Article Text


Agraphia in Kanji after a contusional haemorrhage in the left temporo-occipital lobe
  1. S Maeshima1,
  2. A Ueyoshi1,
  3. T Matsumoto1,
  4. R Okita2,
  5. H Yamaga2,
  6. F Ozaki2,
  7. H Moriwaki2,
  8. P Roger3
  1. 1Department of Physical Medicine and Rehabilitation, Wakayama Medical University, Wakayama, Japan
  2. 2Division of Neurological Surgery, Hidaka General Hospital, Wakayama, Japan
  3. 3School of Communication Sciences and Disorders, University of Sydney, Australia
  1. Correspondence to:
 Dr S Maeshima;

Statistics from

The Japanese language has two different writing systems, Kana (Japanese syllabograms) and Kanji (Japanese ideograms), and Japanese sentences usually consist of combinations of both. Recently, there has been speculation that different pathways are involved in Kanji and Kana reading and writing.1 It has been suggested that the processing of Kanji and Kana involves different intrahemispheric mechanisms, as judged mainly through studies on patients with alexia and agraphia.2 Wernicke's area and its surrounding left middle temporal lobe may play the most important part in Kanji reading when visual information is transmitted by any pathway. We recently had the opportunity to examine a patient who had mild transient aphasia and persistent agraphia of Kanji after a discrete contusional haemorrhage in the left temporo-occipital lobe.

A 66 year old right handed man (pensioner) with 12 years of school education was involved in a traffic accident and sustained a closed head injury. He was admitted to our hospital the next day. He had no history of neurological problems. His parents, brothers, and sisters are all right handed. On examination, he was alert and cooperative. Neurological examination disclosed no motor or sensory disturbance.

Formal language assessment was undertaken 3 days after injury using the standard language test for aphasia (SLTA). His spontaneous speech was fluent. He had no difficulty initiating speech, articulated normally, and did not have logorrhoea. Echolalia was not seen and phonological structure was clear. Object naming was moderately impaired, with considerable paraphasia. Word fluency (animal naming) was 3/min. Repetition was excellent. Reading aloud was good; however, writing of both Kanji and Kana was impaired, and considerable paragraphia was evident.

In addition, he demonstrated constructional apraxia using figure copying of a three dimensional cube but not oral, ideomotor, or ideational apraxia. There were no signs of motor impersistence or unilateral spatial neglect. The score on Raven's coloured progressive matrices was 26/36.

Brain CT on admission showed an area of high density consistent with contusional haemorrhage in the left temporo-occipital lobe. Brain MRI 5 days after admission showed contusional haemorrhage in the left temporo-occipital lobe involving the posterior part of the temporal cortex and the adjacent white matter. There were no abnormalities in the left angular gyrus.

Aphasia was gradually resolved, his scores on SLTA 8 weeks after the onset becoming normal. Despite this, his writing ability remained clearly disturbed. We asked him to read and write on dictation 221 Kanji and 76 Kana, all of which are taught in the first 2 years of primary school in Japan. He was unable to write 83 of the 221 Kanji (37.6%), which he reported that he knew but could not recall. He tended to have difficulty with Kanji characters, which were more complex and had been learnt later (fig 1). Error types included partial substitution (two), partial completion (seven), neographism (29), and no response (45). On the other hand, he made no mistakes in reading Kanji or in reading and writing Kana letters. On the revised Wechsler adult intelligence scale (WAIS-R), his verbal intelligence quotient (IQ) was 105 and his performance IQ was 101, with an overall IQ of 103. Brain MRI 3 months after admission showed a subcortical lesion in the left temporo-occipital lobe (fig 2).

Although this patient experienced aphasia immediately after cerebral contusion, the aphasia resolved over a subsequent 8 week period. He was left with an isolated disturbance of Kanji writing. In the absence of any dementia, aphasia, or disturbance of consciousness, his condition was classified as one of pure agraphia.

Soma et al2 described “pure agraphia of Kanji” in three patients, who had a lesion in the left posterior temporal area extending to the angular gyrus on CT. Although they exhibited amnestic aphasia, alexia, and agraphia in the acute phase, disturbances other than agraphia of Kanji disappeared in a few months. Yokota et al3 also reported a case with pure agraphia of Kanji, and suggested that the process of writing Kanji involves a different pathway from that which mediates Kanji reading in the left temporal lobe.

Iwata1 proposed a hypothetical neuronal mechanism in the writing of Kanji and Kana as follows: the Broca's and motor association areas are the final coordination centres, but the angular gyrus plays an important part in sending the graphic information to the motor areas. However, spontaneous writing and responses to dictation are usually initiated by Wernicke's area, which gives rise to two different pathways to the angular gyrus. One is the auditory somaesthetic association pathway leading directly from Wernicke's area to the angular gyrus; Kana writing mainly depends on the intactness of this route. The second pathway is from Wernicke's area to the occipital lobe by way of the posterioinferior temporal area. This is the pathway involved in selecting the correct Kanji graphemes according to the meaning of the word, and thus recalled visual engrams of letters are sent to the angular gyrus.

Our patient demonstrated that initial amnestic aphasia, and agraphia of Kanji were associated with a lesion in the left temporal lobe. Because the posterior temporal region is located in close proximity to the angular gyrus and Wernicke's area, the pathological process in the first area affects the second two regions in its acute phase. We conclude that the persistent symptom of pure agraphia for Kanji in this study was caused by the left posterior temporal lesion which disconnected the pathway for Kanji writing selectively.

Figure 1

Examples of dictation. This patient showed intact Kana writing and difficulties in Kanji writing: 1, no response; 2, partial completion; 3, partial substitution; 4, neographism; 5, correct.

Figure 2

Brain MRI 3 months after admission showed the subcortical lesions in the left temporo-occipital lobe.


View Abstract

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.