NEURAL CODING OF SPEECH

语音是迄今为止人类最重要的一类声刺激 听众 而听觉神经中的语音编码相对 很好地理解，并提供了耳蜗核的一些信息 神经元，很少有人知道更多的中央神经元的反应 to speech语音stimulus刺激. 我们建议研究单个神经元的反应， 麻醉猫的下丘（IC）对各种各样的 语音和类似语音的声音， 用于英语和其他语言。 一系列的实验将 研究元音的共振峰频率和基频是如何 在IC神经元的放电模式中编码，使用合成刺激， 这些参数在适当的范围内系统地变化 男性和女性的声音。 这些反应将与 用于正弦波语音和调幅音调的那些， 确定对元音的响应是否可以从对 更简单的刺激 另一系列实验将研究 发音、位置和发音方式不同的辅音。 在 特别是，我们将研究如何对辅音的反应受到影响， 前面的上下文。 这些反应将与 连续出现的成对音调，以确定是否 适应、易化和持久抑制等机制 增强对特定语音特征的响应。 本研究 解决了言语感知的基本问题，如神经基础 分类知觉和语音的知觉不变性 在它们的声学特性中面对上下文依赖性。 通过 提供脑干中语音神经编码的直接知识， 这项研究也可能有助于设计用于听觉（耳蜗）的语音处理器。 和脑干）植入物， 为中央听觉的特定部分适当编码的信息 系统 结合空间听觉的研究， 也有助于更好地理解为什么听力受损和老年人 听众在理解他们的讲话时有更大的困难。 竞争的声音比正常的听众，并可能有助于开发新的类型 助听器在嘈杂的环境中表现更好。

Speech is by far the most important class of acoustic stimuli for human listeners. While the coding of speech in the auditory nerve is relatively well understood, and some information is available for cochlear-nucleus neurons, very little is known about the responses of more central neurons to speech stimuli. We propose to study the responses of single neurons in the inferior colliculus (IC) of anesthetized cats to a wide variety of speech and speech-like sounds encompassing the acoustic characteristics used in English and other languages. One series of experiments will examine how the formant frequencies and fundamental frequency of vowels are coded in the discharge patterns of IC neurons, using synthetic stimuli in which these parameters ar systematically varied over a range appropriate for both male and female voices. These responses will be compared with those for sine-wave speech and amplitude-modulated tones in order to determine whether responses to vowels can be predicted from responses to simpler stimuli. Another series of experiments will study the responses to consonants differing in voicing, place, and manner of articulation. In particular, we will examine how the responses to consonants are affected by the preceding context. These responses will be compared with those to pairs of tones presented in succession in order to determine whether mechanisms such as adaptation, facilitation and long-lasting inhibition enhance the response to particular speech features. This research addresses fundamental issues in speech perception such as the neural basis for categorical perception, and the perceptual invariance of speech sounds in the face of context-dependence in their acoustic characteristics. By providing direct knowledge of the neural coding of speech in the brainstem, this research may also help design speech processors for auditory (cochlear and brainstem) implants that would be more effective by providing information appropriately coded for specific parts of the central auditory system. In combination with research on spatial hearing, this work may also lead to a better understanding of why hearing-impaired and elderly listeners have much greater difficulties understanding speech among competing sounds than do normal listeners, and may help develop new kinds of hearing aids that would perform better in noisy environments.