Real-time control of memory encoding

项目摘要 神经活动的波动会影响记忆。随后的内存分析表明， 特定的神经状态预示着更好的工作记忆和长期记忆行为。这些分析是 通常在数据收集之后进行，但实时监测神经波动将实现更多 直接和及时的干预。我们将使用实时脑电(EEG)来跟踪瞬间到- 神经活动的瞬间波动，以便更直接地将大脑信号与行为联系起来，并增强 内存性能。在这个提案中，我们重点关注两个关键的记忆时刻：刺激前(目标1)和 保留间隔期间的主动维护(目标2)。在目标1中，我们将检验前刺激的假设 神经信号(振荡的阿尔法和西塔)预测记忆编码的成功。在实验1中，我们将改变 刺激出现的时间点，这是基于对α和θ功率的实时计算得出的。我们将使用 神经活动作为自变量，在大脑处于以下两种状态时“触发”刺激呈现 有利状态(低α，高θ)或不利状态(高α，低θ)。我们预测 更好的大脑状态将预示着更好的工作记忆和长期记忆精确度 连续报告任务。在实验2中，我们将提供神经反馈来奖励有利的前刺激 大脑状态(低阿尔法，高西塔)。我们预测，上调这些有利的州将导致 增强的内存性能(更精确的内存)。在目标2中，我们将测试以下假设： 活动跟踪信息的主动维护。持续的活动是工作记忆的关键特征，但 最近的证据通过展示活动--沉默的工作记忆来质疑它的作用。在……里面 实验3，我们将使用持续活动(对侧延迟活动，多变量)的实时测量 阿尔法拓扑学)来调整保留间隔的持续时间和工作记忆探头的身份。我们 预测当内存探测时，性能会更好(更快的反应时间，更精确的记忆) 是基于更高的持续活动而触发的。在实验4中，我们将在实验期间提供神经反馈 保持间隔，以奖励更大的持续活动。我们预测，上调这些有利的州 将带来更高的存储精度。在这些实验中，我们将探索通过镜头进行记忆编码 实时脑电触发信息(实验1和实验3)并提供反馈(实验2和4)。这个 拟议的研究将通过跟踪和驱动神经来表征助记表征的命运 编码前(目标1)和编码后(目标2)的活动，以了解神经波动如何 唤起我们记忆中的东西。

Project Summary Fluctuations of neural activity impact memory. Subsequent memory analyses have demonstrated that particular neural states predict better working memory and long-term memory behavior. These analyses are typically conducted after data collection, but monitoring neural fluctuations in real time would enable more direct and timely interventions. We will use real-time electroencephalography (EEG) to track moment-to- moment fluctuations of neural activity in order to more directly link brain signals with behavior, and to enhance memory performance. In this proposal, we focus on two key moments for memories: pre-stimulus (Aim 1) and active maintenance during a retention interval (Aim 2). In Aim 1, we will test the hypothesis that pre-stimulus neural signals (oscillatory alpha and theta) predict memory encoding success. In Experiment 1, we will vary the point of time when the stimuli appear based on real-time calculations of alpha and theta power. We will use neural activity as the independent variable to “trigger” stimulus presentation when the brain is in either advantageous states (low alpha, high theta) or disadvantageous states (high alpha, low theta). We predict that better brain states will predict better working memory and long-term memory precision in a sensitive continuous report task. In Experiment 2, we will provide neurofeedback to reward advantageous pre-stimulus brain states (low alpha, high theta). We predict that up-regulating these advantageous states will lead to enhanced memory performance (more precise memories). In Aim 2, we will test the hypothesis that sustained activity tracks active maintenance of information. Sustained activity is a key signature of working memory, but recent evidence has questioned its role through the demonstration of activity-silent working memory. In Experiment 3, we will use real-time measures of sustained activity (contralateral delay activity, multivariate alpha topography) to adjust the duration of a retention interval and the identity of working memory probes. We predict that performance will be better (quicker reaction times, more precise memories) when memory probes are triggered based on higher sustained activity. In Experiment 4, we will provide neurofeedback during the retention interval to reward greater sustained activity. We predict that up-regulating these advantageous states will lead to greater memory precision. Across these experiments, we will explore memory encoding via the lens of real-time EEG to trigger information (Experiments 1 & 3) and provide feedback (Experiments 2 & 4). The proposed research will characterize the fate of mnemonic representations by tracking and driving neural activity both pre-encoding (Aim 1) and post-encoding (Aim 2), in order to understand how neural fluctuations give rise to what we remember.