Improving the diagnostic accuracy of primary progressive aphasia using cognitive tests

Primary progressive aphasias (PPA) are rare younger-onset dementias that primarily affect speech and language functions. Three subtypes of PPA have been identified and are classified based on the presence or absence of core language, imaging, and clinical findings: logopenic (lv-PPA), non-fluent (nfv-PPA) and semantic variant (sv-PPA). Differentiating these variants is important for treatment and planning; however, assessment of the nfv-PPA and lv-PPA variants’ speech and language is often difficult for clinicians with limited expertise. As such, simple language, cognitive, and behavioural tests are needed to aid diagnosis. This thesis comprises five studies that investigates this issue. Chapter 3 showed that, relative to the other PPA variants, lv-PPA is particularly impaired on a simple measure of visuospatial short-term memory, and that this impairment relates to the grey matter intensity decrease of the parietal lobe and surrounding regions. Chapter 4 expanded on these findings using a novel, computerised visuospatial working memory test. Using the cognitive subdomain scores of a cognitive screening tool (ACE-III), Chapter 5 found that the cognitive profiles of the PPA variants were distinct at baseline visit and over time. Crucially, cognitive decline in lv-PPA was more widespread and rapid than in nfv-PPA and sv-PPA. Chapter 6 demonstrated that PPA profiling is enhanced when evaluating the individual items of the ACE-III, and an interactive online ACE-III PPA diagnostic calculator was built for clinicians. Finally, Chapter 7 established that the mood, behaviour, and everyday skills profiles of the PPA variants are distinct at initial presentation and over the disease course. The work within this thesis demonstrates that the evolving cognitive and behavioural profiles of the PPA variants are distinct and extend beyond the primary disorder of language. Accordingly, assessing these skills in clinical practice can improve the diagnostic accuracy of the PPA variants