Evidence for a reference frame transformation of vestibular signal contributions to

36 To contribute appropriately to voluntary reaching during body motion, vestibular signals must be 37 transformed from a headto a body-centered reference frame. We quantitatively investigated the 38 evidence for this transformation during online reach execution by using galvanic vestibular 39 stimulation (GVS) to simulate rotation about a head-fixed, roughly naso-occipital axis as human 40 subjects made planar reaching movements to a remembered location with their head in different 41 orientations. If vestibular signals that contribute to reach execution have been transformed from a 42 headto a body-centered reference frame, the same stimulation should be interpreted as body tilt 43 with the head upright but as vertical-axis rotation with the head inclined forward. Consequently, 44 GVS should perturb reach trajectories in a head-orientation-dependent way. Consistent with this 45 prediction, GVS applied during reach execution induced trajectory deviations that were 46 significantly larger with the head forward as compared to upright. Only with the head forward 47 were trajectories consistently deviated in opposite directions for rightward versus leftward 48 simulated rotation, as appropriate to compensate for body vertical-axis rotation. These results 49 demonstrate that vestibular signals contributing to online reach execution have indeed been 50 transformed from a headto a body-centered reference frame. Reach deviation amplitudes were 51 comparable to those predicted for ideal compensation for body rotation using a biomechanical 52 limb model. Finally, by comparing the effects of applying GVS during reach execution versus 53 prior to reach onset we also provide evidence that spatially transformed vestibular signals 54 contribute to at least partially distinct compensation mechanisms for body motion during reach 55 planning versus execution. 56