Academic journal article Psychonomic Bulletin & Review

Perceptual Expertise Enhances the Resolution but Not the Number of Representations in Working Memory

Academic journal article Psychonomic Bulletin & Review

Perceptual Expertise Enhances the Resolution but Not the Number of Representations in Working Memory

Article excerpt

Despite its central role in cognition, capacity in visual working memory is restricted to about three or four items. Curby and Gauthier (2007) examined whether perceptual expertise can help to overcome this limit by enabling more efficient coding of visual information. In line with this, they observed higher capacity estimates for upright than for inverted faces, suggesting that perceptual expertise enhances visual working memory. In the present work, we examined whether the improved capacity estimates for upright faces indicates an increased number of "slots" in working memory, or improved resolution within the existing slots. Our results suggest that perceptual expertise enhances the resolution but not the number of representations that can be held in working memory. These results clarify the effects of perceptual expertise in working memory and support recent suggestions that number and resolution represent distinct facets of working memory ability.

Working memory enables the maintenance of a limited number of representations in a rapidly accessible and easily manipulated state. This form of memory storage is widely acknowledged as a core component of most complex cognitive activities (e.g., ACT-R, Anderson, 1993; EPIC, Meyer & Kieras, 1997). Moreover, individual differences in working memory capacity are correlated with a variety of measures of general intelligence (Cowan et al., 2005). Thus, there has been widespread interest in understanding the factors that determine working memory capacity.

The present work focuses on how capacity in visual working memory is influenced by perceptual expertise. Perceptual expertise can be defined as an enhanced ability to carry out subordinate-level discriminations, typically as a result of extensive experience with a specific category of stimuli (e.g., Gauthier & Tarr, 2002; Tanaka & Taylor, 1991). It has been suggested that perceptual expertise enables more efficient mental representations of information (e.g., Gobet et al., 2001; Humphreys, Hodsoll, & Campbell, 2005; Moore, Cohen, & Ranganath, 2006), perhaps through "holistic" encoding that takes into account the relationships between individual features rather than just the features themselves. This raises the intriguing possibility that such efficient representations occupy less "space" in visual working memory, such that a greater number of objects can be stored. Curby and Gauthier (2007) examined this hypothesis by measuring change detection performance for faces, a class of stimuli for which the typical observer has perceptual expertise. In this paradigm (e.g., Luck & Vogel, 1997; Pashler, 1988), subjects are asked to remember an array of items over a brief delay and then to indicate whether any of these items has changed in a subsequent test display. The number of objects that can be simultaneously maintained is estimated by measuring change detection accuracy as a function of the number items in the sample array (Cowan, 2001; Pashler, 1988). Curby and Gauthier found that when subjects were given adequate time to encode these complex stimuli, capacity estimates were significantly larger for upright faces than for inverted faces. In addition, Curby and Gauthier found that the influence of orientation was larger for faces than for cars, suggesting that the advantage for upright faces may have resulted from holistic encoding processes that are typically associated with faces but not with cars. Thus, it was suggested that holistic processing enabled the formation of more efficient memory representations, thereby allowing a larger number of objects to be held in working memory.

Curby and Gauthier (2007) also pointed out, however, that holistic processes might have enabled an increased number of features to be encoded into each of the object representations stored in memory. In this case, change detection performance may have been enhanced for upright relative to inverted faces even though the same number of each stimulus type was stored. …

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.