Extensive Reading Affects Second Language Listening Proficiency: an fNIRS Study

doi:10.21203/rs.3.rs-2245621/v1

Download PDF

Article

Extensive Reading Affects Second Language Listening Proficiency: an fNIRS Study

https://doi.org/10.21203/rs.3.rs-2245621/v1

This work is licensed under a CC BY 4.0 License

Version 1

posted

You are reading this latest preprint version

One of the difficulties in learning English as a foreign language (EFL) is obtaining sufficient input in the language. The importance of input in EFL settings has been discussed for decades, and extensive reading (ER) has been proposed to supply input to learners of English. Many studies have provided evidence that ER enhances learners' reading proficiency, vocabulary, and grammar. Interestingly, some studies have reported the effect of ER on listening besides reading skills. The current study aimed to clarify the effect of ER on listening ability from both behavioral and neurophysiological points of view. The participants were university students in Japan who took the same compulsory English class and were required to complete 150,000 words of ER per semester. To examine the effect of ER, measurements were obtained at three time points: the beginning of the spring semester, the end of the spring semester, and the end of the fall semester. Our results showed that ER training affects both behavioral performance and brain activity during foreign language listening and that ER training enables listening processing with less energy. The findings demonstrate that ER training affects real-time processing during foreign language listening tasks.

Biological sciences/Neuroscience

Biological sciences/Psychology

Extensive reading (ER) is considered one of the best means of learning English as a foreign language (EFL), particularly for learners who live in a chronically input-poor environment, such as Japan [1]. Throughout the six years of English language education in junior and senior high school, students read only around 30,000 English words [2]. Moreover, in EFL settings, learners tend to learn English through grammar-translation methods with an emphasis on remembering grammatical rules and vocabulary. Consequently, the inefficacy of learning and the lack of communicative competence are often demonstrated [3]. Major standardized proficiency tests, such as the Test of English for International Communication (TOEIC) [4], the International English Language Testing System (IELTS, 2020) [5], and the Test of English as a Foreign Language (TOEFL) [6], reveal the low English proficiency of Japanese learners. Since English is now a world language or the lingua franca [7], the lack of input in the language must be compensated for in order to improve Japanese learners’ proficiency in English. Language teachers have recognized ER as an effective way to increase input and have adopted it for a long time. Many studies have examined its benefits, such as reading speed, reading comprehension [8–10], vocabulary improvement, building healthy reading habits, and enhancing motivation for language learning [11,8,12]. A meta-analysis[13] reviewed 24 studies and identified four areas of interest: reading speed, reading comprehension, vocabulary, and grammar.

ER is defined as reading 1) for fun, 2) easy and comprehensible books and 3) a large number of books. One of its most important elements is that readers should choose books based on their interests [14]. As is often the case with ER, books such as those with which graded readers are familiar are written with adjusted levels of vocabulary and grammar for non-native speakers of English [12]. Some of these books start with only one English word per book, and gradually the levels increase to a larger number of words and more complex sentences. By reading these easy books, learners gradually enhance their reading abilities. Many studies have demonstrated the effects of such books on reading speed and comprehension [15,16, 13, 17]. Reading speed is often measured by calculating the number of words per minute (WPM). One study showed that a greater amount of reading leads to a higher reading rate [9,10, 18]. A strong correlation between the number of words and WPM (r = .773) has also been reported [19]. Improvements in reading comprehension have also been observed through ER training. After 11 weeks of ER training, the participants showed improved comprehension ability in general reading ; however, the study did not provide the exact number of words read by participants [20]. Another study reported that one semester of ER training (Mean = 133,450 words, standard deviation (SD) = 118,214 words) showed a statistically significant improvement in the scores of reading comprehension on the foreign language assessment test [17]. Thus, much ER research has been conducted on vocabulary acquisition, reading speed, and reading comprehension.

Interestingly, some studies found that ER enhances listening ability [17, 21–23], and many language teachers would agree with this phenomenon through their empirical experiences of ER teaching. A previous study found that ER hours had a higher correlation with the score for TOEIC listening comprehension (r = .93) than with the reading score (r = .75) [24]. It is not difficult to explain why reading training improves reading speed and comprehension; however, the mechanism of improving listening ability with ER must be clarified. The recent meta-analyses of ER mentioned above [13, 25] did not address listening, and previous studies have only examined indirect behavioral data. The current research aimed to examine, through both behavioral and neurophysiological points of view, whether ER training has a direct effect on improving listening ability.

