Evidence for Two Attentional Components in Visual Working Memory. Allen, R., J., Baddeley, A., D., & Hitch, G., J.
Evidence for Two Attentional Components in Visual Working Memory [pdf]Paper  abstract   bibtex   
How does executive attentional control contribute to memory for sequences of visual objects, and what does this reveal about storage and processing in working memory? Three experiments examined the impact of a concurrent executive load (backward counting) on memory for sequences of individually presented visual objects. Experiments 1 and 2 found disruptive concurrent load effects of equivalent magnitude on memory for shapes, colors, and colored shape conjunctions (as measured by single-probe recognition). These effects were present only for Items 1 and 2 in a 3-item sequence; the final item was always impervious to this disruption. This pattern of findings was precisely replicated in Experiment 3 when using a cued verbal recall measure of shape–color binding, with error analysis providing additional insights concerning attention-related loss of early-sequence items. These findings indicate an important role for executive processes in maintaining representations of earlier encountered stimuli in an active form alongside privileged storage of the most recent stimulus. Research primarily using simultaneously presented arrays has shown substantial negative impacts of executive load on the rec-ognition and recall of visual objects (e.g., Allen, Baddeley & Hitch, 2006; Dell'Acqua & Jolicoueur, 2000). It is likely that executive control is also important during the encoding of se-quences and their retention in working memory. Furthermore, by analogy with studies of verbal short-term memory, it is possible that the impact of executive load may vary depending on serial position, potentially throwing light on the underlying cognitive processes. We investigate this by studying the effect of an exec-utive load on the retention of sequences of shapes, colors, and bound objects. The use of simultaneous presentation of all to-be-remembered items in the majority of recent studies examining visual memory (e.g., Luck & Vogel, 1997; Wheeler & Treisman, 2002) is sensible in that information from visual scenes involves parallel presenta-tion of an array of objects and features. However, for anything other than brief presentations, processing quickly becomes sequen-tial, with a series of eye movements being used to pick off crucial features of the array. Furthermore, naturalistic perception often involves changing scenes that are inherently sequential. If, as in verbal memory, sequential order of processing were important, then it would be valuable to take this into account. A useful alternative in this case is sequential stimulus presentation. Though memory for simultaneous and sequential displays is likely to involve similar mechanisms, the former method compresses these into a single, limited time period, meaning that it is often not able to draw clear distinctions between the possible operations of different components. In contrast, for an object sequence, each item is encoded in turn before being retained. Subsequent items are then presented. Examination of performance for items at each position in a sequence thus provides the opportunity to identify separable processes contributing to visual working memory. In the study of verbal memory, serial position effects have often proved informative, with serial recall favoring primacy and free recall recency, an effect that is particularly marked with auditory pre-sentation (e.g., Conrad & Hull, 1968; Murdock, 1966). Sequential presentation might be similarly informative for visual working memory. In a classic series of studies, Phillips (1974; Phillips & Badde-ley, 1971; Phillips & Christie, 1977a,1977b) demonstrated short-term forgetting of visual matrix patterns. When a sequence of patterns was presented and then probed via the change detection method, Phillips and Christie (1977a) found a recency effect of one item, with all other items detected at a much lower level (though still above chance). These results were attributed to separable contributions of long-term memory and visual short-term memory. However, this explanation is unlikely to apply to other observa-tions of recency effects (e.g., Allen et al., 2006; Parmentier, Tremblay, & Jones, 2004), where the repeated reuse of items from the same limited experimental set on each trial renders long-term memory uninformative (Endress & Potter, 2013). This leaves open the possibility that differential performance across a sequence reflects different components operating within working memory. In particular, it may be that the most recently encountered item retains a privileged status in working memory, being temporarily retained in a relatively automatic manner without the need for additional executive support. Such temporary retention is found in verbal short-term memory, where recency is unaffected by a

Downloads: 0