Abstract
Vision contributes to balance, and vision is thought to dominate vestibular (or inertial) information in the perception of linear self-motion (Lishman & Lee, 1973; Mittelstaedt & Mittelstaedt, 2001). Is the visual system more sensitive to linear whole-field accelerations than are inertial systems? This would be surprising because the visual system is regarded as being much less sensitive to local acceleration signals than to local velocity signals (e.g., Eagle, 1996). We measured discrimination thresholds (JNDs) for peak velocity and for peak acceleration using immersive whole-field visual signals for linear motion (in an HMD) as well as non-visual whole-body inertial experiences (on a motorized cart). In both modalities, motion stimuli were developed in which peak acceleration and peak velocity were decoupled. The initial acceleration profiles in each case were roughly Gaussian, while the resulting velocity profiles were S-shaped. Acceleration duration was varied between 1 and 1.5 seconds, so that peak velocity could not be used to substitute for peak acceleration, nor could acceleration substitute for velocity without precise temporal integration. In a 2 X 2 design, stimuli were either visual (virtual hallway presented in a 60 deg FOV HMD) or inertial (on a computer-controlled cart), and judgments were either of peak velocity or peak acceleration. Observers made comparisons to an internal standard, with feedback. Despite the feedback, observers in the inertial experiments confounded velocity with acceleration, and JNDs for peak velocity discrimination from inertial senses were about 10% of the standard, whereas JNDs for peak acceleration were about 5%. Conversely JNDs for peak visual velocity were about 5% of the standard, while those for acceleration were about 10%. Evidently, visual superiority is limited to the perception of velocity. Visual and vestibular sensitivities may be complementary in perceiving accelerative and non-accelerative phases of self-motion.