EFFECTIVENESS OF ROBOT-ASSISTED HAND MOVEMENT TRAINING AFTER STROKE

项目概要/摘要 该项目的广泛、长期、科学目标是确定影响行为的行为和神经因素。 确定中风后机器人手运动训练的效果。实现的重要社会影响 这一目标将是设计更有效的机器人康复锻炼技术，这将使 中风患者的运动恢复速度超出了当前方法所能达到的水平。在我们的 在之前的资助中，我们发现手指本体感觉受损的参与者（用一种新颖的机器人量化方法 协议）并没有从机器人手指训练中获得功能上的好处。福利减少也与 体感系统损伤和异常激活。我们的工作假设是手指 本体感觉完整性是机器人辅助的门户，因为它允许这种辅助刺激 类似赫布的学习机制。在这个“赫布假设”的基础上，我们添加了另外两个假设： 通过对本体感觉受损的人进行有针对性的训练来提高本体感觉能力将增强 随后的机器人手指训练的有效性以及本体感觉完整性调节自发的自我感觉 正式培训之外的培训。我们将用新型机器人进行一系列实验来检验这些假设 手指训练装置（“FINGER”）可帮助参与者做出不同的握法来演奏音乐剧中的音符 类似于《吉他英雄》的电脑游戏。我们将招募手部运动缺陷至少六级的参与者 中风后数月，将手指本体感觉完好的参与者随机分配至目标 1，以及 那些手指本体感觉受损的人参加目标 2。目标 1 是确定赫布的大小 受益于机器人辅助运动训练。机器人辅助增强了本体感受输入，但通常 还可以提高奖励，因为它可以提高任务的成功率。我们将确定利益的大小 增强的本体感受输入，超出了由于增强的奖励而带来的好处。目标 2 是确定 通过有针对性的机器人本体感觉训练可以改善手指本体感觉，从而 增强对后续机器人手指运动训练的反应。目标 3 是确定数学模型 预测目标 1 和 2 中参与者对不同形式的培训的反应。模型 输入将是基线行为和人口统计测量、病变与感觉和运动结构的重叠 （通过解剖 MRI），以及感觉运动网络的激活（通过功能磁共振成像和脑电图）。我们的目标是提供进一步 洞察机器人训练反应的高差异，并确定谁对本体感受反应最好 培训。目标 4 是确定本体感觉受损在减少日常生活中自发用手使用方面的作用 使用新型可穿戴传感器，因为在机器人治疗过程之外用手使用可能会发挥强大的作用 训练效果。对这些假设的检验将深入了解如何操纵独特的资源 每个患者通过基于机器人的运动训练保留感觉运动可塑性。

Project Summary/Abstract The broad, long-term, scientific objective of this project is to identify the behavioral and neural factors that determine the efficacy of robotic hand movement training after stroke. An important societal impact of achieving this objective will be the design of more effective robotic rehabilitation exercise technology, which will allow people with a stroke to increase their movement recovery beyond that possible with current approaches. In our previous grant, we found that participants with impaired finger proprioception (quantified robotically with a novel protocol) did not achieve a functional benefit from robotic finger training. Reduced benefit also correlated with injury to and abnormal activation of the somatosensory system. Our working hypothesis is that finger proprioceptive integrity is a gateway for robotic assistance because it allows such assistance to stimulate a Hebbian-like learning mechanism. Building on this “Hebbian Hypothesis”, we add two other hypotheses: improving proprioceptive capacity with targeted training for those with impaired proprioception will enhance the effectiveness of subsequent robotic finger training, and proprioceptive integrity modulates spontaneous self- training outside of formal training. We will test these hypotheses in a series of experiments with a novel robotic finger training device (“FINGER”) that assists participants in making different grips to play notes in a musical computer game similar to Guitar Hero. We will recruit participants with hand movement deficits at least six months after stroke and randomize the participants with intact finger proprioception to participate in Aim 1, and those with impaired finger proprioception to participate in Aim 2. Aim 1 is to identify the magnitude of the Hebbian benefit from robot-assisted movement training. Robotic assistance enhances proprioceptive input, but it typically also enhances reward, because it increases task success. We will determine the magnitude of the benefit of the enhanced proprioceptive input, beyond the benefit due to enhanced reward. Aim 2 is to determine the extent to which finger proprioception can be improved through targeted robotic proprioceptive training, and thereby enhance response to subsequent robotic finger movement training. Aim 3 is to identify mathematical models that predict the response to the different forms of training experienced by the participants in Aim 1 and 2. The model inputs will be baseline behavioral and demographic measures, lesion overlap with sensory and motor structures (via anatomical MRI), and activation of sensory-motor networks (via fMRI and EEG). We aim to provide further insight into the high variance in robotic training response, and determine who responds best to proprioceptive training. Aim 4 is to identify the role of impaired proprioception in decreasing spontaneous hand use in daily life using a novel wearable sensor, because hand use outside of robotic therapy sessions likely exerts a powerful training effect. Testing of these hypotheses will provide insight into how to manipulate the unique resources for sensory motor plasticity that each patient retains using robot-based movement training.