近年研究指出，中風患者於接受跑步機行走訓練後可以改善行走能力；而接受額外背向行走訓練的中風患者在行走能力、平衡能力上更具有正面療效。中風患者則是因大腦功能受損造成行動不便，而人類行走的機制與大腦皮質的活動又具有密切相關。故期望能藉由腦電圖訊號來觀察正向與背向行走對大腦皮質活動的影響，為行走復健訓練提供更多的生理資訊。
由於腦電圖的訊號非常微弱，極易受到外界訊號的干擾。若要建造良好的量測環境則會因量測成本較高使得普及化困難。本文使用一般商用跑步機以及量測電路作為地面反作用力與腦電圖的量測工具，藉由適應性濾波器消除行走所產生的行走干擾波，並利用獨立成分分析消除如心電圖、眼動圖、肌電圖等干擾訊號。最後再利用統計分析的方法，如峰值、峰態、頻譜能量等特徵來移除受雜訊汙染的腦電訊號。
行走中的腦電訊號增加適應性濾波器濾波後，1.5~8.5Hz 的頻帶能量與行走干擾波的諧波能量均有顯著下降(p-value<0.05)。將濾波後的腦電圖依照σ、α、β等腦波頻帶計算各別的腦波能量以及對稱性，藉由假設分析檢定來判別正向與背向行走的腦波是否具有差異。α波能量於正向行走後，在F3、P3、P4、O1、O2具有顯著減少(p-value<0.05)；於背向行走後，在F4、C3、C4、P3、P4具有顯著減少(p-value<0.05)；於背向行走時，在C3、C4頻道的α2與σ能量顯著低於正向行走(p-value<0.05)。
透過訊號處理以及統計分析的方法可以發現，正向行走與背向行走在不同頻道的α波與β波能量變化具有明顯差異。未來可利用此方法進一步改良與發展復健所需的生理指標與腦機介面。

A recent study indicated that chronic stroke patient’s balance ability and gait performance could be improved by additional backward walking training. Due to the locomotive ability and cerebral cortex activity are closely related, this study measured the electroencephalograph (EEG) to assess the effect of backward and forward walking on cortical activities.
However, the signal strength of EEG is too weak to prevent signal contamination and it’s very expensive to build a friendly EEG measuring environment. This study attempts to modify a commercial treadmill to be the ground reaction force and EEG measuring device. We applied a band-limited adaptive filter to reduce the gait-related artifacts, performed independent component analysis to eliminate ECG, EOG and EMG artifacts. Next, we used statistical artifacts features to qualify the epoch’s quality, such as extreme value, kurtosis and spectral power outliers.
With adaptive filtering, it’s significantly reduced EEG spectral power in 1.5 to 8.5Hz frequency range and gait-related harmonics during walking (p-value<0.05). We compared the differences of EEG rhythms between forward and backward walking by hypothesis testing. After forward walking, Alpha activity was significantly decreased in F3, P3, P4, O1, O2; and after backward walking, Alpha activity significantly decrease in F4, C3, C4, P3, P4. During backward walking, the α2 and σ activity were significantly lower than forward walking in C3 and C4.
These results demonstrated the different EEG features between forward and backward walking and this measuring platform has the potential to develop more physiological features for rehabilitation.

