Journal of the Neurological Sciences 178 (2000) 42–51
www.elsevier.com / locate / jns

Alexia caused by a fusiform or posterior inferior temporal lesion
a,

b

b

b

c

Yasuhisa Sakurai *, Sosuke Takeuchi , Toshihiro Takada , Emiko Horiuchi , Hirofumi Nakase ,
Manabu Sakuta d
b

a
Department of Neurology, Mitsui Memorial Hospital, Tokyo, Japan
Department of Neurology, School of Medicine, University of Tokyo, Tokyo, Japan
c
Department of Neurology, Toranomon Branch Hospital, Kawasaki, Japan
d
Department of Neurology, Japanese Red Cross Medical Center, Tokyo, Japan

Received 18 April 2000; received in revised form 16 June 2000; accepted 26 June 2000

Abstract
We evaluated the alexia and agraphia of three patients with different lesions using Japanese kanji (morphograms) and kana
(phonograms) and made a lesion-to-symptom analysis. Patient 1 (pure alexia for both kanji and kana and minor agraphia for kanji after a
fusiform lesion) made more paragraphic errors for kanji, whereas patient 2 (alexia with agraphia for kanji after a posterior inferior
temporal lesion) showed severe reading and writing disturbances and more agraphic errors for kanji. Brodmann Area 37 was affected in
both patients, but in patient 2 the lesion was located lateral to that in patient 1. Patient 3 showed agraphia without alexia after restricted
lesion to the angular gyrus. We believe that pure alexia (patient 1) results from a disconnection between the medial fusiform gyrus and
posterior inferior temporal area (the lateral fusiform and inferior temporal gyri), whereas alexia with agraphia for kanji (patient 2),
corresponding to lexical agraphia in Western countries, results from damage to the posterior inferior temporal area, in which whole-word
images of words are thought to be stored. Furthermore, restricted lesion in the angular gyrus (patient 3) does not produce alexia; the
alexic symptom of ‘‘angular’’ alexia with agraphia may be the result of damage to the adjacent lateral occipital gyri.  2000 Elsevier
Science B.V. All rights reserved.
Keywords: Pure alexia; Letter-by-letter reading; Alexia with agraphia; Kanji; Kana; Angular gyrus; Lateral occipital gyri

1. Introduction
Alexia with agraphia caused by a lesion in the angular
gyrus and pure alexia (alexia without agraphia) caused by
lesion to the medial occipital gyri and the splenium were
first described by Dejerine [1,2]. He thought that the visual
images of letters and words were stored in the angular
gyrus and that any lesion which interrupted visual input to
the angular gyrus may cause alexia [1,3]. Geschwind [4]
also stressed the importance of the angular gyrus in
reading. In his view, the angular gyrus constitutes a cross-

*Corresponding author. Tel.: 181-3-3862-9111; fax: 181-3-56879765.
E-mail address: ysakurai@mitsuihosp.or.jp (Y. Sakurai).

modal association area that turns written language into
spoken language and vice versa.
In addition to the classical (spleno–occipital) type of
Dejerine’s pure alexia, a type of pure alexia subsequent to
damage to the subcortical white matter [5–10] or to the
cortical region [6,11–13] in the left temporo–occipital
junction or the occipital lobe has been reported. Greenblatt
[14] called this non-classical type [15] of pure alexia
‘‘preangular alexia’’, and the alexia caused by subcortical
lesion to the angular gyrus ‘‘subangular alexia’’. He
considered the subcortical white matter ventral to the
angular gyrus to be an important structure in subangular
alexia. The features of reading impairment in the nonclassical type of pure alexia have only rarely been described in neurological studies [8,10].
On the other hand, because of the existence of two

0022-510X / 00 / $ – see front matter  2000 Elsevier Science B.V. All rights reserved.
PII: S0022-510X( 00 )00363-4

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

distinct writing systems in the Japanese; i.e., kanji (morphograms) and kana (phonograms), Japanese alexic patients can show dissociative disturbances in reading between kanji and kana. It is generally believed that patients
with dyslexia with dysgraphia from a lesion in the angular
gyrus can read kanji better than kana, whereas patients
with dyslexia with dysgraphia from a posterior inferior
temporal lesion can read kana better than kanji [16–18].
Kanji are graphically complicated and have meaning as
well as two kinds of reading: on-reading which conveys
the phonetic value; and kun-reading which conveys the
meaning (e.g.,
[sui, mizu], water). In contrast, kana are
graphically simple and have one definite phonetic value
and no intrinsic meaning (e.g.,
[sa]). We argued that
kanji are an extreme form of irregular words (those that do
not obey English spelling rules) in that a whole-word
image is more concerned in reading and writing, whereas
kana words (or non-words) are an extreme form of regular
words (or pseudowords) in that grapheme-to-phoneme
correspondence is almost complete [19]. Thus the relative
difference of reading disturbances between regular words
and irregular words in European languages can present as
the definite dissociation of reading difficulties between
kanji and kana in Japanese, which may help to understand
the mechanism of reading in general.
We report three patients with reading and writing
disturbances; pure alexia after a fusiform lesion, alexia
with agraphia after a posterior inferior temporal lesion, and
agraphia after an angular lesion (Fig. 1 shows the anatomical relations among gyri discussed in this paper). We
examined the reading and writing abilities of these patients
using kanji and kana materials and correlated the results

43

with the patients’ lesions. The purpose of this study is to
show a neuroanatomical basis for reading applicable not
only to Japanese but to languages in general.

