Continuous Monitoring of Sleep-Related Biomarkers via a Nearable Solution Based on Fiber Bragg Grating Technology

Area of Science:

• Biomedical Engineering
• Wearable Technology
• Sleep Science

Background:

• Continuous monitoring of sleep-related biomarkers is crucial for diagnosing sleep disorders.
• Existing methods often require specialized equipment or invasive procedures.
• There is a need for non-invasive, user-friendly solutions for sleep monitoring in clinical and home settings.

Purpose of the Study:

• To develop and evaluate a nearable solution utilizing fiber Bragg grating (FBG) technology embedded in a mattress for continuous sleep monitoring.
• To assess the accuracy of the FBG-based mattress in detecting bed occupancy, sleeping posture, respiratory rate (RR), and heart rate (HR).
• To explore the potential of machine learning algorithms for analyzing sensor data and estimating sleep biomarkers.

Main Methods:

• A mattress embedded with thirteen FBG sensors was developed using biocompatible silicone.
• Ten participants underwent simulated sleeping conditions to test performance across various postures and respiratory patterns.
• Machine learning algorithms (Decision Tree, SVM, Naïve-Bayes, CNNs) were employed to analyze sensor data for biomarker detection and estimation.

Main Results:

• 100% accuracy was achieved in detecting bed occupancy.
• Support Vector Machine (SVM) achieved 78.4% accuracy in distinguishing axial from lateral sleeping positions; CNNs achieved 75.9% for left vs. right positions.
• Respiratory rate (RR) and heart rate (HR) were estimated with mean absolute errors below 0.7 breaths/min and 4 bpm, respectively.

Conclusions:

• The FBG-based mattress system demonstrates high accuracy in monitoring key sleep-related biomarkers.
• The proposed technology shows significant potential for comprehensive sleep disorder evaluation in both clinical and home environments.
• This innovative nearable solution offers a non-invasive and continuous approach to sleep health monitoring.