Representations Yielding Task-Dependent Behavior

描述（由申请人提供）：本研究的主要目标是促进对任务依赖行为的表征的理解。婴儿和儿童几乎在任何能力上都可能表现得早熟或有限，这取决于他们所承担的任务。例如，在违背期望的研究中，3.5个月大的婴儿似乎能够理解隐藏物体的持续存在，但婴儿直到9个月左右才能够手动搜索被咬合器隐藏的物体。同样，孩子可以回答口头询问如何在不同的情况下灵活应对，但在实际行为中却没有表现出这种灵活性。这种任务依赖行为或分离，提出了关于我们认知系统的组织、发展和训练的重要问题。本项目通过以下问题来调查这些问题：

DESCRIPTION (provided by applicant): The primary goal of this research is to advance understanding of the representations underlying task-dependent behavior. Infants and children can appear precocious or limited in almost any ability depending on the task administered to them. For example, infants as young as 3.5 months seem to understand the continued existence of hidden objects in violation-of-expectation studies, yet infants fail to manually search for objects hidden by occluders until around 9 months. Similarly, children can answer verbal queries about how to respond flexibly in different situations, yet they fail to show such flexibility in their actual behaviors. Such task-dependent behaviors, or dissociations, raise important questions about the organization, development, and training of our cognitive systems. This project investigates these issues via the following questions: 1.) What kinds of representations lead to task-dependent behaviors, specifically: How do graded representations contribute to task-dependent behaviors? How do distinct types of representations (active and latent) contribute to task-dependent behaviors? How do graded and distinct types of representations interact to influence behavior? 2.) How general is the graded, active-latent framework proposed for understanding infants' and children's task-dependent behaviors, specifically: What revisions are required for computational models of this framework to more fully account for behavioral and biological findings? What are the potential educational implications from this framework? What is the relevance of this framework for understanding adult behavior? The proposed studies investigate these questions through the integration of behavioral testing, neural network modeling, and findings from cognitive neuroscience. The proposed work provides a coherent program for advancing our understanding of the factors contributing to task-dependent behavior -- an essential step toward characterizing the nature of representations of knowledge, and in turn informing theory and its application to both typical and special populations.