Fig 1.
Analysis of electromyographic (EMG) activity.
Abductor pollicis brevis (APB) muscle of a typical participant. The raw EMG trace for the APB shows phasic activity. The raw data were rectified and smoothed using a 5-Hz Butterworth low-pass filter. The phases during which muscle activity was ≥5% maximum voluntary contraction (MVC) were defined as on-phases. For each trial, the duration of each on-phase (within the limits of the two-way arrow) and the areas of the on-phases between the trace and the line representing the 5% MVC level (shaded area) were calculated.
Fig 2.
Typical rectified and smoothed electromyographic activity of the abductor pollicis brevis muscle.
This figure shows that temporal and quantitative variabilities in muscle electrical activity are clearly reduced from the first (trial 1) to the final trial (trial 30).
Fig 3.
Results of behavioral changes.
(A) Changes in the ball rotation time from trial 1 to 30. (B) Comparison of the mean (± SE) ball rotation time between the three stages (initial, trials 1–5; middle, trials 13–17; and final, trials 26–30) (C) Changes in number of errors from trial 1 to 30. (D) Comparison of the mean (± SE) number of errors between the three stages. (** p < 0.01, * p < 0.05).
Fig 4.
Changes in the number of electromyography (EMG) on-phases and total EMG area.
(A-C) Comparison of the mean (± SE) number of EMG on-phases between the three stages (initial, trials 1–5; middle, trials 13–17; and final, trials 26–30) in the abductor pollicis brevis (APB) (A), first dorsal interosseous (FDI) (B), and extensor digitorum (ED) (C) muscles. (D–F) Differences in the mean (± SE) EMG total area of the APB, FDI and ED muscles between the three stages are shown in panels (D), (E), and (F), respectively. n = 37, 37, and 36 for APB, FDI, and ED, respectively (**; p < 0.01, *; p < 0.05).
Fig 5.
Changes in electromyography (EMG) variability.
(A–C) Comparison of the mean (± SE) coefficient of variation (CV) of EMG on-phase duration between the three stages (initial, trials 1–5; middle, trials 13–17; and final, trials 26–30) in the abductor pollicis brevis (APB) (A), first dorsal interosseous (FDI) (B), and extensor digitorum (ED) (C) muscles. (D–F) The mean CV of EMG on-phase area for the APB, FDI, and ED muscles for the three stages are shown in panels (D), (E), and (F), respectively. n = 37, 37, and 36 for APB, FDI, and ED, respectively (** p < 0.01, * p < 0.05).
Fig 6.
Correlations between ball rotation rime and electromyography (EMG) variability.
(A–C) Spearman’s correlation coefficient between the ball rotation time and coefficient of variation (CV) of EMG on-phase duration for the abductor pollicis brevis (APB) (A), first dorsal interosseous (FDI) (B), and extensor digitorum (ED) (C) muscles. (D–F) Spearman’s correlation coefficient between the ball rotation time and CV of EMG on-phase area for the APB (D), FDI (E), and ED (F) muscles. Each closed circle represents the data for a single trial in one participant. The data for a total of 30 trials per participant are presented. n = 1110, 1110, and 1080 for APB, FDI, and ED, respectively.
Fig 7.
Correlations between the improvements in ball rotation time and reduction in electromyography (EMG) variability.
(A–C) Spearman’s correlation coefficient between the percent change (Δ) in ball rotation time and coefficient of variation (ΔCV) of EMG on-phase duration in abductor pollicis brevis (APB) (A), first dorsal interosseous (FDI) (B), and extensor digitorum (ED) (C) muscles. (D–F) Spearman’s correlation coefficient between Δ ball rotation time and ΔCV of EMG on-phase area in the APB (D), FDI (E) and ED (F) muscles.