Using Machine Learning Methods to Search for EEG and Genetic Markers of Depressive Disorder

Depression is one of the most common mental disorders. Therefore, the development of new methods for early diagnosis of depression is a highly relevant task. It is well known that depression has both a genetic predisposition and greatly depends on the patient's life background. Therefore, the analysis of only genetic markers of depression is usually unsuccessful, because it does not take into account the physiological state of a person at the time of examination. The aim of our study was to develop an algorithm for the joint analysis of a collection of genetic and neurophysiological data collected from healthy people and patients with depression to identify genetic markers and neurophysiological correlates of pathology. As EEG indicators, we considered the amplitudes of evoked potentials under the conditions of performing tasks in the stop-signal paradigm. We applied machine learning algorithms that allow us to identify single nucleotide polymorphisms associated with the risk of depression.