A team from Mount Sinai Hospital has developed an AI tool to monitor infant movements in the Neonatal Intensive Care Unit (NICU). Using a deep learning algorithm, this AI can analyze video footage of infants, tracking their movements to find important neurologic signs. This tool may lead to continuous, minimally invasive monitoring in NICUs, offering real-time insights into infant health.
Each year, over 300,000 newborns are admitted to NICUs across the U.S. Infant alertness often indicates overall brain health, but traditional neurologic monitoring has been limited. Unlike heart and lung function monitoring, neurologic status checks in NICUs are often done through physical exams, which can miss subtle changes. Mount Sinai’s team believes their “Pose AI” system, trained to track specific body points from video, could help fill this gap. This technology has previously been successful in fields like sports and robotics.
The team trained Pose AI using millions of seconds of video from 115 infants. The AI accurately tracked the infants’ movements and could identify two conditions—sedation and cerebral dysfunction—with high accuracy. The study also showed that the AI worked effectively under different lighting conditions and was reliable across various angles.
Though some NICUs already use video cameras, few apply AI to analyze this footage. The Mount Sinai team hopes this tool can allow quicker interventions when neurologic changes occur, potentially improving patient outcomes. However, more testing is needed across other hospitals and with various video equipment.
Mount Sinai is committed to exploring AI tools to enhance patient care, from reducing hospital stays to supporting early diagnosis. The team plans to test Pose AI in additional NICUs and explore its application for other age groups and health conditions.
Source: news-medical
