Meade Gabriela, Grainger Jonathan, Midgley Katherine J, Emmorey Karen, Holcomb Phillip J
Joint Doctoral Program in Language and Communicative Disorders, San Diego State University & University of California, San Diego, United States.
Laboratoire de Psychologie Cognitive, CNRS & Aix-Marseille University, France.
Brain Res. 2018 Apr 15;1685:29-41. doi: 10.1016/j.brainres.2018.01.029. Epub 2018 Jan 31.
Interactive-activation models posit that visual word recognition involves co-activation of orthographic neighbors (e.g., note, node) and competition among them via lateral inhibitory connections. Behavioral evidence of this lexical competition comes from masked priming paradigms, in which target words elicit slower responses when preceded by a neighbor (e.g., note-NODE) than when preceded by an unrelated word (e.g., kiss-NODE). In the present study, we used ERPs to investigate how masked high frequency word primes influence processing of low frequency word and pseudoword targets. Word targets preceded by a neighbor prime elicited larger negativities within the N400 window than those preceded by an unrelated prime across bilateral anterior sites, which we call a reversed N400 priming effect. Consistent with the behavioral literature, the size of the reversed N400 priming effect was larger for targets from high-density orthographic neighborhoods and for participants who scored higher on a behavioral measure of spelling recognition. Indeed, the opposite effect (i.e., smaller negativities within the N400 window for word targets preceded by a neighbor) was observed for words from low-density orthographic neighborhoods and for less-skilled spellers. Traditional priming was also observed within the N250 window for word targets and within both the N250 or N400 windows for pseudoword targets. The specificity of the reversed N400 priming effect to situations in which both words have precise lexical representations suggests that it, like the behavioral interference effect, indexes lexical competition during visual word recognition.
交互激活模型认为，视觉单词识别涉及正字法邻词（如note、node）的共同激活以及它们之间通过侧向抑制连接产生的竞争。这种词汇竞争的行为证据来自掩蔽启动范式，在该范式中，当目标词之前出现一个邻词（如note - NODE）时，其反应速度比之前出现一个无关词（如kiss - NODE）时要慢。在本研究中，我们使用事件相关电位（ERP）来探究掩蔽高频词启动如何影响低频词和假词目标的加工。在双侧前脑部位，邻词启动的单词目标在N400时间窗内比无关启动的单词目标引发更大的负波，我们将其称为反向N400启动效应。与行为学文献一致，对于来自高密度正字法邻域的目标词以及在拼写识别行为测量中得分较高的参与者，反向N400启动效应的大小更大。事实上，对于来自低密度正字法邻域的单词以及拼写能力较差的人，观察到了相反的效应（即邻词启动的单词目标在N400时间窗内负波较小）。在N250时间窗内，单词目标也出现了传统的启动效应，对于假词目标，在N250或N400时间窗内都出现了传统的启动效应。反向N400启动效应对于两个单词都有精确词汇表征的情况具有特异性，这表明它与行为干扰效应一样，在视觉单词识别过程中索引词汇竞争。