However, a meta-analysis of ER [13] found no areas of interest in listening ability. Another recent meta-analysis of ER [25] found that 15% of the research on listening skills did not examine its efficacy. While we found some research reporting the effect of ER on listening competence, these studies have drawbacks in the research method; no study has measured the behavioral data of listening activity that indicates the effect of ER. However, one study is of particular significance. Ten months of ER training and its effects on reading and listening abilities have been examined [21]. The participants were novice learners of English and senior high school students and showed statistically significant improvements in listening comprehension. The author inferred that the skills enforced in reading training may transfer to listening skills. Another study involved 26 college students who underwent ER training for approximately a semester (Mean = 133,450 words, SD = 118,214 words) and examined their English proficiency using the Computerized Assessment System for English Communication test [17]. The pre and post scores were analyzed using a paired t-test, which showed significant increase both in the reading section (t(25) = -3.095, p = .005, d = − .607) and in the listening section (t(25) = -3.094, p = .005, d = − .607). This was also observed in a large-scale experiment, where the authors reported that the average score of all seven faculties showed an increase in the reading test, of whom three showed a significant increase in the listening test (p < .001) [23]. When we read letters, we first observe a word’s sound called the phonological coding, and then extract the meaning from our mental lexicon that matches the sound [26,27]. All studies mentioned above dealt with static data measured as a post-test of ER training, such as proficiency assessment test scores; however, in order to ascertain the improvement of the psychological skill itself, it is necessary to collect and analyze processing data during a listening activity. Therefore, the current study measures brain activity (BA), reaction time (RT), and the accuracy rate (AR) during the listening task as processing data during a listening activity.

To measure brain activity during the processing of L2 listening in real time, the current study used a two-channel NIRS (near infrared spectroscopy) that can measure BA in the dorsolateral prefrontal cortex (DLPFC). Unlike functional magnetic resonance imaging (fMRI) which requires a large measurement system or full-head NIRS which has many channels, this small two-channel NIRS is an easy-to-use device for measuring BA. Recent fMRI studies have revealed that L2 processing is involved in dorsal attention networks, including the DLPFC [28,29]. In novice L2 learners’ lexical processing, increased activity in the DLPFC has been reported [28], and L2 and L1 are considered to be processed in a language network that includes the DLPFC [29,30]. In L2 acquisition studies, decreased BA was observed in familiar letters compared with unfamiliar ones [31] ; more advanced learners of English showed less activation of the brain than elementary learners; the advanced group showed less activation in processing familiar words than in processing unfamiliar words [32]. These findings suggest that the higher the proficiency gain, the lesser the energy needed for language processing. L2 verbal processing is carried out in Brodmann Area 46 (in the DLPFC) and used for language-related working memory (WM) [33, 34, 28]. Based on these findings, the current research aimed to examine whether ER improved real-time L2 listening processing performance (i.e., RTs and ARs) and whether ER impacts BA in the DLPFC during L2 listening processing.

Trial settings

This study involved examining the effect of ER on learning EFL by monitoring BA in the prefrontal cortex (PFC). The experiment was conducted from April 2018 to February 2019 at Gunma University, Maebashi City, Gunma Prefecture, Japan. The purpose and procedure of the experiment were explained to the participants, and their written informed consent was obtained. Monetary compensation was given for each experiment and for handing over personal records of test scores and ER. This study was approved by the Ethics Committee of the Faculty of International Studies of Bunkyo University and conducted in accordance with the principles of the Declaration of Helsinki (1991). To observe the continual training effect, three experiments (pre, post 1, and post 2) were conducted .

Participants

Forty-three college students (first-year students of a national university) took part in this project from two different classes of a compulsory subject, English I. They had learned EFL and had taken at least six years of formal English education in Japanese junior high and high school. Their English proficiency was measured using the Listening and Reading test in TOEIC Institutional Program (IP). All participants belonged to the science and engineering department and took the same compulsory English classes. They were from the same university and department and were taking the same English course, showing that environmental and ability differences among the participants were small. On the other hand, since this study was conducted through the compulsory subject, a control group who were not given an intervention could not be made.

