What does learning look like in a bumblebee brain?

Figure 1 from Li Li et al. a) the arena set-up with bees exposed to 10 flower colors; b) the different experimental treatments; c) the frontal-view of a whole bumblebee brain; d-f) enlarged views of the portions of bumblebee brains involved in vision; g) enlarged so you can see the labeled microglomeruli; h) microglomeruli structure.

This week's paper is in the Proceedings of the Royal Society of London B, lead-authored by Li Li, a PhD student in the Chittka group at Queen Mary University London. Neurons connect to (and thus communicate) with other neurons at regions called synapses. Microglomeruli are places were lots of synapses come together (synaptic complexes). This study looked at the link between bee learning of flower colors, and the density of microglomeruli in the bee's brains.

Li Li and her colleagues tested bumblebees in five different treatments. All bees were first pre-trained to forage from colorless (clear) artificial flowers. For some of the bees that was it, they were dissected to establish a baseline of microglomeruli density with no exposure to colors (Treatment 1). The next group of bees were trained to distinguish between 10 different colors, five of which had sugar solution and five had nasty tasting quinine (like what's in your tonic water; Treatment 2). The third group of bees were trained to distinguish between only two colors to examine the effect of the number of colors bees had to learn on microglomeruli density (Treatment 3). Group four continued to forage on clear flowers for as long as the other bees were learning flower colors to control for the effect of amount of foraging time on microglomeruli density (Treatment 4). In the final treatment the bees continued to forage on clear flowers, but they were surrounded by the colored flowers (with no rewards) to control for the effect of exposure to lots of flower colors (Treatment 5). In treatments 2-5 after the bees were trained they were then given memory tests to determine how well they performed at learning the task. You can see all these treatments detailed in Fig 1b above. 

In the ten color discrimination task (Treatment 2), Li Li and her colleagues found that learning rate and memory performance correlated with microglomeruli density in the collar part of the bee brain (Fig 1c above). This result, however, does not distinguish between correlation and causation. That is to say, it could be that better learners brains' increase more in microglomeruli density, or that bees with better learning abilities are the ones that already have higher microglomeruli densities. In comparing bees that had color learning versus those with no color learning (Treatments 2 and 1), they found that bees in the color learning treatment had higher microglomeruli densities and larger brain calyxes (the brain part in Fig 1d) in total. There are possibilities other than color learning directly to explain this pattern, however, and using the other treatments Li Li and her colleagues were able to show that the brain size is simply a product of the foraging activity the bees did (comparing treatments 4 and 2), and that the microglomeruli density is a product of exposure to colors not learning necessarily (comparing treatments 5 and 2). So in summary, there is a correlation between density of these microglomeruli and bee learning and memory. But causation is tricky, learning colors does seem to increase microglomeruli density, but this effect may largely be due simply to exposure to the colors.