Consideration of movement errors under the signal-dependent noise hypothesis on repetitive arm movements

Abstract

Human movements vary on trial-by-trial due to signal-dependent noise that is added to motor commands and provokes errors of movements. In some motor tasks, humans have to produce precise movements with overcoming of the noise. This study aimed at considering the motor criteria taking into account movement errors. Three factors affecting movement errors in arm repetitive movements were investigated: the viscous force field, hand laterality, and signal-dependent noise. In the measurement experiment, the participants were asked to move their right and left hands between the target areas specified as the motor task. Pseudo null, positive (resist), and negative (assist) viscous force fields were set in the workspace of the motor task. A two-way analysis of variance involving two factors, hand (right and left) and viscosity (null, resist, and assist), in the movement error was conducted, and it revealed significant main effects of hand and viscosity, and a significant interaction between them. The data obtained from the measurement suggested consistent characteristics for the shape of the paths of the movement trajectories. A simulation experiment was conducted to investigate the shape characteristics which are regarded as the nature of the human motor control system. Under the assumption of signal-dependent noise, results of the experiment revealed that the shape characteristics are associated with lower variance of a movement error directly relating the task achievement rather than an error of a movement endpoint. The lower variance of the movement error is of greater significance as the criteria of motor planning compared to the smoothness of the movement.