臺灣博碩士論文加值系統

睡眠紡錘波是透過腦電圖（EEG）測量, 主要是在睡眠期間非快速眼動（NREM）第二階段測量腦活動的短暫振動西格瑪頻率範圍（11-16Hz）。這些振動具有很大的生物和臨床意義，它們在各種學習與認知功能開發學習領域及複雜的神經系統中是重要的生物標記。通常，睡眠紡錘波由睡眠臨床專家目測腦電信號來辨識判定。這個過程非常耗時，而且不同專家之間的結果並不一致。為了解決這個問題，目前腦科學家已經發展了許多自動化睡眠紡錘波檢測方法。然而，在不同研究中, 這些自動睡眠紡錘波的檢測方法表現並不盡相同, 這主要有兩個原因：（1）缺乏共同的基準測試數據庫，（2）缺乏被腦科學界普遍接受的評估指標。在本研究中，我們專注於解決第二個問題，提出在多目標優化的環境中評估睡眠紡錘波檢測的效能。我們實驗假設，使用Pareto fronts來導出評估度量將提高自動睡眠紡錘波檢測。我們使用盛行於工程優化用用途的多目標演化演算法（MOEA），Strength Pareto Evolutionary Algorithm（SPEA2）來優化六種現有的以頻率為基準的睡眠紡錘波檢測演算法。它們包括三個傅立葉，一個連續小波變換（CWT）和兩個希爾伯特 - 黃變換（HHT）演算法。我們還探討了三種混合型方法。在使用公開取得的DREAMS和MASS數據庫進行了訓練和測試，兩種新的傅立葉與HHT演算法的混合型方法顯示出顯著的效能提升，F1分數達0.726-0.737的高準確度。

Sleep spindles are brief bursts of brain activity in the sigma frequency range (11–16 Hz) measured by electroencephalography (EEG) mostly during non-rapid eye movement (NREM) stage 2 sleep. These oscillations are of great biological and clinical interests because they potentially play an important role in identifying and characterizing the processes of various neurological disorders. Conventionally, sleep spindles are identified by expert sleep clinicians via visual inspection of EEG signals. The process is laborious and the results are inconsistent among different experts. To resolve the problem, numerous computerized methods have been developed to automate the process of sleep spindle identification. Still, the performance of these automated sleep spindle detection methods varies inconsistently from study to study. There are two reasons: (1) the lack of common benchmark databases, and (2) the lack of commonly accepted evaluation metrics. In this study, we focus on tackling the second problem by proposing to evaluate the performance of a spindle detector in a multi-objective optimization context and hypothesize that using the resultant Pareto fronts for deriving evaluation metrics will improve automatic sleep spindle detection. We use a popular multi-objective evolutionary algorithm (MOEA), the Strength Pareto Evolutionary Algorithm (SPEA2), to optimize six existing frequency-based sleep spindle detection algorithms. They include three Fourier, one continuous wavelet transform (CWT), and two Hilbert-Huang transform (HHT) based algorithms. We also explore three hybrid approaches. Trained and tested on open-access DREAMS and MASS databases, two new hybrid methods of combining Fourier with HHT algorithms show significant performance improvement with F1-scores of 0.726–0.737.

