根據台灣睡眠醫學會（TSSM）2017年的一項調查，台灣1/10的人口患有慢性失眠症，患病率為11.3%，超過200萬人患有睡眠障礙。對於睡眠障礙患者，目前的臨床檢測方法是使用睡眠多導睡眠圖（PSG）記錄患者整夜的生理信號，然後人工對睡眠階段進行判讀，以供專家或醫生診斷參考。手動判讀是一項非常主觀且耗時的任務。經驗豐富的睡眠技師對報告進行判讀需要 2-3 個小時。在醫院的睡眠中心，每個月有一百多條數據需要睡眠技師打分。判讀結果上報會影響醫生對病情的判斷和治療。因此，TSSM的睡眠中心認證標準，即認證實施細則中的第一、五項“質量控制”，明確規定擁有兩名以上睡眠技師的專業睡眠機構需要至少每六個月評估一次睡眠技師判讀的可靠性，這顯示了正確判讀和一致性的重要性。
本研究提出了一套智慧睡眠輔助判讀系統應用於睡眠技師評估訓練與人機協同自動判讀。該系統藉由雲端平台提供技師進行判讀並透過一連串流程設計，比對技師判讀差異後提供交流討論，爾後再進行二次判讀，最終給予分析報告。經兩次前驅研究號證實，藉此平台及流程能讓技師一致性有效提升。遂於今年六月推廣至全台，本系、成大睡眠中心與台灣睡眠醫學學會共同舉辦「2021智慧睡眠判讀訓練工作坊-全台首次運用智慧睡眠雲端平台進行精準睡眠判讀訓練」，全程使用本系統進行判讀及得到比對結果。最終推廣至全台35家睡眠中心共計81位技師使用本系統，技師間睡眠期的判讀一致性由原本73.3%上升至80.6%，其中睡眠指標如N1比例、N2比例、N3比例之改善有顯著差異，其餘如睡眠效率、入眠時間等睡眠指標也都有更趨近於答案的趨勢。
除了協助提高睡眠技師間的判讀可靠性，同時，該系統匯集了多位睡眠技師對本次研究中睡眠判讀歧義的討論，重新整理並提供睡眠技師作為再培訓的參考。並且利用這些討論結果，使用可以整合專家知識的基於規則的決策樹，以美國睡眠醫學會 (AASM) 制定的標準為基礎，並輔以與睡眠技師討論結果有方向修改。因此，本研究開發了一種結合人機協同的界面，對系統自動判讀結果中不太可靠的部分進行標記。睡眠技師可以針對必要部分進行修正，有效實用地使用自動判讀系統，減少判讀時間。希望該系統能夠起到輔助睡眠睡眠技師的作用，利用該系統進行定期測試，提高睡眠技師的判讀一致性，並且有效減輕他們的工作量。

According to a 2017 survey by the Taiwan Society of Sleep Medicine (TSSM), 1/10 of Taiwan’s population suffers from chronic insomnia, with a prevalence rate of 11.3%, and more than 2 million people suffer from sleep disorders. For patients with sleep disorders, current clinical detection method uses sleep polysomnography (PSG) to record patient's physiological signals throughout the night and then manually score sleep stages for diagnosis. Manual scoring is very subjective and time-consuming. It takes 2-3 hours for an experienced sleep technician to interpret a report. In the hospital's sleep center, there are hundreds of data sets needed to be scored by sleep technicians every month. Reports of scoring results will affect doctor's judgments and treatments. Therefore, TSSM’s sleep center certification standards, namely the first and fifth "quality control" in the certification implementation rules, clearly stipulate that professional sleep institutions with more than two sleep technicians assess reliability between technicians at least every six months.
