Replication Data for: The relation between complexity and resilient motor performance and the effects of differential learning

The first aim of this study was to test the possible relation between complexity and resilient motor performance (i.e., performance while being perturbed). The second aim was to test whether complexity and resilient performance improve through differential learning. To address our aims, we designed two parallel experiments involving a motor task, in which participants moved a stick with their non-dominant hand along a slider. Participants could score points by moving a cursor as fast and accurately as possible between two boxes, positioned on the right- and left side of the screen in front of them. In a first session, we determined the complexity by analyzing the temporal structure of variation in the box-to-box movement intervals with a Detrended Fluctuation Analysis. Then, we introduced perturbations to the task: We altered the tracking speed of the cursor relative to the stick-movements briefly (i.e., 4 seconds) at intervals of 1 minute (Experiment 1), or we induced a prolonged change of the tracking speed each minute (Experiment 2). Subsequently, participants had three sessions of either classical learning or differential learning. Participants in the classical learning condition were trained to perform the ideal movement pattern, whereas those in the differential learning condition had to perform additional and irrelevant movements. Finally, we conducted a posttest that was the same as the first session. In both experiments, results showed moderate positive correlations between complexity and points scored (i.e., box touches) in the perturbation-period of the first session. Across the two experiments, only differential learning led to a higher complexity index (i.e., more prominent patterns of pink noise) from baseline to post-test. Unexpectedly, the classical learning group improved more in their resilient performance than the differential learning group. Both the dataset and the scripts used for the two studies are provided here.

本研究的首要目标为检验运动复杂度与弹性运动表现（resilient motor performance，即受扰动时的运动表现）之间的潜在关联；次要目标为探究运动复杂度与弹性运动表现是否可通过差异学习（differential learning）得到提升。为达成上述研究目标，我们设计了两项平行实验，均包含一项运动任务：受试者使用非利手（non-dominant hand）持棒，沿滑块轨道移动。实验要求受试者以尽可能快且精准的速度，将光标在屏幕前方左右两侧的两个方框之间来回移动，以此获取积分。在首轮测试环节中，我们通过去趋势波动分析（Detrended Fluctuation Analysis）分析箱间移动间隔的变异时间结构，以此量化运动复杂度。随后我们向任务引入扰动：在实验1中，我们会每隔1分钟短暂（即持续4秒）调整光标相对于棒移动的追踪速度；在实验2中，我们则会在每分钟时对追踪速度进行持续性调整。随后受试者被分为经典学习（classical learning）组与差异学习组，两组均需完成三轮训练。其中经典学习组的受试者需训练至掌握理想运动模式，而差异学习组的受试者则需完成额外的无关运动任务。最终我们开展了与首轮测试环节完全一致的后测。两项实验的结果均显示，在首轮测试环节的扰动阶段，运动复杂度与所得积分（即触碰方框的次数）呈中等程度的正相关。综合两项实验来看，仅差异学习组可从基线水平至后测阶段获得更高的复杂度指数（即更显著的粉红噪声（pink noise）模式）。出乎意料的是，经典学习组的弹性运动表现提升幅度反而高于差异学习组。本研究的数据集与两项实验所用的代码脚本均已在此处提供。