Date of Defense

4-25-2019

Date of Graduation

4-2019

Department

Speech, Language and Hearing Science

First Advisor

Stephen Tasko

Second Advisor

Ronald Kelley

Abstract

Purpose: Individuals who stutter can often exhibit speech patterns that sound unnatural compared to individuals who do not stutter. Unnaturalness can be due to several underlying factors including the presence of stuttering itself and the unintended consequences of treatment approaches that alter speech motor patterns. Understanding factors that impact speech naturalness within the stuttering population can help improve the clinical management of the disorder. The current study investigates the relationship between pause duration, speech rate and stuttering frequency, and listener ratings of speech naturalness in a group of adults who stutter (AWS) before and after participation in a stuttering treatment program.

Methods: Participants include 34 individuals who stutter, who were enrolled in a group-based, intensive, four-week fluency-shaping stuttering treatment program. One-minute audio video recordings were obtained from monologue speaking activities recorded prior to and following stuttering treatment. The audio track from each video sample was extracted and submitted to an acoustic analysis software to extract two measures. Pause duration was determined by identifying the temporal extent of all pauses greater than 250 ms. Fluent syllable rate was determined by measuring the overall duration and syllable count drawn from three perceptually fluent runs of speech in each video sample. These measures were combined with speech naturalness ratings and stuttering frequency data reported in previous studies conducted on this participant pool. Paired t-tests were used to assess pre vs. post-treatment differences in mean pause time, fluent syllable rate, mean naturalness ratings and stuttering frequency. Linear regression was used to identify associations between pause time, syllable rate, stuttering frequency, and speech naturalness.

Results: Results revealed that following participation in stuttering treatment, the speech samples were rated as more natural, (t: 1.92; p=0.03), exhibited a dramatically reduced stuttering frequency (t: 10.18; p<0.00005), had significantly longer pause durations (t: -4.1145; p=.0001), and showed no change in fluent syllable rate. Linear regression results indicated that prior to treatment, stuttering frequency was the strongest predictor of speech naturalness, accounting for over 60 % of the variance in naturalness ratings. Pause duration and syllable rate provided little predictive value for pre-treatment speech naturalness ratings. Following treatment, pause duration was the strongest predictor of speech naturalness, accounting for more than 50 % of the variance in naturalness ratings. Syllable rate and stuttering frequency were not significantly associated with speech naturalness ratings.

Conclusion: While there is a great deal of individual variation, speech samples of persons who stutter are commonly rated as unnatural sounding to listeners. This is true even after participating in some treatment programs designed to reduce or eliminate stuttering. Prior to treatment, stuttering frequency is the best predictor of speech naturalness. However, following treatment, stuttering frequency ceases to have predictive value. Instead, it was found that the duration of pausing between successive runs of speech is a much better predictor of speech naturalness following treatment. These two findings suggest that, within the same group of speakers, speech naturalness ratings are influenced by different kinds of factors. One interpretation is that following treatment, participants used longer speech pauses to plan the integration of the fluency enhancing speech targets. The result was relatively fluent, but still unnatural sounding speech. Results from the current study may provide future speech-language pathologists with information contributing to improved speech naturalness as an outcome of intervention in stuttering treatment programs.

Access Setting

Honors Thesis-Open Access

Hausman Thesis PowerPoint FINAL.pdf (913 kB)
Defense Presentation

Share

COinS