Auditory processing and speech understanding

我过去的研究表明，即使声音足够大，其时间属性的微小扭曲也会影响听众吸收信息的能力，尤其是当他们必须整合信息才能理解或记住它时。然而，语言语境可以减轻信号失真的有害影响；也就是说，自下而上（耳朵到大脑）和自上而下（大脑到耳朵）的倾听方面的相互作用调节着表现。声音的时间属性包括精细结构（音高）的快速变化和声音包络（节奏）的缓慢变化。我们研究了临床上听力正常的老年人，因为他们使用时间线索的问题可以解释他们在噪音中听力困难的原因，而年轻人在使用精细结构线索的时间扭曲来模拟老年听众的时间处理问题时表现相似。人们对不同频率的语音能量与语音可理解性之间的关系了解甚多，但对时间线索的作用知之甚少。此外，时间线索对于谈话者身份和情绪等非语言信息的交流也很重要。本研究将探索时间线索如何促进语言和非语言（如情感、说话者身份）信息的提取，以及这些共存的信息类型如何相互作用。一项惊人的初步发现是，在控制了强度之后，用恐惧的声音说话比用悲伤的声音说话更容易被理解。一种假设是，情绪信息会改变听者对语言信息的关注。另一种假设是，由于情感和语言线索是同时表达的，说话者的动作产生了互动。提出的研究旨在发现语言（说什么）和非语言（怎么说）信息的同时传播如何影响口语的产生和理解。这项研究对设计通信设备（如助听器）的工程师非常感兴趣，也对心理学家、语言学家和其他希望了解人类在通信中如何使用感官信息的人感兴趣。

My past research has shown that even when sound is loud enough, micro distortions in its temporal properties affect how well listeners uptake information, especially when they must integrate information to comprehend or remember it. However, linguistic context can mitigate the deleterious effects of signal distortions; i.e., interactions between bottom-up (ear to brain) and top-down (brain to ear) aspects of listening modulate performance. The temporal properties of sound include fast changes in the fine structure (pitch) and slow changes in the sound envelope (rhythm). Older adults with clinically normal hearing have been studied because their problems using temporal cues may explain their difficulties listening in noise and younger adults perform similarly when tested using temporal distortions of fine structure cues to simulate the temporal processing problems of older listeners. Much is known about how the amount of speech energy in different frequencies relates to speech intelligibility, but little is known about the role of temporal cues. In addition, temporal cues are important for the communication of non-linguistic information such as talker identity and emotion. The proposed research will discover how temporal cues contribute to the extraction of linguistic and non-linguistic (e.g., emotion, talker identity) information and how these co-existing types of information interact. A striking preliminary finding is that, after controlling for intensity, words are understood better when spoken in a fearful voice than in a sad voice. One hypothesis is that emotional information alters how the listener attends to linguistic information. Another hypothesis is that because emotional and linguistic cues are expressed simultaneously, the motor acts performed by a talker give rise to interactions. The proposed research aims to discover how the simultaneous communication of linguistic (what is said) and non-linguistic (how it is said) information influences the production and comprehension of spoken language. This research is of great interest to engineers designing communication devices (e.g., hearing aids), as well as being of interest to psychologists, linguists and others who wish to understand how humans use sensory information during communication.