Real-time feedback in video-based motor learning: A pilot study exploring innovative training methods

Abstract

Video-based training has proven useful for motor learning, particularly when combined with motion feedback. However, the integration of real-time feedback into instructional videos has not been sufficiently explored. This study aimed to develop and explore innovative real-time feedback methods to enhance video-based motor learning. Twenty-seven participants (15 women, 12 men) were assigned to three feedback groups and one control group, who learned a choreography in an initial pilot study. The feedback groups received real-time comparisons of their own motions with those of an instructor. Group A was provided with a proportionally adjusted virtual instructor skeleton superimposed on their movements. Group B’s motions were transparently overlaid on the instructor's video. Group C viewed the instructor’s demonstration alongside a mirror view displayed of themselves. Group D (control) trained using only the instructor’s video, mimicking home-based tutorial formats. Motion tests performed without feedback revealed adaptation across all groups. Temporal motion adaptation was highest in Group A, while spatial motion adaptation was highest in Group B. Findings suggest that motion superimposition is a promising approach for visualizing motion discrepancies. Each method exhibited unique characteristics in the learning process, including different learning curves (e.g., Group A showing adaptation in the second half of the training) and varying levels of adaptation across different exercises and body parts (e.g., Group B experienced arm motion adaptation in squats). While these novel real-time feedback techniques demonstrate potential, further research is required to examine the relationships between feedback modalities and motor learning outcomes, specifically regarding the visualization of motion comparisons.