Life-Span Development of Brain Network Integration Assessed with Phase Lag Index Connectivity and Minimum Spanning Tree Graphs

Graph analysis of electroencephalography (EEG) has previously revealed developmental increases in connectivity between distant brain areas and a decrease in randomness and increased integration in the brain network with concurrent increased modularity. Comparisons of graph parameters across age groups, however, may be confounded with network degree distributions. In this study, we analyzed graph parameters from minimum spanning tree (MST) graphs and compared their developmental trajectories to those of graph parameters based on full graphs published previously. MST graphs are constructed by selecting only the strongest available connections avoiding loops, resulting in a backbone graph that is thought to reflect the major qualitative properties of the network, while allowing a better comparison across age groups by avoiding the degree of distribution confound. EEG was recorded in a large (n = 1500) population-based sample aged 5-71 years. Connectivity was assessed using phase lag index to reduce effects of volume conduction. Connectivity in the MST graph increased significantly from childhood to adolescence, continuing to grow nonsignificantly into adulthood, and decreasing significantly about 57 years of age. Leaf number, degree, degree correlation, and maximum centrality from the MST graph indicated a pattern of increased integration and decreased randomness from childhood into early adulthood. The observed development in network topology suggested that maturation at the neuronal level is aimed to increase connectivity as well as increase integration of the brain network. We confirm that brain network connectivity shows quantitative changes across the life span and additionally demonstrate parallel qualitative changes in the connectivity pattern.