DYNAMIC MODELS FOR LATENCY ACCURACY RELATIONS IN MEMORY

内存中延迟精度关系的动态模型

• 批准号: 6538935
• 负责人: MICHAEL J WENGER
• 金额: $ 6.82万
• 依托单位: UNIVERSITY OF NOTRE DAME
• 依托单位国家: 美国
• 财政年份: 2001
• 资助国家: 美国
• 起止时间: 2001-04-01 至 2004-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/6538935
• 关键词: clinical research; form /pattern perception; human subject; memory; model design /development; psychological models; visual perception; visual stimulus

Explorations of the nature of psychological information and the cognitive processes that operate on information typically use either response latency or probability as a primary dependent measure. This is due partly to the clear difficulties associated with interpreting (for example) latencies in reduced-accuracy conditions or response probabilities when performance is near ceiling. Over the past few decades, the separate bodies of literature dealing with these two measures of performance have developed rather refined approaches to assessing aspects of psychosocial (on the basis of response probabilities) and the processes that operate on that information (on the basis of response latencies). The proposed project is intended to provide initial steps in developing and testing a theoretical language to relate response latency and probability in a variety of cognitive tasks. The theory will make use of linear dynamic systems formulations, augmented with stochastic components. It will be developed explicitly to make contact with two existing bodies of theory, each of which has explored fundamental characteristics of cognition. The first-systems factorial technology (e.g., Townsend & Nozawa, 1995)- addresses elementary processes characteristics, while the second-general recognition theory (e.g., Ashby & Townsend, 1986) addresses aspects of independence in psychological information. In addition, the work will be applied to developing and testing alternative representations for an hypothesis concerning memory for meaningful visual objects, such as faces.

对心理信息的本质和对信息进行操作的认知过程的探索通常使用反应潜伏期或概率作为主要的相关度量。这部分是由于与解释（例如）在降低精度条件下的延迟或响应概率时，性能接近上限的明显困难。在过去的几十年里，处理这两个性能指标的单独的文献机构已经开发出相当完善的方法来评估心理社会方面（基于响应概率）和操作该信息的过程（基于响应延迟）。该项目旨在提供开发和测试理论语言的初始步骤，以将各种认知任务中的响应延迟和概率联系起来。该理论将利用线性动态系统公式，增加随机成分。它将被明确地发展，以便与两个现有的理论体系联系起来，每一个理论体系都探索了认知的基本特征。第一系统阶乘技术（例如，汤森和野泽，1995年）-地址的基本过程的特点，而第二一般识别理论（例如，Ashby &汤森，1986）解决了心理信息中独立性的各个方面。此外，这项工作将被应用于开发和测试替代表示的假设有关的记忆有意义的视觉对象，如面孔。