Abstract
Motor variability is inevitable in human body movements and has been addressed from various perspectives in motor neuroscience and biomechanics: it may originate from variability in neural activities, or it may reflect a large number of degrees of freedom inherent in our body movements. How to evaluate motor variability is thus a fundamental question. Previous methods have quantified (at least) two striking features of motor variability: smaller variability in the task-relevant dimension than in the task-irrelevant dimension and a low-dimensional structure often referred to as synergy or principal components. However, the previous methods cannot be used to quantify these features simultaneously and are applicable only under certain limited conditions (e.g., one method does not consider how the motion changes over time, and another does not consider how each motion is relevant to performance). Here, we propose a flexible and straightforward machine learning technique for quantifying task-relevant variability, task-irrelevant variability, and the relevance of each principal component to task performance while considering how the motion changes over time and its relevance to task performance in a data-driven manner. Our method reveals the following novel property: in motor adaptation, the modulation of these different aspects of motor variability differs depending on the perturbation schedule.
Highlights
In our daily lives, we repeatedly perform various desired movements, such as grasping a cup, throwing a ball, and playing the piano
We propose a flexible and straightforward machine learning framework for evaluating movement variability that combines the advantages of the various previous techniques: our framework can be used to evaluate task-relevant and task-irrelevant variabilities even when the average kinematics or task parameters are changing while considering how motion changes over time and the task function and can reveal how each synergy is relevant to task performance by means of an extension of principal component analysis (PCA)
The current study relied on linear regression to determine the relationship between motion data X ∈ RT×D and performance data d ∈ RT×1 based on the expression h = Xw, where T and D denote the number of trials and the number of variables in the motion data, respectively; h ∈ RT×1 is the predicted performance; and w ∈ RD×1 is the best set of linear coefficients for predicting the performance[30]
Summary
We repeatedly perform various desired movements, such as grasping a cup, throwing a ball, and playing the piano. Our motor systems somehow resolve these difficulties (i.e., variability and the large number of DoFs) to generate the desired movements It remains unclear how this taming of movement variability is achieved, one possible answer lies in the decomposition of motor variability into task-relevant and task-irrelevant variabilities. In the uncontrolled manifold (UCM) approach, task-relevant and task-irrelevant variabilities are evaluated mainly in terms of joint angles and angular velocities. This method focuses on the kinematic parameters relevant to task achievement, such as the hip joint position in stand-and-sit motions[11] or the hand position in arm-reaching movements[18]. In the TNC and GEM methods, the task-relevant and task-irrelevant variabilities are evaluated based on such task functions
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
