Abstract
Objects frequently have a hierarchical organization (tree-branch-leaf). How do we select the level to be attended? This has been explored with compound letters: a global letter built from local letters. One explanation, backed by much empirical support, is that attentional competition is biased toward certain spatial frequency (SF) bands across all locations and objects (a SF filter). This view assumes that the global and local letters are carried respectively by low and high SF bands, and that the bias can persist over time. Here we advocate a complementary view in which perception of hierarchical level is determined by how we represent letters in object-files. Although many properties bound to an object-file (i.e., position, color, even shape) can mutate without affecting its persistence over time, we posit that same object-file cannot be used to store information from different hierarchical levels. Thus, selection of level would be independent from locations but not from the way objects are represented at each moment. These views were contrasted via an attentional blink paradigm that presented letters within compound figures, but only one level at a time. Attending to two letters in rapid succession was easier if they were at the same-compared to different-levels, as predicted by both accounts. However, only the object-file account was able to explain why it was easier to report two targets on the same moving object compared to the same targets on distinct objects. The interference of different masks on target recognition was also easier to predict by the object-file account than by an SF filter. The methods introduced here allowed us to investigate attention to hierarchical levels and to object-files within the same empirical framework. The data suggests that SF information is used to structure the internal organization of object representations, a process understood best by integrating object-file theory with previous models of hierarchical perception.
Highlights
IntroductionOBJECT BASED ATTENTION we can choose to attend to anything that happens at a given spatial location (i.e., the goal zone in a match of the FIFA World Cup), or to a specific feature (i.e., find the black uniforms in the playfield), we often focus on visual objects
OBJECT BASED ATTENTION we can choose to attend to anything that happens at a given spatial location, or to a specific feature, we often focus on visual objects
This test assumes that there should be little competition between two pieces of information arising within the same unit of attentional selection, in contrast to strong competition when these pieces originate from distinct units (Duncan, 1984; Kravitz and Behrmann, 2011)
Summary
OBJECT BASED ATTENTION we can choose to attend to anything that happens at a given spatial location (i.e., the goal zone in a match of the FIFA World Cup), or to a specific feature (i.e., find the black uniforms in the playfield), we often focus on visual objects. The last alternative is especially sensible from an ecological point of view, given that most of our interactions with the world are precisely directed at objects (i.e., we grasp/eat/avoid/or-flee-from objects or we boo at them if they fail to score a goal). These alternatives for defining the units of selection are known respectively as spatial-based, feature-based and object-based attention (Serences et al, 2004). Note that TTT elegantly keeps several confounding factors other than attention (such as the number of perceptual decisions, working memory load, and response competition) constant across the focused/divided attention comparison
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have