Changes to white matter and cognition following adult traumatic brain injury

Abstract

The cognitive impairments that frequently occur following a traumatic brain injury (TBI) are thought to be primarily caused by diffuse white matter (WM) injury. Computed tomography and magnetic resonance imaging have traditionally been used to determine the location and extent of this WM damage, however they better detect macroscopic damage, while underestimating the amount of microstructural damage to WM. Diffusion tensor imaging (DTI), on the other hand, better detects microstructural WM changes by measuring the diffusion of water molecules. DTI has been used to examine WM changes following TBI, however findings are mixed, with the location and magnitude of the changes varying between studies. Similarly, disparate findings concerning the relationship between DTI findings and cognitive outcomes have been reported. Finally, the traditional methods used to analyse DTI data are limited in areas of the brain that contain more than one WM tract. The overarching aim of this thesis was therefore to examine WM changes and cognitive outcomes following adult TBI. Four studies were completed to address these aims. Specifically, two meta-analyses were performed to synthesise the findings from studies that 1) used DTI to examine the location and extent of WM changes following adult TBI; and, 2) examined the relationship between DTI findings and cognitive outcomes following adult TBI. The third study aimed to determine whether the findings from the meta-analyses — which were primarily based on small samples (most studies had fewer than 30 participants) — were replicated in large TBI and control samples. Finally, the same diffusion data were analysed using a recently developed method of analysis, known as fixel-based analysis (FBA), in order to determine whether it detected micro- and macro-structural WM changes in individual WM tracts following TBI. The first study (Chapter 3) meta-analysed the findings from 44 studies that used DTI to examine adult TBI to determine the location and extent of WM changes. The findings indicated that WM changes, reflected in lower fractional anisotropy (FA) and higher mean diffusivity (MD),were evident in a very large number of brain regions following both mild and moderate to severe TBI, with more severe injuries leading to more prominent WM changes. These regions included the corpus callosum, fornix, superior longitudinal fasciculus, internal capsule, occipital white matter, centrum semiovale, and thalamic radiations. The second study (Chapter 4) meta-analysed 20 studies that examined the relationship between DTI findings and cognition following adult TBI. It was found that lower FA and higher MD from a number of brain regions were related to poorer cognitive functioning, particularly in the domains of memory and attention. These regions were similar to those that were identified in the first meta-analysis and included the corpus callosum, fornix and superior longitudinal fasciculus. However, most findings were based on single studies with relatively small samples (60% of studies from the two meta-analyses had fewer than 30 participants), limiting the conclusions that could be drawn. Thus, in the third study (Chapter 5), a large sample of adults with mild to severe TBIs (N=165) and a healthy and orthopaedic control group (N=106) underwent DTI and cognitive testing. Based on the findings from the meta-analyses, FA and MD were calculated using a region of interest approach for the corpus callosum (genu, body, splenium), fornix and superior longitudinal fasciculus and participants completed tests of memory, attention and executive functioning. Although mild TBI was not associated with significant WM or cognitive changes, people with moderate to severe TBI displayed large WM alterations (all regions) and poorer cognitive performance. The final study (Chapter 6) analysed the same diffusion data examined in Study 3 using a novel method of analysis known as FBA. This emerging methodology is designed to overcome the main limitation of traditional DTI methods of analysis: that these methods are inaccurate in regions containing crossing fibres. Again, the mild group did not show evidence of WM changes, but the moderate to severe group displayed considerable changes in widespread WM tracts, reflecting fewer axons (reduced fibre density) and a reduction in cross-sectional area. Similarly,the mild group did not show slower reaction times, but the moderate to severe group did, although the fixel findings were not associated with reaction times/processing speed. This thesis showed that WM changes are widespread following moderate to severe TBI and can be detected using DTI and FBA. Findings following mild TBI, however, are less clear and warrant further research. TBI is a complex and multifaceted injury that does not have a typical pattern of damage that is readily captured using a single neuroimaging analysis technique. Research is now needed to determine whether DTI and/or FBA can predict long-term cognitive outcomes.