Mechanisms of accented speech recognition in native and non-native listeners: Biological insights

母语和非母语听众的重音语音识别机制：生物学见解

• 批准号: 10375557
• 负责人: Jennifer Krizman
• 金额: $ 16万
• 依托单位: NORTHWESTERN UNIVERSITY
• 依托单位国家: 美国
• 财政年份: 2021
• 资助国家: 美国
• 起止时间: 2021-04-01 至 2024-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/10375557
• 关键词: Accent; Acoustics; Adult; Attention; Auditory; Biological; Businesses; Cognitive; Communication; Communication impairment; Complex; Cueing for speech; Data; Dependence; Development; Elements; Evaluation; Familiarity; Frequencies; Goals; Grain; Grouping; Human; Individual; Individual Differences; Language; Language Disorders; Linguistics; Measures; Noise; Outcome; Participant; Performance; Persons; Process; Role; Signal Transduction; Societies; Speech; Speech Acoustics; Stimulus; Testing; Vocabulary; aged; auditory processing; auditory stimulus; base; bilingualism; cohort; executive function; experience; experimental study; hearing impairment; improved; insight; language processing; middle age; neuromechanism; neurophysiology; non-Native; racial and ethnic; racial diversity; relating to nervous system; remediation; response; skills; socioeconomics; sound; speech in noise; speech recognition

PROJECT SUMMARY The goal of this study is to identify the auditory, cognitive, and linguistic processes that support foreign-accented speech recognition for native and non-native listeners. With increasing globalization, foreign-accented speech occurs regularly in real-world human communication and is often a cause of miscommunication. Yet little is known about how non-native speech alters the mechanisms of speech recognition. To understand the mechanisms that support accented-speech recognition and their language-dependent plasticity, 60 middle-aged adults from three different language backgrounds (native English monolingual, Spanish-English bilingual, and Mandarin-English bilingual) are tested on accented-speech recognition, the frequency-following response (FFR), and a battery of cognitive and linguistic measures. The accented-speech recognition test uses sentences of varying linguistic complexity spoken by native-English monolinguals, Spanish-English bilinguals, and Mandarin- English bilinguals, allowing evaluation of non-native speech recognition in individuals whose language backgrounds either match or do not match the talkers. The FFR is a neurophysiological response to complex auditory stimuli that provides fine-grained detail about how multifaceted, overlapping components of sound (e.g., fundamental frequency and harmonics) are transformed into discrete neural components. Thus, the FFR can identify how individual sound components contribute to non-native speech recognition. Additionally, a comprehensive battery of cognitive and linguistic processes thought to be important for non-native speech recognition are measured on these participants. Together, these tests delineate the contribution of auditory, cognitive, and linguistic processes on foreign-accented speech recognition for native and non-native listeners. By identifying the neural mechanisms underlying accented-speech recognition and language-based plasticity of these mechanisms, this project provides the groundwork for developing strategies to improve accented speech recognition. Furthermore, outcomes from this project will aid in our overarching goal of understanding the shared and separate mechanisms that support all types of degraded-speech recognition and how these mechanisms are influenced by language experience. Identifying mechanisms of degraded-speech recognition and their experience dependency will identify ways to improve communication in everyday settings, and will facilitate development of remediation strategies for individuals whose speech recognition difficulties are exacerbated, such as individuals with a hearing impairment or language disorder.

项目摘要 本研究的目的是确定支持外国口音的听觉、认知和语言过程 为母语和非母语听众提供语音识别。随着全球化的发展，外国口音的讲话 在现实世界的人类交流中经常发生，并且经常是错误沟通的原因。但很少有人 了解非母语语音如何改变语音识别机制。了解 支持重音语音识别的机制及其语言依赖的可塑性，60名中年人 来自三种不同语言背景的成年人（母语英语单语，西班牙语-英语双语， 普通话-英语双语）进行了重读语音识别，频率跟随反应（FFR）， 以及一系列认知和语言测试。重音语音识别测试使用以下句子 母语为英语的单语者、西班牙语-英语双语者和普通话者所讲的语言复杂性各不相同， 英语双语者，允许评估其语言 背景匹配或不匹配讲话者。血流储备分数是对复杂的 听觉刺激提供关于声音的多面、重叠分量（例如， 基频和谐波）被变换成离散的神经分量。因此，FFR可以 识别各个声音成分对非母语语音识别的贡献。此外， 一套综合的认知和语言过程，被认为对非母语语言很重要 对这些参与者的认可。总之，这些测试描绘了听觉， 认知和语言过程的外国口音的语音识别为母语和非母语的听众。 通过识别重音语音识别和语言可塑性的神经机制， 这些机制，这个项目提供了基础，发展战略，以改善口音讲话 识别.此外，该项目的成果将有助于我们实现理解共同的 以及支持所有类型的降级语音识别的独立机制，以及这些机制如何 受语言经验的影响。退化语音识别的识别机制及其应用 经验依赖将确定如何改善日常环境中的沟通，并将促进 为语音识别困难加剧的个人制定补救策略， 例如具有听力损伤或语言障碍的个体。