- Top of page
- Conflict of Interest
Although brain activities that are associated with first (L1) and second language (L2) comprehension have been examined and compared (Klein et al. 1995; Kim et al. 1997; Perani et al. 1998; Wartenburger et al. 2003; Rüschemeyer et al. 2005, 2006; Crinion et al. 2006; Yokoyama et al. 2006b, 2009; Abutalebi et al. 2008), less attention has been paid to the cross-linguistic influence of L1 on L2 processing (Jeong et al. 2007). As cross-linguistic variations among L1 induce distinct cortical activation patterns during L1 word recognition (Yokoyama et al. 2006a; Bick et al. 2010), differences between L1 and L2 may affect cortical activation during L2 processing. In particular, it has been shown that different neural substrates are associated with word reading between alphabetic (e.g., English), phonographic (e.g., Japanese), and logographic (e.g., Chinese) languages (Sakurai et al. 2000; Nakamura et al. 2005; Tan et al. 2005; Hu et al. 2010). The findings of these studies have led us to hypothesize that cross-linguistic variations in orthography between L1 and L2 induce differential brain activations during L2 word reading.
To the best of our knowledge, this possibility has been directly investigated in only a single functional magnetic resonance imaging (fMRI) study, in which the brain activities of native English speakers were scanned during a word reading task in English, while those of early Chinese-English bilinguals were scanned during word reading tasks in both Chinese and English (Tan et al. 2003). That study found that, despite processing two different language stimuli, the Chinese-English bilinguals exhibited similar activation patterns in the left middle frontal gyrus, which is well known as the brain region that is involved in the processing of Chinese Hanji words, during the processing of L1 (Chinese) and L2 (English) (French and Jacquet 2004; Siok et al. 2004, 2008). Additionally, despite processing identical English words, Chinese-English bilinguals and English natives displayed different cortical activation patterns. The authors interpreted these findings as an indication that the influence of L1 orthographic experience during development determines cortical activation during L2 word reading processing. However, because the brain activities of L2 learners of different L1 backgrounds were not directly compared, the study did not determine whether English natives process L1 and L2 words in an identical manner.
Here, we used fMRI to test the hypothesis of whether the influence of L1 orthographic experience during development determines cortical activation during L2 word reading processing. According to previous findings, Chinese native speakers use the left middle frontal gyrus to read Chinese Hanji words, which are logographic, as described above. If Tan et al. (2003)'s hypothesis is correct, the left middle frontal gyrus will be used, even when Chinese learners read words that are written in the phonographic system, which do not activate the left middle frontal gyrus in the case of L1 reading (Sakurai et al. 2000). In addition, if the hypothesis is correct, the left middle frontal gyrus will not be activated when native speakers who have a phonographic system as their L1 read L2 words that are written in the phonographic system. To this end, the current fMRI study compared the brain activities of Chinese and Korean learners during a word-reading task of Japanese Kana (L2). In terms of orthography, Korean Hangul is similar to Japanese Kana because they are both phonographic (i.e., a single letter is mapped onto a single sound unit but is generally not mapped onto a morpheme or meaning; the alphabetic system is included in this category), whereas Chinese Hanji is logographic (i.e., a single letter is mapped onto a single word or morpheme) and therefore markedly differs orthographically from Japanese Kana.
- Top of page
- Conflict of Interest
Chinese (n = 10) and Korean (n = 7) learners were evaluated for their response times and accuracy rates in a task involving the reading of actual and pseudo Japanese (L2) words. The two groups of learners showed no significant differences in either their accuracy rates (P > 0.1) or response times (P > 0.1) in the L2 word-reading task. Both groups showed significantly longer response times to pseudowords compared to that to actual words (P < 0.05), although no differences in the accuracy rates were detected between word types (P > 0.1). The behavioral data of the two learner groups are summarized in Table 1.