2. Case reports

2.1. Patient 1
The patient’s acute phase observations are reported
elsewhere [20].
In May 1995, a 73-year-old, right-handed man, who was
a university graduate and engineer, experienced floating
sensation in the head. He consulted a physician and
underwent an MRI examination in September of that year.
Cerebral infarction in the left medial occipital gyri was
detected (Fig. 2). At this time, his reading ability was
normal, and he was not aware of any visual field defect. In
April 1996, he suddenly found that he could not read
characters, and the next day realized that he could not read
any Japanese kana symbol, whereas he could read numerals and some kanji symbols. He still could write kana
symbols with ease. He entered a hospital in Yokohama,
where a CT scan showed a fresh infarction in the left
occipital lobe. Because of persistent reading impairment, in
May 1996 he was admitted to the Department of Neurology of the Japanese Red Cross Medical Center, Tokyo.
The neurological examination on admission, showed
right upper homonymous quadrantanopsia (later confirmed
by Goldmann perimetry), reading and writing impairment,
and minimal anomia. He read slowly and painfully even
words that he had just written. He could read three of five

Fig. 1. Anatomical relations among gyri in the occipital and temporal lobes (Drawn from Damasio [21]). Nomenclature is based on Dejerine [22] and
Talairach and Tournoux [23]. Coronal planes through the fusiform and inferior temporal gyri (Area 37; (a)) and the angular gyrus (b) are shown. An area
with oblique lines denotes the fusiform gyrus. Middle occipital gyrus is located just posterior to the angular gyrus. Abbreviations. SMG, Supramarginal
gyrus; AG, Angular gyrus; MO, Middle occipital gyrus; IO, Inferior occipital gyrus; Fu, Fusiform gyrus; IT, Inferior temporal gyrus; Li, Lingual gyrus.

44

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

Fig. 2. MRI T2-weighted axial images of patient 1, 4 months after the first stroke (September 1995; a–c), and axial and coronal images 45 days after the
second stroke (June 1996; d–h). The left side of each image is the patient’s right side (also in Figs. 3 and 4). In the first stroke, the lesion is restricted to the
left lingual gyrus and part of the cuneate gyrus (a–c; arrow, 0.5 tesla, TR / TE53000 / 100, also seen in Fig. h; arrow). In the second stroke, a high-intensity
area with ringed low intensity inside, indicative of hemorrhagic infarction, is located mainly in the fusiform gyrus but extends laterally to the inferior
temporal gyrus and posteriorly to the inferior occipital gyrus and subcortical area of the angular gyrus (d–h; 0.5 tesla, TR / TE52500 / 100).

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

kana characters. Whenever he had difficulty reading, he
used kinesthetic reading (i.e., tracing characters with the
fingers). He made some verbal paragraphia in kanji when
writing to dictation.
The Western aphasia battery (WAB; Japanese edition)
administered 1 month after on-set showed fluent spontaneous speech without paraphasia (Table 1). Auditory
comprehension, repetition, and naming were all perfect.
The patient could not, however, read some kanji and kana
symbols, including one- or two-character kanji nouns,
one-character kanji verbs with kana suffixes, and kana
function words, resulting in a poor performance for
sentence comprehension. Kinesthetic reading was not
always successful. His writing was nearly normal, but
there was impaired recall and verbal paragraphia of some
kanji symbols. In all, his language profile was rated as
pure alexia (strictly speaking, alexia with minor agraphia
for kanji). The digit span forward score was four. The
copying of sentences, drawing of figures, and color naming
were normal. The Panel D15 test showed intact color
vision.
Forty days after on-set, his reading impairment gradually began to improve, and he became able to read newspapers, although with some effort.

2.1.1. Neuroradiological examination
Lesion location was determined from T1- and T2weighted axial and coronal MRI images with a center-tocenter slice distance of 7–12 mm. Each gyrus was
identified using the atlases of Damasio [21], Dejerine [22],
and Talairach and Tournoux [23]. T2-weighted images at
the time of the first stroke showed a high-intensity area in
the left medial occipital gyri, including the lingual gyrus

Table 1
Neuropsychological test scores for the patients 1 and 2
Time after on-set (months)

1

64

Patient 1

Patient 2

WAIS-R
Verbal IQ
Performance IQ

113
101

–
–

WAB
Spontaneous speech
Information content
Fluency
Naming total
Repetition
Comprehension total
Reading total
Comprehension of sentences
Spelled kanji character recognition
Oral spelling of kanji characters
Writing total
Copying
Kanji writing from dictation
Kana writing from dictation

