Attention To Objects A Demonstrations of Object Based Attention

In addition to selecting regions of space, recent research has demonstrated that objects can be attended and selected independently of their locations. Although some investigators have argued that attention is object-based instead of space-based, the emerging consensus is that space-based and object-based attentional systems coexist. A more subtle issue, however, is whether the objects selected by attention are low-level retinotopically defined regions formed by gestalt grouping processes or higher level invariant objects formed by object representation processes. Again, the emerging consensus is that both types of object selection may exist.

1. Grouped Array Selection

One mechanism for object-based attention involves attending to perceptual objects that are defined in a spatial reference frame—a ''grouped array.'' The grouped array is a spatiotopic map in which locations or features are grouped according to gestalt principles such as similarity (e.g., features similar in color group with one another), closure (e.g., features that form closed shapes group with one another), or figure-ground relations (e.g., figures are closer to the viewer and are more salient than grounds). Selection from a grouped array representation involves attending to the locations of items that are grouped together.

Results from several behavioral studies are consistent with grouped array selection. Two strategies have been used to study this form of object selection. One strategy is to manipulate whether stimulus objects, such as letters, are grouped together based on secondary features, such as a common color or direction of motion. For example, if observers are asked to attend and categorize a target letter presented at fixation, their responses are influenced by adjacent flanking letters. Flanking letters that are consistent with the response to the target decrease observers' response times, and flanking letters that are inconsistent with the response to the target letter increase response times. The effect of flankers depends on object grouping factors: Flankers that group with the target letter influence responses more strongly than flankers that do not group with the target (Fig. 4). Presumably, the gestalt grouping principles define a set of letters as a related group, and this group is attended as a single unit. The flankers within this single perceptual group then decrease or increase response times to the target letter. Grouping effects on targets and flankers have been reported with several gestalt principles, including similarity, common fate, connectedness, and good continuation.

A second strategy for studying object-based attention is to manipulate whether two attended features fall on the same object or on different objects or, alternatively, whether attention shifts within an object or across objects. One widely used paradigm is illustrated in Fig. 5. In this paradigm, observers view

Figure 4 Displays used to study the influence of gestalt organization on visual attention. Observers report whether the central letter is an H or a T. (a) Grouping via similarity; the nontarget Hs group with the target H. Because the flankers that group with the target are compatible with one another, observers would classify the target letter quickly. (b) Grouping via good continuation; the nontarget Ts group with the target H. Because the grouped flankers are incompatible with the target H, observers would classify the target letter slowly.

Figure 4 Displays used to study the influence of gestalt organization on visual attention. Observers report whether the central letter is an H or a T. (a) Grouping via similarity; the nontarget Hs group with the target H. Because the flankers that group with the target are compatible with one another, observers would classify the target letter quickly. (b) Grouping via good continuation; the nontarget Ts group with the target H. Because the grouped flankers are incompatible with the target H, observers would classify the target letter slowly.

two rectangles oriented either horizontally or vertically. One end of one of the rectangles is precued with a brief flash, and this precue is followed by a target that requires a keypress response. The target usually appears at the cued location; when it appears at an uncued location, it may appear within the same object as the cued location or in the other object. Both of these uncued locations are the same spatial distance from the

Figure 5 Spatial precuing task used to study object-based attention. Two rectangles appear, and the end of one is precued with a peripheral flash. After a delay, a target appears at one of three locations: a, a validly cued target; b, an invalidly cued target that appears in the cued object; c, an invalidly cued object that appears in the uncued object. Observers more quickly detect invalidly cued targets appearing in the cued rectangle faster than those appearing in the uncued rectangle.

Figure 5 Spatial precuing task used to study object-based attention. Two rectangles appear, and the end of one is precued with a peripheral flash. After a delay, a target appears at one of three locations: a, a validly cued target; b, an invalidly cued target that appears in the cued object; c, an invalidly cued object that appears in the uncued object. Observers more quickly detect invalidly cued targets appearing in the cued rectangle faster than those appearing in the uncued rectangle.

precued region and the target is equally unlikely to appear at either of them, but observers are faster to respond to targets appearing in the uncued end of the cued rectangle than in either end of the uncued rectangle. Thus, attention appears to cover the entire cued rectangle even though only one end was cued. Similar results have been obtained with tasks that do not require spatial precues. In displays containing two rectangles that are overlapped to form an ''X,'' observers are faster to discriminate features on the same rectangle than on different rectangles even though the spatial distances are similar.

Note that spatial location is centrally important in selection from a grouped array because this representation is spatiotopic. A handful of studies have demonstrated the importance of location by manipulating both grouping principles and spatial position. These studies have demonstrated that both grouping principles and spatial position influence attentional selection. In the cued detection task depicted in Fig. 5, moving the rectangles closer to one another reduces the cost of switching attention from the cued rectangle to the uncued rectangle, although attention continues to shift faster within an object than between objects. On the basis of such results, it may be possible to explain many demonstrations of ''object-based attention'' as occurring within a spatially formatted representation.

2. Object-Centered Selection

A second mechanism for object-based attention involves attending to objects that are defined in an object-centered reference frame, which represents the parts and features of an object in relation to a reference point on the object. Coding parts and features in reference to the object allows the relative positions of the parts and features to be constant as the object changes spatial position and retinal size. A person's head is above the torso irrespective of where the person appears (e.g., left or right visual field) or how distant the person is from the viewer.

Because an object-centered reference frame is relatively insensitive to spatial position, spatial manipulations should not influence this form of object-based selection. There is evidence for selection from a late, object-centered representation from a discrimination task that requires observers to focus attention on a single object or divide attention between two objects —a box and a line (Fig. 6). Observers report the values of two features, such as whether a gap is on the left or right side of a box. Sometimes, the two features are

Figure 6 Object stimuli used to study object-based attention. Observers are more accurate in reporting attributes from the same object (e.g., box height and side of gap) than attributes from different objects (e.g., box height and tilt of line).

located on the same object (both features on the box) and sometimes on different objects (one feature on the box and one on the line). Performance is object-based in this task because features on the same object are reported more accurately than are features on different objects. However, unlike grouped array selection, performance in this task is not influenced by the spatial distance between the box and line, suggesting the objects are selected and not the locations occupied by the objects.

All About Alzheimers

All About Alzheimers

The comprehensive new ebook All About Alzheimers puts everything into perspective. Youll gain insight and awareness into the disease. Learn how to maintain the patients emotional health. Discover tactics you can use to deal with constant life changes. Find out how counselors can help, and when they should intervene. Learn safety precautions that can protect you, your family and your loved one. All About Alzheimers will truly empower you.

Get My Free Ebook


Post a comment