The Sleep Well Baby project: an automated real-time sleep–wake state prediction algorithm in preterm infants

Association between sleep stages and brain microstructure in preterm infants: Insights from DTI analysis

STUDY OBJECTIVES

The aim of this study was to investigate the relationship between sleep stages and neural microstructure - measured using diffusion tensor imaging - of the posterior limb of the internal capsule and corticospinal tract in preterm infants.

METHODS

A retrospective cohort of 50 preterm infants born between 24 + 4 and 29 + 3 weeks gestational age was included in the study. Sleep stages were continuously measured for 5-7 consecutive days between 29 + 0 and 31 + 6 weeks postmenstrual age using an in-house-developed, and recently published, automated sleep staging algorithm based on routinely measured heart rate and respiratory rate. Additionally, a diffusion tensor imaging scan was conducted at term equivalent age as part of standard care. Region of interest analysis of the posterior limb of the internal capsule was performed, and tractography was used to analyze the corticospinal tract. The association between sleep and white matter microstructure of the posterior limb of the internal capsule and corticospinal tract was examined using a multiple linear regression model, adjusted for potential confounders.

RESULTS

The results of the analyses revealed an interaction effect between sleep stage and days of invasive ventilation on the fractional anisotropy of the left and right posterior limb of the internal capsule (β = 0.04, FDR-adjusted p = 0.001 and β = 0.04, FDR-adjusted p = 0.02, respectively). Furthermore, an interaction effect between sleep stage and days of invasive ventilation was observed for the radial diffusivity of the mean of the left and right PLIC (β = -4.1e-05, FDR-adjusted p = 0.04).

CONCLUSIONS

Previous research has shown that, in very preterm infants, invasive ventilation has a negative effect on white matter tract maturation throughout the brain. A positive association between active sleep and white matter microstructure of the posterior limb of the internal capsule, may indicate a protective role of sleep in this vulnerable population.

Sleep as a driver of pre- and postnatal brain development

In 1966, Howard Roffwarg proposed the ontogenic sleep hypothesis, relating neural plasticity and development to rapid eye movement (REM) sleep, a hypothesis that current fetal and neonatal sleep research is still exploring. Recently, technological advances have enabled researchers to automatically quantify neonatal sleep architecture, which has caused a resurgence of research in this field as attempts are made to further elucidate the important role of sleep in pre- and postnatal brain development. This article will review our current understanding of the role of sleep as a driver of brain development and identify possible areas for future research. IMPACT: The evidence to date suggests that Roffwarg's ontogenesis hypothesis of sleep and brain development is correct. A better understanding of the relationship between sleep and the development of functional connectivity is needed. Reliable, non-invasive tools to assess sleep in the NICU and at home need to be tested in a real-world environment and the best way to promote healthy sleep needs to be understood before clinical trials promoting and optimizing sleep quality in neonates could be undertaken.

Characterising the motion and cardiorespiratory interaction of preterm infants can improve the classification of their sleep state

AIM

This study aimed to classify quiet sleep, active sleep and wake states in preterm infants by analysing cardiorespiratory signals obtained from routine patient monitors.

METHODS

We studied eight preterm infants, with an average postmenstrual age of 32.3 ± 2.4 weeks, in a neonatal intensive care unit in the Netherlands. Electrocardiography and chest impedance respiratory signals were recorded. After filtering and R-peak detection, cardiorespiratory features and motion and cardiorespiratory interaction features were extracted, based on previous research. An extremely randomised trees algorithm was used for classification and performance was evaluated using leave-one-patient-out cross-validation and Cohen's kappa coefficient.

RESULTS

A sleep expert annotated 4731 30-second epochs (39.4 h) and active sleep, quiet sleep and wake accounted for 73.3%, 12.6% and 14.1% respectively. Using all features, and the extremely randomised trees algorithm, the binary discrimination between active and quiet sleep was better than between other states. Incorporating motion and cardiorespiratory interaction features improved the classification of all sleep states (kappa 0.38 ± 0.09) than analyses without these features (kappa 0.31 ± 0.11).

CONCLUSION

Cardiorespiratory interactions contributed to detecting quiet sleep and motion features contributed to detecting wake states. This combination improved the automated classifications of sleep states.

