Weather     Live Markets

A team of clinicians, scientists, and engineers at Mount Sinai developed a deep learning pose-recognition algorithm to track movements and identify neurologic metrics in infants in the neonatal intensive care unit (NICU). The AI-based tool, known as “Pose AI,” was trained on video feeds of infants in the NICU and published in Lancet’s eClinicalMedicine. This innovation could provide a minimally invasive and scalable method for continuous neurologic monitoring in NICUs, offering real-time insights into infant health that were previously unavailable.

Over 300,000 newborns are admitted to NICUs in the United States each year, making accurate neurologic monitoring crucial. Neurological deterioration in NICUs can occur unexpectedly and have devastating consequences. While cardiorespiratory telemetry is commonly used to monitor heart and lung function in NICU babies, neurotelemetry has been elusive despite years of work in EEG and specialized neuro-NICUs. Neurologic status is currently evaluated intermittently using imprecise physical exams that may overlook subtle changes, highlighting the need for a more continuous and accurate monitoring method.

The Mount Sinai team hypothesized that using a computer vision method like Pose AI to track infant movement could predict neurologic changes in the NICU. By training the AI algorithm on video footage from infants undergoing continuous video EEG monitoring, they were able to accurately track infant landmarks and predict critical conditions such as sedation and cerebral dysfunction. This approach has the potential to detect neurologic changes early, allowing for faster interventions and improved outcomes for infants in the NICU.

The researchers were surprised by the effectiveness of Pose AI across various lighting conditions and angles, as well as its association with gestational and postnatal age. They emphasized that this AI-based approach does not replace physician and nursing assessments in the NICU but rather enhances them by providing a continuous readout that can prompt intervention when necessary. The ultimate vision is to have cameras continuously monitoring infants in the NICU, with AI providing a neuro-telemetry strip similar to heart rate or respiratory monitoring, alerting clinicians to changes in sedation levels or cerebral dysfunction.

Although the study was conducted at a single institution, the Mount Sinai team plans to evaluate the algorithm and neurologic predictions on video data from other NICUs and institutions. They also aim to conduct clinical trials to assess the impact of this technology on care and explore its applications for other neurological conditions and adult populations. This effort aligns with Mount Sinai’s commitment to leveraging artificial intelligence to advance patient care, as AI tools have already shown promise in improving outcomes and enhancing clinical decision-making across various areas of healthcare.

In conclusion, the development of Pose AI represents a significant advancement in the field of neonatal neurology and NICU care. By harnessing the power of deep learning and computer vision, this technology has the potential to revolutionize how infant health is monitored and managed in NICUs, providing clinicians with valuable real-time insights and prompting early interventions that can ultimately improve outcomes for newborns. The Mount Sinai team’s innovative approach demonstrates the value of incorporating AI into healthcare settings to enhance patient care and highlights the importance of further research and validation to ensure the effectiveness and reliability of these tools in clinical practice.

Share.
Exit mobile version