Table 1. Behavioral data of the two learner groups for legal and pseudowords
|Accuracy rate (%)|
|Reaction time (msec)|
In the fMRI imaging results, the left parietal, bilateral frontal, temporal, and occipital cortices were significantly activated in both the Chinese and Korean learner groups (Fig. 1). In order to exclude the possibility that the results of the direct comparison were affected by the different proficiency levels in L2 word reading between the two learner groups, we used vocabulary test scores as a confounding covariate. In a direct comparison of the fMRI results between the two groups, Chinese learners showed significantly greater activation in the left middle frontal gyrus, and this activation survived at the P value (P < 0.05) that was corrected by the small volume correction (Fig. 2 and Table 2). In addition, to confirm that the left middle frontal activation we observed is not due to the L2 proficiency level in L2 word reading, we tested the correlation between the vocabulary test scores and brain activation. The vocabulary test scores positively correlated with brain activation of the left superior frontal gyrus and inferior temporal gyrus during the L2 word reading task and negatively correlated with the activation of the right middle and inferior frontal gyri and precentral gyrus (Fig. 3 and Table 2), indicating that the left middle frontal activation observed in the group comparison was not due to a proficiency effect of L2 word reading.
Table 2. Results of fMRI data analysis
|Anatomical area||L/R|| F || k || P || x || y || z |
|Chinese vs. Korean|
|Middle frontal gyrus||L||33.81||11||0.001||−39||32||43|
|Anatomical area||L/R|| T || k || P || x || y || z |
| Superior frontal gyrus||L||4.71||12||0.001||−15||41||55|
| Inferior temporal gyrus||L||4.47||6||0.001||−42||−43||−17|
| Middle frontal gyrus||R||4.5||10||0.001||27||53||22|
| Precentral gyrus||R||4.45||8||0.001||45||5||49|
| Opercular part of the inferior frontal gyrus||R||4.13||13||0.001||45||20||34|
Figure 1. A surface rendering of word reading for Chinese and Korean learners. The upper and lower figures show the imaging results for Chinese and Korean learners, respectively. The significance threshold was set at P < 0.001 in height (uncorrected) for visual purposes.
Download figure to PowerPoint
Figure 2. Differential brain activation during second language (L2) word reading between Chinese and Korean learners. Functional magnetic resonance imaging (fMRI) results of the left middle frontal gyrus. The graphs show the activation profiles for Chinese (red) and Korean (blue) learners. The statistical threshold was set at P < 0.05, corrected by small volume correction.
Download figure to PowerPoint
Figure 3. Brain activation during L2 word reading that was correlated with vocabulary test scores. The figures show the imaging results of brain activation that was correlated with vocabulary test scores, as evaluated by single regression and correlation analyses. (A) Positive correlation was detected in the left superior frontal gyrus and inferior temporal gyrus and (B) negative correlation was found in the right middle and inferior frontal gyri and precentral gyrus. The significance threshold was set at P < 0.001 in height (uncorrected) for visual purposes.
Download figure to PowerPoint
- Top of page
- Conflict of Interest
In the present fMRI study, we tested the hypothesis that L1 orthographic experience during development determines cortical activation during L2 word reading processing. Notably, we found that the Chinese learners showed significantly greater activation in the left middle frontal gyrus than Korean learners during L2 Japanese phonographic word reading (Fig. 2 and Table 2). Our findings strongly supported Tan et al. (2003)'s hypothesis that cross-linguistic differences in L1 orthography affect the cortical processing of L2 word reading in L2 learners.
Because we controlled for differences in age, AOA of Japanese (L2), and L2 vocabulary proficiency level, which are known to affect brain activation during L2 processing (see 'Methods'), the observed activation patterns of the left middle frontal region were independent of the activation that was elicited by these factors. Additionally, no differences in the behavioral performances in the reading task were identified between the two groups. Because these factors cannot account for the differences in brain activation, our results indicated that differential cortical activation was induced by orthographic differences in the L1 writing system, namely phonographic Hangul for Korean and logographic Hanji for Chinese. Although the number of subjects that was included in our study was limited due to the highly specialized population, previous brain activation studies that have a similar purpose and design have used a similar number of subjects (Wartenburger et al. 2003; Yokoyama et al. 2009). However, we detected statistically robust differences with correction for multiple comparisons between the two learner groups with a random-effect model, which enabled us to generalize the observed results. Hence, our results can be interpreted as an indication that cross-linguistic differences in L1 orthography affect the cortical processing of L2 word reading in L2 learners.
