University of Southern California Dissertations and Theses

Cognitive-linguistic factors and brain morphology predict individual differences in form-sound association learning: two samples from English-speaking and Chinese-speaking university students

(USC Thesis Other)

Cognitive-linguistic factors and brain morphology predict individual differences in form-sound association learning: two samples from English-speaking and Chinese-speaking university students

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content
i

Cognitive-linguistic factors and brain morphology predict
individual differences in form-sound association learning--
two samples from English-speaking and Chinese-speaking
university students

by
Miao Wei
_______________________________________________________
A Dissertation Presented to the
Faculty of the Graduate School
University of Southern California

In Partial Fulfillment of the
Requirements of the Degree
Doctor of Philosophy
(Psychology)
December 2015

ii

Acknowledgments
As this doctoral project has come to an end, I would like to acknowledge those people
who have assisted me, encouraged me and supported me along the way.
My sincere thanks to my advisors Dr. Zhong-Lin Lu and Dr. Frank Manis. Dr. Zhong-Lin
Lu inspired me to explore the structural brain changes that may co-occur with second language
acquisition. His profound knowledge of MRI image processing, statistics and academic writing
has guided me to the end of this journey. Thanks to your commitment and patience! Dr. Manis
provided me with generous support during graduate school. Thanks for your understanding,
optimism and confidence in me!
I would also like to thank my doctoral committee members: Dr. Bosco Tjan, Dr. Stephen
Read and Dr. Elsi Kaiser for their valuable guidance and comments throughout my qualification
and dissertation projects.
Many thanks to my colleagues, former lab mates, collaborators, and friends: Mingxia
Zhang, Qinghua He, Leilei Mei, Feng Xue, Jongsoo Baek, Xiangrui Li, Carlos Cabrera, Jiajuan
Liu, Changbing Huang, Yukai Zhao, Yue Leng, Jin Li, Gui Xue, Chuansheng Chen, Anand
Joshi, Rachel Beattie, Suzanne Houston, Corianne Rogalsky, So Young Choi, Jia Li, Jinxia Ma,
Zhisen Jiang, Pinglei Bao, Jun Wang, Jie Yang. All of you have contributed to my spirit as a
doctoral candidate.
I want to express my warmest thanks to my family. Thanks to my parents and in-laws for
everything. Thanks to my husband, Yuan Zhang for his love and care since 1998. Thanks to my
sons Ryan and Kyle for the hugs. I love you all.

iii

Table of Contents

Title page ............................................................................................................................ i
Acknowledgments .............................................................................................................. ii
List of figures .................................................................................................................... vi
List of tables ..................................................................................................................... vii
Abstract ............................................................................................................................... 1
1. Introduction .................................................................................................................... 2
1.1 Why we need to predict form-sound association learning? ...................................... 2
1.2 Major factors in L2 acquisition ................................................................................ 3
1.2.1 Overview ......................................................................................................... 3
1.2.2 L1 background ................................................................................................ 3
1.2.3 L1 and L2 interaction ...................................................................................... 4
1.2.4 Behavioral predictors ...................................................................................... 6
1.2.5 Neuroanatomical predictors .......................................................................... 11
1.3 The rationale of using Hangul artificial form-sound association
learning paradigm and a 2 × 2 design ..................................................................... 15
1.4 Hypotheses ............................................................................................................. 17
2. Material and Methods ................................................................................................... 19
2.1 Participants ............................................................................................................. 19
2.2 Behavioral evaluation ............................................................................................ 19
2.2.1 Raven’s Advanced Progressive Matrices ...................................................... 20
2.2.2 Woodcock Reading Mastery Tests-Revised ................................................. 20

iv

2.2.3 Test of Word Reading Efficiency ................................................................. 21
2.2.4 Comprehensive Test of Phonological Processing ........................................ 22
2.2.5 Chinese group additional tests ...................................................................... 23
2.3 Artificial language, training procedure and outcome measures ............................. 23
2.3.1. Learning materials ....................................................................................... 23
2.3.2 Training procedure ........................................................................................ 25
2.3.3 Phonological learning outcome measures ..................................................... 26
2.4 Image acquisition ................................................................................................... 28
2.5 Image data processing: voxel-based morphometry (VBM) ................................... 28
2.6 Statistical analysis .................................................................................................. 28
2.6.1 Behavioral predictors .................................................................................... 28
2.6.2 Neural predictors ........................................................................................... 29
2.6.3 Hierarchical regression analysis ................................................................... 30
3. Results .......................................................................................................................... 31
3.1 Behavioral predictors ............................................................................................. 31
3.1.1 Descriptive statistics ..................................................................................... 31
3.1.2 ANOVA ........................................................................................................ 32
3.1.3 Correlations between the cognitive-linguistic
factors and outcome variables ....................................................................... 33
3.2 Neural predictors .................................................................................................... 38
3.3 Hierarchical regression analysis ............................................................................ 40
4. Discussion ..................................................................................................................... 45
4.1 Why English-speaking students learned faster than
Chinese-speaking students in alphabetic learning condition ................................. 45

v

4.2 Behavioral predictors ............................................................................................. 46
4.2.1 Native English speaker ................................................................................. 46
4.2.2 Native Chinese speaker ................................................................................. 48
4.2.3 About Raven ................................................................................................. 49
4.3 Neural predictors .................................................................................................... 50
4.3.1 Native English speaker ................................................................................. 50
4.3.2 Native Chinese speaker ................................................................................. 51
5. Conclusions .................................................................................................................. 54
Abbreviations ................................................................................................................... 55
References ........................................................................................................................ 57

vi

List of figures
Page
Figure 1. Training materials: 60 Hangul characters ......................................................... 25
Figure 2. λ fits in English-speaking group ....................................................................... 27
Figure 3. λ fits in Chinese-speaking group ....................................................................... 27
Figure 4. Interaction plot for λ of discrimination (A) and naming task (B) ..................... 33
Figure 5. Significant correlations between grey matter volumes
and discrimination task growth rates ................................................................. 38

vii

List of tables
Page
Table 1. Descriptive statistics for measures ..................................................................... 32
Table 2. Correlations among the principal variables in English group ............................ 35
Table 3. Correlations among the principal variables in Chinese group ............................ 37
Table 4. Brain regions showing significant correlations between
GM volumes and learning rates .......................................................................... 40
Table 5. English alphabetic group hierarchical regression predicting learning rate ......... 42
Table 6. Chinese alphabetic group hierarchical regression predicting learning rate ........ 43
Table 7. Chinese logographic group hierarchical regression predicting learning rate ..... 44

1

Abstract
The present study assessed the role of multiple behavioral and neuroanatomical measures
in the second language (L2) phonological learning in two samples of university students: native
English and native Chinese. The artificial language materials based on Korean Hangul were
learned under either alphabetic or logographic condition over eight days. Key findings:
Pseudoword reading efficiency was associated with alphabetic phonological learning in native
English speakers, which is consistent with the theory that language transfer would occur easily
between two alphabetic phonologies. MRI Voxel-Based Morphometry (VBM) revealed that the
left SMG was correlated with English group alphabetic phonology learning and the left IFG
ob
was correlated with English group logographic phonology learning while the left /right MTG and
right HG were correlated with Chinese group alphabetic phonology learning. These findings
support the dual-route model of reading and the assimilation/accommodation hypotheses.
We suggest that the aforementioned factors may serve as good predictors at least at the
beginning stage of phonology learning of a new language.
Keywords

Phonology; reading; second language acquisition; VBM; structural MRI

2

1. Introduction
1.1 Why we need to predict form-sound association learning?
In today’s world, a fundamental skill for success in life is the ability to read and absorb
new knowledge. Currently, more than half of the world’s populations speak a second language
(Grosjean & Li, 2013), and it has seemingly become more and more valuable to be proficient in
a second language. Or, as professor Frank Smith, a psycholinguist, put it, “One language sets
you in a corridor for life. Two languages open every door along the way.”
Accumulating evidence of factors affecting L2 reading comprehension has been
published over the years (Melby-Lervåg & Lervåg, 2014). However, an important step in reading
a new language is to recognize and decode print and to associate form to its sound (implicitly
and explicitly). Successful perception of foreign sound is the lowest level of linguistic perception
and is the basic skill for “upstream” language processes such as comprehension (Golestani,
2014). We are interested in how this visual form-sound association is initiated and how it is
modulated by culture, especially L1 experience, and neural anatomy, since fewer studies have
tried to integrate behavioral and neuroanatomical data to predict the efficiency in learning the
phonology in a second language. It is apparent that people vary in the ability to learn new skills
(Ventura-Campos et al., 2013). In the field of L2 phonology acquisition, if educators have
predictive information available before any learning begins, they would be able to decide
whether specific instructions are needed which is of vast impact for educational purposes. In
addition, reading difficulties appear to be strongly related to deficits in phonological processing
in alphabetic languages. Therefore, research on the form-sound learning prediction has important
theoretical and practical implications.

3

1.2 Major factors in L2 acquisition
1.2.1 Overview
In the field of L2 acquisition, multiple questions are investigated. For example, whether
L1 and L2 operate on the same brain mechanism; bilingual’s advantages and disadvantages; age
and proficiency in L2 acquisition, language transfers etc. have all been of major interest and
received enormous research over the years. Based on our specific goal to predict form-sound
association learning, we will discuss the aspects of L1 background, L1 and L2 interaction,
behavioral and neural predictors of L2 phonology learning in the following sections of
introduction.
1.2.2 L1 background
Writing systems can be classified into two categories: (1) phonography and (2)
morphography (or logography) (Koda, 1989). In the phonography writing system, such as
English and Russian, the unit is the phoneme, or the smallest unit of a word. In the
morphography writing system, such as Chinese and Japanese Kanji, the representational unit is a
morpheme and it is semantically oriented (Koda, 1989). Depending on the nature of the writing
system, different phonological and orthographic L1 processing skills are involved (Wang,
Perfetti, & Liu, 2005).
In the phonography writing system, English has been widely used and more children have
been educated in English as their L2 to make sure that they are citizens of the “global village”. In
shallow alphabetic orthographies, such as Greek, German, Finnish, Welsh, Spanish, Italian
Dutch and Serbo-Croatian, the relationship between grapheme and phoneme is mainly one to one
(Chow, McBride-Chang, & Burgess, 2005; Katz & Frost, 1992). In contrast, English is not a

4

completely transparent orthography, but rather a deep orthography with complex correspondence
between letters and sounds.
In the morphography writing system, the Chinese and Japanese Kanji scripts are
logographic in which characters correspond to syllables and morphemes rather than phonemes
(X. Chen, Xu, Nguyen, Hong, & Wang, 2010; Perfetti et al., 2007). Chinese is one of the most
popular languages in the world and the third most commonly spoken language after English and
Spanish in the U.S. While the majority of the world’s orthographic systems are alphabetic in
nature, Chinese provides the clearest contrast of what happens for the orthographies that are not
alphabetic (MacWhinney, 1995). Mandarin, one of the major dialects among Chinese
population, is the official language in the People’s Republic of China. Without further
declaration, Chinese in this paper is equivalent to Mandarin. Mandarin has some features that
distinguish it from English: (1) Visual-complexity: Chinese characters contain complex
structures based on a rectangular layout instead of a linear layout. The compositional relationship
of radicals to form characters in Chinese is fundamentally different from how letters form words
in the English language (Wang et al., 2005). In Chinese, the visual-orthographic strategy is used
to access the meaning of Chinese characters (Hanzi), which may lead to more visual processing
demand, (2) Orthographic depth: Chinese is considered a deep orthography. Instead of an
alphabet it has graphic forms that map onto morphemes, making phonological assembly not
possible for a single Chinese character (Perfetti et al., 2007). While there are radicals that can
provide some partial information about pronunciations or meanings of the character, these clues
are not reliable (Ehrich, Zhang, Mu, & Ehrich, 2013).
1.2.3 L1 and L2 interaction
Behaviorally, language transfer within the same linguistic system is commonly observed

5

(Durgunoǧlu, Nagy, & Hancin-Bhatt, 1993). For example, learning to read in two alphabetic
languages such as Spanish and English occur more naturally. Besides, learning to read a
shallower (Spanish) orthography is easier than learning to read a deep orthography (English).
When L1 and L2 differ not only in the depth of orthography but also in their orthographic
backgrounds, can we assume that differences in L2 word reading strategies are due to L1
orthographic background differences? It has been shown that native Korean speakers with
alphabetic L1 literacy backgrounds relied more on phonological information when processing
printed English words, whilst Chinese ESL learners relied more on visual-based orthographic
information in identifying English words than their Korean counterparts (Wang, Koda, &
Perfetti, 2003). Similarly, Chikamatsu (1996) examined the effects of L1 background on L2
word learning strategies. Japanese kana (Japanese alphabetic script) were given to native English
and native Chinese learners of Japanese. The results indicated that Chinese subjects relied more
on the visual information in kana words while English subjects utilized the phonological
information in kana words more than Chinese subjects. Taken together, these studies indicated
that the difference in L2 recognition strategies was due to the L1 orthographic background
difference.
There is evidence that native language determines both cognitive (Chikamatsu, 1996;
Wang et al., 2003) and neural (Fiez, 2000; Nakada, Fujii, & Kwee, 2001; Tan et al., 2003)
strategies involved in the processing of a second language. In other words, previous L1 language
experience may prepare the cortex for later L2 acquisition. Perfetti et al. (2007) proposed a
contrasting model to account for the effects of L1 on neural mechanisms of L2 learning, the
assimilation and accommodation hypotheses. In this model, when a reader acquires a new
writing system, does the brain assimilate L2 to L1 and use the L1 network to support L2? Or

