The Influence of Temporal Spacing on the Cognitive and Neural Systems Supporting Feedback Learning

奖励和厌恶的经历会对以后的决策产生强烈的影响。例如，当我们在苹果和香蕉之间做出选择时，我们的决定依赖于建立在许多经验之上的价值观。总的来说，随着时间的推移，通过学习哪些刺激和行为通常会带来有利的结果，我们可以在这个复杂多变的世界中做出适应性的选择。在过去的几十年里，神经科学研究揭示了支持从奖励反馈中学习的神经机制，证明了纹状体和中脑多巴胺系统的关键作用。简单的反馈学习范式正越来越多地被扩展到理解情绪和精神障碍以及成瘾的学习功能障碍。然而，本研究忽略了一个重要特征，即时间对学习的影响：人类反馈学习范式中的事件只相隔几秒钟，而日常环境中的学习事件几乎总是相隔很长一段时间。重要的是，20世纪关于反馈学习的早期研究发现，学习事件的间隔大大提高了学习的速度，这表明潜在的学习机制发生了定量或定性的转变。然而，值得注意的是，学习事件之间的间隔对人类奖励学习的影响尚未得到研究。此外，人类和动物研究还没有检查间隔对学习有益影响的神经基础。考虑到间隔定义了日常环境中的大多数学习条件，我们对反馈学习机制的理解可能是不完整的。本研究旨在了解时间间隔对反馈学习影响的认知和神经机制。在我担任斯坦福大学研究员期间，实验1和2（下文将对此进行回顾）通过控制单次或几周内学习事件之间的间隔，研究了基于奖励的反馈学习的行为和神经特征。我们发现由于间隔时间，学习成绩显著提高。在伦敦大学学院，拟议的实验3和4将通过检查间隔学习对重度抑郁症患者的行为影响，使用MEG检查健康年轻人的神经活动，并测试如何取消学习良好的关联，为这项工作提供重要的扩展。计划中的研究我们将采取重要的步骤来翻译这项工作，以了解情绪障碍中的学习功能障碍，并且扩展将允许在新机构完成这项工作。总的来说，这些研究将解决我们对反馈学习的基本过程的知识的巨大差距，对我们对情绪障碍和成瘾的学习的理解具有潜在的广泛意义。

Rewarding and aversive experiences exert a strong influence on later decision making. When making a choice between an apple and a banana, for example, our decision relies on values built across many experiences. In general, by learning over time which stimuli and actions regularly lead to favorable outcomes, we can make adaptive choices in our complex and changing world. Over the past few decades, neuroscience research has revealed the neural mechanisms supporting learning from reward feedback, demonstrating a critical role for the striatum and midbrain dopamine system. Simple feedback learning paradigms are increasingly being extended to understand learning dysfunctions in mood and psychiatric disorders as well as addiction. However, one important characteristic that this research ignores is the effect of time on learning: events in human feedback learning paradigms are only separated by several seconds, while learning events in the everyday environment are almost always separated by long periods of time. Importantly, early studies on feedback learning in the 20th century found that spacing of learning events strongly increased the rate of learning, suggesting a quantitative or qualitative shift in the underlying learning mechanisms. Remarkably, however, the effect of spacing between learning events on human reward learning has not been investigated. Further, human and animal research has not examined the neural basis of the beneficial effect of spacing on learning. Given that spacing defines most learning conditions in the everyday environment, our understanding of feedback learning mechanisms is likely to be incomplete. The current amended research proposal aims to understand the cognitive and neural mechanisms underlying the effect of temporal spacing on feedback learning. During my fellowship at Stanford, Experiments 1 & 2, reviewed below, examined behavioral and neural characteristics of reward-based feedback learning by manipulating spacing between learning events in a single session or across weeks. We found striking increases in learning due to spacing. At UCL, the proposed Experiments 3 & 4 will provide critical extensions to this work by examining behavioral effects of spaced learning in patients with major depressive disorder, neural activity in healthy young adults using MEG, and test how well-learned associations can be unlearned. The planned research we will take important steps in translating this work to understand learning dysfunctions in mood disorders, and the extension will allow for the completion of this work in the new institution. Overall, these studies will address a large gap in our knowledge of the fundamental processes of feedback learning, with potentially broad implications for our understanding of learning in mood disorders and addiction.