Perception requires sensory signals to be framed in their current environment and recent history. To incorporate this context, cortical feedback modulates sensory input by attenuating predictable background and amplifying unexpected signals. The latter response, known as deviance detection, flags relevant information for ongoing perceptual constructions, yet is disrupted in neuropsychiatric disorders. Deviance detection appears across many species, including rodents, and can serve as a translational tool for investigating sensory context processing in health and disease. While rodent studies have mapped molecular, cellular, and circuit-level bases for this neural signature, when and how deviance detection arises during development remains unclear. Adolescence encompasses the final wave of maturation in prefrontal (i.e., feedback) regions and the onset of many neuropsychiatric disorders, representing a plausible period for the development of deviance detection. To test this, deviance detection was measured via local field potentials in primary visual cortex during unexpected stimuli in peri-adolescent and adult mice. Enhanced responses to deviant stimuli (i.e., deviance detection) were observed in adult but not peri-adolescent mice, suggesting that deviance detection emerges during the adolescent period (Ch. 2). Subsequent investigations addressed how deviance detection develops in adolescence across three independent substrates: i) microglia, which are thought to prune synapses in development (Ch. 2), ii) Kalrn, a cytoskeletal regulator which influences morphology in deviance-detecting cells (Ch. 3), and iii) top-down input to V1, which is required for adult deviance detection (Ch. 4). Transient microglial depletion in development did not impact deviance detection, suggesting that this response can develop independently of microglia. In contrast, deviance detection was abolished in mice containing a Kalrn missense mutation linked to adolescent-onset dendritic loss. Finally, reducntions in top-down input to superficial V1 was identified in adolescence. Altogether, these data reveal i) the development of deviance detection in adolescence, ii) necessary (i.e., Kalrn) and dispensable (i.e., microglia) components of the cortical milieu for the development of deviance detection, and iii) likely pruning of top-down input in superficial V1 during adolescence, which may contribute to emerging deviance detection circuitry. These findings underscore the developmental value of deviance detection in translational research and invite further exploration of fronto-sensory development.