中文摘要 I
ABSTRACT II
TABLE OF CONTENTS III
LIST OF TABLES VI
LIST OF FIGURES VII
CHAPTER 1 INTRODUCTION 1
CHAPTER 2 MATERIAL AND METHODS 6
2.1 DATA 6
2.1.1 DREAM 6
2.1.2 MASS 7
2.2 PERFORMANCE EVALUATION 8
2.3 PARETO FRONT-DERIVED PERFORMANCE METRICS 10
2.4 ROLLING BALL EMD 14
2.4.1 MOTIVATION 14
2.4.2 ALPHA-SHAPE AND DELAUNAY TRIANGULATION 16
2.4.3 ROLLING BALL SIFTING 18
2.5 SIX SIMPLEX DETECTORS 21
2.6 THREE HYBRIDIZATION DETECTORS 27
2.7 SUBSAMPLE STRATEGY 29
2.8 STRENGTH PARETO EVOLUTIONARY ALGORITHM (SPEA2) 29
2.8.1 FITNESS ASSIGNMENT 31
2.8.2 ENVIRONMENTAL SELECTION 32
2.8.3 SPEA2 MODULE 33
2.9 STATISTICS 33
2.10 SOFTWARE IMPLEMENTATION 35
CHAPTER 3 RESULTS 37
3.1 SPINDLE DETECTION PERFORMANCE ON DREAMS DATABASE 37
3.2 SPINDLE DETECTOR HOLD-OUT VALIDATION ON MASS DATABASE 40
3.3 SPINDLE DETECTOR 3-FOLD CROSS-VALIDATION ON MASS DATABASE 47
3.4 OPTIMIZED OPERATING PARAMETERS 47
3.5 STATISTICAL ANALYSIS 54
3.6 COMPUTATION TIME 60
CHAPTER 4 DISCUSSION AND CONCLUSION 61
REFERENCE 66
SUPPLEMENTAL MATERIAL 71

[1] Adamczyk, M., Genzel, L., Dresler, M., Steiger, A., and Friess, E. (2015). Automatic sleep spindle detection and genetic influence estimation using continuous wavelet transform. Front. Hum. Neurosci. 9:624. doi: 10.3389/fnhum.2015.00624
[2] Bergmann, T. O., Mölle, M., Diedrichs, J., Born, J., and Siebner, H. R. (2012). Sleep spindle-related reactivation of category-specific cortical regions after learning face-scene associations. Neuroimage, 59(3), 2733-2742. doi: 10.1016/j.neuroimage.2011.10.036
[3] Bódizs, R., Kis, T., Lázár, A. S., Havran, L., Rigó, P., Clemens, Z., et al. (2005). Prediction of general mental ability based on neural oscillation measures of sleep. J. Sleep Res. 14(3), 285-92. doi: 10.1111/j.1365-2869.2005.00472.x
[4] Bódizs, R., Körmendi, J., Rigó, P. & Lázár, A.S. (2009). The individual adjustment method of sleep spindle analysis: methodological improvements and roots in the fingerprint paradigm. J. Neurosci. Methods. 178(1), 205-213. doi: 10.1016/j.jneumeth.2008.11.006
[5] Brun, M., Xu, Q., and Dougherty, E. R. (2008). Which is better: holdout or full-sample classifier design? EURASIP Journal on Bioinformatics and Systems Biology 8:297945. doi: 10.1155/2008/297945
[6] Causa, L., Held, C. M., Causa, J., Estévez, P. A., Perez, C. A., Chamorro, R., et al. (2010). Automated sleep-spindle detection in healthy children Polysomnograms. IEEE Trans. Biomed. Eng. 57(9), 2135-2146. doi: 10.1109/TBME.2010.2052924
[7] Crowley, K., Trinder, J., Kim, Y., Carrington, M., and Colrain, I. M. (2002). The effects of normal aging on sleep spindle and K-complex production. Clin. Neurophysiol. 113, 1615-1622. doi: 10.1016/S1388-2457(02)00237-7
[8] Deb, K., Pratap, A., Agarwal, S., and Meyarivan, T. (2002). A fast and elitist multiobjective genetic algorithm: NSGA-II. IEEE Trans. Evolutionary Computation 6(2), 182-197. doi: 10.1109/4235.996017
[9] Deb, K., Mohan, M., and Mishra, S. (2005). Evaluation the epsilon-domination based multi objective evolutionary algorithm for a quick computation of Pareto-optimal solutions. Evolutionary Computation 13(4), 501-525. doi: 10.1162/106365605774666895
[10] Devuyst, S., Dutoit, T., Stenuit, P., and Kerkhofs M. (2011). “Automatic sleep spindles detection-overview and development of a standard proposal assessment method,” in Proceedings of the annual International Conference of the IEEE Engineering in Medicine and Biology Society (Boston,MA), 1713-1716.
[11] Devuyst, S. (2013). The DREAMS Sleep Spindles Database. Available online at: http://www.tcts.fpms.ac.be/~devuyst/Databases/DatabaseSpindles/
[12] Doncieux, S., Mouret, J.-B., and Bredeche, N. (eds.) (2011). New Horizons in Evolutionary Robotics. Springer, Berlin-Heidelberg. doi: 10.1007/978-3-642-18273-3
[13] Edelsbrunner, H., Kirkpatrick, D. G. and Seidel, R. (1983). On the shape of a set of points in the plane. IEEE Trans. Inf. Theory, IT-29(4): 551–559.
[14] Edelsbrunner, H. and M¨ucke, E. P. (1994). Three-dimensional alpha shapes, ACM Trans. Graph., 13(1): 43–72.
