Why do we keep failing in identifying reliable biological markers in depression?

Despite major depression is a raising and wide spread health problem, the knowledge about its biological bases is limited. This not only has consequences epistemologically speaking, but also has clinical implications. However, considering the whole literature on biological bases of depression, it seems that a precise or unique biological marker enough specific and sensible to be used in clinical practice or to better understand the psychopathology of the disorder is not available and research has produced highly heterogeneous results. Moreover, the studied markers are often assessed with expensive and not commonly available techniques, limiting their possible application in clinical routine. Here we hypothesize that the emerging clinical and biological heterogeneity mainly reflect the fact that depression, as most of psychiatric disorders, is not a unique condition, but the sum of different entities with similar symptoms. Such entities may rely on different and only partially overlapping biological features. Therefore, rather, than continue to invest on highly inconsistent studies, we sustained the necessity of a shift in the study paradigm. Namely, future researches may focus on a bottom-up approach considering patients with homogeneous biological markers and evaluating whether these markers correspond to precise clinical features. Unsupervised supported vector machines and other data driven approaches may represent the best tool for such type of research. Pursuing the idea of a clinical exploitation of this type of research we also considered markers derived from electrocardiography (like the heart rate variability) the most promising given the relatively low cost, simplicity and availability of this technique.