A Software Platform for Sensor-based Movement Disorder Recognition

基于传感器的运动障碍识别软件平台

• 批准号: 9321913
• 负责人: Gianluca De Luca
• 金额: $ 56.65万
• 依托单位: ALTEC, INC.
• 依托单位国家: 美国
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-09-30 至 2018-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/9321913
• 关键词: Accelerometer; Activities of Daily Living; Affect; Algorithms; American; Bradykinesia; Calibration; Cerebral Palsy; Clinical; Complement; Complex; Computer software; Data; Databases; Development; Devices; Disease; Dyskinetic syndrome; Dystonia; Effectiveness; Enhancement Technology; Essential Tremor; Freezing; Future; Gait; Gait abnormality; Generic Drugs; Goals; Guidelines; Health Care Sector; Home environment; Hybrids; Individual; Involuntary Movements; Kinetics; Laboratories; Lead; Left; Limb structure; Measurement; Miniaturization; Modality; Monitor; Morphologic artifacts; Motor; Movement; Movement Disorders; Muscle; Nature; Neurologic; Paper; Parkinson Disease; Patient Self-Report; Patients; Phase; Population; Posture; Process; Productivity; Reporting; Research; Research Personnel; Resolution; Restless Legs Syndrome; Risk; Rotation; Scientist; Severities; Surface; System; Tablets; Techniques; Technology; Testing; Time; Training; Tremor; Update; Walking; Wireless Technology; base; brain behavior; commercialization; data acquisition; data management; design; handheld mobile device; human subject; improved; instrument; motor symptom; nervous system disorder; new technology; novel; patient monitoring device; portability; prospective; protocol development; prototype; public health relevance; sensor; signal processing; soft tissue; systems research; tool; usability

 DESCRIPTION (provided by applicant): The goal of this Phase II is to enhance the availability of advanced brain and behavior research tools [PA-14-250] by developing an automated sensor-based means of tracking the presence and severity of a broad spectrum of movement disorders during unscripted activities of daily living. The continuously updated and interpreted information from body-worn sensors will provide accurate, objective, and high resolution (1 s.) measurement of motor symptom severity of tremor, dyskinesia, bradykinesia, freezing and gait disorders in Parkinson's disease and postural/kinetic tremor in essential tremor. It will allow researchers to assess the oftentimes complex and dynamic nature of movement disorders, which is poorly captured by the current standard of self-reports and pencil-and-paper instruments. Advances in wearable sensor technology have facilitated such a solution, but there are currently no movement disorder recognition devices capable of interpreting sensor data from non- scripted activity in an effective manner for the more than 45 million people in the U.S. with movement disorders. Our approach is unique in that we are developing a generic Application Generator (AG) software platform containing signal processing modules that can be readily configured to provide automated recognition for different disorders without the need to prepare separate algorithms from scratch for each. Phase I established a proof of concept by developing a rudimentary AG platform that achieved automatic recognition of tremor, dyskinesia and freezing-of-gait in patients with Parkinson's disease (PD) from novel hybrid sensors that provided both muscle activity and movement data through surface electromyographic (sEMG) and accelerometer recordings. Phase II will continue the development to include a broader range of PD movement disorders, as well as other neurological conditions. Aim 1 will create an enhanced AG Platform by incorporating combined sEMG and inertial measurement unit (IMU) sensors to more completely describe involuntary movements and reduce the risk when tracking additional disorders. Human subject testing will provide a sensor database for testing IMU sensor accuracy and minimizing soft tissue artifacts. The Phase I recognition algorithms will be updated using the enhanced platform. Aim 2 will use the enhanced platform to develop new recognition applications that track bradykinesia and gait disorders in PD, and postural and kinetic tremors in patients with essential tremor. Our goal is to achieve error rates < 5% during unconstrained monitoring conditions with user- independent algorithms. Aim 3 will deliver a portable pre-commercial device with the requisite hardware, software, user interface, and report generator to effectively monitor PD, essential tremor, and sitting/standing/walking activity. The system will collect and process sEMG/IMU data using a tablet PC to enhance usability. Movement disorder experts and prospective end-users will guide the Phase II development and assist us with future commercialization plans for other neurological conditions such as cerebral palsy, dystonia, ALS, and restless leg syndrome. It will also form the basis for a patient-operable device for clinical use.

 描述（由申请人提供）：第二阶段的目标是通过开发一种基于传感器的自动化方法来跟踪日常生活活动中广泛运动障碍的存在和严重程度，从而提高先进大脑和行为研究工具 [PA-14-250] 的可用性。来自身体佩戴传感器的不断更新和解释的信息将为帕金森病中的震颤、运动障碍、运动迟缓、冻结和步态障碍以及原发性震颤中的姿势/运动性震颤的运动症状严重程度提供准确、客观和高分辨率（1秒）的测量。它将使研究人员能够评估运动障碍通常复杂和动态的性质，而当前的自我报告和纸笔工具标准很难捕捉到这一点。可穿戴传感器技术的进步促进了这样的解决方案，但目前还没有运动障碍识别设备能够为美国超过 4500 万患有运动障碍的人有效地解释来自非脚本活动的传感器数据。我们的方法是独特的，因为我们正在开发一个通用应用程序生成器（AG）软件平台，其中包含信号处理模块，可以轻松配置这些模块以提供对不同疾病的自动识别，而无需从头开始为每种疾病准备单独的算法。第一阶段通过开发一个基本的 AG 平台建立了概念验证，该平台通过新型混合传感器实现了对帕金森病 (PD) 患者的震颤、运动障碍和步态冻结的自动识别，这些传感器通过表面肌电图 (sEMG) 和加速度计记录提供肌肉活动和运动数据。第二阶段将继续开发，以涵盖更广泛的 PD 运动障碍以及其他神经系统疾病。目标 1 将通过整合 sEMG 和惯性测量单元 (IMU) 传感器来创建增强型 AG 平台，以更完整地描述不自主运动并降低跟踪其他疾病时的风险。人体测试将提供传感器数据库，用于测试 IMU 传感器精度并最大限度地减少软组织伪影。第一阶段的识别算法将使用增强的平台进行更新。 Aim 2 将使用增强的平台开发新的识别应用程序，用于跟踪 PD 中的运动迟缓和步态障碍，以及原发性震颤患者的姿势和运动性震颤。我们的目标是在不受约束的监控条件下使用独立于用户的算法实现错误率 < 5%。 Aim 3 将提供一款便携式预商用设备，配备必要的硬件、软件、用户界面和报告生成器，以有效监测 PD、特发性震颤和坐/站/行走活动。该系统将使用平板电脑收集和处理 sEMG/IMU 数据，以增强可用性。运动障碍专家和潜在的最终用户将指导二期开发，并协助我们制定针对脑瘫、肌张力障碍、ALS 和不宁腿综合征等其他神经系统疾病的未来商业化计划。它还将构成患者可手术的基础 供临床使用的设备。