Journal of Neurology, Neurosurgery, and Psychiatry 1987;50:1125-1129

Alexia with agraphia of kanji (Japanese
morphograms)
MITSURU KAWAMURA,* KEIZO HIRAYAMA,* KEIKO HASEGAWA,f
NOBUYOSHI TAKAHASHI,* AKIRAYAMAURAt
From the Departments of Neurology, Brain Research Institute,* Rehabilitation,t and Neurosurgery,t School of
Medicine, Chiba University, Chiba, Japan

The case of the right-handed young Japanese woman with alexia with agraphia of kanji
(the Japanese morphograms) due to a small circumscribed haematoma in the left posterior inferior
temporal gyrus is described. Her chief complaint was the inability to read and write kanji. Detailed
examination showed that her alexia with agraphia was much more predominant for kanji than kana
(the Japanese syllabograms). These facts suggest that the processing of kanji and kana involves
different intrahemispheric mechanisms.
SUMMARY

Recent neuropsychological studies have revealed that
the left cerebral hemisphere is responsible for the
reading and writing of both kanji (which are the
structurally complex morphograms in written
Japanese with often several phonetic readings) and
kana (which are the relatively simple syllabograms
with unambiguous phonetic readings) (ref I and literature citations therein). Both kanji and kana are
used in many sorts of printed matter, such as newspapers, magazines and books. In schooling, all kana
(46 characters) and some graphically simple kanji are
learned by the end of the first grade of elementary
school. Standard use of kanji (approximately 1900
characters) prescribed by the Ministry of Education
are to be learned by the end of 9th grade. All kanji can
be represented in kana. For example, IL wind (the
first kanji character used in the sentence #4 of fig IA)
is also represented as bff /kaze/ in kana. In writing
sentences, kanji (0) and kana (A) are combined
.A
4'
together as shown in the example.
man walks. (RA:male A:human W walk 0): gramatical formative which makes the preceding noun
function as an adjective <: conjugated part of a
verb). Major lexical morphemes are usually written in
kanji while conjugated endings of verbs, adjectives as

A5

well as function words are always written in kana.
It has been suggested, however, that the processing
of kanji and kana involves different intrahemispheric
mechanisms, although very few clinical reports have
directly addressed this question. In the present study,
we describe a young Japanese woman who showed
alexia with agraphia for kanji due to a small circumscribed haematoma in the left posterior inferior temporal gyrus. The location of the lesion has been
identified by magnetic resonance imaging (MRI).
Selective deficits in the reading and writing of kanji,
as found in this case, are particularly important for
understanding the possibly different mechanisms
employed in the processing of kanji (morphograms)
and kana (syllabograms).
Case report
A 32 year old right-handed woman, educated through high
school, suddenly developed a headache on her left side on
4 October 1985. She became blind in the right half of her
visual field for about an hour, at the same time. Thereafter,
she found that she could not read the kanji in newspapers
and could not write in her diary in kanji. She had no
difficulty in the reading or writing of kana.

Neurological findings On admission (14 November 1985),

she was fully conscious and oriented. Pursuit and saccadic
ocular movements, optokinetic nystaginus and corrected
visual acuity were normal. Goldmann perimetry confirmed
full visual fields. There was no visual inattention to double
Address for reprint requests: Mitsuru Kawamura, MD, Department simultaneous stimulation. Visually guided reaching with
of Neurology, Brain Research Institute, School of Medicine, Chiba each upper extremity into each hemifield (visuo-motor coorUniversity, 1-8-1 Inohana, Chiba 280, Japan.
dination) was normal. Other cranial nerve functions. limb
muscle power, coordination, muscle stretch reflexes, cutaneReceived 12 September 1986 and in revised form 7 January 1987.
ous sensations and gait were intact.
Accepted 19 January 1987

1125

Kawamura, Hirayama, Hasegawa, Takahashi, Yamaura

1126
1

3

2

4

5

2

1

3

4

5

)t '~l

.n

P\"

of
v.

3-

I . 'P
.3 ?
t- gAj:
.
-

-3.0 ):.

P'

',/

%.I

~1

7 <,LI

\i~
'

t

1

$ p.

'

J"

T-c
0

r-

V.
--

Fig 1 Handwriting samples from dictation. (A) Preoperatively, she wrote the corresponding kana (circles) instead of kanji.
There were no paragraphic errors. (B) Postoperatively, she wrote kanji which she could not write preoperatively (circles). All
of the kanji errors in the figure were ofkanji used in verbs, whereas kanji nouns were written correctly. A significant trend of
this kind, however, was not found. (1) He had his hat blown off. (2) He picked it up with a stick. (3) A man is walking.
(4) The wind started to blow. (5) He fell into the river.