Since accuracy and coherence are important factors in sleep assessment reports, we propose an intelligent computer- aided sleep scoring system for inter-scorer reliability enhancement and human-machine collaborative scoring. A cloud platform is provided for scoring and comparing inconsistencies between technicians through a serial programming. After communication, experts conduct their second scoring and write their final analysis report. It has been confirmed by two pilot studies that the cloud platform and the serial programming can effectively improve scoring consistency between technicians. Since then, we had been promoted the application across the country in June this year (2021). The Sleep Medicine Center of National Cheng Kung University Hospital and the Taiwan Society of Sleep Medicine jointly organized the " 2021 Intelligence PSG Scoring Training Workshop" first use the Cloud Platform for accurate sleep scoring training in Taiwan. The system performs scoring and obtains comparative results, moreover, it was promoted to a total of 81 technicians in 35 sleep centers in Taiwan. The consistency of scoring of sleep period among technicians increased from 73.3% to 80.6%. Sleep indicators such as N1 ratio, N2 ratio, and N3 ratio have been improved significantly; other sleep indicators such as sleep efficiency, sleep onset time also have a trend to approach.
In addition to helping to improve scoring reliability among sleep technicians, the system has also gathered several discussions from sleep technicians on sleep scoring ambiguities. It can provide a reference for retraining for sleep technicians. By using rule-based decision tree based on the standards set by the American Academy of Sleep Medicine (AASM), and supplemented by the opinions of sleep technicians, we developed a human-machine collaboration interface to mark the least reliable parts of the system's automatic scoring results. Sleep technicians can effectively compare different scoring results and can also efficiently apply the automatic scoring system. Finally, we hope that this system can play an important role in assisting sleep technicians by automatically improving the consistency of scoring to reduce their workload.

Agarwal, R., & Gotman, J. (2001). Computer-assisted sleep staging. IEEE Transactions on Biomedical Engineering, 48(12), 1412-1423.
Anderer, P., Gruber, G., Parapatics, S., Woertz, M., Miazhynskaia, T., Klösch, G., . . . Danker-Hopfe, H. (2005). An E-health solution for automatic sleep classification according to Rechtschaffen and Kales: validation study of the Somnolyzer 24× 7 utilizing the Siesta database. Neuropsychobiology, 51(3), 115-133.
Chang, T.-H. (2017). An Automatic Sleep Scoring System for Accurate Estimation of Various Sleep Measures. (Master), National Cheng Kung University, Taiwan.
Danker-Hopfe H, Anderer P, Zeitlhofer J, et al. (2009) Interrater reliability for sleep scoring according to the Rechtschaffen & Kales and the new AASM standard. J Sleep Res ,18,74-84.
Duman, F., Erdamar, A., Erogul, O., Telatar, Z., & Yetkin, S. (2009). Efficient sleep spindle detection algorithm with decision tree. Expert Systems with Applications, 36(6), 9980-9985.
Iber, C. (2007). The AASM manual for the scoring of sleep and associated events: rules, terminology and technical specifications: American Academy of Sleep Medicine.
Imtiaz, S. A., & Rodriguez-Villegas, E. (2015). Automatic sleep staging using state machine-controlled decision trees. Paper presented at the Engineering in Medicine and Biology Society (EMBC), 2015 37th Annual International Conference of the IEEE.
Jansen, B. H., & Dawant, B. M. (1989). Knowledge-based approach to sleep EEG analysis-a feasibility study. IEEE Transactions on Biomedical Engineering, 36(5), 510-518.
Jorgensen, G., Downey, C., Goldin, J. et al. (2020) An Australasian Commentary on the AASM Manual for the Scoring of Sleep and Associated Events. Sleep Biol. Rhythms 18, 163–185.
Kuo, C.-E., Liang, S.-F., Lee, Y.-C., Cherng, F.-Y., Lin, W.-C., Chen, P.-Y., . . . Shaw, F.-Z. (2014). An EOG-based automatic sleep scoring system and its related application in sleep environmental control. Paper presented at the International Conference on Physiological Computing Systems.
Li, X., Cui, L., Tao, S., Chen, J., Zhang, X., & Zhang, G.-Q. (2017). HyCLASSS: A Hybrid Classifier for Automatic Sleep Stage Scoring. IEEE Journal of Biomedical and Health Informatics.
