A Pilot Clinical Study to Understand the Relationship between General Movements and Ultra-Short-Term HRV of Neonates

Ziyan Wu, Yiming Zhong, Chuchu Liao, Xiaoyan Song, Qiqiong Wang, Wenjin Wang

Confirmation: I have read and agree with the IEEE BHI 2025 conference submission's policy on behalf of myself and my co-authors.

Keywords: NICU, Health Monitoring, General Movements, Ultra-short-term HRV, ECG, Camera

TL;DR: This study synchronously monitors GMs and USTHRV in a neonatal intensive care unit (NICU) to understand their physiological relationship, demonstrating the potential of UST-HRV in neonatal health monitoring.

Abstract: Heart rate variability (HRV) reflects the regulation of the infant autonomic nervous system in neonatal care, and ultra-short-term (UST) HRV provides a faster response with higher time resolution. General movements (GMs) are indicators for the evaluation of neonatal neurological development. However, movements are often considered as a source of artifacts in HRV measurement and their physiological significance has been overlooked. This study synchronously monitors GMs and USTHRV in a neonatal intensive care unit (NICU) to understand their physiological relationship, demonstrating the potential of UST-HRV in neonatal health monitoring. UST-HRV (a total of nine HRV parameters including RMSSD, SDNN, pNN20, LF, HF, LF/HF, SD1, SD2, SD1/SD2) is extracted from denoised electrocardiography (ECG) signals, and GMs are measured by an RGB camera with the optical flow method. Our clinical study shows that LF has a strong temporal correlation with GMs, with an average Pearson correlation coefficient of -0.623. Significant changes (p<0.05 for t-statistic in the linear mixed effects model) in UST-HRV are observed in SDNN, SD2, SD1/SD2, LF, HF and LF/HF before, during and after GMs. Such relationship indicates that the variation of UST-HRV is a quick and sensitive indicator for state changes such as GMs, providing an indication for clinical evaluation in dynamic events.

Track: 4. Clinical Informatics

Registration Id: RKN4DNVTDMQ

Submission Number: 38