Objective assessment of vocal fatigue in laboratory and real-world settings

项目摘要/摘要 约30%的美国成年人在一生中受到语音障碍的影响，约有2500万人 在任何给定的时间点经历语音障碍，导致社会成本(失去工作，医疗费用， 等)估计每年135亿美元。发声疲劳(VF)被认为是一种病因和/或反应性成分 在最常见的发声障碍中，也是最常见的与发声有关的个人抱怨之一 依靠自己的声音谋生的人(例如，教师、歌手等)。以前对VF的定义各不相同 最近试图达成共识，将VF描述为一个多方面的概念，涉及个人的自我- 感觉到的症状(例如，增加的努力和不适)和/或与发声功能恶化相关的 一个人试图满足他/她的声音需求。先前的研究表明，说话人与说话人之间的高对比度 说话人在从VF中恢复时的可变性，但与如此高的可变性相关的因素尚不确定。 不幸的是，关于发声机制的原因和影响的客观信息很少。 VF的发生，限制了预防和临床处理这种常见疾病的努力。 该项目的目标是(1)使用多模式测量方法来全面、 客观地描述VF对发声功能的渐进性影响，(2)量化潜在的发声-休息 有助于室颤进展和恢复的行为，以及(3)识别发声功能和 可以解释VF恢复轨迹中观察到的高变异性的发声行为参数。这个 这项研究的目标是使用现有的动态语音监测数据集和 新数据。现有的数据集包括87个典型发声个体的动态语音记录(对于 共889天)，123例发声功能亢进患者(共763天)。新数据将包括 控制良好的室内语音加载方案，以及使用最先进的无线技术进行为期三天的内场监控 监控系统。实验室会议将包括定期获取发声功能的多模式数据(高 快速视频内窥镜、空气动力学、电声门图、声学和颈面加速度计) 加载协议的进展。机器学习的统计能力与两种新的 声带耗散能量(反映发声功能)的动态测量与发声休息时间序列 Ratio(反映发声行为的发声和休息期的时间序列)，以量化进程 以及根据一个人先前累积的发声行为和发声功能恢复室颤。 实现该项目的目标将为开发新的临床预防工具奠定基础， 评估和缓解VF，这对要求较高的职业的专业人员特别有价值 语音使用。

Project Summary/Abstract Approximately 30% of US adults are affected by a voice disorder during their lives, with about 25 million people experiencing a voice disorder at any given point in time, resulting in societal costs (lost work, medical expenses, etc.) estimated at $13.5 billion annually. Vocal fatigue (VF) is viewed as an etiological and/or reactive component in most common voice disorders and is also among the most common voice-related complaints of individuals who rely on their voices to make a living (e.g., teachers, singers, etc.). Previous definitions of VF have varied with a recent attempt at consensus describing VF as a multifaceted concept that involves an individual’s self- perceived symptoms (e.g., increased effort and discomfort) and/or a deterioration in vocal function associated with an individual’s attempt to meet his/her vocal demands. Prior studies have demonstrated high speaker-to- speaker variability in recovery from VF, but the factors associated with such high variability are not determined. Unfortunately, there is a paucity of objective information about the causes and impact on phonatory mechanisms of VF, limiting efforts to prevent and clinically manage this common complaint. The objectives of this project are (1) to use a multi-modal measurement approach to comprehensively, and objectively, describe the progressive impact of VF on vocal function, (2) to quantify the underlying voicing-resting behaviors that contribute to the progression of VF and its recovery, and (3) to identify the vocal function and vocal behavior parameters that could account for the observed high variability in VF recovery trajectories. The objectives of this study are pursued using a combination of an existing ambulatory voice monitoring dataset and new data. The existing dataset includes ambulatory voice recordings from 87 vocally typical individuals (for a total of 889 days) and 123 patients with vocal hyperfunction (for a total of 763 days). The new data will include a well-controlled inlab vocal loading protocol, and three days of infield monitoring using a state-of-the-art wireless monitoring system. The inlab session will include periodic multi-modal data acquisition of vocal function (high- speed videoendoscopy, aerodynamics, electroglottography, acoustics, and neck-surface accelerometry) during the progression of the loading protocol. Statistical power of machine learning is combined with two novel ambulatory measures of vocal fold dissipated energy (reflecting vocal function) and a time series voicing-resting ratio (the temporal sequencing of phonatory and resting periods reflecting vocal behavior) to quantify progression and recovery of VF in terms of a person’s prior cumulative vocal behavior and vocal function. Achieving the goals of this project will lay the groundwork for the development of new clinical tools for preventing, assessing, and alleviating VF, which will be particularly valuable for professionals in occupations requiring heavy voice use.