Non-linear Galilean vestibular receptive fields

We present a set of formulas for the receptive fields of the vestibular neurons that are motivated by Galilean invariance. We show that these formulas explain non-trivial data in neurophysiology, and suggest new hypothesis to be tested in dynamical 3D conditions. Moreover our model offers a way for neuronal computing with 3D displacements, which is reputed to be hard, underlying the vestibular reflexes. This computation is presented in a Bayesian framework. The basis of the model is the necessity of living bodies to work invariantly in space-time, allied to the necessary discreteness of neuronal transmission.