Quantifying bi-variate coordination variability during longitudinal motor learning of a complex skill

Biofeedback (BFb) can enhance the motor learning process by guiding skill exploration. Too much BFb, however, can foster dependency leading to skill retention deficits once removed. A reducing BFb schedule could negate dependency effects, however limited methodologies exist to assess the effectiveness of an intervention during application. This research proposes a new bi-variate method (CI2Area) to quantify coordination variability (CoordVar) as a measure of skill exploration during a motor learning intervention. Thirty-two participants were introduced to a novel explosive-lunge task. A BFb group (n = 16) were provided with visual BFb on rear hip, knee and ankle joint extension magnitudes and timing during a 26-week reducing schedule BFb intervention. CoordVar of hip-knee and knee-ankle angular velocities were quantified by calculating the area encompassed by the 95% confidence intervals of joint coupling angular-velocity bi-variate plots (CI2Area). Linear regressions were fitted to group and individual CoordVar longitudinal data. The BFb was effective in successfully altering whole limb technique within just two sessions, and these changes were retained. The BFb group demonstrated a continual increase of CoordVar throughout the intervention, showing continual skill exploration strategies, while the Control group remained unchanged. Gradually increasing time between sessions, using a longitudinally reducing BFb schedule, successfully negates dependency effects on BFb while also encouraging motor learning. Manipulating time between sessions allows for the provision of a high frequency of 100% BFb without fostering dependency. The CI2Area method was able to detect individual exploration strategies and could be used in the future to direct individual intervention modifications. © 2019 Elsevier Ltd