Deep learning approach for ECG-based automated sleep state classification in preterm infants

Jan Werth (Corresponding author), Mustafa Radha, Peter Andriessen, Ronald Aarts, Xi Long

Research output: Contribution to journalArticleAcademicpeer-review

Abstract

Preterm infant neuronal development is related to the distribution of their sleep states. The distribution changes throughout development. Automated sleep state monitoring can become a powerful aid for development monitoring in preterm infants. Three datasets including 34 preterm infants and a total of 18,018 30 s manually annotated sleep intervals (sleep-epochs) were analyzed in this study. The annotation of sleep states includes active sleep, quiet sleep, intermediate sleep, wake, and caretaking. Four different recurrent neuronal network architectures were compared for two-state, three-state, and all-state analysis. A sequential network was used to compare long- and short-term memory and gated recurrent unit models. The other network architectures were based on the popular ResNet and ResNext architectures utilizing residual connection for more depth. The most essential sleep states, active and quiet sleep, could be separated with a kappa of 0.43 ± 0.08. Quiet versus caretaking and wake showed a kappa of 0.44 ± 0.01. The three state classifications of active versus quiet versus intermediate sleep resulted in a kappa of 0.35 ± 0.07 and active versus quiet versus wake and caretaking resulted in a kappa of 0.33 ± 0.04. The all-state classification was underperforming with probably due to difficulty in separating subtle differences between all states and a lack of sufficient training data for the minority classes.
LanguageEnglish
Article number101663
Number of pages10
JournalBiomedical Signal Processing and Control
Volume56
DOIs
StatePublished - 1 Feb 2020

Fingerprint

Electrocardiography
Premature Infants
Sleep
Learning
Network architecture
Deep learning
Polysomnography
Monitoring
Child Development
Short-Term Memory
Data storage equipment

Keywords

  • Automated sleep state classification
  • Deep learning
  • Electrocardiography
  • Gated recurrent unit
  • Long short-term memory
  • Preterm infants
  • Recurrent neural networks
  • Sleep

Cite this

@article{615296f6fb494475a26e52cb750c696f,
title = "Deep learning approach for ECG-based automated sleep state classification in preterm infants",
abstract = "Preterm infant neuronal development is related to the distribution of their sleep states. The distribution changes throughout development. Automated sleep state monitoring can become a powerful aid for development monitoring in preterm infants. Three datasets including 34 preterm infants and a total of 18,018 30 s manually annotated sleep intervals (sleep-epochs) were analyzed in this study. The annotation of sleep states includes active sleep, quiet sleep, intermediate sleep, wake, and caretaking. Four different recurrent neuronal network architectures were compared for two-state, three-state, and all-state analysis. A sequential network was used to compare long- and short-term memory and gated recurrent unit models. The other network architectures were based on the popular ResNet and ResNext architectures utilizing residual connection for more depth. The most essential sleep states, active and quiet sleep, could be separated with a kappa of 0.43 ± 0.08. Quiet versus caretaking and wake showed a kappa of 0.44 ± 0.01. The three state classifications of active versus quiet versus intermediate sleep resulted in a kappa of 0.35 ± 0.07 and active versus quiet versus wake and caretaking resulted in a kappa of 0.33 ± 0.04. The all-state classification was underperforming with probably due to difficulty in separating subtle differences between all states and a lack of sufficient training data for the minority classes.",
keywords = "Automated sleep state classification, Deep learning, Electrocardiography, Gated recurrent unit, Long short-term memory, Preterm infants, Recurrent neural networks, Sleep",
author = "Jan Werth and Mustafa Radha and Peter Andriessen and Ronald Aarts and Xi Long",
year = "2020",
month = "2",
day = "1",
doi = "10.1016/j.bspc.2019.101663",
language = "English",
volume = "56",
journal = "Biomedical Signal Processing and Control",
issn = "1746-8094",
publisher = "Elsevier",

}

Deep learning approach for ECG-based automated sleep state classification in preterm infants. / Werth, Jan (Corresponding author); Radha, Mustafa; Andriessen, Peter; Aarts, Ronald; Long, Xi.

In: Biomedical Signal Processing and Control, Vol. 56, 101663, 01.02.2020.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Deep learning approach for ECG-based automated sleep state classification in preterm infants

AU - Werth,Jan

AU - Radha,Mustafa

AU - Andriessen,Peter

AU - Aarts,Ronald

AU - Long,Xi

PY - 2020/2/1

Y1 - 2020/2/1

N2 - Preterm infant neuronal development is related to the distribution of their sleep states. The distribution changes throughout development. Automated sleep state monitoring can become a powerful aid for development monitoring in preterm infants. Three datasets including 34 preterm infants and a total of 18,018 30 s manually annotated sleep intervals (sleep-epochs) were analyzed in this study. The annotation of sleep states includes active sleep, quiet sleep, intermediate sleep, wake, and caretaking. Four different recurrent neuronal network architectures were compared for two-state, three-state, and all-state analysis. A sequential network was used to compare long- and short-term memory and gated recurrent unit models. The other network architectures were based on the popular ResNet and ResNext architectures utilizing residual connection for more depth. The most essential sleep states, active and quiet sleep, could be separated with a kappa of 0.43 ± 0.08. Quiet versus caretaking and wake showed a kappa of 0.44 ± 0.01. The three state classifications of active versus quiet versus intermediate sleep resulted in a kappa of 0.35 ± 0.07 and active versus quiet versus wake and caretaking resulted in a kappa of 0.33 ± 0.04. The all-state classification was underperforming with probably due to difficulty in separating subtle differences between all states and a lack of sufficient training data for the minority classes.

AB - Preterm infant neuronal development is related to the distribution of their sleep states. The distribution changes throughout development. Automated sleep state monitoring can become a powerful aid for development monitoring in preterm infants. Three datasets including 34 preterm infants and a total of 18,018 30 s manually annotated sleep intervals (sleep-epochs) were analyzed in this study. The annotation of sleep states includes active sleep, quiet sleep, intermediate sleep, wake, and caretaking. Four different recurrent neuronal network architectures were compared for two-state, three-state, and all-state analysis. A sequential network was used to compare long- and short-term memory and gated recurrent unit models. The other network architectures were based on the popular ResNet and ResNext architectures utilizing residual connection for more depth. The most essential sleep states, active and quiet sleep, could be separated with a kappa of 0.43 ± 0.08. Quiet versus caretaking and wake showed a kappa of 0.44 ± 0.01. The three state classifications of active versus quiet versus intermediate sleep resulted in a kappa of 0.35 ± 0.07 and active versus quiet versus wake and caretaking resulted in a kappa of 0.33 ± 0.04. The all-state classification was underperforming with probably due to difficulty in separating subtle differences between all states and a lack of sufficient training data for the minority classes.

KW - Automated sleep state classification

KW - Deep learning

KW - Electrocardiography

KW - Gated recurrent unit

KW - Long short-term memory

KW - Preterm infants

KW - Recurrent neural networks

KW - Sleep

UR - http://www.scopus.com/inward/record.url?scp=85072868522&partnerID=8YFLogxK

U2 - 10.1016/j.bspc.2019.101663

DO - 10.1016/j.bspc.2019.101663

M3 - Article

VL - 56

JO - Biomedical Signal Processing and Control

T2 - Biomedical Signal Processing and Control

JF - Biomedical Signal Processing and Control

SN - 1746-8094

M1 - 101663

ER -