Researchers at Lund University in Sweden have discovered that bumblebees have an incredible navigational memory. Rickesh Patel and his colleagues did it by tracking the walking behavior of foraging bees. Here’s what he told Chris Smith…
Rickesh – It’s been known for a long time that bees are very good at constructing straight paths from one feeder to their home and back to that feeder. And another thing that people have noticed is that bees are really good at constructing more complex navigation paths. For example, bees have been observed constructing a shortcut from one feeder to another even though they’ve never actually walked that path. Or so it’s thought. Also, bees fly to multiple flower beds in one flight and over time are able to make really optimized circuits or direct circuits between those feeders before returning home. So these complex navigation tasks require a little bit more than just following a single straight path back home. Again, bees can use landmarks to navigate, but it seems that this system gets better over time, which might suggest that maybe it’s these straight paths that are shared, which might explain how they can accomplish these kinds of tasks.
Chris – It sounds like you’re saying that bees have a lot of memories of places connected by a straight line, but they can link together these different memories, the different paths from one place to another, and basically create a map in their brain.
Rickesh – I think you can think of it as a web where there are these nodes, familiar places and these straight lines that connect these nodes. That’s just an idea. Nobody really knows if that’s true.
Chris – But what experiments have you conducted that have convinced you that this is probably the case?
Rickesh – To study how bees are able to keep track of these directions and distances in their brains, I had to construct a situation where bees were only using vector navigation, these straight-line paths, and not using visual features, which bees are also pretty good at. And since they travel for miles, I had to come up with a method for bees to navigate very short distances in the lab. So I took bumblebees and had them navigate very short distances of about a meter and a half in a symmetrical arena, and I forced them to run in the arena. I found that when bees ran, they formed these vector paths, these straight-line paths, to find their way home and to food. Now for this study, with a control experiment, I found that bees surprisingly seemed to be able to remember the straight-line paths they had formed during the previous days of foraging and used those paths to find their way home.
Chris – And you consider that long-term memory because they seem to retain it over a period of days? They refer back to the same pathways or fall back on them, which shows that they have stored them. That’s the connection between A and B that I’m tracking. And they use that over a period of time?
Rickesh – Exactly, yes. We tuned the working memory that I talked about earlier to provide a behavioral statement: if the bee uses its working memory, it will fly in one direction, but if it uses its long-term memory, it will fly in another direction. We found that we can predict when the bee will fly in one of those two predicted directions, which confirms that they are actually using navigation vectors, these directions and distances that they need to travel, that are stored in long-term memory. And they can recall that memory when they are in that familiar spot in the arena.
Chris – Is the reason they work like that because they have a pretty small brain? So it’s kind of an optimal way to store a lot of information. You have these threads that connect position A and position B, position X and position Y, and if you store all of these things as simple vectors, it’s very easy to store that information in the small number of neurons that they have.
Rickesh – That is exactly right.
Chris – If they can do it so well and efficiently while only storing a very small amount of information, can we learn something from this to develop better navigational instruments or systems ourselves? Can we steal it?
Rickesh – That’s one of the goals of this line of research. I’m working with collaborators here in Lund who are studying the detailed neuroanatomy of insect brains in general. You can identify which neurons are communicating with other neurons and we can create a circuit map. From that anatomical information we can build models of how vectors might be constructed in the brains of these insects. So if we can figure that out and have the behavioral evidence to support it, we might be able to build robots that can perform very similar navigation tasks to those insects using a very small circuit. These robots might be useful in situations where what we currently rely on, like GPS, might not work. For example, if you can build an autonomous robot that can find people in a disaster area like an earthquake or a war zone, it could potentially go into a cave or rubble, find someone and communicate where they are, and maybe find its way back to a rescue team. This is just an idea, but there are potential applications where we can take advantage of what evolution has enabled these tiny animals to do complex tasks really well.