10 / 10
10 / 10
10 / 10
10 / 10
10 / 10
4.55 / 10
6 / 40
2/6
0/6
9.8 / 10
10 / 10
5.5 / 6
6/6

9 / 10
9 / 10
3.5 / 10
9.4 / 10
8.9 / 10
7.12 / 10
30 / 40
0/6
0/6
6.95 / 10
10 / 10
0/6
4.5 / 6

45

and part of the cuneate gyrus (Fig. 2a–c). T1- and T2weighted axial images 45 days after the second attack
showed an abnormal intensity lesion with low-(T2) and
high-(T1) intensity rings in the left occipito–temporal and
occipital areas, indicative of a hemorrhagic infarction (Fig.
2d–f). T2-weighted coronal images showed that this lesion
involved the left lingual (Brodmann Areas 17 and 18),
fusiform (Areas 18, 19, and 37), parahippocampal, inferior
temporal, and inferior lateral occipital gyri, probably
affecting the optic radiation and the inferior longitudinal
fasciculus, and that the lesion extended posteriorly and
upward to the subcortical white matter under the angular
gyrus (Fig. 2g and h). Single photon emission computed
tomography done with iodoamphetamine labeled with 123 I
( 123 I-IMP-SPECT) showed hypoperfusion throughout the
left occipital lobe, in the posterior part of the middle and
inferior temporal gyri, and in the angular gyrus.

2.2. Patient 2
This patient also has been described elsewhere [24].
Here we summarize the case report and give follow-up
data
In June 1991, a 71-year-old, right-handed man, a senior
high school graduate, suddenly realized that he could not
remember the names of objects or his family and could not
read newspapers or write anything but his own name. A
subcortical hemorrhage in the left posterior inferior temporal area was diagnosed by CT. Neuropsychological examination 3 months after on-set showed alexia with agraphia
that was preferentially disturbed for kanji, and severe
anomia. The patient read character-by-character sentences
written in kanji and kana in the WAB test. He could not
read kanji words and showed phonemic paralexia for kana
verbs and verbal paralexia for kanji verbs with kana
suffixes. Of note is that he could point to objects correctly,
even though he could not read aloud the names of the
objects written in kanji (understanding without phonology), characteristic of deep dyslexia [25]. He produced four
semantic paralexia errors in a 100-kanji character reading
test (described later). The digit span forward score was
five. Copying of sentences, drawing of figures, and color
naming were intact. Because the patient could explain how
to use objects and could correctly point to them on
command almost all the time, visual agnosia was ruled out
as a cause of anomia. MRI done 4 months after on-set
showed the lesion involved the left lateral fusiform and
inferior temporal gyri (Area 37) and extended subcortically
to the anterior third of the temporal lobe. Follow-up MRI 6
years later showed the same lesion with additional periventricular hyperintensity (Fig. 3; see also the figures in
Sakurai et al. [24]). Alexia with agraphia for kanji and
anomia persisted at the time of the re-evaluation, 5 years
after on-set, but character-by-character reading had disappeared (Table 1). In the WAB test of sentence reading,
he still could not read some kanji words, sometimes

46

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

Fig. 3. MRI T2-weighted axial (a) and coronal (b) images of patient 2, 6 years after on-set. A low-intensity area, suggestive of hemosiderin deposition,
with medial high-intensity is present in the left inferior temporal gyrus and lateral side of the fusiform gyrus and extends forward in the temporal lobe.
Note that this lesion is located more laterally than the lesion in patient 1 (0.5 tesla, TR / TE52500 / 120).

omitted one kana or kanji character in a word and
sometimes added a kana functor. Anomia and reading
scores for both kanji and kana were almost unchanged,
whereas his kana writing was worse than at re-examination
made 3 years before (from 77% to 58% on the 100-word
writing test). In addition, he had difficulty reading single
kana characters, and misread seven of 46 kana characters.

2.3. Patient 3
We have reported the patient in another context (patient
6) [26]. A short summary is given here.
In June 1996, a 61-year-old, right-handed man, a
university graduate and president of an advertising company, found that he could not remember the names of
well-known persons and things. A cerebral infarction in
the left parietal lobe was diagnosed by CT. The neuropsychological examination showed Gerstmann syndrome
(agraphia, acalculia, finger agnosia, and left-right disorientation) and slight anomia that converged to agraphia,
acalculia, and minimal anomia 2 months after on-set. The
digit span forward score was four. Copying of sentences,
drawing of figures, and color naming were intact. MRI
done 40 days after on-set showed the lesion involved the
left angular gyrus with the subcortical white matter and the
inferior part of the supramarginal gyrus, and that it

extended to the posterior horn of the lateral ventricle (Fig.
4; see also the figures in Sakurai et al. [26]).

