Title page for ETD etd-06262012-110651

Type of Document Dissertation
Author Wei, Zhenyi
Author's Email Address zwei2@lsu.edu
URN etd-06262012-110651
Title Dynamic Characterization of Vocal Fold Vibrations
Degree Doctor of Philosophy (Ph.D.)
Department Mechanical Engineering
Advisory Committee
Advisor Name Title
de Queiroz, Marcio Committee Chair
Gu, Guoxiang Committee Member
Kunduk, Melda Committee Member
Pang, Su-Seng Committee Member
Beck, Stephen David Dean's Representative
  • vocal folds
  • vibration
  • dynamics
  • modeling
  • high-speed digital imaging
Date of Defense 2012-05-09
Availability unrestricted

An emerging trend among voice specialists is the use of quantitative protocols for the diagnosis and treatment of voice disorders. Vocal fold vibrations are directly related to voice quality. This research is devoted to providing an objective means of characterizing these vibrations. Our goal is to develop a dynamic model of vocal fold vibration, and map the parameter space of the model to a class of voice disorders; thus, furthering the assessment and diagnosis of voice disorder in clinical settings.

To this end, this dissertation introduces a new seven-mass biomechanical model for the vibration of vocal folds. The model is based on the body-cover layer concept of the vocal fold biomechanics, and segments the cover layer into three masses along the longitudinal direction of the vocal fold. This segmentation facilitates the model comparison with the motion of the vocal glottis contour derived from modern high-speed digital imaging systems. The model simulation is compared to 14 sets of experimental data from human subjects with healthy vocal folds and pathological vocal folds including nodule, polyp, and unilateral paralysis. We also propose a semi-empirical two-stage procedure for tuning the parameters so that the model response matches as closely as possible the experimental data in the time and frequency domains. The first stage involves the manual coarse tuning of parameters based on limited data to expedite the process. The second stage is an automatic (or manual) fine tuning process on a subset of the parameters tuned in the first stage based on a larger amount of data.

Once an ‘optimal’ set of model parameters has been identified, two model-based factors, quantifying the asymmetry between left and right vocal folds and anterior and posterior segments of the vocal folds, are introduced and calculated for each of the 14 cases. The two factors form an asymmetry plane. Based on the value of the asymmetry factors for the 14 cases, the plane is subdivided into four regions corresponding to healthy vocal folds, nodule, polyp, and unilateral paralysis. This yields a clear visual aid for clinicians, correlating the model parameters to voice quality.

  Filename       Size       Approximate Download Time (Hours:Minutes:Seconds) 
 28.8 Modem   56K Modem   ISDN (64 Kb)   ISDN (128 Kb)   Higher-speed Access 
  wei_diss.pdf 4.27 Mb 00:19:46 00:10:10 00:08:54 00:04:27 00:00:22

Browse All Available ETDs by ( Author | Department )

If you have questions or technical problems, please Contact LSU-ETD Support.