I. Introduction
At present functional magnetic resonance imaging (fMRI) is one of the most popular methods of studying the human brain in neurophysiology. Many projects, such as the Human Connectome Project [1], the Human Brain Project [2], and others collect a large amount of fMRI data to analyze the activity of the human brain. One of the most popular group of problems solved with the help of machine learning using fMRI data is classification. The whole wide range of classification problems can be divided into two subgroups: (1) the classification of groups of people and (2) the classification of the brain state of people.