Speech Production Errors and Their Relationship to Speech Intelligibility Across Levels of Articulatory Severity in Children with Spastic Cerebral Palsy

Objectives: This study aimed to investigate differences in the types of speech production errors according to the levels of articulation severity in children with spastic cerebral palsy (CP) and to examine the relationship between these error types and speech intelligibility. Methods: The participants were 36 children with spastic CP (20 boys, 16 girls) with a mean chronological age of 10.17 years. They were divided into three groups based on articulation severity: mild (n=12), moderate (n=9), and severe (n=15). The Assessment of Phonology & Articulation for Children (APAC) was administered, and speech production errors were classified as omission, substitution, distortion, prolongation, and non-speech errors. Speech intelligibility was analyzed as the percentage of words correctly understood by listeners. Results: The pattern of speech production errors differed significantly according to articulatory severity. Significant group differences were found in omission, prolongation, non-speech, and substitution errors, and the severe group showed significantly higher frequencies of omission and non-speech errors than other error types. Furthermore, withingroup analyses revealed that substitution was the predominant error type in the mild group, whereas substitution and omission were the major error types in the moderate group. In contrast, omission and non-speech errors were most prominent in the severe group, indicating that greater articulatory severity was associated with increased difficulty in consonant production and increased non-speech vocalizations. In addition, omission, prolongation, and non-speech errors showed significant negative correlations with intelligibility. A hierarchical multiple regression analysis revealed that omission and non-speech errors were the strongest predictors of intelligibility, even after controlling for articulatory severity. Conclusion: These findings suggest that intelligibility is influenced not only by severity but also by specific error types, particularly omission and non-speech errors, indicating the need to consider both factors in intervention.