Research lead: Dr Elise Baker
Research area: Allied Health
Team: Dr Weicong Li, Prof Caroline Jones, Dr Saeed Afshar, A/Prof Yi Guo, and A/Prof Mark Antonio (Western Sydney University), Prof Mary Alt (University of Arizona), Prof Natalie Munro (Southern Cross University), Dr Rosemary Hodges (Word Stars Speech Pathology), Dr Sarah Masso (University of Sydney), and Ms Katrina Tosi (South Western Sydney Local Health District)
What was the health problem that led you to carry out your research?
Communication difficulties in childhood are widespread, affecting about 1 in 10 children globally. In Australia, around 1 in 5 children experience being late to talk. Early intervention is crucial as communication problems can lead to behavioural issues, learning delays, low self-esteem and poorer literacy outcomes. Tools are essential to screen, assess and intervene in these cases to help children develop their communication skills. VAULT is an input play-based intervention in which speech pathologists say predetermined treatment words at a prescribed frequency (e.g. say 90 productions each of put, baby, and chair) during a 30-minute therapy session. VAULT has demonstrated its success as an intervention. However, counting the dose
of 90 words per half hour is not feasible for speech pathologists as they engage toddlers in varied play activities. Therefore, a real-time automated counting tool is required to ensure treatment fidelity (correct dose needed for the intervention).
Describe the research achievement and its impact.
We are exploring how automatic speech recognition (ASR) technologies can support the assessment and diagnosis of communication delays. We are doing this by automating the count of specific treatment words during VAULT sessions. The ASR tool provides real-time visual feedback in the form of a graph on a tablet or smart phone. This visual feedback ensures therapists can interact with children and maintain intervention intensity without being distracted by word counting. Dr. Baker developed the STeWoT prototype using Python through interdisciplinary collaboration, which was then transformed into a user-friendly interface called WombatWords with Ben Blain and his team at Neurabuild. Accessible on iPhones, iPads and computers, the tool
automates and enhances the efficiency of VAULT by ensuring fidelity.
The next step is usability testing with speech pathologists from local health districts in NSW and families so that it can be widely adopted.
How was the work unique or pioneering?
WombatWords, used with VAULT, could aid the speech pathology community to provide VAULT with optimal fidelity in delivering the intervention, ultimately benefiting children. The aim is to extend its use to coach parents in administering VAULT at
home.
What is the broader implication of the research?
Delayed talking is a global issue with significant educational, social and wellbeing impacts. Speech pathologists have long sought to enhance treatment effectiveness, particularly before children start school. Effective, efficient, and equitable speech pathology interventions like VAULT, supported by tools like WombatWords, can mitigate the adverse effects of late talking. This research has the potential to benefit thousands of Australian children and millions more worldwide, providing them with a better start in school by improving their ability to communicate and be understood.