6

does the brain change its network and accommodate the new features of L2? In Perfetti’s opinion,
the neural mechanisms of L2 reading engaged “some of each”, meaning both assimilation and
accommodation were involved, with assimilation as the default first step and accommodation as
the ability to adapt to changes in the new writing system by adding neural resources that support
specific requirements. Evidence supporting both hypotheses is available. Consistent with the
assimilation hypothesis, when Chinese subjects were processing English, the left middle frontal
gyrus (MFG, usually involved in Chinese reading) was activated instead of superior temporal
gyrus (STG, usually activated in English reading). The Chinese subjects were applying a strategy
of processing Chinese to English processing (Tan et al., 2003). Similar results have been
reported by Nakada et al. (2001) on Japanese and English native speakers. On the other hand, in
English speakers who are learning Chinese, the bilateral visual word form areas and left MFG
were activated when processing Chinese, which supported the accommodation hypothesis
(Abrahamsson & Hyltenstam, 2009).
To summarize, existing research has suggested the possibility that native language may
shape the cognitive and neural mechanisms involved in the acquisition of a second language.
1.2.4 Behavioral predictors
At the most general level, research has focused on the determinants of success in the
acquisition of a second language since the early 1950s (Randhawa & Korpan, 1973). In one
study, sex was found to play a key role in L2 learning in that females performed better than
males in both syntax and semantics L2 tasks (Andreou, Vlachos, & Andreou, 2005). A more
recent study reported that general statistical-learning ability is an important factor as it may
predict the speed and success of L2 learning (Frost, Siegelman, Narkiss, & Afek, 2013).
Research also suggests that musical intelligence is a determinant of success in foreign language

7

phonetic performance (Zybert & Stępień, 2009). In addition to the aforementioned factors, an
attempt was also made to provide a framework in which predictors of L2 acquisition in general
can be divided roughly into two groups: sociolinguistic factors and cognitive-linguistic factors
(Jia, Gottardo, Koh, Chen, & Pasquarella, 2014). Under the umbrella of sociolinguistic factors,
there are acculturation (Jia et al., 2014), motivation for L2 learning (Manolopoulou-Sergi, 2004),
and level of L2 anxiety (MacIntyre, Baker, Clément, & Donovan, 2002). Cognitive-linguistic
factors include elements such as metacognition (Pishghadam & Khajavy, 2013), L2 aptitude (R.
Sparks & Ganschow, 2001), L1 phonological awareness (PA) (Lindsey, Manis, & Bailey, 2003;
Yeung & Chan, 2013), L1 word/pseudoword reading efficiency (Verhoeven, 1990), and
phonological working memory (Meschyan & Hernandez, 2002). Overall, L2 aptitude and L1
linguistic variables have been found to be strong predictors of L2 proficiency and achievement
(Skehan, 2002).
Among the above cognitive-linguistic factors, phonological awareness (PA, a broader
term which includes phonemic awareness) has been widely studied and has been established as a
strong predictor in L2 phonology acquisition. For instance, it is well documented that L1
phonological awareness played an important role in predicting L2 reading performance and it
was transferable across languages, such as two alphabetic orthographies with orthographies
varying in depth (Chiappe & Siegel, 1999; Cisero & Royer, 1995; Durgunoǧlu et al., 1993; Sun-
Alperin & Wang, 2011). Furthermore, it was transferable between one non-alphabetic L1
(Chinese) and one alphabetic L2 (English)(X. Chen et al., 2010; Chow et al., 2005; Gottardo,
Yan, Siegel, & Wade-Woolley, 2001; Yeung & Chan, 2013)
However, a reciprocal factor of PA, the role of word and pseudoword reading skills in L2
phonological learning remains largely unknown. Because of the highly correlated relationship

8

between PA and reading skills, PA may offer little unique information of reading once a current
reading measure is available (Hogan, Catts, & Little, 2005). Word reading is the ability to
translate printed words into a speech code and is usually evaluated by the accuracy and speed of
reading aloud (Melby-Lervåg, Lyster, & Hulme, 2012). Pseudoword reading is a sub-word level
measure of phonological decoding ability. There is relatively limited literature on whether
word/pseudoword reading skills is transferable. In a study of 37 bilingual Portuguese-Canadian
children aged 9–12 by Da Fontoura and Siegel (1995), English was the main instructional
language at school while Portuguese was the language spoken at home. They found significant
correlations among word and pseudoword reading accuracy, working memory, and syntactic
awareness between Portuguese (L1) and English (L2). Similarly, Lindsey et al. (2003) found
correlations between Spanish (L1) word reading and English (L2) word reading scores, which
suggests that cross-language transfer may happen in word decoding (besides the transfer of PA
reported in this paper). Geva and Siegel (2000) found moderately high positive correlations
between L1 (English) and L2 (Hebrew) word recognition and word attack (pseudoword reading)
measures. Thus, not only did word/pseudoword reading accuracy transfer between two
alphabetic languages, but the bidirectional feature of the transfer was also revealed (deep
orthography ⇄ shallow orthography). Yet, it seems that more research reflected a positive
transfer from the more predictable grapheme–phoneme conversation rules of alphabetic
languages (e.g. Portuguese and Spanish) to the very opaque orthography alphabetic language
(e.g. English).
Presumably, if bilinguals were given more time to read in their L2, they would be able to
use orthographic information more accurately that may lead to better reading performance.
(Geva, Wade-Woolley, & Shany, 1997). This entails the assumption that it is not enough to

9

focus on L2 word/psueoword reading accuracy alone, but also efficiency. Verhoeven (1990) used
a word reading efficiency measure in which the average number of words read correctly per
minute was calculated and it was found that word reading efficiency is predictive of Turkish
Children learning Dutch. Similarly, August, Calderon, and Carlo (2001) reported that Spanish
phonemic awareness, Spanish letter identification, and Spanish word and pseudoword reading
efficiency of 2
nd
graders were reliable predictors of performance on parallel tasks in English at
the end of 3
rd
and 4
th
grades in Spanish-English bilingual children, and hence, transferable from
Spanish (L1) to English (L2). Geva et al. (1997) presented a study in which word reading
accuracy and speed were highly correlated in a group of 66 children learning to read English
(L1) and Hebrew (L2). In summary, as is the case in word/pseudoword reading accuracy,
evidence exists that there is a two-way transfer in word/pseudoword reading efficiency in
alphabetic languages.
The question now comes to whether there is cross-language transfer of word/pseudoword
reading skills between alphabetic languages and non-alphabetic (e.g., logographic) languages.
Gottardo et al. (2001) administered parallel word reading measures of Cantonese (L1) and
English (L2) in 65 native Cantonese-speaking children but failed to find correlations between L1
and L2 word reading tasks. Conversely, in a study conducted on 53 Hong Kong elementary
school students, researchers found that Chinese (L1) word reading test scores were significantly
correlated (r=0.58, p<0.001) with English (L2) word reading (Keung & Ho, 2009). Similarly, in
Baum and Titone (2014)’s study, Chinese (L1) character reading score was correlated (r=0.27,
p<0.05) with English (L2) pseudoword reading. The findings of word/pseudoword reading skill
transfer between alphabetic and logographic language are mixed.

10

To the best of our knowledge, no study has examined cross-language transfer of word
reading efficiency between languages that come from entirely different orthographic systems
(e.g., English, Korean Hangul, Chinese). Our study design (please see details in 1.3) will not
only include both word reading accuracy and efficiency measures, but also test the effects
between different orthographic systems that will broaden the current view on this topic.
In addition to PA and reading skills, researchers have studied the memory component in
L2 phonological learning. Gathercole and Baddeley (1990) provided convincing evidence that
phonological short-term memory is a salient contributor to the learning of novel words. They
divided 118 5~6 years-old children into two groups based on their scores in a non-word
repetition task and found that the higher phonological memory group learned the unfamiliar
names of toys (e.g., Piemas) more rapidly and retained them longer. Likewise, Meschyan and
Hernandez (2002) reported that non-word repetition skill in English was predictive of second
language decoding, vocabulary, grammar, and reading comprehension skills in college-age
adults. In a study with Finnish children, Service (1992) found that pseudoword repetition task
scores could predict English (L2) learning, suggesting that the ability to represent unfamiliar
phonological information in working memory underlies the acquisition of new items in a foreign
language (Service, 1992). In a longitudinal study with 160 Finnish-speaking elementary school
children, Dufva and Voeten reported that both word recognition and phonological memory tests
had a significant role in learning English as a foreign language (Dufva & Voeten, 1999). These
studies supported the hypothesis that better phonological short-term memory could lead to better
foreign language phonological processing.
To summarize, these findings lend support to the selection of our predictors in L2 learning
in the English group: word/pseudoword reading accuracy and efficiency tests, and phonological

11

short-term memory tests. Similarly, in our Chinese group, we had word reading accuracy and
efficiency tests, and phonological working memory tests in their native language (Chinese).
Additionally, we added word/pseudoword reading accuracy and efficiency tests in their L2
(English) to study cross-language transfer between languages. In order to read a Chinese pseudo-
character correctly, the reader needs to know the sound cues provided by the phonetic
component of a Chinese character. Since this ability is unlikely transferrable to our training
materials based on Hangul language, we didn’t include Chinese pseudo-character reading
accuracy and efficiency tests.
1.2.5 Neuroanatomical predictors
The brain is the source of behaviors and can change with short experience, contrary to
theories that the brain plasticity may only appear after longitudinal exposure (days to years), Sagi
et al. (2012) provided the first evidence that rapid structural plasticity of white matter can be
observed by DTI after only two hours of training in human. However, researchers also caution
that individual variation in brain anatomy should not be ascribed only to environmental/training
conditions, but also to pre-existing neuroanatomical differences (Zatorre, Fields, & Johansen-
Berg, 2012). Evidence from the twin study paradigm showed that genetic factors significantly
influenced cortical structure in Broca’s and Wernicke’s language areas, as well as frontal brain
regions (Thompson et al., 2001). Further evidence of pre-existing brain anatomical differences
predicting individual performances on tasks are enhanced by non-invasive neuroimaging
procedures, such as Magnetic Resonance Imaging (MRI). Here, we would like to investigate,
anatomically, whether there are any “L2 talent” areas that may exist prior to any training to assist
us in the prediction of L2 phonological learning success.
We followed two research lines to explore which neuroanatomical predictors of L2

12

phonology learning may appear in the current study. First, the Dual Route Cascaded (DRC)
model (Coltheart, Rastle, Perry, Langdon, & Ziegler, 2001) is a model of visual word recognition
and reading aloud. This model is composed of two distinct routes of phonological access. The
first route refers to alphabetic phonology (sublexical/indirect route) in which visual words are
transformed into phonology through grapheme-to-phoneme correspondences (GPC). The GPC
rule means that we link a letter or a letter sequence directly to a single phoneme, such as “i" in
“file”. Alphabetical languages typically rely on this type of phonology with exceptions. It is also
believed that English low-frequency words and pseudo-words are accessed through the route of
alphabetic phonology. The second route refers to logographic phonology (lexical/direct route),
which is used for English high-frequency words and irregular words. In logographic languages
such as Chinese, phonological access typically relies on the direct route as well. Another model,
the connectionist model (Harm & Seidenberg, 2004; Seidenberg, 2005) is a statistical model
based on brain-based differences. In this study, our goal was to study L2 phonology learning
across highly contrastive scripts. Since a universal model of L2 reading which can accommodate
readers from both alphabetic and logographic L1 backgrounds is not available (Ehrich et al.,
2013), we focused on the DRC model of reading. In order to evaluate the dual route theory from
the perspective of brain mapping, Jobard, Crivello, and Tzourio-Mazoyer (2003) conducted a
meta-analysis of 35 neuroimaging studies and found that alphabetic phonological processing
relies on the dorsal route (in DRC model) including the superior temporal gyrus (STG), the
middle temporal gyrus (MTG), the supramarginal gyrus (SMG), and the pars opercularis of the
inferior frontal gyrus (IFG
op
). The logographic phonological processing relies on the ventral
route including the left pars triangularis (IFG
tri
), inferior temporal gyrus (ITG), and the posterior
part of the MTG. The integration of cognitive and neural accounts of the dual route model has

13

proved to be extremely difficult because the construct of the model was not intended to yield
claims about the neural fMRI activity and structural differences (Taylor, Rastle, & Davis, 2013).
Despite this difficulty, evidence is mounting to support the application of dual route reading
model from the neural viewpoint. For example, using fMRI, Thuy et al. (2004) found that Kana
(phonogram) required phonological recoding via dorsal route while Kanji (morphogram) reading
involved more visual orthographic retrieval through the ventral route. In a different study,
Chinese Pinying (known as Chinese alphabet) and characters were used as the contrasting
material of phonograms and morphograms, and researchers reported that reading pinyin led to
greater activations in the bilateral inferior parietal lobule (IPL), the precuneus, the STG and the
left anterior MTG. On the other hand, activation in the bilateral superior frontal gyrus (SFG),
bilateral cuneus, left fusiform gyrus, the left posterior MTG, the right IFG were greater for non-
alphabetic Chinese character reading (Babcock, Stowe, Maloof, Brovetto, & Ullman, 2012).
Moreover, in one fMRI study from our research group, we reported that, in forty-three native
English speakers there was stronger activation in the left SMG (extending to superior occipital
gyrus) and the left precentral gyrus (extending to IFG) for alphabetic phonology, but greater
involvement of the right orbital frontal cortex, right angular gyrus (AG), right MTG, anterior
cingulate cortex, posterior cingulate cortex for logographic phonology (Mei et al., 2014).
Second, to refine our prior hypothesis of possible good neuroanatomical predictors of L2
phonology acquisition, we tracked down one recently published review on neuroanatomical
predictors. Compared to fMRI, relatively fewer structural MRI studies have been conducted to
examine the individual differences in brain anatomical configuration related to L2 phonological
acquisition. Li, Legault, and Litcofsky (2014) provided a good review of studies on anatomical
predictors of L2 learning. Several brain regions were associated with L2 phonological learning

