Egashira Yuka, Kaga Yoshimi, Gunji Atsuko, Kita Yosuke, Kimura Motohiro, Hironaga Naruhito, Takeichi Hiroshige, Hayashi Sayuri, Kaneko Yuu, Takahashi Hidetoshi, Hanakawa Takashi, Okada Takashi, Inagaki Masumi
Department of Developmental Disorders, National Institute of Mental Health, National Center of Neurology and Psychiatry (NCNP), Kodaira, Japan.
Department of Pediatrics, Faculty of Medicine, University of Yamanashi, Chuo, Japan.
Front Hum Neurosci. 2022 Aug 15;16:913945. doi: 10.3389/fnhum.2022.913945. eCollection 2022.
Reading fluency is based on the automatic visual recognition of words. As a manifestation of the automatic processing of words, an automatic deviance detection of visual word stimuli can be observed in the early stages of visual recognition. To clarify whether this phenomenon occurs with Japanese kanji compounds-since their lexicality is related to semantic association-we investigated the brain response by utilizing three types of deviants: differences in font type, lexically correct or incorrect Japanese kanji compound words and pseudo-kanji characters modified from correct and incorrect compounds. We employed magnetoencephalography (MEG) to evaluate the spatiotemporal profiles of the related brain regions. The study included 22 adult native Japanese speakers (16 females). The abovementioned three kinds of stimuli containing 20% deviants were presented during the MEG measurement. Activity in the occipital pole region of the brain was observed upon the detection of font-type deviance within 250 ms of stimulus onset. Although no significant activity upon detecting lexically correct/incorrect kanji compounds or pseudo-kanji character deviations was observed, the activity in the posterior transverse region of the collateral sulcus (pCoS)-which is a fusiform neighboring area-was larger when detecting lexically correct kanji compounds than when detecting pseudo-kanji characters. Taken together, these results support the notion that the automatic detection of deviance in kanji compounds may be limited to a low-level feature, such as the stimulus stroke thickness.
阅读流畅性基于对单词的自动视觉识别。作为单词自动处理的一种表现，在视觉识别的早期阶段可以观察到对视觉单词刺激的自动偏差检测。为了弄清楚这种现象是否也出现在日语汉字复合词中（因为它们的词汇性与语义关联有关），我们利用三种类型的偏差刺激来研究大脑反应：字体类型差异、词汇正确或错误的日语汉字复合词，以及从正确和错误复合词修改而来的假汉字。我们采用脑磁图（MEG）来评估相关脑区的时空特征。该研究纳入了22名以日语为母语的成年受试者（16名女性）。在MEG测量期间呈现上述三种包含20%偏差刺激的类型。在刺激开始后250毫秒内检测到字体类型偏差时，观察到大脑枕极区域有活动。虽然在检测词汇正确/错误的汉字复合词或假汉字偏差时未观察到显著活动，但在检测词汇正确的汉字复合词时，作为梭状回相邻区域的 collateral sulcus后横区域（pCoS）的活动比检测假汉字时更大。综上所述，这些结果支持这样一种观点，即汉字复合词中偏差的自动检测可能仅限于低水平特征，如刺激笔画粗细。