論文審定書 i
致謝 ii
摘要 iii
Abstract iv
Contents v
List of Figures ix
List of Tables xi
Chapter 1 Introduction 1
1.1 Background 1
1.2 Motivation 1
1.3 Organization of This Thesis 3
Chapter 2 EEG and Artifacts 4
2.1 Brain and Electroencephalogram 4
2.1.1 EEG Rhythms 5
2.1.1 EEG Asymmetry 6
2.2 Signal Artifacts 7
2.2.1 Intrinstic Artifacts 7
2.2.2 Gait-related Artifacts 8
Chapter 3 Experiment Equipment and Methods 10
3.1 Force Treadmill 10
3.1.1 Treadmill Architecture 10
3.1.2 Treadmill Modeling 11
3.1.3 Ground Reaction Force Estimation 11
3.2 ECG Measuring Device and Method 12
3.3 EEG Measuring Device and Method 13
3.4 Experiment Flow 13
Chapter 4 EEG Artifacts Removal 15
4.1 Signal Preprocessing 16
4.1.1 Synchronization 17
4.2 Gait-Related Artifacts Removal 19
4.2.1 LMS Adaptive Filters 19
4.2.2 Band Limited LMS Adaptive Filter 20
4.3 Visual Inspection 21
4.4 Independent Component Analysis 22
4.4.1 Blind Source Separation 22
4.4.2 Entropy and Mutual Information 23
4.4.3 Information Maximization ICA (Infomax ICA) 25
4.4.4 Removing ICA Components 26
4.5 Statistical Artifacts Removal 28
4.5.1 Epochs 28
4.5.2 Time-Domain Statistical Analysis 30
4.5.3 Frequency-Domain Statistical Analysis 32
Chapter 5 EEG Spectrum Analysis 33
5.1 EEG Power Spectrum Density Estimation 33
5.2 EEG Rhythms Analysis 34
5.2.1 EEG Rhythms 34
5.2.2 EEG Rhythm Increment 35
5.2.2 EEG Rhythm Asymmetry 35
5.3 Statistical Hypothesis Test 35
5.2.1 Hypothesis Tests for Dependent Samples 37
Chapter 6 Experimental Results 38
6.1 Performance of Gait-related Artifact Removal 38
6.1.1 Spectral Power Ratio 38
6.1.2 Mode Power Ratio 38
6.1.3 Results 40
6.2 EEG Rhythm Analysis 42
6.2.1 Forward Walking 42
6.2.2 Backward Walking 43
6.2.3 Forward Walking versus Backward Walking 44
6.3 EEG Asymmetry Analysis 47
6.3.1 Forward Walking 47
6.3.2 Backward Walking 48
6.3.3 Forward Walking versus Backward Walking 48
Chapter 7 Discussion and Future Works 50
7.1 Performance of Gait-related Artifacts Removal 50
7.2 Different EEG Patterns in Forward and Backward Walking 51
Reference 54
Appendix 60
Appendix I Gait-related Artifacts 60
Appendix II EEG Rhythms (Relative Power) 66
Appendix III EEG Asymmetry (Relative Power) 74
Appendix IV EEG Amplitude Spectrum (Absolute Power) 76

1. 張耿維，2014，跑步機背向行走訓練對慢性中風患者平衡能力、走路表現與心肺適能之療效，碩士論文，高雄醫學大學運動醫學研究所。
2. Grasso R, Bianchi L, Lacquaniti F. Motor patterns for human gait: backward versus forward locomotion. J Neurophysiol. 1998;80(4):1868-85.
3. Winter DA, Pluck N, Yang JF. Backward walking: a simple reversal of forward walking? J Mot Behav. 1989;21(3):291-305.
4. Hao WY, Chen Y. Backward walking training improves balance in school-aged boys. Sports Med Arthrosc Rehabil Ther Technol. 2011;3:24.
5. Cipriani DJ, Armstrong CW, Gaul S. Backward walking at three levels of treadmill inclination: an electromyographic and kinematic analysis. J Orthop Sports Phys Ther. 1995;22(3):95-102.
6. Flynn TW, Soutas-Little RW. Mechanical power and muscle action during forward and backward running. J Orthop Sports Phys Ther. 1993;17(2):108-12.
7. Flynn TW, Soutas-Little RW. Patellofemoral joint compressive forces in forward and backward running. J Orthop Sports Phys Ther. 1995;21(5):277-82.
8. Drew T, Marigold DS. Taking the next step: cortical contributions to the control of locomotion. Curr Opin Neurobiol. 2015;33C:25-33.
9. Luft AR, Smith GV, Forrester L, Whitall J, Macko RF, Hauser TK, Goldberg AP, Hanley DF. Comparing brain activation associated with isolated upper and lower limb movement across corresponding joints. Hum Brain Mapp. 2002;17(2):131-40.
