Classification of brain activity using synolitic networks

Because the brain is an extremely complex hypernet of interacting macroscopic subnetworks, full-scale analysis of brain activity is a daunting task. Nevertheless, this task can be greatly simplified by analysing the correspondence between various patterns of macroscopic brain activity, for example, through functional magnetic resonance imaging (fMRI) scans, and the performance of particular cognitive tasks or pathological states. The purpose of this work is to present and validate a methodology of representing fMRI data in the form of graphs that effectively convey valuable insights into the interconnectedness of brain region activity for subsequent classification purposes. Methods. This paper explores the application of synolitic networks in the analysis of brain activity. We propose a method for constructing a graph, the vertices of which reflect fMRI voxels’ values, and the edges and edge weights reflect the relationships between fMRI voxels. Results and Conclusion. Based on the classification of fMRI data by graph properties, the effectiveness of the method in conveying important information for classification in the construction of graphs was shown.

cognitive processes, functional magnetic resonance imaging, synolitic networks, graphs, classification, machine learning