An Integrated Model of Set Shifting and Maintenance

Serial attention involves adopting a task set (or intention to do a task), maintaining that set for a time, then shifting set. When set must be shifted often, serial attention is difficult because old sets cannot be forgotten instantly and will interfere with the current one. This memory perspective on set shifting is the basis for a computational cognitive model that explains a variety of effects in serial attention, some well-documented (e.g., RT switch cost) and some novel (e.g., within-run slowing). The mechanisms producing these effects are tightly integrated with each other and with memory theory. The model absorbs extant accounts of set shifting, including failure to prepare (De Jong, Berendsen, & Cools, 1999) and exogenous cueing (Rogers & Monsell, 1995), and promises to unify the effects presented here with other set-shifting phenomena.