Segments and syllables in the treatment of apraxia of speech: An investigation of learning and transfer effects

Abstract

Background: In the treatment of severe apraxia of speech (AOS), single phonemes are often used as training targets. However, until now it remains unclear if segment production can be learned and if patients show transfer effects when a segment trained in isolation is embedded in syllables and words. Speech motor learning in normal language acquisition is based on syllabic rather than segmental learning mechanisms. The syllable is also assumed to be the basic unit of articulatory programming in normal speech.Aims: In the present study, we investigated the effectiveness of learning single segments as compared to whole syllables in patients with AOS. Besides immediate learning effects, we were particularly interested in transfer effects on trained targets embedded into larger untrained units and in maintenance effects.Methods & Procedures: Four patients with severe AOS participated in this study. Target segments and target syllables were trained in two separate phases. To examine learning effects, we presented trained and untrained items in isolation immediately before and after a learning phase. Transfer effects of the segmental learning were examined by presenting syllables which contained the target and control phonemes. The transfer of syllabic learning was assessed on the basis of two-syllabic words containing the target and control syllables.Main Contributions: In the segmental condition, only one patient showed a significant learning effect. Transfer effects onto untrained syllables containing the trained consonants were missing in all patients. After the syllabic training, three patients produced significantly lower error rates on the trained, but not on the control syllables. In two of them, the improvements were still maintained when the syllables were embedded in two-syllabic words.Conclusions: The lack of segmental learning effects in three patients may be due to the fact that single consonants are artificial speech units we normally do not produce in isolation. The complete absence of transfer effects suggests that intrasyllabic coarticulation is part of the phonetic plan and cannot be mastered by speech apraxic patients without extra exercise. The observed learning effects after the syllabic learning confirm that syllables constitute more natural speech motor units which are easier to re-acquire than single consonants. Furthermore, the embedding of learned monosyllables into unlearned two-syllabic words requires less extra speech motor planning than the embedding of a trained segment into the coarticulated context of a syllable.