Spatiomotor dynamics of hand movements during the drawing of memory-guided trajectories without visual feedback.

Abstract

Although the underlying principles of the spatiomotor dynamics during human movement execution are now broadly understood to conform to a minimum jerk principle, the question addressed in the present analysis is whether different principles operate during human drawing movements without visual input, deriving from studies of the Likova Cognitive-Kinesthetic Memory-Drawing Training. For two groups of participants, completely blind, and sighted but temporarily blindfolded, this analysis shows that the consensus model of arm-motion kinematics as a simple one-third power relationship of drawing speed to the local curvature of the line being drawn is not a sufficient characterization of their coupling. Instead, the drawing dynamics without visual feedback conform to a hyperbolic power relationship, with a coupling power of approximately 1.0 for regions of the highest curvature, asymptoting to curvature-independence for regions of shallow curvature, for both blind and blindfolded groups. Thus, the asymptotic power was much higher than the one-third power predicted by the minimum jerk principle. In detail, the maximum-velocity asymptote for both groups averaged about 6 cm/s for drawing from memory, increasing to more than twice as fast for mindless scribbling. We conclude that the more elaborate operating principle of a hyperbolic saturation function, with a power asymptote of about 1.0, may be interpreted as an adaptive implementation approximating the Minimum Jerk Principle of the simple one-third power law relating velocity and curvature.NEW & NOTEWORTHY This study reevaluates the one-third power law proposed to govern arm-motion kinematics relating drawing speed to the local curvature of the line being drawn. For complex drawings guided by memory without visual feedback, we find that the relationship is better characterized as a steeper power function that asymptotes to a constant speed for shallow curvatures, empirically approximating the predictions of the minimum jerk principle.