Abstract
In a familiar environment, memory representations may reduce attentional demands for visual information to guide behavior. Despite earlier evidence for sparse memory representations of complex scenes, more recent evidence has revealed the existence of more extensive memory representations (Castelhano & Henderson, 2007; Hollingworth, 2009). Experiments using 2D images have shown that visual search is facilitated by a scene preview, whether or not future search targets are present, suggesting that both memory for context and object-to-scene binding facilitate visual search in 2D images. However, memory representations during ongoing daily behaviors in the real world are a consequence of a very different exposure history than in typical experiments. In this study we examined visual search in an immersive virtual environment to determine (1) whether pre-exposure to the scene context and targets play a role in visual search, and (2) what aspects of the context might benefit visual search. The virtual environment was composed of two rooms; one room was explored for 1 minute before the search trials while the other was not. Early search targets were geometric objects, while later trials featured previously present realistic apartment objects. Pre-exposure to one of the rooms facilitated search for geometric objects (by approximately 25%), but only when the targets were present during exploration and only during early searches. When apartment objects are continuously visible, with opportunities for incidental fixations, further experience in the environment led to small search facilitation for those objects nearby previous targets. However, there was no facilitation for more distant objects despite having been present in the environment for over 5 minutes of immersive experience. Together the results suggest that prior exposure in immersive 3D environments leads to a small advantage in subsequent search but only when future search targets are also present, and the primary determinant of search time is context-relevant experience.
Meeting abstract presented at VSS 2014