185 Longitudinal tau PET and atrophy and parcellation qualities were manually inspected for gross abnormalities. For cortical thickness we used all cortical gray matter ROIs as available in the DKT atlas. Statistical analysis All analyses were performed for Aβ+ and Aβ- individuals separately. To establish whether imaging markers changed over time, we performed linear mixed effects models (LMMs) with global BPND, R1 or cortical thickness as dependent variable and time as determinant, adjusted for age-at-PET and sex (using both the follow-up and baseline-only sample). To assess differences in demographic variables and BPND, R1 and cortical thickness, between Aβ- and Aβ+ individuals two-sample t-tests were used. Associations between baseline BPND in our three regions of interest (Braak I, Braak III/IV and Braak V/VI) and cortical thickness or R1 over time (in all cortical regions available in the brain templates) were assessed using LMMs with random intercepts and fixed slopes, adjusted for age, sex and time between baseline and follow-up assessments. For these analyses individuals from both the follow-up and baseline-only samples were used (ntotal=147). Next, the associations between annual change in BPND and longitudinal cortical thickness or R1 were assessed using the same LMMs, now with annual change in BPND as predictor and additionally adjusted for baseline BPND. Therefore, annual change for BPND was calculated for all three regions of interest (Braak I, III/IV and V/VI) by subtracting the baseline value from the follow-up value, and dividing by the time between measurements in years. For these analyses only individuals from the follow-up sample were used (n=61). Results are reported both with and without the Benjamini-Hochberg False Discovery Rate (FDR) correction with a Q value of 5%. All statistical analyses were performed in Rstudio v4.0.3 and results are visualized using forest plots of the effect sizes and their respective confidence intervals. The regional associations are displayed using the ggseg R package for FDR-surviving cortical thickness ROIs only (since ggseg does not yet support the atlases used for R1). RESULTS Participants Out of the 61 individuals who underwent follow-up, 35 were Aβ+ and 26 Aβ- at baseline (Table 1, Figure 1). Baseline demographics and characteristics of the total baseline sample (including the baseline-only sample, n=147) are shown in Supplementary Table 1. Relative to the total baseline sample, the follow-up sample consisted of younger patients, with relatively higher MMSE scores and higher levels of global tau PET BPND. In the follow-up sample, Aβ+ individuals were more often 7
RkJQdWJsaXNoZXIy MjY0ODMw