UC Berkeley

Executive control comprises a set of neural processes that are critical for goal-directed behaviors, such as attention and working memory. Such complex behaviors likely rely on the flexible communication within and between brain sub-networks, or modules. In this thesis, I present several projects that use graph theoretical methods to describe the brain as a complex network and examine the role of brain network organization in supporting aspects of executive control processing. Across these projects, I use a variety of methods, such as functional magnetic resonance imaging (fMRI), functional connectivity, and cognitive training.

The first chapter of this thesis examines the influence of attention demands on brain network organization. Here, I show that brain network modules become more integrated during the processing of relevant stimuli compared with the processing of irrelevant distractors during a working memory N-back task. The strength of this reconfiguration is related to faster task performance, suggesting its importance in executive control.

The second chapter of this thesis examines the influence of aging on reconfiguration of brain networks due to executive control (i.e., N-back load) demands. Here, I demonstrate that older adults exhibit changes in brain network organization at lower levels of demand compared with young adults. Further, brain network reconfiguration from a task-free ‘resting-state’ to an N-back task is related to better task performance and greater structural connectivity of a core frontal-posterior white matter tract.

The final two chapters of this thesis examine how brain network organization can predict gains in executive control processing after cognitive interventions. The first project shows that the extent of segregation of brain network modules (i.e., higher network modularity) is predictive of greater training-related cognitive gains in older adults. The second project extends this predictive framework to healthy young adults.

In sum, these projects demonstrate that brain networks flexibly reconfigure depending on executive control demands and that aspects of brain networks are predictive of training-related gains of executive control. These findings provide insight into how properties of large-scale brain networks allow for executive control processing.