Sleep apnea is a common sleep disorder that causes repeated breathing interruption during sleep. The performance of automated apnea detection methods based on respiratory signals depend on the signals considered and feature extraction methods. Moreover, feature engineering techniques are highly dependent on the experts’ experience and their prior knowledge about different physiological signals and conditions of the subjects. To overcome these problems, a novel deep recurrent neural network (RNN) framework is developed for automated feature extraction and detection of apnea events from single respiratory channel inputs. Long short-term memory (LSTM) and bidirectional long short-term memory (BiLSTM) are investigated to develop the proposed deep RNN model. The proposed framework is evaluated over three respiration signals: Oronasal thermal airflow (FlowTh), nasal pressure (NPRE), and abdominal respiratory inductance plethysmography (ABD). To demonstrate our results, we use polysomnography (PSG) data of 17 patients with obstructive, central, and mixed apnea events. Our results indicate the effectiveness of the proposed framework in automatic extraction for temporal features and automated detection of apneic events over the different respiratory signals considered in this study. Using a deep BiLSTM-based detection model, the NPRE signal achieved the highest overall detection results with true positive rate (sensitivity) = 90.3%, true negative rate (specificity) = 83.7%, and area under receiver operator characteristic curve = 92.4%. The present results contribute a new deep learning approach for automated detection of sleep apnea events from single channel respiration signals that can potentially serve as a helpful and alternative tool for the traditional PSG method.

Purpose The present study compared the effects of mandibular advancement therapy (MAD) with continuous positive airway pressure therapy (CPAP) on daytime cardiac autonomic modulation in a wide range of obstructive sleep apnea (OSA) patients under controlled conditions in a randomized, two-period crossover trial. Methods Forty OSA patients underwent treatment with MAD and with CPAP for 12 weeks each. At baseline and after each treatment period, patients were assessed by polysomnography as well as by a daytime cardiac autonomic function test that measured heart rate variability (HRV), continuous blood pressure (BP), and baroreceptor sensitivity (BRS) under conditions of spontaneous breathing, with breathing at 6, 12, and 15/min. Results Both CPAP and MAD therapy substantially eliminated apneas and hypopneas. CPAP had a greater effect. During daytime with all four conditions of controlled breathing, three-minute mean values of continuous diastolic BP were significantly reduced for both MAD and CPAP therapy. At the same time, selective increases due to therapy with MAD were found for HRV high frequency (HF) values. No changes were observed for BRS in either therapy mode. Conclusions These findings indicate that both MAD and CPAP result in similar beneficial changes in cardiac autonomic function during daytime, especially in blood pressure. CPAP is more effective than MAD in eliminating respiratory events.

In obstructive sleep apnea (OSA) esophageal pressure (Pes) is the gold standard for measurement of respiratory effort, and respiratory inductance plethysmography (RIP) is considered an accepted measurement technique. However, the use of RIP could lead to limited accuracy in certain cases and therefore suprasternal pressure (SSP) monitoring might improve the reliability of OSA diagnosis. We aimed to use SSP for the visual characterization of respiratory events in adults and compared results to those obtained by RIP from polysomnography (PSG). In patients with OSA, a 1-night SSP recording using the PneaVoX sensor (Cidelec, Sainte-Gemmes-sur-Loire, France) was done. In parallel, PSG was performed according to American Academy of Sleep Medicine criteria. A subgroup of patients agreed to have Pes measurement in addition. Characterizations of apneas as obstructive, central, and mixed as well as hypopneas as central and obstructive were done by visual evaluation of SSP, RIP, and Pes in random order by two independent scores (S1 and S2). The sensitivity and specificity of characterization by SSP compared to RIP and to Pes were calculated. Synchronous recordings of SSP and PSG were analyzed from n = 34 patients with OSA (AHI 34.1 ± 24.2 events/h); 9 of them had synchronized Pes monitoring as well. Interscorer agreement for apnea characterization as obstructive, central, and mixed based on SSP, RIP, and Pes were found, with R² values from 0.91-0.99. The sensitivity of SSP in apnea characterization with reference to RIP (S1/S2) was 91.5%/92.3% for obstructive, 82.7%/76.2% for central, and 87.4%/79.9% for mixed. The sensitivity of SSP in apnea characterization with reference to Pes was (S1/S2) 93.1%/92.1% for obstructive, 80.8%/81.6% for central, and 91.7%/90.8% for mixed. Hypopnea was only classified for the nine patients with Pes. This study demonstrated a good agreement in the detection of respiratory effort with the SSP signal using the PneaVoX sensor compared to the RIP belts signals as well as to the Pes signal. These findings were consistently found by two independent scorers. In summary, results suggest that SSP is a reliable signal for the classification of respiratory events and could be used as an additional tool for OSA characterization in clinical practice.

The scoring of sleep stages is an essential part of sleep studies. The main objective of this research is to provide an algorithm for the automatic classification of sleep stages using signals that may be obtained in a non-obtrusive way. After reviewing the relevant research, the authors selected a multinomial logistic regression as the basis for their approach. Several parameters were derived from movement and breathing signals, and their combinations were investigated to develop an accurate and stable algorithm. The algorithm was implemented to produce successful results: the accuracy of the recognition of Wake/NREM/REM stages is equal to 73%, with Cohen's kappa of 0.44 for the analyzed 19324 sleep epochs of 30 seconds each. This approach has the advantage of using the only movement and breathing signals, which can be recorded with less effort than heart or brainwave signals, and requiring only four derived parameters for the calculations. Therefore, the new system is a significant improvement for non-obtrusive sleep stage identification compared to existing approaches.