Chaos in human phonation and its measurement

点击翻译按钮获取中文摘要

PROJECT SUMMARY Voice disorders caused by physical injury, voice misuse, physiological disease, and surgery represent a significant social cost, resulting in severe functional and psychological limitations on the lives of millions of Americans. Acoustic analysis represents a promising tool for clinical assessment of voice disorders; however, it has not been widely applied outside of research despite its low cost, objectivity, and non-invasive nature. The intrinsic complexity of voice production and aperiodicity of disordered voice necessitates that acoustic analysis be nonlinear. Current nonlinear parameters are capable of quantifying both normal and disordered phonation; however, existing nonlinear methods fail to comprehensively describe the non-stationary, dynamic elements comprising voice signals. The long-term goal of the proposed research is to expand the utility of nonlinear dynamic analysis in clinical practice as an objective means for detecting vocal dysfunction and monitoring voice rehabilitation following treatment intervention. In Aim 1, excised larynx models will be employed to simulate and investigate physiological conditions contributing to breathy and rough phonation, including extreme subglottal pressures, vocal fold elongation, asymmetrical tension, and benign mass lesions. Diffusive chaos and intrinsic dimension nonlinear analysis will be implemented to evaluate the four different type components present in the corresponding acoustical output and to subsequently construct voice type component profiles (VTCP). Quantitatively describing the VTCP will provide more descriptive information pertaining to the physiological location and conditions underlying pathological phonation. Aim 2 will focus on applying VTCP nonlinear analysis to normal and disordered patients in a variety of voice tasks, such as vowel and consonant utterances and connected speech, and assessment of clinical treatment interventions, including surgical treatment of benign mass lesions and vocal fold paralysis and rehabilitative voice training for esophageal and tracheoesophageal speech. To facilitate wider and more extensive utility of nonlinear dynamic analyses, we will assess the concurrent validity of the VTCPs with the clinically established technique of auditory-perceptual assessment of voice quality. The likely impact of this research is that innovative methods of diffusive chaos and intrinsic dimension nonlinear analysis will provide significant improvements over previous acoustical parameters in both computationally efficiency and voice quality description. Moreover, acoustical analysis of disordered phonation in excised larynges will provide a physical model to examine the physiological and biomechanical conditions underlying pathologic voice production, which remains poorly understood. Lastly, the clinical utility of VTCP analysis as a diagnostic tool for voice disorder detection will be demonstrated.