Pain is multidimensional, including sensory-discriminative, affective-motivational, and cognitive-evaluative components. Although the concept of pain is learned through life, it is not known when and how the brain networks that are required to encode these different dimensions of pain develop. Using the 2 largest available databases of brain magnetic resonance images-the developing Human Connectome Project and the Human Connectome Project-we have mapped the development of the pain connectome-the neural network required for pain perception-in infants from 26 to 42 weeks of postmenstrual age (PMA, n = 372), compared with adults (n = 98). Partial correlation analysis of resting BOLD signal between pairwise combinations of 12 pain-related brain regions showed that overall functional connectivity is significantly weaker before 32 weeks PMA compared with adults. However, over the following weeks, significantly different developmental trajectories emerge across pain connectome subnetworks. The first subnetwork to reach adult levels in strength and proportion of connections is the sensory-discriminative subnetwork (34-36 weeks PMA), followed by the affective-motivational subnetwork (36-38 weeks PMA), while the cognitive-evaluative subnetwork has still not reached adult levels at term. This study reveals a previously unknown pattern of early development of the infrastructure necessary to encode different components of pain experience. Newborn neural pathways required for mature pain processing in the brain are incomplete in newborns compared with adults, particularly regarding the emotional and evaluative aspects of pain. The rapid age-related changes suggest that pain processing, and consequently pain experience, changes rapidly over this developmental period and unlikely to be the same as in adults, even at term.