A rat navigating a maze has to rank somewhere near the top of science tropes. Now, scientists report that they’ve developed an analogous test for humans—one that involves driving through a virtual landscape in a simulated car. The advance, they say, may provide a more sensitive measure for detecting early signs of Alzheimer’s disease.

“I think it’s a very well-done study,” says Keith Vossel, a translational neuroscientist at the University of California, San Francisco (UCSF), who was not involved with the work.

In the rodent version of the so-called Morris Maze Test, researchers fill a large cylindrical container with water and place a platform just above the waterline. A scientist then places a rat into the tank, and the rodent must swim to the platform to avoid drowning. The experimenter then raises the water level just above the height of the platform and adds a compound to the water to make it opaque. The trial is repeated, but now the rat must find the platform without seeing it, using only its memory of where the safe zone exists relative to the tank’s walls and the surrounding environment. In subsequent trials, researchers place the rat at different starting points along the tank’s edge, but the platform stays put. In essence, the task requires the rat to move to a specific but invisible location within a circular arena from different starting points.

Previous work has shown that rats bred to develop symptoms that model Alzheimer’s disease perform worse at the maze. Specifically, rodents genetically modified to express human amyloid precursor protein (hAPP), which can lead to the debilitating plaques that form in the brains of Alzheimer’s patients, seem to struggle to find the hidden platform relative to their healthy peers. The test is useful in rats, but obviously there are ethical concerns when it comes to placing humans (especially those who we suspect may be prone to Alzheimer’s) into a tank filled with water where the only means of survival is to find a hidden platform.

So UCSF neuropsychologist Katherine Possin and her team at Gladstone Institutes in San Francisco created a Morris Maze Test for humans. Instead of a water tank, the human version relies on a simple driving simulator—a steering wheel and a gas pedal—connected to a monitor. Like the rat test before the water is raised, participants are first trained to drive to a specific location, marked by a lavender box, in a virtual countryside, which includes landmarks such as mountains, lakes, and the sun.

After the volunteers complete their training, the lavender box disappears. In its place, players are told, is a “buried treasure.” Users can’t see the treasure, but when they drive over it a notice pops up, letting them know they’ve hit the spot. “Every time they start from a new position, but the treasure is always buried in the same place relative to the trees and mountains,” Possin explains.

The researchers tested their virtual Morris Maze on people with mild cognitive impairments due to Alzheimer’s, which may manifest in the form of memory and cognition problems. Then they compared the results with the genetically modified mice, as well with healthy humans and mice.

All groups of both species showed improvement over the course of 10 to 12 trials, but the healthy participants consistently reached the target with a more economic route, the team reports on today in the Journal of Clinical Investigation. More importantly, within the afflicted groups, both the mice and humans had similarly poor performance in the hidden target trials, making the Morris Maze Test a useful tool for comparing our two species, the researchers say.

Performance on the Morris Maze Test can be scored in a variety of ways, but the scientists here opted to measure success by distance traveled, with the best scores being the shortest, reflecting a more direct path to the goal area. After each trial, the participants were ranked relative to one another and then all their ranks across all trials were averaged. “That’s basically it,” Possin says. “Then you can subject that to any statistical test.”

The human Morris Maze Test may give researchers a more uniform playing field for comparing human brains to the mouse models often used to test drugs and other therapies. The authors are hopeful that, because we can use a similar test across species, improvements will translate as well. In the past, a drug’s efficacy would often be measured with one test in mice and another in humans.

The field of Alzheimer’s research is riddled with stories of drugs that showed promise in mice, but failed in humans. Because of its sensitivity, Vossel thinks the new test may have been able to show small improvements that researchers may have missed in past drug trials.

“We currently use what I think of as sledgehammer approaches. They’re not very precise at learning about patient’s abilities to learn and remember information,” he says. “They mostly rely on pencil-on-paper tests that rely on verbal memory, for instance retelling a story or recalling a list of words. These kinds of tasks don’t seem to translate as well into real world function as something like navigation would.”

In a new test for Alzheimer’s disease, participants must navigate a 3D landscape in search of buried treasure.

DOI: 10.1126/science.aae0250