Electronic Department, Montefiore Institute, University of Liège (ULg), Building B28, Grande Traverse, Sart-Tilman, B4000, Liège, Belgium.
Sleep Breath. 2012 Jun;16(2):535-42. doi: 10.1007/s11325-011-0539-4. Epub 2011 Jun 11.
Estimating the total sleep time in home recording devices is necessary to avoid underestimation of the indices reflecting sleep apnea and hypopnea syndrome severity, e.g., the apnea-hypopnea index (AHI). A new method to distinguish sleep from wake using jaw movement signal processing is assessed.
In this prospective study, jaw movement signal was recorded using the Somnolter (SMN) portable monitoring device synchronously with polysomnography (PSG) in consecutive patients complaining about a lack of recovery sleep. The automated sleep/wake scoring method is based on frequency and complexity analysis of the jaw movement signal. This computed scoring was compared with the PSG hypnogram, the two total sleep times (TST(PSG) and TST(SMN)) as well.
The mean and standard deviation (in minutes) of TST(PSG) on the whole dataset (n = 124) were 407 ± 95.6, while these statistics were 394.2 ± 99.3 for TST(SMN). The Bland and Altman analysis of the difference between the two TST was 12.8 ± 57.3 min. The sensitivity and specificity (in percent) were 85.3 and 65.5 globally. The efficiency decreased slightly when AHI lies between 15 and 30, but remained similar for lower or greater AHI. In the 24 patients with insomnia/depression diagnosis, a mean difference in TST of -3.3 min, a standard deviation of 58.2 min, a sensitivity of 86.3%, and a specificity of 66.2% were found.
Mandible movement recording and its dedicated signal processing for sleep/wake recognition improve sleep disorder index accuracy by assessing the total sleep time. Such a feature is welcome in home screening methods.
在家庭记录设备中估计总睡眠时间对于避免低估反映睡眠呼吸暂停低通气综合征严重程度的指数（例如，呼吸暂停低通气指数[AHI]）非常重要。评估了一种使用下颌运动信号处理来区分睡眠和清醒的新方法。
在这项前瞻性研究中，使用 Somnolter（SMN）便携式监测设备同步记录下颌运动信号，该设备与连续抱怨睡眠恢复不足的患者进行多导睡眠图（PSG）同步记录。自动睡眠/觉醒评分方法基于下颌运动信号的频率和复杂度分析。将此计算评分与 PSG 催眠图、两个总睡眠时间（TST（PSG）和 TST（SMN））进行比较。
整个数据集（n=124）的 TST（PSG）平均值和标准差（以分钟为单位）分别为 407±95.6 和 394.2±99.3。两种 TST 之间差异的 Bland 和 Altman 分析为 12.8±57.3 min。总体灵敏度和特异性（以百分比表示）分别为 85.3%和 65.5%。当 AHI 在 15 到 30 之间时，效率略有下降，但对于较低或较高的 AHI，效率仍然相似。在 24 例失眠/抑郁诊断患者中，TST 平均差异为-3.3 min，标准差为 58.2 min，灵敏度为 86.3%，特异性为 66.2%。
下颌运动记录及其专用的睡眠/觉醒识别信号处理通过评估总睡眠时间来提高睡眠障碍指数的准确性。这种功能在家庭筛查方法中很受欢迎。
