Human standing postural control adjusts to biomechanical constraints: Is the CNS using multiple control plans, or a single flexible control plan?

Abstract

For a variety of postural perturbations, balance disorder patients exhibit difficulties in producing appropriate postural adjustments governed by the central nervous system (CNS), but the nature and design of the postural control are unclear. This study addresses the biomechanics involved in postural adjustments, and proposes an interpretation that posture control can be described as a feedback scheme with scalable gains, rather than as a large number of feedforward responses triggered by a perturbation. Thirteen healthy young adults were subjected to backward perturbations of a support surface ranging from 1.5cm to 15cm applied for 273msec in two different initial postures: upright and forward lean. The results showed that kinematics and joint torques trajectories gradually scaled as a function of perturbation magnitudes and initial leans before the constraints imposed by maximum allowable ankle torque became active. This indicates that the biomechanical constraints are represented in a soft form. A linear time-invariant feedback control model was able to reproduce the adaptive postural responses to the above perturbations. Optimization was used to identify a linear feedback gain matrix for each trial. These gains were then implemented with gain scheduling based on perturbation magnitude and initial leans in a multivariate manner. To provide a parsimonious model of postural adjustments to biomechanical constraints, we developed an optimal control model penalizing the violation of the constraints within the objective function. The existence of a global objective reproducing postural adjustments implies that CNS control can be interpreted as a feedback process that takes into account body dynamics and biomechanical constraints. To examine the mechanism of multisensory processing in postural balance, we examined the frequency response function (FRF) of visually induced postural sway to sinusoidal visual field stimulus ranging from 0.075Hz to 1Hz. An optimal estimator was able to reproduce the FRF of the data. The results suggest that the CNS makes use of an internal representation of body dynamics, and integrates sensory information to estimate body orientation and movement.