Intervention

In an ER training, the participants were assigned to read 150,000 words as one of the requirements for their credits during the 15-week semester. There were about 100,000 books in the library, from which they were allowed to choose according to their interests and proficiency. The participants were required to keep a record on Moodle, take 10 comprehension quizzes after reading each book, and get 60% or more correct answers. The instructors checked their records. The same amount of ER and procedure was repeated in the fall semester. As a result, the participants read approximately 300,000 words in total. Measurements were conducted at three time points: the beginning of the spring semester in April, the end of the spring semester in July, and the end of the fall semester in January. They did not have any prior experience in ER before they entered the university. Therefore, we could have a baseline to examine the effect of ER.

English proficiency

The TOEIC Listening and Reading test was used to objectively assess the proficiency of the participants. The TOEIC test was developed by the Educational Testing Service (ETS), United States (US), and consists of a 45-minute listening section (100 questions) and a 75-minute reading section (100 questions). It is widely used to measure English proficiency in Asia, particularly in Japan and South Korea. In 2018, 2,456,000 students took the test, and the average score of first-year college students was 440 out of 990. The mean score of the participants was 431.43 (SD = 19.16), which is considered as indicating the average proficiency of Japanese college students.

Listening task for functional near-infrared spectroscopy (fNIRS) experiment

A listening task was conducted through E-Prime 2.0. The listening materials were extracted from an EFL textbook, “Timed Reading for Fluency (Book 1)” [35]. All the texts were 200 words each, and the readability was carefully controlled (written with 800 headwords). The participants listened to the whole text by pressing the Enter key and answered seven True/False questions about the content of the same text. The RTs of the listening comprehension questions and their ARs were calculated. The participants listened to the passage on their headphones. The fixation point was shown at the center of the laptop computer monitor. The sound was played at an appropriate volume for each participant using a computer. The topics of the text which they read and to which they listened were counterbalanced among the participants. The same stimuli were not given to each participant among pre, post 1, and post 2.

Trial session

A trial session was conducted just before the main experiments were conducted so that the participants could become proficient in the measurement task. The procedure was the same as the main experiments, and two items were given to each participant in this trial session. These items in the trial session did not overlap with those in the main experimental sessions. The trial session had to allow all participants to participate in the experiments even in post 1 after familiarizing themselves with the task itself. Therefore, we assumed that this trial session could avoid a practice effect/repetition effect on our main experimental tasks among pre, post 1, and post 2 [36] .

Data exclusion

The participants who did not participate in all three experiments and whose data were not correctly measured were excluded from the analysis. The data of 24 participants (20 male and 4 female participants, aged 18.13, SD = 0.448) were analyzed (Table 1). The main reason for the decrease in the number of participants was the failure of data collection at the post 1 experiment, not the loss of many participants’ motivation to continue ER.

Table 1

Participants’ demographic data
Participants (N = 24)
	Mean	SD
Age	18.13	0.448
Sex	Male 20: Female 4

fNIRS data

During the trials, prefrontal BA was measured using a two-channel near-infrared spectroscopy (NIRS) device (HOT-1000, a wavelength of 810 nm, sampling frequency 10 Hz, NeU Corp.). HOT-1000 has two dual source-detector optode sets and measures the concentration change of total hemoglobin (total Hb) for each source-detector [37]. It was set onto the forehead/frontal pole of the DLPFC (Broadman Area 10) of both the left and right hemispheres. The NIRS signals were collected while the participants listened to the stimulus. The whole data were baseline corrected by subtracting the average NIRS signals of both the preceding and post-rest periods, 15 s each, for a total of 30 s. We then averaged the NIRS signals in each experiment: pre, post 1, and post 2.

Analysis

We tested the expectation that BA would change along with the improvement in listening comprehension of the target language (English) as the participants gained the number of words during the ER training. Familiarity and mastery of skill have been found to decrease the load on BA [38]. To validate the results by analyzing longitudinal data, a generalized linear mixed-effects model (GLMM ANOVA) in SPSS was used. GLMMs are fundamentally flexible because they model both the mean and covariance structures [39]. We investigated the contribution of the fixed factors (the three time points and the left and right sides of the brain) and their interaction with behavioral and BA data. A post-hoc test was conducted using the Bonferroni method to extract the differences between the two periods. Missing values were excluded from the analysis. Statistical significance was set at p < 0.05, and all analyses were conducted using SPSS 25 (IBM SPSS 2017, Chicago, IL, US).

Results

Participant description

There were 24 participants (20 male and 4 female participants), aged 18 to 20 years (Mean = 18.13, SD = 0.448; Table 1).