14

including Heschl’s gyrus (HG) and left parietal regions. HG was associated with successful
phonetic learning (Golestani, Molko, Dehaene, LeBihan, & Pallier, 2007). On the basis of the
previous studies, native English-speaking subjects learned to associate picture patterns with
pseudoword identification and more successful learners had larger left (not the right) HG volume
(Wong et al., 2008). Sutherland et al. (2012) hypothesized that an individual's frequency-
modulation threshold (the threshold to detect a frequency modulation of auditory stimuli) would
correlate with gray-matter density in left HG, and they collected anatomical MRI data from
participants who were tested and scanned at three time points: at 10, 11.5 and 13 years of age.
Using voxel-based morphometry, they found that frequency modulation threshold was
significantly correlated with gray-matter density in left HG at the age of 10 years, but this
correlation weakened with age increase.
Golestani and her colleagues presented a foreign speech sound learning study. Twenty-
one native French speakers were asked to repeat Farsi voiced /q/ in the context of 6 different
consonant-vowel syllables (sound /q/ followed by –a, -o,-e, -i,-u,-A) and in the context of 6
different bisyllabic nonwords (sound /q/ followed by -azA, -orme, -ese, -ise,-ulum, -Ali). VBM
analysis revealed individuals who could more accurately pronounce the foreign sound had higher
WM density in the bilateral inferior parietal regions as well as left insula and left prefrontal
cortex (Golestani & Pallier, 2007). In the same vein, using Hindi stimuli, Golestani, Paus, and
Zatorre (2002) reported faster phonetic learners had more white matter in parietal regions,
especially in the left hemisphere, and associated this greater asymmetry in the WM as greater
myelination allowing more efficient neural processing of non-native speech sounds.
Taken together, the aforementioned findings highlighted the importance of left HG and
left parietal region in L2 phonological acquisition and the direction was the larger GM/WM

15

(density and volume), the more successfully participants learn the non-native phoneme.
Nevertheless, it is noteworthy that Golestani’s studies focused on alphabetic phonological
learning while Wong’s study could be considered as partial logographic phonological learning.
Little is known about whether anatomical predictors exist uniquely for alphabetic and
logographic phonological learning in English and Chinese participants. It is imperative to extend
the previous work to examine individual differences in learning form-sound associations. It is
especially valuable to study individual differences under different training conditions and across
people speaking different native languages.
In the present study, we are interested in the brain structural predispositions associated
with alphabetic and logographic phonology learning. Our approach will give us the power to
examine how different language experience would modulate a new phonological learning
process.
1.3 The rationale of using Hangul artificial form-sound association learning paradigm and
a 2 × 2 design
We developed a training paradigm to train native English and native Chinese speakers to
learn an artificial Korean based language. Both groups were further divided into two groups to be
presented with different learning conditions. One is to learn this artificial language via GPC rule,
namely, to teach participants individual letter sound first and then to combine the letter sound
and form the pronunciation of the entire word (alphabetic learning condition). Another condition
is to have them learn the pronunciation of the word directly by memorizing them (logographic
learning condition). This setting lead us to a 2 (L1 status: English, Chinese)× 2 (learning
condition: alphabetic and logographic) design in which English-speaking participants learned the
targeted language in alphabetic and logographic condition, with the former condition closer to

16

their own language learning habit. Similarly, when Chinese group participants studied the
Korean based targeted language from the logographic condition, presumably that would be less
alien to them. The artificial Korean Hangul language system has English-like sub-lexical GPC
rules and Chinese-like logographic visual features. It is valuable and interesting to have a Hangul
task in that: (1) Vast research has been focused on how people learn English as a second
language, but English native speakers learning another language has been relatively less
examined. Especially, how English-speaking students would respond to both alphabetic and
logographic language characteristics is yet to be documented; (2) The participants in alphabetic
learning condition in the current study are learning to read in a super shallow orthography with
one to one grapheme/phoneme correspondence between the symbol and its sound while English
group participants read in a deep alphabetic orthography and Chinese group participants read in a
non-alphabetic orthography. While semantic information was excluded for the study, using a
language that can be learned to read either from alphabetic or from logographic phonology with
the same visual forms made it possible for us to study whether different or same prediction
mechanisms would apply in participants whose native languages vary in orthography
transparency. Another related issue would be whether L2 (English) decoding ability would
predict L3 (Hangul) reading skills in our Chinese group since Chinese college students were
exposed to English learning as their L2 from first grade; (3) Comparing to traditional foreign
language learning prediction studies, the merit of our study is that we can reduce the effects of
sociolinguistic factors such as motivation and anxiety since all experiments were finished in lab
settings. As was mentioned by Hirshorn and Fiez (2014) the artificial linguistic learning setting
allows us to minimize the problems of covariance while controlling the language experience of
natural language learners. In addition, the predictive anatomical differences in English-speaking

17

Hangul learners and Chinese-speaking Hangul learners might relate to different ortho-semantic
connections as well as different ortho-phonetic connections. We can isolate and study ortho-
phonetic processing in reading within our study paradigm since semantic information was not
provided in the artificial language. (4) The longitudinal design is another merit of this study since
we can obtain the rate of learning growth (λ), and further understand whether variations in pre-
selected variables can predict the differences in growth rates.
In summary, the design of the current study would be able to further our understanding of
phonological information processing and explore possible causal influences on L2 acquisition.
1.4 Hypotheses
A cross-sectional design can allow us to answer the following questions (1) How are
participants’ L1 associated with the cognitive and neural substrates involved in form-sound
association learning of a new language? (2) What are the behavioral and neuroanatomical
predictors in the acquisition of a new language? We further hypothesized that: (1) In native
English speakers, English word reading skills (accuracy and efficiency) would predict Hangul
alphabetic learning in line with past research if both L1 and L2 were all in alphabetic
orthography. (2) In native Chinese speakers, their Chinese word reading ability would predict
Hangul logographic learning based on the past research when L1 and L2 are of the same
orthographic system. (3) We hypothesized that visual-auditory memory and phonological
working memory measures may predict alphabetic and logographic Hangul learning outcome in
both English and Chinese groups since the early phonological information registration should
operate regardless of orthography system. (4) Consistent with the dual route model, alphabetic
phonology processing should be correlated with grey matter volume in the left STG, left MTG,
left SMG, and left IFG
op
, while logographic phonology processing relies on left ITG, left IFG
tri

18

and the posterior part of the left MTG. If assimilation was involved, the correlations of
neuroanatomical measures and learning growth rates in English-speakers learning alphabetic
Hangul would reflect the recruitment of L1’s phonology related regions (left STG, left MTG, left
SMG, and left IFG
op
). In the same vein, if accommodation was involved as well, native English-
speakers who were under logographic learning condition should demonstrate the inclination to
recruit logographic phonology processing regions such as left ITG, left IFG
tri
and the posterior
part of the left MTG. In contrast, Chinese-speakers should rely on logographic phonology
processing regions under logographic Hangul learning condition and recruit alphabetic
phonology processing regions when under alphabetic Hangul learning condition. Within the
aforementioned regions, the most likely correlations of structural and learning measures that
would be unveiled should exist in the HG (HG is a part in STG) or left parietal regions (such as
left SMG) since they were repeatedly reported.

19

2. Materials and Methods
2.1 Participants
One hundred and one native English-speaking students from the University of Southern
California and the University of California, Irvine (42 males and 59 females; mean
age=20.9±2.5; range=18~30) and forty-three native Chinese-speaking students from Beijing
Normal University (22 males and 21 females; mean age=22.1±1.8; range=19~25) were recruited
for the study. Both English and Chinese groups were divided into two subgroups to receive either
alphabetic or logographic phonological training (English group: 60 in alphabetic training and 41
in logographic training; Chinese group: 22 in alphabetic training and 21 in logographic training).
All participants had no Korean language experience before the study and were right-handed
according to the Edinburgh Inventory (Babcock et al., 2012). Participants in the study had
normal or corrected-to-normal vision and no history of head injury, psychiatric or neurological
disorders. Informed consent was obtained and procedures were approved by the IRBs in all sites.
2.2 Behavioral evaluation
There were eight potential predictors in the English group and ten predictors in the
Chinese groups. These cognitive-linguistic predictors included Raven score, visual-auditory
learning, reading accuracy, reading efficiency and phonological working memory measures. All
variables were administered individually with instructions by trained experimenters: (a) Raven’s
Advanced Progressive Matrices (Raven); (b) Woodcock Reading Mastery Tests-Revised
(WRMT-R): Visual-Auditory Learning (VAL), Word Identification (WI) and Word Attack
(WA); (c) Test of Word Reading Efficiency (TOWRE): Sight Word Efficiency (SWE) and the
Phonetic Decoding Efficiency (PDE); (d) Comprehensive Test of Phonological Processing
(CTOPP): Memory of Digits (MD) and Non-word Repetition (NR). WRMT-R and TOWRE was
administered and responded to in English in both groups. In this manner, reading accuracy

20

(WRMT-R) and efficiency (TOWRE) in English were measures of Chinese participants’ L2
ability while they represented English group’s L1 skills. CTOPP was tested in their native
languages (English group in English; Chinese group in Mandarin).
In addition, the Chinese group had two supplementary tests, which are Chinese Word
Identification (C-WI) and Chinese Sight Word Efficiency (C-SWE). These two variables were
the measures of Chinese group’s L1 capacity.
Other than word reading accuracy (WRMT-R), efficiency (TOWRE) and phonological
short-term memory (CTOPP) tasks, we also collected age, gender and Raven score from all
participants for these variables to be controlled in future analyses.
2.2.1 Raven’s Advanced Progressive Matrices
Raven’s Advanced Progressive Matrices (Raven, 1990), a computer-based test, was
adopted to estimate general cognitive ability. This test contained thirty-six test items and was
scored automatically 40 minutes after the test. Twelve additional items were introduced for
practice before the main test. For each item, participants were asked to identify the missing
element from eight possible alternatives, the one that best completed a pattern. Each correctly
answered test item would earn one point. The maximum score was 36.
2.2.2 Woodcock Reading Mastery Tests-Revised
Woodcock Reading Mastery Tests-Revised (WRMT-R) (Woodcock, 1987) was
administered to evaluate visual-auditory learning and word decoding abilities. We selected three
subtests from form G: Visual-Auditory Learning (VAL), Word Identification (WI) and Word
Attack (WA). The Visual-Auditory Learning test has 26 symbols and their names, as well as 2
symbols for word endings (-ing, -s). It tests the participant’s ability to associate unfamiliar visual
stimuli with familiar oral words and then to translate symbols to form a sentence. In this task, the

21

experimenters first taught the symbols and their meaning to participants. Then the page was
turned and the participants only saw the newly learned symbols and were supposed to read these
symbols in a sentence. More and more symbols and their corresponding words were given
during the test and the sentence test previously learned and newly learned items each time after
new symbols were learned. This task had a large component of visual-auditory short-term
memory in it. Since the Visual-Auditory learning (VAL) test directly assesses the ability to form
associations between visual stimuli and oral responses, which is analogous to the form-sound
association learning, we also included VAL in the present study.
The WI test requires participants to read 106 real words that are increasingly difficult to
read (e.g. is, you… zeitgeist). The raw score from this test is the number of words read correctly.
The WI test had an internal consistency reliability of 0.92 (Woodcock, 1998).
The WA test has 45 pseudowords that are again increasingly difficult (e.g. dee,
ap…pnomocher) and participants were instructed to use their phonic analysis/decoding skills to
read them. The raw score from this test is the number of pseudowords read correctly. The WA
test had an internal consistency reliability of 0.91 (Woodcock, 1998).
2.2.3 Test of Word Reading Efficiency
Test of Word Reading Efficiency (TOWRE) (Torgesen, 1999) measures individual’s
ability to pronounce printed words and non-words accurately and fluently. It includes two
subtests: the Sight Word Efficiency (SWE) test and the Phonetic Decoding Efficiency (PDE)
test. Each test has two alternate forms (Form A and Form B). Students were asked to read as
many of the words/non-words on the list as they could in 45 seconds. SWE is a measure of
accuracy and fluency in reading phonetically regular and irregular words in a list of 104 words
(e.g. up, cat…transient). The total number of words read accurately indicates their sight word

22

reading efficiency. We added the scores of both Form A and Form B in SWE and divided by 2 to
generate a single measure for SWE. PDE was used to assess students’ phonemic decoding
efficiency and it has 63 pronounceable non-words (e.g. ip, ga…emulbatate) to be sounded out.
Scoring of PDE test followed the same rules as SWE. Note that above WI and WA didn’t
measure reading efficiency while SWE and PDE did.
2.2.4 Comprehensive Test of Phonological Processing

Comprehensive Test of Phonological Processing (CTOPP), developed by Wagner,
Torgesen, and Rashotte (1999), is a widely used screener of reading problems for children,
adolescents and young adults. CTOPP tests whether a participant has weaknesses in the
phonological processing abilities in three domains: phonological awareness (Elision, Blending
Words and Sound Matching), rapid naming [Rapid Object Naming (RON) and Rapid Color
Naming (RCN)] and phonological working memory [Memory of digits (MD) and Non-word
Repetition (NR)]. We chose MD and NR out of a total of seven subtests.
We used a recorder to play standard sounds of numbers and non-words in phonological
working memory tests. A total of 27 lines of digits were played for MD test (the number of digits
increases from each line: e.g. 16, 72, 521… 3759628146). Participants were asked to repeat the
numbers in the same order after they hear them and one point was earned with each trial repeated
without error. The test was discontinued if examinee missed 3 trials in a row. Similarly, a total of
18 lines of non-words (e.g. jup, zid…, shaburiehuvoimush) were played for the non-word
repetition task with the same scoring rules.
The Chinese version of phonological working memory tests were adopted from the
English version of CTOPP, but were conducted in Mandarin.

