Working memory in the brain.

concept of working memory is ubiquitous in research concerned with temporary storage and on-line processing in humans and animals. The replacement by Baddeley and Hitch (1974) of the concept of short-term memory focused largely on retention of word lists, with the broader concept of a multiple component working memory, was remarkable in redefining a major area of scientific endeavour. It has been remarkable in stimulating substantial lines of research yielding insight into verbal and visuo-spatial temporary storage through behavioural studies of healthy and brain damaged human adults and children. Recent years have seen genuine progress in the understanding of executive functions, previously consigned to a largely intractable homunculus. It has also been remarkable in its resilience over more than three decades of scientific interrogation, and of robust declarations of its demise papers in this Special Issue were inspired (although not exclusively so) by some of the contemporary research presented at that conference. This followed on from the first such International Conference on Working Memory held in 1994 in Cambridge, UK to mark 20 years since the landmark paper. From the 2004 Kyoto conference, it was clear that the working memory concept has survived very successfully like all powerful ideas because of its relative simplicity and its combination of predictive and explanatory power. However, like the memes of Dawkins (1976), working memory as an idea has evolved into several variations (reviewed in Miyake and Shah, 1999); the original trio of phonological loop, visuo-spatial scratchpad and central executive is now a quartet with the addition of an episodic buffer (Baddeley, 2000) included to account for temporary retention of integrated representations which are multi-sensory and have semantic content. In contrast is the concept of a domain general working memory (e.g., Daneman and Merikle, 1996; Kane and Engle, 2002) with a focus on individual differences in mental ability. Other approaches have developed computational models of overall working memory function (Barnard, 1999; Lovett et al., 1999) or of domain specific functions (review in Burgess and Hitch, 2005). The 2004 Kyoto Conference also reflected the exponential rise in the use of measures of brain activity to track the time course and the neuroanatomical networks associated with working memory function. However, whether functional imaging studies with PET, fMRI, ERP or MEG will help assess the relative utility of the range of working memory theories on offer any more effectively than behavioural studies remains a topic of some …