A flexible deep learning architecture for temporal sleep stage classification using accelerometry and photoplethysmography
Research output: Contribution to journal › Journal article › Research › peer-review
Standard
A flexible deep learning architecture for temporal sleep stage classification using accelerometry and photoplethysmography. / Olsen, Mads; Zeitzer, Jamie M.; Richardson, Risa N.; Davidenko, Polina; Jennum, Poul J.; Sorensen, Helge B.D.; Mignot, Emmanuel.
In: IEEE Transactions on Biomedical Engineering, Vol. 70, No. 1, 2023, p. 228-237.Research output: Contribution to journal › Journal article › Research › peer-review
Harvard
APA
Vancouver
Author
Bibtex
}
RIS
TY - JOUR
T1 - A flexible deep learning architecture for temporal sleep stage classification using accelerometry and photoplethysmography
AU - Olsen, Mads
AU - Zeitzer, Jamie M.
AU - Richardson, Risa N.
AU - Davidenko, Polina
AU - Jennum, Poul J.
AU - Sorensen, Helge B.D.
AU - Mignot, Emmanuel
N1 - Publisher Copyright: IEEE
PY - 2023
Y1 - 2023
N2 - Wrist-worn consumer sleep technologies (CST) that contain accelerometers (ACC) and photoplethysmography (PPG) are increasingly common and hold great potential to function as out-of-clinic (OOC) sleep monitoring systems. However, very few validation studies exist because raw data from CSTs are rarely made accessible for external use. We present a deep neural network (DNN) with a strong temporal core, inspired by U-Net, that can process multivariate time series inputs with different dimensionality to predict sleep stages (wake, light-, deep-, and REM sleep) using ACC and PPG signals from nocturnal recordings. The DNN was trained and tested on 3 internal datasets, comprising raw data both from clinical and wrist-worn devices from 301 recordings (PSG-PPG: 266, Wrist-worn PPG: 35). External validation was performed on a hold-out test dataset containing 35 recordings comprising only raw data from a wrist-worn CST. An accuracy = 0.71±0.09, 0.76±0.07, 0.73±0.06, and κ = 0.58±0.13, 0.64±0.09, 0.59±0.09 was achieved on the internal test sets. Our experiments show that spectral preprocessing yields superior performance when compared to surrogate-, feature-, raw data-based preparation. Combining both modalities produce the overall best performance, although PPG proved to be the most impactful and was the only modality capable of detecting REM sleep well. Including ACC improved model precision to wake and sleep metric estimation. Increasing input segment size improved performance consistently; the best performance was achieved using 1024 epochs (∼8.5 hrs.). An accuracy = 0.69±0.13 and κ = 0.58±0.18 was achieved on the hold-out test dataset, proving the generalizability and robustness of our approach to raw data collected with a wrist-worn CST.
AB - Wrist-worn consumer sleep technologies (CST) that contain accelerometers (ACC) and photoplethysmography (PPG) are increasingly common and hold great potential to function as out-of-clinic (OOC) sleep monitoring systems. However, very few validation studies exist because raw data from CSTs are rarely made accessible for external use. We present a deep neural network (DNN) with a strong temporal core, inspired by U-Net, that can process multivariate time series inputs with different dimensionality to predict sleep stages (wake, light-, deep-, and REM sleep) using ACC and PPG signals from nocturnal recordings. The DNN was trained and tested on 3 internal datasets, comprising raw data both from clinical and wrist-worn devices from 301 recordings (PSG-PPG: 266, Wrist-worn PPG: 35). External validation was performed on a hold-out test dataset containing 35 recordings comprising only raw data from a wrist-worn CST. An accuracy = 0.71±0.09, 0.76±0.07, 0.73±0.06, and κ = 0.58±0.13, 0.64±0.09, 0.59±0.09 was achieved on the internal test sets. Our experiments show that spectral preprocessing yields superior performance when compared to surrogate-, feature-, raw data-based preparation. Combining both modalities produce the overall best performance, although PPG proved to be the most impactful and was the only modality capable of detecting REM sleep well. Including ACC improved model precision to wake and sleep metric estimation. Increasing input segment size improved performance consistently; the best performance was achieved using 1024 epochs (∼8.5 hrs.). An accuracy = 0.69±0.13 and κ = 0.58±0.18 was achieved on the hold-out test dataset, proving the generalizability and robustness of our approach to raw data collected with a wrist-worn CST.
KW - Classification algorithms
KW - consumer sleep technologies
KW - Decoding
KW - Deep learning
KW - deep learning
KW - Feature extraction
KW - mHealth
KW - Rapid eye movement sleep
KW - Recording
KW - Sleep apnea
KW - sleep stage classification
KW - wrist actigraphy
U2 - 10.1109/TBME.2022.3187945
DO - 10.1109/TBME.2022.3187945
M3 - Journal article
C2 - 35786544
AN - SCOPUS:85134212549
VL - 70
SP - 228
EP - 237
JO - IEEE Transactions on Biomedical Engineering
JF - IEEE Transactions on Biomedical Engineering
SN - 0018-9294
IS - 1
ER -
ID: 324665181