Artificial Intelligence Monitoring of Neurological Status From Patient Videos in the Neuroscience Intensive Care Unit

Le résumé fourni par la source

BACKGROUND AND OBJECTIVES: The neurological examination is pivotal in assessing patients with neurological conditions but has severe limitations: It can vary between examiners, may not discern subtle or subacute changes, and being intermittent can delay recognition of new deficits. Most intensive care units/hospitals lack subspecialized neurocritical care services, exacerbating these problems. We hypothesized that artificial intelligence (AI) pose estimation, a machine learning approach to track patient position, could provide a continuous and relevant method of neurological monitoring. METHODS: We retrospectively collected video segments from patients in the neuroscience intensive care unit (NSICU) who underwent video-electroencephalography at a large, urban hospital between July, 2024 and January, 2025. We externally validated 2 leading AI pose estimation models, ViTPose and Meta Sapiens. We then developed a robust movement index and evaluated its correlation with 2 measures of consciousness obtained through hourly physical examinations, the Glasgow Coma Scale (GCS), and Richmond Agitation Sedation Scale (RASS). RESULTS: We collected 998 520 video minutes from 119 patients. ViTPose demonstrated superior performance to Sapiens across multiple metrics, so we used ViTPose to calculate a computer vision movement index (λ MI ). We observed higher movement with increasing GCS (GCS 3-8 λ MI = 0.52, GCS 9-13 λ MI = 0.70, GCS 14 λ MI = 3.52, and GCS 15 λ MI = 10.99, P = .01), a 21-fold increase from the lowest to highest tranche. We also observed 10-fold higher movement in awake/agitated patients (RASS >-1 λ MI = 6.59) compared with those who were asleep/sedated (RASS ≤ -1 λ MI = 0.67, P = .005). Taken together, we developed a novel computer vision movement index and demonstrated expected correlations with GCS and RASS scores in NSICU patients. CONCLUSION: We show that AI pose estimation can provide minimally invasive, continuous, and clinically relevant neuromonitoring in critically ill patients. Neurological conditions account for the highest global disease burden and AI pose estimation may be a low-cost, explainable, and scalable AI solution to address this pressing need for neuro-telemetry.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.