Working memory impairment after traumatic brain injury (TBI) is common and can lead to long-term disability. Predicting which patients are likely to develop impairments and providing effective treatments is currently challenging. This is in part due to the heterogeneity in damage acquired across the population.

Diffusion tensor imaging (DTI) provides a means to quantify white matter damage associated with working memory impairment after TBI. I therefore investigated whether the development of a diagnostic pipeline using DTI could sensitively detect axonal injury in individual patients and provide insight into the impairments observed. Next, I applied a graph theoretical approach to determine whether distinct patterns of white matter damage were associated with working memory impairment and whether these measures could predict impairment in subacute TBI. Finally, I examined whether the application of machine learning to neuroimaging data could predict the effect of methylphenidate on working memory in patients and whether multimodal imaging improved prediction models.

I found that almost a third of patients with no visible MRI damage had evidence of axonal injury when using my diagnostic pipeline and that patients diagnosed with impairment had significantly poorer cognitive and functional outcomes. I found that patterns of white matter damage associated with alterations to the structural topology of the working memory network were related to working memory impairments after TBI and that network measures could accurately predict working memory outcomes in a novel, subacute TBI population. Finally, I demonstrated that the prediction of treatment effects on working memory performance could be achieved, but only when multimodal imaging was used.

These findings provide mechanistic insights into the development of working memory impairment after TBI and a diagnostic pipeline that can account for heterogeneity in the population. This work also identifies clinically useful prognostic markers, and provides evidence to support the development of more tailored treatments after TBI.