Interdisciplinary Program in Medical and Biological Engineering, Seoul National University, Graduate School, Korea.
Physiol Meas. 2009 Dec;30(12):1327-40. doi: 10.1088/0967-3334/30/12/003. Epub 2009 Oct 28.
Polysomnography (PSG) is currently considered the gold standard for assessing sleep quality. However, the numerous sensors that must be attached to the subject can disturb sleep and limit monitoring to within hospitals and sleep clinics. If data could be obtained without such constraints, sleep monitoring would be more convenient and could be extended to ordinary homes. During rapid-eye-movement (REM) sleep, respiration rate and variability are known to be greater than in other sleep stages. Hence, we calculated the average rate and variability of respiration in an epoch (30 s) by applying appropriate smoothing algorithms. Increased and irregular respiratory patterns during REM sleep were extracted using adaptive and linear thresholds. When both parameters simultaneously showed higher values than the thresholds, the epochs were assumed to belong to REM sleep. Thermocouples and piezoelectric-type belts were used to acquire respiratory signals. Thirteen healthy adults and nine obstructive sleep apnea (OSA) patients participated in this study. Kappa statistics showed a substantial agreement (kappa > 0.60) between the standard and respiration-based methods. One-way ANOVA analysis showed no significant difference between the techniques for total REM sleep. This approach can also be applied to the non-intrusive measurement of respiration signals, making it possible to automatically detect REM sleep without disturbing the subject.
多导睡眠图（PSG）目前被认为是评估睡眠质量的金标准。然而，必须附着在受试者身上的众多传感器会干扰睡眠，并将监测限制在医院和睡眠诊所内。如果可以在没有这些限制的情况下获得数据，那么睡眠监测将更加方便，并可以扩展到普通家庭。在快速眼动（REM）睡眠期间，呼吸率和可变性已知大于其他睡眠阶段。因此，我们通过应用适当的平滑算法，在一个时段（30 秒）内计算呼吸的平均速率和可变性。使用自适应和线性阈值提取 REM 睡眠期间呼吸的不规则和增加的模式。当两个参数同时显示出高于阈值的更高值时，就假设该时段属于 REM 睡眠。热电偶和压电式皮带用于获取呼吸信号。13 名健康成年人和 9 名阻塞性睡眠呼吸暂停（OSA）患者参加了这项研究。Kappa 统计显示，标准方法和基于呼吸的方法之间存在高度一致性（kappa > 0.60）。单向方差分析显示，两种技术之间在总 REM 睡眠方面没有显著差异。这种方法也可以应用于非侵入性呼吸信号测量，从而可以在不干扰受试者的情况下自动检测 REM 睡眠。
