Brain & Language 127 (2013) 1–5

Contents lists available at ScienceDirect

Brain & Language
journal homepage: www.elsevier.com/locate/b&l

Short Communication

Pure agraphia in Romaji after left inferior frontal gyrus infarction: A case
of selective deficit in syllable-to-grapheme conversion in Japanese
Kaeko Ogura ⇑, Toshikatsu Fujii, Kyoko Suzuki, Etsuro Mori
Department of Behavioral Neurology and Cognitive Neuroscience, Tohoku University Graduate School of Medicine, Sendai, Japan

a r t i c l e

i n f o

Article history:
Accepted 24 June 2013
Available online 15 August 2013
Keywords:
Sound-to-letter conversion
Japanese writing system
Pure agraphia
Left frontal lobe
Romaji

a b s t r a c t
The Japanese writing system employs two distinct categories of characters: Kana and Kanji. The difference between Kana and Kanji writing corresponds roughly to that between phonological and lexical systems in Western languages. When typing, most Japanese use alphabetical characters based purely on
phonological rules. In particular, the Romaji system is used, in which a syllable consists of a single vowel,
a consonant–vowel compound (e.g., ka, ki), or a sequence of consonant–y(semivowel)–vowel (e.g., kya,
kyu). We describe a right-handed Japanese patient who developed pure agraphia that affected Romaji
writing but preserved Kana and Kanji writing and who had a lesion in the left pars opercularis of the inferior frontal gyrus that extended to the anterior precentral gyrus. The patient demonstrated literal paragraphia in spelling Romaji across modalities. Our findings suggested that the patient’s agraphia in Romaji
after a confined left frontal infarction was manifested by a selective impairment in syllable-to-grapheme
conversion.
Ó 2013 Elsevier Inc. All rights reserved.

1. Introduction
Cultural differences influence cognitive processes and thus the
organization of neural networks in humans, and one such domain
of influence is the writing system (Paulesu et al., 2000; Siok, Perfetti, Jin, & Tan, 2004). Current models for the writing system in
Western languages postulate that there are two parallel routes:
the phonological route and the lexical route (Ellis, 1988; Margolin
1984; Roeltgen, 1993). The former relies on a phoneme-to-grapheme conversion process, which is important for writing regularly
spelled words (or non-words), whereas the latter relies on a
whole-word retrieval process, which enables one to write familiar
words regardless of their spelling regularity. Focal brain damage
can affect spelling in different ways: dysfunctioning of the phonological and lexical routes gives rise to phonological agraphia and
lexical agraphia, respectively (Beauvois & Dérouesné, 1981; Hillis,
Chang, Breese, & Heidler, 2004; Marien, Pickut, Engelborghs,
Martin, & De Deyn, 2001; Rapcsak, Arthur, & Rubens, 1988;
Roeltgen, Gonzalez Rothi, & Heilman, 1986; Roeltgen, Sevush, &
Heilman, 1983; Shallice, 1981).
The Japanese writing system employs two distinct categories of
characters: Kanji (morphogram) and Kana (syllabogram). The
relative difference comprising the regular-irregular dichotomy in

⇑ Corresponding author. Address: Department of Behavioral Neurology and
Cognitive Neuroscience, Tohoku University Graduate School of Medicine, 2-1,
Seiryo-machi, Aoba-ku, Sendai 980-8575, Japan. Fax: +81 22 717 7360.
E-mail address: ogura-kaeko@umin.ac.jp (K. Ogura).
0093-934X/$ - see front matter Ó 2013 Elsevier Inc. All rights reserved.
http://dx.doi.org/10.1016/j.bandl.2013.06.004

