Mechanisms underlying the control and generalization of sensorimotor learning

为了科普不断变化的环境，人类大脑已经进化出一种令人印象深刻的能力，可以灵活地调整运动指令，从一个动作到下一个动作。这种学习是任何智能系统的标志，也是人类认知和行为的核心。然而，支持感觉运动学习和更新的机制，以及使获得的技能容易转移到新的环境中，仍然知之甚少。本文提出的研究计划的主要目标是：（1）更好地理解个体学习感觉运动任务的机制和策略;（2）探索如何改善和更好地推广这种学习;（3）更好地理解支持感觉运动学习和更新的神经回路。拟议的研究包括三个独立的项目。 项目一： 许多成功的技能的一个重要特征是它们可以推广到新的环境中。然而，这种泛化可能被最大化的方式却很少受到关注，并且对工作场所，音乐和体育训练具有重要的实际好处。项目1，通过操纵参与者的感官反馈，将确定如何将用一只手获得的技能最好地推广到另一只未经训练的手。此外，使用神经刺激工具，我们将评估这种概括背后的因果神经机制。 项目二： 社会往往认为，虽然要防止和补救信息的遗忘或丢失，但要练习和改进信息的保留。虽然这可能对我们的许多陈述性记忆是正确的（例如，记住家庭成员、事实、数字等），我们的运动系统有能力抑制某些类型的学习，这是至关重要的，这些学习可能会干扰整体任务的表现。项目2将检验这样一个假设：在与学习两个独立任务相关的干扰效应超过与学习其中每一个任务相关的增益的情况下，运动系统将智能地抑制学习价值较低的任务。 项目三： 运动学习经常使用适应任务进行检查，其中要求个体对视觉反馈中的不寻常力量或变化做出反应。然而，重要的学习也发生在标准的日常任务中，比如第一次获得举起物体的重量。我们最近表明，物体的重量信息，通过提升，是在腹侧视觉通路。然而，这种表征在物体运动相互作用中的因果作用仍然不清楚。项目3将使用功能性MRI和神经刺激工具来因果评估腹侧视觉通路在体重编码中的作用，并确定腹侧视觉区活动的扰动如何破坏额顶叶皮层中互连区域的活动。

To cope with a continuously changing environment, the human brain has evolved an impressive ability to flexibly adjust, from one movement to the next, its motor commands. Such learning is a hallmark of any intelligent system and is central to human cognition and behaviour. Yet, the mechanisms that support sensorimotor learning and updating, and which allow acquired skills to be readily transferred to new contexts, remains poorly understood. The broad aims of the research program proposed here are to (1) better understand the mechanisms and strategies through which individuals learn sensorimotor tasks, (2) explore how this learning can be improved and better generalized, and (3) better understand the neural circuitry that supports sensorimotor learning and updating. The proposed research consists of three separate projects. Project 1: An important feature of the success of many well-learned skills is that they generalize to new contexts. Yet, the ways in which such generalization might be maximized has received very little attention and has important practical benefits for workplace, musical and sports training. PROJECT 1, by manipulating participant sensory feedback, will determine how skills acquired with one hand can be best generalized to the other, untrained hand. In addition, using neurostimulation tools, we will assess the causal neural mechanisms underlying this generalization. Project 2: Society often takes the view that, whereas the forgetting or loss of information is to be prevented and remedied, the retention of information is to be practised and improved. While this may be true of many of our declarative memories (e.g., remembering family members, facts, figures, etc.), it is critical that our motor systems have the capacity to suppress certain types of learning that can interfere with overall task performance. PROJECT 2 will test the hypothesis that the motor system, in situations in which the interference effects associated with learning two separate tasks overrides the gains associated with learning each of those tasks, will intelligently suppress learning the less valued task. Project 3: Motor learning is often examined using adaptation tasks in which individuals are required to respond to unusual forces or changes in visual feedback. However, significant learning also occurs in standard everyday tasks, as when acquiring the weight of an object lifted for the first time. We recently showed that object weight information, gained through lifting, is represented in the ventral visual pathway. However, the causal role of this representation in object motor interactions remains unclear. PROJECT 3 will use functional MRI and neurostimulation tools to causally assess the role of the ventral visual pathway in weight encoding and determine how perturbing the activity of ventral visual areas disrupts the activity of interconnected regions in frontoparietal cortex.