Mice Seek Information for Its Own Sake, Brain Study Shows

Aristotle opened his Metaphysics with a simple claim: all humans naturally desire to know. A new study suggests he was onto something larger than he knew. Mice, it turns out, want information too—and they want it badly enough to give up food and water to get it.

Researchers at Columbia University trained water-deprived mice to choose between two ports. At one port, they would smell odors that reliably predicted whether a sip of water was coming. At the other, the odors meant nothing. The water reward was identical at both locations. Yet when given the choice, the mice picked the informative port 78 percent of the time. They approached it faster, by an average of 220 milliseconds. They waited for the predictive cue before drinking, rather than lapping immediately. The behavior persisted even when researchers reduced the water payoff at the information port—the mice kept choosing it, so long as the shortfall stayed between 4 and 6 microliters.

This is not a small finding. It means mice value knowledge for its own sake, independent of any material gain. The work, published in Nature Neuroscience in July, extends a long scientific tradition stretching back to B.F. Skinner's operant conditioning experiments in the 1930s. Humans and other primates have shown similar information-seeking behavior in controlled settings. Now the mouse joins that list—and that matters, because mice come with tools that primates do not.

Jennifer Bussell, the postdoctoral neuroscientist who led the work in the Axel lab, spent considerable time convincing herself and her colleagues that they were actually measuring curiosity, not some confounding factor. "It took many iterations of the task," she says. The precision was necessary. Once she had it, the neural picture became clear. Using calcium imaging to watch excitatory neurons in the orbitofrontal cortex—a region involved in decision-making and value assessment—the team found that while some neurons responded to both information and reward, others lit up distinctly for one or the other. The brain, it seems, keeps these desires in separate accounts.

Joe Kable, a psychology professor at the University of Pennsylvania who was not involved in the study, calls this "the major contribution of this work." Knowing that mice value information the way humans do opens a door. "It makes it possible to study this particular cognitive process of information seeking in an animal model that has many, many, many more tools than the tools that we have to investigate this in humans and other primates." Ben Hayden, a neurosurgeon at Baylor College of Medicine who has conducted similar work in monkeys, echoes the sentiment. "You can do much better neuroscience with a mouse than you can with a monkey," he says, for the purpose of mapping neural circuits.

Bussell is already planning the next phase. She wants to use activity-dependent labeling and tracing—techniques that are straightforward in mice but difficult or impossible in other animals—to follow the neural pathways more precisely. She also plans to look beyond the orbitofrontal cortex at other regions like the medial prefrontal cortex and the insula. The question driving this work is fundamental: what is curiosity, at the level of the brain? Why do creatures seek to know? The mouse, it turns out, is willing to help us find out.