臺灣博碩士論文加值系統

在過去的研究中，單擺試驗法在臨床上已經普遍應用於量化中風患者下肢膝關節痙攣程度，但並未應用至量化上肢肘關節痙攣程度；本研究的目的即為發展一簡單的輔助儀器以幫助受測者進行上肢肘關節單擺試驗，從生理現象建立的生物力學模型，透過拘束最佳化方法估測出肘關節神經肌肉系統模型的參數，再藉由本研究室先前所發展的痙攣量測系統來做參數驗證的工作，以期能以簡單的單擺試驗幫助神經科臨床醫師量化中風患者上肢肘關節痙攣程度。
本研究建立了三種生物力學模型：(1)簡化模型：去除了複雜的肌肉牽張反射機制，只包含肘關節軟組織的扭剛度與扭阻尼特性；(2)純速度反射模型：由簡化模型加入與速度有關的牽張反射機制；(3)加成模型：由純速度反射模型再加入與位置有關的牽張反射機制，這是生理上認為完整的牽張反射機制；參數估測的方法以使用拘束最佳化方法為主，而加成模型的參數眾多，再以遺傳基因演算法來比較加成模型的參數估測結果。
由上述模型參數估測結果發現：從簡化模型所得之系統扭阻尼比，可以成功地區分出常人與患者的患側與健側之間的差異；純速度反射模型在系統加入與速度有關的牽張反射機制後，使得患者患側肌肉軟組織扭阻尼比簡化模型下降許多，顯示與速度有關的牽張反射機制其對系統的影響力；在加成模型中，企圖區分肌肉痙攣與其牽張速度及伸長量變化的關係時，在最佳化過程中出現了多重解，而目前尚無法則可以分辨
何組解是最佳解，且複雜模型的各組解其模擬結果也不比簡單模型更好。
綜合本研究三種模型的參數估測結果發現，統計上患者患健側的軟組織扭剛度都會比常人來得大；患者患側軟組織的扭阻尼會比健側略大，但是並無統計上的差異；而在純速度反射模型中，患者患側與速度有關的牽張反射參數B會比健側大，在加成模型就無法區別了，不過與位置有關的牽張反射機制在加成模型中的作用也一直不大，還是由扭阻尼與B值在影響整個系統。
總之太過複雜的模型對參數估測並無多大的幫助，反而受到多重解的困擾；患者患側與健側的軟組織扭剛度皆會比常人大，符合臨床上患者健側可能也有類似僵直現象的情況；而患者患側與健側的軟組織扭阻尼的改變，並無統計上的差異，與伸長量改變的部分其作用也並不明顯，所以我們推論肌肉痙攣與牽張速度有關的部分對實驗的影響會比與伸長量有關的部分來的大，符合一般生理學者所提出的理論。

Pendulum test is a common clinical technique for assessing spasticity on knee joints of stroke patients, but it has not been applied on elbow joints. The goal of this study is to design a simple instrument for performing pendulum tests on upper limbs. Based on biomechanical models of elbow, one can obtain the physiological parameters of the models by using a constrained optimization method (sequential quadratic programming) and the models are validated by using a sophisticated spasticity measurement system developed in previous researches. We hope that the simple pendulum test can assist neurologists to assess spasticity in stroke patients.
Three biomechanical models are utilized in this study: (1) simple model: with removal of muscle stretch reflex mechanisms, only rotational stiffness and rotational damping coefficient of soft tissue of elbow joint are considered; (2) pure velocity-stretch-reflex model: by adding velocity-dependent stretch reflex component to the simple model; (3) additive model: by adding position-dependent stretch reflex component to the pure velocity- stretch-reflex model. The last one is a complete model from physiological point of view. Parameters of the models are estimated by using the sequential quadratic programming method. To explore global minimum solutions of the parameter estimation problem, genetic algorithm is employed and results are compared with those of sequential quadratic programming. 10 normal subjects and 10 stroke patients are recruited in this study.
