Influence of dynamic tilts on the perception of earth-vertical.

The aim of this study was to test the hypothesis that optimal activation of both the semicircular canals and the otoliths provides reliable vestibular cues about self-orientation in space. For this, we measured the ability of subjects to estimate the subjective vertical immediately, 20 s and 90 s after a rapid tilt (180 degrees /s(2)) from upright into different roll positions between 90 degrees left and right side down. Subjects had to estimate the earth-vertical and earth-horizontal direction in the dark by (a) setting a luminous line, (b) performing saccades, and (c) verbally declaring body position relative to gravity. The mean error curves from the three paradigms showed consistent E (Müller)- and A (Aubert)-effects, which did not significantly change over time. Horizontal and vertical saccade tasks exhibited different response characteristics, as previously reported by others, which likely reflect different computation mechanisms. The verbal estimation paradigm yielded complementary results to those of the luminous line paradigm and vertical saccade task. The E-effect of the luminous line and the vertical saccade paradigm might be explained by a bias towards earth-vertical due to interactions of vestibular and neck afferent signals. The invariably small A-effect of the luminous line and the vertical saccade paradigm probably reflects somatosensory signals that had relatively weak influence in our experiments. We conclude that phasic activation of the vestibular system reduces the influence of non-vestibular cues observed in low tilt velocity or static experiments. Although this activation generates an E-effect, the total error in the range of +/-90 degrees is reduced.