Taking a computational approach to aging: the SPAN theory of working memory.

The decline of working memory capacity associated with normal adult aging is well-known. What is less well established is the cause of this decline. One prominent proposal is that working memory decline is caused by a reduction in basic information-processing speed, but this account has lacked a demonstration that general slowing is computationally sufficient to produce a decrease in working memory capacity. This article presents a production system theory of working memory (SPAN) based on established mechanisms: slowing, decay, and displacement. Models of 2 tasks--digit symbol and computation span--which have been prominent in research on slowing, are presented in detail. These models demonstrate that slowing is sufficient to produce differences in these tasks, and they provide a quantitative match to observed young-old differences as well. This advance for slowing theory also demonstrates the viability of computational tools in aging research.