Cognitive variability, brain aging, and cognitive decline in late-life major depression.

OBJECTIVES: Older adults with late-life major depression (LLMD) are at increased risk of dementia. Dispersion, or within-person performance variability across cognitive tests, is a potential marker of cognitive decline. This study examined group differences in dispersion between LLMD and nondepressed healthy controls (HC) and investigated whether dispersion was a predictor of cognitive performance 1 year later in LLMD. We also explored demographic, clinical, and structural imaging correlates of dispersion in LLMD and HC. We hypothesized that dispersion would be greater in LLMD compared with HC and would be associated with worse cognitive performance 1 year later in LLMD. DESIGN: Participants were enrolled in the Neurobiology of Late-Life Depression, a naturalistic longitudinal investigation of the predictors of poor illness course in LLMD. PARTICIPANTS: The baseline sample consisted of 121 older adults with LLMD and 39 HC; of these subjects, 94 LLMD and 35 HC underwent magnetic resonance imaging (MRI). One-year cognitive data were available for 107 LLMD patients. MEASUREMENTS: All participants underwent detailed clinical and structural MRI at baseline. LLMD participants also completed a comprehensive cognitive evaluation 1 year later. RESULTS: Higher test dispersion was evident in LLMD when compared with nondepressed controls. Greater baseline dispersion predicted 1-year cognitive decline in LLMD patients even when controlling for baseline cognitive functioning and demographic and clinical confounders. Dispersion was correlated with white matter lesions in LLMD but not HC. Dispersion was also correlated with anxiety in both LLMD and HC. CONCLUSIONS: Dispersion is a marker of neurocognitive integrity that requires further exploration in LLMD.

Duke Scholars

Author Lihong Wang Psychiatry & Behavioral Sciences, Adult Psychiatry & Psychol ...

Author David Carl Steffens Psychiatry & Behavioral Sciences, Adult Psychiatry & Psychol ...