Abstract
How do we construct a seamless visual experience despite changes in the current view? Recent studies suggest that overlapping visual content between individual views plays a key role in forming panoramic scene memory. What reference points are used when integrating individual views into a panoramic representation? In the current study, we tested the effect of the overlapping visual experience in two different types of reference points: viewer-centered and scene-centered. In each trial, participants viewed two short clips sampled from a 360° panoramic environment. In the viewer-centered condition, the view changed as if the viewer rotated the head to survey the environment in a fixed position. In the scene-centered condition, the view changed as if the viewer revolved around an invisible pivot inside the scene. In each condition, the two clips either shared a portion of the visual content (Overlap Condition) or not (No-Overlap Condition). In the following memory test, participants were asked to determine whether the two snapshot images from the opposite poles of the studied environment were drawn from the same place. Our results show higher panoramic scene memory performance for the images presented in the Overlap Condition compared to the No-Overlap Condition, replicating previous findings. Importantly, the overlap effect was significantly greater in the viewer-centered than in the scene-centered reference point condition. These results demonstrate that reference point information is utilized along with visual overlap cues when integrating scenes in memory, suggesting a potential role of a viewer-centered reference frame in constructing a coherent visual experience.