[15] Eschenko, O., Mölle, M., Born, J., and Sara, S. J. (2006). Elevated sleep spindle density after learning of after retrieval in rats. J. Neurosci. 26(50), 12914-20. doi:10.1523/JNEUROSCI.3175-06.2006
[16] Ferrarelli, F., Huber, R., Peterson, M. J., Massimini, M., Murphy, M., Riedner B. A., et al. (2007). Reduced sleep spindle activity in schizophrenia patients. Am. J. Psychiatry.164(3), 483-492. doi: 10.1176/ajp.2007.164.3.483
[17] Fogel, S. M., Nader, R., Cote, K. A., and Smith, C. T. (2007). Sleep spindles and learning potential. Behav. Neurosci. 121, 1-10. doi: 10.1037/0735-7044.121.1.1
[18] Geiger, A., Huber, R., Kurth, S., Ringli, M., Jenni, O. G., and Achermann, P. (2011). The sleep EEG as a marker of intellectual ability in school age children. Sleep 34(2), 181-189.
[19] Goutte, C., and Gaussier, E. (2005). “A probabilistic interpretation of precision, recall and F-score, with implication for evaluation,” in Proceedings of 27th European conference on IR research (ECIR 2005), 345–359.
[20] Gruber, R., Wise, M. S., Frenette, S., Knäauper, B., Boom, A., Fontil, L., et al. (2013). The association between sleep spindles and IQ in healthy school-age children. Int. J. Psychophysiol. 89, 229-40. doi: 10.1016/j.ijpsycho.2013.03.018
[21] Huang, N. E., Shen, Z., Long, S. R., Wu, M. C., Shih, H. H., Zheng, Q., et al. (1998). The empirical mode decomposition and the Hilbert spectrum for nonlinear and non-stationary time series analysis. Proc. Roy. Soc. Lond. 454A, 903-995. doi: 10.1098/rspa.1998.0193
[22] Huang, N. E., Wu, Z., Long, S. R., Arnold, K. C., Chen, X., and Blank, K. (2009). On instantaneous frequency. Advances in Adaptive Data Analysis 1(2), 177-229. doi: 10.1142/S1793536909000096
[23] Huang, A., Li, J., Summers, R. M., Petrick, N., and Hara, A. K. (2010). Improving polyp detection algorithms for CT colonography: pareto front approach. Pattern Recognition Letters 31(11), 1461-1469. doi: 10.1016/j.patrec.2010.03.013
[24] Huang, A., Liu, M. Y., and Yu, T. W. (2015). Bandpass empirical mode decomposition using a rolling ball algorithm. Advances in Adaptive Data Analysis 7(1):1550003. doi: 10.1142/S179353691550003X
[25] Huang, A., Lee, C.-W., and Liu, H.-M. (2016). Rolling ball sifting algorithm for the augmented visual inspection of carotid bruit auscultation. Sci. Rep. 6:30179. doi: 10.1038/srep30179
[26] Huupponen, E., Gómez-Herrero, G., Saastamoinen, A., Värri, A., Hasan, J., and Himanen S. L. (2007). Development and comparison of four sleep spindle detection methods. Artificial Intelligence in Medicine 40(3), 157-170. doi: 10.1016/j.artmed.2007.04.003
[27] Iber, C., Ancoli-Israel, S., Chesson, A. L., and Quan, S. F. (2007). The AASM manual for the scoring of sleep and associated events: rules, terminology, and technical specifications. Westchester, IL: American Academy of Sleep Medicine.
