Fusion of DET and Time-Frequency Analysis for Obstructive Sleep Apnea Screening

Abstract

Obstructive Sleep Apnea (OSA) is a common disease, whose main feature is repeated episodes of apnea and hypopnea during sleep, and always leading to waking up from sleep. Traditional Heart Rate Variability (HRV) analysis methods for OSA screening are divided into time domain analysis and frequency domain analysis. The time domain analysis has been widely used, it uses the statistical discrete trend analysis method to analyze the variation of heart rate and RR. The law of frequency domain analysis is to use the time domain signal of HRV to do spectrum analysis. However, traditional HRV analysis methods are linear analysis methods, and there are shortcomings such as ignoring the short-term volatility of data and being unable to measure complexity. Therefore, this experiment proposes to perform time domain analysis, frequency domain analysis, and nonlinear DET analysis on the PPG signals of 15 OSA patients and 15 normal people, a total of 30 subjects. Significance, specificity, and accuracy are contrasted. Finally, when the embedding dimension m=6 and the scale factor s=4, the DET method gets better results than the traditional time-domain and frequency-domain analysis methods on significance, accuracy and specificity of OSA screening. This provides a fresh perspective for OSA screening.