Automated sleep apnea detection from snoring and carotid pulse signals using an innovative neck wearable piezoelectric sensor

Abstract

This study introduces an innovative wearable neck piezoelectric sensor (NPS) that measures snoring vibrations and carotid pulsations, offering a significant advancement in sleep apnea syndrome (SAS) diagnosis. Utilizing advanced algorithms like discrete wavelet transform and dynamic thresholding, the NPS detects snoring events with 83% accuracy, comparable to polysomnography, and calculates key metrics such as the snoring index (SI) and normalized snoring vibration energy (SVE%). Unlike traditional methods, the SVE% from NPS directly correlates with subjective assessments of snoring severity. It also measures carotid pulsation metrics such as pulse rate and the standard deviation of normal-to-normal intervals, achieving 85% accuracy in sleep phase determination against polysomnography. Moreover, NPS surpasses traditional methods in SI and SVE% accuracy, closely aligning with clinical evaluations of SAS severity. This user-friendly technology automates the measurement of critical snoring metrics, transforming SAS diagnosis and treatment by enhancing accessibility and efficiency for healthcare providers and patients.