3. Methods and results

3.1. Special tests
To evaluate the patients’ reading and writing ability
quantitatively, we administered a special test (Table 2), the
same as that used in a previous study [24]. The task was to
read aloud and write from dictation 100 kanji (one-character words) and their corresponding kana (readings of kanji,
or kana transcription of kanji words), all of which are
taught in the first 3 years of Japanese primary school. The
kanji and kana materials consisted of an equal number of
concrete nouns, abstract nouns, adjectives, and verbs.
Because kanji characters usually have on- and kun-readings (described in the introduction), and the on-readings
contain many homophones, we required the patients to
identify kanji characters by the kun-reading, the transcription of which was adopted in the kana reading test.
Table 2 shows scores and times required for reading for
each patient. Errors were classified into non-response,
visual, phonological, semantic, etc. (Table 3).
Results show that for patient 1, reading improved to the
normal range, although the act of reading took a long time.

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

47

Fig. 4. MRI T2-weighted axial (a) and coronal (b) images of patient 3, 40 days after on-set. The lesion is restricted to the cortico-subcortical structures of
the left angular gyrus and to part of the inferior supramarginal gyrus (1.5 tesla, TR / TE53000 / 102).

Errors consisted of two concrete nouns and one verb for
kanji and two abstract nouns and one adjective for kana.
He actively used the strategy of kinesthetic facilitation and
succeeded in reading most of the time. In contrast, his
writing scores were lower in the kanji task than in the kana
task (P,0.0001 by Fisher’s exact method). Half of the
errors were visually similar to the target (Table 4). He
could write all of the three kanji correctly that he could not
read.
Patient 2 showed severe reading and writing impairment
of kanji with relative sparing of kana reading and writing
(reading, P,0.0001; writing, P,0.0001 between kanji and
kana by Fisher’s exact method). Kinesthetic reading was
not always successful. Most kanji reading and writing

errors were the non-response type (impaired word recall).
All the circumlocution and semantic errors in kanji reading
were associated semantically with the correct answer: e.g.,
‘‘white’’ (incorrect) for ‘‘color’’ (correct), and ‘‘something
to eat’’ for ‘‘drink’’. In kanji reading, the number of
correct responses for concrete nouns was 13 / 25, abstract
nouns 13 / 25, adjectives 11 / 25, and verbs 8 / 25. In addition, this patient could not write kanji words that he could
not read.
Patient 3 read both kanji and kana very well, although
he spent more time reading than the control subjects (time
was not measured). In contrast, he had lower scores for
writing kanji words than kana words (P,0.0001 by
Fisher’s exact method) because of impaired word recall

Table 2
Special tests for reading and writing (100-single character kanji and the corresponding kana)
Patient
Symptom
Time after on-set (months)

1
Pure alexia (recovery phase)
2

2
Alexia with agraphia
61–64

3
Agraphia
2

Controls (n511)a
mean score (S.D.)
Mean time for reading

Kanji reading
Kana reading
Kanji writing
Kana writing

97 (7 min; 5 / 8)b
97 (5 min)
77 (35 min)
100 (12 min)

45 (22 min; 7 / 59)
95 (4 min)
6 (12 min)
58 (15 min)

100 c
99
29
97

99.6 (1.2)
99.6 (0.5)
95.9 (3.0)
99.3 (0.9)

a

1 min 32 s
1 min 13 s
10 min 11 s
8 min 7 s

All normal volunteers (ten men and one woman, ages 61–78, mean 68-years-old) gave their informed consent.
Values are % correct (and time spent in the task; correct responses with kinesthetic reading). The fraction denotes the number of correct responses
versus the total attempts at kinesthetic reading; e.g., 5 / 8 means that five of the eight kinesthetic reading trials were correct.
c
For patient 3, time for reading was not measured.
b

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

48

Table 3
Types of reading errors in 100-character kanji or 100-word kana test (values are the number of errors)
Patient

1

2

3

34
6
5
3
2
5

Total errors

3

55

Kana (n5100)
Phonological b
Phonological and / or visual b
Unrelated b

2
1

3
1
1

1

Total errors

3

5

1

Kanji (n5100)
Non-response
On response a
Circumlocution
Visual a
Semantic a
Unrelated a

3

a

On response: the patient read the kanji with its on-reading (the reading that conveys its phonetic value according to the Chinese pronunciation) but
could not give its kun-reading (the reading that conveys its meaning in the Japanese language). Visual response: changing the kanji into the reading of
another kanji that has a visual resemblance to the correct answer; e.g.,
([kiku], hear)→
([hiraku], open). Semantic errors: those associated
semantically with the target word; e.g.,
([iro], color)→ ([shiroi], white). Unrelated response: the changing of a kanji into another kanji that has no
visual or phonological similarity to the correct answer.
b
Phonological response: one or more characters in a kana word were substituted for other kana (phonemic paralexia). Phonological and / or visual
response: changing one character into another visually similar to the target character; e.g.,
([aoi], blue)→
([ooi], numerous). Unrelated
response: the changing of a kana word into another kana word that has no visual or phonological similarity; e.g.,
([hayai], fast) →
([aoi],
blue).

