🧬 biology

Spatial proximity heuristics structure the foraging behavior of Callithrix jacchus in the Caatinga: evidence for spatial cognition in a Neotropical primate

This study demonstrates that *Callithrix jacchus* marmosets in the Caatinga utilize low-cost spatial proximity heuristics, rather than complex cognitive maps or Lévy walks, to optimize their foraging trajectories, with seasonal phenology playing a secondary, complementary role.

Original authors: Gessica Rafaelly Dantas da Silva, Igor Eloi, Arrilton Araújo

Published 2026-09-08

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Gessica Rafaelly Dantas da Silva, Igor Eloi, Arrilton Araújo

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Every day, animals must solve a complex puzzle: where to go next to find food. For primates living in forests, this is not just a matter of wandering until something is spotted. It requires a mental map, a way to remember where resources are, when they are available, and how to get there efficiently. Scientists have long debated how these animals navigate. Do they rely on a sophisticated internal map of their world, or do they simply follow simple rules, like always heading toward the nearest known food source? Understanding this helps us see how the brain of a small animal can solve problems that seem to require complex planning. It also reveals how the environment shapes the way an animal thinks, forcing it to choose between remembering every detail or using a quick, energy-saving shortcut.

In the dry, thorny scrublands of Brazil's Caatinga region, a team of researchers set out to watch how common marmosets, tiny monkeys about the size of a squirrel, solve this daily puzzle. These monkeys live in a harsh environment where food is scattered and seasonal. The researchers wanted to know if the monkeys were just moving randomly, or if they were using specific mental strategies to get from one feeding spot to another. To find out, they followed three different groups of wild marmosets for nearly six months. They used GPS trackers to record the monkeys' movements every minute and watched them closely to see exactly where they stopped to eat. By mapping out these feeding spots, the scientists could see the paths the monkeys took between them, turning raw movement data into a story about decision-making.

The first thing the team checked was whether the monkeys' movements looked like a random walk, a pattern often seen in animals that are searching blindly for food. They found that the monkeys did not move randomly. Their paths were not the erratic, short steps of a blind search, nor did they follow the specific mathematical pattern known as a Lévy walk, which some scientists had predicted might be used by animals in sparse environments. Instead, the monkeys moved in a way that suggested they knew where they were going. They were not just stumbling through the brush; they were making choices.

To understand what those choices were, the researchers built a series of computer simulations to test different theories. They compared the real paths the monkeys took against what would happen if a monkey had no memory at all and just picked a random spot to go to next. The results were clear: the monkeys were far better at choosing their destinations than a random guesser would be. The monkeys consistently chose to travel to the nearest available feeding spot. This simple rule, known as the nearest-neighbor rule, explained the vast majority of their movements. It turned out that these small primates were not trying to calculate the perfect route to visit every single food source in their territory. Instead, they were using a low-cost strategy: look at the map in your head, find the closest place you know has food, and go there.

The study also looked at whether the monkeys remembered not just where the food was, but when it was available. In the Caatinga, some plants fruit only at certain times of the year. The researchers tested if the monkeys would ignore a nearby spot if they knew it was out of season, or if they would stick to the nearest one regardless. They found that while the monkeys did have a slight awareness of the seasons, it was a minor factor. The distance to the food was the most important thing. Even if a more distant spot might have had better food, the monkeys almost always chose the closer one. This suggests that for these small animals, saving energy on travel is more critical than finding the absolute best meal at any given moment.

This research changes how we view the intelligence of small primates. For a long time, scientists thought that only large-brained animals, like chimpanzees or gorillas, could use spatial memory to navigate their world. This study shows that even the smallest monkeys in the Americas use a form of spatial memory to guide their daily lives. They do not need a perfect, complex mental map of their entire home. Instead, they rely on a simple, efficient rule that works well in their specific environment. By choosing the nearest known resource, they save energy and survive in a difficult landscape. The findings suggest that this kind of smart, simple thinking is likely common among many small primates, allowing them to thrive without needing to be the biggest or the most complex thinkers in the forest.

Technical Summary: Spatial Proximity Heuristics in Callithrix jacchus Foraging

Problem Statement
Understanding how primates decide where to move during foraging is central to elucidating spatial cognition. While Optimal Foraging Theory (OFT) provides a normative framework for maximizing energy gain, it often fails to specify the cognitive rules or spatial representations animals use to navigate heterogeneous environments. A critical gap exists in distinguishing between stochastic movement patterns (e.g., random walks, Lévy walks) and the underlying decision-making processes that generate them. Specifically, it remains unclear whether the foraging trajectories of small-bodied Neotropical primates, such as the common marmoset (Callithrix jacchus), are driven by complex spatial memory and planning or emerge from simple interactions with the immediate spatial distribution of resources. This study addresses the question of whether C. jacchus displacements are organized by decision rules consistent with accumulated spatial information or if they result from search processes shaped solely by immediate resource distribution.

