Trauma-predictive brain network connectivity adaptively responds to mild acute stress

Abstract

Past traumatic experiences shape neural responses to future stress, but the mechanisms underlying this dynamic interaction remain unclear. Here, we assessed how trauma-related brain networks respond to current acute stress in real time. Using a machine learning approach, we trained and tested brain functional connectivity networks to predict past trauma exposure in a community sample of adults ( N = 170). We then evaluated the response of these trauma-predictive brain networks to an acute emotional and physiological stressor in a subsample of participants ( N = 92) and to a pharmacological manipulation (hydrocortisone) in an independent crossover study ( N = 27). We found that connectome-based predictive modeling successfully predicted past trauma exposure. The network associated with greater trauma exposure showed high involvement of salience network connections, with model prediction driven by connectivity in the medial frontal cortex, salience network, motor regions, default mode network, and cerebellum. Notably, connectivity in this trauma-predictive network was significantly attenuated following an acute stressor relative to control. A similar pattern was observed in the pharmacology sample, with decreased connectivity under hydrocortisone compared to placebo within this trauma-predictive network. Finally, attenuated trauma-predictive network connectivity after acute stress was associated with lower depressive symptoms in the stress-exposed group, but not control group. These findings suggest that stress may induce blunted connectivity of brain networks related to past traumatic experiences and that the reduced engagement of these trauma-predictive brain networks may facilitate adaptive regulatory response to later stress.