Processing of complex sounds at high frequencies

摘要 音高是自然声音的一个基本知觉维度，而音高知觉对于 理解语音和音乐，并分离同时发出的声音。音调知觉会随着 声音中的分量频率增加超过2-3 khz，这一趋势通常被认为反映了 2-3 kHz以上听神经的相锁定相应丧失为颞叶细微结构(TFS)。 然而，最近的心理物理和生理发现，包括有证据表明准确的音调知觉 在高频下是可能的，而在低于以下频率时，人类的相位锁定可能会退化 之前认为的，都对这一联系提出了质疑。拟议研究的目标1测试了一种假设 跨频率比较或信息整合方面的普遍缺陷，而不是相位的下降 锁定，可以解释为什么音调知觉会随着频率的增加而降低。为此，性能将是 在不依赖于跨频率比较的任务中进行测量，这一假设将预测 在低频和高频下的同等性能，以及在需要跨 频率比较，其中该假设预测在高频下的性能比在 低频率。拟议研究的目标2验证了这种跨频率缺陷的假设 比较还可以预测并发声音的分离应该在高频下受到损害， 即使在分离提示没有通过TFS信息编码时也是如此。为此，我们将衡量 幅度调制和起始异步提示促进了单个或多个目标分离 复杂色调中的组件。在这里，我们的假设预测，在 频率很高，特别是与不需要隔离的匹配控制任务相比。这个 行为结果，将借助模拟听神经的理想观察者分析进行解释 反应，将提供有价值的洞察神经机制的基础复杂的音调感知和 并存的声音隔离。

Abstract Pitch is a fundamental perceptual dimension of natural sounds, and pitch perception is crucial for understanding speech and music and for segregating concurrent sounds. Pitch perception degrades as component frequencies in a sound increase beyond 2-3 kHz, a trend that is usually assumed to reflect a corresponding loss of phase locking to temporal fine structure (TFS) in the auditory nerve above 2-3 kHz. However, recent psychophysical and physiological findings, including evidence that accurate pitch perception is possible at high frequencies and that phase locking may degrade in humans at lower frequencies than previously believed, have cast doubt on this link. Aim 1 of the proposed research tests the hypothesis that a general deficit in across-frequency comparison or integration of information, instead of the roll-off of phase locking, may explain why pitch perception degrades as frequency increases. To this end, performance will be measured in tasks that do not depend on across-frequency comparisons, where this hypothesis would predict equivalent performance at low and high frequencies, and in tasks that are designed to require across- frequency comparisons, where this hypothesis would predict poorer performance at high frequencies than at low frequencies. Aim 2 of the proposed research tests the hypothesis that this deficit in across-frequency comparison might also predict that segregation of concurrent sounds should be impaired at high frequencies, even when the segregation cues are not encoded via TFS information. To this end, we will measure the extent to which amplitude modulation and onset asynchrony cues promote segregation of single or multiple target components from within a complex tone. Here, our hypothesis predicts that performance should be poorer at high frequencies, especially when compared to matched control tasks that require no segregation. The behavioral results, which will be interpreted with the aid of ideal observer analysis of simulated auditory nerve responses, will provide valuable insight into the neural mechanisms that underlie complex pitch perception and concurrent sound segregation.