Methodology
The research was conducted over 167 days (April 2024–March 2025) in the Floresta Nacional de Açu, a semi-arid Caatinga environment in Brazil. Three wild social groups of C. jacchus were tracked using GPS units recording locations every 60 seconds.

Data Processing and Patch Definition:
- Feeding events were recorded via focal animal sampling.
- Foraging patches were algorithmically delimited using the DBSCAN (Density-Based Spatial Clustering of Applications with Noise) algorithm with a 30-meter search radius. This radius accounted for GPS positional dispersion and the functional size of resource units.
- Movements were defined as transitions between the centroids of distinct, consecutive patches, filtering out within-patch movements to isolate directional travel.
Statistical Modeling of Movement Patterns:
- The study first tested whether displacement lengths conformed to stochastic models: Brownian motion, Lévy walks, and log-normal distributions.
- Model fitting utilized maximum likelihood estimation and log-likelihood ratio (LLR) tests, with goodness-of-fit assessed via bootstrap resampling.
Heuristic Process Modeling:
- To infer the decision rules generating the observed patterns, the authors compared actual displacements against "virtual models" representing varying levels of cognitive complexity:
  - Mere Random Walk (MRW): A null model representing a forager without spatial memory, selecting destinations randomly from available patches.
  - Nearest Neighbor (NN): A model simulating a forager with spatial memory that moves to the closest unvisited patch.
  - Temporal NN: An extension of the NN rule incorporating a phenological constraint (selecting only patches visited within a 30-day window).
  - Gravity Rule: A model weighting destination choice by a ratio of patch attractiveness (feeding history) to distance ( $i/d^b$ ).
- Predictive accuracy was evaluated using prediction errors (absolute difference between observed and expected distances) and Cohen's $d_z$ effect sizes.

Key Results

Movement Distribution: Displacement lengths exhibited heavy-tailed distributions that did not conform to a classical Lévy walk. The power-law model was rejected in favor of log-normal and exponential models. This suggests movement is multiscale and bounded by spatial constraints rather than scale-invariant.
Superiority of Spatial Memory: The Nearest Neighbor (NN) heuristic significantly outperformed the null Random Walk (MRW) model. The median prediction error decreased from approximately 217 meters (null) to 39 meters (NN), with a large effect size ( $d_z = 1.15$ ). This indicates that marmosets utilize accumulated spatial information to minimize travel costs.
Role of Temporal Memory: Incorporating a phenological constraint (Temporal NN) yielded a small but statistically significant improvement over the simple NN rule ( $d_z = 0.37$ ). This suggests that while spatial proximity is the dominant criterion, memory regarding seasonal patch use plays a complementary, albeit minor, role.
Distance vs. Value: The gravity rule analysis revealed that distance was weighted approximately twice as heavily as patch attractiveness ( $a/b \approx 0.57$ ). The improvement of the gravity rule over the simple NN rule was marginal, reinforcing that proximity is the primary driver of destination choice.
Overshooting Behavior: Observed median displacements (~~55 m) exceeded the distance to the nearest available patch (~~13–31 m), indicating that marmosets do not strictly minimize distance to the absolute nearest neighbor but follow a structured heuristic that occasionally overshoots immediate options.

Significance and Claims
The paper claims to provide evidence that Callithrix jacchus, one of the smallest Neotropical primates, structures its foraging behavior using low-cost spatial heuristics functionally analogous to those found in larger-bodied primates.

Distinction Between Pattern and Process: The study emphasizes that the geometric structure of movement (log-normal distribution) does not imply random search. Instead, the structured pattern emerges from a process of distance minimization guided by spatial memory.
Cognitive Economy: The results support the hypothesis that efficient navigation does not require complex cognitive maps or multi-step planning (e.g., solving the Traveling Salesman Problem optimally). Instead, simple local rules, such as the nearest-neighbor heuristic, generate near-efficient trajectories.
Ecological Validity: By validating these heuristics in a natural, seasonal environment with unmanipulated resource distributions, the study extends findings from controlled experiments to wild populations.
Generalizability: The authors suggest that the reliance on proximity heuristics may be a widespread trait within the genus Callithrix and other callitrichids inhabiting heterogeneous environments, challenging the notion that complex spatial cognition is exclusive to larger-brained primates.

The study concludes that C. jacchus navigates its environment as a "known and structured space," utilizing spatial memory to reduce travel costs, with temporal information serving as a secondary, complementary factor.

Technical Summary: Spatial Proximity Heuristics in Callithrix jacchus Foraging

More like this