The role of noradrenaline in signaling prediction error and promoting learning

预测性学习是高度适应性的；它允许我们和其他动物使用环境中的信息来预测和准备具有生物学意义的事件。同样重要的是，当环境条件发生变化时，生物体对即将发生的事件的预测不再准确，能够更新以前的学习。这个项目考察了神经递质去甲肾上腺素(NA)在这种特殊形式的错误驱动学习中的作用。检测和编码事件之间的关系，特别是在面对新信息时修改先前的学习，对于自适应、灵活的行为控制至关重要。然而，我们仍然对使这一点成为可能的神经系统知之甚少。虽然大量研究表明多巴胺(DA)与最初对奖励事件的学习有关，但科比特博士最近进行的研究表明，更新这种学习，特别是在灭绝(当预期的强化被省略)时，涉及去甲肾上腺素(NA)。关于NA在发出奖励预测错误信号方面的更广泛作用以及这一活动如何转化为新的学习，关键问题仍然存在。这项研究计划将采用多学科方法。我们将使用一系列专门为操纵预测误差而设计的复杂行为任务。这些行为方法将与最先进的神经科学方法相结合，这些方法记录(体内电生理学)和操纵(光遗传学；药物治疗)正在进行的行为任务中的神经活动，具有其他方法无法超越的时间和空间精确度，以定义NA在信号传递、预测误差和驱动学习中的作用。这项工作将引起广泛的兴趣，因为它解决了关于动机、唤醒和强化学习的基本问题。这个项目意义重大，因为建立和更新关于世界事件之间关系的知识是适应行为的基本能力，而未能灵活地修改以前学习到的反应是包括痴呆症、抑郁症、药物成瘾和注意缺陷多动障碍(ADHD)在内的一些神经精神疾病的核心。要理解这种病理，需要深入了解健康大脑中学习和注意力的神经基础。如果NA在一般错误驱动的学习或在条件改变时更新学习的具体过程中发挥重要作用，那么这对我们对学习的神经生物学的概念有重大影响。此外，该研究计划提供了一个多学科的培训环境，为博士后研究员、研究生和本科生提供特定的角色，并为受训人员提供大量机会来发展广泛的研究技能，从问题和假设的概念、实验设计到数据收集和解释，最后是手稿准备。

Predictive learning is highly adaptive; it allows us and other animals to use information in the environment to anticipate and prepare for biologically significant events. Equally important is the ability to update previous learning when environmental conditions change, and the organism's predictions about impending events are no longer accurate. This project examines the role of the neurotransmitter noradrenaline (NA) in this particular form of error-driven learning. Detecting and encoding relationships between events and, in particular, modifying previous learning when confronted with new information is essential for adaptive, flexible control of behaviour. However, we still know very little about the neural systems that make this possible. While substantial research implicates dopamine (DA) in initial learning about rewarding events, recent research conducted by Dr. Corbit has shown that updating this learning, notably in extinction (when expected reinforcement is omitted), involves noradrenaline (NA). Critical questions remain regarding the broader role of NA in in signaling reward prediction errors and how this activity translates into new learning. This research program will adopt a multidisciplinary approach. We will use a battery of sophisticated behavioural tasks specifically designed to manipulate prediction error. These behavioural methods will be combined with state of the art neuroscience methods that record (in vivo electrophysiology) and manipulate (optogenetics; pharmacological treatments) neural activity during ongoing behavioural tasks with temporal and spatial precision unsurpassed by other methods to define the role of NA in signaling prediction error and driving learning. This work will be of wide interest because it addresses basic questions about motivation, arousal and reinforcement learning. This project is significant because establishing and updating knowledge about the relationship between events in the world is a fundamental capacity essential for adaptive behaviour and failure to flexibly modify previously learned responses is central to a number of neuropsychiatric diseases including dementia, depression, drug addiction and attention deficit hyperactivity disorder (ADHD). To understand pathology of this kind requires insight into the neural bases of learning and attention in the healthy brain. If NA plays a major role in either error-driven learning in general or specifically in the updating of learning when conditions change, then this has major implications for our conception of the neurobiology of learning. In addition, this research program provides a multidisciplinary training environment, with specified roles for postdoctoral fellows, graduate students and undergraduate students and numerous opportunities for trainees to develop a wide range of research skills from the conception of the question and hypothesis, design of experiments, to data collection and interpretation, and finally manuscript preparation.