Fluorodeoxyglucose autoradiography, quantitative image analysis, and a multivariate tool (partial least squares) were used to assess distributed patterns of brain activation in postnatal day 17 and day 12 rat pups engaged in extinction of instrumental behavior. Pups were trained in a straight alley runway on an alternating reward schedule, or on a pseudorandom reward schedule, injected with fluorodeoxyglucose, and then shifted to continuous nonreward (extinction). Another group at each age served as handled controls. Day 17 pups trained on the alternating schedule demonstrated faster extinction rates compared to those trained on the pseudorandom schedule, a phenomenon known as the partial reinforcement extinction effect. No differences were found between day 12 groups. Partial least-squares analysis revealed age-related increases in fluorodeoxyglucose uptake across all three training conditions in the cingulate and frontal cortices, amygdala, midline thalamic nuclei, cerebellum, and in several brainstem regions. Training-related increases common to both age groups were found in the orbital frontal cortex, limbic thalamus, gigantocellular reticular nucleus, the somatosensory system, and cerebellum. Age-dependent training effects were found in the interpositus and medial cerebellar nuclei wherein fluorodeoxyglucose uptake increased in the day 12 alternation and pseudorandom groups relative to controls. Day 12 pups trained on the alternating schedule demonstrated increased uptake in the anterior dorsal thalamus relative to pseudorandom and control pups. Hence, a large-scale neural system comprised by somatosensory, cerebellar, and brainstem regions govern extinction behavior in preweanling rats. Recruitment of limbic structures may allow the older pups to modify extinction behavior based on prior learning.