Neuropsychologicalfindings Neither hemineglect (bisection
of a line, copying the picture of a flower and Albert's linecrossing test), nor constructional apraxia (copying the figure
of a diamond shape, a Greek cross and a cube) were found.
Facial recognition, identification of overlapping figures
(Poppelreuter), colour naming, limb and facial praxis to
command, left-right discrimination, finger naming and simple verbal calculations were intact. Her Wechsler Adult
Intelligence Scale was at an average level (Verbal IQ = 94,
Performance IQ = 122, Full Scale IQ = 105).
Verbal language Spontaneous speech was well articulated
and fluent without paraphasic errors. There were occasional
word-finding pauses. The Western Aphasia Battery (WAB)2
profile of speech characteristics was rated as normal, and
there were no errors on the WAB tests of Spontaneous
Speech, Auditory Comprehension, or Repetition. The score
on the Token test was almost perfect (164/165).
Standarised naming tasks The scores of the WAB tests of
Naming or Naming tests of 100 pictures3were high (93%
and 90%, respectively). In contrast to the prompt responses
in Spontaneous Speech, Auditory Comprehension and Repetition, she was slow in performing these naming tests. Naming of certain pictures followed a latency of up to 10 seconds.
Standardised reading and writing tasks Her reading and
writing scores on the WAB tests were at the high level (93%,
98-5%, respectively) of aphasic controls, but failed to reflect
her specific difficulties in kanji. There were only a few errors
in reading and writing kana words, but she was very slow in

reading and writing kanji on the WAB tests. Reading of
some kanji followed a latency of up to 10 seconds and more
often, she could not read them at all. Her explanation for her
inability to read kanji was that she had forgotten the meaning and pronounciation of the kanji. When she wrote sentences which included both kanji and kana on dictation, she
wrote the corresponding kana instead of kanji (fig 1, preoperative). When she was asked to write the kanji which she
had previously replaced with kana, she was unable to do so,
but there were few paralexic or paragraphic errors. She
could not obtain the correct answer even when the index
finger was allowed to trace the shape of the kanji (kinesthetic
facilitation). Copying of written sentences was intact.
Special tests for reading and writing to compare the disturbances of kanji with those of kana She was asked to read
aloud and write in response to dictation 416 kanji and 104
kana letters, all of which she had learned in the first three
years of primary school. In reading kanji, she made 14 mistakes out of 416 trials (3.4%); in writing kanji, she made 32
mistakes out of 416 (7-7%). No mistakes in reading or writing kana letters were made. In contrast to the prompt
responses made in the kana tasks, her responses to the kanji
tasks were often very slow, even when she eventually proved
capable of reading and writing them. There were few paralexic and paragraphic errors. When she could not read or
write a kanji she reported that she had forgotten the character.

Anatomical localisation Computed tomography (CT) scan

Alexia with agraphia of kanji (Japanese morphograms)
)
(_I_N_)3XMAW

Fig 2 (Legend on next page).

1127

Kawamura, Hirayama, Hasegawa, Takahashi, Yamaura

1128

Fig 2 MRI (55 days after symptom onset) (inversion recovery technique). (a)
Horizontal view [time of repetition (TR) 2080 ms, time of inversion (TI)
500 ms], (b) Coronal view (TR 1500 ns, TI500ms), (c) Left lateral sagittal
view (60 mmfrom midline) TR 2080 ms, TI 500 nis) . All MRI showed a ring-like
high intensity area involving the cortex and the subcortical white matter ofthe
posterior part in the left inferior temporal gyrus, (1) lateral cerebralfissure,
(2) superior temporal sulcus.

(on 26 November 1985) showed a discrete, non-enhanced
haematoma without significant mass effect. The haematoma
involved the cortex and subcortical white matter of the left
inferior temporal gyrus. The angular gyrus was spared.
Magnetic resonance imaging (MRI) (Picker International
Ltd, VISTR-MR 0 5T) was performed with a 0-26 Tesla
superconducting magnet, using the inversion recovery technique (on 29 November 1985). The width of the slice was
10mm and the patient's head was positioned supine. Horizontal (fig 2a), coronal (fig 2b) and left lateral sagittal imagings (fig 2c) showed a ring-like high intensity area involving
the cortex and the subcortical white matter of the posterior
part in the left inferior temporal gyrus. There were no abnormalities in the left angular gyrus. The right hemisphere was
normal.
Surgical operation andfollow-up On 5 December 1985 the
subcortical haematoma was evacuated from the posterior
inferior part of the left temporal lobe. The haematomacap
sule was separated from the cerebral tissue and the hae-

