Using Artificial Intelligence (AI) to Create Within Speaker Stimuli for Perception Training of Children with RSSD
Presentation Type
Poster
Faculty Advisor
Elaine Hitchcock
Access Type
Event
Start Date
26-4-2024 9:45 AM
End Date
26-4-2024 10:44 AM
Description
When a speech sound disorder (SSD) extends past 8 years of age, these errors can be classified as a residual speech sound disorder (RSSD). Research has shown that production training alone may not be sufficient in resolving children’s residual speech errors and the addition of perceptual training may be necessary to build accurate representations of errored sounds [1]. The multifaceted nature of perception [3, 4] suggests the likelihood of a bidirectional relationship with speech production accuracy [2, 5, 6]. To date, research outcomes have not readily translated to clinical guidelines because of methodological differences across studies. Such differences impact the generalizability of findings, specifically for between-speaker vs. within-speaker training stimuli. However, barriers to creating customized within-speaker stimuli have limited its investigation, particularly for children with RSSD who are unable to produce correct versions of errored phonemes [7]. In this pilot study, 100 errored fricative productions from a single participant with an /s/ distortion were artificially corrected using an AI software program. Four expert listeners rated the corrected tokens for perceptual accuracy. Those with unanimous agreement were combined with naturally produced stimuli and included in a second listening task for accuracy and source differentiation (AI vs. natural) ratings. The project aimed to investigate the use of AI stimuli to create perceptual training modules to enhance generalization outcomes. Preliminary ratings from blinded listeners reveal a lack of significance between listeners for accuracy or sound source identification suggesting AI can be a useful tool in creating perceptual training stimuli for children with RSSD.
Using Artificial Intelligence (AI) to Create Within Speaker Stimuli for Perception Training of Children with RSSD
When a speech sound disorder (SSD) extends past 8 years of age, these errors can be classified as a residual speech sound disorder (RSSD). Research has shown that production training alone may not be sufficient in resolving children’s residual speech errors and the addition of perceptual training may be necessary to build accurate representations of errored sounds [1]. The multifaceted nature of perception [3, 4] suggests the likelihood of a bidirectional relationship with speech production accuracy [2, 5, 6]. To date, research outcomes have not readily translated to clinical guidelines because of methodological differences across studies. Such differences impact the generalizability of findings, specifically for between-speaker vs. within-speaker training stimuli. However, barriers to creating customized within-speaker stimuli have limited its investigation, particularly for children with RSSD who are unable to produce correct versions of errored phonemes [7]. In this pilot study, 100 errored fricative productions from a single participant with an /s/ distortion were artificially corrected using an AI software program. Four expert listeners rated the corrected tokens for perceptual accuracy. Those with unanimous agreement were combined with naturally produced stimuli and included in a second listening task for accuracy and source differentiation (AI vs. natural) ratings. The project aimed to investigate the use of AI stimuli to create perceptual training modules to enhance generalization outcomes. Preliminary ratings from blinded listeners reveal a lack of significance between listeners for accuracy or sound source identification suggesting AI can be a useful tool in creating perceptual training stimuli for children with RSSD.