Western languages can appear to be a qualitative difference
between Kana and Kanji. For instance, some Japanese patients with
left frontal agraphia have shown particular difficulty in writing
Kana with literal paragraphia (Abe, Yokoyama, Torifuji, & Yanagihara, 1993; Sakurai, Matsumura, Iwatsubo, & Momose, 1997; Tohgi
et al., 1995), and these patients are considered to be showing phonological agraphia. However, in terms of linguistics, the phonological level of sound-to-letter conversion differs between Western
regular words (phoneme-to-alphabet conversion) and Kana words
(syllable-to-syllabogram conversion) (Iwata, 1984).
Most Japanese use another writing system for typing, an alphabetic writing system called Romaji, which is based on phonological
rules learnt in elementary school. A syllable in Romaji consists of a
single vowel, a consonant–vowel compound (e.g., ka, ki), or a sequence of consonant-y(semivowel)-vowel (e.g., kya, kyu) (Hattori,
1979). When typing with a software program for Romaji input, the
moment an alphabetical letter string makes up a syllable unit, it
automatically changes into a Kana character. Therefore, a whole
word written in Romaji does not appear on the monitor. Japanese
rarely use Romaji in handwriting, and Romaji words are seldom
seen in Japanese printed materials. Consequently, the syllableto-grapheme conversion can develop in Romaji writing without
the development of whole-word imagery. Selective damage of
the syllable-to-grapheme conversion system may result in impaired Romaji writing with preserved Kana and Kanji writing.
Neither the mechanisms nor the neural substrates of the Romaji
writing system have been investigated as of yet. We report here
on a patient with isolated agraphia in Romaji after a left frontal

2

K. Ogura et al. / Brain & Language 127 (2013) 1–5

infarct, indicating that there are two distinct sound-to-letter conversion systems and lexical routes developed in Japanese culture
and that the left frontal region engages purely in sound-to-letter
conversion without engaging in the associated whole-word
imagery.
2. Case report
2.1. Case description
The patient was a 51-year-old, right-handed Japanese clerical
worker with a 12th-grade education. He had been hypertensive
for 12 years, for which he had received treatment. A year previously, he had suffered from the Tolosa-Hunt syndrome, a painful
ophthalmoplegia caused by nonspecific inflammation of the cavernous sinus, and he had been treated with steroids. It is not
known whether this condition had been associated with the onset
of his infarction. He had been skilled at typing using Romaji input,
having had 20 years’ experience, but he suddenly realized that he
was unable to type despite being able to operate a keyboard and
to read alphabetical letters, and he was admitted to a nearby hospital. General physical and neurological examinations were unremarkable, but brain magnetic resonance imaging (MRI)
demonstrated a small infarct confined to the left frontal lobe. After
conservative therapy, he was referred to our hospital for further
assessment and rehabilitation. On admission, the patient was alert
and well-oriented. He showed no cranial nerve deficit and had no
hemiparesis or ataxia. The muscle tone of his arms and legs was
normal, as were his muscle stretch reflexes, and no pathological reflexes were elicited. Sensitivity to light touch, pain, temperature,
vibration, and joint motion was also normal. A brain MRI performed 28 days after the onset of symptoms demonstrated a small
lesion confined to the pars opercularis of the inferior frontal gyrus
and its subcortical area, extending to the foot of the anterior precentral gyrus in the left hemisphere (Fig. 1). No other lesion was
detected.
2.2. Neuropsychological examination
Neuropsychological evaluations were performed between one
and two months after onset. Detailed data for the standard neuropsychological tests are shown in Table 1. The patient’s intelligence
and memory were normal. His performance of executive function
was assessed by the Wisconsin Card Sorting Test, the Trail Making
Test, and verbal fluency. The ability was impaired, which may not
affect language performance as evidenced by the results of the
Western Aphasia Battery (WAB) (Kertesz, 1982) Japanese version

