Bipolar Proteomic Assay Validation Study

This is a hypothesis-driven confirmatory study to validate the diagnostic signature (model) for distinguishing BDI from MDD that also aims to optimize the models to discriminate BDII from MDD and BDI. A binary classification model, using linear discriminant analysis and based on 13 a priori-defined proteomic markers will aim to distinguish BDI from MDD. An alternative binary classification model based on multiple logistic regression and using 10 a priori -defined proteomic markers will aim for the same result. To improve the predictive performance of the signatures, items from self-report mood rating scales and treatment-emergent changes in proteomic markers will be analyzed. In addition, the study will examine if baseline or early treatment-emergent changes in proteomic markers predict treatment response.