23

2.2.5 Chinese group additional tests
The Chinese group had two additional tests. In the Chinese Word Identification test (C-
WI), participants were asked to read 40 very low-frequency Chinese characters and the
performance in this test was scored by adding the total number of characters that were correctly
read (Liu, Shu, & Li, 2007). The Chinese Sight Word Efficiency test (C-SWE) was modified
from its English counterpart and 104 items were selected from Chinese character
psycholinguistic norms (Liu et al., 2007). The total number read correctly within 45 seconds
indicates their sight word reading efficiency in Chinese.
2.3 Artificial language, training procedure and outcome measures
2.3.1. Learning materials
We created an artificial language by adopting the visual forms and phonologies from
Korean Hangul characters. There are 24 letters in Hangul, in which 14 are consonants and 10 are
vowels. Hangul letters are grouped into blocks to form characters, with each letter having its own
sound. Assembling letter sounds together forms the pronunciation of the whole character. For
instance, 헙 is composed of three distinct letters: ㅎ [h], ㅓ [eo] and ㅂ [b].
We chose 22 Hangul letters (12 consonants and 10 vowels) to form 90 characters that are
easy to pronounce by both native English and Chinese speakers. Within the 90 characters, we
used 60 characters for training. The remaining 30 were used for examination purposes. The
standard sounds of the letters and characters were recorded from a native Korean female speaker
and were normalized to the same length (600ms) and loudness.
The artificial language could be learned through either alphabetic or logographic form-
sound associations (see Fig.1 for the Hangul character examples). In the alphabetic group, each
word contained at least one consonant and one vowel. The average usage of each letter (either a

24

consonant or a vowel) in the training materials was 8 times (range: 5-17 times). Participants
randomly assigned to the alphabetic group were first taught the pronunciation of the letters and
then to assemble the phonology of the characters from letters. The Hangul alphabetic learning
condition could be described as “super-shallow” orthography in which there was a direct “one to
one” correspondence between letters and their sounds. The reading of the alphabetic group was
similar to Korean language reading in the real world with their original sounds. In contrast, in the
logographic group, subjects were asked to memorize the arbitrary association between each word
and its pronunciation. To prevent the subjects from implicitly learning the GPC rule in the
logographic phonology condition, the visual forms were not paired with their original
pronunciations. Instead, we shuffled the same set of sounds and visual forms so that it would be
impossible for the “logographic phonology” group to learn and read the characters by assembling
letter sounds together.
A key component of reading is to associate visual words with their sounds. Due to our
specific focus on predictors of learning form-sound association, we employed Hangul because it
has both alphabetic orthography and logographic visual features (C. Chen, Chen, Xue, Mei, &
Dong, 2009). While semantic information was excluded for the study, participants could learn to
read either from alphabetic or from logographic phonology with the same visual forms. Since
English is based on assembling letters to form sound, and Chinese is built on assigning a certain
sound to a character, using a language that can be learned both ways made it possible for us to
not only study the variables that can predict form-sound learning success in both English and
Chinese native speaking groups, but also to study the cross-linguistic transfers of phonological
skills.

25

Figure 1. Training materials: 60 Hangul characters
2.3.2 Training procedure
We trained participants in eight-day learning sessions with one session (roughly one
hour) per day. Participants learned to read 60 Hangul characters on a computer program. In the
alphabetic learning condition, participants were provided with instructions on the pronunciations
of the vowels and consonants and the GPC rules. In the logographic learning condition
participants were asked to memorize pronunciations of the whole characters.
We had different combinations of tasks in one learning session to facilitate efficient
learning. There were Letter Sound (memorize the sound of letters in alphabetic condition), Word
Sound (memorize the sound of whole word in logographic condition), Massed Learning (all 60
characters on display to compare and learn), Free Recall (name characters shown on the screen at
one’s own speed), Fast Naming (name characters within limited time span), Discrimination
(select the correct pronunciation of a character from four potential pronunciations in 3 seconds
and the character would disappear after 3 seconds), Naming (test whether participant can name
the characters correctly). Letter Sound, Word Sound and Massed Learning were for participants
to carefully listen, imitate, and compare their own pronunciations with the standards. Free
Recall, Fast Naming and Discrimination tasks are aimed to increase the automaticity of
connections between the sounds and their visual forms. The Naming task was at the end of each

26

day’s training and participants’ naming was recorded. We added 30 more new (never-learned)
words in the Naming task for the alphabetic training groups in order to test the ability of the
participants to apply the GPC rules instead of memorizing the pronunciations. Both learned and
new items had to be named within 4 seconds in the first four days of training and 3 seconds in
last four days of training. The characters would disappear from the screen after the allowed time
span and would be counted as wrong pronunciations.
2.3.3 Phonological learning outcome measures
We determined the correct ratio (CR) by dividing the number of correct answers by the
number of items that were presented during either discrimination or naming task. Thus, we
obtained 8-time-point CR results of discrimination and naming tasks at the end of the training. In
addition, participants in alphabetic groups also had CR of “naming new item” task, which
directly reflected the participants’ ability to apply the GPC rules to form sound for 30 never-
learned Hangul characters. Since the discrimination and pronunciations of items were generated
under time pressure, we assumed that the measures of CR had both reading accuracy and
efficiency component in it.
An exponential function was used to fit our non-linear learning curve which was
comprised of 8-time-point CR: y = a + b (1 – exp (-λ* t)), where the shape of the curve is
determined by a single parameter λ. This function was derived from the work of Lu, Williamson,
and Kaufman (1992) in which λ=1/tau. We used λ as the learning growth rate where larger λ
indicated faster learning.
Taken together, λ of discrimination (λ
D
) and naming task (λ
N
) was used in logographic
groups while λ of discrimination task (λ
D
), naming task (λ
N
) and naming new item task (λ
NN
) was
used in alphabetic groups (see subject # 1 in the group as examples of curve fits in figure 2 and

27

3). In figure 2 and 3, the x axis represented day 1 to day 8 and the y axis represented the CR. The
first day’s CR was used as a variable of no interest in the regression model of λ. Thus, we finally
provided the learning growth rate as dependent variable of our artificial language phonological
learning.

Figure 2. λ fits in English-speaking group

Figure 3. λ fits in Chinese-speaking group

English alphabetic group Naming new item task subject #1 English alphabetic group Naming task subject #1
English alphabetic group Discrimination task subject #1
English logographic group Discrimination task subject #1 English logographic group Naming task subject #1
Chinese alphabetic group Discrimination task subject #1
Chinese alphabetic group Naming task subject #1
Chinese alphabetic group Naming new item task subject #1
Chinese logographic group Discrimination task subject #1 Chinese logographic group Naming task subject #1

28

2.4 Image acquisition
Image data was collected from two scanners of the same type: 3.0 T Siemens Magnetom
Trio MRI scanners, located in the Dana and David Dornsife Cognitive Neuroscience Imaging
Center at the University of Southern California and in the Brain Imaging Center at Beijing
Normal University. We acquired 3D, T1-weighted images using the Magnetization Prepared
Rapid Gradient Echo (MP-RAGE) sequence on both sites. The MR imaging protocol was the
same for both scanners: TR 2530 ms, TE 3.09 ms, TI 800 ms, flip angle 10°, field of view 256 ×
256 mm, slice thickness 1 mm, voxel dimensions: 1.0x1.0x1.0 mm, acquisition time 10 m 48 s.
2.5 Image data processing: voxel-based morphometry (VBM)
Structural images were first analyzed with the Statistical Parametric Mapping (SPM 12)
statistical package (Welcome Department of Cognitive Neurology, Institute of Neurology,
London, England). Brains were tissue-segmented, non-linearly registered to the Montreal
Neurological Institute (MNI) space and "modulated" to correct for local expansion (or
contraction). The modulated grey matter images were then smoothed with an isotropic Gaussian
kernel with a sigma of 8 mm. Intracranial volumes were calculated by adding up grey matter,
white matter and CSF volumes together and were controlled in the analyses. The ultimate
measure of this procedure was grey matter volume and all results are reported in MNI space.
2.6 Statistical Analysis
2.6.1 Behavioral predictors
Statistical analyses of behavioral measures were performed using the SPSS software
package (SPSS version 21). The detailed data analyses steps are described below. We had two
groups, English-speakers and Chinese speakers, with each group having two sub-groups based
on the learning conditions to which they were assigned (alphabetic and logographic).

29

First, to gain a broader understanding of group differences in learning results, we
performed a group×learning condition (2×2) factorial analysis on learning rates. Second,
Pearson’s correlations between all independent and dependent variables were examined. We
reported all significant correlational results. Finally, results of significant correlations would then
be entered in the regression model to predict learning outcomes (λ).
2.6.2 Neural predictors
A general linear model was used to examine the associations between whole- brain grey
matter volumes and learning outcome variables. The learning growth rates of the discrimination
task, naming task and naming new item task were entered into the GLM model one by one with
age, Raven score, the first day’s performance and total intracranial volume as covariates of no
interest. We could better understand the learning growth if the initiated performance (the first
day’s performance) was controlled. Because larger brains are likely to have larger brain
structures, so in VBM it is useful to take in to account that how the regional volumes are likely to
vary as a function of whole brain volume. Thus, we excluded the impact of variables of no
interest on grey matter volume. We also applied a cerebellum excluded grey matter mask to limit
the regions we tested. Statistical significance threshold for the whole-brain analyses was
determined by Monte Carlo simulations using 3dClustSim in AFNI (Cox, 1996). All reported
clusters were significant at p=0.005 (Bonferroni correction: 0.05/10 analyses, 10 analyses
included λ
D
, λ
N
of English group and Chinese group in logographic learning condition, and λ
D
,
λ
N
together with λ
NN
of English group and Chinese group in alphabetic learning condition)
corrected for multiple comparisons, and were in MNI space. The results were visualized using
xjView toolbox (http://www.alivelearn.net/xjview).

30

Volumes of significant clusters were extracted and subsequently used to implement
hierarchical regression analyses to provide a more integrated and refined picture of how
behavioral and neural predictors can be combined to predict form-sound association learning (see
2.6.3).
2.6.3 Hierarchical regression analysis
Hierarchical regression analyses were conducted to evaluate predictors of the learning
rates of the discrimination task, the naming task and the naming new item task. Age and Raven
score were entered in the model to control for their impact at the first step. Next, cognitive-
linguistic predictors were entered to study their individual contribution to the dependent
variables. Lastly, anatomical predictors were entered into the hierarchical regression models.

31

3. Results
We aimed to identify the behavioral and neural anatomical factors that can be used to
predict the growth rate of phonological learning under both alphabetic and logographic
conditions in English and Chinese groups. We were interested in whether the variances of
learning success in English and Chinese participants come from different predictor variables.
Namely, which predictors in the English group would contribute to alphabetic learning and
which predictors in the Chinese group would contribute to alphabetic learning? Similarly, which
predictors in the English group would contribute to the logographic learning and which
predictors in the Chinese group would contribute to the logographic learning? Moreover, would
predictors vary by tasks (discrimination task, naming task and naming new item task)?
We found that reading accuracy and efficiency of pseudoword predicted the growth rate
of the discrimination task in the English group’s alphabetic phonology learning. Also, English
word reading accuracy predicted the growth rate of discrimination task in the Chinese group’s
logographic phonology learning. In addition, we found that memory measures in both groups
contributed to the growth rate of learning. The grey matter volumes of a few classic language
regions were associated with alphabetic and logographic phonology learning, including the
SMG, MTG, HG, IPL and IFG.
3.1 Behavioral predictors
3.1.1 Descriptive statistics
Table 1 shows an overview of age, Raven score and behavioral measures of interest in
both English and Chinese groups. As shown, English group participants had higher scores in
English reading tasks.

32

Table 1. Descriptive statistics for measures

English group

Chinese group
Alphabetic Logographic Alphabetic Logographic
Mean (SD) Mean (SD) Mean (SD) Mean (SD)
Age
21.4(2.8) 20.1(1.6) Age 21.8(1.8) 22.3(1.9)
Raven
25.7(5.8) 25.9(3.9) Raven 27.5(3.5) 27.8(4.5)
WRMT-R
Visual-Auditory Learning 123.8(11.9) 123.7(8.6) Visual-Auditory Learning 124.1(8.7) 123.6(9.2)
English Word Identification 98.7(4.1) 99.0(4.3) English Word Identification 39.4(12.2) 41.7(11.4)
English Word Attack 39.1(3.4) 39.3(4.3) English Word Attack 30.3(5.3) 29.3(5.7)
TOWRE
English Sight Word Efficiency 197.1(13.4) 197.9(11.3) English Sight Word Efficiency 72.1(6.1) 74.4(7.3)
English Phonemic Decoding Efficiency 113.3(10.2) 114(10.4) English Phonemic Decoding Efficiency 42.1(7.7) 43.4(7.3)
CTOPP
English Memory of Digits 16.9(3.0) 16.9(3.0) Chinese Memory of Digits 12.3(2.9) 12.5(3.2)
English Non-word Repetition 13.6(2.9) 13.8(2.5) Chinese Non-word Repetition 13.3(3.0) 12.8(3.4)

Chinese Word Identification 24.7(7.3) 25.4(5.6)

Chinese Sight Word Efficiency 80.6(12.3) 85.0 (12.3)
Note: WRMT-R=Woodcock Reading Mastery Tests- Revised; TOWRE=Test of Word
Reading Efficiency; CTOPP= Comprehensive Test of Phonological Processing.

