Modeling Correlated Signals in Memory Systems

项目概要 实验室测量短期和长期记忆的主要方法是 简单的识别记忆测试，其中向人们呈现单个项目并 询问在实验过程中是否看到该物品。最近的工作在 然而，目击者领域表明，当人们看到多个相似的 测试项目（而不是测试单个项目）可以提高他们的记忆性能。 然而，目前还没有关于如何呈现相似项目的机制解释 在测试中可能会提高记忆力。此外，尚未对类似物品的这种影响进行正式研究。 已经超出了目击者记忆的范围。它实际上是未经审查的 其他记忆系统的背景，特别是长期记忆与短期记忆，或刺激 在不同方面是相似的（即，跨视觉维度和语义维度）。此举的目的 因此，该提案有两个方面。总体目标是理论性的：开发和测试理论 测试中记忆信号之间的相关性如何从共享神经网络中产生的框架 群体代表项目，并且可以促进记忆。项目不仅会问如何 目击者领域的可概括的先前工作是，但这是该研究的一个主要新颖方面 项目解决这个问题的方法是利用计算模型来开发一个 综合理论框架有助于解决具体如何相似性的问题 测试中呈现的项目之间有利于记忆。第二个目标是实证性的：检验 测试中呈现的刺激之间的相似性会影响记忆，这一发现的普遍性如何， 跨越长期和短期记忆系统，以及与视觉和视觉相关的刺激 语义维度。因此，该项目将是第一个在整个研究中测试这些影响的项目 不同的记忆系统，以及基于其相关的项目的研究 语义或视觉相似性。无论获得什么实证结果，这一工作 将对记忆研究人员和从业者产生广泛的影响，因为它将首先 时间，深入了解相似项目的影响的一般性或边界条件 记忆力测试；此外，当前的工作将为认知神经研究人员提供 用于检查这些效应的神经特征的概念和计算工具。

Project Summary The dominant method for measuring short- and long-term memory in the laboratory is a straightforward recognition memory test in which people are presented with a single item and asked whether the item was or was not seen during the experiment. Recent work in the eyewitness domain suggests, however, that when people are presented with multiple similar items at test —instead of a single item at test— their memory performance is facilitated. However, currently there is no mechanistic explanation regarding how presenting similar items at test may improve memory. Furthermore, the formal study of this effect of similar items has not been extended outside of the context of eyewitness memory. It is virtually unexamined in the context of other memory systems, particularly, long- versus short-term memory, or stimuli that are similar in different ways (i.e., across visual versus semantic dimensions). The aim of this proposal is thus two-fold. The overarching aim is theoretical: To develop and test a theoretical framework for how correlations amongst memory signals at test may arise from shared neural populations representing items, and may facilitate memory. Not only will the project ask how generalizable previous work from the eyewitness domain is, but a major novel aspect of the project approach to this problem is that it will leverage computational modeling to develop an integrative theoretical framework that helps address the question of how, specifically, similarity between items presented at test facilitate memory. The second aim is empirical: To examine how generalizable the finding is that similarity amongst stimuli presented at test affects memory, across the long- and short-term memory systems, and for stimuli that are related on visual and semantic dimensions. Thus, the project will be the first to test these effects across the study of different memory systems, and across the study of items that are related based on their semantic or visual similarity. Regardless of what empirical results are obtained, this line of work will have broad implications for memory researchers and practitioners because it will, for the first time, provide insight into the generality or boundary conditions for the effects of similar items in memory test; furthermore, the current work will equip cognitive neural researchers with the conceptual and computational tools to examine the neural signatures of these effects.