An electroencephalography-based sleep index and supervised machine learning as a suitable tool for automated sleep classification in children

STUDY OBJECTIVES

Although sleep is frequently disrupted in the pediatric intensive care unit, it is currently not possible to perform real-time sleep monitoring at the bedside. In this study, spectral band powers of electroencephalography data are used to derive a simple index for sleep classification.

METHODS

Retrospective study at Erasmus MC Sophia Children's Hospital, using hospital-based polysomnography recordings obtained in non-critically ill children between 2017 and 2021. Six age categories were defined: 6-12 months, 1-3 years, 3-5 years, 5-9 years, 9-13 years, and 13-18 years. Candidate index measures were derived by calculating spectral band powers in different frequent frequency bands of smoothed electroencephalography. With the best performing index, sleep classification models were developed for two, three, and four states via decision tree and five-fold nested cross-validation. Model performance was assessed across age categories and electroencephalography channels.

RESULTS

In total 90 patients with polysomnography were included, with a mean (standard deviation) recording length of 10.3 (1.1) hours. The best performance was obtained with the gamma to delta spectral power ratio of the F4-A1 and F3-A1 channels with smoothing. Balanced accuracy was 0.88, 0.74, and 0.57 for two-, three-, and four-state classification. Across age categories, balanced accuracy ranged between 0.83 and 0.92 and 0.72 and 0.77 for two- and three-state classification, respectively.

CONCLUSIONS

We propose an interpretable and generalizable sleep index derived from single-channel electroencephalography for automated sleep monitoring at the bedside in non-critically ill children ages 6 months to 18 years, with good performance for two- and three-state classification.

CITATION

van Twist E, Hiemstra FW, Cramer ABG, et al. An electroencephalography-based sleep index and supervised machine learning as a suitable tool for automated sleep classification in children. J Clin Sleep Med. 2024;20(3):389-397.

Machine Learning-Derived Active Sleep as an Early Predictor of White Matter Development in Preterm Infants

White matter dysmaturation is commonly seen in preterm infants admitted to the neonatal intensive care unit (NICU). Animal research has shown that active sleep is essential for early brain plasticity. This study aimed to determine the potential of active sleep as an early predictor for subsequent white matter development in preterm infants. Using heart and respiratory rates routinely monitored in the NICU, we developed a machine learning-based automated sleep stage classifier in a cohort of 25 preterm infants (12 females). The automated classifier was subsequently applied to a study cohort of 58 preterm infants (31 females) to extract active sleep percentage over 5-7 consecutive days during 29-32 weeks of postmenstrual age. Each of the 58 infants underwent high-quality T2-weighted magnetic resonance brain imaging at term-equivalent age, which was used to measure the total white matter volume. The association between active sleep percentage and white matter volume was examined using a multiple linear regression model adjusted for potential confounders. Using the automated classifier with a superior sleep classification performance [mean area under the receiver operating characteristic curve (AUROC) = 0.87, 95% CI 0.83-0.92], we found that a higher active sleep percentage during the preterm period was significantly associated with an increased white matter volume at term-equivalent age [β = 0.31, 95% CI 0.09-0.53, false discovery rate (FDR)-adjusted p-value = 0.021]. Our results extend the positive association between active sleep and early brain development found in animal research to human preterm infants and emphasize the potential benefit of sleep preservation in the NICU setting.

Evaluation of Sleep Practices and Knowledge in Neonatal Healthcare

BACKGROUND

Developmental care is designed to optimize early brain maturation by integrating procedures that support a healing environment. Protecting preterm sleep is important in developmental care. However, it is unclear to what extent healthcare professionals are aware of the importance of sleep and how sleep is currently implemented in the day-to-day care in the neonatal intensive care unit (NICU).

PURPOSE

Identifying the current state of knowledge among healthcare professionals regarding neonatal sleep and how this is transferred to practice.

METHODS

A survey was distributed among Dutch healthcare professionals. Three categories of data were sought, including (1) demographics of respondents; (2) questions relating to sleep practices; and (3) objective knowledge questions relating to sleep physiology and importance of sleep. Data were analyzed using Spearman's rho test and Cramer's V test. Furthermore, frequency tables and qualitative analyses were employed.