Table 4
Types of writing errors in the 100-character kanji and 100-word kana tests (values are number of errors)
Patient
Kanji (n5100)
Non-response
Constructional a
Visual a
Visual and / or semantic a
Phonological a
Semantic a
Unrelated a
Neologism
Total errors

1

2

3

5

90

67
2
1

1
1
2

1

94

71

9
1
2
4
2
23

Kana (n5100)
Constructional b
Phonological b
Phonological and / or visual b
Unrelated b

31
9
2

Total errors

42

a

1
2

3

Constructional response: the omission or addition of a component of a kanji; e.g.,
([tori], bird)→ . Visual response: substitution of another kanji
that has a visual resemblance to the correct answer; e.g.,
([oto], sound) → ([mukashi], old times). Some visual errors were semantically associated
with the target (visual and / or semantic); e.g.,
([kao], face)→ ([atama], head). Phonological response: substitution of another kanji with a phonetic
value similar to the correct answer; e.g.,
([kazu], digit)→ ([katsu], active). Semantic errors: those associated semantically with the target word; e.g.,
([hayai], fast)→ ([isogu], hurry). Unrelated response: substitution of another kanji that has no visual or phonological similarity to the correct answer.
b
Constructional response: the omission or addition of a component of a kana. Phonological response: one or more characters of a kana word were
substituted for other kana (phonemic paragraphia). Phonological and / or visual: changing one character into another visually similar to the target character;
e.g.,
([kuroi], black)→
([kurai], dark). Unrelated response: substitution of another kana word with no visual or phonological similarity;
e.g.,
([huyu], winter)→
([umi], sea).

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

and made a few phonological errors in writing kana words
(Table 4).

4. Discussion
Patient 1 presented with pure alexia (alexia for both
kanji and kana and minor agraphia for kanji). Anatomically, he is classified as having non-classical type preangular
alexia [15,27]. In Japan, pure alexia is frequently accompanied by slight agraphia for kanji [28], but why this
agraphia coexists with pure alexia remains unknown. For
our patient, it was because the lesion involved the adjacent
lateral side of the fusiform gyrus and the inferior temporal
gyrus (posterior inferior temporal area); the type of lesion
which causes alexia with agraphia for kanji [24,29], or
pure agraphia for kanji [30]. His condition was semiologically similar to what is called group A alexia with agraphia
produced by a posterior inferior temporal lesion, in which
kana character (or word) reading is more impaired than
kanji character (or word) reading [24].
Patient 2 had alexia with agraphia for kanji accompanied by severe anomia caused by a posterior inferior
temporal lesion (group C alexia with agraphia, as described in Sakurai et al. [24]). Alexia with agraphia for
kanji corresponds to the category of lexical agraphia
[30,31] in European countries, in which irregular word
spelling is selectively impaired. As this symptom is
thought to derive from deficient visual images of words,
we originally designated it ‘‘morphologic’’ agraphia, but
because ‘‘morphologic’’ has a specific usage in linguistics,
hereafter we use ‘‘orthographic’’ agraphia to refer to this
state. He could write his own name at the on-set, either
because the whole-word image of his name was so robust,
or because he did it almost automatically (with little access
to the orthography).
Patient 3 first showed Gerstmann syndrome with
anomia, with agraphia, acalculia, and minimal anomia
remaining. A lesion in the angular gyrus is known to cause
alexia with agraphia, but he did not show alexia throughout
the clinical course (discussed later). Agraphia was more
pronounced for kanji words than for kana words.
Next we correlate the alexic symptoms of the three
patients with their lesions and discuss kinesthetic reading
in alexia.

4.1. Neuroanatomical basis of alexia
According to Greenblatt’s classification [27], patients 1
and 2 showed preangular alexia. Their lesions were similar
to spleno–occipital type alexia without hemianopsia [7],
except that there was no damage to the splenium. Patient 1
had his first stroke in the left medial occipital area (mainly
the lingual gyrus). At that time, right upper quadrantanopsia may have occurred, even though he was not aware of it,
and visual information arriving in the left lingual gyrus

