Feasibility of Technology Enabled Speech Disorder Screening

Area of Science:

• Pediatric Health
• Speech Pathology
• Computational Linguistics

Background:

• Speech disorders affect 1 in 20 Australian children, impacting literacy and academic success.
• Early identification of speech disorders is vital to mitigate long-term educational and social costs.
• Current screening methods may lack accessibility and efficiency for widespread early detection.

Purpose of the Study:

• To develop and test a prototype screening and decision support tool for assessing speech disorders in young children.
• To leverage technology for automated phoneme classification, distinguishing normal from disordered speech.
• To lay the groundwork for a mobile application for accessible, early speech disorder screening.

Main Methods:

• Development of a prototype tool integrating speech signal processing and machine learning algorithms.
• Utilizing expert knowledge alongside AI for automatic classification of phonemes.
• Conducting feasibility tests to evaluate the prototype's performance in assessing speech disorders.

Main Results:

• Successful development of a functional prototype for speech disorder assessment.
• Demonstrated capability of the prototype to automatically classify phonemes.
• Feasibility tests indicated potential for accurate and efficient speech disorder screening.

Conclusions:

• The developed prototype shows promise for early and accessible speech disorder screening.
• Further development towards a mobile tool could significantly enhance early detection rates.
• Automated speech analysis holds potential to reduce the burden of speech disorders on children and society.