Syllabic tone variation by Sepedi speakers with dysarthia

Abstract

Background: Speech production in Bantu languages places great demands on neuromotor control, because unique speech motor behaviours such as syllabic tone variation and the aspiration of speech sounds require an additional level of vocal fold control compared to speech production in Germanic languages. As these motor behaviours play an important role in differentiating the meaning of words (Van der Merwe & Le Roux, 2014a), neuromotor speech disorders such as dysarthria may have a greater impact on communication in Bantu languages than in Germanic languages. The focus of this study was on syllabic tone variation in Bantu language speakers with dysarthria compared to typical speakers. Sepedi was the Bantu language investigated. Syllabic tone variation refers to pitch level changes for every syllable of words in a tone language (Zerbian & Barnard, 2008a) and requires manipulation of vocal fold length and mass over and above the voicing or devoicing of sounds within words. These pitch changes convey the lexical and grammatical meaning of words and may differentiate between the meanings of two orthographically identical words (Zerbian & Barnard, 2008a). Studies on lexical tone variation in speakers with dysarthria to date have focused mostly on the tone languages of Asia and Scandinavia (Kadyamusuma, De Blesser, & Mayer, 2011). No studies of tone variation in Bantu language speakers with dysarthria were found. Furthermore, past research only regarded tone variation in monosyllabic words, with no reference to how tone would be affected across bisyllabic words and within each of the two syllables of these words. No inquiries were made into the tone variation ability of speakers with dysarthria when producing short utterances compared to longer utterances and mostly speakers with congenital dysarthria were used as research participants. These shortcomings needed to be addressed to gain a more holistic and accurate view of the extent to which tone variation is a challenge for Bantu language speakers with dysarthria. Aims: The first aim of the study was to determine whether a difference exists between typical Sepedi speakers and Sepedi speakers with dysarthria, in their ability to vary tone across CVCV words with a HL tone pattern. The second aim of the study was to determine whether a difference in tone variation exists between short and longer utterances in typical Sepedi speakers and Sepedi speakers with dysarthria. Method: A quasi-experimental, between-group comparison was used in the study. Speech samples were obtained from a control group of five typical Sepedi speakers and from an experimental group of four Sepedi speakers with dysarthria. These speech samples consisted of 20 consonant-vowel-consonant-vowel (CVCV) words with high-low (HL) tone variation produced in three- and also in six- /seven-syllable utterances (resulting in a total of 40 words). The speech samples were analysed acoustically using Praat software. To achieve the first aim, the following acoustic measures were obtained from the 40 words produced by participants: (1) Mean fundamental frequency (F0) of syllable 1 (S1) and syllable 2 (S2), (2) Change in F0 across words from the highest F0 point of S1 to the lowest F0 point of S2, (3) Intrasyllabic change in F0 within S1 and S2. To achieve the second aim of the study, the change in F0 across words in short utterances was compared to the change in F0 across words in longer utterances for the typical speakers and speakers with dysarthria. Results: Wilcoxon rank tests were used for statistical analyses. Descriptive statistics were performed and median values were used to achieve research aims. All of the control participants and participants with dysarthria produced S1 with a higher mean F0 than S2, as was appropriate for the HL tone pattern ascribed to the target words. For most of the individuals from both groups, the mean F0 of S1 was significantly higher than the mean F0 of S2. However, one participant from each group produced an insignificant difference between the mean F0 values of the two syllables. The control group produced slightly greater median F0 changes across the words and within S1 than the dysarthria group, but the differences between the speaker groups for the change in F0 across words and the change in F0 within S1 were insignificant. In contrast to this, the control group produced a significantly smaller median change in F0 within S2 than the dysarthria group. Individual speakers from both groups produced unique patterns of F0 changes for all aspects of tone variation (change in F0 across words and changes in F0 within S1 and S2). Both speaker groups produced a significantly greater median change in F0 across words in short utterances compared to long utterances. The difference in the change in F0 across words between short and long utterances was significantly greater for the control group than for the dysarthria group. Conclusions: The speakers with dysarthria in the study maintained the ability to vary tone across bisyllabic words with an HL tone pattern. The dysarthria group only differed significantly from the control group with regard to the extent of tone reduction in the second syllable. This finding may point to possible difficulties in the required graded relaxation of the vocal folds. Individual differences in F0 changes were found for both typical speakers and speakers with dysarthria, indicating that unique tone variation patterns may normally exist for all speakers. For both control and dysarthria groups, greater tone variation was observed in short compared to longer utterances. The role of increased utterance length in decreased F0 variation was greater for the typical speakers than for the individuals with dysarthria.