How mice learn to think out of the box — a neuroscience surprise | Emory News

Skip to main content

const url = 'https://template.aws.emory.edu/shared-assets/emergency/alert-feed.php';

async function load() {<br>try {<br>const res = await fetch(url);<br>const text = (await res.text()).trim();<br>body = text;<br>$nextTick(() => {<br>$store.header.update()<br>$store.header.updateOffset()<br>});<br>} catch (e) {<br>console.error('Failed to load alert feed', e);<br>await load();<br>// poll every 60s<br>setInterval(load, 60_000);

})()"

Featured Series

Explore Emory’s Featured Series for inspiring stories, groundbreaking research, and compelling perspectives from our vibrant university community.

Explore All Series

$el.querySelector('input[name=\'search\']')?.focus(), 300)

Search

How mice learn to think out of the box — a neuroscience surprise

August 7, 2026

Carol Clark

Thinking outside of the proverbial box appears to require suppressing a brain region actively promoting an old habit. Photo: Getty Images, Bilanol.

The prefrontal cortex, a brain region long associated with learning more flexible behaviors, can sometimes block the ability to think outside the box, a new study in mice shows. Science Advances published the finding by biologists at Emory University.<br>“It’s a surprising result,” says Robert Liu, senior author of the study and Emory professor of biology. “We demonstrated that for a particular naturalistic behavior in mice, the prefrontal cortex blocks adopting a new and better strategy for solving a problem. Thinking out of the proverbial ‘box’ requires suppressing this executive decision-making part of the brain.”<br>The finding may provide insights into the mechanisms involved in human neurodiversity and some cognitive disorders. It adds nuance to the understanding of the role of the prefrontal cortex — which is often described as the brain’s management system, involved in controlling working memory, decision-making, flexible thinking and emotional responses.

Focusing on the moment<br>Animals, like people, often stick with what worked well in the past to achieve a goal — known as a win-stay strategy. When a smarter strategy comes along, ideally you learn to adopt it.<br>“One analogy is learning to play a game, such as poker,” says Kai Lu, first author of the Science Advances paper and a postdoctoral fellow in the Liu Lab.<br>New poker players, Lu explains, focus more on mathematical probabilities and thinking about the value of the cards in their hands and on the table. For optimal play, however, they need to learn to adapt their strategy during each game, based on subtle cues from other players, to factor in the possibility of a bluff.<br>While one strategy is based primarily on memory, the other relies more on external sensory cues.

Studying a natural behavior<br>The Liu lab investigates the functional, mechanistic, developmental and evolutionary origins of stimulus-elicited behaviors. It uses the stimulus of sound in the laboratory model of rodents, combining experimental, computational, and chemogenetic techniques.<br>Typically, laboratory experiments to study neural mechanisms for learning are geared simply towards making new associations by reinforcing success or punishing failure. The role of predispositions to guide a behavior are usually not considered.<br>The Liu lab wanted to home in on the neural mechanisms as a mouse learns from scratch to displace a well-worn decision-strategy with a more efficient one. They devised experiments based on a natural behavior of the mice — retrieving displaced mouse pups to bring them back to the nest.<br>“The female mice have a default strategy for searching for the pups, probably built up over evolution and experience, which is just to go back to where they last found a pup,” Liu says. “But they can also learn to do better by following a reliable sound cue that tells them where they can find the pup.”

Learning to shift a strategy<br>To explore the neural mechanisms behind shifting from this default strategy, the researchers conducted experiments using a T-shaped maze. An adult female mouse was placed in the “nest” at the base of the “T,” while an artificial sound played as a kind of beacon to attract the mouse to either the right or left arm of the “T,” signaling where the experimenter would deliver a pup to reward the correct choice.<br>As expected, the adult female mice would initially return to the arm of the “T” where they last found a pup — the default, win-stay strategy — regardless of where the sound was located. Over repeated trials, however, they showed a gradual strategy shift, learning to override the inefficient default and use the sound cue to correctly choose the sound side first and receive the pup faster. Half the cohort of 12 female adults in the experiments made this shift by day four, while all of them learned to use the sound by day eight.<br>The adult mice were implanted with silicon probes to their auditory cortex and to the medial prefrontal cortex, to allow the researchers to collect data...