Depression and aging: insights from brain age prediction models

The impact of depression on brain aging remains unclear, but both have been linked to stressful life events. Shared biological pathways may underlie structural brain changes. Clarifying these relationships could advance understanding of underlying mechanisms and inform treatment approaches.

Structural MRI scans of 190 participants (controls, n = 110, clinically diagnosed with major depressive disorder [MDD], n = 80), from the REDEEM dataset, were input into three pretrained brain age prediction models: brainageR, DeepBrainNet, and pyment. Prediction accuracy was compared in controls to identify the optimal model. DeepBrainNet demonstrated the highest accuracy and was selected for subsequent analysis. Brain-predicted age difference (brain-PAD) was calculated as predicted age minus chronological age. Linear regression examined the effects of MDD diagnosis, childhood maltreatment, and cortisol awakening response on brain-PAD.

Depressed participants reported greater childhood maltreatment but a similar cortisol awakening response. An Age × Group interaction (β = 0.34, 95% CI: 0.15-0.53, p < 0.001) indicated older adults with MDD exhibited greater positive deviations from normative brain age predictions, suggesting nonuniform brain aging across the lifespan. Cortisol awakening response showed a negative association with brain-PAD (β = -0.01, 95% CI: -0.01 to -0.00, p = 0.041), indicating higher HPA-axis reactivity was linked to younger-appearing brains. Females showed lower brain-PAD than males, reflecting younger-appearing brains.

MDD was associated with age-dependent differences in brain-PAD. The protective association between cortisol awakening response and brain age highlights the importance of integrating stress biomarkers to better understand neural aging mechanisms in depression.