Pre-test

The results of the behavioral data are presented in Table 2. The participants took the TOEIC test before beginning ER. The mean total score was 414.79 (SD = 132.33), reading section score was 181.67 (SD = 75.18), and listening section score was 233.13 (SD = 67.14). This score was slightly lower than the score of Japanese first-year college students (440, 2018). In the listening task, the AR of the comprehension tests was measured, and the total time taken to answer the comprehension questions was calculated as RT. The AR was 54.69% (SD = 8.82), and RT was 7560.04 ms (SD = 2094.11).

Table 2

Descriptive data of behavioral data
Criteria	Pre (n = 24)		Post 1 (n = 24)		Post 2 (n = 24)
Criteria	Mean	SD	Mean	SD	Mean	SD
TOEIC Total	414.79	132.33	465.00	141.56	466.46	155.15
TOEIC Reading	181.67	75.18	198.54	72.42	197.08	76.07
TOEIC Listening	233.13	67.14	266.46	76.34	269.38	85.43
Listening Accuracy Rate	54.69	8.82	54.58	7.82	56.47	6.63
Listening Response Time	7560.04	2094.11	6080.22	2079.29	6194.27	2228.17

Post 1

After 15 weeks of ER, Post1 tests were carried out. The total score on the TOEIC test was 465 (SD = 141.56), the reading section score was 198.54 (SD = 72.42), and the listening section score was 266.46 (SD = 76.34). Listening test AR was 54.58% (SD = 7.82), and RT was 6080.22 ms (2079.29) (Table 2).

Post 2

At the end of another 15 weeks of ER, the total score on the TOEIC test was 466.46 (SD = 151.15), the reading section score was 197.08 (SD = 76.08), and the listening section score was 269.38 (SD = 85.43). The listening test AR was 56.47% (SD = 6.63), and RT was 6194.27 ms (2228.17) (Table 2).

GLMM analysis

Number of words read in ER training

Statistically significant increases emerged among the three time points in the proportions of Post1 (148,583, SD = 3,6739) words and Post2 (289,040, SD = 85,809) words from baseline on word counts in ER (F(2,46) = 216.14, p = .000). A Bonferroni post-hoc test was performed and was significant for pre vs. post 1 (p = .000, 95% CI = [-183,130, -114,040]), pre vs. post 2 (p = .000, 95% CI = [-323,590, -254,490]), and Post1 vs. Post2 (p = .000, 95% CI = [-175,010, -105,910]) between any of the groups.

Listening task results

The RT in the listening task performed at the three time points was Pre: 7560.04 ms, SD = 2094.11 ms; Post 1: 6080.22 ms, SD = 2079.29; and Post2: 6194.27 ms, SD = 2228.17 with a significant difference (F(2,46) = 10.32, p = .000). After the Bonferroni post-hoc test, there was a significance decrease in RT at post 1 (p = .001, 95% CI = [579.096, 2380.556]) and post 2 (p = .001, 95% CI = [465.041, 2266.501]) compared with pre. There was no significant difference between post1 and post2 (p = 1.000, 95% CI = [-1014.79, 786.674]) (Fig. 3).

The AR of the listening task performed at the three time points was pre: 54.69, SD = 8.82; post 1: 54.58, SD = 7.82; and post 2: 56.47, SD = 6.63. There were no significant differences (F(2,46) = 7.28, p = .002)(Fig. 2).

BA results

In BA during the listening task, there were significant differences among the three time points: pre: left tHb (0.0111, SD = 0.11), right t-Hb (-0.0104, SD = 0.12); Post1: left t-Hb (-0.0375, SD = 0.13), right t-Hb (-0.0295, SD = 0.13); and Post2: left t-Hb (-0.1101, SD = 0.22), right t-Hb (-0.0851, SD = 0.12); (F(2,46) = 10.32, p = .000). The main effect of the three time points on BA revealed a significant decrease (F(2,115) = 7.362, p = .001). After the Bonferroni post-hoc test, there was a significant decrease in BA at Post 2 (p = .001, 95% CI = [0.035, 0.161]) compared with pre and at post 2 (p = .045, 95% CI = [0.001, 0.127]) compared with Post1. There was no significant difference between pre and post 1 (p = 0.584, 95% CI = [-0.029, 0.097]) (Table 3).

