丁香园AV

Reading Across Writing Systems:
The System Accommodation Hypothesis
in Cross-Linguistic Visual Word Recognition

Introduction

Although visual word recognition has received much interest in the field of psycholinguistics, a considerable amount of the research done on this topic has focused on languages written in the Latin alphabet, particularly the English language. As a result, many of the basic assumptions that have been made about visual word recognition do not reflect universal, cross-linguistic truths: for example, although left hemisphere lateralization, which is found in languages such as English, has long been believed to be a given feature of language processing (e.g., Maurer & McCandliss, 2008), counterevidence has been found in non-alphabetic languages such as Hindi (Singh & Rao, 2014) and varieties of Chinese (e.g., Hsiao & Lam, 2013). The emergence of research from languages in different writing systems continues to challenge our presumptions about how words are visually identified, signalling a necessary change in how we study visual word recognition as a whole.

In order to address this issue, this paper will use the System Accommodation Hypothesis (Perfetti & Liu, 2005) to explain how different writing systems are accommodated by cognitive processes and networks. After introducing this model, a literature review will demonstrate both universal and writing system-specific properties of visual word recognition. From the vast research done across writing systems, we can conclude that the unique visual characteristics of individual writing systems, as well as those characteristics on the level of orthography, have a direct impact on a reader鈥檚 cognitive processes and the neural systems underlying them.

Writing Systems, Orthographies, and Scripts

Although the terms 鈥渨riting system,鈥� 鈥渙rthography,鈥� and 鈥渟cript鈥� are often used interchangeably, it is important to keep in mind that they refer to different ideas. A writing system represents the fundamental design in the relationship between writing and language (Perfetti & Liu, 2005). An alphabetic writing system, for example, maps a single grapheme (letter) to a single phoneme, whereas a logographic writing system maps a single grapheme (character) to a morpheme or a whole word. On a smaller scale, orthographies apply these writing systems to individual languages (Perfetti & Liu, 2005). Although English, French, Swedish, and other European languages use the Latin alphabet, thus utilizing the same writing system, the ways in which this alphabet is implemented in each individual language varies. The term script is used to refer to the visual aspect of writing (Perfetti & Liu, 2005), although this factor is excluded from the writing-language relationship.

Now that the distinction between writing systems, orthographies, and scripts has been established, it is possible to discuss the different major writing system types. Although the categorizations of each may vary depending on the source, the general consensus seems to be that there are five distinctive writing system types: abjads, abugidas (or alphasyllabaries), alphabets, logographies, and syllabaries. Note that these categorizations are subject to change as research into different writing systems continues.

The abjad writing system is used in Semitic languages such as Arabic, Hebrew, and Syriac. Like alphabets, abjads are phonemic, but they differ in that they are consonantal: letters correspond to consonants, but not to vowels (Share & Bar-On, 2018). Abjads may leave vowels entirely up to the reader鈥檚 interpretation, or they may depict vowels in another form. Hebrew, for example, may represent vowels as diacritics or as dual-purpose letters which double as vowels. This form of Hebrew is referred to as pointed, but its unpointed counterpart, in which vowels are inconsistent in or entirely absent from a given representation, is considered standard (Lahoud et al., 2023).

The abugida writing system, also referred to as an alphasyllabary, is employed in Canadian syllabic, Ethiopic, and Indic languages. In abugidas, each grapheme represents a syllable unit consisting of a consonant and a vowel phoneme. For some languages descended from the Brahmi script, including Malayalam, this vowel is the phoneme /a/, which may be realized with different pronunciations (Vasudevan et al., 2023). In Kannada, from the same language family as Malayalam, orthographic units called akshara may encode vowels other than /a/ as well as consonant clusters (Nag & Snowling, 2012).

The most salient feature of alphabetic writing systems is the correspondence between a letter and a phoneme, although the consistency of this correspondence may vary from language to language. Alphabets are generally thought of in terms of Latin-derived alphabetic languages, but other languages also utilize alphabetic writing systems. Korean, for example, is generally classified as an alphabetic language due to its high letter-phoneme correspondence (e.g., Pae, 2020), although it utilizes syllabic blocks that more closely resemble logographic characters.

