How the brain separates information about multiple objects despite overlap in the neurons responsive to each item is not well understood. It has recently been proposed that when more than one stimulus is present, single neurons can fluctuate between coding one vs. the other(s) across some time period, i.e. a form of neural multiplexing in the time domain (Caruso et al., 2018). However, it is not known (a) whether such fluctuations occur specifically when two distinct objects are present and not in other cases, and (b) how fluctuations in individual neurons might be coordinated with one another to ensure the representation of both objects across the population. Here we investigated these questions in visual cortex (V1, V4) using two visual gratings that could be presented adjacent to one another (two objects), or superimposed to form a single “plaid” object. We found fluctuating activity in V1 only for the two-object case. The fluctuations were coordinated across the neural population to produce a pattern of V1 noise correlations that had not previously been detected with single stimuli: distinct distributions of positive and negative values, depending on whether the two neurons in the pair had similar or different preferences for the individual grating stimuli. A similar pattern was also seen in V4 for adjacent stimuli, but was not observed in either structure for single stimuli or when the two gratings were superimposed and formed a single object. Importantly, the bimodal correlation patterns were most pronounced among pairs of neurons showing the strongest evidence for multiplexing. These findings suggest multiple stimuli evoke different response dynamics than those evoked by single stimuli, lending support to the multiplexing hypothesis and suggesting a means by which information about multiple objects can be preserved despite the apparent coarseness of sensory coding.