CAREER: Role of Spatiotemporal and Identity Continuity in Object Constancy

University Of Texas At Arlington, Arlington TX

Investigators

Abstract

Philosophers, artists, and scientists for centuries have thought about how people organize the continuous flow of information reaching our sensory organs into discrete, individuated perceptual objects. As advances have been made in understanding vision, this topic has reemerged in cognitive psychology, artificial intelligence, and neuroscience. An object is individuated when the observer (1) perceives its features as belonging to a connected unit, (2) perceives its features as separate from things that are not part of the object, and (3) perceives that it is or is not something that was present sometime in the recent or distant past. An object that is believed to be the same even when it undergoes changes exhibits what is called object constancy. Much of what we know about object individuation and constancy in adult perception has been based on self-report methods. This research will explore the nature of object individuation and object constancy in tasks in which observers are attending to an object but are unlikely to be reasoning about its individuality or constancy. The research will employ a variety of tasks that have been developed in recent years to demonstrate ways in which attention to objects differs from attention to the space that objects occupy. The research will determine whether attention to objects is affected by disruptions to identity continuity or spatiotemporal continuity (i.e., the degree to which it is consistently present and moves through space in a physically possible way). By exploring which discontinuities disrupt attention to objects, the research will establish which continuities are critical to object constancy. In real scenes, spatiotemporal and identity discontinuity rarely are independent of each other. In contrast, this research will take advantage of media special effects that can dissociate them. For example, spatiotemporal discontinuity will sometimes be accomplished by an identifiable object disappearing unexpectedly and reappearing abruptly, and identity discontinuity will sometimes be accomplished by one object morphing smoothly into another. One part of the project will build on previous findings that once one part of an object is attended, it is easier to detect a subsequently presented target feature when it is elsewhere within the attended object rather than equidistant from the attended location but in another object. This project will investigate what disrupts this same-object benefit. For instance, spatiotemporal changes or identity discontinuities will be introduced to the cued or uncued object between the time the cue is presented and the time the target is presented, to see if the same-object benefit is disrupted. The research will also look for disruptions in speed and accuracy of responses to the display. Finally, it will also explore how eye movements are affected by the discontinuities. The direct goal of this research is to understand the fundamental nature of object constancy and individuation. However the results of the research will also have implications for engineering robotic systems that use machine vision to guide manipulation of dynamic objects. Such robotic systems need to differentiate which parts of visual information belong to coherent objects and which belong to different objects. In a different realm, knowledge of what sorts of changes distract people who are processing dynamic displays will aid designers of videos understand how special effects can be used to guide people's attention.