Perception and Production Errors in Normal Hearing Children, Pediatric Cochlear Implant Users, and Children Listening to Vocoder Simulations

OBJECTIVES: This study investigates the relationship between speech perception and production, focusing on consonant error patterns in children using cochlear implants (CIs), children with normal hearing (NH), and NH children listening to vocoder simulations (NHV). The goals were to determine how perception and production errors align within and across groups, to identify feature- and consonant-specific error patterns, and to assess the relative contributions of developmental constraints, signal degradation, and auditory experience. DESIGN: The study included three groups: 20 children with CIs, ages 5 to 9 years, implanted before the age of three; and two hearing-age-matched control groups of 20 children each, ages 4 to 8 years, presented with natural speech (NH) or listening to nine-channel tone vocoder simulations (NHV). To evaluate their speech perception and production abilities, participants were assessed using a modified version of the California Consonant Test. RESULTS: Relationships between perception and production: Moderate correlations were observed between perception and production across all groups, with slightly weaker correlations in the NHV group, supporting the interconnected nature of these processes. Error rates across groups: No statistically significant differences were observed in mean error rates between the CI and NH groups. However, the NHV group exhibited significantly poorer perception abilities, with higher error rates compared with both the NH and CI groups. NHV listeners also showed significantly higher error rates in perception than in production. This diminished perception in the NHV group likely reflects limited adaptation to the vocoder's degraded auditory signal. Error patterns across consonant features: All groups made fewer errors for stop consonants compared with fricatives in perception and production. In addition, all groups demonstrated fewer errors with voiceless consonants compared with voiced ones in perception and production, with some variability across groups. These results indicate that consonant features such as manner and voicing systematically influence performance. Individual consonant errors: The study revealed discrepancies between perception and production errors for individual consonants. Fricatives (e.g., /v/, /θ/, /z/) were associated with the highest error rates across groups. CI children frequently substituted [d] for /ɡ/, a rare pattern among NH children. Their most common production error was substituting [f] for /θ/, while NH children often confused /θ/ and /f/ or substituted [f] for /v/. These findings indicate that CI children's errors are more variable and sometimes atypical compared with NH peers, whereas NHV errors were broader and less systematic. CONCLUSIONS: The study's results emphasize that while overall error rates are similar between CI and NH groups, the error patterns for individual consonants can differ considerably. This indicates that interventions for children with CIs should focus on the specific error patterns observed. The poorer performance of the NHV group in perception underscores the limitations of vocoder simulations and suggests they may not fully capture the complexities of CI use. Clinical practices should incorporate these insights to develop more effective techniques tailored to the specific needs of children with CIs. Further research is needed to deepen our understanding of the interplay between perception and production errors in CI users to refine clinical protocols accordingly.