As the 61st minute of the 2019 Women’s World Cup Final between the U.S. and the Netherlands began, Megan Rapinoe stood at the edge of the penalty box, stoically awaiting the referee’s whistle. An hour of attack and counterattack in the sweltering heat and under the anxious gaze of tens of thousands of fans had exhausted both sides but had yet to produce a goal for either. At the sound of the whistle, Rapinoe took a centering breath, trotted forward and skipped the ball into the back of the net, breaking the tie. As the stadium burst into exultation, Rapinoe headed for the sidelines; she had already taken 10 steps when her calm finally yielded to the unmistakable expression of pure joy. It was a beautiful moment and a reminder that while the spoils go to the winners, there are yet more powerful forces—in our biology, in our minds—that motivate us to play in the first place.

Playing is a universal human behavior and has therefore long been a subject of intense scientific interest. Nevertheless, because play is unprompted and natural—characteristics that do not usually lend themselves to laboratory work—much about its nature has remained mysterious. But in a thrilling study published recently in Science, experimenters concocted a work-around for this dilemma: they taught rats how to play a common childhood game. And in doing so, they made a series of discoveries suggesting that play is an even deeper part of our nature than previously thought.

How deep in our nature is play? It would be useful to begin by defining exactly what “play” is. Dutch historian and cultural theorist Johan Huizinga, in his now classic Homo Ludens, tried to do just that. Among other things, he argued that play must be voluntary: gladiatorial combat, in many cases, should be disqualified because its participants may have been forced into the arena. And play must occur in a space and time in which the rules are different from those in real life. Taking a time-out during a game is a way to leave that “magical circle.”

Play also needs to be internally motivated and should carry no material interest—players may grow stronger or faster, but play should not feed, clothe or pay them. In that sense, most collegiate athletes—at least for now—are still playing. Most important of all, play should be fun. To formalize this notion, if play serves some behavioral or evolutionary function, then the neural circuits of the brain involved in motivation and reward should be active during its occurrence.

What is the function of play? In making that assessment, it helps to remember that humans are not the only ones that do it. According to Homo Ludens, play predates human culture. “Animals have not waited for man to teach them their playing,” Huizinga wrote. He had a point: playing is a widespread behavior among animals, from dogs catching Frisbees to cats playing with, well, just about anything. Some types of play may involve learning to work cooperatively with a group for survival. Predators might engage in sparring or chasing games to simultaneously train and explore. Other types of play help animals learn how to follow complex rules, how to switch roles or even how to build a theory of mind. In general, games are critically important in establishing healthy social interactions, and failing to play them can result in inappropriate aggression, anxiety and social isolation.

Because of this role for playing in social learning, the most important games may be the ones we play when we are young. Take, for example, hide-and-seek—a game that has been passed down by oral tradition all over the world since ancient times. The fact that it is both ancient and widespread is an argument in favor of its importance. But hide-and-seek’s roots may lie deeper yet: even rats can play it. And true to the definition of play, they seem to do it just because they like it.

In an attempt to understand the neuroscience of play, a group of scientists trained rats to play games of cross-species hide-and-seek. In each game, the human experimenter began by placing the rat in a small box. If the lid of the box was closed, then the rat was the “seeker” and needed to “count off” in the box before setting out to find the experimenter, who had several objects to hide behind. If the lid of the box was left open, then the rat was the “hider” and learned to quickly leave the box to find a hiding place before the human experimenter began pursuit. In both scenarios, rats were rewarded only with social interaction.

During games when the rats were hiding, they preferred to be behind opaque rather than transparent objects, and they tended to stay very quiet until they were found—a commonsense strategy for those wishing to stay hidden—and basked in the accomplishment of besting a foe. On being found, rats often re-hid and awaited being discovered again, delaying their reward for playing. In contrast, during games when rats were seeking, they frequently vocalized and showed no preference for opaque versus transparent hiding places.

Rats also showed evidence for memory-guided search strategies. In some trials, experimenters returned to the same hiding place over and over again. Rats were quick to catch on and tended to find the experimenters faster in later trials, suggesting conscious access to game history. Finally, by recording from the brains of rats while they played hide-and-seek, the experimenters identified a series of neurons in the prefrontal cortex—a brain region associated with abstract coding of reward, motivation and rules—whose activities were correlated with specific phases of the game. Together, these results indicate that rats can learn the rules of hide-and-seek and that these rules have corresponding signatures in the their brain. In other words, hide-and-seek may be evolutionarily ancient.

These findings come as quite a surprise, which, in the world of research, means they have opened several new avenues of study. One primary reason for surprise is the complexity of the game the rats were playing. In a statistical sense, hide-and-seek could be thought of as a game of location inference. To be a good hider or seeker is to predict where the other individual will be hiding or looking and to exploit any prior knowledge (“Remy always hides in the chef’s hat”). Indeed, the rats showed evidence for such behavior by returning to previously used locations, suggesting this paradigm could be leveraged to understand how we make inferences about the actions and intentions of others. Knowing rats can play hide-and-seek should therefore motivate us to ask what the boundaries on game complexity are for different animals—and whether, by seeking out those boundaries, we might better understand animal intelligence.

Another reason these findings are surprising has more to do with our understanding of what could still be hiding in the human mind. As Huizinga might have said, play is freedom. By acknowledging that nonhuman animals can exercise a type of freedom that seems very human, this research may chip away at the idea of certain human freedoms as exceptional. As scary as that sounds, it moves us closer to understanding ourselves, which is the ultimate goal of neuroscience.

But perhaps the greatest surprise in this story comes from the boldness in how it was told. Scientific research often relies on strict experimental protocols, controls and constraints. There is a reason for this practice: by minimizing variability, experimenters can build high-confidence predictions about specific processes. But just as there are benefits to studying the unusual, there are benefits to studying the ordinary in unusual ways. By examining an act of freedom, this study has taught us much about being free. In doing so, it has attempted to corner an age-old question: Why do we play? The answer may have less to do with fame, glory, money or power than with Rapinoe’s rapturous smile as she jogged free of the penalty box this past July: we play because it is in our nature to do so.