Abstract
Most object-based attention research over the past two decades has been conducted with simple, geometric objects (e.g., rectangles, semicircles, trapezoids, crossed lines, letters). This research has shown that, under specific circumstances, objects guide attentional selection in a predictable manner (Chen & Cave, 2006, 2008; Goldsmith & Yeari, 2003; Martinez et al., 2006; Richard, Lee, & Vecera, 2008; Shomstein & Behrmann, 2006, 2008; Shomstein & Yantis, 2002, 2004). Object-based guidance of attentional selection with simple objects is then assumed to transfer from the lab setting to complex objects in the real world. Here, we test this assumption by conducting a series of experiments that investigate whether real-world objects guide attention similarly to simple objects, and whether a set of circumstances that are known to yield differences in object-based guidance in simple objects (probability imbalances, reward biases) behaves similarly with real-world objects. It is not readily expected that object-based attention will be observed with real-world objects as those objects have more complex low-level (i.e., complexity of edges, frequency composition) and high-level (semantic information, learned associations) properties. We used a set of simple rectangles, 50 real-world objects with horizontal primary axis, and 50 real-world objects with a vertical primary axis (manipulated between subjects). We observed a robust object-based effect for real-world objects (p< .002), that was in fact greater than that observed for rectangles (p< .05). Additionally, we probed the robustness of this effect with various biases against the same object location (e.g., reward and probability). Our findings extend the literature on object-based attention to include real-world objects, and suggest that objects guide attentional selection even in complex environments that are full of low- and high-level variability.
Meeting abstract presented at VSS 2016