A breakthrough discovery at the Kavli Institute at the Norwegian University of Science and Technology has uncovered a previously unknown function of grid cells, the specialized neurons that help the brain map space.

Discovered in 2005 by Nobel Laureates May-Britt and Edvard Moser, grid cells lay the foundation for building mental maps of your surroundings and for tracking your precise positions within these landscapes. As you move around your environment, your movements are being followed by grid cells on an internal map.

But now it turns out that grid cells also perform rapid, rhythmic sweeps into the space ahead of the animal. These sweeps act almost like an antenna, allowing the animal to probe the environment ahead of it. This revelation reshapes our understanding of spatial navigation in the brain.

Beyond the Brain’s GPS: A Hidden Pattern Emerges

Scientists have long believed that grid cells acted like a GPS pin, faithfully marking a location at any time.

But the Kavli Institute researchers found a much more dynamic process at play: Grid cells alternate between tracking an animal’s real-time position and scanning the environment ahead in a highly regular pattern—sweeping 30 degrees to the right, then 30 degrees to the left—at a rapid pace of ten times per second.

These rhythmic sweeps create a more efficient way to anchor locations relative to one another, providing a richer and more adaptable navigation system than previously imagined.

This breakthrough, made by researchers Abraham Zelalem Vollan, Rich Gardner, May-Britt Moser, and Edvard Moser, was published in Nature on February 3, 2025.

For decades, these sweeps remained undetected, trapped inside minuscule pockets of time in the grid cell data.

The reason? Technological limitations.

Past recording tools lacked the capacity and temporal resolution to show the rapid fluxes within the large-scale coordination of whole populations of grid cells in real time.

That changed with the introduction of Neuropixels 2.0—a revolutionary neurotechnology capable of recording thousands of neural interactions with millisecond precision.

“Previous research studies at the Kavli Institute have identified the components of the sense of place, such as cell types and neural circuits involved. What we wanted to find out was how the mental maps are used in real-time,” Gardner said.

He said that understanding the dynamics, processes, and changes that happen in the overall neural activity in the network can show how the brain actually uses this machinery for navigation.

To understand exactly what has going on, the researchers used a method called decoding. A sweep lasts 125 milliseconds and corresponds to what is called a theta wave in brain data.

The new Neuropixels probe can record interactions from thousands of cells across the network down to the millisecond level while the rat navigates in an awake state, is in REM sleep, or in deep sleep.

For each time block of brain data uploaded by the probe, the computer decoded which mental-map-location the rat’s grid cells were focused on. The computer then correlated the grid cells’ mental map location with the rat’s actual location in the physical landscape.

Using the new Neuropixels probe, researchers decoded how a rat’s mental map dynamically shifts during navigation.

To their astonishment, they found that the map did not perfectly align with the rat’s actual location.

Instead, the map was out of sync with the rat’s location in a very regular manner. Deep in the millisecond timescale, the grid cells sent waves of neural activity through a series of grid cells coding for neighbouring locations in an outwards movement.

The sweeps start powerfully with the grid cells coding for the rat’s actual location in space, sweeping forward to the right, die out at the furthest end of the sweep, before the sweeps emerge anew at the rat’s position, sweeping out to the other side.

These rhythmic sweeps suggest that grid cells do more than simply pinpoint the animal’s position—they actively explore the environment in real-time, continuously updating its internal representation of space.

“In the old research data, this entire dynamic is merged into the large, fat grid fields we know,” says May-Britt.

“When we subtract sweeping data from cell activity coding for the rat’s self-location, the grid fields become smaller and more precise,” she said.

Why do grid cells scan the environment in this particular way? And why the specific angle, length and alternating pattern?

The answer may lie in nature itself.

“Some bat species use echolocation to navigate, emitting sound waves in alternating directions to scan their surroundings,” said Edvard Moser. “This pattern is a strikingly similar to that of grid cell sweeps.”

He uses his hands to show how the coordinated grid signal shoots out from the forehead like focused spotlights, alternating to the right and left.

The length of the sweeps could be explained by previous findings: our mental maps are doughnut-shaped.

“Sweeps extend up to half a loop around the doughnut-shaped map, and never beyond. This way, grid cells avoid overlapping into other areas of the mental map,” Abraham Vollan said.

