CAREER: Neuromechanics of human-robot interaction via robot-assisted in-vivo imaging of neuromuscular function

机器人在神经康复、性能增强、机器人辅助制造和动力假肢等方面越来越多地被用来支持人类的运动功能。虽然在过去20年中，技术能力的提高令人印象深刻，但由于对人-机器人辅助运动功能过程中人的组成部分缺乏基本了解，进步一直受到限制。这个职业项目将通过非侵入性地测量大脑功能、肌肉力量和僵硬来解决这一限制，同时受试者与机器人交互，在磁共振扫描仪内执行手腕任务。这一框架将用于研究神经运动障碍如何影响肌肉协调，为机器人辅助运动功能的精细神经肌肉骨骼模型以及用于提高表现和神经康复的个性化人-机器人交互策略铺平道路。该项目将为研究生和本科生提供生物力学、神经科学和机器人交叉问题方面的培训，帮助创建一支随时准备应对学术界和工业界康复工程研究的复杂挑战的多学科劳动力队伍。开发的计算模型将成为一个新的推广计划的基础，该计划将在神经肌肉建模的背景下引入游戏，该计划将应用于初中生和高中生，以提高他们接受STEM教育的兴趣。研究人员的长期研究目标是开发新的以神经科学为基础的干预措施，用于神经运动障碍后的运动康复。为了实现这一目标，研究人员目前的研究集中在1)开发和验证用于人类辅助和康复的机器人技术，以及2)将这些新技术用于模拟健康和受损的人类运动控制的实验研究。这个职业项目寻求使用与MRI兼容的机器人，在需要受试者在MR扫描仪中执行动态点到点和等长腕部任务的过程中，通过功能磁共振成像(FMRI)非侵入性测量大脑功能，通过磁共振弹性成像(MRE)测量肌肉功能。将神经力学建模与先进的实验方法和假设驱动的实验相结合，该框架将用于量化人体内在的神经肌肉动力学如何影响人与人的物理交互，并研究在现实神经肌肉动力学存在的情况下人-机器人交互的最优控制策略。该项目建立在研究人员成功开发了一系列与MRI兼容的机器人设备和多肌肉磁共振弹性成像(MM-MRE)成像技术以确定肌肉协调的基本标准的基础上。这项研究计划分为两个阶段。第一个重点是通过结合建模、行为实验和神经成像的多方面方法来研究强调力和阻抗控制的神经机制。将开发一个计算框架，用于分析手腕指向运动的力和阻抗控制，其中包括准确的肌肉骨骼动力学。这一框架将在健康受试者身上进行的实验中得到验证，这些受试者学习在稳定或不稳定的任务中控制力或阻抗。将确定学习产生任务执行所需的适当力和阻抗的大脑区域。一项使用功能磁共振成像和与磁共振成像兼容的手腕机器人的神经成像实验将被用来测试关于动态任务中力和阻抗的神经表示的假设，具体地说，“力和阻抗在不同的皮质区域受到控制”，以及“肌肉共同激活的表示与快速学习状态的表示重叠。”第二个重点是使用MM-MRE来估计前臂多个肌肉的力学性质。这项新技术将肌肉特异性MRE测量与通过与MRI兼容的仪器手柄获得的关节扭矩测量相结合。这项技术将用于确定适用于手和手腕任务的肌肉协调标准，并量化这些标准在中风后个体中的变化。基于获得的独特测量，目前被接受的短程刚度模型尚未直接在体内和人体内进行验证，将使用高功率MRI兼容机器人应用的机器人扰动进行测试。该奖项反映了NSF的法定使命，并已通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Robots are used more and more to support human motor function in neurorehabilitation, performance augmentation, robot-assisted manufacturing, and powered prosthetics. While improvement in technological capabilities has been impressive during the last 20 years, progress has been limited by poor fundamental understanding of the human component during human-robot-assisted motor function. This CAREER project will address this limitation by noninvasively measuring brain function and muscle force and stiffness while a subject interacts with a robot to perform wrist tasks within a magnetic resonance scanner. This framework will be used to study how neuromotor impairment affects muscle coordination, paving the way for refined neuromusculoskeletal models of robot-assisted motor function and for personalized human-robot interaction strategies for performance augmentation and neurorehabilitation. This project will provide training for graduate and undergraduate students in problems at the intersection between biomechanics, neuroscience, and robotics, helping create a multidisciplinary workforce ready to tackle the complex challenges of rehabilitation engineering research in both academia and industry. The computational model developed will form the basis of a new outreach program that introduces gaming within the context of neuromuscular modeling, which will be administered to middle and high school students to raise interest in pursuing STEM education.The investigator’s long-term research goal is to develop novel neuroscience-grounded interventions for movement rehabilitation after neuromotor disorders. Toward this goal, the investigator’s current research is focused on 1) developing and validating robotic technologies for human assistance and rehabilitation and on 2) using such novel technologies in experimental studies modeling the healthy and impaired human motor control. This CAREER project seeks to use MRI-compatible robots to non-invasively measure brain function via functional magnetic resonance imaging (fMRI) and muscle function via magnetic resonance elastography(MRE) during motor tasks that require subjects to perform dynamic point-to-point and isometric wrist tasks in an MR scanner. Combining neuromechanical modeling with advanced experimental methods and with hypothesis-driven experiments, this framework will be used to quantify how the intrinsic neuromuscular dynamics of the human body affect physical human-interaction and to study optimal control policies for human-robot interaction in the presence of realistic neuromuscular dynamics. The project builds on the investigator’s success in developing a family of MRI-compatible robotic devices and a multi-muscle magnetic resonance elastography (MM-MRE) imaging technique for identifying fundamental criteria of muscle-coordination. The research plan is divided into two thrusts. The FIRST thrust focuses on studying the neural mechanisms underscoring force and impedance control using a multifaceted approach that combines modeling, behavioral experiments, and neuroimaging. A computational framework will be developed for analyzing force and impedance control for wrist pointing movements that includes accurate musculoskeletal dynamics. This framework will be validated in experiments conducted on healthy subjects that learn to control force or impedance in stable or unstable tasks. Brain regions involved in learning to generate proper force and impedance required for task execution will be identified. A neuroimaging experiment that uses fMRI and an MRI-compatible wrist robot will be used to test hypotheses on the neural representation of force and impedance during dynamic tasks, specifically that "force and impedance are controlled in different cortical regions" and that "the representation of muscle co-activation overlaps with the representation of the fast learning state." The SECOND thrust focuses on using MM-MRE to estimate the mechanical properties of multiple muscles in the forearm. The new technique integrates muscle-specific MRE measurements with measurements of joint torque obtained via an MRI-compatible instrumented handle. This technique will be used to identify criteria of muscle coordination that apply for tasks of the hand and wrist and to quantify how these criteria change in post-stroke individuals. Based on the unique measurements obtained, the currently accepted short-range stiffness model, not yet directly validated in-vivo and in-humans, will be tested using robotic perturbations applied by a high power MRI-compatible robot.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.