Session 4 - Symposium: Leveraging machine learning to understand and support the everyday experiences of young children with Down syndrome

Organiser: Hana D’Souza, Cardiff University, UK

Contact: Hana D’Souza (DSouzaH@cardiff.ac.uk)

Over the last decade, machine learning has advanced at an unprecedented pace, with algorithms reaching new heights of sophistication (particularly noticeable with ChatGPT). While these developments have transformed fields like cybersecurity and healthcare, their application in understanding and supporting the development of children with Down syndrome remains relatively unexplored.

In this symposium, we introduce our initial steps in leveraging machine learning in our research. In the first talk, Charlotte will share novel findings from our head-mounted eye-tracking study of parent-child interaction and explain how machine learning can help us understand how young children with Down syndrome engage with objects. In the second talk, Teodor will provide an update on our ongoing head-mounted camera study and explain how we can harness machine learning to learn about the everyday experiences of children with Down syndrome. In the third and final talk, Becky will introduce our plans for exploring parents’ and practitioners’ views on the utility of these technological advancements beyond research, towards providing better support for young children with Down syndrome in clinical practice.

By integrating machine learning into developmental research, we aim to open new avenues for tailored support, ultimately improving the everyday experiences of young children with Down syndrome.

Symposium structure:

• A brief introduction (Hana D’Souza)
• Talk 1: How young children with Down syndrome experience objects during play (Charlotte Bocchetta)
• Talk 2: Everyday experiences of children with Down syndrome during the first five years of life (Teodor Y. Nikolov)
• Talk 3: Bringing research to practice: Exploring parents’ and practitioners’ views on the use of technology to support clinical practice for young children with Down syndrome (Becky Hardiman)
• Brief discussion (Hana D’Souza)

How young children with Down syndrome experience objects during play

Charlotte Bocchetta¹, Craig D J Thompson¹, Ziye Zhang¹, Catalina Suarez-Rivera^2,3, Yu-Kun Lai¹, Chen Yu⁴, & Hana D’Souza¹

¹Cardiff University, UK; ²New York University, USA; ³University College London, UK; ⁴University of Texas at Austin, USA

Contact: Charlotte Bocchetta (BocchettaC@cardiff.ac.uk)

Young children learn about the world by interacting with objects around them. When they hold and explore objects, they create moments where a single object stands out in their view. 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. However, children with Down syndrome often face motor challenges that may affect these experiences.

Our study aims to understand how motor abilities influence how children with Down syndrome experience objects and how parents’ actions shape their experiences. We compared 15 typically developing toddlers (17-27 months) and 15 young children with Down syndrome (36-58 months), matching them on developmental ability. Each child and their parent played with new objects while wearing head-mounted eye-trackers to record their views. To analyse how the objects appeared in their visual field, we developed a cutting-edge computer model to track object size. We also examined how children and parents handled the objects to see how these interactions influenced object sizes in the view.

This research highlights the importance of considering motor abilities and parent involvement in the everyday experiences of children with Down syndrome. Understanding how young children with Down syndrome engage with objects can help us to provide more optimal support for their early learning and development.

Everyday experiences of children with Down syndrome during the first five years of life

Teodor Y. Nikolov, Yu-Kun Lai, & Hana D’Souza

Cardiff University, UK

Contact: Teodor Y. Nikolov (NikolovTY@cardiff.ac.uk)

Many children with Down syndrome experience motor difficulties. Surprisingly, we understand very little about how these difficulties impact their everyday experiences and influence their learning. In this talk, we provide an update on the study we introduced at the Down Syndrome Research Forum last year (2024), which uses our custom-made TinyExplorer headgear. This lightweight and user-friendly head-mounted camera records what young children with Down syndrome hear (auditory experiences) and see (visual experiences) from their own perspective.

We will share preliminary findings from this ongoing study which aims to capture the first five years of life of children with Down syndrome. We will also explain how we are using machine learning to extract important novel information from children’s views.

We are hoping this study will shed light on the developmental trajectories of auditory and visual experiences during this early period of development. Given the critical role of early motor abilities (including eye movements, reaching, sitting, and locomotion) in facilitating interaction with the environment, motor difficulties can potentially limit the diversity and richness of experiences available to some children. This, in turn, can inform the development of targeted support designed to enhance learning opportunities and facilitate overall developmental outcomes.

Bringing research to practice: Exploring parents’ and practitioners’ views on the use of technology to support clinical practice for young children with Down syndrome

Becky Hardiman, Charlotte Bocchetta, Nicola Birdsey, & Hana D’Souza

Cardiff University, UK

Contact: Becky Hardiman (HardimanB@cardiff.ac.uk)

New technologies, like head-mounted cameras and machine learning, are helping researchers better understand how young children with Down syndrome learn and explore their surroundings. These tools allow us to see the world from the child’s perspective, providing valuable insights into early development. However, it is important to consider how these technologies can be used beyond research and whether they could help support early learning and development for children with Down syndrome. It is paramount to consult those with lived and professional experience to guide and set priorities for the development of the potential clinical use of such technologies.

In this talk, we will share our plans for exploring parents' and professionals' views on the potential use of the head-mounted cameras and machine learning to gather information about how young children with Down syndrome engage with their everyday environments for clinical practice. We are planning to conduct two Delphi studies (30 participants per panel)—one with parents and another with allied health professionals—to consult upon the use of head-mounted cameras and machine learning. This will include an initial round of semi-structured interviews with individual members of each Delphi panel. Themes emerging from these interviews will then be used to generate questions for participants to rate (e.g., on the strength of agreement with different statements). Finally, panel attendees will be provided with a summary of their own ratings against the group average, allowing an opportunity for them to reflect on, and potentially amend, their responses based on exposure to others’ opinions. The aim is to gather a panel consensus to help guide the future development and use of head-mounted cameras and machine learning technology in clinical practice.

This research is in its early stages (design and recruitment). We welcome attendees’ thoughts on the project and expression of interest in potential participation.