Navigating surprises in neuronal maps

By Nora Brackbill

We’ve all felt it — you’ve been walking forever, but your favorite restaurant is still blocks away. Other times, it sneaks up on you, and you’re surprised at how quickly you got there.

What happens when our sense of movement disagrees with external landmarks? In a recent paper in Nature Neuroscience, Stanford researchers found that our internal maps adjust to match those landmarks, as long as they’re roughly where we expect them. But if they’re too far off, we throw the whole map out.

“Navigation is something that pretty much all animals have to do to survive, so it’s a really basic, essential skill. At the same time, it’s a very interesting computational problem,” says lead author Malcolm Campbell, a graduate student working in the lab of Lisa Giocomo, assistant professor of neurobiology.

Giocomo’s team studies neurons in the brain that help us navigate, relying on a combination of landmarks and sense of movement. One type is called a grid cell, because it responds when we are near specific places in our environment, and those places form a grid across the entire environment. Each grid cell has its own, slightly different grid of locations that it responds to, creating a coordinate system for our internal map.

In a previous study, her team found these grids tend to drift and become unreliable over time, but an encounter with a landmark can reset them. Imagine walking down a hallway — you might lose track of how far you’ve gone until you pass something you know, a landmark like a door or a poster. But posters can move, so Giocomo wondered how these cells would respond if a landmark was not where it was expected.

The researchers recorded activity of grid cells in mice placed in a virtual reality environment. “Everyone always asks whether we put tiny goggles on mice. I wish that were the case, but no. We have mice running on a treadmill which is surrounded by computer screens. The mouse’s running causes it to move down a virtual hallway, like in a video game,” Campbell says. They could make landmarks appear earlier or later than expected based on the distance run on the treadmill.

Surprisingly, when the landmarks appeared later, the grid cells responded to entirely new locations. The mice were mostly using locomotion to estimate their own speed and movement on the treadmill, so the landmarks appeared far from where the mice expected, and the grid cells responded as if they were in a new environment. But when the landmarks flashed by earlier than expected, the mice relied on those exciting visual cues (rather than locomotion) to estimate their movement on the treadmill, something the researchers confirmed by recording different neurons that track the mouse’s speed. That estimate was more consistent with the landmarks, so the grids could shift to match. As Campbell describes it, “there was this transition between ‘oh, ok, I still roughly get it, I am going to shift around to accommodate’ versus ‘I give up entirely.’”

This flexibility explains how we can know if our sense of our own movement is just a bit off or if the landmark is gone and we are in a new place all together, and not get hopelessly lost in either situation. Navigation is an essential computational skill that is closely related to memory and learning, and getting lost is an early symptom of Alzheimer’s, so it’s important to understand how it works. The good news is that if our favorite restaurant moves, we will be able to navigate our way to more food! We might just be a little hungry on the way.