The distance covered by sweeps is thus limited by the brain’s own doughnut-shaped rule for how grid cells are allowed to act within the brain’s GPS.

The brain has at least three to five different modules of grid cells in the brain, corresponding to three to five different doughnut-shaped mental maps, each tasked with covering different scales of space.

The sweeps travel the same distance on all doughnut-shaped maps. But since the mental maps relate to physical space in different scales, the grid cells that build our large-scale maps cover larger distances in our physical landscape than grid cells that generate more precise small-scale maps.

To find the explanation for why the brain prefers to scan the surroundings with these narrow antenna-like sweeps, Gardner built an artificial agent (a computer model based on artificial intelligence).

The small”robot” tested different ways to map an area it moved through and found that the optimal strategy followed a characteristic herringbone pattern.

It arrived at exactly the same rule as the sweeps in the rat brain: right, left, right at a 30-degree angle. This strategy turns out to be the most efficient for mapping an area in the shortest possible time with minimal overlap.

“If we look at echolocation in bats, antennae, whiskers, or our eyes located on each side of the skull, we recognize a principle where these two alternating angles for sampling or probing space recur,” Gardner said.

The computer model arrived at exactly the same principle for optimal mapping.

“It makes sense to think that evolution would have positioned our sensory organs and mental faculties in a way that exploits this principle,” he said.

The researchers found sweeps both when the rat navigated and in REM sleep when the rat did not receive sensory input from the outside world.

“Maybe it’s navigating in its dreams,” Edvard Moser said. During deep sleep, a state where the brain does not generate theta waves, the sweeps were rarer and more irregular.

“The sweeps also probe over cliffs and through walls, so it’s clear that the signal is not primarily about where the rat plans to go,” May-Britt Moser said.

“The vector principle in the sweeps suggests that the sweeps may be a way to build more robust maps. The precise scans contain very systematic information about direction and distance to places and may be a way for the brain not only to focus on individual locations, but also to relate and anchor these places to their surrounding environments,” Gardner said.

Edvard Moser said that the results suggest that the sweeps could be a fundamental mechanism hardcoded into the network – a kind of brain algorithm.

“What I think we have found is a repetitive and stereotypical process that continuously occurs in the brain’s mental maps, helping to map the surroundings of a rat running around,” Vollan added.

This map may be important for recalling memories of the rat’s surroundings and for creating new maps of environments the rat has not been in before, he said.

The study was conducted in rats, but might a similar mechanism exist in humans?

Given that we share the same brain structures, cell types, and rhythmic neural processes, the researchers believe this is likely.

“I believe you will find something similar to sweeps in humans too,” Vollan said.

“We have the same brain areas, cell types, the same functions such as memory and navigation, and we also have this rhythm, albeit a bit weaker. So, the question is whether we will see the exact same pattern? Humans, for example, are more visually driven. We can use our gaze to explore places at a distance,” he said.

Perhaps the sweeps in humans are more closely tied to where an individual directs their gaze in space.

Researchers have found visual place cells in both birds and monkeys, which are activated depending on where they look.

So, the place cell fires not where the animal is, but where the animal’s attention is focused.
“Sweeps are a fundamental mechanism at the cell population level that can begin to explain this focus,” Vollan said.

“Previously, we have shown that grid cells have a geometry in space—the hexagonal coordinate system. They also have a geometry in network activity—the donut. And now we have found that grid cells also have a geometry in time,” Edvard Moser said.
Here’s the hexagonal coordinate system:

“The sweeps change our idea of grid cells and navigation in the sense that we now know that grid cells don’t just code for the current position in isolation but also relate these positions to each surrounding position, precisely coded into the mental map,” Edvard Moser said.

Many questions remain unanswered, and the Kavli researchers do not intend to let go of the sweeps anytime soon.

“We are already working on several studies that will provide new answers to the things we are itching to find out,” May-Britt Moser said.

Seven long years of tough research were rewarded with a champagne toast and a publication in Nature.

“Our mentor Per Andersen taught us to celebrate our victories together. We have brought that tradition with us to the Kavli Institute,” said May-Britt Moser.

“Today we celebrate not only excellent research but also a good research collaboration. It has been a pleasure to work with such fine, generous, and talented young researchers as Rich and Abraham,” she said.