This home-based system enables reliable quantification of motor symptoms in neurological disorders using short videos of patients performing simple movements recorded with a smartphone and machine learning algorithms. Motor symptoms like bradykinesia, tremor, and gait impairments are a hallmark of many neurological disorders, like Parkinson’s disease. Clinicians typically rely on in-person evaluations using standardized rating scales, which depend on subjective observation and are often limited by inter-rater variability and infrequent clinic visits.
Subtle motor changes, especially in early disease stages, can easily go undetected, delaying diagnosis and intervention. Objective measurement tools, such as motion sensors or wearable devices, are available but often require specialized hardware or technical expertise and can be costly or inconvenient for routine use.
Researchers have developed a digital health platform that leverages state of the art machine learning algorithms to transform standard consumer devices into clinically validated tools for quantifying motor symptoms in neurological disorders. This software achieves a level of accuracy and accessibility that surpasses traditional in-clinic assessments, with tiered classification models for nuanced disease staging. This makes it reliable and allows clinicians to have access to accurate assessments of motor symptoms often and easily. This platform also includes a user-friendly mobile application that provides real-time feedback to help patients perform the different movements, empowering clinicians and patients to manage neurological disorders more efficiently.
Clinically validated motor assessment software enabling clinicians, researchers, and patients to objectively detect, monitor, and track progression of motor symptoms in neurological disorders from any location, using only a standard video recording device, such as a smartphone
This tool for motor assessments utilizes a predictive data analysis engine that analyzes simple patient video recordings of standardized motor tasks, such as finger tapping or gait, and extracts precise features like amplitude, speed, and variability. The system then analyzes the motor data collected from patients to generate accurate outputs such as symptom severity scores and longitudinal progression metrics. A tiered classification algorithm, designed to reflect the actual progression of disease, distinguishes between healthy, early, moderate, and advanced states, using the most relevant movement features at each stage. By employing commonly available devices such as smartphones, this tool provides an accessible, easy-to-use platform that can be used by clinicians and patients to manage neuromuscular disorders. Validated against expert clinical ratings, the assessment tool achieves high accuracy in detecting even the earliest signs of disease, often before symptoms are clinically apparent. The software, VisionMD, is open-source, cross-platform, and processes all data locally for privacy. The software’s ability to support rapid, automated analysis and output clinically meaningful scores makes it a valuable tool for early detection, therapy monitoring, and longitudinal disease tracking.