The demands of parenthood are so considerable that it’s fair to wonder why any adult takes on the challenge. Mammalian babies are especially helpless—and among mammals, only humans can see beyond individual sacrifice to understand a species’s survival depends on caring for its young.

Yet there is remarkable consistency in the way all mammals change their behavior upon becoming parents. Suddenly they are motivated to care for their young, and know how to feed and shelter, nurture and protect new babies. Parents also give up a lot of adult social interaction, whether it is mating with other mice or going barhopping with friends. “What this means is that there is this instinctive or genetically programmed aspect to the drive to take care of offspring,” says neuroscientist Catherine Dulac of Harvard University. But if a complicated and variable behavior like parenting is hardwired, how would that work? Reporting in Nature this week, Dulac, also a Howard Hughes Medical Institute investigator, and her colleagues have provided a wiring diagram of the brain-wide circuit that coordinates parenting behavior in mice. The study marks the first deconstruction of the architecture of a brain circuit underlying a complex social behavior.

The circuit they describe resembles the hub-and-spoke flight-routing system used by airlines and relies on a type of neuron that expresses the signaling molecule galanin. A relatively small number of these galanin neurons form a parenting command center—the medial preoptic area (MPOA)—in the hypothalamus, a brain structure responsible for controlling everything from appetite to sex drive. Responding to sensory input received from all over the brain, the neurons at the hub send distinct messages to at least 20 downstream subsets of galanin neurons. Like an airport terminal serving passengers according to their destinations, these subsets of cells, which the researchers dub “pools,” handle different facets of parenting behavior such as motor control of grooming or the motivation to parent at all.

Whether or not the same circuit exists in humans is unknown. However, both the hypothalamus, one of the oldest brain areas, and parenting behavior generally show remarkable consistency across species in spite of other differences. In scientific terms, they are evolutionarily conserved. “Because infants are very vulnerable, these behaviors are unlikely to be left to trial and error,” says Johannes Kohl, a postdoctoral fellow in Dulac’s lab and lead author of the new study. “That’s why we think that something similar might be happening in humans.”

The paper represents an important step in advancing our understanding of how social behavior is assembled in the brain, says biologist Lauren O’Connell of Stanford University who studies the evolution of parenting behavior but was not involved in this research. “This work has shown us where to look in the brain. It’s provided a path moving forward.”

A large body of work had previously identified individual brain areas involved in parental behavior including the medial preoptic area. Most of those studies were done by lesioning, destroying small bits of neural tissue and assessing the resulting behavioral changes. That technique left many unanswered questions about the particular cell types involved and how they were connected. In 2014 Dulac’s laboratory discovered the molecular identity of a population of galanin neurons in the MPOA and their role in parenting behavior. Using genetic and optical techniques to turn their activity on or off, the researchers changed the way adults responded to pups. For instance, activating galanin neurons in virgin male mice, who normally attack and kill pups, elicited grooming behaviors instead.

Having established a pivotal role for this set of neurons, the lab set out to answer the next question. “The big mystery was how only about 6,000 neurons out of 100 million in the mouse brain can orchestrate [parenting] behavior,” Kohl says. After hypothesizing the existence of a brain-wide circuit, they started to deconstruct it using what Dulac calls “a panoply” of modern neuroscience techniques. These included optogenetics, in which light activates particular neurons, and imaging of animals while they were awake and actively parenting.

By tracing the long fibers called axons that extend out from neurons in the hub region of the hypothalamus, the team closely investigated the function and anatomy of four of the 20 pools they had found. One set of axons projected to the periacqueductal gray in the midbrain, an area involved in motor behaviors. Manipulating that pool of neurons affected grooming behavior, although not the parents’ retrieval of pups. A pool in the ventral tegmental area (VTA), a reward-related region, controlled the motivation to engage in parenting tasks. When the VTA neurons were stimulated, both male and female mice worked harder to cross barriers placed between them and their pups. A third projection to the medial amygdala (MeA), a locus for emotion, inhibited competing social interactions with other adults such as male aggression against other males. The fourth projection led to the paraventricular hypothalamic nucleus, a key area for neuroendocrine control. This area modulated parenting-related hormones such as oxytocin, vasopressin and corticotropin-releasing hormone. “Even a complex behavior can be decomposed into its components and those components can then be mapped onto discrete, non-overlapping sets of neurons,” Kohl says.

Intriguingly, such a circuit is reminiscent of the way motor neurons control movement from the spinal cord. “When you walk or run or swim, each muscle needs to be controlled by a distinct pool of neurons. It’s the coordinated function of these different pools that [allows] animals to move in a fluid manner,” Dulac says. “Something like this is happening for parenting.” Now that such a pattern has been found twice, it is possible other brain circuits share the same logic. “This is a first model,” Dulac says. “I hope it’s going to be a motif found in other behavioral control. But who knows?”

Another significant discovery found that parenting circuitry was nearly identical in males and females, although it was not always active in males. The researchers concluded the galanin neurons perform different computations of inputs according to the animal’s sex and reproductive state.

Although it is a long way off, this work might someday have clinical applications. In a condition like postpartum depression, for example, the motivational component of parenting is distorted. Assuming a similar circuit exists in humans, Dulac says, “it would be interesting to see how these neurons are functioning in models of stress and depression and whether one could potentially intervene to help relieve those symptoms.”

For now this research adds considerably to our basic understanding of how the brain works, and how its connectivity relates to specific behaviors, says neuroendocrinologist Robert Bridges of Tufts University, a pioneer in the neurobiology of parenting. He added that this work stands out as the “most meticulous examination” to date of the workings of brain circuits dedicated to child-rearing.