BRAIN AGE ESTIMATION IN LATE-LIFE DEPRESSION: ASSOCIATION WITH COGNITIVE PERFORMANCE AND DISABILITY

Abstract

Introduction Mental health disorders are associated with several biological markers of accelerated aging. Recent advances in machine learning can provide an estimation of brain age that may differ from actual, chronological age in the setting of neuropsychiatric disease and accelerated aging. We examined whether depressed elders exhibited a calculated brain age that was older than expected when compared with chronological age, and if so whether that difference was related to cognitive performance and disability Methods 154 elders (118 depressed, 36 never-depressed) enrolled in three studies with common entry criteria, undergoing structural brain MRI, a cognitive battery assessing episodic memory, executive function, processing speed, and working memory, and assessment of disability using the WHODAS 2.0. A machine-learning model estimated brain age based on MRI data, which was compared to chronological age to yield a brain-age gap (BAG). (BAG; relative difference between calculated age and chronological age). Analyses examined whether BAG was greater in depressed elders and whether it was associated with cognitive performance and disability. We also examined the depressed cohort alone to test for interactive effects between BAG and depression severity on function. Results After adjustment for covariates, late-life depression was associated with a higher brain-age gap (Wald χ2=8.84, p=0.0029), or a calculated age greater than chronological age. Across all subjects, a higher brain-age gap was associated with poorer episodic memory performance (Wald χ2=4.10, p=0.0430). Among depressed adults, higher BAG was associated with slower processing speed (Wald χ2=4.43, p=0.0354.) We also observed an interaction between BAG and depression severity, where increasing depression severity in context of higher BAG resulted in poorer performance in domains of executive function (Wald χ2=5.89, p=0.0152) and working memory (Wald χ2=4.47, p=0.0346.) Across all subjects, a higher BAG was also associated with greater disability (Wald χ2=6.00, p=0.0143). Conclusions Late-life depression is associated with a higher BAG, indicating brains appear older than expected by chronological age alone. Higher BAGs have cognitive and functional implications and may moderate the effect of depression severity on cognitive performance. This research was funded by: NIH grants R01MH102246, R21MH099218