In this interview, Dr. Anna Schmitt speaks with Dr. Decarli about her latest research on the early development of mathematical skills. Based on the study Number sense at 12 months predicts 4-year-olds’ maths skills (Decarli et al., 2023), Dr. Decarli explains how infants’ ability to perceive quantities at 12 months can predict their later math abilities at 4 years of age. This research sheds light on the foundations of numerical cognition and offers valuable insights for scientists, educators, and parents interested in early childhood development.*
*Decarli, G., Zingaro, D., Surian, L., & Piazza, M. (2023). Number sense at 12 months predicts 4‐year‐olds’ maths skills. Developmental Science, 26(6).
Pour avoir accès à l’étude complète, voici un accès direct : https://doi.org/10.1111/desc.13386
Could you explain to us the objective of your study?
Understanding and addressing the typical developmental trajectories of mathematical acquisition is essential, as mathematical competence is crucial for daily activities, enabling individuals to engage with numerical data, solve complex problems that involve quantities and probability, and make informed decisions (Geary, 2011). Some authors have proposed that formal math skills develop on the basis of a dedicated neurocognitive system that supports the ability to represent the approximate number of objects in sets, called the Approximate Number System (ANS). This system is thought to represent numerosity with a precision (referred to as the ANS acuity) that varies across individuals and that is subject to change with maturation and math learning.
The majority of the studies that assessed the link between ANS and math skills have used correlational approaches, capitalizing on inter-individual differences. However, they mostly tested children who have already been exposed to some form of math education (e.g., verbal counting principles). This makes it difficult to determine the direction of causality between ANS acuity and symbolic math skills. Therefore, the objective of our study was to investigate whether infants’ early numerosity acuity serves as a foundation for later mathematical abilities. Specifically, we aimed to determine if the ANS acuity measured at 12 months of age, thus way before any form of math learning, would be a reliable and specific predictor of symbolic mathematical skills at 4 years, independent of general intelligence or inhibitory skills. This longitudinal approach was designed to replicate and expand on previous research, using a different cultural and linguistic sample, as well as diverse tasks, to test the specificity and robustness of this developmental relationship.
What gaps in the scientific literature did your study address?
Our study builds upon the groundbreaking but standalone work of Starr and colleagues (2013), who first showed that ANS acuity measured at 6 months longitudinally predicted symbolic math achievement at 3.5 years. With our study, we sought to replicate their important findings while introducing new elements to further investigate this topic. Specifically, we conducted a longitudinal study in which we first tested 12-month-old infants on both a numerosity perception task and a perceptual face perception task. The latter was used as a control task to assess the specificity of the potential correlation between numerosity perception and math, ensuring that the link was not guided by more general perceptual abilities. Notably, and differently to Starr et al., who used different control tasks across participants (either a color detection task or a size detection task), we employed the same control task (face recognition) for all participants. This uniform approach ensured that participants were consistently assessed in the same non-numerical perceptual capacity.
We then re-evaluated the same participants at 4 years of age on a comprehensive set of non-symbolic and symbolic formal math tasks, as well as general processing skills, including measures of general intelligence, a face discrimination task as well as inhibitory abilities. The inclusion of an inhibitory task was another novelty of our study. Indeed, recent literature suggests that inhibitory control may play a key role in both numerosity comparison tasks (by helping suppress irrelevant visual information in favor of numerical one) and formal math tasks (by supporting complex calculations). According to this view, the observed correlation between ANS acuity and symbolic math performance could be explained by the shared reliance on inhibitory control.
In sum, the assessment of a new sample from a different linguistic and cultural background compared to Starr et al., the use of a face perception control task for all infants at 12 months, and the introduction of inhibitory skills test, allow us to offer new advancements in understanding the role of early numerosity perception in the development of mathematical skills.
What type of methodological design did you use, and why?
