Large-Scale Network Dynamics of Visual Working Memory

A core set of fronto-parietal brain regions is implicated in a wealth of cognitive functions. Researchers have suggested that working memory, a fundamental element in many higher-order processes, is the underlying mechanism supported by this network. However, the idea that the same fronto-parietal network underlies the qualitatively different tasks employed throughout visual working memory research is contentious. Instead, systems neuroscience has adopted the view that cognition arises out of the dynamic interaction of several large-scale networks. A hierarchical split divides these networks into an extrinsic system, which governs attention to the external environment, and an intrinsic system, which guides internally-directed processes. Given visual working memory involves a two-way connection between perceptual input and internal representations, the current dissertation uses converging methodologies to explore whether tasks that vary in their exogenous and endogenous attentional demands are likely supported by different network dynamics.A quantitative meta-analysis used stress as a paradigm to investigate the differential effects of exogenous and endogenous distraction on visual working memory task performance. This analysis was followed by a controlled stress study that examined whether endogenous distraction, instigated by a psychosocial stressor, differentially influenced visual maintenance versus mental rotation. Finally, an electroencephalographic study was conducted where participants were required to store visual information despite an ongoing external distractor. Taken together, the data presented from these three studies suggest key differences between maintenance and manipulation that may reflect the variable interplay between extrinsic and intrinsic networks. These data support the idea that the maintenance of visual information depends on right-lateralized regions of the dorsal attention network, while mental rotation recruits additional regions of the default mode and central executive networks. This work emphasizes the importance of implementing synergistic protocols while investigating cognitive function, and supports the view that the human brain consists of multiple interacting networks. It also corroborates the idea that the functional connectivity patterns accompanying cognitive state changes can impose processing constraints, and that understanding these constraints can allow us to predict behavioural impairments likely to arise in various circumstances. Similar investigations of large-scale network dynamics can provide a framework for understanding fundamental aspects of cognitive function.