Related Concept Videos
Stages of Sleep
Before sleep begins, in wakefulness, the brain exhibits primarily beta waves, which are high in frequency and low in amplitude, indicating alertness...
Mechanical Ventilation III: Noninvasive Ventilation
Noninvasive Positive-Pressure Ventilation (NIPPV)
Ventilatory Modes
There are three ventilatory modes: full support, partial support, and spontaneous. These are described below.
Full Support Modes
Full support modes include controlled mechanical ventilation, continuous mandatory...
Assessment of Diffusion and Perfusion
The Role of Diffusion in Respiration
Diffusion is the process by which molecules move from an area of higher concentration to an area of lower concentration. In the respiratory system, this principle...
Cardiopulmonary Resuscitation II: ACLS Airway Management
Pulse Oximetry
Purpose
Average SpO2 values are greater than 95%. If the readings fall below 90%, it indicates that...
You might also read
Related Articles
Articles linked to this work by shared authors, journal, and citation graph.
Predicting sleep state from continuous positive airway pressure flow in patients with obstructive sleep apnea.
Automatic sleep staging from CPAP airflow using a dual fusion multi-period convolutional neural network.
Latent classes of sleep deficiency and correlates among patients receiving methadone treatment: A longitudinal study.
Related Experiment Video
Updated: Jun 27, 2026

Multi-Modal Home Sleep Monitoring in Older Adults
Published on: January 26, 2019
Sleep Stage Classification During CPAP Therapy from CPAP-Airflow and Wearable Fingertip Signals.
Hsin-Yu Chen1, Aatif Husain2, Andrey V Zinchuk3
1PranaQ Pte. Ltd., Singapore 018937, Singapore.
Combining Continuous Positive Airway Pressure (CPAP) airflow and photoplethysmography (PPG) signals improves sleep stage detection accuracy. This fusion enhances sleep monitoring and auto-CPAP titration for obstructive sleep apnea-hypopnea syndrome (OSAHS) patients.
Area of Science:
- Biomedical Engineering
- Sleep Medicine
- Artificial Intelligence
Background:
- Continuous Positive Airway Pressure (CPAP) is the standard treatment for obstructive sleep apnea-hypopnea syndrome (OSAHS).
- Photoplethysmography (PPG) sensors in wearable devices are used for home sleep apnea testing.
- Both airflow and PPG signals contain rich physiological data for sleep analysis.
Purpose of the Study:
- To develop a method for efficiently estimating sleep dynamics in patients undergoing CPAP treatment.
- To enhance sleep stage detection accuracy by combining CPAP airflow and PPG signals.
- To improve the precision of sleep monitoring and CPAP titration.
Main Methods:
- One-dimensional convolutional neural networks were trained separately for CPAP-airflow and PPG signals.
- A probabilistic ensembling method (late-fusion soft-voting) was used to fuse the outputs of the individual models.
- The fused model predicted sleep stages based on synchronized softmax probability vectors.
Main Results:
- The ensembled model achieved a higher overall Cohen's kappa score (0.587) compared to PPG-only (0.511) and CPAP-airflow-only (0.452) models for three-stage classification.
- The F1-score for REM sleep detection improved to 0.706 with the ensemble method.
- A deep sleep sensitivity of 0.596 was achieved in four-stage classification using probabilistic ensembling.
Conclusions:
- Fusion of CPAP and PPG data enhances sleep stage detection accuracy and precision of sleep monitoring.
- The improved REM identification is particularly significant for clinical applications.
- The algorithm can optimize in-home auto-CPAP titration by addressing REM-related respiratory instability in OSAHS patients.
