Machine learning as the new approach to understand biomarkers of suicidal behavior

Compared to other medical fields, the situation in psychiatry is particularly lacking in terms of identification of biological markers that can complement current clinical interviews. Such markers would enable more objective and rapid clinical diagnosis and allow more accurate monitoring of treatment responses and remission. Current technological developments can provide analyses of various biological marks at a high-throughput scale and at reasonable cost, and therefore such “-omic” studies are also now entering psychiatry research. However, big data demands a whole plethora of new skills in data processing before clinically useful information can be extracted. To date, the classical approaches to data analysis have not really contributed to identification of biomarkers in psychiatry. However, the extensive amount of data might be taken to a higher level if artificial intelligence can be applied, in the shape of machine learning algorithms. Not many studies on machine learning in psychiatry have been published, but we can already see from the handful of studies now available that the potential to build a screening portfolio of biomarkers for different psychopathologies, including suicide, exists.