Laboratory Models, Dynamics, Corpora, Department of Language Science, University Paris Nanterre - Paris Lumières, CNRS, Nanterre, France.
Shanghai Unidt, inc, People's Republic of China.
Neuropsychologia. 2020 Oct;147:107557. doi: 10.1016/j.neuropsychologia.2020.107557. Epub 2020 Jul 15.
The present electroencephalographic (EEG) study was designed to investigate the perception of Mandarin neutral tone (T0) by adult Mandarin speakers. For this purpose, we examined the event-related brain potential (ERP) correlates of T0 processing in two-character Mandarin compounds. Eighteen native Mandarin speakers were tested using a modified oddball paradigm. Sixty strong-strong and 60 strong-weak (T0) disyllabic Mandarin compounds were selected from the Contemporary Chinese Dictionary. Participants performed one explicit task of stress discrimination (i.e., deciding whether the fourth word in sequences of 4 spoken words had the same stress pattern as the previous words or a different one) and one implicit task of passive listening. Behaviorally, accuracy showed that Mandarin listeners were able to correctly discriminate T0 from the canonical strong-strong stress pattern in 87.2% of cases. Neurophysiologically, larger N200 and P200 were observed for the infrequent strong-weak stress pattern compared to the canonical strong-strong one. Critically, a N325 was replicated in Mandarin, with a larger N325 for strong-weak than for strong-strong compounds. Consistently with a previous interpretation proposed by Böcker et al. (1999) for Dutch, we argue that the N325 might be a manifestation of the extraction of stress pattern in Mandarin also. Taken together, the present data on T0 perception in Mandarin are discussed in the context of the Prosody-Assisted-Processing (PAP) model (Isel et al., 2003), a cognitive model of spoken compound processing based on stress-timed languages, which postulates an early involvement of prosody in order to guide the morphological analysis at the lexical level.
本脑电图（EEG）研究旨在探讨成年普通话使用者对普通话中声音中立调（T0）的感知。为此，我们考察了在双音节普通话复合词中 T0 加工的事件相关脑电位（ERP）相关性。18 名母语为普通话的人参加了一项经修改的Oddball 范式测试。从《现代汉语词典》中选择了 60 个强-强（T0）和 60 个强-弱（T0）的双音节普通话复合词。参与者执行了一个压力辨别（即判断四个语音词序列中的第四个词的重音模式是否与前几个词相同或不同）的显性任务和一个被动聆听的隐性任务。行为表现上，准确率表明普通话听众能够正确辨别 87.2%的 T0 与典型的强-强重音模式。神经生理学上，与典型的强-强模式相比，不常见的强-弱重音模式引起了更大的 N200 和 P200。重要的是，在普通话中复制了 N325，与强-强复合词相比，强-弱复合词的 N325 更大。与 Böcker 等人（1999 年）对荷兰语的先前解释一致，我们认为 N325 可能也是普通话中重音模式提取的表现。总的来说，本文对普通话 T0 感知的研究结果在 Prosody-Assisted-Processing（PAP）模型（Isel 等人，2003）的背景下进行了讨论，该模型是一种基于重音语言的口语复合词处理的认知模型，它假设早期涉及韵律以指导词汇水平上的形态分析。
