Abstract
<title>Abstract</title> <p>Psychosocial stress is a central risk factor for adolescent depression and anxiety. Yet the neuroendocrine and neural responses to stress have mostly been studied separately, and whether their relationship is altered in depression remains unclear. We examined 92 adolescents (58 with major depressive disorder, with or without comorbid anxiety, and 34 healthy controls) who completed the Montreal Imaging Stress Task during functional MRI, with resting-state scans before and after stress and repeated collection of salivary cortisol and oxytocin. Patients reported greater subjective stress, yet showed a blunted cortisol response relative to controls. Before stress, patients differed in resting-state functional connectivity of fronto-limbic seeds, mainly in operculo-insular and cingulate cortex. From before to after stress, patients showed increased connectivity of the amygdala, hippocampus, and insula with a right insula and frontal opercular region, alongside reduced amygdala-ventromedial prefrontal cortex (vmPFC) coupling. Oxytocin reactivity did not differ between groups; nonetheless, across all participants, an increased oxytocin response was associated with a greater stress-induced increase in vmPFC-dorsolateral prefrontal cortex connectivity, whereas the cortisol response was not. These results reveal a double dissociation. Cortisol reactivity was related to the clinical group but not connectivity, whereas the oxytocin response was related to connectivity but not clinical group. The signal that identified depressive disorder was distinct from the one linked to the brain's stress response. These findings support dimensional, transdiagnostic models of adolescent depression, and highlight the endogenous oxytocin response as a group-independent correlate of the brain's stress response and a potential target for future intervention research.</p>