Token‐to‐token variability in adult apraxia of speech: A perceptual analysis

Abstract

Background: Many researchers and clinicians have suggested that a definitive characteristic of adult apraxia of speech (AOS) is that the speech sound errors are highly variable. Furthermore, it has been argued that error variability is one means of differentially diagnosing AOS from dysarthria. The findings from studies of error variability have been mixed; however, the speech errors have been elicited in different ways and “variability” is not consistently defined across different studies, making comparisons across studies difficult.

Aims: The purpose of this investigation was to characterise error variability in two speakers with AOS.

Methods & Procedures: Single words were elicited in imitation with repeated trials on the same words over a period of several weeks. The speech samples were transcribed using narrow phonetic transcription and analysed for accuracy of production and consistency of error.

Outcomes & Results: We found that the errors were both consistent and inconsistent, depending on how the errors were analysed, e.g., certain sounds were consistently produced in error. In addition, our data suggest that consistency of error may be related to the severity of the AOS.

Conclusions: Future research should be aimed at standardising the definitions of consistency, and empirically testing the variables that may affect the observed variability. Furthermore, we need more data on error consistency from a large number of well‐described individuals with a variety of neurogenic disorders, as well as continued refinement of models of speech production that can account for the error patterns that are observed.

Keywords:

• Speech
• Apraxia
• Variability
• Motor