3.1.2 ANOVA
In order to have a general idea of how well each group learned the artificial language, a
group×learning condition (2×2) factorial ANOVA analysis was conducted on discrimination and
naming task learning rates (Figure 4: Panel A and Panel B). Panel A demonstrates that in the
discrimination task, no significant interaction between group (English or Chinese) and learning
condition (alphabetic or logographic) was observed, F (1,134)=1.841, p=0.177. There was a
significant main effect for group, F (1,134)=8.402, p=0.004. English-speaking group
(mean=0.828) performed significantly better than the Chinese-speaking group (mean=0.561).
There was also a significant main effect for learning condition: the alphabetic group
(mean=0.874) learned faster than the logographic group (mean=0.591), F (1,134)=7.253,
p=0.008. The results in Panel B indicate that for the naming task, there was no significant
interaction between the effects of group and learning condition, F (1,129)=0.311, p=0.578. There
was a significant main effect for group, F (1,129)=4.466, p=0.037. The English-speaking group

33

(mean=0.447) performed significantly better than the Chinese-speaking group (mean=0.353).
There was also a significant main effect for learning condition: the alphabetic group
(mean=0.502) learned faster than the logographic group (mean=0.298), F (1,129) = 21.394, p <
0.001. We conducted t-test on naming new item task growth rate and found that the English-
speaking group (mean=0.656) learned significantly faster than Chinese-speaking group
(mean=0.447), p=0.033.

Figure 4. Interaction plot for λ of discrimination (A) and naming task (B)
3.1.3 Correlations between the cognitive-linguistic factors and outcome variables
Pearson’s correlations of behavioral predictors and dependent measures are presented in
Table 2 (English group) and Table 3(Chinese group). We found that WA with PDE was
correlated with English group alphabetic phonology learning and WI was correlated with
Chinese group logographic condition learning. Specifically, participants with better word or
pseudoword reading ability would learn novel form-sound associations faster. Memory measures
in both English group and Chinese group contributed to growth of learning.
Table 2 shows that for the English alphabetic learning group, discrimination task learning
growth rate λ was correlated with Raven (r=-0.282, p=0.039), VAL (r=0.356, p=0.007), WA
(r=0.312, p=0.018) and PDE (r=0.415, p=0.001) while naming task learning growth rate λ was
A" B"
D"
N"

34

correlated with Raven (r=0.363, p=0.012), VAL (r=0.310, p=0.032). There was no significant
correlation in English logographic learning group between behavioral predictors and task λs.

35

Table 2 Correlations among the principal variables in English group

1 2 3 4 5 6 7 8 9
A L A L A L A L A L A L A L A L A L
1.Age 1 1
2.Raven .014 -.244 1 1
3.VAL -.181 .107 .209 .150 1 1
4.WI .102 -.008 -.012 .104 .060 .331* 1 1
5.WA .044 -.135 -.034 .203 .295* .242 .545*** .747*** 1 1
6.SWE -.032 -.241 .240 .172 -.087 .095 .298* .167 .087 .094 1 1
7.PDE .102 -.061 -.131 .102 .094 .230 .564*** .716*** .502*** .723*** .341** .174 1 1
8.MD .317* -.195 .019 -.094 .013 .071 .087 .104 -.026 -.160 -.131 .022 .008 -.035 1 1
9.NR

.145 .339* -.219 -.271 .017 .267 .379** .576*** .352** .409** -.109 -.056 .300* .416** .102 .257 1 1
10.λ D .092 -.238 .282* .038 .356** .114 .179 .084 .312* .158 .027 .084 .415** .173 .156 .098 .115 .011
11.λ N -.078 -.159 .363* .197 .310* .079 .049 -.199 .048 -.003 .232 .070 .244 .034 -.178 -.031 -.150 -.104
12.λ NN

-.183 .246 .176 .118 -.014 .290 .151 -.001 -.172
Note: A=alphabetic learning condition; L=logographic learning condition. VAL=Visual-Auditory Learning; WI=Word Identification;
WA=Word Attack; SWE=Sight Word Efficiency; PDE=Phonemic Decoding Efficiency; MD= Memory of Digits; NR= Non-word Repetition.
λ
D
=overall learning rate of discrimination task; λ
N
=overall learning rate of naming task; λ
NN
=overall learning rate of naming new item task.
*p<0.05; **p<0.01; ***p<0.001.

36

In the Chinese alphabetic learning group, there were moderate significant correlations
between the discrimination task λ and VAL (r=0.495, p=0.026) and between naming task λ and
MD (r=0.459, p=0.032). In the Chinese logographic learning group, we found λ
D
was correlated
with Raven (r=0.500, p=0.035), VAL (r=0.497, p=0.036), WI (r=0.473, p=0.048), NR (r=0.513,
p=0.030). λ
N
was correlated with NR (r=0.471, p=0.031).

37

Table 3 Correlations among the principal variables in Chinese group
1 2 3 4 5 6 7 8 9 10 11
A L A L A L A L A L A L A L A L A L A L A L
1.Age 1 1
2.Raven -.201 .179 1 1
3.VAL -.295 -.220 -.025 .049 1 1
4.WI -.040 .114 .134 -.203 -.050 .036 1 1
5.WA .173 -.137 .040 -.064 -.068 .108 .709***

.580**

1 1
6.SWE -.325 -.076 .101 -.200 .292 -.056 .228 .558**

.336 .548*

1 1
7.PDE -.087 -.228 .446*

-.079 -.107 -.202 .315 .363 .505*

.640**

.610**

.806***

1 1
8.MD -.266 -.006 -.145 .338 .180 -.021 -.025 .312 .149 .151 .404 .251 .382 .140 1 1
9.NR .340 -.107 -.292 -.124 -.071 .006 .516*

.706***

.531*

.581**

.215 .428 .071 .254 .066 .449*

1 1
10.C-WI .028 -.108 -.143 -.092 .279 .029 .500* -.082 .603**

.048 .232 .177 .161 .188 .161 -.003 .159 -.332 1 1
11.C-SWE .037 -.573**

.294 -.012 .130 .231 .384 .003 .486*

.261 .421 .361 .379 .336 .240 .045 .379 .293 .296 .218 1 1
12.λ D .064 .064 .114 .500* .495* .497* -.122 .473* .002 .464 -.043 .237 .025 .047 .047 .328 .065 .513* -.011 -.114 .213 .155
13.λ N .139 .169 -.014 .190 .395 .401 .141 .405 .223 .424 .255 .359 .261 .135 .459* .093 .235 .471* .047 -.024 .380 .144
14.λ NN

-.117 .054 .352 .241 .323 .382 .355 .388 .170 .023 .206
Note: A=alphabetic learning condition; L=logographic learning condition. VAL=Visual-Auditory Learning; WI=Word Identification;
WA=Word Attack; SWE=Sight Word Efficiency; PDE=Phonemic Decoding Efficiency; MD= Memory of Digits; NR= Non-word Repetition; C-
WI=Chinese Word Identification; C-SWE=Chinese Sight Word Efficiency. λ
D
=overall learning rate of discrimination task; λ
N
=overall learning
rate of naming task; λ
NN
=overall learning rate of naming new item task. *p<0.05; **p<0.01; ***p<0.001.

38

3.2 Neural predictors
In the English-speaking alphabetic learning condition group, the left SMG cluster’s grey
matter volume had marginally significant positive correlation with discrimination task learning
growth rate, meaning that when this region had larger GM volume, the participants were
relatively faster alphabetic phonology learners. In contrast, the left postcentral gyrus grey matter
volume was negatively correlated with naming new item task growth rate.
In the English-speaking logographic learning condition group, there was only one cluster:
the pars orbitalis of left inferior frontal gyrus (IFG
or
) had significant positive correlation with the
growth rate of the discrimination task. Thus participants who have larger L. IFG GM volumes
were faster learners in the logographic condition learning.

Figure 5. Significant correlations between grey matter volumes and discrimination task
growth rates

Alphabe(c*
Alphabe(c*
Logographic*
E*group* C*group*
E*group* C*group*
A*
B
L*
Fast*learners>slow*learners*
Slow*learners>fast*learners*

39

In the Chinese-speaking alphabetic learning condition group, the grey matter volumes of
the left/right MTG, right HG, right precuneus were positively correlated with the learning growth
rate λ of the discrimination task. It is noteworthy that even though one cluster peaked at the right
HG, it extended to the right SMG. The left inferior parietal lobule was positively correlated with
the learning growth rate λ of the naming new item task. In addition, the grey matter volume of
the left rectus, left superior medial frontal gyrus, right middle cingulate, right paracentral lobule
were negatively correlated with λ
N
.
In the Chinese-speaking logographic learning condition group, negative correlations
between growth rates and volumes were observed. The volumes of the right parahippocampal
gyrus, left calcarine and right posterior cingulate gyrus were negatively correlated with λ of
discrimination task. Right anterior cingulate grey matter volume was marginally significant
correlated with λ of naming task.

40

Table 4 Brain regions showing significant correlations between GM volumes and learning rates
DV Learning
condition
Group Correlation Location Cluster
size
(voxels)
Peak MNI Z
x y z
λ D

Alphabetic E (+) L. supramarginal gyrus 253 -49.5 -45 33 3.7
C (+) L. middle temporal gyrus 2570 -64 -37 1.5 3.9
(+) R. middle temporal gyrus 1123 68 -33 -4.5 3.7
(+) R. Heschl’s gyrus 1888 57 -7.5 6 3.6
(+) R. precuneus 574 15 -49.5 22.5 3.5
Logographic E (+) L. inferior frontal gyrus orbital 789 -24 21 -16.5 3.7
C (-) R. parahippocampal gyrus 732 15 -39 -6 3.6
(-) L. calcarine 612 -12 -81 4.5 4.4
(-) R. posterior cingulate gyrus 637 4.5 -36 28.5 3.7
λ N Alphabetic E none
C (-) L. rectus 1030 -3 19.5 -19.5 3.4
(-) L. superior medial frontal gyrus 1248 -1.5 51 22.5 3.5
(-) R. middle cingulate 883 13.5 10.5 37.5 4.0
(-) R. paracentral lobule 1373 1.5 -33 63 3.8
Logographic E none
C (-) R. anterior cingulate 356 3 28.5 16.5 3.2
λ NN

Alphabetic E (-) L. postcentral gyrus 529 -15 -39 64.5 3.9
C (+) L. inferior parietal lobule 1357 -45 -54 42 4.8
Note: L=left; R= right. E= Native English-speaking group; C=Native Chinese-speaking
group. λ
D
= overall learning rate of discrimination task; λ
N
=overall learning rate of naming
task; λ
NN
=overall learning rate of naming new item task. Other than L. supramarginal
gyrus and R. anterior cingulate (in grey), the rest regions listed above all survived p at
0.005 level whole-brain corrections (0.05/10 =0.005). Inferior parietal lobule= inferior
parietal cortex that did not belong to either supramarginal gyrus or angular gyrus.

3. 3 Hierarchical regression analysis
We included predictors from behavioral and neural analyses for the purpose of
hierarchical regression analysis. Therefore, three sets of hierarchical regression analyses were
conducted to examine the unique contribution of each predictor in Hangul form-sound
association learning. Control variables (age and Raven score) were entered first followed by
cognitive-linguistic factors block and neural predictors block. Thus, both overall model and
relative contribution of each block could be assessed.
The English alphabetic group discrimination task results are shown in Table 5. The total

41

variance explained by the step 1 model was 8.7% [F (2, 51)=2.436, p=0.098] and by step 2
model was 35.8% [F (5, 48) = 5.362, p<0.001]. It can be seen that VAL, WA and PDE accounted
for an additional 27.1% of variance. Finally, neural predictors added another significant 19.8%
variance to the model and leading to a total of 55.6% variance explained in step 3 model [F (6,
47)=9.806, p<0.001]. The best predictor of λ
D
is PDE (β = 0.510) followed by L. SMG cluster
grey matter volume (β = 0.463) and Raven (β =0.314).

42

Table 5. English alphabetic group hierarchical regression predicting learning rate

Discrimination task
R
2
∆R
2
B SE ᵦ t
Step 1 .087 .087
Age .015 .023 .088 .657
Raven .024 .011 .281* 2.098
Step 2 .358*** .271***
Age .016 .020 .094 .788
Raven .024 .010 .278* 2.319
VAL .011 .005 .265* 2.082
WA .006 .020 .040 .285
PDE .019 .006 .397** 2.929
Step 3 .556*** .198***
Age .011 .017 .065 .651
Raven .027 .009 .314** 3.097
VAL .008 .004 .192 1.778
WA -.007 .017 -.050 -.419
PDE .024 .005 .510*** 4.372
L. SMG .015 .003 .463*** 4.573
Note. Statistical significance: * p<.05; **p<.01; ***p<.001
Chinese alphabetic group discrimination task results are presented in Table 6. Only step 3
model was significant [F (7, 12)=8.945, p=0.001]. A total of 83.9% variance was accounted by
seven predictors within which two neural predictors were significant, R. MTG (β = 0.600) and R.
precuneus (β = 0.510).

