Putting Goldfish in the Driver’s Seat

*AUTHOR BIO

Teaching fish to drive has an important lesson to teach us about how we “just know” how to move through our worlds. Teaching a fish to drive sounds like a bizarre, not to mention impossible, task. But researchers at an Israeli university have done just that.

Researchers at an Israeli university taught fish to drive in order to better understand how they—and by extension, all animals—navigate through their surroundings. ©Sachar Givon/Ben Gurion University of the Negev
Researchers at an Israeli university taught fish to drive in order to better understand how they—and by extension, all animals—navigate through their surroundings. ©Sachar Givon/Ben Gurion University of the Negev

In order to find food, shelter, and safety in a new place—in other words, to survive—animals must use navigation skills. Previous research has shown that mammals innately learn and analyze their environments, but what about fish? By studying if fish can learn to drive—rather than just swim—researchers can learn more about neural behavior in the brain: the innate processes, conscious and subconscious, by which animals intuitively move, pay attention, perceive, and reason.


Asking fish to drive around underwater seems odd. But researchers wanted to find out if the way fish learn to swim in a new tank comes from muscle memory—committing a specific motor task to memory via repetition—or if fish somehow consciously analyze and learn about their environment, as other animals do, to move safely through it, states Ronen Segev, a professor of life sciences at Ben-Gurion University of the Negev in Israel. He was part of a Ben-Gurion team that did the study.


The answer might mean that conscious analysis and learning is responsible for the way varied and distant species move through the world, he says. In other words, even small animals, like goldfish, might learn to move through their environment in a way that is similar to mammals.


“Understanding navigation is closely connected to understanding of learning and memory,” Segev says. “When an animal navigates, it needs to learn the structure of the environment and embed it in memory.”


The best way to determine how any animal learns to navigate is to have it move in a way it does not normally move and do so in an unfamiliar place, Segev believes.


The researchers sought to put the fish in an environment that—while underwater—resembles how they would move over ground. Could they maneuver through this new place? And if so, how did they learn how to do it?


To find out, a team of researchers taught fish to drive around an unfamiliar tank. See a video of the goldfish navigating (Courtesy Ben Gurion University of the Negev).


Fish Operated Vehicles


Rats learn to navigate with relative ease, and previous attempts to let them drive have been impressive. Dogs drove successfully in research studies too.

Dogs and rats navigate well in studies. ©Enrique Zavaleta/Pixabay
Dogs and rats navigate well in studies. ©Enrique Zavaleta/Pixabay

If fish could learn to drive, it could mean many animals can analyze their environment and learn to navigate their worlds. And if they can learn to move over land, it may mean the capability is more common and not so “species independent,” as the researchers put it.


As the researchers say: “Given their fundamental role and universal function in the animal kingdom, it makes sense to explore whether space representation and navigation mechanisms are dependent on the species, ecological system, and brain structures, or whether they share general and universal properties.”

"Are space representation and navigation mechanisms dependent on the species, ecological system, and brain structures, or are they general and universal properties?"

They chose to study how fish drive because asking fish to move in a way another species does is what is called “domain transfer methodology.” This means a species must cope with a familiar task, in this case, navigation, in the way another species would. Domain transfer is one of the best ways to study and understand animal navigation, Segev says.


The researchers developed Fish Operated Vehicles (FOV) that served as the “car” in their study. FOVs allowed them to study how a “fish out of water” behaves and moves, even as it is actually in water.


Over the Land, Through the Water


For the study, the researchers put the fish in an unfamiliar place and had them travel along the surface of the tank via the FOV—or “in a terrestrial environment,” as the researchers put it.


They also needed to provide a swimming site new to the fish. “We needed a way to get them away from the small water tanks we have in the lab,” Segev says.


They placed six goldfish, one at a time, in the FOV tank they had set up for the study. The FOV itself was actually a little robot that looked much like a clear skateboard with short sides. The goldfish were safely harnessed into their drivers’ seats—actually more like on the board—and given free reign.


The tank was equipped with LIDAR lasers that measured movement. A camera captured images from above, while a computer processed the information captured by the lasers and the camera. Robots mounted on a platform under the tank also helped trace the FOV’s movements. Moreover, to add another variable, the researchers arranged for light to enter the tank and reflect off the Plexiglass and into the water in a way the fish had never experienced before, Segev says.


“First, the fish need to learn new motor skills to drive the vehicle, since these are very different from the muscle power applied to fins to enable swimming,” he says. “Second, they needed to learn how to navigate the vehicle in an alien environment despite the significant distortion in vision.”


To encourage the fish, “we rewarded a fish when it managed to drive the vehicle towards a target on the tank wall,” Segev says.


At first, the fish moved about randomly. But after about ten to fifteen days of thirty-minute sessions, the fish drove the FOV directly to target, in this case their favorite fish food.


The researchers even changed up the environment by putting the fish in different starting positions and moving the target. They wanted to make sure the fish were not just driving the same route every time after they had learned it.

"The goldfish drove steadily to the target no matter its location. Even when they changed the way it looked the fish moved toward it."

The goldfish drove steadily to the target no matter its location. Even when they changed the way it looked, the fish moved toward it.


In other words, the fish had “internalized its target,” as the researchers put it. They recognized it, they took it into their brains and understood—through neurons firing in the brain—that they needed to head for it to get the reward.


By driving their FOVs, the fish showed they acted not merely from muscle memory but from real-time spatial analysis and navigation. The researchers theorized that the fish may have learned to navigate in the same way land animals like rats and dogs learned to drive.


As Segev says, “the way space is represented in the fish brain and the strategies fish use may be as successful in a terrestrial environment as they are in an aquatic one.”


“The finding hints that movement through space and across environments is the same for all types of animals,” he says. This study was only the first step. Next, the researchers will study the strategies fish use when navigating larger environments.


Those studies should lead to even greater understanding of how animals intuitively move about their worlds, finding food and shelter and avoiding danger, as well as expanding knowledge about how their brains work.


By demonstrating that fish can drive a vehicle across a surface towards a target, the researchers were able to find that the human and the fish brain have, perhaps, more in common than previously thought.

 

*Jean Thilmany is a freelance writer living in St. Paul, MN, who writes frequently about science and engineering topics.