言語機能
Language
O1-9-3-1
統辞構造の計算において上縦束・弓状束を介したトップダウン結合の重要性
The importance of the top-down connection through the superior longitudinal and arcuate fasciculi for the computation of syntactic structures

○太田真理1,2, 福井直樹3,4, 酒井邦嘉1,4,5
○Shinri Ohta1,2, Naoki Fukui3,4, Kuniyoshi L. Sakai1,4,5
東京大院・総合文化・生命環境1, 日本学術振興会特別研究員2, 上智大院・外国語学・言語学3, 東京大院・総合文化・相関基礎5
Dept. of Life Sci., Univ. of Tokyo, Tokyo, Japan1, JSPS Res. Fellow, Tokyo, Japan2, Dept. of Ling., Sophia Univ., Tokyo, Japan3, CREST, Japan Science and Technology Agency, Tokyo, Japan4, Dept. of Basic Sci., Univ. of Tokyo, Tokyo, Japan5

We have reported that the left inferior frontal gyrus (L. F3op/F3t) and supramarginal gyrus (L. SMG) are crucial for syntactic computation; activations in both regions are parametrically modulated by "the Degree of Merger (DoM)", which can be defined as the maximum depth of merged subtrees within an entire sentence. By using diffusion tensor imaging, we identified that the left dorsal pathway of the superior longitudinal and arcuate fasciculi (SLF/AF) connected these regions (N = 15). To identify the direction of the connection modulated by nested sentences with largest DoM among tested conditions (similar to "Taro-ga Hanako-ga utau-to omou"), we performed effective connectivity analyses by using dynamic causal modeling. Here we assumed intrinsic, i.e., task-independent, bidirectional connections between L. F3op/F3t and L. SMG. The models were grouped into three "modulatory families" based on the modulations of the connections. Each family was composed of three "input models" as regards the regions receiving driving inputs. Using Bayesian model selection (BMS), we found that the family, in which the bottom-up connection from L. SMG to L. F3op/F3t was modulated, was the most likely family (exceedance probability = 0.85: N = 18). According to a second BMS for the input models within the family, the model, in which L. F3op/F3t received driving inputs, was the best and highly probable model (exceedance probability = 0.95). For this particular model, we further tested whether the parameter estimates were significantly different from zero. The intrinsic connection from L. F3op/F3t to L. SMG was significantly positive (+0.22; P < 0.0002), indicating that this top-down connection was consistent among the participants. The modulatory effect for the bottom-up connection was inhibitory (−0.17), though it did not reach the significance level. These results indicate that there is a transmission of information about DoM from L. F3op/F3t to L. SMG through SLF/AF.
O1-9-3-2
言語理解に関わる皮質間ネットワーク:皮質電気刺激を用いた後部中側頭回からの皮質間結合性の臨床研究
Language comprehension network: imaging cortico-cortical connectivity from the posterior middle temporal gyrus with cortical stimulation

○松本理器1, 國枝武治2, 下竹昭寛1, 小林勝哉1, 宇佐美清英1, 三國信啓4, 福山秀直3, 高橋良輔1, 池田昭夫1
○Riki Matsumoto1, Takeharu Kunieda2, Akihiro Shimotake1, Katsuya Kobayashi1, Kiyohide Usami1, Nobuhiro Mikuni4, Hidenao Fukuyama3, Ryosuke Takahashi1, Akio Ikeda1
京都大院・医・臨床神経学1, 京都大院・医・脳神経外科2, 京都大院・医・脳機能総合研究センター3, 札幌医大・医・脳神経外科4
Dept. of Neurology, Kyoto Univ., Kyoto1, Dept. of Neurosurgery, Kyoto Univ., Kyoto2, HBRC, Kyoto Univ., Kyoto3, Dept. of Neurosurg, Sapporo Med. Univ, Sapporo4

Introduction
The posterior part of the middle temporal gyrus (pMTG) has been recently highlighted for language comprehension because of the converging evidences from the voxel-based lesion-symptom mapping and fMRI meta-analysis. We aimed to trace the cortico-cortical connectivity from pMTG by means of an electrical tract tracing method of cortico-cortical evoked potentials (CCEPs). The connectivity from pMTG was compared with that from the classical language area in the adjacent posterior superior temporal gyrus (pSTG).
Methods
9 epilepsy patients who underwent invasive presurgical evaluation with subdural electrodes participated in this study (IRB No.443). After functional cortical mapping by 50 Hz electrical stimulation, repetitive single electrical pulses (1 Hz) were applied to 1) the language area in the left pSTG (a total of 14 stimulus sites) and 2) the left pMTG (anatomically identified, 16 stimulus sites). CCEPs time-locked to the stimuli were recorded from the language-dominant, left temporal, frontal and parietal lobes.
Results
In the lateral temporal area, pMTG stimulation elicited CCEPs more frequently in the anterior MTG/ITG while pSTG stimulation showed CCEPs more in the anterior STG (p<0.05). Both pSTG and pMTG stimulation elicited CCEPs from the basal temporal language areas (BTLA). Compared with pSTG stimulation (median of 2 connections/stimulus site), pMTG stimulation showed more divergent connections (median of 5 connections) not only to the lateral and basal temporal areas but also to the inferior parietal lobule (IPL), the posterior middle frontal gyrus (pMFG) and the middle-to-posterior inferior frontal gyrus (mpIFG).
Conclusion
Compared with the classical language area in pSTG, pMTG showed denser connections to the anterior MTG/ITG and more divergent connections to the semantic network (BTLA, aMTG, IPL, pMFG, mpIFG). pMTG together with BTLA likely constitutes a ventral language stream to the frontal lobe.
O1-9-3-3
漢字の各認知要素に対応して左下側頭領域内の活動部位が異なる
Different component consisting Japanese Kanji involve different area with left inferior temporal region

