fMRI of the Medial Temporal Lobe Memory System

内侧颞叶记忆系统的功能磁共振成像在美国国家科学基金会的支持下，斯塔克博士将进行一项为期三年的研究，使用8个功能磁共振成像(FMRI)实验来探索海马区(海马体、齿状回和下丘)以及内嗅觉、周围和旁海马皮层对陈述性记忆的贡献。该项目分为三个分项目，每个分项目的目的是解决一组具体的问题。在第一个子项目中，研究了用于编码和检索的神经底物之间的关系。这里的实验试图确定在识别记忆任务期间是否可以观察到附带的自动编码。这些实验利用了这一发现，即有意编码任务中的活动水平可以预测受试者随后的记忆(布鲁尔等人，1998；瓦格纳等人，1998)。实验1和实验2评估在识别记忆任务期间与新的箔片项目相关联的活动是否能够预测当在扫描仪外执行第二个识别记忆任务时参与者是否能够记住那些项目。对记忆提取过程中编码相关活动的观察将有助于解释文献中的几个无效结果，包括最近的几项研究，这些研究证明在某些条件下海马区有活动，但在另一些条件下不是。第二个子项目(实验3-5)询问海马体是否特别参与需要形成和使用联想的陈述性记忆任务。这一假设的几个变种在文献中得到了很好的表达，但目前的数据并不确定关联/非关联区分是否充分地捕捉到了海马区角色和邻近皮质结构角色之间的划分。实验3使用记住/知道程序来探索这一假设。实验4是连续识别记忆任务的fMRI版本，该任务以前曾用于探索健忘症患者的单一项目和联想记忆。最后，实验5试图确定提取过程中的内侧颞叶活动是否受到构成记忆的联想数量的影响。在第三个子项目中，我们探索了错误记忆的神经基础，假设了解记忆是如何失败的，以及了解真实记忆和错误记忆可能有何不同，将有助于我们进一步理解陈述性记忆系统是如何运作的。实验6是最近一项关于现实监测错误(将记忆错误归因于真实事件而不是想象事件)的ERP研究的fMRI扩展。实验7a和7b也使用了“错误信息”范式来研究记忆来源的扭曲。在这里，我们建议从记忆的原始编码开始追踪与其相关的活动，通过对后续矛盾错误信息的编码，并将这两个阶段的活动与提取原始或后续错误信息的可能性联系起来。这些研究将有助于解决该领域的一个核心问题：错误信息是否会损害或干扰对原始事件的实际记忆，或者是否对原始事件的记忆没有影响。使用功能磁共振成像来解决有关内侧颞叶的问题带来了一系列重大挑战。信号嘈杂和微弱，由于磁场不均匀而有显著的失真，解剖正常化参与者的简单技术往往表现不佳，内侧颞叶经常看起来很活跃，尽管我们尽了最大努力使其活跃。然而，拟议项目中的技术是专门为缓解这些担忧而设计的，并使内侧颞叶声明性记忆的功能磁共振成像更可行。通过专注于内侧颞叶，扫描参数、对齐技术、基线任务等方面的实验设计可以进行调整，以优化数据质量，从而影响结果。

FMRI of Medial Temporal Lobe Memory SystemAbstract With National Science Foundation support, Dr. Stark will conduct a three-year study using eight functional magnetic resonance imaging (fMRI) experiments are proposed to explore the contributions to declarative memory that are made by the hippocampal region (the hippocampus proper, dentate gyrus, and subiculum) and the entorhinal, perirhinal, and parahippocampal cortices. The project is divided into three sub-projects, with each aimed at addressing a specific set of questions. In the first subproject, the relationship between the neural substrates for encoding and retrieval is investigated. The experiments here attempt to determine whether incidental, automatic encoding can be observed during a recognition memory task. These experiments make use of the finding that the level of activity during an intentional encoding task can predict participants' subsequent memory (Brewer et al., 1998; Wagner et al., 1998). Experiments 1 and 2 assess whether activity associated with the novel foil items during a recognition memory task can predict whether participants will be able to remember those items when a second recognition memory task is administered outside the scanner. The observation of encoding-related activity during memory retrieval would help explain several null results in the literature, including several recent studies that demonstrated activity in the hippocampal region under some conditions, but not others. The second sub-project (Experiments 3-5) asks whether the hippocampus is particularly involved in declarative memory tasks that require the formation and use of associations. Several variations of this hypothesis are represented strongly in the literature, but the current data are inconclusive as to whether an associative / non-associative distinction adequately captures the division between the role of the hippocampal region and the role of the adjacent cortical structures. Experiment 3 uses the Remember / Know procedure to explore this hypothesis. Experiment 4 is an fMRI version of a continuous recognition memory task that has been previously used to explore single-item and associative memory in amnesic patients. Finally, Experiment 5 seeks to determine whether medial temporal lobe activity during retrieval is affected by the number of associations that constitute a memory. In the third sub-project, the neural substrates of false memories are explored, with the hypothesis that understanding how memories fail, and understanding how true and false memories might differ, will further our understanding of how the declarative memory system operates. Experiment 6 is an fMRI extension of a recent ERP study of reality monitoring errors (misattributing a memory as originating from a real rather than an imagined event). Experiments 7a and 7b use a "misinformation" paradigm also to study distortions of memory source. Here, we propose to trace the activity associated with a memory from its original encoding, through the encoding of subsequent contradictory misinformation, and to relate activity at both stages to the likelihood of retrieval of either the original or the subsequent misinformation. These studies will help address a central issue in the field: whether misinformation impairs or interferes with the actual memory of the original event, or whether it has no effect on the memory for the original event. Using fMRI to address questions about the medial temporal lobe poses an array of significant challenges. Signals are noisy and weak, there is significant distortion due to magnetic field inhomogeneity, simple techniques to anatomically normalize participants often perform poorly, and the medial temporal lobe often appears to be active, despite our best attempts to make it otherwise. However, the techniques in the proposed projects are specifically designed to alleviate these concerns and to make fMRI of declarative memory in the medial temporal lobes more viable. By focusing on the medial temporal lobes, scanning parameters, alignment techniques, baseline tasks, and other aspects of experimental design can be tuned to optimize the quality of the data, and therefore the impact of the results.