| n recent years, complex networks have provided an increasingly challenging framework for the study of collective behaviors based on the interplay between complexity in the wiring architecture and dynamical properties of the coupled units. Among others, modularity -natural divisions of network nodes into densely connected subgroups- has been found to be a common organizational mechanisms in many networks, ranging from RNA structures to biological organisms and social groups. In the brain, functional and anatomical specializations are evident from physiological, neuropsychological and neuroimaging studies. Nevertheless we find that at the same time, cognitive functions are not encapsulated in isolated modules, but distributed across brain areas. As a result, an important topic that is still under debate in neuroscience is how the brain orchestrate specialized and anatomically distant brain regions to form a web-like structure that is able to combine and manage information at both, local and global scales. To understand how the brain organizes these activities, we have analysed the modular structure -natural divisions of network nodes into densely connected subgroups- of the brain networks extracted from functional magnetic resonance imaging (fMRI) signals. We propose a fully data driven approach that do not rely on a priori choice of a seed brain region nor signal averaging in predefined brain areas. Results confirm that the activity of brain networks presents -at voxel level- functionally differentiated but interacting structures in which a non-random modular structure emerges. We consider that our study provides a meaningful insight into how the brain combines local and global information to integrate neural activities into unified and coherent brain functions: the presence of a modular structure also denotes an increase in the robustness, flexibility, and stability of the brain, while establishes a link between the segregated (local) versus integrated (global) processing of the functional brain activity. Further analysis could help to understand how all these modules cooperate to bridge between the brain processes into different pathological or cognitive states. |
|