Academic journal article Psychonomic Bulletin & Review

Visual Working Memory as the Substrate for Mental Rotation

Academic journal article Psychonomic Bulletin & Review

Visual Working Memory as the Substrate for Mental Rotation

Article excerpt

In mental rotation, a mental representation of an object must be rotated while the actual object remains visible. Where is this representation stored while it is being rotated? To answer this question, observers were asked to perform a mental rotation task during the delay interval of a visual working memory task. When the working memory task required the storage of object features, substantial bidirectional interference was observed between the memory and rotation tasks, and the interference increased with the degree of rotation. However, rotation-dependent interference was not observed when a spatial working memory task was used instead of an object working memory task. Thus, the object working memory subsystem-not the spatial working memory subsystem-provides the buffer in which object representations are stored while they undergo mental rotation. More broadly, the nature of the information being stored-not the nature of the operations performed on this information-may determine which subsystem stores the information.

How is the representation of a visual object stored while mental rotation is performed? Although many studies have focused on the process of mental rotation, previous research has not considered how representations are stored while being rotated. In general, working memory is thought to be used for the temporary storage and manipulation of information, so it seems plausible that object representations are buffered in working memory during mental rotation. Some sort of buffer seems necessary to hold an intermediate representation of the object while it is being rotated because the original object remains visible and therefore controls the low-level sensory representations of the object.

Working memory for visual information can be stored in two dissociable subsystems: a spatial subsystem that stores spatial information and an object subsystem that stores nonspatial object features, such as color and form. This distinction is made on the basis of evidence from behavioral studies (reviewed by Logic, 1995), singleunit recordings (reviewed by Goldman-Rakic, 1996), and neuroimaging experiments (reviewed by Jonides et al., 2003). Although the process of mental rotation is spatial in nature and involves the dorsal stream (Carpenter, Just, Keller, & Eddy, 1999; Gauthier et al., 2002; Jordan, Heinze, Lutz, Kanowski, & Jancke, 2001), it is plausible that the storage of this object representation occurs in the object working memory subsystem. That is, because there is no evidence that the spatial working memory subsystem can represent detailed form information, it is more likely that representations of object form are held in the object working memory subsystem during mental rotation.

To test this hypothesis, we used a dual-task approach that has been used to explore the role of working memory in visual search tasks. Woodman, Vogel, and Luck (2001) asked subjects to perform a visual search task during the retention interval of a working memory task, and performance in this dual-task condition was compared with the performance observed when participants performed the memory and search tasks individually. The working memory task was a change detection task in which observers viewed an array of objects, retained the objects in working memory briefly, and were then presented with a test array that was either identical to the original array or different in the color or shape of one of the objects. Observers simply reported whether or not they detected a change. When the visual search task was performed during the retention interval of this working memory task, the efficiency of the search process was not impaired in comparison with when the visual search task was performed alone. That is, increases in the set size for the visual search task led to longer search reaction times (RTs), but the search rate (i.e., the slope of the RT X set size function) did not vary between the dual-task and single-task conditions of the experiment. …

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.