Automated preterm infant sleep staging using capacitive electrocardiography

Objective: To date, mainly obtrusive methods (e.g. adhesive electrodes in electroencephalography or electrocardiography) have been necessary to determine the preterm infant sleep states. As any obtrusive measure should be avoided in preterm infants because of their immature skin development, we investigated the possibility of automated sleep staging using electrocardiograph signals from non-adhesive capacitive electrocardiography. Approach: Capacitive electrocardiography data from eight different patients with a mean gestational age of 30  ±  2.5 weeks are compared to manually annotated reference signals from classic adhesive electrodes. The sleep annotations were performed by two trained observers based on behavioral observations. Main results: Based on these annotations, classification performance of the preterm infant active and quiet sleep states, based on capacitive electrocardiography signals, showed a kappa value of 0.56  ±  0.20. Adding wake and caretaking into the classification, a performance of kappa 0.44  ±  0.21 was achieved. In-between sleep state performance showed a classification performance of kappa 0.36  ±  0.12. Lastly, a performance for all sleep states of kappa 0.35  ±  0.17 was attained. Significance: Capacitive electrocardiography signals can be utilized to classify the central preterm infant sleep states, active and quiet sleep. With further research based on our results, automated classification of sleep states can become an essential instrument in future intensive neonatal care for continuous brain maturation monitoring. In particular, being able to use capacitive electrocardiography for continuous monitoring is a significant contributor to reducing disruption and harm for this extremely fragile patient group.

[1]  Sabine Van Huffel,et al.  Quiet sleep detection in preterm infants using deep convolutional neural networks , 2018, Journal of neural engineering.

[2]  Ronald M. Aarts,et al.  Unobtrusive assessment of neonatal sleep state based on heart rate variability retrieved from electrocardiography used for regular patient monitoring. , 2017, Early human development.

[3]  S. Vanhatalo,et al.  Automated classification of neonatal sleep states using EEG , 2017, Clinical Neurophysiology.

[4]  Ronald M. Aarts,et al.  Unobtrusive sleep state measurements in preterm infants - A review. , 2017, Sleep medicine reviews.

[5]  Louis Atallah,et al.  Features of Heart Rate Variability Capture Regulatory Changes During Kangaroo Care in Preterm Infants , 2017, The Journal of pediatrics.

[6]  Sabine Van Huffel,et al.  An Automated Quiet Sleep Detection Approach in Preterm Infants as a Gateway to Assess Brain Maturation , 2017, Int. J. Neural Syst..

[7]  W. Fifer,et al.  An automated method for coding sleep states in human infants based on respiratory rate variability. , 2016, Developmental psychobiology.

[8]  M G Signorini,et al.  Novel heart rate parameters for the assessment of autonomic nervous system function in premature infants , 2016, Physiological measurement.

[9]  R Vullings,et al.  Unobtrusive ECG monitoring in the NICU using a capacitive sensing array , 2014, Physiological measurement.

[10]  Alois Knoll,et al.  Gradient boosting machines, a tutorial , 2013, Front. Neurorobot..

[11]  Nathaniel H. Hunt,et al.  The Appropriate Use of Approximate Entropy and Sample Entropy with Short Data Sets , 2012, Annals of Biomedical Engineering.

[12]  M. McHugh.  Interrater reliability: the kappa statistic , 2012, Biochemia medica.

[13]  Massimo Mischi,et al.  Low-complexity R-peak detection for ambulatory fetal monitoring , 2012, Physiological measurement.

[14]  Douglas E. Lake,et al.  Improved entropy rate estimation in physiological data , 2011, 2011 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.

[15]  C H L Peters,et al.  A robust physiology-based source separation method for QRS detection in low amplitude fetal ECG recordings , 2010, Physiological measurement.

[16]  Haibo He,et al.  ADASYN: Adaptive synthetic sampling approach for imbalanced learning , 2008, 2008 IEEE International Joint Conference on Neural Networks (IEEE World Congress on Computational Intelligence).

[17]  A. Gruetzmann,et al.  Novel dry electrodes for ECG monitoring , 2007, Physiological measurement.

[18]  Kenneth Loparo,et al.  Automated State Analyses: Proposed Applications to Neonatal Neurointensive Care , 2005, Journal of clinical neurophysiology : official publication of the American Electroencephalographic Society.

[19]  Thomas Ruf,et al.  The Lomb-Scargle Periodogram in Biological Rhythm Research: Analysis of Incomplete and Unequally Spaced Time-Series , 1999 .

[20]  H. Prechtl,et al.  The behavioural states of the newborn infant (a review). , 1974, Brain research.

[21]  Quiet Sleep , 2020, Encyclopedia of Behavioral Medicine.

[22]  Ronald M. Aarts,et al.  Reduction of Periodic Motion Artifacts in Photoplethysmography , 2017, IEEE Transactions on Biomedical Engineering.

[23]  Ronald M. Aarts,et al.  Estimating actigraphy from motion artifacts in ECG and respiratory effort signals , 2016, Physiological measurement.

[24]  Rik Vullings,et al.  Motion Artifacts in Capacitive ECG Measurements: Reducing the Combined Effect of DC Voltages and Capacitance Changes Using an Injection Signal , 2015, IEEE Transactions on Biomedical Engineering.

[25]  Vladimir Krajca,et al.  Newborn Sleep Stage Classification Using Hybrid Evolutionary Approach , 2007 .

[26]  Nitesh V. Chawla,et al.  SMOTE: Synthetic Minority Over-sampling Technique , 2002, J. Artif. Intell. Res..