Date of Award

Spring 1-1-2016

Document Type

Dissertation

Degree Name

Doctor of Philosophy (PhD)

First Advisor

Marie T. Banich

Second Advisor

Tim Curran

Third Advisor

McKell Carter

Fourth Advisor

Naomi Friedman

Fifth Advisor

Christine Brennan

Abstract

Executive functions (EFs) are a thoroughly characterized set of high-level cognitive abilities that contribute to a person's ability to pursue, maintain, and modify goals. A popular theoretical model of EFs - the unity and diversity model - describes the variety of laboratory tasks used to measure EF as having shared components (unity) and components specific to subsets of tasks (diversity). A fundamental question in the field of cognitive neuroscience is what brain characteristics underlie the wide array of high-level cognitive skills and abilities, like the unity and diversity aspects of EF, that vary across the population but are stable within an individual. The development of methodologies such as functional magnetic resonance imaging (fMRI) has contributed to a surge in interest in this fundamental question. Resting-state fMRI, in particular, characterizes the stable oscillatory behavior of the blood oxygen level dependent signal when individuals are not directed overtly to perform a particular task. As such, resting-state fMRI is a non-invasive and minimally-demanding protocol offering the potential to make predictions about individual differences in cognitive functioning. We present results from several resting-state analyses of two large data sets (n = 91 and 251). Participants in each sample were scanned without being overtly directed toward a particular mental task set and administered a battery of EF tasks that measured three EF constructs from the unity and diversity model - the general ability to maintain goals and task sets (common EF), abilities related to fluidly shifting between task sets (shifting-specific EF), and the ability to add to or delete from the contents of working memory (updating-specific EF). The resting-state analyses used here quantify the dynamics of the resting brain at two complementary spatial scales - at the level of networks and individual regions. Taking the results of these network and regional connectivity analyses together, we provide a new perspective on the neural bases of EFs that complements our understanding of the mechanisms of cognitive control (as assessed using the computational modeling approach) and the areas implicated in online cognitive control (as assessed by the brain mapping approach via fMRI studies of demanding versus non-demanding conditions). First, individual differences in common EF were associated with the intensity of the frontoparietal network at rest and connectivity among networks of regions implicated in high level cognition (both "task-positive" and "task-negative" networks). Because common EF captures variance in a wide array of cognitive tasks, these features perhaps underlie the unity of the neural basis of EFs. The diversity aspects from the unity and diversity model of EF are also associated with variation in separable neural substrates as assessed by resting-state connectivity. Individual differences in shifting-specific EF were associated with a more spatially distributed set of resting-state connectivity measures than task-based neuroimaging studies may lead one to believe, for example, the spatial extent of a network of regions implicated in top-down biasing and the diversity of connections in sensorimotor cortex. Individual differences in updating-specific EF are not supported by variation in resting-state networks in the way common EF and shifting-specific EF are but rather by variation in a limited set of graph theoretic properties of individual regions. Some results for updating-specific EF are novel, such as a less integrated frontal pole being associated with higher updating-specific EF ability, while others are in agreement with findings from the computation modeling literature which suggests updating depends upon connectivity of dorsolateral prefrontal cortex. Our work builds upon those findings by showing that not only do the mechanisms of working memory updating depend upon a certain type of connectivity, but variation in the degree to

Share

COinS