RESULTS

The survey was completed by 427 participants from 34 hospitals in 25 Dutch cities. While healthcare professionals reported sleep to be especially important for neonates admitted in the NICU, low scores were achieved in the area of knowledge of sleep physiology. Most healthcare professionals (91.8%) adapted the timing of elective care procedures to sleep. However, sleep assessments were not based on scientific knowledge. Therefore, the difference between active sleep and wakefulness may often be wrongly assessed. Finally, sleep is rarely discussed between colleagues (27.4% regularly/always) and during rounds (7.5%-14.3% often/always).

IMPLICATIONS

Knowledge about sleep physiology should be increased through education among neonatal healthcare professionals. Furthermore, sleep should be considered more often during rounds and handovers.

The past, current, and future of neonatal intensive care units with artificial intelligence: a systematic review

Machine learning and deep learning are two subsets of artificial intelligence that involve teaching computers to learn and make decisions from any sort of data. Most recent developments in artificial intelligence are coming from deep learning, which has proven revolutionary in almost all fields, from computer vision to health sciences. The effects of deep learning in medicine have changed the conventional ways of clinical application significantly. Although some sub-fields of medicine, such as pediatrics, have been relatively slow in receiving the critical benefits of deep learning, related research in pediatrics has started to accumulate to a significant level, too. Hence, in this paper, we review recently developed machine learning and deep learning-based solutions for neonatology applications. We systematically evaluate the roles of both classical machine learning and deep learning in neonatology applications, define the methodologies, including algorithmic developments, and describe the remaining challenges in the assessment of neonatal diseases by using PRISMA 2020 guidelines. To date, the primary areas of focus in neonatology regarding AI applications have included survival analysis, neuroimaging, analysis of vital parameters and biosignals, and retinopathy of prematurity diagnosis. We have categorically summarized 106 research articles from 1996 to 2022 and discussed their pros and cons, respectively. In this systematic review, we aimed to further enhance the comprehensiveness of the study. We also discuss possible directions for new AI models and the future of neonatology with the rising power of AI, suggesting roadmaps for the integration of AI into neonatal intensive care units.

Combining Cardiorespiratory Signals and Video-Based Actigraphy for Classifying Preterm Infant Sleep States

The classification of sleep state in preterm infants, particularly in distinguishing between active sleep (AS) and quiet sleep (QS), has been investigated using cardiorespiratory information such as electrocardiography (ECG) and respiratory signals. However, accurately differentiating between AS and wake remains challenging; therefore, there is a pressing need to include additional information to further enhance the classification performance. To address the challenge, this study explores the effectiveness of incorporating video-based actigraphy analysis alongside cardiorespiratory signals for classifying the sleep states of preterm infants. The study enrolled eight preterm infants, and a total of 91 features were extracted from ECG, respiratory signals, and video-based actigraphy. By employing an extremely randomized trees (ET) algorithm and leave-one-subject-out cross-validation, a kappa score of 0.33 was achieved for the classification of AS, QS, and wake using cardiorespiratory features only. The kappa score significantly improved to 0.39 when incorporating eight video-based actigraphy features. Furthermore, the classification performance of AS and wake also improved, showing a kappa score increase of 0.21. These suggest that combining video-based actigraphy with cardiorespiratory signals can potentially enhance the performance of sleep-state classification in preterm infants. In addition, we highlighted the distinct strengths and limitations of video-based actigraphy and cardiorespiratory data in classifying specific sleep states.

Effects of the neonatal intensive care environment on circadian health and development of preterm infants

The circadian system in mammals ensures adaptation to the light-dark cycle on Earth and imposes 24-h rhythmicity on metabolic, physiological and behavioral processes. The central circadian pacemaker is located in the brain and is entrained by environmental signals called Zeitgebers. From here, neural, humoral and systemic signals drive rhythms in peripheral clocks in nearly every mammalian tissue. During pregnancy, disruption of the complex interplay between the mother's rhythmic signals and the fetal developing circadian system can lead to long-term health consequences in the offspring. When an infant is born very preterm, it loses the temporal signals received from the mother prematurely and becomes totally dependent on 24/7 care in the Neonatal Intensive Care Unit (NICU), where day/night rhythmicity is usually blurred. In this literature review, we provide an overview of the fetal and neonatal development of the circadian system, and short-term consequences of disruption of this process as occurs in the NICU environment. Moreover, we provide a theoretical and molecular framework of how this disruption could lead to later-life disease. Finally, we discuss studies that aim to improve health outcomes after preterm birth by studying the effects of enhancing rhythmicity in light and noise exposure.