Artificial Intelligence-Based Major Depressive Disorder (MDD) Diagnosis Using Raman Spectroscopic Features of Plasma Exosomes

In vitro diagnosis using biomarkers for major depressive disorder (MDD) can offer considerable advantages in overcoming the lack of objective tests for depression and treating more patients. Plasma exosomes can be novel biomarkers for MDD based on their ability to pass through the blood-brain barrier and offer brain related information. Here, we demonstrate a novel and precise MDD diagnosis using deep learning analysis and surface-enhanced Raman spectroscopy (SERS) of plasma exosomes. Our system is implemented based on 28,000 exosome SERS signals, providing sample-wise prediction results. Notably, this approach shows remarkable performance in predicting 70 test samples unused in the training step, with an area under the curve (AUC) of 0.939, a sensitivity of 91.4%, and a specificity of 88.6%. In addition, we confirm that the diagnostic scores were correlated with the degree of depression. These results show the utility of exosomes as novel biomarkers for MDD diagnosis and suggest a novel approach for prescreening techniques for psychiatric disorders.