Attention and action during play in young children with Down syndrome: Insights from head-mounted eye-tracking and cameras

1. Hana D'Souza (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)

Correspondence: dsouzah@cardiff.ac.uk

Young children learn by interacting with the world around them. For children with Down syndrome, differences in motor development and social interaction may influence these everyday learning experiences. This symposium brings together a set of closely related studies that use child-worn, head-mounted eye-tracking and cameras to study parent-child play from the child’s perspective.

By combining measures of attention, social engagement, and physical action, the presented work offers insights into how young children with Down syndrome engage with objects, people, and shared activities as they unfold moment by moment. The findings highlight the role of parental involvement in supporting children’s attention and exploration, and show how social interaction and physical action shape what children attend to during play.

Overall, the symposium explains how we currently use child-centred technologies to study early experiences in ways that prioritise the child’s point of view, and introduces our plans for future work. By sharing this work with parents, practitioners, and other researchers, the session offers a space to reflect on how child-centred research can inform our understanding of how young children with Down syndrome learn, and how best to support them.

Joint attention boosts sustained attention in young children with and without Down syndrome: Insights from dual head-mounted eye-tracking

1. Qin Liu (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)
2. Catalina Suarez-Rivera (Great Ormond Street Institute of Child Health, University College London, UK, Department of Applied Psychology, New York University, USA)
3. Merideth Gattis (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)
4. Chen Yu (Department of Psychology, University of Texas at Austin, USA)
5. Hana D'Souza (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)

Correspondence: liuq51@cardiff.ac.uk

Sustained attention (SA) and joint attention (JA) are foundational for young children’s early learning and development. Even though they are traditionally conceptualised as distinct processes, emerging evidence indicates that JA provides a critical social pathway for SA in young typically developing children. However, it remains unclear whether JA scaffolds SA in children with Down syndrome. Addressing this question is crucial for informing practitioners and parents about effective support.

Here, we simultaneously investigated SA and JA during free-flowing parent-child object play using dual head-mounted eye-tracking in 15 young children with Down syndrome aged 3.0 to 4.8 years and 15 typically developing toddlers aged 1.4 to 2.3 years and their parents. The groups were matched on ability level using the Mullen Scales of Early Learning.

The results reveal both similarities and differences in attention among young children with Down syndrome and ability-matched typically developing children, as well as between children and their parents. Specifically, typically developing children and children with Down syndrome did not differ significantly in JA. However, children with Down syndrome exhibited significantly shorter durations of SA than their typically developing peers. Importantly, our findings revealed that JA serves as a social process that boosts SA in both typically developing children and children with Down syndrome. This study underscores the social foundations of early SA and highlights that both children with Down syndrome and typically developing children benefit from parental support for attention engagement.

How motor actions of young children with Down syndrome, and their parents, shape what they see

1. Charlotte Bocchetta (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)
2. Craig Thompson (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK, School of Computer Science and Informatics, Cardiff University, UK)
3. Catalina Suarez-Rivera (Great Ormond Street Institute of Child Health, University College London, UK, Department of Applied Psychology, New York University, USA)
4. Chen Yu (Department of Psychology, University of Texas at Austin, USA)
5. Hana D'Souza (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)

Correspondence: BocchettaC@cardiff.ac.uk

Young children learn about the world by interacting with objects around them. Through holding, exploring, and manipulating objects, they create moments in which a single object stands out in their visual scene. These moments help them to focus, recognise objects, and build early language skills. For toddlers who are typically developing, these experiences happen naturally as they move and explore their environment. However, children with Down syndrome often face motor challenges that may affect how these visual learning opportunities unfold.

Our study aimed to understand how motor abilities influence how children with Down syndrome experience objects, specifically how large objects appear in their visual scene, and how parents’ actions shape these experiences. We compared 15 typically developing toddlers (17–27 months) and 15 young children with Down syndrome (36–58 months), matching them on developmental abilities. Each child and their parent played with three objects while wearing head-mounted eye-trackers that captured their first-person visual scenes. To analyse visual experience, we developed a computer model to estimate the sizes of objects within children’s visual scene. We also examined how children and parents handled objects to see how these interactions influenced their visual experience.

Overall, both children with Down syndrome and typically developing children experienced moments in which objects were large in view, and these moments were associated with increased looking to the object. However, objects tended to appear larger for children with Down syndrome, indicating that the objects were closer to their faces. Children with Down syndrome also handled objects significantly less often than typically developing children, whereas their parents handled objects more frequently. Importantly, during these moments when an object was largest in view, children’s attention to objects and object handling did not differ significantly between groups.

Together, the results suggest that ability-matched young children with and without Down syndrome experience differing patterns of object sizes in their visual scene during play, shaped by both their own motor actions and those of their parents. These findings highlight the importance of motor abilities and parent involvement in shaping everyday visual experiences of children with Down syndrome. Future research should examine how these experiences relate to learning, to better support the development of children with Down syndrome.

Detecting tiny hands: Evaluating automated hand detection on a unique headcam dataset of young children with and without Down syndrome

1. Craig Thompson (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK, School of Computer Science and Informatics, Cardiff University, UK)
2. Yu-Kun Lai (School of Computer Science and Informatics, Cardiff University, UK)
3. Hana D'Souza (Centre for Human Developmental Science, School of Psychology, Cardiff University, UK)

Correspondence: ThompsonC21@cardiff.ac.uk

Hand and body movements play a crucial role in shaping what children are able to see, interact with, and learn from. However, for children with Down syndrome, delays in motor development can influence access to objects and learning opportunities compared to typically developing children. The use of head-mounted cameras (headcams) provides a unique insight into these behaviours in everyday environments. However, quantifying hand and body movements across age and developmental groups remains a large challenge, with manual annotation being too time-consuming.

This study aims to evaluate the effectiveness of existing automated hand detection in identifying hands, with the best-performing model to be further trained to achieve optimal performance in headcam videos from children with and without Down syndrome. Six hand detection models were evaluated to determine which models best generalise to naturalistic infant data that often includes occlusions, motion blur, and varied hand positions.

The data used to evaluate the models were collected using the TinyExplorer head-mounted camera gear, worn by 28 children with Down syndrome (6–60 months) and 29 typically developing children (2–30 months). From these recordings (mean duration = 62 minutes), 42,338 frames were sampled at a rate of one frame every 5 seconds and manually annotated to create ground-truth labels and training data.

Preliminary results for evaluation show that two models, 100Days of Hands (100DoH) and BodyHands, demonstrate strong out-of-the-box performance, with both achieving high overall presence accuracy above 94%. Additionally, for frames containing only the wearer’s own hands, 100DoH and BodyHands maintained high presence accuracy of above 84%.

Taken together, some of the models demonstrate strong performance across diverse headcam scenes, providing a foundation for quantifying how children interact with their environment. Our future work will refine detection thresholds and provide additional annotated training datasets to further enhance accuracy and reliability. With accurate hand identification, future applications include the automatic location of hands in child videos, the extraction of hand-object interactions, and providing insights into object exploration behaviour, supporting new ways we can learn about and support everyday experiences of young children with Down syndrome.