Honeybees can be trained to associate odorants to sucrose reward by conditioning the proboscis extension response. Using this paradigm, we have recently shown that bees can solve a side-specific task: they learn simultaneously to discriminate a reinforced odor A from a non-reinforced odor B at one antenna (A+B−) and the reversed problem at the other antenna (A−B+). Side-specific (A+B−/B+A−) conditioning is an interesting tool to measure neurophysiological changes due to olfactory learning because the same odorant is excitatory (CS+) on one brain side and inhibitory (CS−) on the opposite side. In the bee brain, the antennal lobe (AL) is the first olfactory relay where the olfactory memory is established. Using calcium imaging, we compared odor-evoked activity in the functional units, the glomeruli, of the two ALs, both in naive and conditioned individuals. Each odor evoked a different pattern of glomerular activity, which was symmetrical between sides and highly conserved among naive animals. In conditioned bees, response patterns were overall symmetrical but showed more active glomeruli and topical differences between sides. By representing odor vectors in a virtual olfactory space whose dimensions are the responses of 23 identified glomeruli, we found that distances between odor representations on each brain side were significantly higher in conditioned than in naive bees, but only for CS+ and CS−. However, the distance between CS+ and CS− representations was equal to that of naive individuals. Our work suggests that side-specific conditioning decorrelates odor representations between AL sides but not between CS+ and CS− within one AL.