10. Dobkin BH, Firestine A, West M, Saremi K, Woods R. Ankle dorsiflexion as an fMRI paradigm to assay motor control for walking during rehabilitation. Neuroimage. 2004;23(1):370-81.
11. Suzuki M, Miyai I, Ono T, Kubota K. Activities in the frontal cortex and gait performance are modulated by preparation. An fNIRS study. Neuroimage. 2008;39(2):600-7.
12. Harada T, Miyai I, Suzuki M, Kubota K. Gait capacity affects cortical activation patterns related to speed control in the elderly. Exp Brain Res. 2009;193(3):445-54.
13. Severens M, Nienhuis B, Desain P, Duysens J. Feasibility of measuring Event Related Desynchronization with EEG during walking. Conf Proc IEEE Eng Med Biol Soc. 2012;2012:2764-7.
14. Perusqia Hernandez M. A Brain Computer Interface for walking using EEG. Master thesis. Eindhoven University of Technology.2012.
15. Perusqia Hernandez M. Sensorimotor cortical activity during human gait: Forward and backward walking. Research project. Eindhoven University of Technology. 2012.
16. Pfurtscheller G, Lopes da Silva FH. Event-related EEG/MEG synchronization and desynchronization: basic principles. Clin Neurophysiol. 1999;110(11): 1842-57.
17. Brümmer V, Schneider S, Abel T, Vogt T, Strüder HK. Brain cortical activity is influenced by exercise mode and intensity. Med Sci Sports Exerc. 2011;43(10):1863-72.
18. Harmon-Jones E, Gable PA, Peterson CK. The role of asymmetric frontal cortical activity in emotion-related phenomena. A review and update. Biol Psychol. 2010;84(3):451-62.
19. Harmon-Jones E, Harmon-Jones C, Price TF. What is approach motivation? Emotion Review. 2013; 5(3):291-295.
20. Hall EE, Ekkekakis P, Petruzzello SJ. Predicting affective responses to exercise using resting EEG frontal asymmetry: Does intensity matter? Biol Psychol. 2010;83(3):201-6.
21. Petruzzello SJ, Hall EE, Ekkekakis P. Regional brain activation as a biological marker of affective responsivity to acute exercise: Influence of fitness. Psychophysiology. 2001;38(1):99-106.
22. Hall EE, Ekkekakis P, Petruzzello SJ. Regional brain activity and strenuous exercise. Predicting affective responses using EEG asymmetry. Biol Psychol. 2007;75(2):194-200.
23. Vogt T, Schneider S, Brümmer V, Strüder HK. Frontal EEG asymmetry-The effects of sustained walking in the elderly. Neurosci Lett. 2010;485(2):134-7.
24. Goncharova II, McFarland DJ, Vaughan TM, Wolpaw JR. EMG contamination of EEG: spectral and topographical characteristics. Clin Neurophysiol. 2003;114(9):1580-93
25. Jung TP, Makeig S, Humphries C, Lee TW, McKeown MJ, Iragui V, Sejnowski TJ. Removing electroencephalographic artifacts by blind source separation. Psychophysiology. 2000;37(2):163-78.
26. SE. Kerick, KS. Oie, K. McDowell. Assessment of EEG signal quality in motion environments. Army Res. Lab., Rep. ARL-TR-4866. 2009.
27. Gwin JT, Gramann K, Makeig S, Ferris DP. Removal of movement artifact from high-density EEG recorded during walking and running. J Neurophysiol. 2010;103(6):3526-34.
28. Gwin JT, Gramann K, Makeig S, Ferris DP. Electrocortical activity is coupled to gait cycle phase during treadmill walking. Neuroimage. 2011;54(2):1289-96.
29. T Castermans, M Duvinage, M Petieau, T Hoellinger, C D Saedeleer, K Seetharaman, A Bengoetxea, G Cheron, T Dutoit. Optimizing the performances of a P300-based brain–computer interface in ambulatory conditions. IEEE J. Emerg. Sel. Topics Circuits Syst. 2011;1(4):566-577.