Liang, S.-F., Kuo, C.-E., Hu, Y.-H., & Cheng, Y.-S. (2012). A rule-based automatic sleep staging method. Journal of neuroscience methods, 205(1), 169-176.
Liu, G. R., Lin, T. Y., Wu, H. T., Sheu, Y. C., Liu, C. L., Liu, W. T., Yang, M. C., Ni, Y. L., Chou, K. T., Chen, C. H., Wu, D., Lan, C. C., Chiu, K. L., Chiu, H. Y., & Lo, Y. L. (2021). Large-scale assessment of consistency in sleep stage scoring rules among multiple sleep centers using an interpretable machine learning algorithm. Journal of Clinical Sleep Medicine, 17(2), 159-166.
Magalang UJ, Chen NH, Cistulli PA, Fedson AC, Gíslason T, Hillman D, Penzel T, Tamisier R, Tufik S, Phillips G, Pack AI; SAGIC Investigators. (2013). Agreement in the scoring of respiratory events and sleep among international sleep centers. Sleep, 36(4), 591-6.
Mignot, E. (2008). Why we sleep: the temporal organization of recovery. PLoS Biol, 6(4), e106. Development of a human-computer collaborative sleep scoring system for polysomnography recordings
Norman, R. G., Pal, I., Stewart, C., Walsleben, J. A., & Rapoport, D. M. (2000). Interobserver agreement among sleep scorers from different centers in a large dataset. Sleep, 23(7), 901-908.
Park, H., Park, K., & Jeong, D.-U. (2000). Hybrid neural-network and rule-based expert system for automatic sleep stage scoring. Paper presented at the Engineering in Medicine and Biology Society, 2000. Proceedings of the 22nd Annual International Conference of the IEEE.
Penzel T, Zhang X, Fietze I. (2013). Inter-scorer reliability between sleep centers can teach us what to improve in the scoring rules. J Clin Sleep Med, 9(1), 89-91.
Penzel T, Behler PG, von Buttlar M, et al. (2003) Reliability of visual evaluation of sleep stages according to Rechtschaffen and Kales from eight polysomnographs by nine sleep centers. Somnologie,7,49-58.
Rechtschaffen, A., & Kales, A. (1968). A manual of standardized terminology, techniques and scoring system for sleep stages of human subjects.
Richard B.Berry, M. (2017). The AASM Manual for the Scoring of Sleep and Associated Events: rules, terminology and technical specifications (version 2.4 ed.): American Academy of Sleep Medicine.
Rosenberg, R. S., Van Hout, S. (2013). The American Academy of Sleep Medicine inter-scorer reliability program: sleep stage scoring. J Clin Sleep Med, 9(1), 81-87.
Smith, J. R., Negin, M., & Nevis, A. H. (1969). Automatic analysis of sleep electroencephalograms by hybrid computation. IEEE transactions on systems science and cybernetics, 5(4), 278-284.
Tian, J., & Liu, J. (2006). Automated sleep staging by a hybrid system comprising neural network and fuzzy rule-based reasoning. Paper presented at the 2005 IEEE Engineering in Medicine and Biology 27th Annual Conference.
Whitney, C. W., Gottlieb, D. J., Redline, S., Norman, R. G., Dodge, R. R., Shahar, E., . . . Nieto, F. J. (1998). Reliability of scoring respiratory disturbance indices and sleep staging. Sleep, 21(7), 749-757.
Younes M, Raneri J, Hanly P. (2016). Staging Sleep in Polysomnograms: Analysis of Inter-Scorer Variability. J Clin Sleep Med, 12(6), 885-94.
Zhang X, Dong X, Kantelhardt JW, Li J, Zhao L, Garcia C, Glos M, Penzel T, Han F. (2015). Process and outcome for international reliability in sleep scoring. Sleep Breath, 19(1), 191-5.