Effect of position influence and environmental changes in hybrid torque-position BMI.

(A) Decreasing position influence (, see methods) from 200 (n = 32) to 100 (n = 46; 42) to 20 (n = 75) and (B) from 20 (n = 127; 92) to 0 (pure torque control; n = 116; 77) moved the FSV locations away from the movement start-point and towards the target (P<0.05), indicating more ballistic movement profiles. (C) Increased viscous gain-fields (Increased Load, IL, see Table 1; right; n = 74) from before (Equal Load, see Table 1; left; n = 77) moved the FSV locations towards the movement start-point and decreased the FSV vector lengths (P<0.05), indicating decreased movement velocities under high resistance. (D) Introduction of new viscous gain-fields (New Load, NL, see Table 1; right; n = 81) within the bounds of ongoing low- and high- gain fields (Routine Load, see Table 1; left; n = 76) led to no statistically significant differences in the FSV locations or vectors. Same conventions as in Figure 1, E, F are used. Means and standard deviations of the FSV locations and vectors of different sessions (mean FSV location: base of arrow; standard deviation of FSV location: dashed circle, scaled 10%; mean FSV vector: arrow itself; standard deviation of FSV vector: thin solid line circle, scaled 10%) are overlaid in between the FSV plots for easy comparison; blue: from the FSV plot on the left, red: from the FSV plot on the right. The arm was restrained at the bottom left (75° shoulder and 85° elbow angles, A, D) or at the center (25° shoulder and 85° elbow angles, B, C) of the workspace unless indicated as F, free or unrestrained arm.