43

Table 6. Chinese alphabetic group hierarchical regression predicting learning rate

Discrimination task
R
2
∆R
2
B SE ᵦ t
Step 1 .020 .020
Age .011 .031 .088 .360
Raven .008 .015 .130 .533
Step 2 .329 .308*
Age .038 .028 .297 1.337
Raven .009 .013 .144 .692
VAL .015 .006 .593* 2.712
Step 3 0.839*** 0.510**
Age -.009 .021 -.069 -.415
Raven -.008 .008 -.123 -.948
VAL .005 .004 .196 1.327
L. MTG .000 .004 .009 .041
R. MTG .011 .005 .600* 2.342
R. HG .000 .004 -.003 -.013
R. Precuneus .005 .001 .510** 3.458
Note. Statistical significance: * p<.05; **p<.01; ***p<.001
Chinese logographic group discrimination task results are presented in Table 7. In the
first step of the hierarchical multiple regression, age and Raven were entered. This model was
not statistically significant [F (2, 15) = 2.930, p =0.084]. After entry of VAL, WI and NR at step
2 the total variance explained by the model was 76.1% [F (5, 12) = 7.658, p =0.002] with Raven
and VAL as significant predictors. The introduction of neural predictors explained only an
additional 0.3 % variance. When reversing the entry order of behavioral predictors and neural
predictors, they accounted for roughly the same amount of additional variance (behavioral

44

predictors =22.7% R square change, neural predictors=28.2% R square change). In the final
model no predictors make statistically significant contribution.
Table 7. Chinese logographic group hierarchical regression predicting learning rate

Discrimination task
R
2
∆R
2
B SE ᵦ t
Step 1 .281 .281
Age -.019 .024 -.197 -.807
Raven .028 .011 .587* 2.403
Step 2 .761** .480**
Age -.003 .017 -.027 -.151
Raven .023 .008 .485* 2.869
VAL .008 .003 .420* 2.676
WI .003 .003 .224 1.088
NR .021 .011 .402 1.969
Step 3 0.791* 0.029**
Age .016 .025 .164 .638
Raven .016 .011 .339 1.498
VAL .008 .004 .433 2.214
WI .001 .004 .096 .357
NR .020 .013 .392 1.632
R. parahippocampal gyrus .001 .003 .094 .397
L. calcarine -.001 .003 -.134 -.402
R. posterior cingulate gyrus -.003 .006 -.205 -.514
Note. Statistical significance: * p<.05; **p<.01; ***p<.001

45

4. Discussion
This study examined the independent contributions of behavioral and neural factors to
predict variability in phonological acquisition of an artificial Hangul language, and we found that
pseudoword reading efficiency was predictive of alphabetic phonological learning in native
speakers of English, which was consistent with the view that language transfer of reading skills
would occur easily between two alphabetic phonologies. The 2×2
approach adopted in the
present study permitted a fine-grained investigation into the specific dynamic between native
language status and the phonological training paradigm. In addition, the left SMG was correlated
with English group alphabetic phonology, while left and right MTG were correlated with
Chinese group alphabetic phonology learning, partially supporting the dual-route model of
reading and the assimilation/accommodation hypotheses.
4.1 English-speaking students learned faster than Chinese-speaking students in alphabetic
learning condition
We found that the English-speaking group learned significantly faster than Chinese-
speaking group in discrimination (p=0.004), naming (p=0.037) and naming new item (p=0.033)
task under alphabetic learning condition. At the phonological level, there were striking
differences between the English and our artificial alphabetic Hangul learning protocol in that
there were fewer letters in Hangul (22 in total; 12 consonants and 10 vowels) than in English (26
in total; 21 consonants and 5 vowels) and Hangul had an extremely shallow orthography when
English had a deep orthography. Because of its transparent and highly systematic characteristics,
Hangul alphabetic learning presented little challenge to English-speaking students to learn. Thus,
native English-speaking students learned alphabetic Hangul with relative ease.

46

Second, when taking a closer look at the results, we found that the Chinese group did not
perform better than English group under logographic learning conditions. Since Hangul and
Chinese are separate logographic system and did not share similarity of its form-sound
associations, the Chinese-speaking participants were not able to take advantages of their
experience in their native language. In other words, it can be theorized that orthographic
constraints are important for decoding orthographic patterns meaning: Chinese participants learn
slower than the English group under the alphabetic learning condition due to limited grapheme-
phoneme correspondence due to culture constraints. On the other hand, it seems that Chinese
word orthography familiarity did not facilitate Chinese speakers in the logographic learning
condition, which suggests that logographic orthography might work on a one to one basis. These
findings suggest that knowledge in one logographic orthography does not facilitate learning in
another logographic language.
4.2 Behavioral predictors
4.2.1 Native English speakers
Researchers have provided support that there are connections between native language
ability and L2 learning (R. L. Sparks, Patton, Ganschow, Humbach, & Javorsky, 2006). Our first
hypothesis was: in native English speakers, English word reading skills (accuracy and efficiency)
would predict Hangul alphabetic learning in line with past research if both L1 and L2 were all in
alphabetic orthography. Based on this hypothesis, we were able to find that higher scores in
English pseudoword reading accuracy (WA) and efficiency (PDE) would lead to faster learning
of Hangul alphabetic learning in English speakers (Table 2). In support of the findings reported
by Meschyan and Hernandez (2002), there was the same pattern that the first language WA and
second language word-decoding task were correlated. In addition, PDE served as a robust
predictor of alphabetic phonology learning in English speakers (Table 5).

47

However, word reading accuracy (WI) and efficiency (SWE) didn’t have noticeable
correlations with English group alphabetic learning. Comparing to WI and SWE, why WA and
PDE are better predictors in the model? WI and SWE tests tend to measure reading automaticity
while WA and PDE measures decoding ability. Therefore, WA and PDE, especially PDE
appeared to be a more robust predictor in that good pseudoword decoders have the ability to
represent the phonological structures of an unfamiliar language more accurately and efficiently
than poor decoders. Thus, the cross-language transfer that may help the Korean alphabetic
phonology learners in building up word decoding skills was reflected on PDE. Evidence from the
current study supports the use of PDE to screen English-speaking students who are going to learn
an alphabetic L2 for possible performance and help to determine the level of instructional
support that students may require.

VAL was correlated with the growth rates of both discrimination (r=0.356, p<0.01) and
naming tasks (r=0.310, p<0.05) of alphabetic phonology learning in the English group (Table 2)
but didn’t contribute significantly in the hierarchical regression model (Table 5). The VAL
directly assessed the ability to form associations between visual stimuli and oral responses and
had a large memory component in it but comparing to PDE, the contribution to prediction model
was limited since VAL was not a significant predictor in all three hierarchical regression models.
We did not find behavioral predictors of logographic phonology learning in English
native speakers. The results were consistent with previous findings that there are no correlations
between L1 and L2 word/pseudoword reading when L1 and L2 came from entirely different
orthography system (Gottardo et al., 2001).

48

4.2.2 Native Chinese speakers
We hypothesized that in native Chinese speakers, their Chinese word reading ability
would predict Hangul logographic learning based on past research when L1 and L2 are of the
same orthographic system. We examined whether native Chinese-speaking university students’
Chinese reading accuracy and efficiency would explain unique variance in Hangul logographic
learning growth rate. Language transfer didn’t occur between the two logographic scripts.
Further, transfer was not demonstrated between logographic to alphabetic phonology neither in
Chinese speakers (Table 3). Why was L1 (Chinese) word reading accuracy and efficiency not
transferrable to Hangul learning (both alphabetic and logographic)? From Yeung and Chan
(2013), we know that Chinese tone awareness was transferrable from L1 (Chinese) to L2
(English) suggesting a transfer of PA from non-alphabetic to alphabetic languages. Some form of
general phonological processing ability may underpin the L2 reading development, but Chinese-
WI and Chinese-SWE wasn’t able to capture it because Chinese is a logographic language
system with arbitrarily assigned word form and sound associations. So general Chinese PA
ability instead of Chinese word-reading skills may serve as better predictors of L2 learning.
Instead, VAL, MD, NR were associated with both alphabetic and logographic phonology
learning indicating the salient role of short-term memory in acquiring a new language in Chinese
participants (Table 3). Together with VAL’s result from English native speakers (Table 2), the
third hypothesis that visual-auditory memory and phonological working memory measures may
predict alphabetic and logographic Hangul learning outcome in both English and Chinese groups
was supported. This is consistent with a previous notion that phonological short-term memory is
predictive of novel word learning (Brooks & Kempe, 2013; Nicolay & Poncelet, 2013).
However, it is worth mentioning that phonological short-term memory played a more important

49

role in Chinese native speakers than English native speakers in terms of new form-sound
association learning.
The English (L2) WI test score in Chinese group was correlated with the discrimination
task growth rate in logographic condition. This may indicate that Chinese speakers treat English
words and Hangul logographic characters similarly. There has been an increased interest in the
field of third language acquisition and how the three languages interact with one another. Cenoz
(2001) compared Basque-speaking (L1) and Spanish-speaking (L1) children learning English
(L3) and found that the native speakers of Basque (relatively distant from English) showed more
evidence of lexical transfer from their L2 Spanish than the native Spanish speakers do. This
counterintuitive finding was attributed to the combined effect of typological closeness and L2
status. Other factors such as age, proficiency and amount of exposure, must be taken into
consideration when comparing of L3 learners with different language backgrounds.
4.2.3 About Raven
Finally, it is noteworthy that Raven scores played an important part in the prediction of
phonological learning acquisition. According to Pishghadam and Khajavy (2013), Raven
accounted for 12.2% of the variance in foreign language achievement. Intelligence has been
given less attention in the field of L2 learning despite the extensive interest in its role in general
learning (Pishghadam & Khajavy, 2013). But there is evidence that suggests that intelligence can
predict success in L2 achievement. For example, Fahim and Pishghadam (2007) found a positive
correlation between IQ and foreign language achievement. In this study, we found that in the
English speaker group, under the alphabetic learning condition, both the Raven and PDE
significantly contributed to the learning rate of the discrimination task (as shown in Table 5:
model 3). Specifically, we observed that the Raven was a significant predictor in the first, second

50

and third models, while in the first model Raven and age accounted for 8.7% of variance (Table
5). Raven was also positively and significantly correlated with the naming task in English
speakers alphabetic learning condition (Table 2) and the discrimination task in Chinese speakers
logographic learning condition (Table 3).
4.3 Neural predictors
As shown in Table 4, the whole brain analysis yielded the neural substrates correlated
with phonological learning.
4.3.1 Native English speakers
As indicated in Table 4, we found that in the English native speaker group: (1) the left
SMG was positively correlated with learning growth rate λ of discrimination task in alphabetic
Hangul learning, and that (2) the left IFG
or
was positively correlated with learning growth rate λ
of the discrimination task in logographic Hangul learning; We will discuss these findings
separately in the following two paragraphs.
We corrected for the total number of regression analyses that were done
(p=0.05/10=0.005), in addition to the whole brain permutation test. We used a stringent threshold
on the multiple comparison correction issue (p
corrected
=0.005). The L. SMG cluster (Figure 5,
Panel A Left) was marginally significant (the cluster size to survive multiple comparison
correction was 476 voxels while this cluster contained 253 voxels). When adding the L. SMG
cluster in the hierarchical regression model that predicted learning rates of the discrimination
task, the L. SMG cluster’s grey matter volume accounted for significant variance (19.8% R
square change) (Table 5). Specifically, this cluster provided useful information in which the
result was consistent with the neural substrates dual route model where alphabetic phonology
processing should be correlated with grey matter volume in the L. STG, L. MTG, L. SMG, and
L. IFG
op
. The correlations of the L. SMG and learning growth rates of discrimination task

51

reflected the recruitment of L1’s phonology related regions supporting the assimilation
hypothesis.
The second cluster was the L. IFG orbital part (IFG
or
: Figure 5, Panel B Left) that was
correlated with the learning growth rate λ of the discrimination task in logographic Hangul
learning. The IFG
or
was

associated with semantic processing, phonological processing and
working memory. It was not in Jobard et al. (2003)’s neural circuit of dual route model, but was
indicated as ventral pathway (direct route) together with pars triangularis in the recent review by
Angela (2011). Therefore it may reflect the accommodation hypothesis in that English-speaking
participants recruited the ventral pathway. We speculate that this cluster reflected the
phonological and memory processes of logographic phonology learning in English speakers
since no semantic information was involved in the training paradigm.
4.3.2 Native Chinese speakers
Consistent with the dual route model, alphabetic phonology processing positively
correlated with grey matter volume in the left MTG, right MTG, R. HG, R. precuneus (Table 4 &
Figure 5 Panel A, Right). The left MTG cluster (x=-64, y=-37, z=1.5), peaked within the range
of the left MTG cluster range (x=-63±5, y=-30±7, z=4±6) provided by Jobard et al. (2003). This
MTG cluster was associated with the GPC route in Jobard’s study. In other words, we found a
cluster in the left MTG even though our study focused on structural MRI and Jobard’s study
reported fMRI finding summaries. This cluster also reflected the recruitment of L2’s phonology
related regions and thus supported the accommodation theory.
The STG comprises the HG and the planum temporale, which are the regions for
phonetic perception (Li et al., 2014). Our result on the right HG is consistent with previous
results that a larger HG was associated with better phonological learning, but the fact that the

52

cluster was located on the right instead of the left hemisphere may indicate the inclination of
bilateral hemisphere involvement when Chinese speakers process new phonology information.
This tendency resonate with the results reported by Nelson, Liu, Fiez, and Perfetti (2009) in that
in their study when English speakers were learning Chinese, they used heavily left-lateralized
fusiform regions when reading English, but recruited an additional right fusiform region for
viewing Chinese. The fact that this STG cluster extended to MTG also lends support to the GPC
route neural substrates within the framework of the dual route model.
Our results indicated that larger grey matter volume of the left IPL cluster was correlated
with faster naming new item task learning rate in Chinese speakers under the alphabetic learning
condition (Table 4). The left IPL was implicated in phonological working memory, phonological
storage, semantic integration etc (Li et al., 2014). The left IPL was also marked as “language
talent area” by researchers (Della Rosa et al., 2013).
We found a cluster that peaked at the right parahippocampal gyrus that was negatively
correlated with learning rate of discrimination task in Chinese group. The limbic lobe consists of
three cortical areas: the parahippocampal gyrus, the cingulate gyrus, and the subcallosal gyrus.
These cortical regions were adjacent to subcortical structures of the limbic system such as the
hippocampus (Krebs, Weinberg, & Akesson, 2012). From an influential VBM study that was
published by our research group using a large sample of 416 Chinese college students, we know
that the grey matter volumes of bilateral hippocampus were correlated with form-sound
association ability (left: MNI=-28, -32, -6, Z =3.98; right: MNI =24, -36, 4, Z=4.39)(He et al.,
2013). The results from the present study that smaller right parahippocampal gyrus was
associated with faster learning in the discrimination task in the Chinese group may reflect the

53

optimal local neural allocation preferably to right hippocampus to implement maximum
functions.

