Breathable soft bioelectronics for enhanced automatic detection of obstructive sleep apnea

Area of Science:

• Biomedical Engineering
• Artificial Intelligence
• Sleep Medicine

Background:

• Obstructive sleep apnea (OSA) affects 1-5% of children, with higher rates in those with cleft lip/palate and dentofacial deformities.
• Current diagnostic methods like polysomnography are costly, complex, and uncomfortable, limiting access for pediatric patients.
• There is a need for accessible, non-invasive diagnostic tools for pediatric OSA.

Purpose of the Study:

• To introduce a novel wireless, soft, and breathable bioelectronic system for detecting obstructive sleep apnea (OSA) in children.
• To develop a deep learning framework for automatic sleep stage classification and apnea event detection.
• To provide a foundation for diagnosing pediatric OSA and monitoring outcomes of orthognathic surgery.

Main Methods:

• Development of a wearable bioelectronic system with a perforated, deformable structure for improved skin conformity and reduced motion artifacts.
• Measurement of electrophysiological signals from the face using the wearable device.
• Application of a deep learning framework combining multi-stream convolutional neural networks and bi-directional long short-term memory models for data analysis.

Main Results:

• The developed system successfully detects obstructive sleep apnea (OSA) by analyzing electrophysiological signals.
• The deep learning model accurately classifies sleep stages and identifies apnea events.
• The wearable device's design minimizes discomfort and motion artifacts, enhancing data quality.

Conclusions:

• The wireless bioelectronic system presents a promising, accessible solution for diagnosing pediatric obstructive sleep apnea (OSA).
• This technology has the potential to improve early diagnosis and management of OSA in high-risk pediatric populations.
• The system could also be valuable for differentiating pre- and post-operative sleep patterns in patients undergoing orthognathic surgery.