Executive-process interactive control: A unified computational theory for answering 20 questions (and more) about cognitive ageing

Abstract

Although the eVects of ageing on human information processing and performance have been studied extensively, many fundamental questions about cognitive ageing remain to be answered de nitively. For example, what are the sources of age-related slowing? How much is working-memory capacity reduced in older adults? Is time-sharing ability lost with age? Answering such questions requires a uni ed computational theory that characterises the interactive operations of many component mental processes and integrates diverse data on cognitive ageing. Toward ful lling this requirement, an executive-process interactive control (EPIC) architecture has been extended to model performance of both young and older adults. EPIC models yield accurate accounts of ageing eVects on reaction times and accuracy in basic dual-task and working-memory paradigms. From these accounts, it appears that time-sharing ability and working-memory capacity decrease relatively little until after 70 years of age. Before age 70, at least some apparent performance decrements may be attributable to conservative executive processes and ineYcient task procedures rather than decreased ''hardware'' functionality. By clarifying and deepening such insights, uni ed computational theories like EPIC will help answer many questions about cognitive ageing.

Adult Age Differences in the Implicit and Explicit Components of Top-Down Attentional Guidance During Visual Search.

“…Thus, both target activation and distractor inhibition processes were available for explicit top-down guidance, and it remains to be determined whether both types of processes contributed to the type of age-related change we have proposed in the relative influence of explicit and implicit top-down attentional guidance. However, the age difference in Experiment 2 as well as the age constancy in the repetition priming effects in both experiments caution against viewing a deficit in executive control as a general theme of age-related cognitive change (Meyer et al, 2001;West, 1996). Demonstrating an age-related decline in executive control may require a critical level of task complexity or type of cognitive demand, such as detecting subtle regularities in a sequence of events (Howard et al, 2004) or time sharing between different tasks (Salthouse & Miles, 2002).…”

“…Executive control has been a prominent theme in explanations of age-related deficits in many tasks (Meyer, Glass, Mueller, Seymour, & Kieras, 2001;West, 1996), although distinguishing a deficit in executive control from a more basic mechanism such as perceptual speed is difficult. For example, the performance costs associated with the executive process of switching between different tasks (i.e., global task switching) are higher for older adults than for younger adults, beyond what would be expected on the basis of generalized age-related slowing, whereas the costs of selecting or inhibiting specific features within a task (local task switching) are more consistent with a generalized slowing model (Mayr & Liebscher, 2001;Salthouse & Miles, 2002;Verhaeghen & Cerella, 2002; but cf.…”

“…If executive control is a fundamental dimension of age-related cognitive decline (Meyer et al, 2001;West, 1996), then the explicit component of top-down attentional guidance, in particular, should be more pronounced for younger adults than for older adults. In other words, the Maljkovic and Nakayama (1994) and Kristjánsson et al (2002) proposal that repetition priming accounts entirely for the top-down effect is more likely to characterize the performance of older adults than that of younger adults.…”

The neural bases of the effects of item-nonspecific proactive interference in working memory

“…A finding that such trends were also evident in the delay epoch, in contrast, would be interpreted as evidence that the buildup of item-nonspecific PI engages a mechanism that is active during many phases of the working memory task. In the event of the latter outcome, the mechanism of PI resolution might be either tonically active or phasically activated by any processing of an interfering stimulus, whether it be during encoding, rehearsal (e.g., Meyer, Glass, Mueller, Seymour, & Kieras, 2001), or probe evaluation.…”

“…A finding that such trends were also evident in the delay epoch, in contrast, would be interpreted as evidence that the buildup of item-nonspecific PI engages a mechanism that is active during many phases of the working memory task. In the event of the latter outcome, the mechanism of PI resolution might be either tonically active or phasically activated by any processing of an interfering stimulus, whether it be during encoding, rehearsal (e.g., Meyer, Glass, Mueller, Seymour, & Kieras, 2001), or probe evaluation.The independent variables entered into the modified GLM (Worsley & Friston, 1995;Zarahn, Aguirre, & D'Esposito, 1997a) coded the data by epoch (i.e., target, delay, or probe) and by trial position within the block (i.e., 1-12). Low-frequency drift was accounted for in the design matrix (i.e., detrending) with scan effect covariates.…”