Along with alphabetic writing systems, logographies have received a great deal of interest in visual word recognition research. These two writing systems are the most used globally (Pae, 2020), and as such there is a vast amount of evidence for the processes behind reading in both. As mentioned previously, logographies represent a morpheme or a whole word by a single grapheme (character). Chinese scripts are most often cited as examples of logographies: although there are numerous varieties of spoken Chinese, they use the same Modern Standard Chinese writing system (Li, 2019), which is why they tend to be grouped together in visual word recognition research. However, it is important to note that no modern language, including those in the Chinese language family, can be considered a 鈥減ure鈥� logography. A majority of Chinese characters are characterized as semantic-phonetic compounds, which include both semantic and phonetic radicals; only pictographic and indicative characters can be considered purely logographic, as they represent morphemic information alone (Hsiao & Shillcock, 2006). Despite some arguments that Chinese is better classified as a morphosyllabary (e.g., Li, 2019), this paper will categorize Chinese scripts and Chinese-derived scripts, such as the Japanese kanji, as logographies, following the general consensus in writing system research.

The final writing system of interest is the syllabary, in which a single grapheme represents a syllable. Syllabaries share some features with abugidas, which are also called alphasyllabaries due to their representation of syllables by graphemes. However, where abugidas focus on consonants as the visual basis of a grapheme, graphemes in syllabaries do not have such an emphasis. Abugida syllables beginning with the same consonant tend to share visual characteristics in grapheme form; this is possible in syllabaries, but not necessary or formal, and words possessing the same phonemes may not share any visual characteristics as graphemes. In syllabaries, the consonant-vowel unit is common: the syllabic Japanese script kana is mostly made up of this type of syllable (Rayner et al., 2011a).

The System Accommodation Hypothesis

In their 2005 review, Perfetti and Liu examined two universal principles of writing systems in order to form the System Accommodation Hypothesis. The Universal Writing System Constraint (Perfetti, 2003) states that all writing systems encode the specific characteristics of the language which they represent, and therefore certain properties of this language must be accommodated in its visual form. The Universal Phonological Principle (Perfetti, Zhang, & Berent, 1992) states that the reading of a word prompts the activation of phonology at the lowest level that the writing system permits.

After applying these two universals to a set of studies both within and across writing systems, Perfetti and Liu proposed the System Accommodation Hypothesis, which suggests that both the cognitive processes involved in reading as well as the neural structures underlying them adapt in order to accommodate the attributes of a writing system. Event-related potential (ERP) studies of Chinese and English provided especially useful evidence for this hypothesis. By examining these studies, the authors found that the visual processing requirements of a given writing system鈥攆or example, in logographic languages compared to alphabetic languages鈥攎ay influence which brain regions are involved in processing, regardless of whether the reader is a native speaker or a learner. In other words, the visual demands of a logographic writing system had to be accommodated by the reader鈥檚 cognitive processes and neural structures, an effect which was seen in native speakers of both logographic and alphabetic languages.

Although this hypothesis was proposed tentatively in Perfetti and Liu鈥檚 review, it has since become influential in visual word recognition research, with high experimental support and academic interest (e.g., Liu et al., 2007; Perfetti et al., 2007; Pae, 2020). For the purposes of this paper, the System Accommodation Hypothesis is especially significant, as it can be applied both to a singular writing system as well as writing systems in comparison. It allows us to assess how the visual demands of a given writing system are reflected in the visual word recognition of native speakers, but it can also be extended to cross-writing system studies, where differences arise due to writing system demands. Therefore, this hypothesis will prove useful in comparing studies that span across several of the major writing systems discussed previously. We should expect to see more cognitive and neural similarities in writing systems that are visually similar; furthermore, differences that arise in visual word recognition across writing systems should be able to be linked back to each writing system鈥檚 unique visual demands.

Literature Review

Having established the major writing system types and a model by which they may be analyzed, it is possible to consider the different methods, techniques, and technologies used to conduct research into visual word recognition. Although behavioural measures, including reading speed and accuracy, were traditionally used in such studies (Lahoud et al., 2023), newer technologies have provided researchers with a number of ways to study visual word recognition. These can be broadly divided into brain responses, computer modelling, eye movement, literacy acquisition studies, and priming studies. Because many studies have significant results for more than one writing type, the literature review will proceed in subsections based on the method or technology used.

Brain Responses

Brain responses during visual word recognition provide precise measurements of cognitive processing. The continued development of neuroscientific technologies means that brain responses can be measured in a variety of ways.