(Sugishita, 1986), the Wechsler Adult Intelligence Scale-Revised,
and Wechsler Memory Scale-Revised. The WAB is a standardized
instrument for evaluating the language and praxis function of Japanese adults, discerning the presence, degree, and type of aphasia,
and assessing writing ability in both Kana and Kanji. He showed no
apraxia on the WAB. He showed no signs of unilateral spatial neglect or constructional disability. His linguistic ability was assessed
by the WAB. He retained his ability to read words and sentences
written in Kana or Kanji. He could read and write Arabic numerals.
Writing of one- to four-digit numbers was performed without error
during dictation (100% correct on 40 numbers). When writing, he
used his right hand and produced orthographically well-formed
Kana and Kanji, with occasional difficulty writing Kana. He was
also able to read Romaji. In contrast, literal paragraphia was evident in his writing and typing of words in Romaji, even though
he could name, write, and type each of the alphabetic characters.
When typing, he used the fingers of both hands, which is typical
for a person skilled in typing. He showed no other signs of aphasia.
2.3. Special assessments for agraphia
2.3.1. The ability to write a single syllable in Romaji and Kana
To evaluate further the patient’s ability to spell in Romaji as
well as to write in Kana, we first gave him a monosyllable spelling
test. An articulatory test in Japanese includes 101 single syllables.
An examiner dictated the 101 syllables one at a time in a random
order. At 23 days after onset, the patient’s handwriting ability was
experimentally examined. Three days later, his writing, typing
(with no conversion to Kana), and orally spelling of these syllables
in Romaji were simultaneously assessed. On the same day, he was
also asked to transcribe into Romaji 101 Kana characters corresponding to the 101 syllables used in the monosyllable spelling
test. He was also asked to write 101 syllables in Kana from dictation. He committed many errors in all tasks involving Romaji
(Table 2), which were confined to literal paragraphia; the distribution of error types was remarkably similar across the tasks. In almost all cases, consonants were substituted for consonants and
vowels for vowels, and the consonant substitutions were not determined by either phonological or orthographical similarity. The
patient was not aware of all his spelling errors. He was able to
write 94 of the 101 Kana characters correctly and could repeat
all of the 101 syllables and read all of them when written in Romaji
or Kana.
2.3.2. The ability to write words and non-words in Romaji and Kana
Next, we assessed the patient’s ability to write dictated words
and non-words in Romaji and Kana. We presented him with eight

Fig. 1. MRI fluid-attenuated inversion recovery images (a, axial; b, sagittal) performed 28 days after onset, showing an infarct affecting the pars opercularis of the inferior
frontal gyrus and its subcortical area (arrow head), extending to the foot of the anterior precentral gyrus (arrow) in the left hemisphere.

3

K. Ogura et al. / Brain & Language 127 (2013) 1–5
Table 1
Performance on standard neuropsychological tests.
Tests
Language
WAB

Intelligence
WAIS-R

Score

Normative data; mean (SD)

Fluency
Auditory word recognition
Repetition
Object naming
Reading (Kanji and Kana)
Writing (Kanji and Kana)

10
60
100
60
9.8
9.7

10.0 (0)
59.5 (1.3)
98.7 (3.0)
59.2 (2.4)
9.5 (0.8)
9.6 (1.0)

Full IQ
Verbal IQ
Performance IQ

100
92
110
34

34.2 (2.1)

J-RCPM
Memory
Digit span
Spatial span
WMS-R

BVRT
Executive function
WCST

TMT
Verbal fluency
Praxis
WAB

Forward, backward
Forward, backward
General index
Verbal index
Visual index
Attention/concentration

5, 3
5, 3
92
85
112
101

Immediate recall
Delayed recall

9
9

Impaired if < 5
Impaired if < 4

Categories achieved
Perseverative errors
Difficulty maintaining set
Part A
Part B
Animal

1
6
0
39
90
14

3.7 (1.9)
4.2 (5.4)
1.5 (1.5)
32 (8.4)
76.0 (27.9)
16.1 (3.6)

Praxis

60

59.8 (0.7)

WAB: Western Aphasia Battery, WAIS-R: Wechsler Adult Intelligence Scale-Revised, J-RCPM: Japanese Raven’s Colored Progressive Matrices, WMS-R: Wechsler Memory
Scale-Revised, BVRT: Benton’s Visual Retention Test, WCST: Wisconsin Card Sorting Test, TMT: Trail Making Test.

