Extrinsic contributions to movement variability in human object manipulation

Abstract

Human movements vary across repetitions and the structure of this variability can shed light on motor control and learning. Previous studies showed that humans adapt their movements to minimize the effect of their intrinsic variability on the task goal. This study shows that in complex object manipulation there are also extrinsic contributions to variability from the deterministic nonlinear object dynamics that add to the resultant variability in human movements. This extrinsic contribution has two distinct components: the sensitivity of the movement to the object's states and the sensitivity to the interaction forces between hand and object. The hypothesis is that humans adapt to this variability, similar as to their intrinsic variability. This hypothesis is tested in a task that mimics the action of carrying a cup of coffee. The cup of coffee is simplified to a ball rolling in a 2D arc and modeled with cart-and-pendulum dynamics. To test the hypothesis, experimental data from 9 subjects were analyzed, who manipulated the complex object in a virtual environment. The two sensitivity measures were parametrically derived from the cart-and-pendulum model and were then evaluated numerically using the experimentally observed states of the system. Results showed that movement variability decreased with practice, mirroring the decrease in one sensitivity measure. This coincidence is consistent with the hypothesis that these extrinsic contributions affect the movement variability. Although causation cannot be shown, it is regarded important to quantify these extrinsic contributors to movement variability in order to understand the control of manipulation of objects with complex dynamics.