Table 3

Blood flow change (mMmm) in Listening tasks
		Mean	SD	Range
Listening	Pre_Left	0.0111	0.1097	-0.3120–0.2072
	Post1_Left	-0.0375	0.1322	-0.2203–0.1886
	Post2_Left	-0.1101	0.2154	-0.9756–0.0950
	Pre_Right	-0.0104	0.1157	-0.3952–0.1366
	Post1_Right	-0.0295	0.1329	-0.2962–0.3374
	Post2_Right	-0.0851	0.1174	-0.3934–0.0486

However, there was no main effect of left or right hemisphere on BA (F(2,115) = 0.033, p = .867). In addition, there was no interaction between the three time points and left-right BA (F = 0.413, p = .663).

TOEIC results

The total score on the TOEIC tests, which can test overall English proficiency, showed significant differences among the three time points: pre (414.79, SD = 132.33), post 1 (465.00, SD = 141.56), and Post 2 (466.46, SD = 155.15); (F (2,46) = 7.28, p = .002). After the Bonferroni post-hoc test, there was a significant increase is listening section scores at Post 1 (p = .006, 95% CI = [-88.528, -11.888]) compared with pre and at post 2 (p = .005, 95% CI = [-89.987, -13.347]) compared with pre. There was no significant difference between post 1 and post 2 (p = 1.000, 95% CI = [-39.778, 36.862]) (Fig. 1).

The aim of this study was to examine whether ER improved real-time language processing performance (i.e., RTs and ARs) and BA in the DLPFC during foreign language listening tasks. We investigated the effect of an 8-month ER training program on skill improvement during listening and BA in the DLPFC. There was a significant decrease in the RT and TOEIC scores between pre and post 1, but not between post 1 and post 2 (Fig. 3). No significant change in blood flow was observed between pre and post 1; however, a significant decrease was observed between post 1 and post 2 (Fig. 4). These results indicate that ER training affects both behavioral performance and BA during EFL listening tasks.

The current research directly observed real-time processing during foreign language listening tasks, which none of the previous ER studies have done. Our results showed a significant decrease in RT during the listening task between pre and post1, suggesting that ER training improved participants' listening proficiency. This result was supported by the TOEIC as a standardized test and showed a significant increase in the TOEIC listening section between pre and post1. This TOEIC result is in line with previous studies [21, 22, 19, 40] .

Furthermore, a significant decrease in blood flow in the DLPFC was observed between post 1 and post 2. This can be interpreted as learners using less oxyhemoglobin (oxy-HB) when their skills improve. This phenomenon is explained by performance automation with less demand for control of attention [38]. Similar phenomena have been observed in non-motor learning, i.e., language learning, and a decrease in RT and blood flow has been reported [41]. Decreased BA was observed in familiar stimuli letters compared with unfamiliar stimuli [30]. Another study revealed that a group of advanced learners of English showed greater activation in the left PFC while working on low-frequency word translation tasks but no activation with high-familiarity words [42, 32]. In addition, greater involvement of the DLPFC in reading by advanced L2 learners has been reported [43]. Many other studies have reported that L2 is processed in brain regions, including the DLPFC [29]. Therefore, our results may indicat e that a certain amount of ER training enhances listening comprehension with less effort in the DLPFC.

Another possible interpretation is that the factors that affected the decrease of BA are the practice or repetition effects. Actually, the fact that the exposure of the same stimuli decreases BA as “adaptation” is well-established in previous studies on functional brain imaging (e.g., Larssona et al., 2016 for a review) [44]. However, this interpretation may not be fit our result. First, as explained in the method section, all participants were trained in the trial session just before the main experimental sessions for pre, post 1, and post 2. Additionally, the same stimuli were not given to each participant among pre, post 1, and post 2. Second, to our knowledge, there were no previous functional brain studies on adaptation, repetition, or priming in L2 processing, except the one on the priming effects in L1 [45]. In this L1 priming study, there was no decrease in BA in the DLPFC during L1 processing of auditory sentence stimuli. Therefore, the decrease in BA observed in the current study is not considered to be a practice effect/repetition effect.

This result also suggests that ER training improved participants' listening proficiency. Some neurophysiological studies have reported that BA for a certain task decreases in a related area as a result of practice [46]. For example, significantly higher levels of oxy-Hb were used when people played harder pieces on the piano than when playing easier pieces [47]. The use of less oxy-HB means either improved skills or the use of relatively less WM during a performance. With improved proficiency in L2 processing, such as in translation tasks, lesser activation of WM is observed [33]. Increased BA in the DLPFC was reported during lexical processing, which is considered to be WM in L2 processing, such as in relatively difficult picture naming tasks. Our results are in line with these previous findings.