From results of parameter estimation, one may find that, in the simple model, intact side, affected side of stroke patients and normal group can be differentiated by using rotational damping ratios of the model. In the pure velocity-stretch-reflex model, the rotational damping coefficients of patients'' affected side are lower than values obtained in the simple model. It reveals that the contribution of velocity-dependent stretch reflex component is significant, however when attempting to separate the position-dependent and velocity-dependent components of spasticity, one gets multiple solutions from the optimization process. The problem of choosing an optimal solution remains open and parameters obtained by using the simple model might be valid for assessing spasticity.
Statistical analyses show that, the rotational stiffness of both sides of patients is higher than that of normal group. For stroke patients, the rotational damping coefficient of affected side is slightly higher than that of intact side. In the pure velocity-stretch-reflex model, the velocity-dependent parameter of affected side is higher than that of intact side, but it shows no significant difference in the additive model. The contribution of position-dependent stretch reflex component is not significant and the dominant parameters are rotational damping coefficient and velocity-dependent parameter.
In summary, based on our limited subjects and data, we can find that the complete biomechanical model might not be useful for assessments of spasticity due to the problem of multiple solutions. The rotational stiffness of both sides of patients is higher than that of normal group. This finding agrees with increased toughness of intact side found in clinical practice. The change of rotational damping coefficients of both elbows of stroke patients is not significant and the contribution of position-dependent stretch reflex to spasticity is not significant. One may conclude that spasticity is more dependent on stretch velocity, which agrees with physiological hypothesis on spasticity.

第一章 緒論……………………………………………..…………………..1
1-1牽張反射………………………………………………………………..1
1-2肌肉痙攣………………………………………………………………..2
1-3肌肉痙攣臨床半定量評估法……………………….………………….3
1-4文獻回顧………………………………………………………………..4
1-5研究動機與目的…………………………….…………………………7
第二章 方法………………………………………………………………..9
2-1 上肢單擺試驗……………………………………………………….....9
2-1-1 上肢單擺試驗裝置…………………………….………………9
2-1-2 上肢單擺試驗步驟……………………………………………12
2-1-3質心及慣性矩量測……………………………………………13
2-1-4 試驗數據的研判………………………………………………15
2-2 等速牽張試驗………………………………………………………...16
2-2-1 等速牽張試驗裝置……………………………………………16
2-2-2 等速牽張試驗程序……………………………………………17
2-3 模型的建立…………………………………………………………...18
2-4 分析方法……………………………………………………………...22
2-4-1 簡單的分析…….………………………………………………22
2-4-2 拘束最佳化方法.………………………………………………23
2-4-3 遺傳基因演算法………………………………………………26
2-5 等速牽張試驗模擬…………………………………..………………30
2-6 受測者之選擇………………………………………………………...30
第三章 結果………………………………………………………………32
3-1 輔助儀器之校正………………………...……………………..………32
3-2 電子量角器之校正………..…………...………………………………34
3-3 實驗結果……………………….……………………………………...35
3-4 模型參數估測結果………………..……...……………………………39
3-4-1 簡化模型之參數估測………….………………………………39
3-4-2 純速度反射模型之參數估測…….……………………………44
3-4-3 加成模型之參數估測……………….…………………………47
3-5 等速牽張模擬結果………………………...……..……………………52
3-6 遺傳基因演算法結果……………………...………..…………………58
第四章 討論………………………………………………………………61
4-1模型由簡至繁之比較…………………………………………………61
4-2拘束最佳化方法與遺傳基因演算法之異同…………………………62
4-3單擺法與等速牽張法之比較………………………………….………62
4-4 生理上的探討……………………………………………………..…...63
第五章 結論與建議………………………………………………………66
5-1 結論…………………………………………………………………...66
5-2 建議…………………………………………………………………...67
參考文獻…………………………………………………………………..68

[1] C. H. Wu, J. C. Houk, K-Y Young & L. E. Miller, "Nonlinear Damping of Limb Motion," Chapter 13, in Multiple Muscle Systems, edited by J. M. Winters and S.L.-Y. Woo, 1990.
[2] D. A. Winter, "Biomechanics of Human Movement," New York : Wiley, p. 151, 1979.
[3] D. C. Lin & W. Z. Rymer, "A Quantitative Analysis of Pendular Motion of the Lower Leg in Spastic Human Subjects," IEEE Trans. on Biomed. Eng., Vol. 38, pp. 906-918, 1991.
