Mild early-life stress exaggerates the impact of acute stress on corticolimbic resting-state functional connectivity

Abstract

Abundant evidence shows that early-life stress (ELS) predisposes for the development of stress-related psychopathology when exposed to stressors later in life, but the underlying mechanisms remain unclear. To study predisposing effects of mild ELS on stress sensitivity, we examined in a healthy human population the impact of a history of ELS on acute stress-related changes in corticolimbic circuits involved in emotional processing (i.e., amygdala, hippocampus and ventromedial prefrontal cortex [vmPFC]). Healthy young male participants (n = 120) underwent resting-state functional magnetic resonance imaging (fMRI) in two separate sessions (stress induction vs. control). The Childhood Trauma Questionnaire (CTQ) was administered to index self-reported ELS, and stress induction was verified using salivary cortisol, blood pressure, heart rate and subjective affect. Our findings show that self-reported ELS was negatively associated with baseline cortisol, but not with the acute stress-induced cortisol response. Critically, individuals with more self-reported ELS exhibited an exaggerated reduction of functional connectivity in corticolimbic circuits under acute stress. A mediation analysis showed that the association between ELS and stress-induced changes in amygdala-hippocampal connectivity became stronger when controlling for basal cortisol. Our findings show, in a healthy sample, that the effects of mild ELS on functioning of corticolimbic circuits only become apparent when exposed to an acute stressor and may be buffered by adaptations in hypothalamic-pituitary-adrenal axis function. Overall, our findings might reveal a potential mechanism whereby even mild ELS might confer vulnerability to exposure to stressors later in adulthood.

Keywords: HPA axis; cortisol; early-life stress; functional connectivity; stress response.

FIGURE 1

Overview of the study design. Every participant went through two sessions with a counterbalanced order: A stress session in which stressful movie clips were played several times and a neutral session in which neutral movie clips were played at the same time points as the stressful movie clips in stress session. The onset of the first movie clip was defined as t = 0 min, and this movie clip lasted 10 min. The resting‐state scan lasted 6.5 min. Different measures were obtained to verify stress responses at three different time points: Before the session at t = −15 min, after the first task at t = 20 min and before the last task at t = 56 min. Heart rate was recorded throughout resting‐state scanning. BP, blood pressure; DTI, diffusion tensor imaging; PANAS, Positive and Negative Affect Schedule

FIGURE 2

Distribution of CTQ scores. The CTQ scores of 115 participants range from 25 to 56 (M = 32.79, SD = 6.321). CTQ, Childhood Trauma Questionnaire

FIGURE 3

The correlation between basal cortisol (nmol/L) at early afternoon and total CTQ scores across participants. *Bootstrapped 95% confidence interval does not cross zero. CTQ, Childhood Trauma Questionnaire

FIGURE 4

Success of stress induction was verified by different measures at t = 20 min, whereas heart‐rate frequency and variability showed differences (stress − neutral) between two sessions during resting‐state scanning (t = 49 min − t = 56 min). Salivary cortisol, overall BP, self‐reported negative affect and overall heart‐rate frequency were increased due to the stress manipulation, whereas heart‐rate variability was reduced. Error bars represent mean ± SE. ***p < .001; **p < .01; *p < .05. AA, alpha‐amylase; BP, blood pressure

FIGURE 5

Interactive effects of early‐life stress and acute stress on resting‐state functional connectivity. The three panels show negative correlations between total CTQ scores and stress‐induced changes in amygdala–hippocampal connectivity (a), amygdala–vmPFC connectivity (b) and hippocampus–vmPFC connectivity (c). *Bootstrapped 95% confidence interval does not cross zero. a.u., arbitrary units; CTQ, Childhood Trauma Questionnaire; FC, functional connectivity; vmPFC, ventromedial prefrontal cortex

FIGURE 6

Main effects of acute stress on resting‐state functional connectivity. The three panels show connectivity between amygdala and hippocampus (a), between amygdala and vmPFC (b) and between hippocampus and vmPFC (c). Error bars represent mean ± SE. a.u., arbitrary units; vmPFC, ventromedial prefrontal cortex

FIGURE 7

Visualization of mediation model. Path a is the association between CTQ scores and basal cortisol. Path b is the association between basal cortisol and differential amygdala–hippocampus connectivity. Path c is the association between CTQ scores and differential amygdala–hippocampus connectivity. Path c′ shows the direct effect of CTQ scores on differential amygdala–hippocampus connectivity. *Bootstrapped 95% confidence interval does not cross zero. CTQ, Childhood Trauma Questionnaire