49

may not have been processed further. Early visual analysis
in the lingual gyrus may have been completely compensated for by the right homologous area. His second attack
was within the occipital lobe. This time the lesion was
located mainly in the left fusiform gyrus and extended to
the inferior temporal and lateral occipital gyri, including
the optic radiation. Neurological and neuropsychological
examinations showed right upper homonymous quadrantanopsia and pure alexia. Some patients with pure alexia
also have a lesion restricted to the fusiform gyrus, which
lesion sometimes extends to the lateral occipital gyri
(Henderson et al. [8]; Kawamura [15], Case 2; Buxbaum et
al. [32]). This preangular type of pure alexia is interpreted
as being within the framework of disconnection theory [7],
but the classic disconnection theory cannot account for
alexia with agraphia caused by a posterior inferior temporal lesion (patient 2). The fact that visual information from
the right homologous area did not compensate for the
damage to the left fusiform gyrus suggests there is
functional localization for letters or words in the fusiform
gyrus.
Moreover, the main affected site was Brodmann Area 37
in both patients, but the lesion in patient 2 with alexia with
agraphia was located more laterally in the fusiform and
inferior temporal gyri. Compared with patient 1, patient 2
had severer reading and writing disturbances for kanji, and
made more agraphic errors (non-responses due to impaired
recall) for kanji, whereas patient 1 made paragraphic errors
(visual and semantic responses) as well as agraphic errors
for kanji (Table 4).
These findings suggest that both the fusiform gyrus and
the inferior temporal gyrus constitute a contiguous (but
different in detail) functional localization in reading and
writing, visual images of individual kana being stored
medially in the fusiform gyrus, and visual images of
individual kanji and whole-word images of both kana and
kanji words stored more laterally in the fusiform and the
inferior temporal gyri. If so, the reason why patient 1 could
write kanji that he could not read, although with some
paragraphic errors, was because the visual images of the
kanji remained relatively intact in the lateral fusiform and
the inferior temporal gyri (posterior inferior temporal
area), and he used them in writing. Thus pure alexia
(patient 1) results from a disconnection between the medial
fusiform gyrus and the posterior inferior temporal area, in
which access to orthography for words is interrupted,
whereas alexia with agraphia for kanji (patient 2) results
from damage to the posterior inferior temporal area, in
which orthography per se is impaired. This hypothesis
contradicts the disconnection theory, in which the angular
gyrus is regarded as the storage site [1,3] or cross-modal
association area [4].
Furthermore, it should be noted that patient 3 with the
angular lesion showed not alexia but agraphia throughout
the clinical course, although he did have minimal difficulty
in reading kanji words in the acute phase [26]. His lesion

50

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51

was restricted to the angular gyrus and the inferior part of
the supramarginal gyrus. The findings for this patient
suggest that restricted lesion of the angular gyrus causes
not alexia, but agraphia. The concept of alexia with
agraphia caused by an angular lesion has been established
[33–36], but even in Dejerine’s original patient there was a
lesion extending posteriorly to the lateral occipital gyri [1].
Some subsequent studies on angular alexia were of lesions
affecting the occipital lobe [37]. CT and MRI studies from
the 1970s also showed involvement of the lateral occipital
gyri just posterior to the angular gyrus in alexia with
agraphia [18,38]. The accompanying inability to copy
words [1,39] suggests the existence of a more widespread
lesion that involves the parieto–occipital and occipital
areas. When a lesion is confined to the angular gyrus, the
alexic symptom disappears within a few months [40], or
apparently does not appear as in our patient 3. Yamadori
[41] examined five cases of ‘‘angular’’ alexia with agraphia and concluded that clinical data were still insufficient to attribute alexia with agraphia to damage to the
restricted area of the angular gyrus. It is our view that
involvement of the lateral occipital gyri (more precisely,
the middle occipital gyrus just posterior to the angular
gyrus) is essential for the production of ‘‘angular’’ alexia.
Kleist [42] first ascribed the alexic symptom in alexia with
agraphia to lesion extension from the angular gyrus to the
middle occipital gyrus.
This view is supported by our activation studies with
positron emission tomography (PET) [43,44] (for the
location of individual region-of-interest, see Sakurai et al.
[45]). In the reading aloud tasks for kanji words, kana
words and kana non-words, as compared with the fixation
task, we found that the lateral occipital gyri posterior to the
angular gyrus were activated in the kana word and nonword tasks, whereas the angular gyrus was deactivated in
all the tasks. We reanalyzed our data using Statistical
Parametric Mapping [46] and confirmed the above findings
[47]. This fact may explain why kana reading is more
impaired in ‘‘angular’’ alexia with agraphia (described in
Section 1). Deactivation of the angular gyrus in reading
also has been reported in Western countries [48].

4.2. Kinesthetic reading in alexia
Kinesthetic reading (shreibendes Lesen in the German
literature [49]) is often used by Japanese patients with pure
alexia as a strategy for overcoming reading difficulty.
Kinesthetic reading clearly was less helpful for patient 2
with alexia and agraphia than for patient 1 with pure
alexia. This reading strategy is known to be effective for
patients with pure alexia. Although it is not clear whether
kinesthetic reading is effective in alexia with agraphia for
kanji, published reports on alexia with agraphia that refer
to this strategy show that it is not helpful for reading kanji
and kana together (e.g., Kawahata et al. [50]) or kanji
reading alone (Kawamura [15]; Patient TN). This probably

is because in pure alexia the visual image of a word is
mostly preserved access to it being interrupted; whereas, in
alexia with agraphia for kanji, the visual image of the word
itself is damaged, as described in Section 4.1. The fact that
Brodmann Area 37 (a lesion in which causes alexia with
agraphia for kanji) was activated in a PET study not only
by word processing but by tactile reading [51] suggests
this area is important in kinesthetic reading.
The above discussion is based on a small number of
case and activation studies. Further studies therefore are
required to determine the roles of the angular gyrus and the
lateral occipital gyri in reading, and the mechanism of
kinesthetic reading.

