Divided Attention and Memory

Divided Attention Tasks in Perception and Working Memory Research

The dual-task technique has been used extensively to study the general processing capacity of humans. Daniel Kahneman's view of human information processing, in which attention is described as drawing on a single, limited, pool of resources, was challenged based on dual-task research. While some cognitive tasks were shown to be difficult to combine, because they drained a single pool of attentional resources, others could be combined relatively easily, suggesting that multiple pools of attention exist. For example, repeating aloud (shadowing) sentences in a dichotic listening test is hampered when attention is divided with another task involving comprehension of simultaneously presented words. In contrast, Alan Allport and his colleagues found little interference when attention was divided between word shadowing and a task such as playing piano music or recognizing pictures. Along the same lines, Lee Brooks showed that reading, a visual task, was more difficult when performed concurrently with another task that required internal visualization (imagery), than it was when combined with one without a visualization component. Such results were incompatible with a single-resource view and led researchers [Page 261]to propose that humans have multichannel processors or pools of attention resources; it is only when the same ones are required simultaneously for two tasks that performance will decline.

Marcel Kinsbourne and Robert Hicks attempted to understand the results of dual-task experiments by proposing that the degree of interference from simultaneous tasks is an inverse function of the functional distance between cerebral control centers. By contrast, research by Hal Pashler and his colleagues showed that competition for a central response channel forms a bottleneck that can account for much of the slowing that occurs when two tasks, be they memory, perception, or motor tasks, are combined.

The dual-task technique has also been used to elucidate the characteristics of working memory, which provides temporary storage and allows the active manipulation of information necessary for complex cognitive tasks such as arithmetic and reading comprehension. The model of working memory first proposed by Alan Baddeley and Graham Hitch in 1974 was influenced by various DA experiments and drew its support from them. The model posits a limited-capacity central executive component, which coordinates operations of two subsidiary storage systems, the phonological loop and visuospatial sketchpad. These subsystems provide temporary storage/representation for verbal and visuospatial material, respectively. Support for the model came from Robert Logie's studies showing a greater disruptive effect of unattended speech on short-term memory for words or digits than on memory for visuospatial information such as patterns on a visually presented grid, and vice versa if the task involved visual imagery and the interference was spatial.

...