This item is under embargo and not available online per the author's request. For access information, please visit http://libanswers.wustl.edu/faq/5640.
Date of Award
Doctor of Philosophy (PhD)
The control of attention over salient yet irrelevant information is a critical component of goal-directed behavior. Compared to younger adults, older adults often produce larger interference effects in tasks which tap selective attention, a deficit that has typically been viewed as reflecting an age-related decline in attentional control processes. Interference in distinct, attentionally demanding tasks has produced different influences on the characteristics of underlying response time distributions leading to the assumption of different control mechanisms operating across various paradigms. More recently, accumulating research has shown that the magnitude of observed interference can be critically modulated by the congruency of the immediately preceding trial, a phenomenon known as the congruency sequence effect (CSE). This effect is thought to reflect the online modulation of control processes across adjacent trials. Three specific questions regarding the nature of cross-trial effects in attentional control tasks were addressed across three different experiments. First, what stimuli characteristics drive cross-trial changes in attentional control? Second, what component of cognitive processing is responsible for such changes? Third, how do these processes change as a function of healthy aging?
To address these questions, Experiment 1 tested younger and older adults on versions of three attention tasks, Stroop, Simon and flanker, which were designed to eliminate common confounds such as the exact repetition of stimuli across adjacent trials. The results revealed a significant CSE was present in all tasks which was further moderated by age in two. Specifically, the CSE was larger for older adults compared to younger in the Stroop task, but smaller in the Simon task (and trending smaller in flanker). Furthermore, examination of the response time distributions indicated the CSE in Stroop was due to changes in distributional skewing whereas changes in Simon and flanker were better characterized by a simple shift.
Experiment 2 was modeled closely after Experiment 1 with the addition of a response deadline procedure. This was implemented to increase the error rates in each task to better allow for the application of computational modeling (i.e., using the diffusion model). The procedure was successful in speeding up responses equally for each age group and also produced more errors, especially for younger adults. Examination of diffusion model parameters showed intriguing task dissociations such that the previous trial influenced both the non-decision time and drift rate parameters of the diffusion model in Simon but influenced primarily drift rate in both Stroop and flanker. These parameter changes were largely similar across younger and older adults.
Experiment 3 was designed to further probe the characteristics of stimuli that prompt cross-trial changes in control. Stroop stimuli were utilized that were either mostly congruent at the item level (i.e., the word appeared in the corresponding color most of the time), mostly incongruent (the word appeared in a non-corresponding color most of the time), or neutral (items were 50% congruent), while keeping the overall list-wide congruency constant at 50%. This manipulation (the item-specific proportion congruency effect: ISPC) has well-established effects on overall interference such that effects are smaller for the mostly incongruent items. This experiment assessed whether the differential change in conflict would produce CSEs of varying magnitude. This hypothesis was not supported as evidence was obtained for both a robust CSE and ISPC which did not interact in either younger or older adults.
These results support the conclusion that the CSE reflects a very locally driven mechanism which responds to any degree of response conflict regardless of its overall magnitude. This adjustment process appears to decline with age as evidenced by the lower CSEs for older adults in the Simon and flanker tasks. In contrast, in the Stroop task it is hypothesized that a different cross-trial process may be operating, namely the priming of irrelevant processing pathways which is only engaged when relevant and irrelevant dimensions are perceived as an integrated whole.
Chair and Committee
David A. Balota
Todd Braver, Julie Bugg, Janet Duchek, Steven Petersen,
Aschenbrenner, Andrew Jacob, "Dynamic Adjustments of Cognitive Control in Healthy Aging: A Diffusion Model Analysis" (2016). Arts & Sciences Electronic Theses and Dissertations. 804.