Acknowledgements
We are grateful to Dr. Toru Mannen, Mitsui Memorial
Hospital, for his support and encouragement. We also
thank Dr. Sumiko Sasanuma, International University of
Health and Welfare, for teaching us the terminology of
psycholinguistics and to Dr. Meiko Shiina, Department of
Neurology, University of Tokyo, for her help in examining
patient 1’s MRI film.

References
´ ´ verbale avec agraphie, suivi
[1] Dejerine J. Sur un cas de cecite
d’autopsie. Compt Rend Soc Biol 1891;3:197–201.
´
[2] Dejerine J. Contribution a` l’etude
anatomopathologique et clinique
´
´ ´ de cecite
´ ´ verbale. Compt Rend Soc Biol
des differentes
varietes
1892;4:61–90.
´
`
[3] Dejerine J. Semiologie
des affections du systeme
nerveux, 2nd ed,
Paris: Masson et Cie, 1926.
[4] Geschwind N. Disconnexion syndromes in animals and man. Brain
1965;88:237–94, 585–644.
[5] Caffarra P. Alexia without agraphia or hemianopia. Eur Neurol
1987;27:65–71.
[6] Damasio AR, Damasio H. The anatomic basis of pure alexia.
Neurology 1983;33:1573–83.
[7] Greenblatt SH. Subangular alexia without agraphia or hemianopsia.
Brain Lang 1976;3:229–45.
[8] Henderson VW, Friedman RB, Teng EL, Weiner JM. Left hemisphere pathways in reading: Inferences from pure alexia without
hemianopia. Neurology 1985;35:962–8.
[9] Leegaard OF, Riis JØ, Andersen G. ‘‘Pure alexia’’ without
hemianopia or colour anomia. Acta Neurol Scand 1988;78:501–5.
[10] Rosati G, De Bastiani P, Aiello I, Agnetti V. Alexia without
agraphia: a study of a case of verbal alexia without accompanying
colour-naming defect. J Neurol 1984;231:20–5.
[11] Johansson T, Fahlgren H. Alexia without agraphia: Lateral and
medial infarction of left occipital lobe. Neurology 1979;29:390–3.
` lesion
´
´
[12] Verstichel P, Cambier J. Alexie lettre a` lettre apres
hemisph´
´
erique
gauche sans hemianopsie
ni atteinte calleuse. Deux observations. Rev Neurol (Paris) 1997;153:561–8.
[13] Vincent FM, Sadowsky CH, Saunders RL, Reeves AG. Alexia
without agraphia, hemianopia, or color-naming defect: A disconnection syndrome. Neurology 1977;27:689–91.
[14] Greenblatt SH. Left occipital lobectomy and the preangular anatomy
of reading. Brain Lang 1990;38:576–95.

Y. Sakurai et al. / Journal of the Neurological Sciences 178 (2000) 42 – 51
[15] Kawamura M. ‘Non-classical type’ alexia without agraphia (in
Japanese). Shitsugosho Kenkyu 1988;8:185–93.
[16] Iwata M. Kanji versus Kana. Neuropsychological correlates of the
Japanese writing system. Trends Neurosci 1984;7:290–3.
[17] Kawamura M. Localization and symptomatology of pure alexia,
pure agraphia and alexia with agraphia (in Japanese). Shinkeishinrigaku 1990;6:16–24.
[18] Yamadori A. Alexia with agraphia (in Japanese). Shinkeinaika
1979;10:428–36.
[19] Sakurai Y, Matsumura K, Iwatsubo T, Momose T. Frontal pure
agraphia for kanji or kana: Dissociation between morphology and
phonology. Neurology 1997;49:946–52.
[20] Takada T, Sakurai Y, Takeuchi S, Sakuta M. Pure alexia due to a
fusiform gyrus lesion (in Japanese). Rinsho Shinkeigaku
1998;38:154–6.
[21] Damasio H. Human brain anatomy in computerized images, New
York: Oxford University Press, 1995.
[22] Dejerine J. Anatomie des centres nerveux I, Paris: RUEFF et Cie,
1895.
[23] Talairach J, Tournoux P. Co-planar stereotaxic atlas of the human
brain, Stuttgart: Thieme, 1988.
[24] Sakurai Y, Sakai K, Sakuta M, Iwata M. Naming difficulties in
alexia with agraphia for kanji after a left posterior inferior temporal
lesion. J Neurol Neurosurg Psychiatry 1994;57:609–13.
[25] Coltheart M, Patterson K, Marshall JC, editors, Deep dyslexia, 2nd
ed, London: Routledge, 1987.
[26] Sakurai Y, Takeuchi S, Kojima E, Yazawa I, Murayama S, Kaga K,
Momose T, Nakase H, Sakuta M, Kanazawa I. Mechanism of
short-term memory and repetition in conduction aphasia and related
cognitive disorders: a neuropsychological, audiological and neuroimaging study. J Neurol Sci 1998;154:182–93.
[27] Greenblatt SH. Localization of lesions in alexia. In: Kertesz A,
editor, Localization in neuropsychology, New York: Academic
Press, 1983, pp. 323–56.
[28] Torii H, Fukuta T, Koyama Y. A report of three cases of alexia
without agraphia dueto cerebrovascular lesion and semiological
consideration
(in
Japanese).
Seishinshinkeigaku
Zasshi
1972;74:546–76.
[29] Kawamura M, Hirayama K, Hasegawa K, Takahashi N, Yamaura A.
Alexia with agraphia of kanji (Japanese morphograms). J Neurol
Neurosurg Psychiatry 1987;50:1125–9.
[30] Soma Y, Sugishita M, Kitamura K, Maruyama S, Imanaga H.
Lexical agraphia in the Japanese language. Pure agraphia for kanji
due to left posteroinferior temporal lesions. Brain 1989;112:1549–
61.
[31] Croisile B, Trillet M, Laurent B, Latombe D, Schott B. Agraphie
´
´ gauche. Rev Neurol (Paris)
lexicale par hematome
temporo-parietal
1989;145:287–92.
[32] Buxbaum LJ, Coslett HB. Deep dyslexic phenomena in a letter-byletter reader. Brain Lang 1996;54:136–67.
[33] Alajouanine TH, Lhermitte F, Ribaucourt-Ducarne deBL. Les