Table 2
Error distribution (number of errors and percentage error).
Error type

Substitutions

Deletions
Insertions
Substitution + deletion
Total errors

Examples

C?C
V?V
C?C+V?V
C?V
y?C

pe ? me
si ? su
re ? mo
kya ? eya
ryo ? kto
tya ? ta
e ? ke
syo ? su

Hand-writing

Typing

Oral spelling

Transcription

Day-26

Day-26

Day-26

N (%)

N (%)

N (%)

N (%)

25 (44.6)
15 (26.8)
8 (14.3)
1 (1.8)
0 (0)
4 (7.1)
0 (0)
3 (5.4)
56 (100)

19 (63.3)
5 (16.7)
0 (0)
0 (0)
1 (3.3)
3 (10)
1 (3.3)
1 (3.3)
30 (100)

8 (28.6)
13 (46.4)
6 (21.4)
0 (0)
1 (3.6)
0 (0)
0 (0)
0 (0)
28 (100)

22 (57.9)
13 (34.2)
3 (7.9)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
38 (100)

Day-23

Day-26

N (%)
23 (62.2)
12 (32.4)
2 (5.4)
0 (0)
0 (0)
0 (0)
0 (0)
0 (0)
37 (100)

N: number of errors, C: consonant, V: vowel, y: semivowel.

lists of ten stimuli. The first to the fourth lists consisted of one- to
four-syllable words, and the fifth to eighth lists consisted of one- to
four-syllable non-words. He wrote 17 out of 40 (42.5%) words and
23 out of 40 (57.5%) non-words incorrectly in Romaji. We obtained
the following error probabilities for syllable units according to the
number of characters in the words or non-words: 10% (1/10), 40%
(8/20), 30% (10/30), and 18% (7/40) for words, and 30% (3/10), 35%
(7/20), 30% (9/30), and 33% (13/40) for non-words. There was no
significant difference in the percentage of errors for words and
non-words (v2 = 0.87, p = 0.35). The distributions of error types
for both words and non-words were essentially identical to those
in the monosyllable spelling tasks. In writing Kana from dictation,
the patient committed only a few errors (2 out of 40 words, 1 out of
40 non-words, and 3 out of 200 kana characters; 1.5%) and could
read all of the 40 words and 40 non-words when written in Kana
or Romaji.

2.4. Recovery
The patient’s ability to write and type using Romaji improved
gradually. Seven months after onset, his spelling accuracy in Romaji was 89% (90/101) when writing by hand and 97% (98/101) when
typing on the monosyllable task.

3. Discussion
We have described a patient who manifested an impairment in
spelling Romaji while his reading of Romaji was preserved after a
confined infarction in the left inferior frontal gyrus. He could write
and read Kanji, Kana, and Arabic numerals, and he showed no
aphasia, agnosia, apraxia, limb clumsiness, constructional disability, or dementia. Therefore, the patient’s symptom can be

