Acoustic analysis of voice/speech characteristics in nonsymptomatic gene carriers of Huntington's disease: Does the speech/voice differ from normal controls?
The study's purpose was to compare preclinical genetic carriers of Huntington's disease with matched controls to determine whether measures of speech motor variability and timing which have been found to be significantly different in people with manifest HD are significantly different for individuals at risk for developing HD. The measures included sustained phonation for vowels, voice onset time, diadochokinetic variability, duration of utterances and controlled speech rate changes.
Eighteen experimental participants and controls performed diadochokinetic tasks, produced vowels for maximum duration, and repeated sentences at regular and increased speed. Productions were recorded digitally and analyzed with Multispeech/CSL software. Differences between the mean values from controls and experimental participants for each parameter were assessed using Unpaired T-tests.
Intratrain coefficient of variation of duration showed more variability in syllable duration for the experimental group. Mean number of syllables per breath was lower for experimental participants. Experimental group showed lower repetition rates for /pa/, /ta/, /pata/, and /pataka/. Mean duration for vowels was lower for the experimental group. Mean jitter, mean shimmer, and mean NHR (noise-harmonic ratio) were higher for the experimental group.
A backward stepwise logistic regression was performed to determine which significant variables were most important. Three variables were retained in the model: intratrain coefficient of variation of duration (ICVD), mean repetition rate per second for the syllables /pata/, when produced for 15 seconds (PT15rep), and mean noise-harmonic ratio (MNHR). Discriminant analysis was performed, using these three variables, and a prediction equation was constructed. The discriminant function for classifying cases showed approximately 89% accuracy for classifying control participants and 72% accuracy for classifying experimental participants.
Participants with the genetic marker for HD showed increased speech variability, reduced speech rate in structured repetition, and vocal instability with increased jitter, shimmer, and NHR. Speech rate, speech and vocal stability, and vocal endurance may ultimately be used in testing asymptomatic carriers for preclinical symptoms.