[4] D. E. Goldberg, "Genetic Algorithms in Search, Optimization, and Machine Learning,"Addison-Wesley Publishing Company, Inc, 1989.
[5] D. R. Lamb, "The Physiological Basis of Musclar Strength," Chapter 12, in Physiology of Exercise Responses & Adaptations, Second Edition, 1984.
[6] J. C. Houk, W. Z. Rymer & P. E. Crago, "Dependence of Dynamic Response of Spindle Receptors on Muscle Length and Velocity," J. of Neurophysiology, Vol. 46, No. 1, pp. 143-166, July 1981.
[7] J. H. Yang, C. I. Chou & M. S. Ju, "Biomechanical and Biocybernetic Study of Hypertonia in Stroke," The ROC Biomedical Engineering Society 1997 Annual Symposium, pp. 441-442, Chung-Li, Dec. 1997.
[8] J. L. Chen & G. K. Lee, "The Application of Genetic Algorithms in Design Optimization," 12th. Nat. Conf. on Mech. Eng., CSME, Chia-Yi, Nov. 1995.
[9] J. L. Gau, J.-J. J. Chen, M. S. Ju & T. C. Hsueh, "Quantitative Analysis of Stretch Reflexes in Spastic Subjects," Annual Symposium of the Biomedical Engineering Society, R.O.C., pp. 24-25, 1994.
[10] J. L. Li, M. S. Ju, C. C. Lin & K. N. An, "Modeling and Identification of Pendular Motion of the Lower Leg in Normal and Stroke Subjects," Chinese J. of Medical and Biological Engineering, Vol.19, No.1, pp. 1-9, 1999.
[11] J. M. Hidler & W. Zev Rymer, "A Simulation Study of Reflex Instability in Spasticity: Origins of Clonus," IEEE Trans. on Rehabil. Eng., Vol. 7, No. 3, pp. 327-340, September 1999.
[12] K. E. Hagbarth, J. V. H gglund, M. Nordin & E. U. Wallin, "Thixotropic Behaviour of Human Finger Flexor Muscles with Accompanying Changes in Spindle and Reflex Resopnses to Stretch," J. of Physiology 368, pp. 323-342, 1985.
[13] L. Q. Zhang, H. Y. Huang, J. A. Sliwa & W. Zev Rymer, "System Identification of Tendon Reflex Dynamics," IEEE Trans. on Rehabil. Eng., Vol. 7, No. 2, pp. 193-203, June 1999.
[14] M. S. Ju, J. J. J. Chen, H. M. Lee, T. S. L, C. C. Lin & Y. Z. Huang, "Time-Course Analysis of Stretch Reflexes in Hemiparetic Subjects Using an On-Line Spasticity Measurement Systems," J. of Electromyography and Kinesiology, Vol. 10, No. 1. , pp.1-14, 2000.
[15] R. A. Brown, D. A. Lawson, G. C. Leslie, A. Macarthur, W. J. Maclennan, M. E. Mcmurdo, W. J. Mutch & N. J. Part, "Does the Wartenberg pendulum test differentiate quantitatively between spasticity and rigidity? A study in elderly stroke and Parkinsonian Patients," J. of Neurology, Neurosurgery and Psychiatry, Vol. 51, pp. 1178-1186, 1988.
[16] R. K. Powers, J. Marder-Meyer & W. Z. Rymer, "Quantitative Relations Between Hypertonia and Stretch Reflex Threshold in Spastic Hemiparesis," Ann. Neurol., Vol. 23, pp. 115-124, 1988.
[17] R. T. Katz & W. Z. Rymer, "Spastic Hypertonia: Mechanisms and Measurement," Arch. Phys. Med. Rehabil., Vol. 70, pp. 144-155, 1989.
[18] R. Wartenberg, "Pendulousness of the legs as a dignostic test," Neurology, Vol. 1, pp. 18-24, 1951.
[19] Y. S. Hsu, M. S. Ju, T. S. Lin & J.-J. J. Chen, "Quantitative Measurement of Spastic Hypertonic Muscles by Frequency Response Method," Symposium of 1995 R.O.C. Biomechanics Conference, pp. 147-148.
[20] 陳福成, "質心及慣性矩量測," 機械月刊, 第二十四卷第十二期, pp. 321-326, 1998.