NeuroVision™: A Smartphone Application for Neurology and Telemedicine

点击翻译按钮获取中文摘要

The transformative shift in healthcare to in-home telehealth services is projected to deliver long term benefits beyond the COVID pandemic to vulnerable populations, and those underserved due to isolation from specialty care centers. For the neurologist, however, the shift away from in-person care has disrupted the ability to perform comprehensive motor assessments with current telehealth technologies. While video conferencing platforms have enabled some neurological evaluations of speech, facial, and upper body symptoms, most neurological disorders such as Parkinson’s disease (PD) require a comprehensive visual examination of different movement tasks and motor symptoms that are all too easily obstructed by the narrow field-of-view and constrained 2D visual display of standard telehealth video-conferencing interactions. To overcome these limitations and offer quantitative outcomes, our team of computer vision experts is partnering with leading neurologists to develop a remote means of recording and quantifying prescribed neurological motor assessment tasks in the home using a smartphone (3D) depth-sensing camera to inform a subsequent telehealth visit with video samples and corresponding numerical outcome metrics. While developments from our group and others in the area of computer vision have used total body pose estimation algorithms and stationary 3D cameras to measure joint mechanics, our pilot work demonstrates the feasibility of a new approach to derive neurological outcomes of gait, transfer, and fine motor assessment tasks directly from depth data recorded from a nonstationary (handheld) smartphone. In Phase I we will build upon this work to develop simultaneous localization and mapping algorithms in Aim 1 to obtain outcome measures of motor impairment from n=10 participants with PD that achieve +/- 5% error with respect to gold-standard motion capture. Aim 2 will add guided instructions, compliance feedback, and summary reports to the software. The resulting NeuroVision™ prototype will be clinically evaluated for usability and perceived value among n=5 independent neurologists and n=10 of their patients with PD. Smartphone-based recordings will be acquired by the patient’s care assistant or family member using the NeuroVision™ app in our laboratory. The resulting report will be provided to the referring neurologist, to be used to inform a scheduled telehealth visit. Our milestones are to achieve favorable Likert usability ratings (≥8/10) and high ratings of perceived clinical value and acceptance from the neurologists. Focus groups involving the patient and caregiver will provide additional feedback for a designing a more complete Phase II NeuroVision™ system that meets in-home use requirements, supports a broader range of common neurological conditions, and provides a secure eHealth assessment report. The pre-commercial prototype will be tested for usability and clinical value in informing telehealth follow-up visits to meet the needs of a broader population of at-risk patients with PD, Essential Tremor, Stroke, and Huntington’s disease currently in need of safe and reliable access to routine clinical care for health and quality of life.