A Real-Time Algorithm for Sleep Apnea and Hypopnea Detection

In this work, we present a novel rule-based method utilizing single nasal pressure (NP) to diagnose sleep apnea-hypopnea syndrome (SAHS) in real-time. The proposed method has adopted several vital parameters to quantify respiratory patterns and updated all the baselines dynamically. We have investigated thirty-five overnight recordings which are manually annotated by certified physicians and conducted event-by-event comparison and statistical analysis for apnea hypopnea index (AHI). The results are promising with 91.6% accuracy and 91.4% sensitivity for merged apnea-hypopnea detection. Furthermore, calculated AHI obtained by the proposed method highly agrees with manual annotations with Pearson's correlation coefficient as high as 0.98. It is plausible that the proposed method is viable to be incorporated into polysomnography (PSG) or other portable devices for automatic sleep disorder monitoring since all the events detected are with high time resolution.