54

5. Conclusions
This study contributed to the existing literature in a couple of ways: (1) We found the
pseudoword reading efficiency works as a good predictor of alphabetic phonological learning in
English speakers. Thus, the salient role of this task should be addressed during alphabetic L2
learning. (2) The left SMG was correlated with English group alphabetic phonology learning and
the left IFG
ob
was correlated with English group logographic phonology learning while the left
/right MTG and right HG were correlated with Chinese group alphabetic phonology learning,
partially supporting the dual-route model of reading and the assimilation/accommodation
hypotheses.
A few limitations of the study should be addressed. First, we only included two measures
of the Chinese group’s L1 ability, and they were not correlated with learning ability. Future
research might need to include more related variables (syllable task in Chinese, tone awareness
task, sound oddity and deletion task in Chinese, morphological awareness task in Chinese) to
fully cover L1 evaluation. In the present study, we only included a limited number of variables in
order to control the inflated type one error accumulated from the number of statistical tests. In
addition, our modest small sample size in the Chinese group affected our power to reach
significance in behavior and neural results and the generalizability of the findings to a larger
Chinese population. Finally, VBM is a dominant image processing method, but the underlying
mechanism of these grey matter volume individual differences is far from well defined.

55

Abbreviations
CR: Correct Ratio;
C-SWE: Chinese Sight Word Efficiency;
CTOPP: Comprehensive Test of Phonological Processing;
C-WI: Chinese Word Identification;
DRC model: Dual Route Cascaded model;
GPC: Grapheme-phonology Correspondence;
HG: Heschl’s gyrus;
IFG
or:
Orbital part of the Inferior Frontal Gyrus;
IFG
op
: Opercular part of the Inferior Frontal Gyrus;
IFG
tri
: Triangular part of the Inferior Frontal Gyrus;
IPL: Inferior Parietal Lobule;
IQ: Intelligence Quotient;
ITG: Inferior Temporal Gyrus;
MD: Memory of Digits;
MFG: Middle Frontal Gyrus;
MTG: Middle Temporal Gyrus;
NR: Non-word Repetition;
RCN: Rapid Color Naming;
RON: Rapid Object Naming;
SMG: Supramarginal Gyrus;
STG : Superior Temporal Gyrus;
SWE: Sight Word Efficiency;
PA: Phonological Awareness;
PDE: Phonetic Decoding Efficiency;

56

TOWRE: Test of Word Reading Efficiency;
VAL: Visual-Auditory Learning test;
WA: Word Attack;
WI: Word Identification;
WRMT-R: Woodcock Reading Mastery Tests-Revised;

57

References
Abrahamsson, N., & Hyltenstam, K. (2009). Age of Onset and Nativelikeness in a Second
Language: Listener Perception Versus Linguistic Scrutiny. Language Learning, 59(2),
249-306. doi: 10.1111/j.1467-9922.2009.00507.x
Andreou, G., Vlachos, F., & Andreou, E. (2005). Affecting Factors in Second Language
Learning. J Psycholinguist Res, 34(5), 429-438. doi: 10.1007/s10936-005-6202-0
Angela, D. F. (2011). The Brain Basis of Language Processing: From Structure to Function.
Physiological Reviews, 91(4), 1357-1392. doi: 10.1152/physrev.00006.2011
August, D. A., Calderon, M., & Carlo, M. (2001). The transfer of skills from Spanish to English.
NABE News, 11–12, 42.
Babcock, L., Stowe, J. C., Maloof, C. J., Brovetto, C., & Ullman, M. T. (2012). The storage and
composition of inflected forms in adult-learned second language: A study of the influence
of length of residence, age of arrival, sex, and other factors. Bilingualism, 15(4), 820-840.
doi: http://dx.doi.org/10.1017/S1366728912000053
Baum, S., & Titone, D. (2014). Moving toward a neuroplasticity view of bilingualism, executive
control, and aging. Applied Psycholinguistics, 35(5), 857-894. doi: 10.1016/
j.neuroimage.2008.12.05519167507
http://dx.doi.org/10.1017/S0142716414000174
Brooks, P. J., & Kempe, V. (2013). Individual differences in adult foreign language learning:
The mediating effect of metalinguistic awareness. Mem Cognit, 41(2), 281-296. doi:
10.3758/s13421-012-0262-9
Cenoz, J. (2001). The effect of linguistic distance, L2 status and age on cross-linguistic influence
in third language acquisition. Cross-linguistic influence in third language acquisition:
Psycholinguistic perspectives, 111(45), 8-20.
Chen, C., Chen, C., Xue, G., Mei, L., & Dong, Q. (2009). Cultural neurolinguistics (Vol. 178,
pp. 159-171). Netherlands: Elsevier Science & Technology.
Chen, X., Xu, F., Nguyen, T.-K., Hong, G., & Wang, Y. (2010). Effects of Cross-Language
Transfer on First-Language Phonological Awareness and Literacy Skills in Chinese
Children Receiving English Instruction. Journal of Educational Psychology, 102(3), 712-
728. doi: 10.1037/a0018802
Chiappe, P., & Siegel, L. S. (1999). Phonological awareness and reading acquisition in English-
and Punjabi-speaking Canadian children. Journal of Educational Psychology, 91(1), 20.
Chikamatsu, N. (1996). The effects of L1 orthography on L2 word recognition: A study of
American and Chinese learners of Japanese. Studies in Second Language Acquisition,
18(04), 403-432.
Chow, B. W.-Y., McBride-Chang, C., & Burgess, S. (2005). Phonological Processing Skills and
Early Reading Abilities in Hong Kong Chinese Kindergarteners Learning to Read
English as a Second Language. Journal of Educational Psychology, 97(1), 81-87. doi:
http://dx.doi.org/10.1037/0022-0663.97.1.81
Cisero, C. A., & Royer, J. M. (1995). The development and cross-language transfer of
phonological awareness. Contemporary Educational Psychology, 20(3), 275-303.
Coltheart, M., Rastle, K., Perry, C., Langdon, R., & Ziegler, J. (2001). DRC: a dual route
cascaded model of visual word recognition and reading aloud. Psychol Rev, 108(1), 204-
256.
Cox, R. W. (1996). AFNI: software for analysis and visualization of functional magnetic
resonance neuroimages. Comput Biomed Res, 29(3), 162-173.

58

Da Fontoura, H. A., & Siegel, L. S. (1995). Reading, syntactic, and working memory skills of
bilingual Portuguese-English Canadian children. Reading and Writing, 7(1), 139-153.
Della Rosa, P. A., Videsott, G., Borsa, V. M., Canini, M., Weekes, B. S., Franceschini, R., &
Abutalebi, J. (2013). A neural interactive location for multilingual talent. Cortex, 49(2),
605-608. doi: 10.1016/j.cortex.2012.12.001
Dufva, M. I. A., & Voeten, M. J. M. (1999). Native language literacy and phonological memory
as prerequisites for learning English as a foreign language. Applied Psycholinguistics,
20(3), 329-348. doi: 10.1017/S014271649900301X
Durgunoǧlu, A. Y., Nagy, W. E., & Hancin-Bhatt, B. J. (1993). Cross-Language Transfer of
Phonological Awareness. Journal of Educational Psychology, 85(3), 453-465. doi:
10.1037/0022-0663.85.3.453
Ehrich, J. F., Zhang, L. J., Mu, J. C., & Ehrich, L. C. (2013). Are alphabetic language-derived
models of L2 reading relevant to L1 logographic background readers? Language
Awareness, 22(1), 39-55.
Fahim, M., & Pishghadam, R. (2007). On the Role of Emotional, Psychometric, and Verbal
Intelligences in the Academic Achievement of University Students Majoring in English
Language. Asian EFL journal, 9(4).
Fiez, J. A. (2000). Sound and meaning: how native language affects reading strategies. Nature
neuroscience, 3(1), 3-5. doi: 10.1038/71078
Frost, R., Siegelman, N., Narkiss, A., & Afek, L. (2013). What predicts successful literacy
acquisition in a second language? Psychol Sci, 24(7), 1243-1252. doi:
10.1177/0956797612472207
Gathercole, S. E., & Baddeley, A. D. (1990). The role of phonological memory in vocabulary
acquisition: A study of young children learning new names. British Journal of
Psychology, 81(4), 439.
Geva, E., & Siegel, L. S. (2000). Orthographic and cognitive factors in the concurrent
development of basic reading skills in two languages. Reading and Writing, 12(1-2), 1-
30.
Geva, E., Wade-Woolley, L., & Shany, M. (1997). Development of reading efficiency in first
and second language. Scientific Studies of Reading, 1(2), 119-144.
Golestani, N. (2014). Brain structural correlates of individual differences at low- to high-levels
of the language processing hierarchy: A review of new approaches to imaging research.
The International Journal of Bilingualism, 18(1), 6-34.
Golestani, N., Molko, N., Dehaene, S., LeBihan, D., & Pallier, C. (2007). Brain structure
predicts the learning of foreign speech sounds. Cerebral cortex (New York, N.Y. : 1991),
17(3), 575-582. doi: 10.1093/cercor/bhk001
Golestani, N., & Pallier, C. (2007). Anatomical correlates of foreign speech sound production.
Cerebral cortex (New York, N.Y. : 1991), 17(4), 929-934. doi: 10.1093/cercor/bhl003
Golestani, N., Paus, T., & Zatorre, R. J. (2002). Anatomical correlates of learning novel speech
sounds. Neuron, 35(5), 997-1010.
Gottardo, A., Yan, B., Siegel, L. S., & Wade-Woolley, L. (2001). Factors related to English
reading performance in children with Chinese as a first language: More evidence of
cross-language transfer of phonological processing. Journal of Educational Psychology,
93(3), 530.
Grosjean, F., & Li, P. (2013). The psycholinguistics of bilingualism. New York: NY: John
Wiley& Sons, Inc.