4

K. Ogura et al. / Brain & Language 127 (2013) 1–5

interpreted as a focal sign after the onset of infarction, and it was
consistent with the criteria for pure agraphia (Roeltgen, 1993).
Neither the mechanisms nor the neural substrates of the Romaji
writing system have been investigated as of yet. The dissociation of
writing abilities in Kana and Romaji is an important finding. Both
writing systems rely on phonological-level associations (Hattori,
1979; Iwata, 1984). Kana characters are syllabic symbols, and a
spoken syllable can be directly mapped to a Kana character
(Kindaichi, 1988). Consequently, the term syllabograms should
be used for Kana characters, although they are conventionally
called phonograms (Iwata, 1984). When spelling in Romaji, in contrast, it is necessary to decompose a syllable into a grapheme unit
(Hattori, 1979), which is the smallest orthographic unit representing a sound (Wydell & Butterworth, 1999). In addition to this difference at the phonological level of sound-to-letter conversion,
Kana and Romaji writing systems have another important difference: Kana writing not only relies on syllable-to-syllabogram conversion but also on whole-word imagery. For instance, in a patient
with Kana agraphia, writing of Kana non-words was more impaired
than that of Kana words (Kokubo, Suzuki, Yamadori, & Satou,
2001). Therefore, the preservation of Kana writing ability may be
explained by the compensatory use of the lexical route. However,
preserved Kana writing in our patient cannot be explained by the
use of the lexical route because the writing of Kana non-words,
as well as of Kana words, was preserved. Therefore, our findings
substantiate the presence of two independent sound-letter
conversion systems in addition to the lexical route in Japanese:
the syllable-to-syllabogram conversion system (Kana) and the
syllable-to-grapheme conversion system (Romaji).
The patient did not show lexical errors, which most likely
reflects a compensatory use of whole-word imagery that is also observed in Western patients with phonological agraphia. It has been
suggested that the visual word images of the lexical system are
developed from visual exposure (Roeltgen et al., 1986). As mentioned earlier, the Japanese rarely use Romaji in hand-writing,
and Romaji words are seldom seen in Japanese printed materials.
Furthermore, even when typing, a whole word written in Romaji
does not appear on the monitor. Consequently, sound-to-letter
conversion can develop in Romaji writing without whole-word
imagery. The comparable error rates for words and non-words in
writing Romaji in our patient support the notion that the Romaji
writing system has no stored representations of the spellings of
words. There might be an objection that the patient did not develop the lexical route in his writing process. However, he retained his
ability to write Kanji, which heavily relies on the lexical route
(Iwata, 1984). These findings suggest that the lexical route in his
writing process, at least in Kanji and Kana, has been developed
and preserved after an infarction.
Roeltgen’s model of writing proposes two general categories of
functions: linguistic components and motor components (Roeltgen, 1993). The former contains two parallel routes (phonological
and lexical) and a graphemic buffer, whereas the latter includes
graphic output programming and motor programming. Our patient
could write all single alphabetic characters. The comparable error
rates for written spelling, oral spelling, and typing indicate that
his deficit cannot be explained by peripheral motor components,
including graphic output programming and motor programming;
his impairment existed in the phonological route or the graphemic
buffer. There are several neuropsychological studies reporting on
patients with selective damage to the graphemic buffer (Caramazza & Miceli, 1990; Caramazza, Miceli, Villa, & Romani, 1987; Schiller, Greenhall, Shelton, & Caramazza, 2001). These patients could
spell dictated meaningless syllables consisting of a consonant
and vowel flawlessly. Long words were affected more than short
words, and their errors in spelling most often occurred in the final
part of words. The absence of increasing errors for longer stimuli

