Building upon recent results which have shown that visual and vestibular signals combine in a statistically optimal fashion for heading, we investigate the relative weights of visual and vestibular cues during self-motion. To do investigate this, participants performed a 2-interval forced choice task (2IFC) in all conditions in which they were asked to judge “in which of the two intervals did you move more to the right�. To observe the weights assigned to each modality we introduced a discrepancy between the visual and vestibular cues in the standard visual-vestibular heading. The experiment comprised of nine conditions: vestibular alone, four visual alone at different standard headings and four visual-vestibular with small and large conflict levels (Δ=±6° or ±10°) between the visual and vestibular input. We found that even when there is a large conflict between the visual and vestibular cues; participants exhibit statistically optimal reduction of variance of visual and vestibular information. On the other hand, we found that the unimodal cues did not predicted the weights in the combined cue. We conclude that visual and vestibular cue combination is not predicted solely by the reliability of each cue but that there is a prior which lends more weight to the body centric cue.