Notably, there was a significant difference between RT and BA at different times. At pre- post 1, RT became quicker without any decrease in BA; on the contrary, at post1- post 2, there was no significant decline in RT but a significant decrease in BA. While it can be assumed that they usually coincide, a recent study revealed a similar finding to our results [48]. Wang et al. (2022) examined the cross-linguistic influences of English language learners' first language on morphosyntactic processing [48]. They compared three different L1 groups, which were native speakers of Mandarin, Spanish, and English. They found no significant difference in behavioral data, AR, and RT between Mandarin and Spanish, but those groups showed different neural activation patterns. This result suggests that even if there is no difference in behavioral indicators, differences in BA can occur owing to different ways of processing in the brain. Thus, this previous study supports our results.

The current study reports the significant impact of ER training on RT and BA in foreign language real-time listening processing. The question remains as to why foreign language reading training improves foreign language listening skills. This may be due to the very similar psychological processes of reading comprehension and listening. When we read letters silently, we usually sense the need to read aloud in our brains. This is because when humans read a word or sentence aloud or silently, they are supposed to go through the so-called “grapheme-phoneme conversion” process (phonological coding). This means that the pronunciation (sound) of words must be corrected before extracting a word from the mental lexicon. The process of inward rehearsal of the spoken forms of words is known as subvocalization [49, 26]. As already defined, ER is the practice of reading easy books for readers’ proficiency so that they can extract the meaning of words and understand sentences instantly. Thus, through ER training, learners repeatedly go through this subvocalization process, which trains them to accelerate lexical access [31]. This process may be similar to listening to spoken words, and it may consequently enhance listening proficiency. However, to our knowledge, no studies have clarified this hypothesis. Further research is required to investigate this hypothesis.

Limitations

Our study has some limitations. First, the number of participants was only 24, which is insufficient to generalize our findings. Second, the number of words in ER may be insufficient. A larger number of participants and more than 500,00 words (a year-long study) would be preferable for examining the effects of ER. Third, it was difficult to collect behavioral data on listening performance. In this study, we found no change in AR, which might be because of the difficulty of the task, which involved listening to a 200-word-long material (approximately 90 seconds long) with natural speed recording (150 WPM) and answering seven True/False questions. Therefore, it is necessary to use more valid stimuli to measure the BA while listening in future studies. Lastly, as mentioned in the discussion, the current study observed BA only in the DLPFC. Future research needs to observe specific language-related areas, such as Broca’s field and Wernicke’s field. This will reveal if there is a delay in improved proficiency in the DLPFC. Furthermore, a control group that does not do ER should be established to verify the effects of ER. The current study had no control condition because it was conducted in the compulsory subject. Future research should recruit a sufficient number of participants who do out-of-class ER at a sufficient level and those who do not do ER as part of a control group.

Conclusion

We conducted a longitudinal intervention study to examine the direct effects of ER training on EFL listening performance; real-time processing and associated RTs and BA data were examined during foreign language listening tasks. Behavioral data, such as RT, AR, and blood flow changes during the foreign language listening task, were measured. A total of 150,000 words of ER training demonstrated an effect on listening processing speed (RT) and general listening proficiency (TOEIC test score), followed by the neurological improvement that accompanied an additional 150,000 words of ER. The findings proved that reading training affects other language skills, particularly behavioral and neurophysiological listening skills. Further research is necessary to examine why ER has an effect on listening proficiency in more language-specific areas of the brain and with a larger number of participants.

Declarations

8. Conflict of Interest

The author(s) declare no competing interests.

Funding

This work was supported by JSPS KAKENHI (grant number 19H01287).

Author Contributions

KC devised the idea of this study and selected and created the experimental stimuli. KC and AM collected the fNIRS data and other behavioral data. SY performed data analyses. KC, AM, and SY wrote the manuscript. All the authors have read the final version of the manuscript and agreed to its publication.

