SPEECH AND COMPLEX SIGNAL PROCESSING IN THE COCHLEAR NUCLEUS

本项目将研究语音刺激如何在耳蜗腹侧核中呈现，采用平行的生理和心理物理研究，在一种接近人类的低频率阈值的啮齿动物中，栗鼠。耳蜗噪声损伤后的辅音、元音、并发元音和元音的表征将在原音、起音和斩音三种主要响应单元类别中进行研究。声信号的特定特征将在不同类型单元的响应中得到不同的表示。我们假设语音特征在耳蜗核水平的单位神经反应中被稳健编码：单个细胞群向更高的中心传递复杂光谱的不同频谱-时间特征；初级样神经元编码精细的频谱特征并保留时间-地点信息；斩波神经元将码谱特征（如共振峰和锁相）速率到声门频率；起跳神经元既编码语音音高，又因其快速适应而对某些语音进行分割。用不同的单元类型同时表示信号特征是一种可能的结果。理想的观察者计算刺激将被用来量化单个单位活动和行为表现之间的关系。结构/功能研究将在鉴定后使用单细胞标记，将细胞形态与耳蜗前腹侧核的生理和投影模式联系起来。主要的问题是：球形/毛囊细胞是否能以PL/N/PL/O/L模式响应，章鱼细胞是否能以O/I和O/L模式响应，以及C/S/C/T二分类是否存在形态学上的关联。对虚拟空间刺激的反应，同时用于听觉脑干和皮层的研究，也将被记录下来，以便更多地了解这些单位类别对定位意义声音的不同功能。

This project will study how speech stimuli are presented in the ventral cochlear nucleus using parallel physiological and psychophysical studies in a rodent with low frequency thresholds near that of humans, chinchilla laniger. The representation of consonants, vowels, concurrent vowels and vowels after noise damage of the cochlea will be studied in the three principal response categories of units, primary-like, onset and chopper. Particular features of the acoustic signal will be differentially represented in the responses of different classes of units. We hypothesize that phonetic features are robustly encoded in neural responses of units at the level of the cochlear nucleus: Individual cell groups covey different spectro-temporal features of complex spectra to higher centers: primary-like neurons code fine spectral features and preserve time-place information; chopper neurons rate code spectral features such as formants and phase lock to the glottal frequency; onset neurons both code voice pitch and the contribute to segmentation of some speech by virtue of their rapid adaptation. Simultaneous representation of signal features in different unit types is a possible outcome. Ideal observer computational stimulations will be used to quantify the relationship between single- unit activity and behavioral performance. Structure/function studies will employ single-cell labeling after characterization to correlate cell morphology with physiology and projection patterns from anteroventral cochlear nucleus. The main questions are whether globular/bushy cells can respond with PL/N/PL/O/L patterns, whether octopus cells can respond with both O/I and O/L patterns, and whether there is a morphological correlate of the C/S/C/T dichotomy. Responses to virtual-space stimuli, concurrently used for studies in the auditory brainstem and cortex, will also be recorded in order to learn more about the differential functions of these unit categories to sounds of localization significance.