Related Experiment Videos
Automatic Screening to Detect 'At Risk' Child Speech Samples using a Clinical Group Verification framework
Insights
This study shows that speech processing techniques can accurately detect speech sound disorders (SSD) in young children. Early detection through this method aids timely intervention, improving developmental outcomes.
Area of Science:
- Speech-language pathology
- Computational linguistics
- Pediatric audiology
Background:
- Pediatric speech sound disorders (SSD) impact education and social development.
- Early SSD detection is crucial but challenging due to developmental variability.
- Existing screening methods may not fully capture speech production nuances in young children.
Purpose of the Study:
- To explore the feasibility of using computational methods for early detection of pediatric speech sound disorders.
- To develop novel speech processing techniques for analyzing children's speech samples.
- To assess the accuracy of these techniques in identifying children at risk for SSD.
Main Methods:
- Gaussian Mixture Models were applied to analyze speech samples from children aged 3-6 years.
- Speech samples were collected by parents via an iOS application.
- Speech-language pathologists provided expert classification of speech samples as 'at risk' or 'no risk' for SSD.
Main Results:
- Novel distance measures and group scoring techniques demonstrated good subject-level prediction accuracy.
- The developed methods show promise in distinguishing between typical and atypical speech production.
- Computational analysis of speech samples achieved effective screening for potential SSD.
Conclusions:
- Speech processing and speaker verification techniques show potential for modeling and screening pediatric speech sound disorders.
- This approach may offer a scalable and objective method for early SSD identification.
- Further research can refine these computational tools for widespread clinical application.
Abstract:
Pediatric speech sound disorders (SSD) encompass a wide range of speech production deficits that can interfere with children's educational growth, social engagement and employment opportunities. Early detection of SSDs can facilitate timely intervention and minimize the potential for life-long adverse effects, but distinguishing between typical and atypical speech production in preschoolers is challenging due to developmental and individual variability in speech acquisition. In this study we apply Gaussian Mixture Models to speech samples from 3- to 6-year-old children, recorded by parents using an iOS app. Speech-language pathologists previously classified the samples as positive ('at risk' speech, warranting a referral for a speech-language evaluation) or negative ('no risk' speech). In a series of exploratory analyses, novel distance measures and group scoring techniques are developed which show good subject-level prediction accuracy. Our results provide evidence that it may be feasible to use Speech Processing and Speaker Verification techniques to model and screen speech samples from children for possible speech sound disorders.