Employing a single trial motor equivalent analysis for the assessment of motor learning.

Abstract

The uncontrolled manifold analysis (UCM) is a useful technique for motor learning research enabling the classification of movement variability into solutions and errors. Less explored methodological considerations within the UCM framework are the selection of mean configurations outside of the current performance, as found in the Motor Equivalence Analysis, and a single trial approach. In this study, we demonstrated how calculating deviations away from varying mean configurations within the UCM influences the results and interpretations within motor learning experiments. Twelve young adult subjects (9F/3 M, 20.53 ± 1.25 years old) practiced the kettlebell swing over a one-week time period. We compared deviations from the mean configuration across all repetitions, to the mean of the first ten repetitions before practice and to the mean of their last ten repetitions after practice. Results suggested that subjects abandoned their initial mean performance within the first sets of kettlebell swings and reduced their errors and solutions towards what would become their mean performance after practice. They continued to refine their performance 1 week later. Subjects then completed a transfer task, testing their ability to adapt to a water-filled kettlebell. We evaluated deviations from their mean performance with the metal kettlebell and their mean performance with the water-filled kettlebell. Subjects did not reduce errors towards their mean metal kettlebell performance, but instead towards a new performance that matched the dynamics of the water-filled kettlebell. When performance is expected to change, i.e., motor learning, assessing how the variance structure changes with respect to different mean configurations can provide further insight when using a UCM approach.