Network Analysis of Depression Symptoms and Nine Hair-based Hormones Among Chinese Undergraduate Students

Revealing the interactions between depression symptoms and neuroendocrine systems contributes to a better understanding of the neuroendocrine mechanisms underlying depression. However, few studies have focused on this issue. This study employed network analysis to examine the interactions between depression symptoms and the hypothalamic-pituitary-adrenal (HPA) and hypothalamic-pituitary-gonadal (HPG) axes and endocannabinoid system (ECS) and melatonin system (MELS). This study recruited 252 Chinese undergraduates (male/female: 192/60; age range: 17-24 years; mean age = 20.0 years; standard deviation = 1.4 years). Depression symptoms were assessed using the Beck Depression Inventory, and hair concentrations of cortisol, cortisone, dehydroepiandrosterone (DHEA), testosterone, progesterone, N-arachidonylethanolamide (AEA), 1-arachidonoylglycerol (1-AG), melatonin, and N-acetylserotonin (NAS) were measured as biomarkers of the four neuroendocrine systems. The results showed that edgeweights were significant for certain edges between depression symptoms and hormones, with the exception of testosterone. Across the 15 networks, the most central symptoms were “Self-dislike,” “Guilty feeling,” “Lack of satisfaction,” “Work inhibition,” and “Punishment feelings.” The expected influence (EI) of individual symptoms was not impacted by single or multiple hormones, except that 1-AG weakened the EI of “Punishment feelings.” All depression symptoms exhibited increased average EI for cortisol, DHEA, testosterone, melatonin, and NAS, and decreased average EI for AEA and 1-AG, whereas cortisone and progesterone showed no effect. Their average and bridge EI were strengthened by the HPA and HPG axes, MELS, and the combined four-system model, but weakened by the ECS. Overall, interactions between depression symptoms and neuroendocrine systems did not impact the centrality or bridge centrality of specific depression symptoms but influenced their average centrality measures. These findings provide a novel network-based perspective on the multiple neuroendocrine mechanisms underlying depression.