Phonovibrography: development of a computerized system for the diagnosis of dysphonia.

• Applicants: Professor Dr.-Ing. Jörg Lohscheller; Professor Dr. Bernhard Schick
• Subject Area: Otolaryngology, Phoniatrics and Audiology
• Term: from 2008 to 2017
• Project identifier: Deutsche Forschungsgemeinschaft (DFG) - Project number 81282454

The time-variant vibration-patterns of vocal folds are the fundamental characteristic of voice production in the larynx. Among modern imaging techniques, high-speed videoendoscopy constitutes the only imaging technology that captures the vibratory behavior of the rapidly moving vocal folds with sufficient temporal resolution. While imaging techniques improved significantly there is still no clinical approved procedure to analyze vocal fold movement from high-speed videos. For the first time, the recently developed phonovibrogram offers a compact description of vocal fold movement from the previously captured high-speed videos. In this manner, a methodology to classify voice disorders based on phonovibrograms was developed within the initial application phase. In a cross-sectional study, a comprehensive video-database of sustained phonation sequences served as training dataset for the classification system.Based on the developed analyzing techniques, there are two major goals in the following application phase. The novel wavelet-based analyzing technique allows for the first time a comprehensive characterization of stationary as well as non-stationary phonation paradigms. In this context, the classification system will be extended to transient phonation paradigms to gather complementary information for diagnosing purposes. In particular, phonation onset and influences of varying frequency and intensity on vibratory patterns will be investigated. The extension of the feature extraction process to transient paradigms intends to describe laryngeal dynamics comprehensively and therefore helps to optimize the classification procedure.High-speed videoendoscopy is currently employed in a subjective manner to assess the outcome of phonosurgical interventions and voice therapies. Our research will address the issue of objectifying therapy results by means of a computer-aided analysis of endoscopic high-speed videos. To estimate the feasibility of this automated assessment a prospective longitudinal study will be performed to quantize the success of phonosurgeries and voice therapies. For this purpose, 120 phonosurgery and voice therapy patients will be attended through their therapy and investigated regularly. The aim is to develop an evidence-based procedure to assess the outcome of voice therapies and further to specify the intra- and interspecific variability of healthy and pathologic voices.

• DFG Programme: Research Grants