Functional architecture of the cortical network supporting relational memory

我们形成记忆不仅是为了记住过去，也是为了预测未来。后一个过程主要取决于学习事件之间的结构化关系的能力，以及做出超越个人经验的可靠推论的能力。我的研究项目的长期目标是揭示大脑如何形成和稳定事件关系的记忆，然后将它们转化为在新情况下推断出适应性行为。已有文献表明，关系记忆依赖于海马体和内侧前额叶皮层（mPFC）。这些区域通过平行的单突触和多突触通路连接。在接下来的五年里，我的研究小组将寻求揭示这些远程电路中的实时神经动力学，使关系记忆的有效形成和传递表达成为可能。为此，我们将进行平行研究，结合光遗传学、体内电生理学和纤维光度法，并在大鼠中进行成熟的关系记忆测试。首先，我们将研究mPFC控制海马体中关系记忆有效形成的远程通路。迄今为止，大多数研究都集中在记忆巩固和提取过程中海马-前额叶的相互作用上。然而，我们最近的研究表明，当新事件发生时，mPFC网络的活动会跟踪它们的相关性，这表明mPFC的相关性信号可能会调节海马体中这些事件的编码。我们建议确定这种新型相互作用的具体解剖途径。其次，我们将探索mPFC转换关系记忆以推断新情况下适应性行为的神经和电路机制。随着记忆变老，每段记忆中特有的偶然细节大多被遗忘，而常见的潜在模式被保留下来。据推测，这一过程会导致对外部世界的可概括知识的积累，从而可以预测在新情况下的适应性行为。最近在人类和啮齿动物身上的证据表明，整合和预测过程严重依赖于mPFC及其与海马体的相互作用。我们试图识别神经集成活动模式，mPFC网络在基于先前学习的关系信息的新目标导向动作上产生预期信号。同时，我们将研究在关系记忆转化过程中，mPFC与海马体相互作用的远程通路。这些研究将确定支持关系信息有效学习和传递表达的区域间相互作用的解剖学途径、时间和内容。因为关系学习和推理是人类智力的核心，我们的研究将为人类智力能力及其个体差异的基本问题提供生物学见解。

We form memory not only to remember the past but also to predict the future. The latter process critically depends on the ability to learn structured relationships among events and make robust inferences that go beyond one's experiences. The long-term goal of my research program is to uncover how the brain forms and stabilizes memories of event relations and later transforms them to infer adaptive behaviour in a new situation. Established literature suggests that relational memories depend on the hippocampus and medial prefrontal cortex (mPFC). These regions are connected through parallel mono- and multi-synaptic pathways. In the next five years, my research team will seek to uncover real-time neural dynamics within these long-range circuits that enable the effective formation and transitive expression of relational memory. To this end, we will conduct parallel studies by combining optogenetics, in vivo electrophysiology, and fiber photometry with well-established relational memory tests in rats. Firstly, we will investigate long-range pathways through which the mPFC controls the effective formation of relational memory in the hippocampus. To date, most studies have focused on the hippocampal-prefrontal interactions during memory consolidation and retrieval. Our recent studies, however, showed that the mPFC network activity tracks the relevance of novel events as they unfold, suggesting that this mPFC's relevancy signal might modulate the encoding of those events in the hippocampus. We propose to identify specific anatomical pathways underlying this novel interaction. Secondly, we will probe neural and circuit mechanisms by which the mPFC transforms relational memories to infer adaptive behaviour in a novel situation. As memories become older, incidental details unique to each memory are mostly forgotten, whereas common latent patterns are retained. This process is presumed to result in the accumulation of generalizable knowledge of the external world, which allows for predicting adaptive behaviour in a novel situation. Recent evidence in humans and rodents suggest that the integration and prediction processes critically depend on the mPFC and its interaction with the hippocampus. We seek to identify neural ensemble activity patterns with which the mPFC network generates prospective signals on a new goal-directed action based on previously learned relational information. In parallel, we will investigate the long-range pathways through which the mPFC interacts with the hippocampus during the transformation of relational memory. These studies will identify the anatomical pathway, timing, and content of the inter-region interactions supporting the effective learning and transitive expression of relational information. Because relational learning and inference are at the core of human intelligence, our study will provide biological insights that speak to fundamental questions about human intellectual abilities and their individual differences.