Abstract
Both motor prediction theory (Wolpert & Flanagan, 2001) and inter-sensory inhibition theory (Barlow, 1991) suggest that perceived visual speed ought to be reduced during walking. The underlying principle in both cases is the notion that perception should encode departures from expectancy. But only inter-sensory theories, including Wallach's compensation theory (e.g., Wallach & Flaherty, 1975), predict that visual speed should also be reduced during passive self-motion. We used the method of magnitude estimation to create psychophysical functions relating perceived to actual visual speed (presented with an HMD) during active and passive self-motion. We found evidence of subtractive visual speed reduction while walking on a treadmill, while walking on solid ground, and while being rolled forward seated on a cart. The amount of speed subtraction while walking on solid ground was approximately the sum of the amount found for treadmill locomotion (motor component) and the amount found for passive self-motion (vestibular or inertial component). In the case of passive self-motion we also compared situations where the direction of visual and physical motion were consistent and inconsistent. Expanding flow fields showed substantial subtraction during forward passive movement, but were unaffected by backward passive movement. Conversely, contracting flow fields showed the most subtraction during backward passive self-motion. However, the inconsistent case of contracting flow fields during forward self motion also showed some reduction. Evidently, the canonical case of forward self-motion provides the strongest basis for visual speed reduction. No evidence was ever found for speed “addition” in cases where the two motions were inconsistent. Overall, these results are most consistent with associative theories, such as Barlow's, which may be regarded as a subsuming motor prediction theory if efference copy is treated as equivalent to a sensation in this context.
HHMI, Swarthmore College faculty research grant