Differential Angular Expansion in Perceived Direction in Azimuth and Elevation Are Yoked to the Presence of a Perceived Ground Plane

Abstract

It has been proposed that perceived angular direction relative to straight-ahead is exaggerated in perception, and that this exaggeration is greater in elevation (or declination) than in azimuth. Prior research has suggested that exaggerations in elevation may be tied to the presence of a visual ground plane, but there have been mixed results across studies using different methods of dissociation. In the present study, virtual environments were used to dissociate visual from gravitational upright while human participants (N = 128) made explicit angular direction judgments relative to straight ahead. Across these experimental manipulations, observers were positioned either upright (Experiments 1A and 1B) or sideways (Experiment 2), so as to additionally dissociate bodily orientation from gravitational orientation. In conditions in which a virtual environment was perceived as containing a level ground plane, large-scale exaggerations consistent with the visually-specified orientation of the ground plane were observed. In the absence of the perception of a level ground plane, angular exaggerations were relatively small. The ground plane serves as an important reference frame for angular expansion in the perceived visual direction.

Keywords: angular expansion; space perception; visual perception.

Figure 1

(a) The experimental condition in Experiment 1A, in which a vertical edge is straight ahead and there is no ground plane. The target is to the right. Participants sat to the left of the center of the screen to allow a greater field of view to the right, where stimuli were presented. A small zero is present to show where the spoken response appeared once typed in by the experimenter. (b) shows the experimental condition in Experiment 1B, in which the ground plane orientation is reoriented by 90° about straight-ahead. (c) shows the baseline comparison condition, for both Experiments 1A and 1B, in which estimates of azimuth were collected with a (gravitationally) horizontal ground plane present. We refer to this as the upright or normal ground plane.

Figure 2

Mean direction estimates (with standard error bars) for the upright (normal) ground plane condition (black circles) and for the unique experimental conditions (white circles) for (a) Experiment 1A (no ground plane) and for (b) Experiment 1B (sideways ground plane). Only data from the first-tested conditions for each participant are shown in each plot. Note that the best-fit lines shown here are 2-parameter fits, including an intercept, whereas the gain values used for our statistical analyses were based on a single parameter fit (slope only).

Figure 3

Summary data of the angular direction estimates in Experiment 2 by world orientation for those who did the relevant orientation first. For the sideways world, with sideways observers, where physical azimuth directions appeared as elevations above a ground plane, the angular estimates closely match the expected 1.5 gain normally found for perceived elevation direction.

Figure 4

Mean angular gains for (gravitationally-defined) azimuth direction with 95% confidence intervals shown for three different visual worlds. Only data from the first world tested for each participant are shown, so comparisons are all between-subject and are uncontaminated by order effects.