Wang Yun, Zatsiorsky Vladimir M, Latash Mark L
Department of Kinesiology, The Pennsylvania State University, Rec Hall 268N, University Park, 16802, USA.
Exp Brain Res. 2005 Nov;167(2):196-210. doi: 10.1007/s00221-005-0012-3. Epub 2005 Nov 15.
We used the framework of the uncontrolled manifold (UCM) hypothesis to analyze multi-muscle synergies involved in making a step by a standing person. We hypothesized that leg and trunk muscles are organized into stable groups (muscle modes, M-modes) related to shifts of the center of pressure (COP) in the anterior-posterior and medio-lateral directions. Another hypothesis was that the magnitudes of the modes co-vary across repetitive trials to stabilize a certain magnitude of the COP shift in both directions. M-modes were defined using principal component analysis applied to indices of changes in the electromyographic (EMG) activity prior to releasing variable loads that were held by the subject using a pulley system. For the task of releasing the load behind the body three M-modes associated with a backward COP shift were defined. Four M-modes were defined for the task of releasing the load at the body side associated with a lateral COP shift. Multiple regression analysis was used to relate changes in the M-mode magnitudes to COP shifts. EMG changes prior to making a step were quantified over five 100 ms time windows before the lift-off of the stepping leg. Two components of the variance in the M-mode space computed across repetitions of a stepping task were quantified-a component that did not affect the average COP shift in a particular direction (variance within the UCM, V (UCM)), and a component that affected the COP shift (variance orthogonal to the UCM, V (ORT)). V (UCM) was significantly higher than V (ORT) for both directions of the COP shifts. This relation was observed for the M-modes in the stepping leg as well as in the support leg. The stepping leg showed a different time evolution of the ratio V (UCM)/V (ORT) such that the difference between the two variance components disappeared closer to the time of the lift-off. The findings corroborate both main hypotheses. The study supports a view that control of whole-body actions involves grouping the muscles, using fewer elemental variables to scale the muscle activity, and forming synergies in the space of the elemental variables that stabilize time profiles of important performance variables.
我们使用非控制流形(UCM)假说的框架来分析站立者迈出一步时涉及的多肌肉协同作用。我们假设腿部和躯干肌肉被组织成与压力中心(COP)在前后和内外方向上的移动相关的稳定组(肌肉模式,M模式)。另一个假设是,这些模式的大小在重复试验中共同变化,以稳定两个方向上COP移动的特定大小。M模式是通过对肌电图(EMG)活动变化指数应用主成分分析来定义的,这些指数是在受试者使用滑轮系统释放可变负荷之前获取的。对于在身体后方释放负荷的任务,定义了与向后COP移动相关的三种M模式。对于在身体一侧释放负荷的任务,定义了与侧向COP移动相关的四种M模式。使用多元回归分析将M模式大小的变化与COP移动联系起来。在迈步腿离地前的五个100毫秒时间窗口内,对迈步前的EMG变化进行了量化。对跨步任务重复过程中计算出的M模式空间中的两个方差分量进行了量化——一个不影响特定方向上平均COP移动的分量(UCM内的方差,V(UCM)),以及一个影响COP移动的分量(与UCM正交的方差,V(ORT))。对于COP移动的两个方向,V(UCM)均显著高于V(ORT)。在迈步腿和支撑腿的M模式中都观察到了这种关系。迈步腿显示出V(UCM)/V(ORT)比值的不同时间演变,使得两个方差分量之间的差异在接近离地时间时消失。这些发现证实了两个主要假设。该研究支持这样一种观点,即全身动作的控制涉及对肌肉进行分组,使用较少的基本变量来缩放肌肉活动,并在稳定重要性能变量时间分布的基本变量空间中形成协同作用。