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: 8734495
• 负责人: Gianluca De Luca
• 金额: $ 41.79万
• 依托单位: ALTEC, INC.
• 依托单位国家: 美国
• 财政年份: 2013
• 资助国家: 美国
• 起止时间: 2013-09-15 至 2015-08-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/8734495
• 关键词: Affect; Algorithms; American; Bradykinesia; Clinical; Code; Communities; Complement; Complex; Computer software; Custom; Data; Development; Disease; Dyskinetic syndrome; Dystonia; Early Diagnosis; Essential Tremor; Freezing; Future; Gait; Generations; Goals; Health Care Sector; Health Professional; Home environment; Involuntary Movements; Lead; Learning Module; Left; Lower Extremity; Machine Learning; Methods; Metric; Miniaturization; Monitor; Motor; Movement Disorders; Nature; Neurologic; Neurology; Outcome; Parkinson Disease; Parkinsonian Disorders; Patient Monitoring; Patient Self-Report; Patients; Performance; Persons; Phase; Population; Process; Productivity; Quality of life; Questionnaires; Research; Research Infrastructure; Research Personnel; Scientist; Severities; Signal Transduction; Small Business Innovation Research Grant; Solutions; Surface; System; Technology; Technology Transfer; Testing; Time; Training; Tremor; Video Recording; Wireless Technology; Work; Writing; advanced system; base; brain behavior; care delivery; clinical care; cost; data acquisition; design; disorder control; effectiveness research; flexibility; handheld mobile device; human subject; improved; innovation; knowledge base; motor disorder; nervous system disorder; novel therapeutics; prototype; public health relevance; research study; response; sensor; signal processing; software development; tool

DESCRIPTION (provided by applicant): The overall objective of this SBIR project is to develop a pre-commercial prototype system capable of continuously monitoring involuntary movement disorders from a wide spectrum of neurological conditions. The impact of this innovation will enhance the availability of advanced brain and behavior research tools [PA- 11-134] by providing a continuous means of tracking the presence and severity of movement disorders during normal daily activities. This project will transform our unique movement disorder recognition algorithms into custom software that analyzes movement disorders for specific neurological conditions. The information obtained from body worn sensors will provide an accurate and objective means for assessing the complex and changeable nature of movement disorders. This goal cannot by realized using the current method of self-report questionnaires. The research strategy for Phase I will establish the merit and feasibility of this effort by developing an Application Generation (AG) software platform using a framework of configurable signal processing modules to generate custom applications for movement disorder analysis (Aim 1). This approach reduces the effort and enhances the flexibility of designing and testing software solutions for these applications. The AG Platform will be developed using C++ software to implement signal processing and machine-learning software modules that operate within a knowledge-based framework that we have previously developed. In Aim 2 we will utilize the AG platform to generate movement disorder analysis software to evaluate a challenging test-case application: freezing-of-gait in Parkinson's disease (PD). The goal is to attain performance metrics for freezing that are comparable to those we have achieved for tremor and dyskinesia in previous efforts. Phase II will refine the capabilities of the AG platform developed in Phase I. We will augment it with the means to automatically design and train the machine learning algorithms, improve the user-interface, and provide options for viewing and summarizing the results. The improved AG platform will be used to develop customized disorder-analysis software that encompasses the full complement of movement disorders associated with PD (e.g. dystonia, bradykinesia, Parkinsonian gait, tremor, dyskinesia), as well as for other neurological condition, such as Essential Tremor (ET). Firmware will be developed for each custom application to efficiently integrate the analytic software with our existing Trigno wireless sensor data acquisition hardware, which needs to be streamlined for this application. This combined system will be evaluated under research use-case scenarios in Neurology. Phase II will deliver an ambulatory Movement Disorder Monitoring system that not only succeeds in providing state-of-the-art monitoring solutions for PD and Essential Tremor, but has proven technology to develop monitoring solutions for a wide variety of neurological conditions. Future development will transfer this technology to a clinical version of this system.

描述（由申请人提供）：该 SBIR 项目的总体目标是开发一种预商业原型系统，能够持续监测各种神经系统疾病中的不自主运动障碍。这项创新的影响将通过提供连续的方法来跟踪正常日常活动期间运动障碍的存在和严重程度，从而增强先进的大脑和行为研究工具的可用性[PA-11-134]。该项目将把我们独特的运动障碍识别算法转变为定制软件，用于分析特定神经系统疾病的运动障碍。从身体佩戴的传感器获得的信息将为评估运动障碍的复杂性和多变性提供准确和客观的手段。目前的自我报告式问卷调查方法无法实现这一目标。 第一阶段的研究策略将确定该研究的优点和可行性 通过使用可配置信号处理模块框架开发应用程序生成 (AG) 软件平台来生成用于运动障碍分析的自定义应用程序（目标 1）。这种方法减少了为这些应用程序设计和测试软件解决方案的工作量并增强了灵活性。 AG 平台将使用 C++ 软件开发，以实现信号处理和机器学习软件模块，这些模块在我们之前开发的基于知识的框架内运行。在目标 2 中，我们将利用 AG 平台生成运动障碍分析软件来评估具有挑战性的测试用例应用：帕金森病 (PD) 中的冻结步态。我们的目标是获得与我们之前针对震颤和运动障碍所取得的性能指标相当的冻结性能指标。 第二阶段将完善所开发的 AG 平台的功能 在第一阶段。我们将通过自动设计和训练机器学习算法、改进用户界面并提供用于查看和总结结果的选项来增强它。改进后的 AG 平台将用于开发定制的疾病分析软件，该软件涵盖与 PD 相关的所有运动障碍（例如肌张力障碍、运动迟缓、帕金森步态、震颤、运动障碍），以及其他神经系统疾病，例如基本功能障碍震颤（ET）。将为每个定制应用程序开发固件，以有效地将分析软件与我们现有的 Trigno 无线传感器数据采集硬件集成，该硬件需要针对该应用程序进行简化。该组合系统将在神经病学研究用例场景下进行评估。第二阶段将提供动态运动障碍监测系统，该系统不仅成功地为帕金森病和特发性震颤提供最先进的监测解决方案，而且拥有经过验证的技术，可以为各种神经系统疾病开发监测解决方案。未来的开发将将该技术转移到该系统的临床版本。