51

alexies agnosiques et aphasiques. In: Alajouanine TH, editor, Les
´ du lobe occipital, Paris: Masson et Cie, 1960, pp.
grandes activites
235–60.
[34] Benson DF. Alexia. In: Frederiks JAM, editor, Handbook of clinical
neurology, Vol. 45, Amsterdam: Elsevier, 1985, pp. 433–55.
[35] Brown JW. Aphasia, apraxia and agnosia. Clinical and theoretical
aspects, Springfield: Thomas, 1972.
¨
[36] Hoff H, Gloning I, Gloning K. Uber
Alexie. Wien Z Nervenheilkunde 1954;10:149–62.
´
[37] Massary de J. L’alexie. L’Encephale
1932;1:53–78, 134–64.
[38] Ishiai S, Sugishita M, Lee E, Watabiki S, Nakayama T, Kotera M.
Processes of writing in alexia with agraphia and pure agraphia (in
Japanese). Shitsugosho Kenkyu 1993;13:264–71.
´
´ ´ verbale avec agraphie suivi
[39] Serieux
P. Note sur un cas de cecite
d’autopsie. Compt Rend Soc Biol 1892;4:13–8.
[40] Iwata M, Sugishita M, Kawai M, Yamashita N, Toyokura Y.
Agraphia and kinesthetic alexia. Disturbances of visual language
function due to the left angular gyrus lesion (in Japanese). Rinsho
Shinkeigaku 1979;19:462–8.
[41] Yamadori Y. Alexia with agraphia and angular gyrus (in Japanese).
Shitsugosho Kenkyu 1982;2:236–42.
[42] Kleist K. Gehirnpathologie, Leipzig: Barth, 1934.
[43] Sakurai Y, Momose T, Iwata M, Watanabe T, Ishikawa T, Takeda K,
Kanazawa I. Kanji word reading process analysed by positron
emission tomography. NeuroReport 1992;3:445–8.
[44] Sakurai Y, Momose T, Iwata M, Watanabe T, Ishikawa T, Kanazawa
I. Semantic process in kana word reading: activation studies with
positron emission tomography. NeuroReport 1993;4:327–30.
[45] Sakurai Y, Momose T, Iwata M, Ishikawa T, Sato T, Kanazawa I.
Regional cerebral blood flow in the covert reading of kana words: A
comparison with the study of reading aloud tasks. Eur Neurol
1996;36:237–9.
[46] Friston KJ, Frith CD, Liddle PF, Frackowiak RSJ. Comparing
functional (PET) images: The assessment of significant change. J
Cereb Blood Flow Metab 1991;11:690–9.
[47] Sakurai Y, Momose T, Iwata M, Sudo Y, Ohtomo K, Kanazawa I.
Different cortical activity in reading of Kanji words, Kana words
and Kana nonwords. Cogn Brain Res 2000;9:111–5.
[48] Rumsey JM, Horwitz B, Donohue BC, Nace K, Maisog JM,
Andreason P. Phonological and orthographic components of word
recognition. A PET-rCBF study. Brain 1997;120:739–59.
[49] Goldstein K, Gelb A. Psychologische Analysen hirnpathologischer
¨ auf Grund von Untersuchungen Hirnverletzter. Z ges Neurol
Falle
Psychiat 1918;41:1–142.
[50] Kawahata N, Nagata K. Alexia with agraphia due to the left
posterior inferior temporal lobe lesion — Neuropsychological
analysis and its pathogenetic mechanisms. Brain Lang 1988;33:296–
310.
¨
[51] Buchel
C, Price C, Friston K. A multimodal language region in the
ventral visual pathway. Nature 1998;394:274–7.