Functional magnetic resonance imaging (fMRI) allows researchers to identify the brain regions associated with visual word recognition. Singh and Rao (2014) used fMRI to determine which brain regions were activated in native Hindi speakers during reading. They found that Hindi, an abugida written in the Devanagari script, activated a neural network in readers which is associated universally with visual word recognition. Notably, complex nonlinear words (e.g., 啶�啷� 啶�, 鈥渇lower鈥�) were found to additionally activate the right occipito-temporal cortex while the left was already activated, a finding which has also been found in Chinese (Tan et al., 2005), Japanese (Ino et al., 2009), and Korean (Yoon et al., 2005). Visual word recognition in alphabetic languages generally activates the left occipito-temporal cortex (Moore & Price, 1999). Singh and Rao suggested that right hemisphere activity in certain writing systems and orthographies may be due to their increased visual complexity.

Measuring event-related potentials (ERPs) is another popular method in visual word recognition research. ERPs are used to measure a number of different processes: for example, Liu and Perfetti (2003) used the principal component analysis (PCA) in order to measure temporal differences in reading Chinese and English. Analysis of the PCA allowed them to discover that Chinese-English bilinguals experienced faster visual processing and faster shifts to pronunciation processing in Chinese than in English, potentially due to the bilateral hemispheric activation discussed above.

Computer Modelling

Computer modelling is a useful tool for studying visual word recognition as it allows researchers to simulate the brain and its cognitive processes. Hsiao and Lam (2013) applied a hemispheric processing model of face recognition (Hsiao et al., 2008) to examine whether the visual demands of writing systems are responsible for lateralization differences in different languages. The use of a computer model allowed them to manipulate different visual aspects of words in order to see which aspects influenced lateralization, and in what ways.

A left hemisphere (LH) lateralization effect is commonly found in alphabetic languages during visual word recognition. Contrary to the phonological mapping hypothesis (Maurer & McCandliss, 2008), which states that visual word processing is generally LH lateralized in accordance with phonological processing lateralization, Hsiao and Lam (2013) hypothesized that the LH lateralization seen in alphabetic languages is due to lowered visual similarity between words as well as a grapheme decomposition requirement. In other words, because alphabetic letters encode single phonemes, a word made of multiple letters must be decomposed into its respective phonemes during reading. This was contrasted to logographic reading, represented by Chinese characters, which demonstrates a RH lateralization effect for single characters. In order to test their hypothesis, the authors ran two simulations. The first simulation compared the effect of a letter identity mapping task, which is similar to alphabetic reading, and a word identity mapping task, which is similar to logographic reading. The second simulation mapped word images to pronunciations: in the alphabetic reading condition, this meant that each grapheme in a word was systematically mapped to a phoneme; in the logographic reading condition, each word was mapped to a syllabic pronunciation, and grapheme-phoneme relationships were unsystematic. The results of these simulations demonstrated LH lateralization in the alphabetic condition as predicted. The authors suspected that RH lateralization for Chinese reading was due to two main factors: one, Chinese characters tend to be more visually similar to one another than English letters are; and two, Chinese does not require grapheme decomposition. Through computer modelling, Hsiao and Lam were able to conclude that the LH laterality of language processing does not necessarily cause LH laterality in visual word recognition, with a logographic writing system as an example of one that activates RH lateralization.

Eye Movement Tracking

Eye movement tracking is a popular technique in psycholinguistics. It is particularly useful in visual word recognition studies, where fixation durations clearly indicate the cognitive processes involved in word recognition (Rayner, 1998).

Word length and word frequency appear to be common factors influencing fixation periods, observed in Arabic (Paterson et al., 2015) and English (Rayner et al., 2011b), among other languages. Paterson et al. found that longer words resulted in longer and more frequent fixation durations for readers of Arabic. Rayner et al. reached a similar conclusion for English readers, who demonstrated lengthened fixation durations in higher numbers for longer and less predictable words. Eye tracking studies help us to determine which factors, such as word length and frequency, should be expected to have a universal effect on visual word recognition, occurring across writing systems such as abjads and alphabets.

