The brain of any species can be modeled as a network of regions and connections between those regions. Here, I analyze the brain’s large-scale network mechanics. I show that the brain can be divided into subnetworks, each with a discrete function. I provide evidence that the each subnetwork’s processing is mostly modular; however, certain regions exist that perform integrative functions. I characterize this integrative set of regions in depth and discover that this set of regions exists across various species’ brains and even in man-made networks. I also identify an optimal network structure for cognitive processing. Finally, I demonstrate one mechanism for why the brain’s network structure was selected by evolution.