and towards the end of a stimulus in our patient suggests that
the graphemic buffer for Romaji itself functions well. Therefore,
the patient’s impairment existed in the phonological route.
Roeltgen et al. (1983) suggests that the phonological route is
mediated by two sequential processes: (1) a segmentation process,
in which syllable sequences of words or non-words are broken
down into phonemes; and (2) phoneme-to-grapheme conversion,
in which the segmented phonemes are translated into graphemes.
In two sequential processes in the phonological route, the patient’s
segmentation process functioned well, making no phonological errors in spontaneous speech, repetition, oral naming, or reading
(Caramazza, Miceli, & Villa, 1986). Furthermore, his paragraphic errors in Romaji writing occurred within the framework of each syllable, i.e., within a single vowel, a consonant–vowel compound, or
a sequence of consonant-y(semivowel)-vowel. These findings suggest that our patient had a deficit in phoneme-to-grapheme conversion. Now, we will take a close look at phoneme-to-grapheme
conversion in Japanese. Syllabic units play a major role in phonological processing in the Japanese language, but phonemic representations also develop (e.g., Mann, 1986). Instruction in Romaji
writing helps Japanese people to become aware of segmental information at the phoneme level (Hattori, 1979). These previous studies allow us to hypothesize the presence of the segmentation
processes in which syllables are broken down into phonemes in
Japanese. However, this does not mean that phoneme-tographeme conversion indeed participates in the process of spelling
alphabet letters as a mandatory cognitive component in Japanese.
Hence, we should treat the phonological route without dividing
into the two sequential processes, and we embrace the concept
of syllable-to-grapheme conversion instead of phoneme-tographeme conversion.
There have been reports of both Western and Japanese patients
with pure agraphia following damage to the left frontal lobe. Hillis,
Wityk, Barker, and Caramazza (2003), Hillis et al. (2004) reported
on five patients who presented with pure agraphia at the onset
of a stroke, all of whom had lesions and/or hypoperfusion in
Brodmann’s areas (BA) 44, 45, and 6. The impaired phonemeto-grapheme conversion mechanisms in these cases indicate that
parts of BA 44 and 45 may be important for the phonological analysis needed for the phonological route, whereas a graphic output
program may depend on BA 6. Marien et al. (2001) reported on a
case of impaired use of the phoneme-to-grapheme conversion
mechanisms for spelling following a lesion involving the left operculum, precentral gyrus, and insula. Rapp and Caramazza (1997)
presented an analysis of the impaired writing performance of
two dysgraphic subjects. One of them had lesions of the posterior
left temporal and anterior and middle aspects of the peripheral left
occipital lobe and the posterior left frontal region. He made wellformed letter substitution errors in written spelling with relatively
intact oral spelling performance. The patient described by Anderson, Damasio, and Damasio (1990) showed impaired production
of specific letter shapes following lesions in the precentral gyrus.
Rapcsak et al. (1988) reported on a patient with lexical agraphia
from a focal lesion of the left precentral gyrus. These reports seem
to suggest that the inferior posterior frontal region, in particular BA
44 and 45 rather than BA 6, engages with the phoneme-to-grapheme conversion mechanism. Most Japanese patients with agraphia
due to a left frontal lesion have been reported to have a Kana-dominant or Kana-specific writing impairment (Abe et al., 1993; Sakurai et al., 1997; Tohgi et al., 1995). They commonly make Kana
errors such as paragraphia and rarely write badly formed Kana
characters. Therefore, the frontal regions are essential for the
sound-letter conversion (syllable-to-syllabogram conversion)
mechanism in Japanese as well. However, most of the patients
cited above had more extensive lesions and deficits in broader
aspects of the language function. The present patient had no

K. Ogura et al. / Brain & Language 127 (2013) 1–5

speech or reading disorders from disease onset. The findings were
obtained despite the fact that his brain damage was recent, and
thus there had been little time for the functional reorganization
of the neural networks. The patient’s good recovery in terms of
Romaji writing after seven months indicates that his impairment
at stroke onset was an acquired rather than a premorbid deficit.
Taken together, our findings suggest that agraphia in Romaji after
a confined left frontal infarction is a manifestation of a selective
impairment in syllable-to-grapheme conversion.
References
Abe, K., Yokoyama, R., Torifuji, S., & Yanagihara, A. (1993). A jargon agraphia and
selective agraphia of kana resulting from an infarct in the left middle frontal
gyrus. Shinkei Shinrigaku, 9, 196–201 (in Japanese).
Anderson, S. W., Damasio, A. R., & Damasio, H. (1990). Troubled letters but not
numbers. Domain specific cognitive impairments following focal damage in
frontal cortex. Brain, 113, 749–766.
Beauvois, M. F., & Dérouesné, J. (1981). Lexical or orthographic agraphia. Brain, 104,
21–49.
Caramazza, A., & Miceli, G. (1990). The structure of graphemic representations.
Cognition, 37, 243–297.
Caramazza, A., Miceli, G., & Villa, G. (1986). The role of the (output) phonological
buffer in reading, writing, and repetition. Cognitive Neuropsychology, 3, 37–76.
Caramazza, A., Miceli, G., Villa, G., & Romani, C. (1987). The role of the graphemic
buffer in spelling: Evidence from a case of acquired dysgraphia. Cognition, 26,
59–85.
Ellis, A. W. (1988). Normal writing processes and peripheral acquired dysgraphias.
Language and Cognitive Processes, 3, 99–127.
Hattori, S. (1979). Phonology and orthography. Tokyo: Ohsyukan shoten.
Hillis, A. E., Chang, S., Breese, E., & Heidler, J. (2004). The crucial role of posterior
frontal regions in modality specific components of the spelling process.
Neurocase, 10, 175–187.
Hillis, A. E., Wityk, R. J., Barker, P. B., & Caramazza, A. (2003). Neural regions
essential for writing verbs. Nature Neuroscience, 6, 19–20.
Iwata, M. (1984). Kanji versus Kana: Neuropsychological correlates of the Japanese
writing system. Trends in Neurosciences, 7, 290–293.
Kertesz, A. (1982). The western aphasia battery. New York, Grune & Stratton.
Kindaichi, H. (1988). The Japanese language. Tokyo: Iwanami Shinsho.

