Monitoring preterm infants on neonatal intensive care units requires more than continuous measurement of physiological signals such as respiration, heart rate and temperature. According to HUN-REN, objective, around-the-clock observation of spontaneous motor activity and body position is also essential because these behaviors are direct indicators of nervous system maturity and the infant's health. Manual visual observation, however, is time-consuming, subjective, and cannot be maintained 24/7.
Hungarian researchers have developed a camera-based monitoring system that applies artificial intelligence to this problem.
Study setup and system capabilities
- The study that introduced the system analyzed the behaviour of 88 infants housed in incubators and used a total of 8,400 hours of video footage to train the model.
- The research was published in the journal Nature Pediatric Research. Cameras were mounted above the incubators and the system processes the images in real time.
- Deep learning algorithms trained by teams from HUN-REN and the Institute for Computer Science and Control (SZTAKI) can identify an infant's body position with centimetre-level accuracy (for example, whether the baby lies on its back, stomach, or side).
- The model also distinguishes fine, coordinated movements from sudden reactions that indicate stress.
Clinical implications and benefits
The Hungarian AI system continuously records and structures the collected data, providing immediate and reliable feedback to nurses and physicians. The data also enables more detailed mapping of individual sleep–wake cycles. Aligning clinical interventions and care tasks with infants' natural biorhythms has been shown to accelerate development, reduce stress levels, and potentially shorten hospital stays.
Deployment timeline
The authors have not specified when the technology might be put into routine clinical use. The published results, however, provide a foundation for continuous, objective monitoring that could complement conventional physiological measurements in neonatal intensive care.
