A multisensory posture control model of human upright stance.

We present a multisensory postural control model based on experiments where the balance in normal subjects and vestibular loss patients was perturbed by application of external torque produced by force-controlled pull stimuli. The stimuli were applied while subjects stood on a stationary or body-sway-referenced motion platform with eyes closed and auditory cues masked. Excursions of the center of mass (COM) and the center of pressure (COP) were analyzed using a systems analysis approach. The results were compared to an 'inverted pendulum' model of posture control. The model receives input from four sensors: ankle proprioceptors, semicircular canals, otoliths, and plantar pressure sensors (somatosensory graviceptors). Sensor fusion mechanisms are used to yield separate internal representations of foot support motion, gravity, and external torque (pull). These representations are fed as global set point signals into a local control loop based on ankle proprioceptive negative feedback. This set point control upgrades the proprioceptive body-on-foot (support) stabilization into a body-in-space control which compensates for support tilt, gravity, and contact forces. This compensation occurs even when the stimuli are combined or a voluntary lean is superimposed. Model simulations paralleled our experimental findings.