[28] Knowles, J.D., and Corne, D.W. (2000). Approximating the nondominated front using the Pareto archived evolution strategy. Evolutionary Computation 8(2), 149-172. doi: 10.1162/106365600568167
[29] Lajnef, T., Chaibi, S., Eichenlaub, J.-B., Ruby, P. M., Aguera, P.-E., Samet, M., et al. (2015). Sleep spindle and K-complex detection using tunable Q-factor wavelet transform and morphological component analysis. Front. Hum. Neurosci. 9:414. doi: 10.3389/fnhum.2015.00414
[30] Latreille, V., Carrier, J., Lafortune, M., Postuma, R. B., Bertrand, J. A., Panisset, M., et al. (2015). Sleep spindles in Parkinson's disease may predict the development of dementia. Neurobiology of Aging 36(2),1083-90. doi: 10.1016/j.neurobiolaging.2014.09.009
[31] Martin, N., Lafortune, M., Godbout, J., Barakat, M., Robillard, R., Poirier, G., et al. (2013). Topography of age-related changes in sleep spindles. Neurobiology of Aging 34(2), 468-476. doi: 10.1016/j.neurobiolaging.2012.05.020
[32] Messac, A., Ismail-Yahaya, A., and Mattson, C. A. (2003). The Normalized Normal Constraint Method for Generating the Pareto Frontier. Structural and Multidisciplinary Optimization 25(2), 86-98. doi: 10.1007/s00158-002-0276-1
[33] Mölle, M., Marshall, L., Gais, S., and Born, J. (2002). Grouping of spindle activity during slow oscillations in human non-rapid eye movement sleep. J. Neurosci. 22, 10941-10947
[34] O’Reilly, C. (2013). Spyndle. Available online at: https://bitbucket.org/christian_oreilly/spyndle
[35] O’Reilly, C., Gosselin, N., Carrier, J., and Nielsen, T. (2014). Montreal Archive of Sleep Studies: an open-access resource for instrument benchmarking and exploratory research. J. Sleep Res. 23, 628-635. doi: 10.1111/jsr.12169
[36] O’Reilly, C., and Nielsen, T. (2015). Automatic sleep spindle detection: benchmarking with fine temporal resolution using open science tools. Front. Hum. Neurosci. 9:353. doi: 10.3389/fnhum.2015.00353
[37] Popov, A. (2005). SPEA2 for Matlab. Available online at: http://p0p0v.com/science/downloads/MOEA_SPEA2.zip
[38] Schabus, M., Hoedlmoser, K., Pecherstorfer, T., Anderer, P., Gruber, G., Parapatics, S., et al. (2008). Interindividual sleep spindle differences and their relation to learning-related enhancements. Brain Res.1191,127-35. doi: 10.1016/j.brainres.2007.10.106
[39] Schimicek, P., Zeitlhofer, J., Anderer, P., and Saletu, B. (1994). Automatic sleep-spindle detection procedure: aspects of reliability and validity. Clin. Electroencephalogr. 25, 26-29. doi: 10.1177/15500594902500108
[40] Sitnikova, E., Hramov, A. E., Koronovsky, A. A., and van Luijtelaar, G. (2009). Sleep spindles and spike-wave discharges in EEG: their generic features, similarities and distinctions disclosed with Fourier transform and continuous wavelet analysis. J. Neurosci. Methods 180(2), 304-316. doi: 10.1016/j.jneumeth.2009.04.006
[41] Tamminen, J., Payne, J. D. , Stickgold, R., Wamsley, E. J., and Gaskell, M. G. (2010). Sleep spindle activity is associated with the integration of new memories and existing knowledge. J. Neurosci. 30, 14356–14360. doi: 10.1523/JNEUROSCI.3028-10.2010
[42] Tsanas, A., and Clifford, G. D. (2015). Stage-independent, single lead EEG sleep spindle detection using the continuous wavelet transform and local weighted smoothing. Front. Hum. Neurosci. 9:181. doi: 10.3389/fnhum.2015.00181
[43] Wamsley, E. J., Tucker, M. A., Shinn, A. K., Ono, K. E., McKinley, S. K., Ely, A. V., et al. (2012). Reduced sleep spindles and spindle coherence in schizophrenia: mechanisms of impaired memory consolidation? Biol. Psychiatry 71, 154-161. doi: 10.1016/j.biopsych.2011.08.008
[44] Wang, Y.-H., Yeh, C.-H., Young, H.-W., Hu, K., and Lo, M.-T. (2014). On the computational complexity of the empirical mode decomposition algorithm. Phys. A, Stat. Mech. Appl. 400, 159-167. doi: 10.1016/j.physa.2014.01.020
[45] Warby, S. C., Wendt, S. L, Welinder, P., Munk, E. G., Carrillo, O., Sorensen, H. B., et al. (2014). Sleep-spindle detection: crowdsourcing and evaluating performance of experts, non-experts and automated methods. Nature methods11(4), 385-392. doi:10.1038/nmeth.2855
[46] Wendt, S. L., Christensen, J. A., Kempfner, J., Leonthin, H. L., Jennum, P., and Sorensen, B. D. (2012). “Validation of a novel automatic sleep spindle detector with high performance during sleep in middle aged subjects,” in Proceedings of the annual International Conference of the IEEE Engineering in Medicine and Biology Society (San Diego, CA), 4250-4253.
[47] Zitzler, E., Laumanns, M., and Thiele, L. (2001a) SPEA2: Improving the Strength Pareto Evolutionary Algorithm, TIK-Report 103.
[48] Zitzler, E. Laumanns, M., and Thiele, L. (2001b). “SPEA2: improving the strength Pareto evolutionary algorithm,” in Evolutionary Methods for Design, Optimisation and Control with Application to Industrial Problems (EUROGEN 2001), 95-100.