Depression is a prevalent and debilitating psychiatric disorder, with over 25 million adults in the United States experiencing clinically significant depression symptoms (Brody et al., 2018; Ettman et al., 2020). In addition, rates of depression are two times higher in women as compared with men (Brody et al., 2018). Recent epidemiological research further indicates that the prevalence rate of depression has increased by threefold since the start of the COVID-19 pandemic (Bueno-Notivol et al., 2021; Ettman et al., 2020; Shah et al., 2021), with women more adversely affected than men (Liu et al., 2020). Abnormalities within the hypothalamic-pituitary-adrenal (HPA) axis, in particular glucocorticoid insensitivity and chronic hypercortisolemia, have been consistently implicated in the pathogenesis of depression (Akil et al., 1993; Gaffey et al., 2019; Halbreich et al., 1985; Holsboer, 2001; Pariante, 2009; Pariante and Lightman, 2008). Glucocorticoid insensitivity in depression has been associated with glucocorticoid receptor (GR) dysfunction including reduced GR mRNA expression and diminished (GR)-mediated negative feedback on the HPA-axis, which normally functions to inhibit further cortisol release (Anacker et al., 2011a; Hepgul et al., 2013; Nemeroff et al., 1992; Pariante, 2009; Pariante and Miller, 2001). Clinical research also indicates that the beneficial effects of antidepressant medications may be related to changes in GR function in depression (Anacker et al., 2011b; Pariante et al., 2012). Together these findings suggest that treatments that modulate glucocorticoids may be helpful in reducing depressive symptoms. However, few studies have examined the effects of acute glucocorticoid versus placebo administration on brain network connectivity in depression.

Resting-state neuroimaging research indicates that individuals with depression exhibit abnormal connectivity of cortical and subcortical brain regions involved in emotion, emotion regulation, and cognition. In particular, aberrant resting-state functional connectivity (rsFC) within and between brain structures located in the default mode network (DMN), frontoparietal network (FPN), and salience network (SN) has been found in depressed populations (Berman et al., 2011; Greicius et al., 2007; Kaiser et al., 2015a; Manoliu et al., 2014; Mulders et al., 2015; Philippi et al., 2018; Pizzagalli, 2011; Ramasubbu et al., 2014; Sheline et al., 2009; Whitfield-Gabrieli and Ford, 2012; Williams, 2017; Zhu et al., 2012). For instance, studies have frequently reported elevated rsFC within the DMN in individuals with depression and those with greater depressive symptoms (Berman et al., 2011; Greicius et al., 2007; Kaiser et al., 2015a; Philippi et al., 2018; Sheline et al., 2009; Whitfield-Gabrieli and Ford, 2012; Williams, 2017; Zhu et al., 2012). Other research has found increased connectivity between FPN and DMN as well as increased and decreased DMN-SN connectivity in depression (Manoliu et al., 2014; Mulders et al., 2015; Pizzagalli, 2011; Ramasubbu et al., 2014). Together, these studies have identified elevated rsFC of the DMN and altered connectivity between the DMN and the FPN and SN in individuals with depression. Given the aforementioned role of glucocorticoids in depression, it is important to understand whether neuroendocrine changes associated with depression may further affect connectivity of these same brain networks.

Neuroimaging studies suggest that chronically elevated endogenous cortisol is associated with alterations in brain activity and connectivity across similar brain structures implicated in depression. In particular, variation in endogenous cortisol is related to connectivity of the DMN, FPN, and SN in healthy adolescents and adults (Kalafatakis et al., 2021; Thomason et al., 2011; Vaisvaser et al., 2013; Veer et al., 2012; Vogel et al., 2015), as well as those with depression (Peters et al., 2019, 2016). For example, one study investigating endogenous levels of cortisol in individuals with a history of depression found that higher levels of endogenous cortisol were associated with greater rsFC between DMN and FPN as well as between SN and FPN when compared with a healthy control group (Peters et al., 2019). These findings are relevant because greater rsFC between these networks associated with elevated cortisol may further exacerbate depressive symptoms. While this research on endogenous cortisol has provided significant contributions to our knowledge of how glucocorticoids affect brain activity in healthy and depressed groups, there are some methodological limitations. For example, these studies of endogenous cortisol primarily used cross-sectional case control designs as opposed to experimental administration of glucocorticoids with a placebo-controlled crossover design. Such an experimental approach is valuable and may be particularly important for individuals with depression where exogenous cortisol administration has been shown to be beneficial for memory and brain activity (Abercrombie et al., 2018, 2011; Bremner et al., 2004). Only one study to our knowledge has examined the effects of acute cortisol administration on rsFC using a double-blind crossover experimental design in healthy male participants (Henckens et al., 2012). Henckens and colleagues (Henckens et al., 2012) reported that the administration of cortisol enhanced connectivity between SN and DMN regions at rest in healthy males with no history of depression. In depressed populations, a few studies using task-based fMRI suggest that exogenous cortisol can normalize activity and/or connectivity of DMN regions during an emotional task in those with a history of depression (Abercrombie et al., 2011; Rivera-Bonet et al., 2021). Altogether, these results suggest that cortisol can affect brain activity and connectivity of large-scale networks in healthy and depressed populations. In addition, some research has shown correlations between elevated endogenous cortisol and greater depressive symptoms (Holsboer, 2001; Karlović et al., 2012; Nemeroff, 1996; Pariante and Miller, 2001). Further, depressive disorder symptom presentation can be heterogeneous and may result in neurobiological differences (e.g., Antonijevic, 2006; Drysdale et al., 2016). Considering this research and the aforementioned abnormalities in the functioning of the HPA axis in depression, it is possible that the effects of transient cortisol elevation may differ in healthy versus depressed individuals and/or in those with more severe depressive symptoms. Specifically, if cellular glucocorticoid signaling is normal in healthy individuals but is deficient in those with depression and/or greater depressive symptoms (who are more likely to exhibit cellular glucocorticoid insensitivity) then the effects of transient cortisol elevation on rsFC may be different for those with depression and/or more severe depressive symptoms.

To our knowledge, no study has examined the effect of acute cortisol administration on rsFC of the DMN, FPN, and SN in women with varying levels of depression using a double-blind crossover design. The current study aims to investigate changes in rsFC of the DMN, FPN, and SN following administration of exogenous cortisol in women with different depression histories and symptom severities. We hypothesize that individuals with a history of depression or greater depression symptoms will have decreased connectivity within DMN regions, decreased FPN-DMN connectivity, and increased and DMN-SN connectivity after cortisol administration compared to placebo.