Untangling visual and proprioceptive contributions to hand localisation over time

Abstract

Previous studies showed that self-localisation ability involves both vision and proprioception, integrated into a single percept, with the tendency to rely more heavily on visual than proprioceptive cues. Despite the increasing evidence for the importance of vision in localising the hands, the time course of the interaction between vision and proprioception during visual occlusion remains unclear. In particular, we investigated how the brain weighs visual and proprioceptive information in hand localisation over time when the visual cues do not reflect the real position of the hand. We tested three hypotheses: Self-localisations are less accurate when vision and proprioception are incongruent; under the same conditions of incongruence, people first rely on vision and gradually revert to proprioception; if vision is removed immediately prior to hand localisation, accuracy increases. Sixteen participants viewed a video of their hands, under three conditions each undertaken with eyes open or closed: Incongruent conditions (right hand movement seen: inward, right hand real movement: outward), Congruent conditions (movement seen congruent to real movement). The right hand was then hidden from view and participants performed a localisation task whereby a moving vertical arrow was stopped when aligned with the felt position of their middle finger. A second experiment used identical methodology, but with the direction of the arrow switched. Our data showed that, in the Incongruent conditions (both with eyes open and closed), participants perceived their right hand close to its last seen position. Over time, the perceived position of the hand shifted towards the physical position. Closing the eyes before the localisation task increased the accuracy in the Incongruent condition. Crucially, Experiment 2 confirmed the findings and showed that the direction of arrow movement had no effect on hand localisation. Our hypotheses were supported: When vision and proprioception were incongruent, participants were less accurate and initially relied on vision and then proprioception over time. When vision was removed, this shift occurred more quickly. Our findings are relevant in understanding the normal and pathological processes underpinning self-localisation.