Parkinson's disease is neurodegenerative disease that often causes motor speech disorder, named hypokinetic dysarthria. Speech difficulties include decreased volume, hoarseness, monotonous tone and imprecise articulation. These can be the first sign of the illness. Speech and language therapists usually relay on auditory-perceptual judgements for the assessment and treatment of dysarthria. Acoustic analysis of speech could be used for more precise detection of speech deviations, as it is more sensitive and gives more accurate results. One of the most frequently used measures for studying the articulation of vowels in dysarthric speech is Vowel Space Area. However, this measure is highly sensitive to inter-speaker variability. To overcome these disadvantages, scientists introduced two alternative acoustic metrics: Vowel Articulation Index and Formant Centralization Ratio.
The study aimed to use acoustic analysis for exploring vowel articulation in Parkinson’s disease speakers. We included five Slovenian speakers with Parkinson’s disease, three women and two men. They were 53 to 73 years old and were diagnosed between 1 to 10 years ago. Five age-matched and gender-matched healthy speakers were recruited as a control group. Each participant was recorded while reading a passage and in spontaneous speech. Recording samples were then analysed in PRAAT and frequencies of the first and second formant of corner vowels (/i/, /a/, /u/) were measured. We drew the vowel triangle and calculated Triangular Vowel Space Area and Vowel Articulation Index for all speakers. With nonparametric statistical tests we compared data from parkinsonian speakers with data from matched healthy controls.
The results indicate that speakers with PD produce vowel /a/ higher in the oral cavity and vowel /u/ more forward, than matched control speakers. The majority of speakers with Parkinson’s disease had a smaller Vowel Space Area and lower Vowel Articulation Index than healthy speakers, although the differences between groups were not always statistically significant. We found that Vowel Articulation Index differentiated between the two groups in both speech tasks and Vowel Space Area only in reading. The reading task thus seems to be more appropriate for uncovering speech characteristics of persons with Parkinson’s disease.
In the thesis we examined acoustic-articulatory characteristics of vowels in Parkinson’s disease and pointed out the differences between healthy speech articulation. The obtained results represent the beginning of the acoustical analysis of Slovenian dysarthric speech and the starting point for further research. They also contribute to a better understanding of changes in speech due to Parkinson’s disease and are an incentive for speech and language therapists to include objective assessment methods in their everyday work.
|
