Perceptual parameters of speech production in childhood apraxia of speech : the effects of speech motor learning treatment

Abstract

Background. The diagnosis and treatment of childhood apraxia of speech (CAS), a sensorimotor speech disorder in children, is both challenging and controversial. Uncertainty exists regarding the disorder’s underlying nature and its salient features. This uncertainty impacts the formulation of treatment approaches. Numerous approaches for CAS treatment have been suggested, based on different authors’ theoretical conceptualisation of the disorders’ underlying nature and its characteristic features. Research on many of these approaches report improvements in items targeted in treatment, but generalisation to untreated items and maintenance of skills varies. Factors which influence the outcomes of treatment include the treatment methods, as well as the types of targets selected. Authors tend to attribute improvements achieved by participants to the methods employed in a particular treatment. Comparatively little theoretical consideration is given to the selection of targets and the influence this has on treatment, generalisation and maintenance effects. The current study investigated the speech motor learning (SML) approach (Van der Merwe, 1985; 2011), which is based on a comprehensive speech motor planning and programming model (Van der Merwe, 1997; 2009). This model aids one in understanding the nature of CAS and predicts many of its characteristic features. Furthermore, the model provides guidelines for treatment targets in particular, but also for treatment methods. Aims. The primary aim was to determine whether the SML approach would effect a positive change in the speech production of a child with CAS, as measured by whole word accuracy. The sub-aims were to determine whether treatment effects would generalise to untreated non- and real-word stimuli containing: 1) only treated age-appropriate consonants and five vowels (Set 1 sounds); 2) in addition, untreated age-appropriate consonants and the same five vowels (Set 2 sounds); 3) in addition, untreated age-inappropriate consonants and the same five vowels (Set 3 sounds). Method. A single-case experimental design, with multiple baselines and multiple probes across behaviours was implemented. The Participant, a 33-month old boy, presented with a cluster of CAS symptoms. He received 18 sessions of treatment according to the SML approach (Van der Merwe, 1985; 2011) over 9 weeks. Three baseline, nine treatment, and two follow-up probes were performed across the study duration. Probe stimuli comprised untreated non- and real-words containing treated (Set 1) and untreated (Set 2: age-appropriate and Set 3: age-inappropriate) sounds. Whole word accuracy was determined through perceptual analysis. Accuracy scores were compared through effect size (ES) calculations. A post-hoc analysis on error types was performed to explore the effect of improvement on the nature of errors across time. Results. The Participant demonstrated a significant improvement in the accuracy of treated sounds (Set 1), while less or no change was observed in untreated age-appropriate (Set 2) and untreated age-inappropriate (Set 3) consonants respectively. The post-hoc analysis revealed a predominance of phonetic-motoric errors throughout the study. Set 3 real words showed a small increase in phonological errors, with a coinciding decrease in the number of words that were distorted to such an extent that individual sounds could not be distinguished. Conclusions. The SML approach was effective in improving the Participant’s ability to plan and coarticulate sounds, resulting in significantly improved production of treated sounds in untreated nonword exemplars and untreated real words. Generalisation of treatment gains to untreated stimuli provides support for the target selection of the SML approach, in addition to the SML methods as implemented in this study. The predominance of phonetic-motoric errors supports the notion that CAS is a disorder of motor planning and/or programming.