Our sensory system is bombarded with information that can change whimsically and yet we make sense of the flow of the information effortlessly. How does the brain encode such richly dynamic stimuli? Specifically, how does the auditory system encode rich spectral and temporal aspects of the stimulus and how does it depend on the behavioral state of the animal? My study aims to answer these questions within the scope of mouse auditory cortex (ACX) using imaging techniques on various scales. Firstly, I studied how the ACX encodes one temporal aspect of the sound, specifically the onset and the offset. I found that offset responses dominated ACX at high sound levels and their strength depended on auditory cortical fields. Moreover, ACX neurons likely inherit their offset responses from thalamocortical input, which is further processed by local cortical microcircuit. Second, I studied the spectral tuning properties of layer 2/3 neurons in mouse ACX using two-tone stimuli. This study revealed the complex inhibitory sideband structures not only in excitatory and inhibitory neurons, but also in feedforward input from auditory thalamus. These complex structures showed a higher degree of feature selectivity of auditory neurons beyond what is predicted by conventional tuning, and thus auditory cortical responses are highly dependent on the spectral context. These two studies focused on passive listening, but cortical responses could depend on the behavioral state of the animal. The predictive coding theory proposes that sensory cortical responses are a form of error response signaling when sensory input failed to conform with predictions from higher order brain areas. Thus, to study the encoding of spectrotemporally dynamic stimulus under active engagement and to test the predictive coding theory, I designed a novel behavior paradigm that allowed the animal to interact with the sound stimulus and studied the cortical responses to not only the combination of sensory information and the animal’s action but also the introduced perturbation. Together, this dissertation combined advanced imaging techniques and innovations in experimental designs to provide new insight into how ACX encodes sound stimulus under various scenarios.