Eye movement tracking has also corroborated findings from naming time studies, which demonstrate that real words are read more efficiently and with shorter and less frequent durations than pseudowords. This finding, which is referred to as the lexicality effect, has been found in naming time studies in different languages, including alphabets like English (Forster & Chambers, 1973) and syllabaries like Japanese (Besner & Hildebrandt, 1987). By using eye movement tracking, Lahoud et al. (2023) found that both child and adult readers of Hebrew experienced a lexicality effect, as pseudowords elicited longer and more frequent fixation durations than real words. The lexicality effect can then be considered another cross-linguistic effect of visual word recognition, as it occurs in Hebrew, an abjad, as well as other writing systems.

Literacy Acquisition

Literacy acquisition can also be studied in order to better understand a given writing system. Vasudevan et al. (2023) tested children鈥檚 literacy acquisition in Malayalam, a Dravidian abugida, by administering reading and writing tests. They compared their results to the parameters set out by Daniels and Share (2017), who developed a universal framework for assessing writing system complexities. These parameters include: (1) linguistic distance, (2) nonlinearity, (3) visual complexity, (4) historical change, (5) spelling constancy despite morphophonemic alternation, (6) omission of phonological elements, (7) allography, (8) dual purpose letters, (9) ligaturing, and (10) inventory size (Daniels & Share, 2017). Of these ten parameters, Vasudevan et al. found that all but 5, 6, and 8 applied to literacy acquisition in Malayalam. They suggested that these parameters are not relevant to Malayalam as it has a high spelling-sound correspondence; learners of languages such as English, which lacks such a consistent correspondence, may be affected by these dimensions. Furthermore, the authors recommended the addition of an eleventh dimension of word length. As discussed previously, word length is shown to have a universal effect on reading, with longer words being associated with more difficulties in processing. This dimension was also shown to pose difficulties to children learning Malayalam.

Literacy acquisition studies are useful for our purposes because they are directly linked to visual word recognition, indicating which factors of writing systems result in difficulties in cognitive processing. Furthermore, the results of these studies allow us to develop testable parameters in order to determine whether such factors are cross-linguistic or whether they are localized to a particular writing system.

Priming

The last method to be discussed is priming. Priming studies are popular in visual word recognition research as they allow researchers to manipulate certain types of processing. For example, Katz and Feldman (1983) used semantic priming in order to manipulate lexically mediated processing in two alphabetic languages, English and Serbo-Croatian. Unlike English, Serbo-Croatian has a high spelling-sound correspondence. It was therefore hypothesized that Serbo-Croatian readers would rely more on articulatory coding鈥攐r, in terms of the dual-route model (Coltheart et al., 2001), the nonlexical or assembled route鈥攖han English readers, who rely on the lexical or direct route to directly access a lexical representation from the mental lexicon.

The authors found that semantic priming facilitated both lexical decision and naming in English, but only lexical decision in Serbo-Croatian. The implications of this study are interesting not only in the context of so-called 鈥渄eep鈥� orthographies such as English, but also in the broader context of writing systems that lack a direct correspondence between spelling and sound. In logographies, for example, we know that graphemes encode words and morphemes rather than phonemes. We would then expect a reliance on the direct route during visual word recognition; a semantic priming effect for both lexical decision and naming, if found experimentally, would support this expectation.

Hanavan and Coney (2005) used a similar associative priming paradigm to study lateralization effects during visual word recognition in different Japanese scripts. They found different patterns of priming depending on which script was used: kanji, the logographic script, or kana, the syllabic script. Words in kana showed a right visual field/left hemisphere advantage, though they demonstrated a priming effect in either visual field. Kanji characters showed a left visual field/right hemisphere preference, with a stronger priming effect for characters projected to this visual field. Interestingly, responses to kanji stimuli were on average 30 ms faster than those to kana stimuli, which is perhaps indicative of a heightened semantic priming effect for logographic writing systems as proposed above. The priming effects demonstrated in this study support the previously discussed findings regarding lateralization effects across writing systems.  

Conclusion

Visual word recognition research has historically focused on languages written in the Latin alphabet, especially English. For this reason, it should not be surprising that there is an abundance of evidence to support what we know about how English and other Latin-derived alphabetic readers recognize written words. It is important, however, to ensure that visual word recognition findings and subsequent models are not biased toward one language or set of languages.