59

He, Q., Xue, G., Chen, C., Chen, C., Lu, Z. L., & Dong, Q. (2013). Decoding the
neuroanatomical basis of reading ability: a multivoxel morphometric study. J Neurosci,
33(31), 12835-12843. doi: 10.1523/jneurosci.0449-13.2013
Hirshorn, E. A., & Fiez, J. A. (2014). Using artificial orthographies for studying cross-linguistic
differences in the cognitive and neural profiles of reading. Journal of Neurolinguistics,
31(0), 69-85. doi: http://dx.doi.org/10.1016/j.jneuroling.2014.06.006
Hogan, T. P., Catts, H. W., & Little, T. D. (2005). The Relationship Between Phonological
Awareness and Reading: Implications for the Assessment of Phonological Awareness.
Language, speech, and hearing services in schools, 36(4), 285-293.
Jia, F., Gottardo, A., Koh, P. W., Chen, X., & Pasquarella, A. (2014). The Role of Acculturation
in Reading a Second Language: Its Relation to English Literacy Skills in Immigrant
Chinese Adolescents. Reading Research Quarterly, 49(2), 251-261. doi: 10.1002/rrq.69
Jobard, G., Crivello, F., & Tzourio-Mazoyer, N. (2003). Evaluation of the dual route theory of
reading: a metanalysis of 35 neuroimaging studies. Neuroimage, 20(2), 693-712. doi:
10.1016/S1053-8119(03)00343-4
Katz, L., & Frost, R. (1992). The reading process is different for different orthographies: The
orthographic depth hypothesis. Advances in psychology, 94, 67-84.
Keung, Y.-C., & Ho, C. S.-H. (2009). Transfer of reading-related cognitive skills in learning to
read Chinese (L1) and English (L2) among Chinese elementary school children.
Contemporary Educational Psychology, 34(2), 103-112.
Koda, K. (1989). Effects of L1 orthographic representation on L2 phonological coding strategies.
J Psycholinguist Res, 18(2), 201-222.
Krebs, C., Weinberg, J., & Akesson, E. (2012). Lippincott's illustrated review of neuroscience.
Philadelphia: Wolters Kluwer/Lippincott Williams & Wilkins Health.
Li, P., Legault, J., & Litcofsky, K. A. (2014). Neuroplasticity as a function of second language
learning: Anatomical changes in the human brain. Cortex(0). doi:
http://dx.doi.org/10.1016/j.cortex.2014.05.001
Lindsey, K. A., Manis, F. R., & Bailey, C. E. (2003). Prediction of first-grade reading in
Spanish-speaking English-language learners. Journal of Educational Psychology, 95(3),
482-494. doi: http://dx.doi.org/10.1037/0022-0663.95.3.482
Liu, Y., Shu, H., & Li, P. (2007). Word naming and psycholinguistic norms: Chinese. Behav Res
Methods, 39(2), 192-198.
Lu, Z., Williamson, S., & Kaufman, L. (1992). Behavioral lifetime of human auditory sensory
memory predicted by physiological measures. SCIENCE-NEW YORK THEN
WASHINGTON-, 258, 1668-1668.
MacIntyre, P. D., Baker, S. C., Clément, R., & Donovan, L. A. (2002). Sex and age effects on
willingness to communicate, anxiety, perceived competence, and L2 motivation among
junior high school French immersion students. Language Learning, 52(3), 537-564.
MacWhinney, B. (1995). Language-specific prediction in foreign language learning. Language
Testing, 12(3), 292-320.
Manolopoulou-Sergi, E. (2004). Motivation within the information processing model of foreign
language learning. System, 32(3), 427-441. doi: 10.1016/j.system.2004.04.002
Mei, L., Xue, G., Lu, Z. L., He, Q., Zhang, M., Wei, M., . . . Dong, Q. (2014). Artificial language
training reveals the neural substrates underlying addressed and assembled phonologies.
PLoS One, 9(3), e93548. doi: 10.1371/journal.pone.0093548

60

Melby-Lervåg, M., & Lervåg, A. (2014). Reading comprehension and its underlying components
in second-language learners: A meta-analysis of studies comparing first- and second-
language learners. Psychological Bulletin, 140(2), 409-433. doi:
http://dx.doi.org/10.1037/a0033890
Melby-Lervåg, M., Lyster, S.-A. H., & Hulme, C. (2012). Phonological skills and their role in
learning to read: a meta-analytic review. Psychological Bulletin, 138(2), 322. doi:
10.1037/a0026744
Meschyan, G., & Hernandez, A. (2002). Is native-language decoding skill related to second-
language learning? Journal of Educational Psychology, 94(1), 14-22. doi:
http://dx.doi.org/10.1037/0022-0663.94.1.14
Nakada, T., Fujii, Y., & Kwee, I. L. (2001). Brain strategies for reading in the second language
are determined by the first language. Neurosci Res, 40(4), 351-358.
Nelson, J. R., Liu, Y., Fiez, J., & Perfetti, C. A. (2009). Assimilation and accommodation
patterns in ventral occipitotemporal cortex in learning a second writing system. Hum
Brain Mapp, 30(3), 810-820.
Nicolay, A.-C., & Poncelet, M. (2013). Cognitive abilities underlying second-language
vocabulary acquisition in an early second-language immersion education context: A
longitudinal study. J Exp Child Psychol, 115(4), 655-671. doi:
http://dx.doi.org/10.1016/j.jecp.2013.04.002
Perfetti, Liu, Y., Fiez, J., Nelson, J., Bolger, D. J., & Tan, L.-H. (2007). Reading in two writing
systems: Accommodation and assimilation of the brain's reading network. Bilingualism:
Language and Cognition, 10(02), 131. doi: 10.1017/s1366728907002891
Pishghadam, R., & Khajavy, G. H. (2013). Intelligence and metacognition as predictors of
foreign language achievement: A structural equation modeling approach. Learning and
individual differences, 24, 176-181. doi: http://dx.doi.org/10.1016/j.lindif.2012.12.004
Randhawa, B. S., & Korpan, S. M. (1973). Assessment of some significant affective variables
and the prediction of achievement in French. Canadian Journal of Behavioural
Science/Revue canadienne des sciences du comportement, 5(1), 24-33. doi:
http://dx.doi.org/10.1037/h0082325
Raven, J. C. (1990). Advanced Progressive Matrices: Sets I, II. . Oxford: Oxford University.
Sagi, Y., Tavor, I., Hofstetter, S., Tzur-Moryosef, S., Blumenfeld-Katzir, T., & Assaf, Y. (2012).
Learning in the fast lane: new insights into neuroplasticity. Neuron, 73(6), 1195-1203.
doi: 10.1016/j.neuron.2012.01.025
Service, E. (1992). Phonology, Working Memory, and Foreign-language Learning. The
Quarterly Journal of Experimental Psychology Section A, 45(1), 21-50. doi:
10.1080/14640749208401314
Skehan, P. (2002). Theorizing and updating aptitude. In P. Robinson (Ed.), Individual
Differences and Instructed Language Learning (pp. 69-93). Philadelphia, PA, USA: John
Benjamins Publishing Company.
Sparks, R., & Ganschow, L. (2001). Aptitude for learning a foreign language. Annual Review of
Applied Linguistics, 21, 90-111.
Sparks, R. L., Patton, J., Ganschow, L., Humbach, N., & Javorsky, J. (2006). Native language
predictors of foreign language proficiency and foreign language aptitude. Annals of
Dyslexia, 56(1), 129-160. doi: 10.1007/s11881-006-0006-2

61

Sun-Alperin, M. K., & Wang, M. (2011). Cross-language transfer of phonological and
orthographic processing skills from Spanish L1 to English L2. Reading and Writing,
24(5), 591-614. doi: 10.1007/s11145-009-9221-7
Sutherland, M. E., Zatorre, R. J., Watkins, K. E., Herve, P. Y., Leonard, G., Pike, B. G., . . .
Paus, T. (2012). Anatomical correlates of dynamic auditory processing: relationship to
literacy during early adolescence. Neuroimage, 60(2), 1287-1295. doi:
10.1016/j.neuroimage.2012.01.051
Tan, L. H., Spinks, J. A., Feng, C. M., Siok, W. T., Perfetti, C. A., Xiong, J., . . . Gao, J. H.
(2003). Neural systems of second language reading are shaped by native language. Hum
Brain Mapp, 18(3), 158-166. doi: 10.1002/hbm.10089
Taylor, J. S. H., Rastle, K., & Davis, M. H. (2013). Can cognitive models explain brain
activation during word and pseudoword reading? A meta-analysis of 36 neuroimaging
studies. Psychological Bulletin, 139(4), 766-791. doi: http://dx.doi.org/10.1037/a0030266
Thompson, P. M., Cannon, T. D., Narr, K. L., van Erp, T., Poutanen, V. P., Huttunen, M., . . .
Toga, A. W. (2001). Genetic influences on brain structure. Nat Neurosci, 4(12), 1253-
1258. doi: 10.1038/nn758
Thuy, D. H. D., Matsuo, K., Nakamura, K., Toma, K., Oga, T., Nakai, T., . . . Fukuyama, H.
(2004). Implicit and explicit processing of kanji and kana words and non-words studied
with fMRI. Neuroimage, 23(3), 878-889.
Torgesen, J., Wagner, R. Rashotte, C. (1999). Test of word reading efficiency. TX: Pro-Ed:
Austin.
Ventura-Campos, N., Sanjuan, A., Gonzalez, J., Palomar-Garcia, M. A., Rodriguez-Pujadas, A.,
Sebastian-Galles, N., . . . Avila, C. (2013). Spontaneous brain activity predicts learning
ability of foreign sounds. J Neurosci, 33(22), 9295-9305. doi:
10.1523/JNEUROSCI.4655-12.2013
Verhoeven, L. T. (1990). Acquisition of Reading in a Second Language. Reading Research
Quarterly, 25(2), 90-114. doi: 10.2307/747596
Wagner, R., Torgesen, J., & Rashotte, C. (1999). Comprehensive test of phonological processes
(CTOPP). Austin, TX: Pro-Ed.
Wang, M., Koda, K., & Perfetti, C. A. (2003). Alphabetic and nonalphabetic L1 effects in
English word identification: A comparison of Korean and Chinese English L2 learners.
Cognition, 87(2), 129-149.
Wang, M., Perfetti, C. A., & Liu, Y. (2005). Chinese-English biliteracy acquisition: Cross-
language and writing system transfer. Cognition, 97(1), 67-88. doi:
http://dx.doi.org/10.1016/j.cognition.2004.10.001
Wong, P. C. M., Warrier, C. M., Penhune, V. B., Roy, A. K., Sadehh, A., Parrish, T. B., &
Zatorre, R. J. (2008). Volume of left heschl's gyrus and linguistic pitch learning. Cerebral
Cortex, 18(4), 828-836. doi: 10.1093/cercor/bhm115
Woodcock, R. W. (1987). Woodcock Reading Mastery Tests-Revised. MN: American Guidance
Service.: Circle Pines.
Woodcock, R. W. (1998). Woodcock reading mastery tests–Revised/Normative Update. MN:
American Guidance Service: Circle Pines.
Yeung, S. S., & Chan, C. K. K. (2013). Phonological awareness and oral language proficiency in
learning to read English among Chinese kindergarten children in Hong Kong. British
Journal of Educational Psychology, 83(4), 550-568. doi: 10.1111/j.2044-
8279.2012.02082.x

62

Zatorre, R. J., Fields, R. D., & Johansen-Berg, H. (2012). Plasticity in gray and white:
neuroimaging changes in brain structure during learning. Nature neuroscience, 15(4),
528-536. doi: 10.1038/nn.3045
Zybert, J., & Stępień, S. S. (2009). Musical Intelligence and Foreign Language Learning.
Research in Language, 7(1), 99-111. doi: 10.2478/v10015-009-0007-4

Abstract (if available)

Abstract The present study assessed the role of multiple behavioral and neuroanatomical measures in the second language (L2) phonological learning in two samples of university students: native English and native Chinese. The artificial language materials based on Korean Hangul were learned under either alphabetic or logographic condition over eight days. Key findings: Pseudoword reading efficiency was associated with alphabetic phonological learning in native English speakers, which is consistent with the theory that language transfer would occur easily between two alphabetic phonologies. MRI Voxel-Based Morphometry (VBM) revealed that the left SMG was correlated with English group alphabetic phonology learning and the left IFGob was correlated with English group logographic phonology learning while the left /right MTG and right HG were correlated with Chinese group alphabetic phonology learning. These findings support the dual-route model of reading and the assimilation/accommodation hypotheses. We suggest that the aforementioned factors may serve as good predictors at least at the beginning stage of phonology learning of a new language.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

The neural correlates of skilled reading: an MRI investigation of phonological processing

PDF

Phonological activation in Chinese word recognition is contigent on readers’ fluency level: evidence from the perspective of phonetically-informed phonology

PDF

Neurobiological correlates of skilled and disabled reading

PDF

Statistical learning is a two-step process

PDF

Computational models and model-based fMRI studies in motor learning

PDF

Interpretation of pronominal forms across languages

PDF

Preparing English language learners to be college and career ready for the 21st century: the leadership role of secondary school principals in the support of English language learners

PDF

Prosody and informativity: a cross-linguistic investigation

PDF

Math achievement and self-efficacy of linguistically and ethnically diverse high school students: their relationships with English reading and native language proficiency

PDF

Learning semantic types and relations from text

PDF

Technology in the EFL classroom: Effects of a mobile English learning application on Turkish university student achievement and motivation

PDF

Dopamine dependent: examining the link between learning and treatment-resistant depression

PDF

The role of individual variability in tests of functional hearing

Asset Metadata

Creator Wei, Miao (author)

Core Title Cognitive-linguistic factors and brain morphology predict individual differences in form-sound association learning: two samples from English-speaking and Chinese-speaking university students

School College of Letters, Arts and Sciences

Degree Doctor of Philosophy

Degree Program Psychology

Publication Date 09/03/2015

Defense Date 08/10/2015

Publisher University of Southern California (original), University of Southern California. Libraries (digital)

Tag OAI-PMH Harvest,phonology,Reading,second language acquisition,structural MRI,VBM

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Lu, Zhong-Lin (committee chair), Manis, Franklin R. (committee chair), Kaiser, Elsi (committee member), Read, Stephen J. (committee member), Tjan, Bosco S. (committee member)

Creator Email iris.miaowei@gmail.com,weimiao@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c40-174765

Unique identifier UC11275474

Identifier etd-WeiMiao-3871.pdf (filename),usctheses-c40-174765 (legacy record id)

Legacy Identifier etd-WeiMiao-3871.pdf

Dmrecord 174765

Document Type Dissertation

Format application/pdf (imt)

Rights Wei, Miao

Type texts

Source University of Southern California (contributing entity), University of Southern California Dissertations and Theses (collection)

Access Conditions The author retains rights to his/her dissertation, thesis or other graduate work according to U.S. copyright law. Electronic access is being provided by the USC Libraries in agreement with the a...

Repository Name University of Southern California Digital Library

Repository Location USC Digital Library, University of Southern California, University Park Campus MC 2810, 3434 South Grand Avenue, 2nd Floor, Los Angeles, California 90089-2810, USA