Ellis, R.. Principles of instructed language learning. System. 33, 209–224 (2005).
Furukawa, A., and Kanda, M. (eds). The Complete Book Guide for Extensive Reading, 4th ed. (Tokyo: CosmoPier Publishing Company 2013).
Suk, N. The effects of extensive reading on reading comprehension, reading rate, and vocabulary acquisition. Reading Res. Q. 52, 73–89 (2017).
Education Testing Service (2020). 2020 Report on Test Takers Worldwide. https://www.ets.org/s/toeic/pdf/2020-report-on-test-takers-worldwide.pdf [Accessed March 29, 2022]
IELTS (2020). Test Taker Performance. https://www.ielts.org/for-researchers/test-statistics/test-t aker-performance [Accessed March 29, 2022]
Educational Testing Service (2020). TOEFL iBT Test and Score Data Summary 2020.
Crystal, D. English as a Global Language, 2nd ed. (New York: Cambridge University Press, 2012) .
Bell, T. Extensive Reading: Speed and Comprehension. The reading matrix. 1(1) (2001).
Beglar, D., Hunt, A. & Kite, Y.. The effect of pleasure reading on Japanese university EFL learners’ reading rates. Lang. Learn. 62, 665–703 (2012).
He, M. Does extensive reading promote reading speed? The reading matrix. 14, 16–25 (2014).
Nation, P. & Wang Ming-Tzu, K. Graded Readers and Vocabulary. Reading Foreign Lang. 12 (1999).
Nation, I. & Waring, R. Teaching Extensive Reading in Another Language. (New York:Routledge. 2019)
Nakanishi, T. A meta-analysis of extensive reading research. TESOL Q. 49, 6–37 (2015) .
Day, R. R. & Bamford, J.. Extensive Reading in the Second Language Classroom. (Cambridge: Cambridge UniversityPress, 1998)
Imamura, K. The effects of extensive reading for Japanese high school students on their reading and listening abilities, vocabulary and grammar. ARELE. 19, 11–20 (2008) .
Fujii, K. Effects of the 1 year English extensive reading class on the reading consciousness and the results of TOEIC. The Bull. Chube Engl. Lang. Educ. Soc. 42, 197–202 (2013).
Chiba, K. Effectiveness and success factors of teaching extensive reading. The cross-linkage to the world and the future. Soei-sha, 181–203 (2017a).
Takeuchi, H. & BeGole, B. Extensive reading in English education - Efficacy and future challenges. Tottori Kankyo Daigaku Kiyo, 119–127 (2015).
Chiba, K. The effect of extensive reading on reading speed in second language learning. J. Fac. Int. Stud. Bunkyo Univ. 28, 57–65 (2017b).
Yamashita, J. Extensive reading and development of different aspects of L2 proficiency. System. 36, 661–672 (2008).
Fujimori, C. The effects of an extensive reading program on reading and listening comprehension among senior high school students. The Bull. Kanto-Koshin-Etsu Engl. Lang. Educ. Soc. 21, 13–24 (2007).
Mayuzumi, M. & Miyatsu, T. The effectiveness of extensive reading to improve English listening ability. The Iryo-Kango-Kenkyu. 7, 59–65 (2011).
Robb, T. & Kano, M. Effective extensive reading outside the classroom: a large-scale experiment. Reading Foreign Lang. 25, 234–247 (2013).
Krashen, S. & Mason, B. Can second language acquirers reach high levels of proficiency through self-selected reading? An attempt to confirm nation's (2014) results. Int. J. Engl. Lang. Teach. 10, 10–19 (2015).
Hamada, A. Using meta-analysis and propensity score methods to assess treatment effects toward evidence-based practice in extensive reading. Front. Psychol. 11, 617 (2020).
Leinenger, M. Phonological coding during reading. Psychol. Bull. 140, 1534–1555 (2014).
Warren, P. Introducing Psycholinguistics. (Cambridge:Cambridge University Press, 2013).
Hernandez, A. E. & Meschyan, G. Executive function is necessary to enhance lexical processing in a less proficient L2: evidence from fMRI during picture naming. Bilingualism Lang. Cogn. 9, 177–188 (2006).
Jung, J., Nam, K., Cho, H., & Kim, S.. Mapping the neural dynamics of Korean–English bilinguals with medium proficiency during auditory word processing. Front. Psychol. 9, 983 (2018).
Yokoyama, S et al. Cortical activation in the processing of passive sentences in L1 and L2: an fMRI study. Neuroimage 30, 570–579 (2006).
Fisher, J. E., Cortes, C. R., Griego, J. A., & Tagamets, M. A.. Repetition of letter strings leads to activation of and connectivity with word-related regions. NeuroImage. 59, 2839–2849 (2012).
Shinozuka, K et al. Language familiarity and proficiency leads to differential cortical processing during translation between distantly related languages. Front. Hum. Neurosci. 15, 593108 (2021).
