Study Reveals How Infants Adapt Their Learning to Changes in Their Environment
Have you ever wondered how infants respond to the constantly changing world around them? Recent research sheds new light on this question. Even before their first birthday, babies are able to learn and adapt when their environment changes. A new scientific study investigates the early development of these adaptive learning abilities in infants and explores how they adjust their understanding of the world from a very young age.
Note: This article is intended for general information and educational purposes. It summarizes scientific research in accessible language for a broad audience and is not an official scientific press release.
A new study led by researchers from the University of Cambridge, Radboud University, Birkbeck University of London, University of Amsterdam, and the University of Oxford has uncovered how human infants as young as eight months actively adjust their learning strategies in response to changing environments.
Published in Science Advances on June 25, 2025, the study by Francesco Poli and colleagues investigated the developmental origins of adaptive learning by analyzing infants’ pupil responses during a reversal learning task. The authors report that infants not only detect environmental volatility but also use this information to optimize their learning, with individual differences in these abilities related to aspects of infant temperament.
What the Researchers Investigated
The study set out to explore the roots of adaptive learning in human development. The researchers aimed to answer how infants estimate and respond to environmental volatility—a process well-documented in adults but not fully understood in early life. Volatility in this context refers to situations where the underlying rules of the environment change, making previous patterns unreliable.
The research was conducted by a team from the MRC Cognition and Brain Sciences Unit (University of Cambridge), Donders Institute for Brain, Cognition and Behaviour (Radboud University), Centre for Brain and Cognitive Development (Birkbeck, University of London), University of Amsterdam, and the Wellcome Centre for Integrative Neuroimaging (University of Oxford). The experimental design focused on infants aged around 8 months and used a combination of eye-tracking, pupillometry, and computational modeling to reveal the mechanisms of learning under uncertainty.
How the Study Was Conducted
To understand how infants adapt to changing environments, researchers devised a special learning task. As reported by Neuroscience News in a simplified explanation of the experiment, scientists showed babies a screen displaying a colourful monster. Sometimes the monster appeared on one side of the screen, sometimes on the other.
‘For example, we sometimes let the monster appear only on the left side for a while and then suddenly on the other side,’ explained Dr. Poli, one of the study’s authors.
Meanwhile, using eye-tracking technology (a camera in the screen with harmless infrared light), scientists continuously monitored where the baby was looking and how their pupil size changed. The location where the monster was most likely to appear either remained the same for a long time (a stable, predictable environment) or changed frequently (a rapidly changing environment).
The study found that at a certain point, the babies had learned where to look; they were already gazing at the place where they expected the monster to appear. Dr. Poli noted, ‘When the environment changed, they adjusted their viewing behaviour accordingly. They learned whether the monster’s location was stable or not and were able to actively adapt their learning strategy. It was surprising to see that babies learn in such a flexible way.’
In detail, the experiment included 61 infants (mean age 7.7 months) in the initial sample and 30 in a replication sample, with exclusions applied for insufficient data, resulting in 38 and 20 infants per group, respectively. Caregivers were instructed not to interfere, and infants viewed the task in a quiet, controlled room. Gaze and pupil data were collected using a Tobii X300 eye tracker. Each trial consisted of a fixation point, cue boxes, and a target image, with target location probabilities changing over the course of the experiment to induce volatility. Tonic (baseline) pupil size was measured during fixation, and phasic (rapid) changes in pupil size were measured during target presentation.
The researchers also assessed anticipatory looking behavior, a proxy for infants’ predictions about upcoming events. A computational model known as a volatile Kalman filter (VKF) was fitted to the physiological (pupil) data to estimate how infants track volatility and update their beliefs over time.
What Makes This Study New
The authors highlight several novel aspects of their approach:
- “We developed an experimental and computational approach to track infants’ adaptive learning processes via pupil size, an indicator of tonic and phasic noradrenergic activity.”
- Unlike previous research, which has focused on behavioral responses, this study fitted an optimal learner model directly to infants’ physiological data, enabling fine-grained tracking of learning dynamics in preverbal populations.
- The study introduces a method for measuring individual differences in infants’ sensitivity to environmental volatility and links these differences to parental reports of temperament.
- According to the authors, “This adaptive strategy resulted in successful task performance, as evidenced by anticipatory looking toward correct target locations.”
Key Findings from the Study
The main findings, as described by the authors, are as follows:
- “We found that 8-month-old infants’ tonic pupil size mirrored trial-by-trial fluctuations in environmental volatility, while phasic pupil responses revealed that infants used this information to dynamically optimize their learning.”
- “The model estimates of volatility significantly correlated with infants’ tonic pupil size… indicating that infants were successfully tracking environmental volatility.”
- “The model predictions about the most likely target location significantly correlated with infants’ anticipatory looking… indicating that, by tracking environmental volatility, infants flexibly adjusted their predictions about where the target was most likely to appear.”
- “Infants successfully predicted the target locations not only in stable but also in volatile environments. This demonstrates that infants were not simply more uncertain or confused when volatility was high, but were instrumentally using the information about environmental volatility to optimize their learning.”
- “The interaction between volatility and the magnitude of the prediction errors significantly modulated phasic pupil size. When the environment was more volatile, greater prediction errors led to a greater phasic pupil response… Conversely, when the environment was more stable, greater prediction errors led to a reduction in phasic pupil size.”
- “Infants optimized their learning by flexibly weighting the impact of prediction errors, depending on the volatility of the current environment.”
- “The ability to estimate volatility varied significantly across infants, and these individual differences were related to infant temperament, indicating early links between cognitive adaptation and emotional responsivity.”
- “Volatility estimation as indexed by δ values showed a linear negative correlation with regulatory capacity… a quadratic relation with positive affect… and no relation with negative affectivity. This indicated that infants who overestimate volatility are less well able to regulate their emotions and that infants who are better at estimating volatility show more positive affect.”
Authors’ Conclusions
The authors interpret their results as follows:
- “These findings demonstrate that infants actively adapt to environmental change, and that early differences in this capacity may have profound implications for long-term cognitive and psychosocial development.”
- “Our results indicate that infants are not only sensitive to changes in environmental volatility but also actively use this information to optimize their learning. As such, these findings promote an outlook on cognitive development where infants and their adaptive skills play an active role in shaping the effects of the environment on their developing minds.”
- “Individual differences in volatility estimation were related to infant temperament, with infants who overestimate volatility showing poorer regulatory capacity and infants who correctly estimate volatility showing more positive affect.”
- The study acknowledges limitations such as the potential influence of stimulus novelty on pupil size, and suggests that further research is needed to clarify causal pathways.
The authors note that “a final empirical confirmation of this hypothesis would require longitudinal research.” The study acknowledges limitations such as the potential influence of stimulus novelty on pupil size, and suggests that further research is needed to clarify causal pathways.
Full study available at: DOI: 10.1126/sciadv.adu2014
The information in this article is provided for informational purposes only and is not medical advice. For medical advice, please consult your doctor.
