Animals follow hidden algorithms, whether they are Meerkats or Hyenas

Mammals with very different lifestyles appear to follow very similar behavioral patterns, according to a study published this week in the National Sciencesmemy Academy of Sciences, in a study published this week.

Surprising findings suggest that animals can be organized how to sequence their daily activities regardless of species, environment, or individual differences.

Scientists at Max Planck’s Animal Behavior and Partners used an accelerometer (the same motion sensing technology found in smartphones) to track three different mammalian species in their natural habitats to record their movements in detail for extended periods.

A normal pattern appears, scientists expect differences

“We think there will be differences,” said Pranav Minasandra, a postdoctoral researcher at MPI-AB and lead author of the study. After all, the obvious differences are obvious when comparing Meerkats, Coatis and Hyenas that occupy different environmental and ecological roles. “But we find common patterns in how animals switch between behaviors, regardless of species and individual. Their behavior seems to be established on the same hidden algorithm.”

What makes this discovery particularly interesting is how it contradicts traditional animal behavioral wisdom. Most researchers (and most of us indeed) hope that animals will be more likely to change their activities the longer they do something. On the contrary, the opposite is true.

Surprising “lock” effect

The most compelling pattern identified is what researchers call “decreased hazardous function,” which actually means that the longer the animal maintains a specific behavior, the less likely it is to change in the next moment. This self-reinforced behavioral momentum was consistent in all the animals studied.

“This was unexpected,” Minasandra added.

Imagine a hyena walking for 10 minutes in a row. Most people may guess that the hyena will stop over time, and the authors do the same. “We initially thought that the probability of switching behavior increased over time because we thought locking in any behavior was not the best.”

How researchers capture hidden patterns

The research team equipped wild animals with accelerometers to track their movements with excellent accuracy. Research includes:

• Meerkats: Small, burrowed social mammals in the Kalahari Desert
• Coat: Raccoon-sized tree residents in Panama rainforest
• Hyena discovery: Large carnivore roaming Kenya’s savanna

Using machine learning algorithms, researchers transform raw motion data into behavioral states, such as lying down, foraging, or walking. This approach allows them to construct detailed behavior sequences spanning days or weeks.

“This approach allows us to capture detailed behavioral sequences of days or even weeks in a few days or even weeks,” said Ariana Strandburg-Peshkin, a team leader at MPI-AB and senior author of the study.

Predictable unpredictability

The researchers also looked at how current behavior predicts future behaviors—which they call “predictive attenuation.” As expected, the further the prediction accuracy decreases in the future, people try to predict. But how does this predictiveness disappear? The pattern of attenuation follows the mathematical form of all studied animals very consistent.

The authors further examine how current behavior predicts future actions—the concept they call “predictive decay.” Predictive decay reflects the increasing difficulty of predicting behavior, and we further understand the future, mainly due to random, unpredictable changes. The shape of the attenuation map conveys how decision systems at different time scales interact to produce a sequence of behavior in animals. “We found that the pattern of predictive attenuation was very consistent in all the animals studied, meaning shared structures under the surface.”

Why does this similar pattern exist?

What can explain these common behavioral structures in these different species? The researchers proposed two main possibilities:

The first is positive feedback: the longer the animal is in a state (e.g., lying down), whether it is warm, safe or strengthened in society, there will be rewards. Behavior becomes self-reinforced.

The second possibility is multi-time decision making. Animals can be combined with their own rhythm, rather than a single internal clock managing when behavior changes, but blending cues from many processes (internal hunger, external threats, social environments). The interaction of these overlapping signals may result in the observed patterns.

Can these common behavioral structures explain the broader patterns we see in nature, such as the heavy distribution of animal movements known as Levi flights? The authors believe this is possible and has potential implications for how we understand the fundamental aspects of animal behavior.

More questions than answers

This study raises many questions. Do non-social animals show the same pattern? What about different developmental stages or ecological stress? Do these behavioral structures actually give advantages to survival or reproduction?

Regardless of the answer, the study suggests the profoundness of the behavioral organization throughout the animal kingdom.

“This study shows that real animals, whether hunting, hiding or resting, are guided by hidden structures that seem to echo in the branches of life,” said Meg Crofoot, co-author of Meg Crofoot, director of the Department of Animal Social Ecology.

The complete study is available in the Proceedings of the National Academy of Sciences (DOI: 10.1073/pnas.2503962122), published on May 15, 2025.

Related