Table 1 shows the mean reaction and movement time. For reaction time, there was a significant main effect of protocol, F(1,12) = 22.30, p <.001, ƞp2 = 0.65, indicating a shorter time to initiate for execution (marginal M = 388 ms, SE = 21) compared to imagery (marginal M = 828 ms, SE = 105). There was no significant main effect of obstacle, F(1,12) = 0.22, p =.655, ƞp2 = 0.02, nor a significant protocol x obstacle interaction, F(1,12) = 1.64, p =.225, ƞp2 = 0.12.

For movement time, there was a significant main effect of protocol, F(1,12) = 18.57, p =.001, ƞp2 = 0.61, indicating a shorter time within the movement for execution (marginal M = 733 ms, SE = 35) compared to imagery (marginal M = 1143 ms, SE = 95). In addition, there was a significant main effect of obstacle, F(1,12) = 44.28, p <.001, ƞp2 = 0.79, indicating a shorter time for no obstacle (marginal M = 843 ms, SE = 50) compared to obstacle (marginal M = 1033 ms, SE = 60). However, there was no significant protocol x obstacle interaction, F(1,12) = 0.01, p =.925, ƞp2 = 0.00.

For movement time, there was a significant main effect of trial n, F(1,12) = 43.75, p <.001, ƞp2 = 0.79, indicating a shorter time for no obstacle (marginal M = 656 ms, SE = 47) compared to obstacle (marginal M = 859 ms, SE = 61) (Table 2). Meanwhile, there was a significant trial n-1 x trial n interaction, F(1,12) = 11.16, p =.006, ƞp2 = 0.48, although the shorter time in trial n with an obstacle following trial n-1 with no obstacle (marginal M = 830 ms, SE = 57) compared to obstacle (marginal M = 888 ms, SE = 66) failed to reach significance (p =.032, Holm-Bonferroni corrected alpha level = 0.025). There were no other significant main, nor interaction effects involving the factor of protocol [protocol: F(1,12) = 0.22, p =.648, ƞp2 = 0.02; protocol x trial n-1: F(1,12) = 1.95, p =.188, ƞp2 = 0.14; protocol x trial n: F(1,12) = 4.11, p =.065, ƞp2 = 0.26; protocol x trial n-1 x trial n: F(1,12) = 0.60, p =.453, ƞp2 = 0.05].

For early angular deviation, there was a significant main effect of trial n, F(1,12) = 356.95, p <.001, ƞp2 = 0.97, indicating significantly less deviation for no obstacle (marginal M = 2.52°, SE = 4.22) compared to obstacle (marginal M = 52.22°, SE = 2.81). Meanwhile, there was a significant protocol x trial n-1 interaction, F(1,12) = 5.81, p =.033, ƞp2 = 0.33 (Fig. 2a), indicating significantly less deviation following execution at trial n-1 with no obstacle compared to obstacle (p =.0166, Holm-Bonferroni corrected alpha level = 0.017), while there was no such difference following imagery at trial n-1 (p =.26). In addition, there was significantly less deviation following execution compared to imagery at trial n-1 with no obstacle (p =.001, Holm-Bonferroni corrected alpha level = 0.013). There were no other significant main, nor interaction effects involving the factor of protocol [protocol: F(1,12) = 3.37, p =.091, ƞp2 = 0.22; protocol x trial n: F(1,12) = 0.05, p =.823, ƞp2 = 0.00; protocol x trial n-1 x trial n: F(1,12) = 3.61, p =.082, ƞp2 = 0.23].

For maximum height, there was a significant main effect of protocol, F(1,12) = 24.52, p <.001, ƞp2 = 0.67, and trial n, F(1,12) = 356.95, p <.001, ƞp2 = 0.97. Meanwhile, there was a significant protocol x trial n-1 interaction, F(1,12) = 6.87, p =.022, ƞp2 = 0.36 (Fig. 2b), indicating a significantly shorter height following execution at trial n-1 with no obstacle compared to obstacle (p =.004, Holm-Bonferroni corrected alpha level = 0.017), while there was no such difference following imagery at trial n-1 (p =.094). In addition, there was a significantly shorter height following execution compared to imagery at trial n-1 with no obstacle (p <.001, Holm-Bonferroni corrected alpha level = 0.013) and obstacle (p =.009, Holm-Bonferroni corrected alpha level = 0.025). Finally, there was a significant protocol x trial n interaction, F(1,12) = 9.72, p =.009, ƞp2 = 0.45, indicating a significantly shorter height following both execution (p <.001, Holm-Bonferroni corrected alpha level = 0.013) and imagery (p <.001, Holm-Bonferroni corrected alpha level = 0.017) at trial n with no obstacle (execution: marginal M = -2.46 mm, SE = 0.88; imagery: marginal M = 3.54 mm, SE = 0.67) compared to obstacle (execution: marginal M = 41.84 mm, SE = 2.06; imagery: marginal M = 43.69, SE = 2.06), while there was significantly shorter height following execution compared to imagery at trial n with no obstacle (p <.001, Holm-Bonferroni corrected alpha level = 0.025). There was no significant protocol x trial n-1 x trial n interaction, F(1,12) = 2.31, p =.154, ƞp2 = 0.16.

Mean early angular deviation (A) and maximum height (B) as a function of protocol and trial n-1 (see legend). Error bars indicate the standard error of the mean. (*) indicates a significant difference in a pairwise comparison (Holm-Bonferroni corrected)

In the absence of trajectory priming following imagery at trial n-1, it is possible that some participants had poor imagery ability. If so, then we would anticipate a relation between trajectory priming and visual and/or kinaesthetic imagery ability. Therefore, we conducted a supplementary analysis by correlating the difference between obstacle and no obstacle at trial n-1 with imagery ability scores. For early angular deviation, there was no significant relation for kinaesthetic, r(13) = − 0.06, p =.86, nor visual, r(13) = − 0.10, p =.75, imagery ability. For maximum height, there was no significant relation for kinaesthetic, r(13) = 0.46, p =.11, but there was for visual, r(13) = 0.64, p =.02, imagery ability. Further inspection of the latter correlation indicated that it was primarily attributed to a single case, which once removed indicated that there was no longer a significant relation, r(12) = 0.39, p =.21.