OBSERVATION

Phonological Agraphia After Superior Temporal
Gyrus Infarction
Han-Joon Kim, MD; Kon Chu, MD; Kyoung-Min Lee, MD, PhD; Dong Wook Kim, MD; Seong-Ho Park, MD, PhD

Background: Phonological agraphia refers to a condi-

tion in which the ability to write nonwords to dictation
is impaired, while writing words to dictation is preserved, as is oral repetition of the words and nonwords.
This condition has been regarded as reflecting a disconnection within the phonological writing system, and
previous neurolinguistic correlations suggested that the
anterior-inferior supramarginal gyrus was a crucial link
within the system.
Setting: A neurology department of a university hospital.
Patient: A 51-year-old right-handed man presented with

speech disturbances. On initial evaluation of his language, his deficit was consistent with that of conduction aphasia, which improved rapidly to an apparently

P

From the Department of
Neurology and the Clinical
Research Institute, Seoul
National University Hospital,
Neuroscience Research Institute
of SNUMRC (Drs H.-J. Kim,
Chu, Lee, D. W. Kim, and
Park), and the Department of
Neurology, Boramae City
Hospital (Dr Park),
Seoul, Korea.

normal level. A subsequent detailed examination of oral
and written repetition of words and nonwords revealed
a rather selective form of phonological agraphia. A magnetic resonance imaging scan of his brain showed a
focal ischemic lesion at the left posterior superior temporal gyrus and at the underlying white matter.
Conclusions: In contrast to most previously described
patients, this patient showed a selective impairment of
phonological agraphia in association with a focal infarction restricted to the left posterior superior temporal gyrus, suggesting that this region of the brain is an important node within a wider network of areas that subserve
the phonological route for writing.

Arch Neurol. 2002;59:1314-1316

HONOLOGICAL AGRAPHIA is a

spelling disorder characterized by the inability to write
nonwordstodictation,whereas
theabilitytowritewordstodictation and to orally repeat words and nonwords is intact.1 This condition is thought
to be caused by a selective impairment in
associating input phonological representations with output orthographic representations and, thereby, a forced reliance on
lexical memory for accessing orthographic
word-form representations.2 Because persons with phonological agraphia rely on
lexical information in writing, they usually show errors in writing affixes and nonwords, which presumably are not stored in
the mental lexicon, and in writing lowfrequency words, for which the access to the
lexicon is less efficient.2
Classically, the anterior-inferior part
of the left supramarginal gyrus has been
pointed to as the lesion site underlying
phonological agraphia.3 However, there
is considerable variability in the clinical
manifestations among described patients.1,3-8 Some patients had additional difficulties in oral repetition of nonwords,3
and others showed serious problems in
writing real words and nonwords to dictation.3,4,7 The patient described by Bud and

(REPRINTED) ARCH NEUROL / VOL 59, AUG 2002
1314

Kertesz6 showed significantly higher performance in writing words than nonwords to dictation, but oral naming was
extremely poor as well. The reliability of
lesion localization in described patients is
also quite variable. Some reports in the
pre–magnetic resonance imaging era used
computed tomographic scans, which
tended to be less sensitive than magnetic
resonance imaging scans in demarcating
the lesion extent. The variability of clinical findings and lesion localization in previous reports has hindered a more precise symptom-lesion correlation. We
recently encountered a patient with pure
phonological agraphia after infarction in
the left posterior superior temporal gyrus (pSTG). The selective impairment and
precise lesion localization observed in this
patient may lead us a step further in understanding the neurolinguistic mechanisms for writing.
REPORT OF A CASE

CASE DESCRIPTION
A 51-year-old right-handed Korean man
who had been healthy, except for a 2-year
history of well-controlled hypertension,
presented with speech disturbance. Three

WWW.ARCHNEUROL.COM

©2002 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a George Mason University User on 03/17/2016

Fluid-attenuated inversion recovery images obtained on the first hospital day show the high signal intensities restricted within the left posterior superior temporal
gyrus and the underlying white matter.