Continued research into less-studied writing systems will provide us with a better understanding not only of those writing systems and of the languages within them, but also of the truly universal principles of visual word recognition. The literature discussed has shown that some effects, including the lexicality effect, can be considered cross-linguistic, as they are supported by evidence from several writing systems with different visual characteristics and processing demands. On the other hand, effects that were once assumed to be cross-linguistic, such as left hemisphere lateralization, may be challenged or even disproved as evidence from different writing systems emerges. Future development should focus on testing the parameters discussed in the literature review across writing systems, and especially those that are less often researched, in order to better understand effects that have been previously discovered. While continued research into widely studied writing systems, such as alphabets and logographies, is certainly beneficial, increased research into abjads, abugidas, and syllabaries may balance and better explain what has already been discovered. Furthermore, as technology continues to develop, we may see an increased ability to research visual word recognition with the use of improved computer models and artificial intelligence. Such technologies may yield new and significant findings, and it is important to incorporate these discoveries into what we already know about visual word recognition.

In the study of writing system effects, one must also consider the effects of orthography. We have seen in the studies above that languages which fall under the same writing system may still have significantly different visual word recognition processes. Languages vary in many dimensions鈥攁s we have seen, for example, in the different levels of spelling-sound correspondence in English and Serbo-Croatian. Effects may arise on levels of analysis below that of writing systems, and it is important to bear this in mind when studying visual word recognition.

The literature reviewed in this paper provides support for the System Accommodation Hypothesis (Perfetti & Liu, 2005), which suggests that both the cognitive processes and neural structures of reading are influenced by the visual demands of a writing system. In terms of cognitive processes, we have seen that some factors create processing difficulties universally, such as word length and word frequency. Other factors, however, such as spelling constancy and phonological omissions, may apply to certain writing systems or orthographies but not to others, as we saw in Vasudevan et al.鈥檚 (2023) study of Malayalam literacy acquisition. The fact that some writing systems pose different challenges than others on a visual basis corroborates the basic claim made by the System Accommodation Hypothesis. In terms of neural structures, brain responses and computer modelling show that different writing systems can activate different brain regions. Writing systems associated with more visually complex scripts, such as the abugida Devanagari or various logographic scripts, may demonstrate bilateral activation that is generally not seen in other writing systems, such as alphabets. By continuing to pursue research into all writing system types, as well as the unique orthographies and scripts within them, we will acquire an increasingly accurate idea of which factors in visual word recognition are truly universal, and which ones can be attributed to the unique features of a given writing system.

References

Besner, D., & Hildebrandt, N. (1987). Orthographic and Phonological Codes in the Oral Reading of Japanese Kana. Journal of Experimental Psychology: Learning, Memory, and Cognition, 13(2), 335鈥�343.

Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: A dual route cascaded model of visual word recognition and reading aloud. Psychological Review, 108(1), 204鈥�256.

Daniels, P. T., & Share, D. L. (2017). Writing system variation and its consequences for reading and dyslexia. Scientific Studies of Reading, 22(1), 101鈥�116.

Forster, K. I., & Chambers, S. M. (1973). Lexical access and naming time. Journal of Verbal Learning and Verbal Behavior, 12(6), 627-635.

Hanavan, K., & Coney, J. (2005). Hemispheric asymmetry in the processing of Japanese script. Laterality, 10(5), 413鈥�428.

Hsiao, J. H., & Shillcock, R. (2006). Analysis of a Chinese Phonetic Compound Database: Implications for Orthographic Processing. Journal of Psycholinguistic Research, 35(5), 405鈥�426.

Hsiao, J. H., Shieh, D. X., & Cottrell, G. W. (2008). Convergence of the visual field split: hemispheric modeling of face and object recognition. Journal of Cognitive Neuroscience, 20(12), 2298鈥�2307.

Hsiao, J. H., & Lam, S. M. (2013). The Modulation of Visual and Task Characteristics of a Writing System on Hemispheric Lateralization in Visual Word Recognition鈥擜 Computational Exploration. Cognitive Science, 37(5), 861鈥�90.

Ino, T., Nakai, R., Azuma, T., Kimura, T., & Fukuyama, H. (2009). Recognition and reading aloud of Kana and Kanji words: An fMRI study. Brain Research Bulletin, 78(4鈥�5), 232鈥�239.

Katz, L., & Feldman, L. B. (1983). Relation between pronunciation and recognition of printed words in deep and shallow orthographies. Journal of Experimental Psychology: Learning, Memory, and Cognition, 9(1), 157鈥�166.

