Abstract
Boundary extension (BE) is as an error in scene memory, such that participants retrieve details beyond the given boundaries of a scene image. Boundary contraction (BC) is the opposite effect, whereby participants retrieve less context within the boundaries of a given scene image. In the BE literature, there is variability in the types of stimuli that are used, how BE is tested, and the proposed mechanisms underlying the phenomenon. Some research supports the view that BE reflects (re)construction of the scene from an internal representation that was formed, whereas other research supports the view that BE (and BC) emerge from image-based properties, including the number of central objects and whether an object is pictured in close range or from a wider angle. Assessing the effects of prior knowledge and experience of a scene on this bias can help disentangle the role of visual perception and scene construction. The current study tested the influence of familiarity on scene recognition through the comparison of lab-based encoding of images of pre-experimentally familiar (real-world) places with images of unfamiliar places. Participants used a continuous rating scale to indicate how they perceived the boundaries of a test image relative to a previously studied image. There was a tendency for BC across both image conditions, with evidence of maintained, and an instance of greater, BC for familiar than unfamiliar scene images. Importantly, the lack of evidence for increased BE with greater familiarity favours an image-based theoretical account of BE and BC.