In the natural world, individual sensory signals rarely occur in isolation. In audition, salient signals commonly occur within constantly fluctuating, complex soundscapes. A key problem for the auditory system is therefore the differentiation of temporally and spectrally overlapping signals. The mammalian auditory system is excellent at grouping signals into separate objects and can do so using a small number of cues (Bregman, 1990), but the neural mechanisms that underlie these processes are poorly understood. One phenomenon that uses these grouping processes is co-modulation masking release (CMR, Verhey et al., 2012). Coherent amplitude modulation of sound (co-modulation) across a broad range of frequencies (in the form of a broadband signal) can increase the detectability of concurrent, narrow band signals (such as a pure tones, Hall et al., 1984). Here we investigate how neuronal populations in primary auditory cortex (A1) represent distinct overlapping objects, by recording activity evoked by narrowband signals in the presence of a co-modulated broadband signal. High density, multi-site extracellular recordings, and whole cell patch clamp recordings were made from neurons in primary auditory cortex (A1) of anaesthetised mice (female, NMRI, 5-10 weeks). Functional responses to pure tones were used to confirm probe location within A1. Cell responses (defined as changes in firing rate or subthreshold Vm for extra- and intracellularly recorded neurons respectively) could be categorised according to their responses to either the broadband or narrowband signals. While the majority of cells responded only to the onset and/or offset of the broadband signal (~52%), the firing rate of a subpopulation locked to the phase of the modulation (~13%). Also a small percentage of cells represented the tone signal (~13%): either by suppressing their phase locking (~3%), or by sharply increasing their firing rates at the onset (~5%) or offset of the tone (~5%). Surprisingly, tone offset-evoked responses were more sensitive to sound levels than the onset-evoked responses. These results suggest that features of both narrow- and broadband sounds are reliably encoded in two separable, yet overlapping, populations of A1 neurons. Sufficiently increasing the strength of one signal, relative to the other, allows that signal to suppress the response to the other and enhance its own representation within A1. We demonstrate that when using cues of object formation, A1 is able to encode both signals as separately represented objects, even at low signal levels. This mechanism could be used in the selective perception of low level signals in noisy environments.