Lahoud, H., Share, D. L., & Shechter, A. (2023). A developmental study of eye movements in Hebrew word reading: The effects of word familiarity, word length, and reading proficiency. Frontiers in Psychology, 14.

Li, Y. (2019). Written Chinese. In The Chinese Writing System in Asia: An Interdisciplinary Perspective (1st ed., pp. 40鈥�44). Routledge.

Liu, Y., & Perfetti, C. A. (2003). The time course of brain activity in reading English and Chinese: An ERP study of Chinese bilinguals. Human Brain Mapping, 18(3), 167鈥�175.

Liu, Y., Dunlap, S., Fiez, J., & Perfetti, C. (2007). Evidence for neural accommodation to a writing system following learning. Human Brain Mapping, 28(11), 1223鈥�1234.

Maurer, U., & McCandliss, B. D. (2008). The development of visual expertise for words: The contribution of electrophysiology. In E. L. Grigorenko & A. J. Naples (Eds.), Single- word reading: Behavioral and biological perspectives (pp. 43鈥�63). Lawrence Erlbaum Associates Publishers.

Moore, C. J., & Price, C. J. (1999). Three distinct ventral occipitotemporal regions for reading and object naming. Neuroimage, 10(2), 181鈥�192.

Nag, S., & Snowling, M. J. (2012). Reading in an Alphasyllabary: Implications for a Language Universal Theory of Learning to Read. Scientific Studies of Reading, 16(5), 404鈥�423.

Pae, H. K. (2020). Linguistic Evidence for Script Relativity. In Script Effects as the Hidden Drive of the Mind, Cognition, and Culture (pp. 147鈥�74). Springer, Cham.

Paterson, K. B., Almabruk, A. A., McGowan, V. A., White, S. J., & Jordan, T. R. (2015). Effects of word length on eye movement control: the evidence from Arabic. Psychonomic Bulletin & Review, 22(5), 1443鈥�1450.

Perfetti, C. A. (2003). The Universal Grammar of Reading. Scientific Studies of Reading, 7(1), 3鈥�24.

Perfetti, C. A., Zhang, S., & Berent, I. (1992). Reading in English and Chinese: Evidence for a 鈥樷�榰niversal鈥欌�� phonological principle. In R. Frost (Ed.), Orthography, phonology, morphology, and meaning (pp. 227鈥�248). North-Holland.

Perfetti, C. A., & Liu, Y. (2005). Orthography to Phonology and Meaning: Comparisons Across and within Writing Systems. Reading and Writing, 18(3), 193鈥�210.

Rayner, K. (1998). Eye movements in reading and information processing: 20 years of research. Psychological Bulletin, 124(3), 372鈥�422.

Rayner, K., Pollatsek, A., Ashby, J., & Clifton, Jr., C. (2011a). Writing Systems. In The Psychology of Reading (2nd ed., pp. 23鈥�48). Psychology Press.

Rayner, K., Slattery, T. J., Drieghe, D., & Liversedge, S. P. (2011b). Eye movements and word skipping during reading: effects of word length and predictability. Journal of Experimental Psychology: Human Perception and Performance, 37(2), 514鈥�528.

Share, D. L., & Bar-On, A. (2018). Learning to Read a Semitic Abjad: The Triplex Model of Hebrew Reading Development. Journal of Learning Disabilities, 51(5), 444鈥�453.

Singh, N. C., & Rao, C. (2014). Reading in Devanagari: Insights from functional neuroimaging. Indian Journal of Radiology and Imaging, 24(1), 44鈥�50.

Tan, L. H, Laird, A., Li, K., & Fox, P. T. (2005). Neuroanatomical correlates of phonological processing of Chinese characters and alphabetic words: A meta-analysis. Human Brain Mapping, 25(1), 83鈥�91.

Vasudevan, N., Haridas, M., Nedungadi, P., Raman, R., Daniels, P. T., and Share, D. L. (2023). A multi-dimensional framework for characterizing the role of writing system variation in literacy learning: A case study in Malayalam. Reading and Writing.

Yoon, H. W., Cho, K. D., Chung, J. Y., & Park, H. W. (2005). Neural mechanisms of Korean word reading: A functional magnetic resonance imaging study. Neuroscience Letters, 373(3), 206鈥�211.