Abstract
A fundamental challenge in the visual system is that objects and boundaries within a scene are all but stationary. For example, dynamic displacements of potential obstacles across time could either render a previous path unnavigable or provide a new path through which we can travel. How are the subsequent movements of a dynamic object accounted for as we encode navigability in a given space? In this study, we hypothesized that the encoding of object displacement in a navigationally relevant setting would be different from that in a navigationally irrelevant setting. We artificially rendered a 3D environment with a sliding door as a dynamic object, opening gradually as a camera moved forward towards the door. In experiment 1, participants saw a short video of the door opening to a certain width. In experiment 2, the 3D-cue of the rendered environment was minimized by replacing the video with a line drawing of a door opening as a camera moved forward. Experiment 3 further reduced the 3D-cue by simply translating a line drawing of a door horizontally, without a forward movement. After the video, participants were shown a probe image and answered if the door on the probe was narrower or wider than the video that preceded it. The point of subjective equality (PSE) was calculated to measure how participants encoded the dynamic object. We found a gradual decrease of the PSE as more 3D-cues were present in the visual environment. Specifically, participants overestimated the width of the door in simple line drawings (Experiment 2 and 3), suggesting that a representational momentum is in effect. Interestingly, participants underestimated the width of the door in a full 3D scene (Experiment 1), showing an opposite of representational momentum. These preliminary findings suggest differential encodings of dynamic objects in navigationally relevant environment.