matoma was removed in its entirety. Histologically, the cap-

sule was a cavernous haemangioma.
Postoperatively, her ability to read and write considerably
improved. Reading and writing scores on the WAB tests on
16 December 1985 were almost perfect. On the tests for writing sentences which included kanji and kana, she correctly
wrote kanji on 17 December 1985 (fig 1 postoperative). The
naming test of 100 pictures was also perfect with normal
speed. Reading aloud and writing 416 kanji letters and 104
kana letters were rapidly completed without any error on
18 December 1985.

Discussion
Alexia with agraphia was first described by Dejerine4
to occur with
in 891. l-and has. sin been thought
lesions of the left 'anguilar gyrus. 5 6

1129
Alexia with agraphia of kanji (Japanese morphograms)
Our patient showed alexia with agraphia and slight conclude that the left posterior inferior temporal
anomia, and had no other neurological or neuro- gyrus and surrounding areas have a function in visual
psychological manifestations. The location of the word memory which is more specific for kanji (morlesion in our patient was not in the left angular gyrus, phograms) than for kana (syllabograms), and that the
but in a circumscribed region of the left posterior infe- processing of kanji and kana involves different intrarior temporal gyrus. By means of MRI, the lesion was hemispheric mechanisms.
clearly demonstrated. She recovered her ability to
read and write completely following evacuation of the References
haematoma. The above findings show that alexia with
agraphia can occur due to damage to the left poste- 1 Iwata M. Kanji versus kana: neuropsychological correlates of the Japanese writing system. Trends Neurosci
rior inferior temporal gyrus and surrounding areas in
1984;7:290-3.
Japanese patients.
2
Kertesz
A. The Western Aphasia Battery. New York:
Recently several Japanese cases7-10 of alexia with
Grune & Stratton, 1982. (Translated into Japanese,
agraphia due to lesions of the left posterior inferior
Sugishita M etal, Tokyo: Igaku-Shoin, 1986).
temporal gyrus have been reported in Japan. Three of 3 Kamiyama G. Manual of Aphasia Therapy. Tokyo:
these cases,8-10 on which concrete data on the tasks
Gendai-sha, 1971.
were reported, showed difficulties predominantly in 4 Dejerine J. Sur un cas de cecite verbale avec agraphie,
suivi d'autopsie. CR Soc Biol (Paris) 189 1;3:197-201.
dealing with kanji. Iwata' concluded that Japanese
patients with a lesion of the left posterior inferior tem- 5 Greenblatt SH. Localization of lesions in alexia. In:
Kertesz A, ed. Localization in Neuropsychology, 1 st ed.
poral gyrus had disturbances of semantic processing
New York: Academic Press, 1983:323-56.
of kanji-reading and -writing.
A. Alexia. In: Aphasia and Associated Disorders.
The chief complaint of our patient was her inability 6 Kertesz
Taxonomy, Localization and Recovery, 1st ed. New
to read and write kanji. We demonstrated her
York: Grune & Stratton, 1979:191-210.
difficulty by using the WAB tests and other special 7 Yamadori A. Alexia with agraphia and angular gyrus.
tests with more than 400 kanji and 104 kana letters
Higher Brain Function Research (Tokyo) 1982;2:
for evaluation and her errors were found to be much
234-42.
more predominant in kanji than in kana. These facts 8 Oka H, Asano T, Hattori S, Mochio S, Sugishita M.
Pure alexia and alexia with agraphia caused by the left
show that at least some Japanese patients with lesions
inferior temporal and fusiform gyri. Neurol Med
of the left posterior inferior temporal gyrus and sur(Tokyo) 1985;23:73-6.
rounding areas have alexia with agraphia pro- 9 Iwata
M. Neural mechanisms of reading and writing in
dominantly in kanji.
the Japanese language. Funct Neurol (Roma)
Other characteristics of our patient include the fact
1986;1:43-52.
that there were few paralexic or paragraphic errors, 10 Shiota J, Kawamura M, Isono S, Hirayama K. Alexia
which are often seen in alexia with agraphia due to
with agraphia produced by a localized infarction of
left angular gyrus lesions,6 and that there was some
the inferior posterior temporal lobe. Brain Nerve
(Tokyo) 1986;38: 1051-5.
difficulty of recall. Based on the above findings, we