Exploring the hyperconnectivity hypothesis in task-based and rest-based data in moderate to severe traumatic brain injury

Restricted (Penn State Only)
Bernier, Rachel Anne
Graduate Program:
Master of Science
Document Type:
Master Thesis
Date of Defense:
May 12, 2016
Committee Members:
  • Frank Hillary, Thesis Advisor
  • Nancy Anne Coulter Dennis, Committee Member
  • Peter Andrew Arnett, Committee Member
  • Stephen Jeffrey Wilson, Committee Member
  • Traumatic brain injury
  • functional connectivity
  • hyperconnectivity
  • fMRI
  • graph theory
  • network theory
Objective: Hyperconnectivity has been reported following traumatic brain injury (TBI) using functional magnetic resonance imaging (fMRI) techniques. This study sought to understand how disrupted systems adapt to injury both during resting and goal-directed brain states by establishing whether individuals with TBI show increased connectivity compared to healthy controls (HCs), whether the increased connectivity is occurring through dedifferentiation of networks, and finally, how differences in connectivity compared to healthy controls relate to cognitive functioning. Participants and Methods: Graph theory was implemented to examine functional connectivity in 18 individuals with moderate to severe TBI and 18 HCs. Adjacency matrices were created and the weights of surviving edges were compared both within group during rest and task and between groups. The relations between these metrics were then compared with performance on cognitive measures. Results: TBI subjects showed no differences in weighted degree between rest and task, in contrast to HCs who demonstrated increased weighted degree during task compared to rest. Individuals with TBI were hyperconnected during rest compared to HCs, and increased connectivity of nodes involving task-related regions of interest (ROIs) accounted for much of the hyperconnectivity observed during rest in individuals with TBI compared to HCs. There were no significant differences in connectivity during task at the network level in TBI subjects compared to HCs. Increased connectivity during rest between task-related ROIs was associated with worse performance on a measure of set-shifting during task, whereas increased connectivity with nodes of the DMN, typically not associated with task, was associated with better performance on Digit Span Backward. Conclusion: The relation between increased connectivity and cognitive functioning depended on what state and between which networks the hyperconnectivity occurred. Therefore, examining hyperconnectivity at a global level obscures specific relations between hyperconnectivity and cognitive function. Continuing to explore connectivity changes during recovery at the network level during specific cognitive states will elucidate specific connectivity profiles used to better understand and predict recovery following injury.