Further, it is important to discuss the role of the left middle frontal gyrus during L2 word reading in learners who have experience using a logographic writing system such as L1. In fact, there are two main hypotheses for the mechanism. The first is that the difference observed here was caused by a processing demand during L2 word reading when differences existed between the orthography of L1 and L2. Actually, we found different activities in the left middle frontal gyrus between Chinese and Korean learners (Fig. 1), and this region is related to processing demand or control for L2 processing (Pillai et al. 2004). However, it has been previously demonstrated that, compared to Chinese subjects with dyslexia, normal Chinese subjects show better behavioral performance and greater activation of the left middle frontal gyrus during Chinese word reading (Hu et al. 2010). This finding indicates that the left middle frontal gyrus activation that was observed in this study during word reading was not due to neural effort because normal Chinese subjects require less effort and exhibit more activation in this region than do Chinese subjects with dyslexia during reading. Here, no differences in task performance or vocabulary proficiency test scores were detected between the Chinese and Korean learners. In addition, the brain regions that were activated and correlated with vocabulary proficiency test scores differed from those activated in the direct comparison between the two groups of learners (Figs. 2, 3 and Table 2), suggesting that different processing demands activated regions other than the left middle frontal gyrus. Thus, this possible interpretation was negated.
The second hypothesis is that the experience of L1 orthography tunes cortical activation during L2 word reading processing (Tan et al. 2003). In several previous studies, the left middle frontal gyrus was specifically active for the reading of logographic characters (Tan et al. 2003, 2005; Siok et al. 2004, 2008; Hu et al. 2010). In particular, Tan et al. (2005) proposed that the left middle frontal gyrus acts as a phonological processer for logographic characters, whereas the left temporoparietal regions are activated for alphabetic characters using meta-analysis methods. Theoretically, a single logographic character has both semantic and phonological information, whereas a single phonographic character, including the alphabet, has essentially no semantic information. Hence, in logographic writing systems, orthography-to-phonology mapping processes are necessary, which are based on long-term memory. The left middle frontal gyrus may play a role in such a process (Tan et al. 2005). In contrast, in phonographic writing systems, because several characters are combined in a single word, the grapheme-to-phoneme conversion process is necessary to read the word, which is based on rule-based computation. Additionally, Tan et al. (2003) proposed that cross-linguistic differences in L1 orthography affect the cortical processing of L2 word reading in L2 learners; that is, L1 orthographic experience tunes cortical mechanisms for L2 word reading. Our results support this hypothesis because L2 word reading written by phonographic characters (i.e., Kana in the current study) activates the left middle frontal gyrus in Chinese learners who have experience with logographic writing systems such as L1. Additionally, L2 phonographic reading does not activate the left middle frontal gyrus in Korean learners who have experience with phonographic writing systems (i.e., Hungul) such as L1.
Before concluding, our results interestingly showed that vocabulary test scores negatively correlated with the activation of several frontal regions during the L2 word reading task (Figs. 2, 3 and Table 2). Previous studies have reported that proficient L2 learners show less activation in the frontal region than less proficient L2 learners during L2 processing (Chee et al. 2001; Wartenburger et al. 2003; Yokoyama et al. 2009). In addition, a recent longitudinal neuroimaging study of L2 processing has reported that, when L2 proficiency level increases, frontal activation decreases during L2 word processing (Stein et al. 2009). Hence, our results of the negative correlation between vocabulary test scores and frontal activation may reflect less activation of the frontal regions with more efficient frontal control of L2 word reading. Another interpretation is that less activation of the frontal regions may be the result of having more L2 vocabulary because more vocabulary enables the efficient use of cortical resources, which causes a reduction in the activation of the frontal regions (Prat and Just 2011). Of course, this is speculative, and it is hard to determine which interpretation is appropriate to explain our results. Thus, further studies are necessary.
In conclusion, the present fMRI study investigated whether L1 orthography influenced L2 word reading by Chinese and Korean L2 learners of the L2 of Japanese. Although the behavioral performances and AOA did not markedly differ between the two groups, Chinese learners showed greater activation in the left middle frontal gyrus than Korean learners did. These activation results were independent of the activation that was elicited by differences in proficiency levels between the two groups, suggesting that this activity of the left middle frontal gyrus was not due to the different processing demands between the two groups. Our results strongly support Tan et al. (2003)'s hypothesis that the experience of L1 orthography determines cortical activation during L2 word reading processing.