如何支持熟练表现的重要信息，并将有助于更深入地理解阅读和物体识别方面的选择性障碍，这些障碍往往伴随着特定形式的脑损伤。

Skilled cognitive actions, such as reading words, identifying objects, or using a tool, are underpinned by multiple mental representations that often are in conflict with one another. For example, a coffee mug affords a number of different, mutually exclusive grasps (e.g., a closed grasp on the handle, an open grasp around the barrel of the mug, or even a pinch grasp on the rim) and interaction with the mug requires that one grasp be selected over the others for execution. How does this selection occur and what are the consequences for the candidates that are not selected? The objective of my research program is to reveal the knowledge-based influences that evoke candidate actions and determine the outcome of the selection process. My approach is guided by two working hypotheses. First, memory for relevant prior experience, including general knowledge, is unconsciously recruited during task execution, forming the basis for skilled performance. Second, task context influences the recruitment and selection of this knowledge. Across a range of task domains, typically involving word reading, object identification, or sentence comprehension, four methods of measuring components of skilled actions will be applied: (1) speeded identification or classification responses (e.g., identifying objects as quickly as possible), (2) hand actions performed using a specially constructed response apparatus that accommodates a range of articulated hand postures (e.g., performing a cued hand action in the context of a related or unrelated object), (3) tracking eye movements as an observer views a visual display (e.g., searching a picture of a natural scene for a target object), and (4) using sensors attached to fingers and hands to record details of hand shape and position as observers make rapid reach and grasp responses. This research will provide important information about how conceptual knowledge and action representations support skilled performance and will contribute to a deeper understanding of selective impairments in reading and object identification that often accompany specific forms of brain damage.