Integration of visual and physical motion cues for postural control in bipedal stance.

摘要

Visual cues play an important role in controlling our body posture during bipedal stance. However, visual cues are inherently ambiguous. When we perceive movement between ourselves and the world, visual cues alone cannot determine if we moved or the world moved or both moved together. The nervous system, therefore, needs to combine visual cues with physical motion cues to resolve this ambiguity and control posture. Despite numerous investigations, it still remains unclear how the nervous system integrates visual and non-visual indicators of posture. This thesis presents empirical findings that characterize the influence of visual and mechanical perturbations on human postural behavior. This thesis also presents a Bayesian model that formalizes the ambiguity problem encountered by the nervous system when it receives discordant sensory cues. The results demonstrate that movement of the visual surround at velocities greater than those experienced during unperturbed stance exhibits significant influence on body posture. Even in the presence of veridical kinesthetic and vestibular cues, the nervous system utilizes misleading visual cues to regulate body posture. Furthermore, the level of sensory discordance affects the biomechanical strategy of segmental control suggesting that the nervous system recruits lower-limb joints for postural stabilization depending on the intensity of both visual and mechanical perturbations. The success of the Bayesian model in predicting postural reactions of healthy subjects and patients with bilateral vestibular loss tested in this thesis and in published studies suggests that the nervous system integrates visual and physical motion cues within a probabilistic framework in order to resolve ambiguity, estimate and control posture. Thus the results presented in this thesis provide new insights into the neural and biomechanical mechanisms governing postural control in bipedal stance.