LEADER 00000nam  2200325   4500 
001    AAI3245623 
005    20121005142518.5 
008    121005s2006    ||||||||||||||||| ||eng d 
020    9781109842289 
035    (UMI)AAI3245623 
040    UMI|cUMI 
100 1  Patel, Rita R 
245 10 High speed digital imaging and kymographic analyses of 
       vocal fold vibrations 
300    162 p 
500    Source: Dissertation Abstracts International, Volume: 67-
       12, Section: B, page: 7045 
500    Adviser: Diane M. Bless 
502    Thesis (Ph.D.)--The University of Wisconsin - Madison, 
520    High speed digital imaging (HSDI) and kymography have the 
       potential to provide detailed biomechanical assessment of 
       vocal fold vibration because capture rates of up to 8000 
       frames per second, make it is possible to assess cycle-to-
       cycle variations that can be extremely aperiodic, brief, 
       and transitory. The overall purpose of the three studies 
       comprising this dissertation was to study normal and 
       disordered vibratory patterns obtained from HSDI and 
       kymography across gender and pathology groups. Invariant 
       moment based features were extracted from the two 
       dimensional Fourier transform to study pattern recognition
       of digital kymographs across gender and three phonatory 
       tasks (normal pitch normal loudness, normal pitch loud and
       high pitch normal loudness) among participants with and 
       without hoarseness in Study A. Principal component 
       analysis of gender and task analysis across participants 
       without dysphonia (Study A) revealed that 90% of the 
       variance was explained in the plane defined by two 
       features: eigen vector 1 and eigen vector 2, suggesting 
       spatial and temporal vibratory differences in phonatory 
       physiology across normal participants. Analysis across 
       participants with and without dysphonia revealed that 
       approximately 82% of the variance was accounted by eigen 
       vector 1 and eigen vector 2 for both male and female 
       participants suggesting that two features capture most of 
       the salient temporal and spatial features of digital 
       kymographs across gender and phonatory tasks. Study B used
       image processing approaches to extract motion waveforms 
       from HSDI to differentiate adductor spasmodic dysphonia 
       from muscle tension dysphonia. Features of aperiodicities,
       oscillatory breaks (as small as 8 ms), and micro motions 
       were greater in number for the spasmodic dysphonia group 
       whereas hyperfunction was greater in muscle tension 
       dysphonia group. Study C investigated the clinical value 
       of HSDI compared to stroboscopy across three laryngeal 
       pathology groups classified as either disorders of 
       epithelium, subepithelium, or neuromuscular. Results 
       revealed that 63% of participants with laryngeal pathology
       could not be judged on stroboscopy because of the severity
       of the voice disorder, and its concomitant aperiodicity, 
       whereas HSDI resulted in analysis of 100% of the data. 
       Neuromuscular group was most difficult to judge with 
       stroboscopy, followed by epithelial and subepithelial 
       groups in that order 
590    School code: 0262 
650  4 Health Sciences, Speech Pathology 
650  4 Engineering, Biomedical 
690    0460 
690    0541 
710 2  The University of Wisconsin - Madison 
773 0  |tDissertation Abstracts International|g67-12B 
856 40 |uhttp://pqdd.sinica.edu.tw/twdaoapp/servlet/