days before hospital admission, he found himself speaking not as fluently as before. He had difficulties in finding appropriate words and initiating sentences. However, his difficulties were not so severe as to cause
problems in usual conversation. Comprehension was not
affected. On neurological examination on hospital admission, he was alert, with intact orientation. Verbal fluency in spontaneous speech was preserved, except for occasional interruptions due to word-finding difficulties,
and no paraphasia was observed. He could obey 3-step
commands perfectly, suggesting grossly normal auditory comprehension. His spontaneous and dictational
writing of words and short sentences showed no error
in spelling. His reading comprehension was intact, even
with rather long and complex sentences. The only deficit observed in language function was an impairment of
repetition. No abnormality was found in tests of other
cognitive functions, such as praxis, left-right orientation, calculation, finger naming, and spatial attention.
Other parts of the neurological examination, including
examination of the cranial nerves, motor and sensory functions, and reflexes, were unremarkable.
The results of routine blood tests were all within normal limits. A transcranial Doppler study revealed a mild
increase in the flow velocity at the proximal trunk of the
left middle cerebral artery. Magnetic resonance imaging
of the brain performed on the second hospital day revealed a hyperintense lesion in the left pSTG and underlying white matter on T2-weighted and fluid-attenuated
inversion recovery scans (Figure). Magnetic resonance
angiography showed decreased vascularity distal to the
left middle cerebral artery bifurcation.
FURTHER ASSESSMENT OF LANGUAGE
FUNCTION
Language function was further assessed with the Korean version of the Western Aphasia Battery on the fourth
hospital day. While spontaneous speech, auditory comprehension, and naming remained intact, the patient’s oral
repetition was significantly improved in that he had difficulties only in repeating sentences composed of 5 or
more words. His repetition of single words or shorter sentences was normal. His aphasia quotient was 98.6.
Whereas his initial language deficit was definitely compatible with conduction aphasia, the result of the Korean version of the Western Aphasia Battery on the fourth
(REPRINTED) ARCH NEUROL / VOL 59, AUG 2002
1315

hospital day revealed a performance within normal range
using quantitative criteria.
On the same day, to evaluate his performance in repetition more thoroughly, we administered more detailed tests of repetition in auditory and written forms.
First, the patient was asked to orally repeat 15 highfrequency words, 15 low-frequency words, and 15 nonwords immediately after each item was spoken to him.
The syllable length of the words and nonwords ranged
from 1 to 4. The patient repeated all words and nonwords correctly. Second, he was asked to write the words
and nonwords as soon as he heard them. He correctly
wrote all 15 high-frequency words, 14 low-frequency
words, and 7 nonwords. From these tests, selective impairment of writing nonwords to dictation was observed, compatible with phonological agraphia. He wrote
nonword syllables that were similar but different from
what was spoken to him, while he wrote words correctly. Furthermore, he made errors only in dictational
writing, but not in oral repetition. Although no normal
control data are available, a person of his education and
job level would be expected to perform perfectly on these
tests. From a detailed inspection of his errors in writing
nonwords to dictation, it was observed that all but
1 errors were made on second or later syllables of the
nonwords.
COMMENT

Three parallel routes may be involved in writing: a phonological route, a lexical route, and a semantic route.2,3
In writing nonwords, only the phonological route can be
used because nonwords have no lexical or semantic representation. Roeltgen et al3 suggested that the phonological route is mediated by 2 sequential processes: (1)
a segmentation process, in which syllable sequences of
words or nonwords are broken down into phonemes; and
(2) the phoneme-to-grapheme conversion, in which the
segmented phonemes are translated into graphemes. They
suggested that selective loss of either of these 2 processes or both could lead to phonological agraphia. We
could not determine definitely which of the 2 components was impaired in our patient because we did not
evaluate his ability to write a single phoneme to dictation, which would be impaired in the case of a phonemeto-grapheme conversion defect. However, from the finding that our patient erred mostly at the second or later
WWW.ARCHNEUROL.COM

©2002 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a George Mason University User on 03/17/2016

syllables of the presented nonwords, we suggest that his
deficit was more of segmentation than of phoneme-tographeme conversion. Presumably, if the deficit had been
in phoneme-to-grapheme conversion and not in segmentation, he would have erred regardless of the position of
a phoneme within the nonwords.
For neurolinguistic correlates, the left anteriorinferior supramarginal gyrus has been suggested as the
anatomical substrate for phonological agraphia,3 because it was the only area that was damaged in all patients when their computed tomographic scans were superimposed. However, the 4 patients of the study showed
considerable anatomical and clinical variability, and for
each patient, the lesion extended over a larger area into
the frontal operculum, STG, postcentral gyrus, and/or superior parietal lobule. Alexander et al9 suggested that, from
the analysis of their 2 cases of phonological agraphia and
other published cases, phonological agraphia could be
produced by lesions in a wide range of perisylvian cortical regions. Phonological agraphia following a large right
frontal infarction, superior temporal and anterior parietal destruction,5 an extensive left perisylvian infarction,1 a left insula and basal ganglia infarction,7 and a focal left anterior insulo-opercular infarction4 has also been
reported. However, as previously noted, all these patients showed language disturbances beyond the restrictive criteria of phonological agraphia, and lesions in most
patients were quite extensive and presumably encompassed neural substrates of many linguistic functions. In
contrast to these patients, our patient had a selective impairment limited only to writing nonwords to dictation
and a restricted lesion in the left pSTG and underlying
white matter, indicating the importance of the area in writing via the phonological route. However, the route was
not completely destroyed in the patient because he was
able to write about 47% of the nonwords to dictation. This
partial impairment may have resulted because of either
incomplete damage to the left pSTG or the widely distributed nature of the network areas involved in the function, as suggested by the published reports previously
mentioned.
The role of the left pSTG in language function has
been suggested by reports10,11 that a pSTG lesion was associated with conduction aphasia and by studies12,13 that
showed that pSTG stimulation produces aphasic symptoms comparable to conduction aphasia. According to
these studies, pSTG dysfunction leads to impairments of
phonological output lexicon,12 phonological loading process during speech production,13 or phonological judgment of presented stimuli,14 which then may result in conduction aphasia. Contrary to these reports, our patient
showed no impairment in phonological input processing and phonological output processing, as evidenced by
normal performance on oral repetition of nonwords. The
deficit in our patient was limited to phonological writing to dictation. We believe this discrepancy between our
case and the previously reported cases of conduction aphasia is because of a difference in the exact location of the
lesions within the left pSTG. Given that the region seems
to be operative in a wider range of phonological processes, not limited to phonological output in auditory modality, it may as well be composed of several subdivi(REPRINTED) ARCH NEUROL / VOL 59, AUG 2002
1316

