傳統上，要評估睡眠，會使用多項睡眠生理檢查(Polysomnography, PSG)。要進行這項診斷，受試者需要在睡眠實驗室睡一晚，並且在身體以及頭上各處連接電極點。除此之外，這些記錄的資料還需要經過有經驗的專家進行判讀。雖然這是黃金標準，卻不適合用於長期的睡眠監測。因此需要其他替代的裝置來測量睡眠。目前市面上有許多測量睡眠的裝置，但是過去的研究發現這些裝置測量到的睡眠指標與多項睡眠生理檢查相比有偏差。
在本篇研究中，我們使用腕動計來偵測睡眠。為了開發總清醒-睡眠演算法，我們需要人工判讀的多項睡眠生理檢查與相對應的腕動計資料。總共有33名健康成年人與34位患有焦慮或憂鬱症的人參與了這個研究。每位受試者提供了1到4個晚上的睡眠，這些紀錄跨越了幾個月。共蒐集了195筆睡眠資料。
我們搜尋了各項參數來讓每筆資料的Cohen’s kappa 係數能夠最大化。演算法準確度為90.72%，靈敏度為94.98%，特異度為57.70%，Cohen’s kappa係數為0.535。此外，這個演算法所估計的各項睡眠指標(包含睡眠效率、睡眠延遲、入睡後醒來的總時數、總睡眠時間)與專家判讀的PSG之間沒有顯著差異。
這個演算法佔用了很少CPU與RAM的資源。對於要處理長達20小時的睡眠資料，只需要1026個bytes。這項演算法被實作在腕動計當中。韌體的內部快閃記憶體與RAM的使用量分別為26.14 kB以及13.81 kB。這項設備透過行動裝置來控制。清醒-睡眠階段透過低功耗藍芽傳輸到行動裝置。應用程式能夠顯示睡眠階段以及各項睡眠指標。另外，應用程式還能透過折線圖來顯示一週、二週、或是四週的睡眠指標，讓使用者能夠方便追蹤長期的睡眠趨勢。在一般使用下測量電池的續航力。在電池沒電之前，共紀錄了31晚的睡眠資料。因此，這項裝置適合應用於長期睡眠紀錄。

Traditionally, polysomnography (PSG) has been used for sleep monitoring. To conduct this diagnosis, the subject needs to sleep in the laboratory and place electrodes on his/her body and head. In addition, the recorded data needs to be scored by a well-trained expert. Even if this is the gold standard, it cannot be used for long-term sleep monitoring. Therefore, we need an alternative device to monitor sleep. There are many products to choose from on the market. However, previous study has shown that some objective sleep measurements estimated by these devices are biased compared to manually scored PSG.
In this study, the actigraphy device are chosen for sleep monitoring. In order to develop a wake-sleep staging algorithm for actigraphy devices, we need manually scored PSGs and corresponding actigraphy recordings. A total of 33 healthy adults and 34 adults with anxiety or depressive disorder participated in the study. Each subject provided 1 to 4 nights of sleep, and these recordings span several months. A total of 195 nights of sleep were recorded.
Parameters were searched to maximize Cohen's kappa coefficient for each recording. The accuracy, sensitivity, specificity and Cohen's kappa coefficient of the algorithm reached 90.72%, 94.98%, 57.70% and 0.535, respectively. In addition, there is no significant difference between the sleep measurements (SE, SOT, WASO, and TST) estimated by the algorithm and the manually scored PSG.
This algorithm consumes very few CPU and RAM resources. For sleep data processing up to 20 hours, only 1026 bytes are required. The algorithm was implemented in the actigraphy device. The internal flash memory usage and RAM usage of the firmware are 26.14 kB and 13.81 kB, respectively. The device can be controlled by a mobile device. The wake-sleep stages are transmitted to the mobile device via BLE. The stages and sleep measurements are displayed in the mobile application. The application also displays 1-week, 2-week or 4-week sleep measurements in the form of line charts, which are convenient for users to track long-term sleep trends. The battery life is tested under normal use. Before the battery was dead, 31 nights of sleep had been recorded. Therefore, the device is suitable for long-term sleep recording.

Altman, D. G., & Bland, J. M. (1983). Measurement in medicine: the analysis of method comparison studies. Journal of the Royal Statistical Society: Series D (The Statistician), 32(3), 307-317.
Ancoli-Israel, S., Cole, R., Alessi, C., Chambers, M., Moorcroft, W., & Pollak, C. P. (2003). The role of actigraphy in the study of sleep and circadian rhythms. Sleep, 26(3), 342-392.
Bland, J. M., & Altman, D. (1986). Statistical methods for assessing agreement between two methods of clinical measurement. The lancet, 327(8476), 307-310.
Brønd, J. C., Andersen, L. B., & Arvidsson, D. (2017). Generating ActiGraph counts from raw acceleration recorded by an alternative monitor.
Brooks, C. M., Hoffman, B., Suckling, E., Kleyntjens, F., Koenig, E., Coleman, K., & Treumann, H. (1956). Sleep and variations in certain functional activities accompanying cyclic changes in depth of sleep. Journal of applied physiology, 9(1), 97-104.