Rinne, J.O et al. The translating brain: cerebral activation patterns during simultaneous interpreting. Neurosci. Let. 294(2), 85–88 (2000).
Baddeley, A. Working memory and language: an overview. J. Com Dis 36(3), 189–208 (2003).
Nation, I. & Marlarcher, C. Timed Reading for Fluency 1. (Sachse,TX:Seed Learning, 2018).
Grill-Spector, K., Henson, R., & Martin, A. Repetition and the brain: neural models of stimulus-specific effects. Trends Cogn Sci. 10(1), 14–23.
Nouchi, R et al. Dorsolateral prefrontal cortex activity during a brain training game predicts cognitive improvements after four weeks’ brain training game intervention: evidence from a randomized controlled trial. Brain Sci. 10 (2020).
Olesen, P. J., Westerberg, H. & Klingberg, T. Increased prefrontal and parietal activity after training of working memory. Nat. Neurosci. 7, 75–79 (2004).
Cheng, J et al. Real longitudinal data analysis for real people: building a good enough mixed model. Stat. Med. 29, 504–520 (2010).
Yoshii, M. The effects of high input on English learning: changes in learners in an extensive reading and listening class. Kumamoto Kenritsu Daigaku Kiyo, 23,59–71 (2017).
Raichle, M. E et al. Practice-related changes in human brain functional anatomy during nonmotor learning. Cereb. Cortex. 4, 8–26 (1994).
Jost, et al. Effects of continuous theta burst stimulation over the left dlpfc on mother tongue and second language production in late bilinguals: a behavioral and ERP study. Brain Topogr. 33, 504–518 (2020).
Gurunandan, K., Carreiras, M. & Paz-Alonso, P. M. Functional plasticity associated with language learning in adults. NeuroImage. 201, 116040 (2019).
Larsson, J., Slomon,S.G, & Kohn Adam. fMRI adaptation revisited. Cortex 80: 154 – 160 (2016).
Devauchelle, A., Oppenheim, C., Rizzi, L., Dehaene, S. & Pllier, C. Sentence syntax and content in the human temporal lobe: an fMRI adaptation study in auditory and visual modalities. J Cogn Neurosci. 21(5): 1000 – 1012 (2009).
Kelly, A. M. C. & Garavan, H. Human functional neuroimaging of brain changes associated with practice. Cereb. Cortex. 15, 1089–1102 (2005).
Yuksel et al . Learn piano with BACh: an adaptive learning interface that adjusts task difficulty based on brain state, in Proceedings of the 2016 Chi Conference on Human Factors in Computing Systems, 5372–5384 (2016).
Wang, D., Wang, S., Zinszer, B., Sheng, L. & Jasińska, K. Cross-linguistic influences of L1 on L2 morphosyntactic processing: an fNIRS study. J. Neurolinguist. 63, 10163 (2022).
Warren, P. (2013). Introducing Psycholinguistics. Cambridge:Cambridge University Press.
Klingberg, T., Forssberg, H. & Westerberg, H. Training of working memory in children with ADHD. J. Clin. Exp. Neuropsychol. 24, 781–791 (2002).
Acknowledgmentsb;
The authors thank Paul Nation and Casey Malarcher for allowing us to use their text as experimental stimuli. We would also like to take the opportunity to acknowledge the cooperation of Atsuhi Iida and Barry Keith in recruiting participants and providing experimental space. Finally, we are grateful to all the participants of Gunma University.

Fig. 1

TOEIC Score

Fig. 2

Accuracy Rate

Fig. 3

Reaction Time

Fig. 4

Blood Flow Change (t-HB)

Fig. 1

TOEIC Results

Fig. 2

Accuracy Rate in listening tasks

Fig. 3

Reaction Time in the listening tasks

Fig. 4

Blood flow changes

Figure Legends

Figure Captions

TOEIC scores were measured pre, post1, and post2 to see changes in English proficiency. Table 1 shows the results for the three categories of total score, reading score, and listening score.

The AR of the listening task performed at the three time points was measured.

The average reaction time in the listening task was measured (ms).

The concentration changes of total hemoglobin (total Hb) while the participants listened to the stimulus was measured.

No competing interests reported.

Download PDF

Version 1

posted

You are reading this latest preprint version

Extensive Reading Affects Second Language Listening Proficiency: an fNIRS Study

Status:

Version 1

Abstract

1. Introduction

2. Materials and Methods

3. Discussion

Declarations

Declarations

Funding

Author Contributions

References

Unsectioned Figure Details

Unsectioned Paragraphs

Additional Declarations

Status:

Version 1