Predicting change in depressive symptoms using longitudinal regional amyloid and cognition in cognitively unimpaired older adults

Abstract

AbstractBackgroundDepression is a neuropsychiatric symptom of Alzheimer’s disease (AD) and has been linked to greater cortical amyloid burden in preclinical AD. We sought to examine whether amyloid accumulation in brain regions subserving emotional control and changes in cognitive performance may contribute to increased depressive symptomatology over time.Method155 cognitively unimpaired participants who were below the cortical amyloid (<0.86DVR) and clinical depression thresholds (<12 on Geriatric Depression Scale (GDS); mean age = 72.6) at baseline were included from the Harvard Aging Brain Study, with annual GDS and Preclinical Alzheimer Cognitive Composite (PACC) assessment and MRI/PiB‐PET every three years (mean years follow‐up = 7.06). PiB slopes were calculated for FreeSurfer‐defined bilateral regions of interest (ROI) implicated in emotional control: amygdala, medial and lateral orbitofrontal cortices (mOFC, lOFC), superior and middle frontal (MFC) cortices, anterior cingulate (ACC), posterior cingulate (PCC), and isthmus cingulate (IC). Linear mixed‐effects models assessed whether main‐effects of ROI PiB slope and PACC slope predicted longitudinal GDS scores for each ROI, covarying for age, sex, education, and random intercept/slope, adjusted for multiple comparisons. Post‐hoc linear regression models assessed relationships between PiB and PACC slopes.ResultSteeper PiB slopes in the mOFC, MFC, ACC, and IC were associated with increasing GDS scores over time, while a decreasing PACC slope was predictive of increasing GDS scores over time in all regions (Table 1). Post‐hoc analyses indicated that PiB slopes were not significantly correlated with PACC slope (Figure 1).ConclusionIn a cohort of cognitively unimpaired older adults with low cortical amyloid and no/subclinical depressive symptoms at baseline, increasing depressive symptoms over time were significantly associated with both amyloid accumulation in specific regions associated with emotional control (i.e., mOFC, MFC, ACC, and IC) and worsening cognitive performance. Furthermore, changes in cognition were not significantly associated with changes in amyloid accumulation in regions involved in emotional control, suggesting that different factors may be independently influencing trajectories in depression versus cognition over time. These results shed light on the neurobiology of depression in older individuals and underscore the importance of monitoring new and increasing neuropsychiatric symptoms in addition to cognitive symptoms when screening for AD.