Verbal and visual subsystems of working memory.

Two recent PET scanning studies have revealed the anatomical basis of the verbal and visual subsystems of working memory. The term 'working memory' refers to the system responsible for the temporary maintenance of information~nec-essary for performing such cognitive tasks as reasoning, understanding and learning. Evidence from a range of studies based on the functioning of working memory in normal and brain d,amaged patients, led myself and G. Hitch to propose a model for working memory that assumes it has three components. The first is an at-tentional controller, a 'central executive' that is responsible for strategy selection and cognitive control. This central system is aided by two subsidiary slave systems, the 'phonological loop' that is responsible for maintaining and manipulating speech-based information, and the 'visuo-spatial sketchpad' that is responsible for holding and manipulating visual images [ 1~1. The phonological loop The phonological loop is thought to be a system with two components: a brief memory store able to hold acoustic or phonological information, coupled with an articulatory control system capable of maintaining information by sub-vocal rehearsal and also of entering new information into the memory store by means of subvocal naming. Evidence for the phonological memory store comes from the acoustic similarity effect, whereby sequences of items that are similar in sound, such as the words MAN, CAD, MAT, MAP, CAN, are harder to remlember and repeat back than dissimilar words such as PIT, DAY, HOT, COW, PEN, while similarity of meaning between the words has little or no effect [3 1. The articulatory control process is reflected in the presence of the word-length effect, whereby the capacity for immediate m.emory is inversely related to the time it takes to say the relevant words [4]. A recent study by Longoni, Richardson and Aiello 151 has combined these two variables in a single experiment and demonstrated that they appear to operate independently. This study, which involved experiments in both Italian and English, provides further evidence for the presence of separate storage and articulatory-rehearsal components of the phonological loop. Previous evidence obtained from patients with impaired phonological working memories suggested the involvement of the perisylvian region of the left hemisphere, but left unanswered the question of whether the two subsystems are anatomically separable [6]. A recent study by Paulesu, Frith and Frackowiak [7] used positron emission tomography (PET) scanning to investigate the anatomical localization of the phonological component of working memory in normal …