Academic journal article Psychonomic Bulletin & Review

Spatial Short-Term Memory Assists in Maintaining Occluded Objects

Academic journal article Psychonomic Bulletin & Review

Spatial Short-Term Memory Assists in Maintaining Occluded Objects

Article excerpt

We examined the role of spatial short-term memory (SSTM) in maintaining the hidden portions of occluded objects. We measured the degree of maintenance by asking participants to perform an object-based attention task on occluded objects while maintaining four locations in SSTM. SSTM did not interfere with object-based selection generally. Maintenance of occluded portions of objects was prevented when four locations were held in SSTM and these locations did not overlap with the locations of occluded objects. Importantly, when the observers remembered locations that overlapped with the occluded regions of objects, the occluded objects were maintained behind the occluder. These results suggest that SSTM participates in maintaining the occluded sections of objects.

Due to the movement of our body and objects in our environment, our visual system is confronted with visual information that contains objects that are fully visible at one moment and partially or totally occluded by other objects at another moment. However, both primate and nonprimate visual systems are highly efficient in representing and recognizing objects across the variability created by occlusion (for completion in nonhumans, see Fujita, 2001; Kanizsa, Renzi, Conte, Compostela, & Guerani, 1993; Nagasaka & Wasserman, 2008). Visual completion serves as a mediating mechanism, enabling objects to be completed behind occluding surfaces and allowing the occluded object to appear as a single object that continues behind the occluder (Kanizsa, 1975; Michotte, Thinès, & Crabbé, 1964/1991).

Visual completion processes have been shown to operate quickly on the basis of bottom-up image information; in most situations, higher level cognitive operations are not required to infer the shape of an occluded object (e.g., Davis & Driver, 1998; He & Nakayama, 1992; Moore, Grosjean, & Lleras, 2003; Rensink & Enns, 1998). For example, in Figure 1A, observers subjectively experience an elongated occluded horse, although the percept violates the knowledge that such an animal is not possible in the real world (see Kanizsa & Gerbino, 1982, for a similar demonstration). Image cues that support visual completion include good contour continuation (Wouterlood & Boselie, 1992) and cotermination and T-junctions (Rensink & Enns, 1998). Visual completion also has been described as the result of interpolating between the visible segments of occluded objects (Kellman & Shipley, 1991).

More objective results for potent image-based completion come from Pratt and Sekuler (2001), who demonstrated that previous exposure to a scene did not affect completion. In their experiments, observers saw a preview display containing four rectangles, followed by an occluder (Figure 1B). When the occluder was present, only the ends of the four rectangles were visible, and phenomenologically, the display appeared to contain two longer, occluded rectangles (not four separated rectangles). Next, a peripheral cue summoned attention to the end of one of the rectangles, followed by several shapes that appeared inside the rectangles. Observers were asked to report the identity of the largest shape. Such object-based attention tasks reveal faster responses to uncued (or invalidly cued) targets appearing at the other end of the cued object than to those invalidly cued targets appearing on the other, uncued object (Egly, Driver, & Rafal, 1994; Vecera, 1994). Pratt and Sekuler showed that an object-based attentional benefit based on a completed image was present even when observers saw the preview display containing four distinct objects. These findings indicate that image cues such as T-junctions and colinear lines easily override observers' knowledge of four disconnected objects behind an occluder.

Although most research on visual completion supports a bottom-up view, some findings suggest that later visual processes can influence completion or the outputs of completion processes. …

Search by... Author
Show... All Results Primary Sources Peer-reviewed

Oops!

An unknown error has occurred. Please click the button below to reload the page. If the problem persists, please try again in a little while.