Bušek, P., Vaňková, J., Opavský, J., Salinger, J., & Nevšímalová, S. (2005). Spectral analysis of heart rate variability in sleep. Physiol res, 54(4), 369-376.
Cajochen, C., Pischke, J., Aeschbach, D., & Borbély, A. A. (1994). Heart rate dynamics during human sleep. Physiology & behavior, 55(4), 769-774.
Chang, D.-W., Liang, S.-F., Young, C.-P., Shaw, F.-Z., Su, A. W., Liu, Y.-D., . . . Chen, C.-Y. (2011). A versatile wireless portable monitoring system for brain–behavior approaches. IEEE Journal on Emerging and Selected Topics in Circuits and Systems, 1(4), 440-450.
Chang, D.-W., Liu, Y.-D., Young, C.-P., Chen, J.-J., Chen, Y.-H., Chen, C.-Y., . . . Liang, S.-F. (2012). Design and implementation of a modularized polysomnography system. IEEE Transactions on Instrumentation and Measurement, 61(7), 1933-1944.
Chen, K. Y., & Bassett, D. R. (2005). The technology of accelerometry-based activity monitors: current and future. Medicine and science in sports and exercise, 37(11), S490.
Chinoy, E. D., Cuellar, J. A., Huwa, K. E., Jameson, J. T., Watson, C. H., Bessman, S. C., . . . Markwald, R. R. (2021). Performance of seven consumer sleep-tracking devices compared with polysomnography. Sleep, 44(5), zsaa291.
Cohen, J. (1960). A coefficient of agreement for nominal scales. Educational and psychological measurement, 20(1), 37-46.
Cornelissen, G. (2014). Cosinor-based rhythmometry. Theoretical Biology and Medical Modelling, 11(1), 1-24.
Kuo, C.-E., Liu, Y.-C., Chang, D.-W., Young, C.-P., Shaw, F.-Z., & Liang, S.-F. (2016). Development and evaluation of a wearable device for sleep quality assessment. IEEE Transactions on Biomedical Engineering, 64(7), 1547-1557.
Lichstein, K., Durrence, H., Taylor, D., Bush, A., & Riedel, B. (2003). Quantitative criteria for insomnia. Behaviour research and therapy, 41(4), 427-445.
Liu, Y.-S. (2020). Actigraphy platform for objective and long-term sleep assessment. (Master), National Cheng-Kung University, Tainan. Retrieved from https://hdl.handle.net/11296/d76gme
Matsuo, M., Masuda, F., Sumi, Y., Takahashi, M., Yamada, N., Ohira, M. H., . . . Kadotani, H. (2016). Comparisons of portable sleep monitors of different modalities: potential as naturalistic sleep recorders. Frontiers in neurology, 7, 110.
Mendez, M., Matteucci, M., Cerutti, S., Aletti, F., & Bianchi, A. M. (2009). Sleep staging classification based on HRV: time-variant analysis. Paper presented at the 2009 Annual International Conference of the IEEE Engineering in Medicine and Biology Society.
Mullaney, D., Kripke, D., & Messin, S. (1980). Wrist-actigraphic estimation of sleep time. Sleep, 3(1), 83-92.
Provost, F., & Kohavi, R. (1998). Glossary of terms. Journal of Machine Learning, 30(2-3), 271-274.
Sadeh, A., & Acebo, C. (2002). The role of actigraphy in sleep medicine. Sleep medicine reviews, 6(2), 113-124.
Sadeh, A., Hauri, P. J., Kripke, D. F., & Lavie, P. (1995). The role of actigraphy in the evaluation of sleep disorders. Sleep, 18(4), 288-302.
Sasaki, J. E., John, D., & Freedson, P. S. (2011). Validation and comparison of ActiGraph activity monitors. Journal of science and medicine in sport, 14(5), 411-416.
Smith, M. T., & Wegener, S. T. (2003). Measures of sleep: the insomnia severity index, medical outcomes study (MOS) sleep scale, Pittsburgh sleep diary (PSD), and Pittsburgh sleep quality index (PSQI). Arthritis Care & Research: Official Journal of the American College of Rheumatology, 49(S5), S184-S196.
Witting, W., Kwa, I., Eikelenboom, P., Mirmiran, M., & Swaab, D. F. (1990). Alterations in the circadian rest-activity rhythm in aging and Alzheimer's disease. Biological psychiatry, 27(6), 563-572.
Xiao, M., Yan, H., Song, J., Yang, Y., & Yang, X. (2013). Sleep stages classification based on heart rate variability and random forest. Biomedical Signal Processing and Control, 8(6), 624-633.
Yuda, E., Yoshida, Y., Sasanabe, R., Tanaka, H., Shiomi, T., & Hayano, J. (2017). Sleep stage classification by a combination of actigraphic and heart rate signals. Journal of Low Power Electronics and Applications, 7(4), 28.
ŽEmaitytė, D., Varoneckas, G., & Sokolov, E. (1984). Heart rhythm control during sleep. Psychophysiology, 21(3), 279-289.