RESUMEN
Objective:To observe the pronunciation of consonants among children with developmental speech sound disorder and explore the correlation between mispronounced consonants and short-term memory so as to determine the pathogenesis of the disorder.Methods:Thirty-six children with developmental speech sound disorder and aged 4 to 13 years were evaluated. Their pronunciation of consonants at the phoneme and lexical levels was tested to record the error types and error rate. Twelve of the children were then randomly chosen to form a voice disorder group. Another 10 healthy counterparts constituted a control group. The short-term memory of both groups was assessed and any correlation between pronunciation and short-term memory was analyzed.Results:The children with a developmental speech sound disorder differed significantly from the controls in terms of the numbers of errors in articulating blade-alveolar, blade-palatal and velar consonants. On the phoneme level, the highest substitution error rate occurred when pronouncing lingua-palatal consonants (42.86%), followed by supradental consonants (32%). The highest distortion and non-acquisition error rates were with blade-palatal consonants (14%) and lingua-palatal consonants (9.5%). On the vocabulary level, the highest substitution, distortion, ellipsis and non-acquisition error rates appeared when pronouncing lingua-palatal and velar consonants, velar and blade-palatal consonants, supradental consonants as well as blade-palatal consonants. Significant differences were found between the phoneme and lexical levels in the substitution of supradental and blade-palatal consonants as well as in the ellipsis of blade-alveolar consonants. They were moderately associated with pronunciation level. There was, however, no significant difference in working memory span between the two groups, and no significant correlation was observed between working memory span and pronunciation level.Conclusion:The mispronunciation of consonants by children with developmental speech disorders is higher at the lexical than at the phoneme level. They mainly substitute lingua-palatal and velar consonants and elide supradental consonants, which may be related to short-term memory span.