CAREER: A Closed-Loop Control Framework for the Treatment of Chronic Stroke

在美国，每年有超过80万人患中风。这些人中的许多人患有偏瘫，或身体一侧无力。中风后早期，这种偏瘫导致疼痛和力量下降，使任务执行困难。同时，运动能力和神经功能也可能自发恢复。然而，由于这种早期肢体无力，人们最终自愿抑制使用较弱的肢体。因此，出现了一种被称为“不使用”的现象：人们可以做什么和他们选择做什么之间的差异。这是有问题的，因为不使用被认为会导致补偿性运动策略，从而导致损伤增加和额外的医疗并发症。随着时间的推移，当个人在没有临床医生监控的情况下使用自己的运动策略时，不使用是许多现象中的一个例子。目前研究方法的目标是自动测量人们何时使用这种有害的策略，并通过数字设备的反馈来鼓励增加肢体的使用。该系统是一种即时自适应干预（JITAI），是一种基于技术的工具，能够实时确定人们何时进行某些活动，并在他们需要的时间和地点向他们提供反馈，以促进恢复。这种方法可能是针对慢性疾病患者的许多干预措施中的第一个，它利用了被称为控制系统工程的成熟科学原理。该方法将模拟人类的自然行为，并通过使用控制系统策略进行适当的干预。这些研究目标与教育目标紧密结合，因为控制系统工程可以教授给不同学术成熟度的学生。因此，工程和医疗保健专业的不同学生将利用本科阶段的暑期培训项目来评估和模拟技术方法，并通过研究生阶段的新课程将计算和神经科学结合起来。最后，这项研究将代表一个概念的证明，即使用自动化工具来适应人类提供的循证治疗，可以在现实世界中实施慢性健康状况。研究者的激励研究主题是环境辅助生活：在临床或实验室环境之外的现实环境中使用辅助技术来帮助残疾人。针对这一主题，本项目侧重于开发和评估一个将控制系统工程和神经康复相结合的框架，为患有习得性失用的慢性中风人群提供家庭康复。研究计划提出了一种解决神经退行性疾病（NDs）的新方法，将症状视为动力系统的输出。通过将患者视为控制回路中的“系统”，经典控制的应用便于使用反馈（监测患者症状）和控制器（向患者提供输入）将症状驱动到所需状态。该框架促进了对疾病症状进行客观量化和建模的方法学和理论上可辩护的方法，以及为减轻此类症状而进行干预的循证战略。研究计划有三个目标。第一个目标是建立一个中风系统进展的动力系统模型。该模型包括一个不使用前向障碍，描述康复信念如何导致行为，一个感觉运动学习障碍，描述家庭实践行为与现实世界中自发肢体使用之间的关系，以及一个使用自我调节理论将自发肢体使用与一个人对自己能力的信念联系起来的不使用反馈障碍。第二个目标是发展治疗的动力系统模型。该模型包括一个CIMT（约束诱导运动治疗）传递包块，它将治疗信念与CIMT干预成分（例如，动机，强制肢体使用和正强化）联系起来，以及一个使用SMC（滑模控制）方法的基于控制的康复块。非侵入式可穿戴传感器和从慢性偏侧中风患者获得的功能评估数据将用于校准Nonuse和CIMT模型。第三个目标是在参与者家中开发和验证基于非线性控制系统的治疗交付。在为期4周的研究之前和之后，将使用UE-FMA（上肢逃亡-迈耶评估）评估疗效，在此期间，传感器将每天佩戴6小时。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Each year, over 800,000 people suffer from stroke in the United States. Many of these individuals are left with hemiparesis, or weakness on one side of the body. Early after the stroke, this hemiparesis results in pain and reduced strength, rendering task performance difficult. Meanwhile, there may be spontaneous recovery of movement ability and neurological capability. However, due to this early limb weakness, people eventually voluntarily suppress the use of the weaker limb. Thus, a phenomenon emerges known as nonuse: a difference between what people can do and what they choose to do. This is problematic as nonuse is thought to lead to compensating movement strategies, which lead to increased injury and additional medical complications. Nonuse is one example of many phenomena that develop over time as individuals use their own movement strategies when they are not monitored by a clinician. The goal of the current research approach is to automatically measure when people utilize such harmful strategies and to encourage increased limb use through feedback from a digital device. The system is a just-in-time adaptive intervention (JITAI), a technology-based tool capable of determining in real-time when people are performing certain activities and providing them feedback when and where they need it in order to promote recovery. This approach may represent the first of many interventions for people with chronic conditions that develop over time and takes advantage of a well-established scientific principle known as control systems engineering. The approach will model the natural behavior of the human and apply appropriate intervention through the use of control systems strategies. These research goals are closely coupled with educational goals, as control systems engineering can be taught to students at varying levels of academic maturity. As a result, diverse students in both engineering and healthcare will take advantage of summer training programs at the undergraduate level designed to evaluate and simulate the technological approach and through a novel curriculum focused on combining computation and neuroscience at the post-graduate level. Finally, this research will represent a proof of concept for the adaptation of human delivered, evidence-based therapy using automated tools that can be implemented in real world settings for chronic health conditions.The investigator's motivating research theme is ambient assisted living: the use of assistive technologies in real-world settings outside of clinical or laboratory environments to assist people living with disability. Toward this theme, this project focuses on developing and evaluating a framework combining control systems engineering and neurorehabilitation to provide in-home rehabilitation for the chronic stroke population suffering from learned nonuse. The Research Plan presents a novel approach to addressing neurodegenerative disorders (NDs) by treating the symptoms as outputs of a dynamical system. By treating the patient as the 'system' within a control loop, the application of classical control facilitates the use of feedback (to monitor patient symptoms) and a controller (to provide inputs to the patient) to drive symptoms to a desired state. The framework facilitates a methodological and theoretically defensible approach to objective quantification and modeling of disease symptoms, and evidence-based strategies for intervening to mitigate such symptoms. The Research Plan is organized under three aims. The FIRST AIM is to develop a dynamical systems model of stroke system progression. The model includes a Nonuse forward block that describes how behavior results from beliefs about rehabilitation, a Sensorimotor Learning block that describes the relationship between home practice behavior and spontaneous limb use in real world settings and a Nonuse feedback block that relates spontaneous limb use to a person's beliefs about their capability using self-regulation theory. The SECOND AIM is to develop a dynamical systems model of therapy. The model includes a CIMT (constraint induced movement therapy) Transfer Package block that relates beliefs about therapy to the CIMT intervention components (e.g., motivation, forced limb use and positive reinforcement) and a Control-Based Rehabilitation block that uses an SMC (sliding mode control) approach. Noninvasive wearable sensor and functional assessment data obtained from individuals with chronic hemiparetic stroke will be used to calibrate the models of Nonuse and CIMT. The THIRD AIM is to develop and validate the nonlinear control systems-based treatment delivery in participants' homes. Efficacy will be evaluated using the UE-FMA (Upper Extremity Fugi-Meyer Assessment) before and after 4 week studies during which sensors will be worn 6 hours/day.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.