5

Kokubo, K., Suzuki, K., Yamadori, A., & Satou, K. (2001). Pure Kana agraphia as a
manifestation of graphemic buffer impairment. Cortex, 37, 187–195.
Mann, V. A. (1986). Phonological awareness: The role of reading experience.
Cognition, 24, 65–92.
Margolin, D. I. (1984). The neuropsychology of writing and spelling: Semantic,
phonological, motor, and perceptual processes. Quarterly Journal of Experimental
Psychology, 36, 459–489.
Marien, P., Pickut, B. A., Engelborghs, S., Martin, J. J., & De Deyn, P. P. (2001).
Phonological agraphia following a focal anterior insulo-opercular infarction.
Neuropsychologia, 39, 845–855.
Paulesu, E., McCrory, E., Fazio, F., Menoncello, L., Brunswick, N., Cappa, S. F.,
et al. (2000). A cultural effect on brain function. Nature Neuroscience, 8,
91–96.
Rapcsak, S. Z., Arthur, S. A., & Rubens, A. B. (1988). Lexical agraphia from focal lesion
of the left precentral gyrus. Neurology, 38, 1119–1123.
Rapp, B., & Caramazza, A. (1997). From graphemes to abstract letter shapes: levels
of representation in written spelling. Journal of Experimental Psychology Human
Perception and Performance, 23, 1130–1152.
Roeltgen, D. P., Gonzalez Rothi, L., & Heilman, K. M. (1986). Linguistic semantic
agraphia: A dissociation of the lexical spelling system from semantics. Brain and
Language, 27, 257–280.
Roeltgen, D. P. (1993). Agraphia. In K. M. Heilman & E. Valenstein (Eds.), Clinical
neuropsychology (3rd ed., pp. 63–90). New York: Oxford University Press.
Roeltgen, D. P., Sevush, S., & Heilman, K. M. (1983). Phonological agraphia: Writing
by the lexical-semantic route. Neurology, 33, 755–765.
Sakurai, Y., Matsumura, K., Iwatsubo, T., & Momose, T. (1997). Frontal pure agraphia
for kanji or kana: Dissociation between morphology and phonology. Neurology,
49, 946–952.
Schiller, N. O., Greenhall, J. A., Shelton, J. R., & Caramazza, A. (2001). Serial order
effects in spelling errors: Evidence from two dysgraphic patients. Neurocase, 7,
1–14.
Shallice, T. (1981). Phonological agraphia and the lexical route in writing. Brain, 104,
413–429.
Siok, W. T., Perfetti, C. A., Jin, Z., & Tan, L. H. (2004). Biological abnormality of
impaired reading is constrained by culture. Nature, 431, 71–76.
Sugishita, M. (1986). The Western aphasia battery (Japanese edition). Tokyo,
Igakusyoin.
Tohgi, H., Saitoh, K., Takahashi, S., Takahashi, H., Utsugisawa, K., Yonezawa, H., et al.
(1995). Agraphia and acaliculia after a left prefrontal (F1, F2) infarction. Journal
of Neurology, Neurosurgery, and Psychiatry, 58, 629–632.
Wydell, T. N., & Butterworth, B. (1999). A case study of an English–Japanese
bilingual with monolingual dyslexia. Cognition, 70, 273–305.