BradyXplore: Bradykinesia Feature Extraction System

描述（由申请人提供）：目的是设计、实施和临床评估一种便携式、用户佩戴的慢动特征提取系统“BradyXplore”，用于与Great Lake Neurotechnologies （GLN）的Kinesia和Kinesia HomeView技术平台集成，用于客观监测帕金森病（PD）。目前对帕金森病的治疗进行了大量的研究，包括药物干预和深部脑刺激（DBS）。虽然震颤通常是最明显的症状，但运动迟缓可能是最有害的。目前评估运动迟缓的标准是统一帕金森病评定量表，这是一个主观的、定性的排名系统。运动迟缓的评分说明将多个运动特征整合到一个单一的评分中，这增加了可变性，限制了对特定运动迟缓特征是否受到特定治疗的影响的探索。如果能更好地量化不同的运动迟缓特征，可能会开发出针对特定运动迟缓表现的新疗法。提出的创新包括：1)紧凑的、用户穿戴的运动传感器，可以与基于网络的软件一起在家用电脑或平板电脑上使用；2)使用运动学数据来独立评估速度、幅度和节奏的算法；3)基于网络的应用程序，用于系统交付、患者互动和症状报告，所有这些都将促进临床接受和自我维持的商业模式。包含加速度计和陀螺仪的无线运动传感器单元将佩戴在手指和拇指上，并收集同步运动数据。BradyXplore将独立评估标准重复性运动任务的速度、幅度和节奏。传感器单元电池将通过感应式USB充电板充电，从而消除患者对小连接器的不安。当传感器单元被放置在垫子上时，运动学数据将通过无线连接传输到患者的电脑或平板电脑上，然后上传到我们符合hipaa的服务器上。在第一阶段，我们成功升级了现有的运动传感器单元，以减小尺寸并提高灵敏度。算法开发独立评级速度，幅度和节奏。算法输出的分数与临床评分高度相关，并显示出对药物治疗的速度、幅度和节奏的不同反应。第二阶段将改进人体工程学、灵敏度、数据传输和报告。硬件将升级为包括两个同步传感器单元，通过感应垫充电，并通过低功率无线电传输数据。软件将被修改为完全基于网络的，因此患者无需安装任何软件。BradyXplore算法将在临床研究中进一步研究，以证明其与临床评分相比的重测可靠性，并确定DBS的速度、幅度和节律波动是否存在差异。我们假设BradyXplore将1)为慢动症评估提供一个标准化的平台，客观地量化速度、幅度和节奏；2)获得临床医生和患者的高临床接受度；3)有助于开发、评估和优化DBS和药物干预等治疗方法。

DESCRIPTION (provided by applicant): The objective is to design, implement, and clinically assess a portable, user-worn, bradykinesia feature extraction system, BradyXplore", for integration with Great Lake Neurotechnologies' (GLN) Kinesia and Kinesia HomeView technology platforms for objective Parkinson's disease (PD) monitoring. There has been tremendous research into PD treatments including pharmaceutical interventions and deep brain stimulation (DBS). While tremor is often the most visible symptom, bradykinesia can be the most impairing. The current standard in evaluating bradykinesia is the Unified Parkinson's Disease Rating Scale, a subjective, qualitative ranking system. Scoring instructions for bradykinesia integrate multiple movement features into a single score that increases variability and limits exploration into if particular bradykinesia features are influenced by specific treatments. If different bradykinesia features were better quantified, novel therapies may be developed to target specific bradykinesia manifestations. The proposed innovations include 1) compact, user-worn motion sensors that can be used with web-based software on a home computer or tablet, 2) algorithms that use kinematic data to rate speed, amplitude, and rhythm independently, and 3) web-based applications for system delivery, patient interaction, and symptom reporting, all of which will promote clinical acceptance and a self- sustaining business model. Wireless motion sensor units containing accelerometers and gyroscopes will be worn on the finger and thumb and collect synchronized motion data. BradyXplore will independently rate speed, amplitude, and rhythm for the standard repetitive motion tasks. Sensor unit batteries will be recharged by an inductive, USB charge pad to eliminate the need for patients to fidget with small connectors. When sensor units are placed on the pad, kinematic data will be transferred to the patient's computer or tablet via a wireless link and then uploaded to our HIPAA-compliant server. In Phase I, our existing motion sensor unit was successfully upgraded to reduce size and improve sensitivity. Algorithms were developed for independently rating speed, amplitude, and rhythm. The algorithms output scores highly correlated to clinical ratings and demonstrated a differential response of speed, amplitude, and rhythm to medication. Phase II will improve ergonomics, sensitivity, data transfer, and reporting. Hardware will be upgraded to include two synchronized sensor units that recharge via an inductive pad and transfer data via a low-power radio. Software will be modified to be completely web-based so patients need not install any software. BradyXplore algorithms will be further investigated in a clinical study to demonstrate its test-retest reliability compared to clinical ratings and determine if speed, amplitude, and rhythm fluctuate differentially ON DBS. We hypothesize that BradyXplore will 1) provide a standardized platform for bradykinesia assessment that objectively quantifies speed, amplitude, and rhythm, 2) receive high clinical acceptance from both clinicians and patients, and 3) aid in development, evaluation, and optimization of therapies such as DBS and pharmaceutical interventions. PUBLIC HEALTH RELEVANCE: Parkinson's disease affects approximately 1.5 million people in the United States causing motor symptoms, of which one of the most debilitating is bradykinesia (slowed movements). Bradykinesia is currently evaluated using a subjective rating scale that gives a single score taking into account speed, amplitude, fatiguing, hesitations, arrests in movement, and how these variables change over time. The proposed BradyXplore bradykinesia feature extraction system will separately quantify specific features of bradykinesia (speed, amplitude, and rhythm) in the home, which should aid in the development of novel therapies to target a patient's specific bradykinesia manifestations.