Effect of inverted visual acceleration profile on vestibular heading perception.

Visual motion is ambiguous in that it can either represent object motion or self-motion. Visual-vestibular integration is most advantageous during self-motion. The current experiment tests the hypothesis that the visual motion needs to have a motion profile consistent with the inertial motion. To test this, we examined the effect on heading perception when the visual stimulus was consistent with the inertial motion compared to an inverted visual stimulus, which was thus inconsistent with inertial motion. Twenty healthy human subjects (mean age 20 ± 3 years, 13 female) experienced 2s of translation, which they reported as left or right. A synchronized 2s visual heading was offset by 0°, ± 45°, ± 60°, or ±75°. In randomly interleaved trials, the visual motion was consistent with the inertial motion or inverted - it started at the peak velocity, decreased to zero mid-stimulus, and then accelerated back to the peak velocity at the end. When the velocity profile of the visual stimulus matched the velocity profile of inertial motion, the inertial stimulus was biased 10.0 ± 1.8° (mean ± SE) with a 45° visual offset, 8.9 ± 1.7° with a 60° offset, and 9.3° ± 2.5 ± with a 75° offset. When the visual stimulus was inverted so it was inconsistent with the inertial motion, the respective biases were 6.5 ± 1.5°, 5.6 ± 1.7°, and 5.9 ± 2.0°. The biases with the inverted stimulus were significantly smaller (p < 0.0001), demonstrating that the visual motion profile is considered in multisensory integration rather than simple trajectory endpoints.