○樋口大樹1,2, 守口善也2, 勝沼るり2, 大場健太郎2, 寺澤悠理2, 三島和夫2, 宇野彰1
○Hiroki Higuchi1,2, Yoshiya Moriguchi2, Ruri Katsunuma2, Kentaro Oba2, Yuri Terasawa2, Kazuo Mishima2, Akira Uno1
筑波大院・人間総合1, 国立精神・神経医療研究センター 精神保健研究所 精神生理研究部2
Grad. Sch. of Comp Human Sci, University of Tsukuba, Tsukuba, Japan1, Dept. of Psychophysiology, NIMH, NCNP, Tokyo, Japan2

The inferior temporal (IT) region, including the fusiform gyrus, is thought to process orthographic information when reading in alphabetic languages. The detailed neural mechanism of orthographic processing in Japanese Kanji has been scarcely investigated, however. Our hypothesis in this study is that a Kanji consists of four different levels of linguistic components; (1)optical, (2)morphological, (3)orthographical, and (4)semantic component, and these components should be represented as the activation in different brain regions specifically within the IT. We measured brain activity in healthy adults (n=12), using functional MRI (fMRI), in response to four types of stimuli (1)Checkerboard, (2)Artificial character, (3)Pseudo Kanji, and (4)Real Kanji, each of which contains one of four hypothesized components. Checkerboard only contains the photogenic/optical component and is considered as a control condition. Artificial characters contain morphologically same contrast as a Kanji (consisting of fragmented lines) but do not follow orthographical rules. Pseudo Kanji orthographically makes sense but semantically meaningless. Real Kanji is the complete structure as it conveys some semantic meanings. For example, the contrast of pseudo Kanji vs. artificial characters reflects pure orthographic processing. The result shows that, first, neural response in the bilateral IT to real Kanji and pseudo Kanji was stronger than response to Checkerboard as expected from previous studies. What we newly found is that four types of stimuli activate different regions within the left IT, such as three different components are processed in sequential order from (2) to (4), from posterior to anterior part in the left IT. The results indicate that, as with alphabetic language, the left IT plays an important role also in the orthographic processing of Japanese Kanji, and different components consisting of a Kanji character involve different regions within IT.
O1-9-3-4
左半球白質の拡散異方性は第二言語における文法習得能力と関わる:拡散テンソル画像法による研究
White matter integrity in the left hemisphere predicts abilities to acquire syntax in second language: A DTI study

○山本香弥子1,2, 酒井邦嘉1,2
○Kayako Yamamoto1,2, Kuniyoshi L. Sakai1,2
東京大院・総合文化・相関基礎1, CREST, 日本科学技術振興機構2
Dept. of Basic Sci., Univ. of Tokyo, Tokyo, Japan1, CREST, JST, Tokyo, Japan2

Cortical regions involved in language processing are connected by several fibers, but the functional roles of these fibers have been unclear. In our previous study (Hum Brain Mapp 30:2440-2452), we found that the acquisition of linguistic knowledge in second language (L2) is subserved by at least two distinct inferior frontal regions: the left triangular part (F3t) and orbital part (F3O). Recent studies have shown that the arcuate fasciculus (AF) connects the F3t and the temporal cortex, and the inferior fronto-occipito fasciculus (IFOF) connects the F3t and the occipital cortex. Here, we used diffusion tensor imaging (DTI) to reveal the relation between the integrity of these fibers and individual differences in syntactic acquisition. Native Japanese speakers (age: 16–17) performed two error-detection tasks: an English syntactic task (Esyn) and an English spelling task (Espe). Participants who correctly answered more than 80% of Espe were included in the analysis (the accuracy of Esyn: 42–88%), in order to ensure the basic word knowledge in English. For each participant, AF and IFOF were identified in both hemispheres and the regions of interest (ROIs) were selected to extract reliable diffusion measurements; a 2-cm-long segment with most uniform thickness was selected as a ROI in the core portion, i.e. the dorsal part for AF and the narrow part for IFOF. In each ROI, we examined the correlation between the accuracy of Esyn and the mean fractional anisotropy (FA), a metric that describes the degree of diffusion anisotropy and is associated with the properties of white matter such as axonal integrity and myelination. We found a significant positive correlation between the accuracy of Esyn and FA in the left AF (r = 0.48, p<0.05). In contrast, the accuracy showed no significant correlation (p>0.2) with FA in the left IFOF, right AF, or right IFOF. These results suggest an important role of the left AF in L2 as a crucial pathway for acquiring syntactic information.
上部に戻る 前に戻る