Unexpected similarity between honey bee and human social life

Bees and humans are about as different as one can imagine. Yet despite their many differences, surprising similarities in the way they interact socially have started to be recognized in recent years. Now, a team of researchers from the University of Illinois at Urbana-Champaign, building on their previous studies, have experimentally measured honey bee social networks and how they change over time. They found that there are detailed similarities with social networks of humans and that these similarities can be explained completely by new theoretical modeling, which adapts the tools of statistical physics to biology. The theory, confirmed by experiments, implies that there are individual differences between bees, just as there are between humans.

The study, which for the first time measures the magnitude of individual differences in honey bee networking, was carried out by the first author, Ph.D. in physics. student Sang Hyun Choi, post-docs Vikyath D. Rao, Adam R. Hamilton and Tim Gernat, Swanlund Chair of Physics Nigel Goldenfeld and Swanlund Chair of Entomology Gene E. Robinson (GNDP). Goldenfeld and Robinson are also professors at the Carl R. Woese Institute for Genomic Biology in Illinois, of which Robinson is the director. The collaboration included experimental measurements of honey bee social behavior performed by Hamilton, Gernat and Robinson, with data analysis by Rao and theoretical modeling and interpretation by Choi and Goldenfeld. Their findings were published in a recent article in the journal Proceedings of the National Academy of Science.

“Originally, we wanted to use honey bees as a practical social insect to help us find ways to measure and think about complex societies,” Goldenfeld said. “A few years ago, Gene, Tim, Vikyath and I collaborated on a large project that put ‘barcodes’ on bees so that we could automatically monitor wherever they went in the hive, all directions in which they pointed to and each interaction partner. That way we could build a social network over time, called a time network. “

This study, carried out a few years ago, involved high-resolution imaging of honey bees equipped with barcodes, with algorithms detecting interaction events by mapping the position and orientation of the bees in the images. In these studies, researchers focused on trophallaxis – the act of mouth-to-mouth liquid food transfer – to measure social interactions between bees. Trophallaxis is used not only for nutrition but for communication, making it a model system for the study of social interactions.

“We chose to look at trophallaxis because it’s the type of bee social interaction that we can track accurately,” Choi said. “Since honey bees are physically connected to each other through proboscis contact during trophallaxis, we can tell whether they are actually engaging in an interaction or not. In addition, each honey bee is labeled so that we can identify each individual engaged in each interaction event. “

“In our previous work, we asked how long bees spend between events where they encounter other bees, and we showed that they interact non-uniformly,” Goldenfeld said. “Sang Hyun and I took the same dataset, but now we asked a different question: What about the duration of interaction events, not the time between interactions?”

Looking at individual interactions, the time spent varied from short interactions to long interactions. Based on these observations, Choi developed a theory where bees exhibited an individual trait of attractiveness that could be compared to human interaction. For example, humans may prefer to interact with friends or family members rather than strangers.

“We developed a theory for this based on a very simple idea: if a bee interacts with another bee, you can think of it as a kind of ‘virtual spring’ between them,” Goldenfeld said. “The force of the spring is a measure of how attracted they are to each other, so if the spring is weak, the bees will quickly break the spring and leave, perhaps to find another bee to interact with.” If the spring is strong, they can interact longer. We call this theoretical description a minimal model, because it can quantitatively capture the phenomenon of interest without requiring excessive and unnecessary microscopic realism. Non-physicists are often surprised to learn that detailed understanding and predictions can be made with minimal descriptive data. “

Goldenfeld explained that the mathematical framework for their theory came from a branch of physics called statistical mechanics, originally developed to describe gas atoms in a container, and since extended to encompass all states of matter, including including living systems. Choi and Goldenfeld’s theory made correct predictions on the honey bee experimental dataset that was previously collected.

Out of curiosity, the theory was then applied to human datasets, revealing patterns similar to those in the bee dataset. Choi and Goldenfeld then applied an economic measure of wealth and income disparities in humans – called the Gini coefficient – to show that bees display disparities in attractiveness in their social interactions, although not as different as humans. These results indicate a surprising universality of patterns of social interactions in honeybees and humans.

“It’s obvious that human individuals are different, but it’s not so obvious to bees,” Choi said. “Therefore, we examined the inequality in the activity level of honey bees in a way that is independent of our theory to verify that female honey bee workers are indeed different. Previous work in our group has used the Gini coefficient to quantify the inequality in honey bee foraging activity, so we thought this method would also work to examine the inequality in trophallaxis activity. “

“The discovery of such striking similarities between bees and humans has sparked interest in the discovery of universal principles of biology and the mechanisms underlying them,” said Robinson.

The researchers’ findings suggest that complex societies can have surprisingly simple and universal regularities, which can potentially inform how resilient and robust communities emerge from very different social roles and interactions. The researchers predict that their minimal theory could be applied to other eusocial insects since the theory does not imply characteristics specific to honey bees.

In future studies, the same techniques from statistical mechanics can be applied to understand community cohesion through well-characterized pair interactions, Choi and Goldenfeld said.

“It was my first project after joining Nigel’s group, and it took me a long time to find the right way to approach the problem,” said Choi. “It was fun and stimulating to work on such an interdisciplinary project. As a physics student studying biological systems, I had never expected to use concepts from economics.

“It was very exciting to see how simple physical ideas could explain such a complex and seemingly widespread social phenomenon, and also provide information about the organism,” Goldenfeld said. “I was very proud of Sang Hyun to have the persistence and insight to understand this. Like any transdisciplinary science, this was a really difficult problem to solve, but incredibly fascinating when it all came together. This is the kind of breakthrough that presents itself. from the co-location of different scientists in the same laboratory – in this case, the Carl R. Woese Institute for Genomic Biology. ”