Modern study of sensory processing at the cellular and circuit level has shifted focus toward operation within complex networks. The ways in which cortical network activity shapes responses to sensory stimuli is crucial to understanding aberrant cortical processing, as occurs in autism, schizophrenia, tinnitus, and delirium, as well as understanding the mechanisms of loss of consciousness in sleep and anesthesia. In this thesis, I investigate the circuitry underlying network responses to afferent input in auditory cortex. In the first section of the thesis, I focus on epidural recordings of cortical responses to sensory stimuli and challenge the idea that "evoked" activity driven by sensory stimulation is fundamentally separate from "induced" activity generated intrinsically or from phase reset of ongoing oscillations. Instead, variable network responses can explain changes in all three measures. Second, I study the engagement of different auditory cortical layers in response to input from thalamus in auditory thalamocortical brain slices. I show that there is a discordance between anatomical patterns of thalamocortical/interlaminar connectivity and the sequence in which cortical layers are activated. These results support a model of cortical processing in which spiking in the entire column depends on recurrent "UP state" activity. Finally, I study the contribution of specific populations of inhibitory interneurons to cortical network activity. I show that parvalbumin-positive cells affect the onset of network activity in infragranular layers and inhibit the spread of activity to supragranular layers, while somatostatin-positive cells contribute to terminating network activity. The latter effect may be a specialization of auditory cortex, where network events are brief compared to other cortical areas. These results emphasize the importance of intrinsic cortical activity in determining signal flow during sensory processing. The response properties of cortical networks may be poorly predicted from the features of individual synapses and cells, and emergent network properties may dominate the encoding of sensory features and integration of intrinsic and extrinsic information sources. Future experiments will elucidate the details of these population coding strategies and integrative processes, and their regulation by cortical inhibition and exogenous pharmaceutical agents