Nature Podcast: Naked Mole Rat Queen Pheromone and the Intuitive Gamer Model for New Board Games

The Nature Podcast examines two distinct lines of science this week. First, researchers uncover a single chemical cue isopropyl myristate produced by naked mole rat queens that triggers reproductive suppression in workers, helping explain how these colonies maintain a single breeding queen. The second story introduces a cognitive science study on how people reason about unfamiliar games for the first time, proposing an intuitive gamer model to predict behavior and inform AI design. The episode highlights how biology and cognition can illuminate social structure and decision making in humans and other species.

• Isopropyl myristate identified as a queen-specific cue influencing worker reproduction
• Queenless colonies show reduced conflict when the cue is supplied, but conflict returns when it is removed
• New theoretical model of human first-time game reasoning called the intuitive gamer
• Implications for AI that reasons under limited compute and for human-AI collaboration

Naked Mole Rat Queens and a Hidden Social Rule

The episode opens with Sharmini Bundell reporting on naked mole rats, extraordinary social rodents that live in deep tunnels and exhibit a striking caste system. In these colonies there is one reproductive female, the queen, backed by a breeding male, while workers perform foraging, defense, and care for the queen’s offspring. The central mystery has been what social dynamics enable only one queen to remain reproductively active. The investigation centers on the queen’s odors rather than genetic determinants. A team led by Mohammed Khalaf at the Max Delbrück Center collected a wide array of smells from queens and workers across colonies. Amid this search they isolated a compound uniquely produced by queens: isopropyl myristate, a molecule that is volatile and surprisingly inexpensive to acquire. When researchers introduced this cue to queenless colonies, aggression decreased and the previously peaceful social order remained stable. When the cue was withdrawn, fighting resumed and high-ranking females jostled for the queen position. Brain imaging revealed that isopropyl myristate activates olfactory regions in workers, acting as a reproductive signal that suppresses ovarian activity. The finding is described as a potential social “holy grail” for naked mole rat researchers, because it reveals a concrete chemical cue controlling reproductive suppression in a mammal, a mechanism that in bees is mediated by queen pheromones but in naked mole rats was previously unclear.

The discussion emphasizes that this cue is not a genetic trait and does not solely rely on pheromones like those in urine, which complicates analogies to bees. The researchers propose that pregnant workers producing the cue can suppress reproduction in others, creating a social order that persists through pregnancy and maternal signals. The discovery invites broader questions about how such cues evolved and are regulated at genetic and social levels, and whether similar mechanisms exist in other mammals or even humans. The interviewee Melissa Holmes notes that the work opens windows into animal communication and social organization, potentially broadening our understanding of how social species share information beyond the naked mole rat itself. The segment ends by noting that follow-up research will probe evolutionary origins and the genetic controls of this mechanism, as well as possible implications for human social behavior and interspecies communication.

The Intuitive Gamer: How People Think About New Games

The second half of the episode shifts to a study in cognitive science and AI. Katie Collins from MIT presents two aims: (1) to understand how people reason about a game when they are novices, and (2) to understand how people decide whether a game is worth playing in the first place. The researchers designed experiments in which participants read descriptions of games that modify classic rules of Tic Tac Toe or Connect Four, such as requiring two in a row to win on a larger board. Participants were asked to predict outcomes such as which player would win or whether the game would end in a draw, and a separate group judged how fun the game might be. The project uses a computational cognitive framework called Stimuli Code Design to iteratively build and test predictions against human data, in a back-and-forth loop of hypothesis, piloting, and refinement.

The central contribution is a model dubbed the intuitive gamer. This model posits that when faced with a new game, people engage in rapid mental simulations that are fast, probabilistic, and goal-directed, while imagining interactions between the players. The agents are not treated as perfect rational optimizers, but as boundedly rational players whose actions reflect simple, structured reasoning under time and cognitive constraints. The intuitive gamer framework successfully captures how people judge new games and how they play them in initial rounds, offering a bridge between human decision making and AI planning. The study suggests that people have a reusable intuitive theory of games, offering explanations for why people sometimes deviate from traditional game theory’s prediction of fully rational behavior. The researchers stress that their model demonstrates systematicity in novices’ reasoning, a kind of structured thinking that AI can leverage for faster, resource-conscious planning.

The discussion then turns to AI implications. The intuitive gamer approach implies that AI systems could operate with limited compute by emulating a bounded reasoning style that still yields effective judgments about new problems. A future AI could serve as a thought partner, simulating human expectations and offering predictions about when a plan is likely to be fruitful. The researchers also acknowledge a broader challenge: bridging the gap from simple, well-defined games to richer environments like The Sims, which involve open-ended goals and creative play. They see this bridging as a critical area of future work: understanding how people invent and reason about new game rules and how to model that creative process within AI systems.

The episode closes with reflections on the potential to extend these cognitive models to longer-term decisions and open-ended environments, and with a nod to the show notes for further details and links. The two stories together illustrate how the neural and social mechanisms underlying animal behavior and human cognition illuminate broad questions about decision making, cooperation, and the design of intelligent systems.