CAREER: Modeling Response-Time Distributions to Test Theories of Event Memory

对以前经历过的事件的记忆可以有多种形式，比如决定以前是否见过一个人的脸，以及决定以前在哪里见过一个人的脸。不同的理论对这些不同形式的记忆是如何产生的有不同的假设。一些理论认为，回忆以前经历的能力是要么全有要么全无的（要么发生，要么不发生），而另一些理论则认为，信息是以一种从少到多的连续方式积累的。然而，其他理论认为，要么全有要么全无的记忆能力和不断积累信息的过程是共存的，任何一种机制都可以成为特定情况下记忆的基础（这是一个将它们分开的问题）。确定事件的真实状态很重要，原因有很多，包括开发评估目击者记忆的适当方法，开发评估记忆障碍情况下认知功能的适当工具，以及开发用于研究记忆的神经基础的适当记忆任务。这个CAREER项目由一个早期职业阶段的研究者发起，旨在通过施加研究者通常不使用的约束来区分现有的人类事件记忆模型。在理解人类事件记忆方面，以前的建模工作导致了某种程度上的僵局，因为在记忆准确性在不同实验条件下如何变化的实证研究中，不同的模型都能够适应人类数据模式。除了通常使用的记忆精度度量之外，本项目还将使用响应时间度量来进一步约束模型需要容纳的数据，以便更好地强制区分可能的理论替代方案。在一系列实验中，各种实验操作将与这些措施结合使用，以提供进一步的约束，以帮助排除模型，并潜在地确定人类事件记忆最有可能如何运作的候选模型。作为职业规划教育的一部分，研究者将在心理学专业开设一个计算建模专业，为本科生提供一个发展认知建模技能的独特机会。他还将通过网站将研究中的建模工具提供给研究界。

Memory for previously experienced events can take many forms, such as deciding whether a person's face was seen previously versus deciding where a person's face was seen previously. Different theories make different assumptions about how these various forms of memory come about. Whereas some theories assume that the ability to call to mind a prior experience is all-or-none (it either occurs or does not), other theories assume that bits of information accumulate in a continuous manner from less to more. Yet other theories assume that both an all-or-none ability to call something to mind and a process of continuously accumulating information co-exist, and either mechanism can underlie memory in a given situation (it is a matter of teasing them apart). Determining the true state of affairs is important for a number of reasons, including the development of appropriate methods for assessing eyewitness memory, the development of appropriate tools for assessing cognitive functions in cases of memory impairment, and the development of appropriate memory tasks for use in research on the neural underpinnings of memory. This CAREER project by an early career stage investigator aims to differentiate among existing models of human event memory by imposing constraints not typically used by researchers. Previous modeling efforts have led to somewhat of an impasse regarding understanding human event memory because vastly different models are all able to accommodate human data patterns produced in empirical studies of how memory accuracy varies across experimental conditions. In addition to the typically-used accuracy measures of memory, this project will use response time measures to further constrain the data needing to be accommodated by the models in order to better force differentiation among the possible theoretical alternatives. Across a number of experiments, various experimental manipulations will be used in conjunction with these measures to provide even further constraints to help to rule models out and potentially identify a candidate model for how human event memory most likely operates. As part of the education component of the CAREER plan, the investigator will create a computational modeling concentration within the Psychology major, providing undergraduate students with a unique opportunity to develop cognitive modeling skills. He will also make the modeling tools from the research available to the research community via a website.