Abstract：
Nouns and verbs are fundamental grammatical building blocks of languages. A key question is whether and where the noun-verb division was represented in the brain. Previous studies mainly used univariate analyses to examine this issue. However, the interpretation of activated brain regions in univariate analyses may be confounded with general cognitive processing and/or confounding variables. We addressed these limitations by using partial representation similarity analysis (RSA) of Chinese nouns and verbs with different levels of imageability. Participants were asked to complete the 1-back grammatical class probe (GCP; an explicit measure) and the 1-back word probe (WP; an implicit measure) tasks while undergoing functional magnetic resonance imaging. RSA results showed that the activation pattern in the left posterior middle temporal gyrus (LpMTG) was significantly correlated with the grammatical class representational dissimilarity matrix in the GCP task after eliminating the potential confounding variables. Moreover, the LpMTG did not overlap with the frontal-parietal regions that were activated by verbs vs. nouns or the task effect (CRP vs. WP) in univariate analyses. These results highlight the role of LpMTG in distinguishing nouns from verbs rather than general cognitive processing.
摘要：
名词和动词是语言的基本语法组成部分。一个关键问题是在大脑中是否以及在何处表示名词-动词划分。以前的研究主要使用单变量分析来研究这个问题。然而,单变量分析中激活脑区的解释可能与一般认知过程和/或混杂变量相混淆.我们通过对具有不同可想象性水平的汉语名词和动词进行部分表示相似性分析（RSA）来解决这些局限性。要求参与者在进行功能磁共振成像时完成1回语法类探针（GCP；显式测量）和1回单词探针（WP；隐式测量）任务。RSA结果表明，在消除了潜在的混杂变量后，左后颞中回（LpMTG）的激活模式与GCP任务中的语法类别代表性差异矩阵显着相关。此外，LpMTG与动词激活的额叶-顶叶区域不重叠名词或任务效果(CRP与WP)在单变量分析中。这些结果强调了LpMTG在区分名词和动词而不是一般认知过程中的作用。
