Telehealth Neurologic Disease Diagnostic Tool

Accurate diagnosis of neuromuscular conditions like myasthenia gravis remains a challenge in telemedicine.  Clinicians often rely on video feeds that fail to capture subtle but critical symptoms such as ptosis (eyelid droop) and diplopia (double vision) leading to delayed or inaccurate assessments.

To overcome these limitations, researchers at The George Washington University have developed a novel AI-powered diagnostic platform that enhances virtual neurological exams. The system provides real-time analysis, quantitative analysis of ocular muscle function. The system uses a hybrid algorithm combining deep learning and computer vision. It detects and measures ptosis and ocular fatigue from standard video or still images—without requiring a controlled environment or specialized hardware. This innovation enables clinicians to remotely assess neuromuscular weakness with greater precision and automatically integrates findings into electronic medical records. In addition to diagnosing myasthenia gravis, the platform shows promise for evaluating a broad range of neurological conditions—including dementias, multiple sclerosis, stroke, and cranial nerve palsies—in both telehealth and in-person settings.

Figure: AI-based eye segmentation system for telehealth. Machine learning landmarks (red dots), vertical eye opening (blue arrow), and eye area (green) enable remote detection of ptosis and ocular fatigue.

Advantages
• Real-time, accurate detection of neuromuscular symptoms via telemedicine
• Compatible with standard video feeds—no specialized equipment needed
• Seamless integration with medical records
• Broad applicability across neurological disorders

Applications
• Remote diagnosis and monitoring of myasthenia gravis
• Telehealth evaluation of neurological diseases
• In-clinic support for cranial nerve and ocular muscle assessments

Publications

1. Garbey, M., Joerger, G., Lesport, Q., Girma, H., McNett, S., Abu-Rub, M., & Kaminski, H. (2023). A digital telehealth system to compute the Myasthenia Gravis core examination metrics. JMIR Neurotechnol, 2(1), e43387. https://doi.org/10.2196/43387
2. Lesport, Q., Joerger, G., Kaminski, H. J., Girma, H., McNett, S., Abu-Rub, M., & Garbey, M. (2023). Eye segmentation method for telehealth: Application to the Myasthenia Gravis physical examination. Sensors (Basel), 23(18), 7744. https://doi.org/10.3390/s23187744