ABSTRACT
In the past decade, the investigations on the role of vision and vestibular organs in eye movement control, posture control and egomotion perception have emphasized the importance of interactions between sensory signals and the necessity to take into consideration any intervening motor commands in the sensory central processing itself. Most endeavors have focused on the interactions between semicircular canals and visual information. The otolithic information processing, on the other hand, was shown to be more complex and difficult to analyze, very likely because of the ambiguous nature of the otolithic signal (both gravitational and inertial). A comprehensive model, explaining how information originating from visual, otolithic and semicircular canals systems might interact, is still lacking.
In the present work, our attempt was to develop such a model, based on a few basic assumptions. The main hypothesis was that interactions do not occur between primary sensory signals, but rather between central representations of physical variables, such as head or gaze velocity and orientation. Each central representation receives a direct message from only one specific dedicated captor (for instance, the head angular velocity estimator receives a direct input from the semicircular canals), which provides a first approximated measure of the related physical variable. The central representations are then cross-coupled through functional connections that mimic the interrelations between physical variables, thus improving the coherence of the internal estimations.
496Three interdependent examples of this model are given: the head/gaze angular velocity coupling (involving the classical visual-vestibular interactions), the head linear acceleration/velocity coupling (involving the interactions between otolithic and visual informations) and the head angular velocity/orientation coupling (involving the intravestibular interactions). Quantitative simulations of the model are shown to be consistent with most experimental results obtained in a wide range of experimental conditions. A new perceptual illusion, still to be experimentally demonstrated, is also predicted by the model.
