NEURAL NETWORKS FOR SPEECH PERCEPTION IN NOISE

噪声中语音感知的神经网络

• 批准号: 7194239
• 负责人: LEE M MILLER
• 金额: $ 29.03万
• 依托单位: UNIVERSITY OF CALIFORNIA AT DAVIS
• 依托单位国家: 美国
• 财政年份: 2006
• 资助国家: 美国
• 起止时间: 2006-04-01 至 2011-03-31
• 项目状态: 已结题
• 来源: https://reporter.nih.gov/project-details/7194239
• 关键词: Acoustics; Address; American; Area; Auditory; Biological Neural Networks; Brain; Brain region; Broca' s area; Cognitive; Communication; Complex; Comprehension; Condition; Cues; Daily; Devices; Diagnosis; Drug Formulations; Environment; Foundations; Frequencies; Functional Magnetic Resonance Imaging; Hearing; Hearing Aids; Human; Illusions; Insula of Reil; Knowledge; Left; Linguistics; Magnetic Resonance Imaging; Mediating; Mediation; Mental Depression; Methods; Movement; Neurobiology; Noise; Oral cavity; Parietal; Play; Process; Psychophysics; Public Health; Research; Research Personnel; Restaurants; Role; Scientist; Secure; Sensory; Signal Transduction; Social isolation; Specific qualifier value; Specificity; Speech; Speech Intelligibility; Speech Perception; Speech Sound; Stimulus; Structure of superior temporal sulcus; Temporal Lobe; Testing; Time; Visual; Visual Motion; Voice; Work; base; design; hearing impairment; improved; innovation; lexical; member; programs; relating to nervous system; repaired; research study; restoration; selective attention; sound; Acoustics; Address; American; Area; Auditory; Biological Neural Networks; Brain; Brain region; Broca' s area; Cognitive; Communication; Complex; Comprehension; Condition; Cues; Daily; Devices; Diagnosis; Drug Formulations; Environment; Foundations; Frequencies; Functional Magnetic Resonance Imaging; Hearing; Hearing Aids; Human; Illusions; Insula of Reil; Knowledge; Left; Linguistics; Magnetic Resonance Imaging; Mediating; Mediation; Mental Depression; Methods; Movement; Neurobiology; Noise; Oral cavity; Parietal; Play; Process; Psychophysics; Public Health; Research; Research Personnel; Restaurants; Role; Scientist; Secure; Sensory; Signal Transduction; Social isolation; Specific qualifier value; Specificity; Speech; Speech Intelligibility; Speech Perception; Speech Sound; Stimulus; Structure of superior temporal sulcus; Temporal Lobe; Testing; Time; Visual; Visual Motion; Voice; Work; base; design; hearing impairment; improved; innovation; lexical; member; programs; relating to nervous system; repaired; research study; restoration; selective attention; sound

DESCRIPTION (provided by applicant): This research addresses the neural bases of speech perception in noisy environments. With its singular role in communication, speech is perhaps the most important everyday stimulus for human beings. Yet rarely does speech occur under pristine conditions; competing voices, reverberations, and other environmental sounds typically corrupt the signal. This poses a continual challenge for normal listeners and especially for those with hearing loss. Among the 30 million Americans with hearing loss, many suffer depression and social isolation because of their difficulty communicating. In the half-century since the original formulation of the "cocktail party effect", scientists have established three key perceptual/cognitive factors that improve speech intelligibility in a competing background: acoustic cues, audiovisual integration (voice + mouth movements), and linguistic context. However, little is known about how these mechanisms are implemented in the brain, particularly at the level of large-scale functional neural networks. The proposed research uses functional magnetic resonance imaging (fMRI) integrated with psychophysics to address the three factors that determine intelligibility. Innovative neural network analyses test how interactions among brain regions accommodate degraded speech and improve comprehension. Our specific AIMS are to identify the neural networks mediating speech perception in noise, when intelligibility depends on: 1) Acoustic Cues, 2) Audiovisual Integration, and 3) Linguistic Context. This research program comprises a multipronged and highly cohesive body of work that will help secure our understanding of speech perception to its neurobiological foundations. Relevance to public health: We study how our brains understand speech in a noisy background, such as at a restaurant, ballgame, or office. Research like this may someday help to design better hearing aids and similar devices. It may also result in more effective listening strategies, both for those with healthy hearing and especially for those with hearing loss.

描述（由申请人提供）：本研究介绍了在嘈杂环境中语音感知的神经基础。言论在沟通中的奇异作用也许是人类最重要的日常刺激。然而，在原始条件下，很少发生言语。竞争的声音，混响和其他环境听起来通常会破坏信号。这对普通听众，尤其是听力损失的人构成了持续的挑战。在有3000万美国人听力损失的美国人中，许多人由于难以沟通而遭受抑郁和社会孤立。自从“鸡尾酒会效应”的最初表述以来，科学家已经建立了三个关键的感知/认知因素，这些因素在相互竞争的背景下提高了语音清晰度：声音提示，视听性整合（语音 +口腔运动）和语言环境。但是，对于如何在大脑中实施这些机制，尤其是在大规模功能神经网络的水平上，知之甚少。拟议的研究使用与心理物理学集成的功能磁共振成像（fMRI）来解决确定可理解性的三个因素。创新的神经网络分析测试大脑区域之间的相互作用如何适应退化的语音并提高理解。我们的具体目的是确定噪声中介导语音感知的神经网络，而清晰度取决于：1）声音提示，2）视听整合，以及3）语言环境。该研究计划包括一项多重且高度凝聚力的工作，这将有助于确保我们对语音感知对其神经生物学基础的理解。与公共卫生有关：我们研究大脑如何在嘈杂的背景中（例如在餐厅，舞台或办公室）中理解演讲。这样的研究可能有一天有助于设计更好的助听器和类似设备。对于那些有健康听力的人，尤其是听力损失的人来说，这也可能导致更有效的听力策略。