In our research, we implemented a longitudinal study design. In this type of study, the same participants are assessed at multiple time points over an extended period, allowing researchers to track changes and developmental trajectories within the same individuals. In our case, this approach allowed us to assess preverbal infants at 12 months of age and re-test them again when they were 4 years old, measuring the abilities of interest at different stages of development. The main strength of this type of design is that it enables us to observe how early skills develop over time and track individual developmental pathways. By following the same participants across multiple time points, we could establish a temporal link between early numerosity skills and later symbolic acquisitions (such as counting). This provides further support for the hypothesis positing a relation between the two skills.
What were your initial hypotheses, and what motivated them?
Our initial hypothesis was that infants’ ability to perceive and process numerical quantities could play a specific and selective role in the development of later mathematical skills. Specifically, we hypothesized that ANS acuity (and not other perceptual skills, such as face perception) at 12 months of age would predict early symbolic mathematical abilities at 4 years. According to this hypothesis, the ANS helps children connect symbolic representations to their intuitive understanding of quantities, allowing them to attribute meaning to number words and map these symbols onto their pre-existing representations of magnitudes. This hypothesis was motivated by prior research supporting a longitudinal correlation between numerosity perception and math achievement in school-aged children (Libertus, Feigenson, & Halberda, 2011; Halberda, Mazzocco, & Feigenson, 2008; Libertus, Odic, & Halberda, 2012). Additionally, deficits in numerosity comparison tasks, where participants are required to indicate the larger of two quantities, have been observed in children with developmental dyscalculia, a specific learning disorder in math (see Decarli et al., 2020; Decarli et al., 2023).
Another important hypothesis underlying our research was that the link between ANS acuity and math would not be driven by general intelligence or inhibitory skills. In the scientific literature, some authors have proposed that inhibitory skills would play a significant role in the link ANS acuity-math (e.g., Gilmore et al., 2013). Contrary to this theoretical position, we expected to find a correlation between ANS and math that would remain significant even after controlling for domain-general abilities.
What were your study groups, and why were they chosen?
Our study involved a group of 60 infants, who were initially tested at 12 months of age, and a follow-up sample of 40 children from the same group, tested at 4 years of age. We chose this age range to capture two key developmental stages: preverbal infants, who have not yet acquired formal numerical knowledge, and preschool-aged children, who are beginning to engage with symbolic math concepts. The selection of these specific age groups was driven by our goal to explore the precursors of symbolic mathematical knowledge before any formal education. Testing infants at 12 months allowed us to assess their early numerosity acuity, while the follow-up at 4 years provided information into their emerging numerical/symbolic abilities in a critical period for early learning.
What are your main findings?
Our study revealed several key findings. First, we found a significant correlation between ANS acuity (measured using a dots comparison task) and mathematical abilities at 4 years of age. In contrast, we did not observe a correlation between mathematical abilities and inhibitory skills, suggesting that, at this stage of development, these general cognitive skills might not play a significant role in the early mastery of math skills.
When considering the longitudinal data, we found that numerosity acuity measured by an implicit change detection paradigm at 12 months significantly correlated with children’s performance on an explicit dots’ comparison task at 4 years, indicating great reliability of the different measures in time. More importantly, the key finding was that we replicated the significant longitudinal correlation between ANS acuity at 12 months and performance in a standardized symbolic math test at 4 years as initially observed by Starr et al. (even though the ages of our participants were slightly different compared to those of the original study). This link was robust, as it remained significant even after controlling for general intelligence and inhibitory skills. Furthermore, we demonstrated that this link was specific to numerosity perception: neither our control task, a face recognition skill measured at 12 months, did predict later math performance, nor the ANS at 12 months did predict face processing skills at 4 years.
To sum up, these results suggest that numerosity acuity correlates with math abilities in preschoolers and that early numerosity perception at 12 months can be considered a reliable and specific predictor of later math skills at 4 years. Crucially, this relationship was not mediated by domain-general abilities, such as IQ or inhibitory control.
