Clinical assessment

We used a structured clinical assessment to record the presence or absence of symptoms and signs typically seen in FTLD syndromes, including all clinical features in the current consensus diagnostic criteria (Supplementary material) (Rascovsky et al., 2007; Bensimon et al., 2009; Gorno-Tempini et al., 2011; Armstrong et al., 2013; Höglinger et al., 2017). Each patient’s primary diagnosis was made according to these criteria, with reference to the dominant features at the time of presentation and assessment. Patients with a mixed PPA, who met the diagnostic criteria for PPA but not one of the three subtypes (Gorno-Tempini et al., 2011) were grouped with lvPPA for this study, in view of the low numbers and the association of both phenotypes with Alzheimer’s pathology (Sajjadi et al., 2012). For patients who met several sub-diagnostic criteria we grouped ‘probable’ and ‘possible’ diagnoses together, and classified by the dominant phenotype or formal MAX rules where available (Grimm et al., 2019). We reapplied the other diagnostic criteria to each patient to assess if he or she met the diagnostic criteria for any of the other FTLD syndromes (excepting the ‘mutual exclusivity’ clause included in several criteria). Patients completed the Addenbrooke’s Cognitive Examination-Revised (ACE-R) wherever possible (Mioshi et al., 2006) and a carer’s assessment was obtained using the Cambridge Behavioural Inventory (CBI-R) (Wear et al., 2008). At the time of writing, 49 participants have undergone post-mortem examination, via the Cambridge Brain Bank.