Bartal expected to see this activity in the rescuer rats, because human empathy appears in these areas. But she was surprised that even the ones that didn’t rescue their cage mates showed the neural traces. “The rats actually process the fact that there’s a rat in distress—that he’s trapped, that he’s unhappy,” she says. “And they activate this empathy system, whether they help or not.”

If that same machinery fires in all cases, but the behavior between in-group and out-group pairs differs, what gives? The difference seemed to lie elsewhere, including in the nucleus accumbens, which deals in carrot-and-stick-type neurotransmitters like dopamine, serotonin, and GABA. “It’s active when you eat something yummy, or when you win money, or have sex,” Bartal says.

It’s often called the brain’s reward center, she adds, “but today, there’s more understanding that it’s not as simple of a picture.” A newer view of the nucleus accumbens’ dopamine ties it to anticipating a reward and motivating its pursuit. “The brain’s main function is to get you to approach stuff that’s good for your survival, and avoid stuff that’s bad for your survival,” Bartal says.

She repeated her experiment to focus on this area using a method called fiber photometry, which let her team monitor neural chatter in living rats. They injected the animals’ accumbens with genetic material that made the neurons fluoresce whenever a synapse spiked. Then they implanted fiber-optic strands to observe those bursts of light while watching the rats scurry around. And indeed, the rats who freed their roommates showed the most activity in the nucleus accumbens. Signals of that activity peaked just as they approached to open the door with their snouts. This told Bartal that, for the free-roaming rats, the salient moment was releasing the restraint, rather than playing with their friend.

Bartal lastly wiretapped the rats’ nucleus accumbens with a dye that traces where electrical signals originate. She wanted to find where that motivation to help first arises. (If a hungry rat searches for pizza in a New York subway, their gustatory cortex would page the accumbens.) By taking brain slices from the animals shortly after they performed the rescue task and observing which regions the dye had reached that overlapped with c-Fos-expressing pockets, she could tell which parts of the brain had been talking to one another.

Bartal traced the calls into the motivational hub during the rodent rescue missions and found a caller she recognized: the anterior cingulate cortex. She suspects this points to a line of communication between empathy and reward that could be important for understanding compassionate behavior. But it’s still too early “to completely outline the entire microcircuitry that’s involved,” she says. “That’s what we’re working on now.”

“This is a fantastic study,” writes Stanford University neurobiologist Robert Sapolsky in an email to WIRED. Sapolsky, who was not involved in the study, wrote the book Behave: The Biology of Humans at Our Best and Worst, which describes what motivates human behavior—namely ubiquitous categorizations of “us” versus “them.”

The team’s results tell us tons about ourselves, according to Sapolsky, because experts would predict identical results in human brains: an us/them distinction, an anterior cingulate making demands, and the accumbens fueling motivation. Running such detailed brain experiments would be untenable in humans, and showing that this plays out in rats offers a bittersweet message, he feels. The good news, Sapolsky writes, is that “the roots of our ability to help, to empathize, is not the product of Sunday morning sermons. It’s older than our humanness, older than our primateness; its legacy long predates us as a species.” The bad news is that our tendency to them-ify those around us is also ancient.