Graph theory provides a framework for exploring brain network organization in normal and pathological conditions.13,14,37 Graph theoretical analysis to fMRI, EEG/MEG and DTI data can model the whole brain as a single network and investigate its properties such as network structure, modularity, and robustness to damage (Panel 2).14 The healthy human brain is thought to be organized into a ‘small-world’ topology,38 a network architecture that combines an efficient balance between local (short range) and global (long range) connectivity. This small-world configuration is considered better suited for information transfer and thus presumably for cognitive processing than the topology of ‘random’ or ‘regular’ networks.39 Graph theory can also extract functional subnetworks (‘modules’) and quantify interactions between them by using data-driven modularity algorithms.40 Another area of graph theory is devoted to the investigation of highly connected (‘hub’) nodes, since these regions are critical for network integrity (Panel 2).
