Potential vulnerability and resilience to accelerated brain aging in women exposed to stressful life events: insights from the brain age prediction model.

Brain age prediction models consistently reveal accelerated brain aging in psychiatric disorders, yet associations with stress, independent of formal psychiatric diagnoses, remain uncertain. This study investigated the relationships of emotional and alcohol-use symptoms, common and often comorbid stress-related symptoms, and resilience with brain aging using high-resolution structural MRI data from 520 women who experienced stressful life events. Participants were divided into four groups based on the presence of emotional and alcohol-use symptoms: no symptoms (n = 287), emotional symptoms only (n = 93), alcohol-use symptoms only (n = 79), or both symptoms (n = 61). Individual brain age gap (BAG)-the difference between predicted brain age and chronological age-was calculated using a deep learning-based brain age prediction model. Individual and interactive associations of the presence of two symptoms with BAG were assessed using two-way ANCOVA. Relationships of a continuous composite symptom score integrating both symptoms and resilience with BAG were evaluated. Participants with both symptoms exhibited significantly larger BAG than the other groups, with a statistically significant interaction between two symptom domains (p = 0.017). Across the full sample, composite symptom scores were positively associated with BAG (β = 0.16, p = 0.004), with an even stronger association within individuals with both symptoms (β = 0.34, p < 0.001). Conversely, higher resilience was linked to smaller BAG across all participants (β = -0.10, p = 0.046). The negative association between resilience and BAG was statistically mediated by the composite symptom score (b = -0.011, p = 0.010). These findings may suggest a synergistic, more-than-additive association between stress-related symptoms and accelerated brain aging, as well as a potentially buffering association of resilience.