Do Ants Work Together Like Brain Networks?

By D.C. Demetre •  Updated: 07/28/22 •  6 min read

A new study claims that colonies of ants work together like brain networks of neurons. Dr. Daniel Kronauer, who studies evolution in insect societies, created a new method to analyze decision-making in ant colonies. Researchers learned that when a colony gets abandoned due to rising temperatures, the decision is related to the magnitude of the heat increase and the size of the ant group.

Every individual ant feels the rising heat beneath its tarsus but keeps calm and carries on until, abruptly, the ants reverse course. The entire group rushes out like one organism as a decision to evacuate has been made. It is almost like the colony of ants has one single collective mind.

The study hints that ants may fuse sensory input with an analysis of their colony’s population to make a collective decision. This would represent a process analogous to neural computations producing decisions.

Why Do Ants Work Together?

Ants are social creatures. They live in colonies, where each ant has a specific job. Ants work together to build a nest and forage for food.

They work together cooperatively to better defend themselves against threats from insects, animals, other ant colonies, and even humans. They work together as a team because evolution has decided that there is greater strength in numbers instead of working as individuals. They have had millions of years to get things right regarding survival.

One thing you may not know is that all worker ants are female – this sisterhood is responsible for the concordant operation of the colony. A young female ant’s final destiny to become a worker or queen is determined more by their diet than genetics. Any female ant larva can become the queen – those that do get diets richer in protein.

How Do Ants Decide

leafcutter ant
Decision-making is a series of computations meant to maximize benefits and minimize costs. For example, one common type of decision-making is called sensory response thresholding.

Sensory response thresholding is when an animal has to detect sensory input like heat over a certain level to react with particular energy- costly behaviour, like moving away. If the temperature rise isn’t significant enough, it won’t be worth the energy expended.

Kronauer and postdoctoral associate Asaf Gal developed a system to disturb an ant colony with precisely controlled temperature increases. They marked each insect with different coloured dots to track the responses of individual ants and the entire colony and recorded the movements with a tracking camera.

As expected, colonies of a set size of 36 workers and 18 larvae dependably evacuated their nest when the temperature hit around 34 degrees Celsius. This finding makes sense, Kronauer said, because if you become too uncomfortable, you leave.

But it was surprising that the ants were not simply responding to the temperature. When they raised the colony size from 10 to 200 individuals, the temperature needed to trigger the evacuation decision increased. Colonies of 200 individuals, for instance, held out until temperatures exceeded 36 degrees.

Ant Pheromones

Every single ant doesn’t know the size of its colony is changing, so how can their decisions depend on it? The researchers theorize that the explanation involves how ant pheromones work. Ant pheromones are the invisible messengers that pass information between ants, and they may just multiply their effect when more ants are present.

“What we’ve been able to do so far is to perturb the system and measure the output precisely. In the long term, the idea is to reverse engineer the system to deduce its inner workings in more and more detail,”

Kronauer said.

What Are Ant Pheromones Made Of?

Ant pheromones are trail pheromones, chemicals secreted by certain insects that serve as territorial markings to other insects and animals. Trail pheromones’ other purpose is to lead other ants to a food source. Other types of pheromones include alarm pheromones, aggregation pheromones, releaser pheromones and sex pheromones.

The exact chemical composition of the pheromone varies with the species of ant. Entomologists are just beginning to understand these chemicals, but one of the important ingredients in the pheromone trails leaf-cutter ants leave is a type of chemical called pyrazines.

Surprisingly, the pyrazine may not actually be produced by the ants, but instead, come from a bacterium named Serratia marcescens that lives in the leaf-cutter ant’s gut.

Is An Ant Colony Like A Neural Network?

The computing term “neural network” is actually not a very good model for how biological brains work. The difference between artificial and real neural networks has been unfortunately blurred in the media.

For one thing, neural network computing methods are highly data-intensive. For example in the task of computer vision, such as the ImageNet challenge, neural models train on classifying 1000 categories of objects using millions of training images. Human brain processing is much more data-efficient and can extrapolate from sparse data signals far more efficiently.

Also, neural networks are relatively poor at the transfer of knowledge. A neural network can attain expert status at a single simple video game, for example, but upon slight alteration of the game state, its capacity to play the game falls apart.

Kronauer approached understanding an ant colony as a “cognitive-like system that perceives inputs and then translates them into behavioural outputs,” Our frontal cortex, the portion of the brain used in both decision-making and voluntary movement, may have a recurrent neural network structure, to thank for those functions, a 2021 University of Wyoming study found.

“The biggest surprise is that RNNs not only exist in our brain, but they are constructed with much more delicate function and, yet, highly efficient in processing sequential inputs,”

said co-author Qian-Quan Sun. Artificial recurrent neural networks (RNNs) are incorporated into popular applications such as Siri, Google Voice Search and Google Translate and are deep-learning algorithms that are used for problems, such as language translation, natural language processing, speech recognition and image captioning.

So if we can say that our brain is somewhat like a neural network in some of its functions and that an ant colony is sort of like one brain when it comes to certain behaviour, then we could perhaps say that an ant colony is like a neural network, if we wanted to stretch it. It’s just another way of looking at how ants work together. But it only goes so far.

Sources:

  1. Asaf Gal, Daniel J. C. Kronauer. The emergence of a collective sensory response threshold in ant colonies. Proceedings of the National Academy of Sciences, 2022; 119 (23) DOI: 10.1073/pnas.2123076119
  2. Asquero. (Aug. 22, 2020) Advantages and Disadvantages of Artificial Neural Networks. asquero.com
  3. C. R. Smith, et al. Caste Determination in a Polymorphic Social Insect: Nutritional, Social, and Genetic Factors. The American Naturalist Volume 172, Number 4
  4. Silva-Junior, E.A., Ruzzini, A.C., Paludo, C.R. et al. Pyrazines from bacteria and ants: convergent chemistry within an ecological niche. Sci Rep 8, 2595 (2018).
  5. Wilson, E. O., Bossert, W. H. (1963). Chemical communication among animals. Recent Progress in Hormone Research, 19, 673-716
  6. Yihan Wang, Qian-Quan Sun. A long-range, recurrent neuronal network linking the emotion regions with the somatic motor cortex. Cell Reports, 2021; 36 (12): 109733 DOI: 10.1016/j.celrep.2021.109733