. Anxiety disorders are the most common class of psychiatric disorder ? annually affecting 18.1% of the U.S. population. Dysregulation of the hormonal stress response is thought to be central in the pathogenesis of these disorders. Davis et al. (2010) proposed that the amygdala and bed nucleus of the stria terminalis (BNST) play complementary roles in regulating the stress system, based primarily on rodent research. This model proposes that the centromedial amygdala (CM) mediates the immediate response to threat (i.e. fear), while the bed nucleus of the stria terminalis (BNST) mediates the sustained response (i.e. anxiety). The BNST may be a more proximal driver of the human stress response than the CM given that 1) the BNST exhibits robust direct connections to the paraventricular hypothalamus ? a key node in initiating the HPA-axis ? while the CM does not, 2) much of the stress individuals experience is associated with anxiety about past or potential threat rather than fear of immediate danger, and 3) past findings on whether amygdala activity predicts biomarkers of stress reactivity are mixed. As such, investigating how the function and connectivity of the human BNST relates to the stress system will provide a foundation for studying the contributions of individual stress system inputs to the individual features of anxiety disorders, uncover potential targets for pharmaceutical treatments of these disorders, and provide neural biomarkers of susceptibility to psychopathology that will be useful in the assessment of psychotherapeutic interventions. However, while the human BNST responds to stressful stimuli, there is a dearth of research on the consequences of BNST function and connectivity on the human stress system. Furthermore, little is known about how interventions that reduce stress ? such as mindfulness based stress reduction (MBSR) training ? affect activity and connectivity of the BNST. The proposed research will explore the relationship between the BNST and the human stress response using two large, existing datasets (N=141 & 140) across three aims.
Aim 1 will test the hypothesis that a greater BNST BOLD response to negatively-valence (vs. neutral) images will predict increased stress reactivity, as indexed by the stress biomarkers cortisol and salivary ?-amylase, during a later Trier Social Stress Test (TSST), after controlling for the CM BOLD response.
Aim 2 will use dynamic causal modeling to test the hypothesis that an 8-week MBSR training course (vs. a health enhancement program, a validated active control condition) will result in increased inhibition of the BNST via the ventromedial prefrontal cortex (an area that inhibits the stress response through connections with the BNST in rodents), decreased activation of the BNST via the CM and basolateral amygdala, and decrease activation of the periventricular hypothalamus via the BNST during threat anticipation.
Aim 3 will use diffusion tensor imaging to test the hypothesis that greater white matter integrity in tracts connecting the vmPFC and BNST will be associated with reduced daily cortisol output, while greater white matter integrity in tracts connecting BNST with the amygdala and hypothalamus will be associated with increased cortisol output.
Investigating the neural inputs that drive activity of the hormonal stress system is critical for understanding anxiety disorders, given that dysregulation of this system is thought play a role in the pathogenesis of these disorders. While rodent research suggests that the bed nucleus of the stria terminalis (BNST) and centromedial amygdala play complimentary roles in driving the stress response, the role of the BNST in the human stress response has not been examined. The proposed research will lay a foundation for future work on the etiology of stress system changes associated with anxiety disorders by examining the relationship of BNST function and connectivity (using diffusion tensor imaging) with acute reactivity and daily function of the stress system, as well as the effect of mindfulness based stress reduction training on BNST functional connectivity (using dynamic causal modeling) within two large extant datasets (N=141 & 140).
