hide
Free keywords:
-
Abstract:
Introduction:
Acute and chronic stress are important factors in the etiology of affective disorders. Most research has focused on alterations of the HPA-axis response in affective disorders. However, increasing evidence has shown that psychological and physiological stress responses are not necessarily correlated as distinct networks may drive the different stress dimensions. Nonetheless, only few studies have investigated alterations of the neural stress response, its dynamics, and its recovery in patients with stress-related disorders. In this study, we characterized the neural and psychological stress response in a heterogenous patient sample and healthy individuals (HC), as variance in the stress response could be critical in determining phenotypic heterogeneity within patients.
Methods:
We assessed the neural stress response of 56 HC and 57 patients diagnosed with heterogeneous affective disorders including depression and anxiety. Participants were scanned on a 3T scanner while psychosocial stress was induced with an adapted [1] version of the Montreal imaging stress task [2]. Psychological and physiological stress reactivity was measured, including pulse-rate, skin conductance and cortisol. First, we tested differences in the neural stress response (Stress-PreStress) and stress recovery (Post-Stress-PreStress) between patients and HC using univariate, voxel-wise t-tests. To capture individual changes in brain activity independent of their directionality, we assessed the similarity (correlation between β-estimates from 268 regions[3]) of the brain response between the three conditions for each participant. Subsequently, we tested differences of similarity between patients and HC. Lastly, we applied leave-one-out cross-validated linear regression models to determine if region-of-interest based stress-induced brain activity profiles were sufficient to predict the individual psychological stress response.
Results:
Stress-induction was successful with increased negative and decreased positive (ps < .001) emotions after stress across all participants. On the neural level, stress led to stronger deactivations in the posterior cingulate cortex, insulae, and bilateral superior frontal and temporal gyri and activation in the occipital lobe (pFWE <.05, figure 1). Patients differed from HC in their subjective stress experience, reporting stronger increases in negative (p =.010) and decreases in positive emotions (p =.017). In contrast, stress-induced neural activation was not different in patients compared to HC on the group level. However, similarity of activation patterns Pre- and PostStress was lower in patients (p =.04, figure 2). Crucially, both extremes, namely highly similar or dissimilar activation patterns PostStress compared to PreStress were associated with greater subjectively experienced stress (multivariate p =.024). Furthermore, both positive (p =.029) as well as negative (p =.030) subjective stress experience was predicted better than chance with cross-validated regression models based on activation maps and these quantitative predictions were also associated with stress-induced changes in similarity (multivariate p =.0004).
Conclusions:
Patients with affective disorders were characterized by lower similarity of the brain response before and after stress, indicating a slowed stress recovery. Since there were no group differences in stress response in specific regions or directions, this suggests considerable heterogeneity in the individual characteristics of the neural stress response within patients. Moreover, absent or strong changes in brain responses after stress were associated with stronger subjective effects of stress. Thus, taking individual characteristics of patients, such as their subjective, autonomous, and endocrine stress response or genetic background, personality traits, and environmental factors into account will be critical to find underlying neural changes in affective disorders and define more homogeneous subgroups of patients.
