Oxygen saturation regularity analysis in the diagnosis of obstructive sleep apnea
Artificial Intelligence in Medicine
Real-time detection of apneas on a PDA
IEEE Transactions on Information Technology in Biomedicine
LIBSVM: A library for support vector machines
ACM Transactions on Intelligent Systems and Technology (TIST)
| Hi-index | 0.00 |
Sleep apnea, which is a common sleep disorder characterized by the repetitive cessation of breathing during sleep, can result in various diseases, including headaches, hypertension, stroke and cardiac arrest, as well as produce severe consequences such as impaired concentration and traffic accidents. A traditional diagnosis method of sleep apnea is polysomnography, which can only be conducted in sleep center with specialized personals, thus is expensive and inconvenient. Moreover, it is only used for understanding the conditions, without treatment function. Some other methods or devices have been developed to alleviate sleep apnea, such as continuous positive airway pressure (CPAP) and intraoral mandibular advancement device and surgery. However, they only provide a treatment method without detection or monitoring function. There is no existing device which can provide both apnea detection and treatment functionality. In this paper, we propose and implement a smart phone-based auto-adjustable pillow system, which enables both sleep apnea detection and treatment. Sleep apnea events can be detected in real-time using the blood oxygen sensor, accordingly, the height and shape of the pillow can be automatically adjusted to terminate the sleep apnea event. On the other hand, after the adjustment, the sensor can continuously monitor the blood oxygen signal to evaluate the effectiveness of the pillow adjustment and to help in selecting a suitable adjustment scheme. Therefore, a real-time feedback control system is formed. Besides, compared with existing diagnosis or treatment devices, our system is non-invasive, inexpensive and portable, which can be used at home or during traveling. In this paper, a real-time sleep apnea detection and classification algorithm is proposed to decide whether the pillow should be adjusted or not. We also design a real-time feedback pillow adjustment algorithm, to decide when and how to adjust the pillow and how to evaluate the effectiveness of the adjustment. We conducted experiments on 40 patients, which demonstrate that using our novel smart pillow system, both the sleep apnea duration and the number of sleep apnea events are dramatically reduced by more than 50%.