Biomarkers and Mental Disorders: A Relevance Analysis Using a Random Forest Algorithm.

Depression and anxiety are mental health disorders that significantly impact global public health, affecting more than 280 million people with depression and 301 million with anxiety worldwide. These conditions impair individuals' ability to engage in economic and personal activities and can lead to severe outcomes, such as suicide. Current research suggests that inflammatory cytokines, such as interleukin-1β (IL-1β), interleukin-6 (IL-6), and tumor necrosis factor (TNF), play crucial roles in the pathophysiology of these disorders, influencing neurotransmitters. Elevated cortisol levels, typically associated with anxiety, worsen these conditions through dysregulation of the hypothalamic-pituitary-adrenal (HPA) axis. Additionally, vitamin D deficiency has been linked to reduced production of dopamine and norepinephrine, hormones involved in depressive symptoms. This study utilized the Random Forest machine learning algorithm along with cross-validation to assess the importance of various biomarkers, including IL-1β, IL-6, IL-8, TNF, cortisol, vitamin D, NT-proBNP, CK-MB, troponin, myoglobin, and C-reactive protein (CRP), in volunteers of both sexes diagnosed with mental disorders. A single sample from each of the 96 participants was analyzed, consisting of 50 women and 46 men. The results revealed sex-specific differences in biomarker relevance, with vitamin D, CRP, and D-dimer being the most predictive for depression in men, while IL-6, CRP, and vitamin D were significant in women. For anxiety, vitamin D and myoglobin were important biomarkers in men, while IL-8 and vitamin D were key in women. The methodological strategy adopted, based on the use of Random Forest and cross-validation assessment, not only confirmed the robustness of the model but also reliably identified the most important biomarkers for the outcomes studied.