Abstract Though depression impacts over 10% of the oldest-old, the contribution of brain health on changes in depression is unknown. This study included 225 participants from LifeAfter90 with magnetic resonance imaging, amyloid positron emission tomography (PET), and repeated Geriatric Depression Scale (GDS) data. We fit linear mixed-effects models for associations of neuroimaging markers with depressive symptom change adjusted for age, gender, race and ethnicity, and education. At baseline, average GDS scores were 2.6 (SD = 2.3) out of 15 possible points, average age was 93.3 (SD = 2.3) years, 56% female, and 22% Black, 25% Asian, 18% Hispanic/Latino, 28% Non-Hispanic White, and 7% as Multiracial/other. Greater baseline occipital volume was associated with slower GDS increase (β = − 0.29; 95% CI − 0.54, − 0.05). Greater baseline free water was associated with faster GDS increase (β = 0.31; 95% CI 0.05, 0.57). Estimates for other neuroimaging biomarkers leaned towards slower, but non-significant, increases in GDS. Estimates were not in the expected direction for amyloid PET and null for frontal cortex volume. Our findings suggest that lower cortical volume and greater white matter injury is associated with increases in depressive symptoms among the oldest old, even for those with low levels of baseline depressive symptoms.
Posis et al. (Sat,) studied this question.