Spontaneous behavior predicts colony membership, individual identity, and rank in naked mole-rat societies

Animal societies are organized across scales, from moment-to-moment actions of individuals to dynamic group structure. A central challenge in sociobiology is to understand how behaviors of individuals shape group structure. Here, using machine vision to map recurrent fine-scale behavioral modules in the eusocial naked mole-rat ( Heterocephalus glaber ), we show that spontaneous behavior in isolation predicts group-level social organization, including colony membership and individual identity. We determined dominance relationships and found highly linear hierarchies, with ranks stable over months to years. Remarkably, fine-scale behavioral profiles in isolation correctly predicted pairwise dominance relationships, allowing reconstruction of rank order within each colony. Models trained within one colony predicted rank poorly in others, suggesting that naked mole-rats express social position through behavioral patterns specific to each colony. Together, our results connect fine-scale behavior to social structure maintained over years and show that, even when an animal is alone, its behavior carries signatures of its identity, colony, and place within that society.