A new scientific study shows that ants function as a “liquid brain” — a system with no “leader” or command centre, but rather a form of coordination, a collective intelligence that emerges from the dynamic interaction of many individuals. The research has demonstrated this empirically and developed a mathematical model that replicates the behaviour. It has also shown the importance of heterogeneity in individual movements — specifically exploration and exploitation — for efficient food gathering.
A team from the Centre for Advanced Studies of Blanes (CEAB-CSIC) has analysed both the individual and collective behaviour of the Mediterranean ant species Aphaenogaster senilis. These ants do not primarily communicate through pheromone trails that others follow, as is common in many ant species. Instead, they use movement and direct contact through their antennae. Additionally, the number of ants involved in food search is small, making communication more challenging. These characteristics make them especially interesting for deciphering the specific mechanisms that allow them to function as a single organism with a shared brain.
To study how they achieve such efficient cooperation, the research team designed a specific experimental setup (a structure significantly larger than those used in previous studies, maze-like, honeycomb-patterned, and equipped with a high-resolution recording system) that replicates natural conditions. They observed and recorded the behaviours — particularly the movements — of hundreds of ants, especially regarding food search and exploitation. Food was placed in various locations, and the researchers tracked how each ant behaved and moved, how they transmitted information to one another, and how these processes related to the overall efficiency of the colony. The study results have been published in the journal Proceedings of the National Academy of Sciences (PNAS).
The article presents key empirical evidence on the operation of the ants’ “liquid brain” — a form of intelligence with no centralized control. It shows that the diversity of individual movements, and the ability to adjust the proportion of different behavioural roles depending on environmental conditions, are essential to colony efficiency. In the food-gathering process, two main roles are identified: exploration (finding new resources — scouts) and exploitation (gathering found resources — recruits). The study has established an objective method to distinguish these behaviours and has precisely quantified their movements. Based on this data, the researchers created a neural model where ants activate depending on the frequency of contact with neighbours and move heterogeneously, reproducing observed patterns both individually and collectively. This model has helped demonstrate that the relative proportions of scout and recruit ants lead to different food-gathering dynamics and connectivity, modulating colony efficiency. The study suggests that this species may flexibly vary the proportion of scouts and recruits, adjusting the exploration-exploitation trade-off in response to environmental conditions — an essential adaptive feature for colony success.
Pol Fernández-López, CEAB-CSIC researcher and first author of the study, explains: “The key to being more or less successful in acquiring resources is adjusting the ‘mix’ between the two roles: individual autonomous behaviour, and above all, heterogeneous movement, is what enables dynamic, adaptive cooperation depending on what’s happening in the environment. With different movement proportions, ant colonies manage to optimize exploration (searching for new resources) and exploitation (using what’s already found) without any individual having a global view of what needs to be done.”
Frederic Bartumeus, also a CEAB-CSIC researcher and co-author, adds: “We’ve shown with real data that an ant colony works like a ‘liquid brain’. Each ant acts like a neuron, intermittently activated by contact with its neighbours. There is no ‘boss’, no one directing — coordination and intelligence arise from their connections. These connections are dynamic but also have spatial and temporal structure. This connectivity is an emergent, organized and dynamic property, depending on the movement patterns of scouts and recruits, and on how they integrate information from their neighbours.”
In summary, this study reveals that in “liquid” or diluted cognitive systems — like insect societies or the immune system — movement is key to keeping the system connected and coordinated. In ants, the two movement types play a crucial role in how information spreads and how quickly new food sources are found and exploited.
The study provides new, valuable insights into social insect collective behaviour and the nature of decentralized collective intelligence — knowledge that may be applicable beyond biology and ecology, in areas such as multi-agent robotics or internet search optimization algorithms.
