Fargier Raphaël, Laganaro Marina
FPSE, University of Geneva, 42, Bd du Pont d'Arve, 1211, Geneva, Switzerland.
Brain Topogr. 2017 Mar;30(2):182-197. doi: 10.1007/s10548-016-0504-4. Epub 2016 Jun 22.
Picture naming tasks are largely used to elicit the production of specific words and sentences in psycholinguistic and neuroimaging research. However, the generation of lexical concepts from a visual input is clearly not the exclusive way speech production is triggered. In inferential speech encoding, the concept is not provided from a visual input, but is elaborated though semantic and/or episodic associations. It is therefore likely that the cognitive operations leading to lexical selection and word encoding are different in inferential and referential expressive language. In particular, in picture naming lexical selection might ensue from a simple association between a perceptual visual representation and a word with minimal semantic processes, whereas richer semantic associations are involved in lexical retrieval in inferential situations. Here we address this hypothesis by analyzing ERP correlates during word production in a referential and an inferential task. The participants produced the same words elicited from pictures or from short written definitions. The two tasks displayed similar electrophysiological patterns only in the time-period preceding the verbal response. In the stimulus-locked ERPs waveform amplitudes and periods of stable global electrophysiological patterns differed across tasks after the P100 component and until 400-500 ms, suggesting the involvement of different, task-specific neural networks. Based on the analysis of the time-windows affected by specific semantic and lexical variables in each task, we conclude that lexical selection is underpinned by a different set of conceptual and brain processes, with semantic processes clearly preceding word retrieval in naming from definition whereas the semantic information is enriched in parallel with word retrieval in picture naming.
在心理语言学和神经成像研究中，图片命名任务主要用于引发特定单词和句子的生成。然而，从视觉输入生成词汇概念显然不是触发言语产生的唯一方式。在推理言语编码中，概念不是从视觉输入中提供的，而是通过语义和/或情景联想来阐述的。因此，在推理和指称性表达语言中，导致词汇选择和单词编码的认知操作可能是不同的。特别是，在图片命名中，词汇选择可能源于感知视觉表征与单词之间的简单关联，语义过程最少，而在推理情境中的词汇检索则涉及更丰富的语义联想。在这里，我们通过分析在指称性任务和推理任务中单词产生过程中的ERP相关性来验证这一假设。参与者说出从图片或简短书面定义中引出的相同单词。这两个任务仅在言语反应之前的时间段内显示出相似的电生理模式。在刺激锁定的ERP波形中，P100成分之后直到400 - 500毫秒，稳定的全局电生理模式的振幅和周期在不同任务之间存在差异，这表明涉及不同的、特定于任务的神经网络。基于对每个任务中受特定语义和词汇变量影响的时间窗口的分析，我们得出结论，词汇选择由不同的概念和大脑过程支持，在根据定义命名时，语义过程明显先于单词检索，而在图片命名中，语义信息与单词检索同时得到丰富。
