Purchase this article with an account.
Christopher Dickinson, David Crane, J. Chapman Munn, Reiss Powell, Leah Stephens, Jordan Todd; Object Orientation Influences False Memory for the Shape of a View. Journal of Vision 2011;11(11):1134. doi: https://doi.org/10.1167/11.11.1134.
Download citation file:
© ARVO (1962-2015); The Authors (2016-present)
Intraub's (2010) multisource model of scene perception suggests that representations of scene views incorporate both bottom-up and top-down input. We examined the contribution of bottom-up information to memory for unseen layout beyond the edges of a view (boundary extension; BE) by testing whether the way an object fills the space of a view influences BE. Stimuli were photographs of overhead views of scenes showing a single object against a natural ground surface (e.g., a hammer on gravel); all stimulus views were square. Objects filled views in one dimension but not the other (e.g., a screwdriver, flashlight). Observers (N = 39) viewed a sequence of 12 close-up views for 15 s each (views alternated with 1-s masks); half the objects were oriented vertically; the other half horizontally. All objects appeared both horizontally and vertically across participants. Horizontal-vertical viewing information was made comparable by placing square frames in front of monitors; stimuli were viewed within a square cutout in the frames' centers. After viewing the 12-picture sequence, observers were instructed to adjust each picture's four view-boundaries to recreate the original view (initial border placement in test views was identical to that in stimulus views). An increase in the overall area of views would indicate BE. In addition to finding significant BE across observers (mean increase in area - vertical objects: 17.2%; horizontal objects: 14.9%; both ps < .001), we found that object orientation influenced BE: for both object orientations, significantly more BE occurred along the object's longer axis (both ps < .001). In addition, for vertical objects, significant BE was found along both object axes (width: 5.1%; height: 10.2%); for horizontal objects, significant BE was found only along the longer axis (width: 12.2%; height: 1.1%). The results suggest that BE is influenced by how objects fill the space of a view, with extension along the pictures' horizontal and vertical axes being differentially affected by object orientation.
This PDF is available to Subscribers Only