Do they support your research hypotheses? Are they consistent with the scientific literature, or do they differ?
Our findings generally support our hypotheses and align with a previous similar study in the field. In particular, we replicated the key longitudinal results of Starr and colleagues, confirming that early ANS acuity is a longitudinal precursor of later mathematical skills. This consistency reinforces the evidence for a developmental association between early numerosity perception and formal math abilities from infancy through childhood.
In line with our hypotheses, but in contrast with some of the literature, our results did not provide evidence for a role of inhibitory skills in math acquisition. This finding contrasts with some theories suggesting that inhibitory skills could account for the observed link between ANS and math performance. For example, Gilmore and colleagues (2013) proposed that dot comparison tasks rely on inhibitory control, as participants must suppress responses to non-numerical features such as size or density. According to their view, individual differences in mathematical performance would be explained by differences in inhibition rather than by numerical representations per se. Our data do not support this hypothesis. We found no correlation between mathematical abilities and inhibitory skills, and the observed link between ANS acuity and symbolic math persisted even after controlling for inhibition.
Can the results of this study be useful in real life for teachers, school psychologists, medical doctors or even parents? If so, to what extent?
The results of our study provide valuable information for people involved in a child’s early development, including educators, psychologists, and parents. Our findings show that, even in the first months of life, there are already inter-individual differences in numerosity perception that may lead to varying levels of mathematical achievement later in childhood. This knowledge could help in identifying children who may be at risk of math-related learning difficulties and could contribute to the early detection of difficulties that might result in later struggles with math. Moreover, the results of our study could help educators and psychologists develop targeted and timely interventions aimed at strengthening these skills as soon as possible in the preschool age. Finally, for parents, our findings emphasize the importance of encouraging children’s engagement with quantities and numbers from a very young age. Everyday activities, such as comparing quantities, could help foster children’s intuitive number sense, potentially supporting their later mathematical development.
What potential links or connections do you see with your study and the scientific studies conducted in relation to Magrid?
Cornu, V., Schiltz, C., Pazouki, T., & Martin, R. (2017b). Training early visuo-spatial abilities : A controlled classroom-based intervention study. Applied Developmental Science, 23(1), 1‑21. https://doi.org/10.1080/10888691.2016.1276835
Jung, S., Meinhardt, A., Braeuning, D., Roesch, S., Cornu, V., Pazouki, T., Schiltz, C., Lonnemann, J., & Moeller, K. (2020c). Hierarchical Development of Early Visual-Spatial Abilities – A Taxonomy Based Assessment Using the MaGrid App. Frontiers In Psychology, 11. https://doi.org/10.3389/fpsyg.2020.00871
Pazouki, T., Cornu, V., Sonnleitner, P., Schiltz, C., Fischbach, A., & Martin, R. (2018d). MaGrid : A Language-Neutral Early Mathematical Training and Learning Application. International Journal Of Emerging Technologies In Learning (iJET), 13(08), 4. https://doi.org/10.3991/ijet.v13i08.8271
There are clear links between our study and the work conducted in relation to the Magrid project, as both focus on the early precursors of math. While our study aims to assess these foundational skills, Magrid mainly focuses on training them through a language-neutral application. Both approaches emphasize the importance of numerosity perception as a key factor in later math acquisition. These complementary approaches highlight the critical role of early numerical cognition in supporting children’s mathematical development and suggest potential educational applications for both assessment and intervention.
What is the subject of your current scientific research?
I am currently working on several research projects that explore different aspects of learning. In particular, my current research focuses on investigating the similarities and differences between dyslexia and dyscalculia at both the cognitive and neural levels. I am also exploring the role of emotions, both positive and negative, in the learning process, analyzing how emotional states can influence children’s academic performance and cognitive development. Finally, I am concluding a study on domain-general and domain-specific precursors of mathematical skills to identify the early cognitive factors that contribute to math acquisition.
