Date of Award

Spring 1-1-2010

Document Type


Degree Name

Doctor of Philosophy (PhD)


Psychology & Neuroscience

First Advisor

Yuko Munakata

Second Advisor

Tim Curran

Third Advisor

Eliana Colunga


We are constantly bombarded with myriad pieces of information, of which only a portion is directly relevant to our immediate experiences. What determines how much irrelevant information we filter, and how quickly we can adjust the amount of filtering? This dissertation explores the possibility that working memory (WM), an ability to actively maintain task-relevant information, plays a critical role in dynamically adjusting filtering strategy, based on task demands. High filtering could result from upregulating top-down control, leading to processing only the task-relevant information. Low filtering could result from loosening this control, to allow for a larger amount of information to be processed. This account was tested in three different experiments, with both adults and six-year-old children. Experiment 1 showed that high WM could support both high and low filtering, within the same adult participants. High WM was associated with high filtering in a paradigm where filtering task-irrelevant information was advantageous, because filtering the distractors reduced WM demand. In contrast, high WM was associated with low filtering in a task-switching paradigm, where low filtering was advantageous, because currently irrelevant features became relevant on subsequent switch trials. Experiment 2 modified the high-filtering-demand task of Experiment 1 to have both high- and low-filtering-demand versions, in order to manipulate filtering demand within the same paradigm. Results were difficult to interpret definitively due to participants' poor compliance with the stated instructions; nonetheless, they point to sensitivity of early attention to subtle changes in experimental setup. Experiment 3 tested filtering across two additional tasks in both adults and children, in order to assess the role of working memory in filtering in more robust paradigms, to test for dynamic changes in filtering strategy within the context of the same task, and to test the developmental origins of the relationship between WM and filtering. Results from both children and adults were mostly consistent with the dynamic filtering theory, but several important caveats are discussed. Despite these shortcomings, results from these experiments provide an important advance in understanding the role of WM in dynamically adjusting filtering strategies based on task demands.