30. M Duvinage, T Castermans, M Petieau, G Cheron, T Dutoit. Are current gait-related artifact removal techniques useful for low-complexity BCIs?. IJCNN. 2012; 1-7.
31. 黃瀜緯、李懿軒、蔡瀛慶、洪志遠、嚴成文，整合於跑步機地面反作用力的量測裝置，中國機械工程學會第三十一屆全國學術研討會。Taichung, Taiwan, Dec. 7-8, 2012.
32. VK. Ingle, JG. Proakis. Digital signal processing using MATLAB. Cengage Learning, 2012. pp.493-498.
33. de Chazal P, Heneghan C, Sheridan E, Reilly R, Nolan P, O'Malley M. Automated Processing of the single-lead electrocardiogram for the detection of obstructive sleep apnoea. IEEE Trans Biomed Eng. 2003;50(6):686-96.
34. JC. Principe, NR. Euliano, WC. Lefebvre. Neural and adaptive systems, Fundamentals through simulations. John Wiley & Sons, 2000.
35. Delorme A, Makeig S. EEGLAB Wikitorial. Available: http://sccn.ucsd.edu/eeglab/, 2011. (Adapted from: Delorme A, Makeig S. EEGLAB: an open source toolbox for analysis of single-trial EEG dynamics. J Neurosci Methods. 2004;134(1):9-21).
36. Bell AJ, Sejnowski TJ. An information-maximization approach to blind separation and blind deconvolution. Neural Comput. 1995;7(6):1129-59.
37. Lee TW, Girolami M, Sejnowski TJ. ICA using an extended infomax algorithm for mixed subgaussian and supergaussian source. Neural Comput. 1999;11(2):417-41.
38. Delorme A, Sejnowski T, Makeig S. Enhanced detection of artifacts in EEG data using higher-order statistics and independent component analysis. Neuroimage. 2007;34(4):1443-9.
39. Buzsáki G, Mizuseki K. The log-dynamic brain: how skewed distributions affect network operations. Nat Rev Neurosci. 2014;15(4):264-78.
40. Pivik RT, Broughton RJ, Coppola R, Davidson RJ, Fox N, Nuwer MR. Guidelines for the recording and quantitative analysis of electroencephalographic activity in research contexts. Psychophysiology. 1993;30(6):547-58.
41. 張皓，2013，心電圖訊號波型的頻譜分析，碩士論文，國立中山大學機械與機電工程學系。
42. Sterman MB. Physiological origins and functional correlates of EEG rhythmic activities: implications for self-regulation. Biofeedback Self Regul. 1996; 21(1):3-33.
43. VB. Pavlenko, SV. Chernyi, DG. Goubkina. EEG correlates of anxiety and emotional stability in adult healthy subjects. Neurophysiology. 2009;41(5):337-345.
44. Crabbe JB, Dishman RK. Brain electrocortical activity during and after exercise: A quantitative synthesis. Psychophysiology. 2004;41(4):563-74.
45. Tanaka H, Monahan KD, Seals DR. Age-predicted maximal heart rate revisited, J Am Coll Cardiol. 2001 37(1):153-6.
46. Grimshaw GM, Carmel D. An asymmetric inhibition model of hemispheric differences in emotional processing. Front Psychol. 2014;5:489.
47. Hall EE, Ekkekakis P, Van Landuyt LM, Petruzzello SJ. Resting frontal asymmetry predicts self-selected walking speed but not affective responses to a short walk.Res Q Exerc Sport. 2000;71(1):74-9.
48. Pfurtscheller G,Neuper C. Movement and ERD/ERS. The Bereitschaftspotential. 2003; pp.191-206.
49. Juri D. Kropotov. Quantitative EEG, Event-Related Potentials and Neurotherapy. Elsevier, 2009. pp.29-76.
50. Baumeister J, Reinecke K, Liesen H, Weiss M. Cortical activity of skilled performance in a complex sports related motor task. Eur J Appl Physiol. 2008;104(4):625-310.