Visual flow is interpreted relative to multisegment postural control.

To control upright stance, the human nervous system must estimate the movements of multiple body segments based on multisensory information. To investigate how visual information contributes to such multisegmental estimation, participants were exposed to 3 types of visual-scene movement: translation in the anteroposterior direction, rotation about the ankle joint, and rotation about the hip joint. Trunk and leg responses were larger for rotational than for translational movements, but only at lower stimulus frequencies. Based on a feedback-control theoretical framework, these results indicated that visual inputs distinguish between translation and rotation of the head. Also, visual condition effects were similar for the leg and trunk segments, suggesting a control strategy with a single control signal that determines the activation of all muscles.