Date of Award:

5-2024

Document Type:

Thesis

Degree Name:

Master of Science (MS)

Department:

Communicative Disorders and Deaf Education

Committee Chair(s)

Annalise R. Fletcher

Committee

Annalise R. Fletcher

Committee

Stephanie A. Borrie

Committee

Alan A. Wisler

Abstract

Hypokinetic dysarthria is a motor speech disorder that occurs in patients with Parkinson’s disease (PD) and affects not only their speech intelligibility but also how their attitudes and emotions are perceived by listeners. People with PD have been judged as less happy, involved, friendly, and interested based only on their speech samples. A lack of speech expressiveness is one of the characteristics that is likely to be related to these negative listener judgments. Specifically, it has been suggested that a lack of fundamental frequency (F0) variation reduces speakers’ ability to express various emotions. To investigate whether speech expressiveness is related to F0 variation, it is necessary to accurately measure how F0 changes across sentences. However, existing software produces a lot of tracking errors when measuring F0, so the accuracy of various statistical measures of F0 is questionable. The current study answered two research questions. First, we evaluated one of the most widely used software applications in the field of speech disorders to assess how tracking errors affected measurements of F0 mean and standard deviation in healthy speakers and people with PD. This was done by manually annotating all errors and calculating summary statistics that included and excluded these values. Second, we explored different statistical approaches to measuring F0 variability to determine which measurements had the strongest relationship with speech expressiveness. To address this question, listeners were presented with 90 sentences, read by both healthy speakers and speakers with PD, and were asked to rate “How expressive is this speaker?” using a visual analog scale. Statistical analysis evaluated the relationship between acoustic and perceptual variables and compared the results 1) before and after excluding tracking errors in F0 measurement; 2) when six different statistical approaches to measuring F0 variability were substituted into the model. Results demonstrated that tracking errors in the F0 contour significantly affect F0 summary statistics. This finding highlights the importance of manually screening and removing outliers when evaluating group differences in F0 statistics. Results also demonstrated that some (but not all) of the statistical approaches for measuring F0 variability show significant relationships with listeners’ perceptions of speech expressiveness. A larger number of approaches showed a significant relationship with expressiveness scores when tracking errors were excluded from the analysis and the approaches that best accounted for listener ratings provided some normalization of sex-related differences in frequency values. These findings offer insight into the type of changes in F0 that are most relevant to listeners in their perception of speech expressiveness.

Checksum

cf54a48021ef9e4c2e6e44c4da12c3de

Download

Share

COinS

Copyright for this work is retained by the student. If you have any questions regarding the inclusion of this work in the Digital Commons, please email us at .