Latash Mark L, Shim Jae Kun, Zatsiorsky Vladimir M
Department of Kinesiology, The Pennsylvania State University, University Park, PA 16802, USA.
Exp Brain Res. 2004 Nov;159(1):65-71. doi: 10.1007/s00221-004-1933-y. Epub 2004 Jul 28.
We studied whether characteristics of individual finger force profiles covaried across repetitions of a quick force pulse production task to stabilize the required magnitude and timing of the peak force. Subjects produced series of quick force pulses by pressing with all four fingers of the right hand on force sensors under the instruction to keep the magnitude of the peak of total force at 15 N and reach the force peaks at prescribed times. Individual finger force pulses were then reshuffled across trials to create a surrogate data set. The surrogate data set showed a lower average peak force with a larger dispersion. This finding has been interpreted as pointing at predominantly negative covariation among finger force pulses in the actual data that stabilized the required magnitude of the total force, a force synergy. The difference between the actual and surrogate data sets was significant early into the pulse time, starting about 40 ms after the pulse initiation. This finding points at a central nature of the negative covariation without a major role played by visual or proprioceptive feedback. In contrast, the surrogate data set showed smaller dispersion of the timing of the total peak force, suggesting positive covariation of the timings of individual finger force pulses in the actual data interpreted as the lack of a timing synergy. These results have been confirmed with principal component (PC) analysis. The first PC for the timing of the individual finger peak forces accounted for over 90% of the total variance for the actual data set and for under 40% of the total variance for the surrogate data set. The fourth PC for the magnitudes of the finger forces accounted for under 4% of the total variance for the actual data set and for over 15% of the variance for the surrogate data set. The data are interpreted within the uncontrolled manifold hypothesis; they support the hierarchical control scheme suggested by Schöner.
我们研究了在快速力脉冲产生任务的重复过程中，单个手指力轮廓的特征是否会协同变化，以稳定所需的峰值力大小和时间。受试者右手的所有四根手指按压力传感器，产生一系列快速力脉冲，指令是将总力峰值大小保持在15牛，并在规定时间达到力峰值。然后在各试验中对单个手指力脉冲进行重新排列，以创建一个替代数据集。替代数据集显示平均峰值力较低且离散度较大。这一发现被解释为表明实际数据中手指力脉冲之间主要存在负协同变化，从而稳定了所需的总力大小，即一种力协同作用。实际数据集和替代数据集之间的差异在脉冲开始后约40毫秒开始的脉冲时间早期就很显著。这一发现表明负协同变化具有核心性质，视觉或本体感觉反馈并未起主要作用。相比之下，替代数据集显示总峰值力时间的离散度较小，这表明实际数据中单个手指力脉冲时间存在正协同变化，被解释为缺乏时间协同作用。这些结果已通过主成分（PC）分析得到证实。单个手指峰值力时间的第一主成分占实际数据集总方差的90%以上，而占替代数据集总方差的不到40%。手指力大小的第四主成分占实际数据集总方差的不到4%，而占替代数据集方差的15%以上。这些数据是在非受控流形假设内进行解释的；它们支持了舍纳提出的分层控制方案。
