Related Experiment Video
Updated: Jun 7, 2025
How to Administer Near-Infrared Spectroscopy in Critically ill Neonates, Infants, and Children
Published on: August 19, 2020
Near-Infrared Spectroscopy for Neonatal Sleep Classification
Naser Hakimi1, Emad Arasteh1, Maren Zahn2,3
1Department of Neonatology, Wilhelmina Children's Hospital, University Medical Center Utrecht, Lundlaan 6, 3584 EA Utrecht, The Netherlands.
Insights
This study shows that near-infrared spectroscopy (NIRS) can accurately classify sleep states in preterm infants using a deep convolutional neural network (CNN), aiding neurological development monitoring. This NIRS-based method offers a portable solution for neonatal sleep assessment.
Area of Science:
- Neonatal neurology and sleep science.
- Biomedical engineering and signal processing.
Background:
- Sleep states, active sleep (AS) and quiet sleep (QS), are critical for brain development in preterm infants.
- Monitoring neonatal sleep is essential for promoting neurological maturation and well-being, especially in high-risk infants.
- Respiratory rate (RR) and heart rate (HR) are key indicators in neonatal sleep assessment systems.
Purpose of the Study:
- To introduce a comprehensive sleep classification approach for preterm infants using high-frequency near-infrared spectroscopy (NIRS) signals.
- To evaluate the performance of a deep convolutional neural network (CNN) model for classifying active sleep (AS) and quiet sleep (QS) states.
- To compare the CNN model's efficacy against benchmark machine learning classifiers.
Main Methods:
- High-frequency NIRS signals (100 Hz) were recorded from nine preterm infants in a neonatal intensive care unit.
- Eight features, including HR, RR, motion parameters, and neural activity proxies, were extracted from NIRS signals.
- A deep convolutional neural network (CNN) was trained and validated using two cross-validation approaches to classify sleep states.
Main Results:
- The CNN model achieved high performance metrics, including 88% accuracy, 94% balanced accuracy, 91% F1-score, 95% Kappa, and 96% AUC-ROC in data pooling cross-validation.
- Random Forest (RF) and XGBoost (XGB) classifiers showed comparable accuracy to the CNN model.
- The study confirmed the feasibility of extracting NIRS-based HR and RR for effective neonatal sleep state assessment.
Conclusions:
- High-frequency NIRS data, combined with extracted HR and RR, provides a viable method for assessing neonatal sleep states, even in intensive care settings.
- The user-friendly, portable, and less complex NIRS approach has potential applications beyond the NICU.
- This research offers a promising advancement in neonatal sleep assessment, contributing to improved infant health and developmental outcomes.
Abstract:
Sleep, notably active sleep (AS) and quiet sleep (QS), plays a pivotal role in the brain development and gradual maturation of (pre) term infants. Monitoring their sleep patterns is imperative, as it can serve as a tool in promoting neurological maturation and well-being, particularly important in preterm infants who are at an increased risk of immature brain development. An accurate classification of neonatal sleep states can contribute to optimizing treatments for high-risk infants, with respiratory rate (RR) and heart rate (HR) serving as key components in sleep assessment systems for neonates. Recent studies have demonstrated the feasibility of extracting both RR and HR using near-infrared spectroscopy (NIRS) in neonates. This study introduces a comprehensive sleep classification approach leveraging high-frequency NIRS signals recorded at a sampling rate of 100 Hz from a cohort of nine preterm infants admitted to a neonatal intensive care unit. Eight distinct features were extracted from the raw NIRS signals, including HR, RR, motion-related parameters, and proxies for neural activity. These features served as inputs for a deep convolutional neural network (CNN) model designed for the classification of AS and QS sleep states. The performance of the proposed CNN model was evaluated using two cross-validation approaches: ten-fold cross-validation of data pooling and five-fold cross-validation, where each fold contains two independently recorded NIRS data. The accuracy, balanced accuracy, F1-score, Kappa, and AUC-ROC (Area Under the Curve of the Receiver Operating Characteristic) were employed to assess the classifier performance. In addition, comparative analyses against six benchmark classifiers, comprising K-Nearest Neighbors, Naive Bayes, Support Vector Machines, Random Forest (RF), AdaBoost, and XGBoost (XGB), were conducted. Our results reveal the CNN model's superior performance, achieving an average accuracy of 88%, a balanced accuracy of 94%, an F1-score of 91%, Kappa of 95%, and an AUC-ROC of 96% in data pooling cross-validation. Furthermore, in both cross-validation methods, RF and XGB demonstrated accuracy levels closely comparable to the CNN classifier. These findings underscore the feasibility of leveraging high-frequency NIRS data, coupled with NIRS-based HR and RR extraction, for assessing sleep states in neonates, even in an intensive care setting. The user-friendliness, portability, and reduced sensor complexity of the approach suggest its potential applications in various less-demanding settings. This research thus presents a promising avenue for advancing neonatal sleep assessment and its implications for infant health and development.
