Nonlinear analysis of physiologic and pathologic tremors

• Main Authors: Yue-Loong Hsin, 辛裕隆
• Other Authors: Terry B. J. Kuo
• Format: Others
• Language: zh-TW
• Online Access: http://ndltd.ncl.edu.tw/handle/98295430618556533660

碩士 === 慈濟大學 === 神經科學研究所 === 89 === Introduction Tremor manifests with visible oscillation of body part, the rhythmic character is suited to different algorithms to analyze it. Power spectral technique is the most common method. Variance of tremor frequency helps to classify the tremor types and distinguish central and peripheral mechanism. But tremor does not always keep in stationary condition. The problem uncovers the limitation of spectral analysis. So we tried to apply the developed estimators of dynamic to quantify the complexity of tremor. Methods The present study was designed to assess the applicability of non-linear estimators in tremor analysis. We collected the electromyography (EMG) signals from volunteers for enhanced physiologic tremor (n=6), essential tremor (n=6) and Parkinson’s patients with rest tremor (n=6). Surface EMG was applied to extensor muscle of tremor under outstretching, hand extending (added stiffness of limb), and 500 g weighted condition (added muscular inertia). We calculated the values of non-linear estimators with approximate entropy (ApEn), correlation dimensions (D2), and largest Lyapunov exponent (LLE). Results In outstretching condition, enhanced PT had higher value of ApEn (mean ? SEM, 0.61 ? 0.04) than pathologic tremors (P<0.05). ET had higher value of ApEn than rest tremor of PD (P<0.05, 0.54 vs. 0.43). In contrast to decreased ApEn by added stiffness and inertia in enhanced PT, added stiffness increased ApEn and added inertia decreased ApEn in ET. Certainly, action paradigm had higher complexity than resting state in PD. The other non-linear estimators did not provide strong evidence to measure the complexities, except in action condition of PD. Discussions Added inertia and stiffness alternate mechanical oscillations of tremor, those ways are helpful to distinguish tremor from the participated rule of central oscillations. Theoretically tremor is influenced by central and peripheral mechanism, and manifested frequency is determined by the dominant mechanism. Undoubtedly, PT has major component of musculoskeletal reflex system. Aggravation of PT is considered to be the explanation of essential tremor, which is largely modulated by central oscillator than PT. In the patient with Parkinson’s disease whose peak power frequency is resistant to external loading, this property reflects the motor units are strongly modulated by central component. Increase of ApEn, D2 and LLE in action condition implies the participation of peripheral reflex mechanism. Conclusions The principal focus of this article described regularity statistics, especial the approximate entropy. The different destruction of interaction between central and peripheral oscillator results in different pathologic tremor and also decreases complexity. ApEn, D2 and LLE can reflect the alternation. Furthermore, change of ApEn corresponds to dynamic processing in individual. It is worth to explore the applicability of non-linear estimation in dynamic neurophysiology, particular the tremor.