The Hong Kong Polytechnic University, Department of Chinese and Bilingual Studies, Hong Kong, China; Research Centre for Language, Cognition, and Neuroscience, The Hong Kong Polytechnic University, Hong Kong, China.
The Hong Kong Polytechnic University, Department of Chinese and Bilingual Studies, Hong Kong, China; Research Centre for Language, Cognition, and Neuroscience, The Hong Kong Polytechnic University, Hong Kong, China.
Neuroimage Clin. 2019;23:101814. doi: 10.1016/j.nicl.2019.101814. Epub 2019 Apr 3.
Despite the lack of invariance in the mapping between the acoustic signal and phonological representation, typical listeners are capable of using information of a talker's vocal characteristics to recognize phonemes, a process known as "talker normalization". The current study investigated the time course of talker normalization in typical listeners and individuals with congenital amusia, a neurodevelopmental disorder of refined pitch processing. We examined the event-related potentials (ERPs) underling lexical tone processing in 24 Cantonese-speaking amusics and 24 typical listeners (controls) in two conditions: blocked-talker and mixed-talker conditions. The results demonstrated that for typical listeners, effects of talker variability can be observed as early as in the N1 time-window (100-150 ms), with the N1 amplitude reduced in the mixed-talker condition. Significant effects were also found in later components: the N2b/c peaked significantly earlier and the P3a and P3b amplitude was enhanced in the blocked-talker condition relative to the mixed-talker condition, especially for the tone pair that is more difficult to discriminate. These results suggest that the blocked-talker mode of stimulus presentation probably facilitates auditory processing and requires less attentional effort with easier speech categorization than the mixed-talker condition, providing neural evidence for the "active control theory". On the other hand, amusics exhibited comparable N1 amplitude to controls in both conditions, but deviated from controls in later components. They demonstrated overall later N2b/c peak latency significantly reduced P3a amplitude in the blocked-talker condition and reduced P3b amplitude irrespective of talker conditions. These results suggest that the amusic brain was intact in the auditory processing of talker normalization processes, as reflected by the comparable N1 amplitude, but exhibited reduced automatic attentional switch to tone changes in the blocked-talker condition, as captured by the reduced P3a amplitude, which presumably underlies a previously reported perceptual "anchoring" deficit in amusics. Altogether, these findings revealed the time course of talker normalization processes in typical listeners and extended the finding that conscious pitch processing is impaired in the amusic brain.
尽管在声学信号和语音表示之间的映射缺乏不变性,但典型的听众能够利用说话者的声音特征信息来识别音位,这一过程被称为“说话者归一化”。本研究调查了典型听众和先天性失歌症患者(一种精细音高处理的神经发育障碍)中说话者归一化的时间进程。我们在两种条件下检查了 24 名粤语失歌症患者和 24 名典型听众(对照组)的词汇声调处理的事件相关电位(ERP):分组说话者条件和混合说话者条件。结果表明,对于典型听众,说话者可变性的影响早在 N1 时间窗口(100-150ms)中即可观察到,在混合说话者条件下 N1 振幅减小。在后续成分中也发现了显著的影响:N2b/c 更早地达到峰值,并且在分组说话者条件下 P3a 和 P3b 的振幅相对于混合说话者条件增强,特别是对于更难区分的声调对。这些结果表明,与混合说话者条件相比,分组说话者模式的刺激呈现可能促进听觉处理,并在更简单的言语分类中需要较少的注意力努力,为“主动控制理论”提供了神经证据。另一方面,失歌症患者在两种条件下的 N1 振幅与对照组相当,但在后续成分中与对照组不同。他们表现出在分组说话者条件下 P3a 振幅总体减小和无论说话者条件如何 P3b 振幅减小的 N2b/c 峰值潜伏期显著延迟。这些结果表明,失歌症患者的大脑在听觉处理说话者归一化过程中是完整的,这反映在可比的 N1 振幅中,但在分组说话者条件下表现出对音高变化的自动注意转换减少,这可以通过减少的 P3a 振幅来捕捉,这大概是失歌症患者先前报告的知觉“锚定”缺陷的基础。总之,这些发现揭示了典型听众中说话者归一化过程的时间进程,并扩展了失歌症患者大脑中意识音高处理受损的发现。