Do bee brains show the path to better navigation systems?

Researchers at Lund University in Sweden have discovered that bumblebees have incredible navigational memory. Rickesh Patel and colleagues did it by tracking the walking behaviour of foraging bees. Here’s what he told Chris Smith…

Rickesh - For a long time, people have known that bees are very good at constructing these straight line paths from a food location to home and back to that food location again. And one other thing that people have noticed is that bees are really good at making more complex navigational paths. So for instance, bees have been seen to make a shortcut from one food location to another, even though they've never really travelled that path before. At least that's what is thought to be the case. Also, bees will go to multiple flower patches along one flight and, over time, they're able to make really optimised circuits or direct circuits between these food locations before returning home. So these complex navigational tasks require a bit more than just following a single straight line path back to home. Again, bees can use landmarks to navigate, but it seems that this system gets better over time, which may indicate that perhaps it's these straight line paths that are being used together that might explain how they're able to do these sorts of tasks.

Chris - It sounds from what you're saying like you regard the bees as having a load of memories of places that are connected by a straight line, but they can link up those different memories, different paths from one place to another, to almost create a map in their brains.

Rickesh - I think you can think of it like a web in which there are these nodes, familiar places, and these straight lines that connect these nodes together. This is just an idea. No one actually knows if this is true.

Chris - But what experiments were you able to do that would convince you or has convinced you that this is probably what's going on?

Rickesh - In order to study how bees are able to keep track of these directions and distances in their brains, I needed to be able to construct a situation in which bees were only using vector navigation, these straight line paths, and not using any visual features which bees are also quite good at doing. And since they're travelling over kilometres, I needed to devise a way of making bees navigate in the lab over very short distances. So I used bumblebees and I made them navigate over very small distances about a metre and a half in a symmetrical arena, and I forced them to walk in the arena. I found that, while walking, bees were making these vector paths, these straight line paths to find home and food. Now for this study, I found with a control experiment that surprisingly it seemed that bees were able to remember the straight line paths that they had constructed during previous days of foraging and would use those paths to navigate home.

Chris - And you regard this as a long-term memory because they seem to preserve it over days at a time? They'll resort or relapse back to these same paths showing that they've stored, this is the connection between A and B that I follow. And they keep using that over a period of time?

Rickesh - Exactly, yes. So the working memory that I talked about before, we caused the working memory to give one behavioural readout: the bee, if it was using its working memory will go one direction, but if it's using its long-term memory it would go a different one. We found that we can predict when the bee goes in either of these two predicted directions, which gives support that they are actually using navigational vectors, these directions and distances they need to travel, that's stored in long-term memory. And they can recall this memory when they're at that familiar location in the arena.

Chris - Does the reason they operate like this reflect the fact that they do have quite a small brain? So it's kind of an optimal way of storing a lot of information. You've got these threads connecting position A and position B, position X and position Y, and if you've got all these things stored as simple vectors, it's very easy to store that information in a small number of nerve cells which they've got.

Rickesh - That's exactly right.

Chris - Is there anything we can learn from this about making better navigational tools or systems ourselves if they can do this and do it so well and so efficiently with a very small amount of information being stored? Can we steal that?

Rickesh - That is one of the goals of this avenue of research. So I have collaborators that I'm working with here at Lund who study the detailed neuroanatomy of insect brains in general, and they can identify which neurons communicate with other neurons and we can build a circuit map. And from this anatomical information we can construct models for how vectors might be built in the brain of these insects. So if we can figure this out, and if we have the behavioural evidence to support this, we can then potentially create robots that can do very similar navigational tasks as these insects with a very small circuit. These robots can be useful for situations in which perhaps what we currently rely on, like GPS, do not work. So for example, if you can create an autonomous robot to find people in a disaster zone like an earthquake or a war zone, they could potentially go in a cavern or in rubble, find someone and communicate where they are and find their way back to a rescue team, perhaps. This is just an idea, but there are potential applications in which we can take advantage of what evolution has built in these really tiny animals to do complex tasks really well.