sions with roles slightly different from each other. This
possibility of functional parcellation of the area, however, requires further studies armed with better spatial
resolution than was possible in the present study.
In summary, the present case, in which a selective
lesion in the left pSTG was associated with phonological agraphia with little or no impairment in other aspects of the linguistic functions, indicates that this area
plays quite an important role in phonological linkage of
auditory input with orthographic output, as needed for
writing nonwords to dictation.
Accepted for publication February 27, 2002.
Author contributions: Study concept and design (Drs
H.-J. Kim, Chu, Lee, and Park); acquisition of data (Drs
H.-J. Kim and D. W. Kim); analysis and interpretation of
data (Drs H.-J. Kim, Chu, Lee, and D. W. Kim); drafting
of the manuscript (Drs H.-J. Kim and Lee); critical revision of the manuscript for important intellectual content (Drs
Chu, Lee, D. W. Kim, and Park); administrative, technical, and material support (Drs H.-J. Kim, Lee, D. W. Kim,
and Park); study supervision (Drs Chu, Lee, and Park).
This study was supported in part by grant R01-200000167 from Korea Science and Engineering Foundation.
Corresponding author and reprints: Kyoung-Min Lee,
MD, PhD, Department of Neurology, Seoul National University Hospital, 28 Yongon-Dong Chongno-gu, Seoul 110744, Korea (e-mail: kminlee@snu.ac.kr).
REFERENCES
1. Shallice T. Phonological agraphia and the lexical route in writing. Brain. 1981;
104:413-429.
2. Martin N. Recovery and treatment of acquired reading and spelling disorders.
In: Stemmer B, Whitaker H, eds. Handbook of Neurolinguistics. San Diego, Calif:
Academic Press Inc; 1998:559-572.
3. Roeltgen DP, Sevush S, Heilman KM. Phonological agraphia: writing by the lexicalsemantic route. Neurology. 1983;33:755-765.
4. Marien P, Pickut BA, Engelborghs S, Martin J, De Deyn PP. Phonological agraphia
following a focal anterior insulo-opercular infarction. Neuropsychologia. 2001;
39:845-855.
5. Bolla-Wilson K, Speedie LJ, Robinson RG. Phonological agraphia in a lefthanded patient after a right-hemisphere lesion. Neurology. 1985;35:17781781.
6. Bud D, Kertesz A. Evidence for lexicographic processing in a patient with preserved written over oral single word naming. Brain. 1982;105:697-717.
7. Kim H, Na DL. Dissociation of pure Korean words and Chinese-derivative words
in phonological dysgraphia. Brain Lang. 2000;74:134-137.
8. Iribarren IC, Jarema G, Lecours AR. Two different dysgraphic syndromes in a
regular orthography, spanish. Brain Lang. 2001;77:166-175.
9. Alexander MP, Friedman RB, Loverso F, Fischer RS. Lesion localization of phonological agraphia. Brain Lang. 1992;43:83-95.
10. Benson DF, Sheremata WA, Bouchard R, Segarra JM, Price D, Geschwind N. Conduction aphasia: a clinicopathological study. Arch Neurol. 1973;28:339-346.
11. Damasio H, Damasio A. The anatomical basis of conduction aphasia. Brain. 1980;
103:337-350.
12. Anderson JM, Gilmore R, Roper S, et al. Conduction aphasia and the arcuate
fasciculus: a reexamination of the Wernicke-Geschwind model. Brain Lang. 1999;
70:1-12.
13. Hickok G, Erhard P, Kassubek J, et al. A functional magnetic resonance imaging
study of the role of left posterior superior temporal gyrus in speech production:
implications for the explanation of conduction aphasia. Neurosci Lett. 2000;287:
156-160.
14. Fujimaki N, Miyauchi S, Putz B, et al. Functional magnetic resonance imaging of
neural activity related to orthographic, phonological, and lexico-semantic judgments of visually presented characters and words. Hum Brain Mapp. 1999;8:
44-59.

WWW.ARCHNEUROL.COM

©2002 American Medical Association. All rights reserved.
Downloaded From: http://archneur.jamanetwork.com/ by a George Mason University User on 03/17/2016