Journal of Neurology, Neurosurgery, and Psychiatry 1994;57:649-651

649

SHORT REPORT

Alphabetical paragraphia in a limited middle
cerebral artery stroke
Martin Catala, Bertrand Fontaine, Gerald Rancurel

Abstract
A Yugoslavian perfectly bilingual for
French and Serb had a limited left
middle cerebral artery stroke. He developed a peculiar dysgraphia characterised
by the use of Latin characters (French
spelling) to transcribe Serb phonemes
that would normally have been spelt in
the Cyrillic alphabet. This dysgraphia
was likely to be due to an impairment of
the allographic procedure. It is concluded that allographs of the two alphabets are produced concomitantly in
bialphabetical patients during the allographic procedure.
(3 Neurol Neurosurg Psychiatry 1994;57:649-65 1)

Services de

Neurologie, Hopital
de La Salpetriere,
47, Bd de l'Hopital
75651 Paris Cedex 13
France
M Catala
B Fontaine
G Rancurel
Correspondence to:
Dr Martin Catala,
Service du Pr Brunet,
H6pital de La Salpetriere,
47, Bd de l'H6pital,
75651 Paris Cedex 13,
France.
Received 3 August 1992
and in final revised form
1 June 1993.
Accepted 27 July 1993

Writing is a complex procedure that has
different sequential stages. During the first
part of the process, each word is transformed
into a sequence of abstract letters. This transformation may be produced by two different
routes in alphabetical languages': the phonological route and the lexical route. These
routes may be impaired by damage to the
nervous system, giving rise to a central
dysgraphia. The second part of the process
allows the transformation of each abstract or
conceptual letter form, namely the grapheme,
(for example, "F") into fonts (or allographs)
that may be executed in various ways (for
example, F,fJ). This step is produced letter
by letter and the process is still poorly understood. There are at least three successive
steps: the sequence of graphemes is
temporarily stored in a working memory
device called the graphemic buffer; during the
allographic procedure, each grapheme is
transformed into allographic forms, one of
which will be chosen according to the context
and the position of the letter in the string;
and, finally, the adequate allograph is
converted into a motor plan. 1-3

In fact, the actual processing of written text
may be more complex, the different groups of
characters (numbers v letters, vowels v consonants) having different status.4 Indeed, two
patients with selective paragraphia exclusively
concerning vowels have already been
reported.5 The author suggested that vowels
and consonants were produced by two different cerebral systems, but he was unable to

recognise a unique structure in the CNS
devoted to the processing of vowels, as the
first patient had a left parietal infarct and the
second a left frontal infarct. In this paper we
report a case of a Yugoslavian man with a left
frontal infarct associated with a selective
dysgraphia. This clinical presentation has
been analysed in terms of the mechanisms
involved in the allograph selection for a
particular grapheme.
Case report
A 52-year-old right-handed man, born in
Yugoslavia and living in France for 25 years,
perfectly bilingual in Serb and French, was
admitted to hospital for a rapidly reversing
right brachiofacial hemiparesis with aphasia.
He had smoked two packs of cigarettes daily
for 15 years and had stopped smoking five
years before admission. He had mild
untreated hypertension.
The evening before admission to hospital,
he had a sudden right hemiparesis and aphasia. Motor disturbance predominated in the
brachiofacial territory and disappeared within
30 minutes. The oral language deficit was
characterised by a non-fluent aphasia, both in
French and in Serb, and impaired naming.
There was no paraphasia or comprehension
difficulty. The language deficit cleared within
three hours. The next morning, neurological
examination disclosed a mild dysarthria and
brisk right sided reflexes without Babinski
sign. Sensation and visual fields were normal.
Two days after admission the patient
complained of a peculiar dysgraphia noted
while solving Serbian crosswords. Short
words (one or two syllables) were correctly
written, whereas long words were written
with literal paragraphia. For example, he
wrote KOHCTYCION for KOHCTYCI4OH
(constitution), HARO,IUN14 for HAPOIHIU4
(nation), IIYTOVATI4 instead of HIYTOBATI4 (to travel). Analysis of the paragraphia
was performed on 30 words and showed that
they corresponded to the substitution of a
Cyrillic character by its Latin counterpart.
Only three pairs of characters were concerned: ((B,V), (P,R), (H,N)). He was aware
of his errors and could correct them after
reading his words. Praxis, reading, and oral
spelling (both in Serb and in French) were
normal. The dysgraphia lasted four days.

650

Catala, Fontaine, Rancurel

((B,V), (H,N), (P,R), (C,S)) corresponds to
pairs of characters in which the Cyrillic one

can be used as a Latin character with a different phonological correspondence. Literal substitution, in our patient, belonged to this last

Figure Cerebral CT showing leftfrontal subcortical
hypodensity in the distribution of the superficial middle
cerebral artery.

Cerebral CT without contrast injection
showed a hypodensity in the left frontal subcortical territory under the middle frontal
gyrus6 in the distribution of the left superficial
middle cerebral artery (figure). Ultrasonic
angiology of the cervical arteries did not show
any significant vascular lesion. Transthoracic
echocardiography revealed a non-obstructive
cardiomyopathy with low left ventricular
flow. Cardiac catheterisation confirmed a
non-obstructive cardiomyopathy with moderate impairment of left ventricular function
and showed a left intraventricular thrombus.
It was concluded that the stroke was likely
due to an embolism from the intraventricular
thrombus.
Discussion
Our patient was Serb and wrote with the
Cyrillic alphabet. After arriving in France 25
years previously, he learned French and the
Latin alphabet and became bilingual and
"bialphabetical". Consequently, phonemes
common to Serb and French corresponded to
both Cyrillic and Latin characters.
Our patient presented a "pure agraphia"
characterised by literal substitutions; he wrote
the Latin character instead of the Cyrillic
one. For phonemes common to both
languages, the correspondence between
Cyrillic and Latin alphabets allows the pairs
of characters to be classified into three
groups. The first group (for example, (,,D),
(14,I), ('D,F)) corresponds to pairs of characters in which the Cyrillic one is specific and
never used in the latin alphabet. The second
group (for example, (A,A), (K,K), (M,M))
corresponds to characters used both in
Cyrillic and Latin writing. The third group

group and only implicated consonants. The
specificity for consonants is explained by the
table of correspondence between the two
alphabets and not by an impairment of a
putative device specific for consonants. The
substitutions concerned pairs of characters
(Cyrillic and Latin) that were evoked by the
same grapheme, and we consider that they
represent different allographs of the same
grapheme. Our case suggests that Cyrillic and
Latin allographs are generated concomitantly
by the same grapheme during the allographic
procedure. This type of substitution corresponds to a peripheral agraphia, the problem
lying within the allographic procedure. 1-3
Because of their production, we termed these
errors alphabetical paragraphia. Such allographic substitutions have already been noted
within a more global writing disorder: A
patient reported by Katz7 often intermixed
upper and lower letters within a word and a
patient reported by Kartsounis8 wrote L for 1
and Q for q. The localisation of the lesion in
these two cases was not stated. Paragraphia
appeared for long words (three or more syllables) as occurred for the patients reported
by Katz7 and Piccirilli et al.9 The increase in
error rate with the length of the words was
considered by these authors to be an indicator of mnesic impairment involving the
graphemic buffer. The literal substitutions by
our patient could be explained by a mnesic
disturbance affecting the device devoted to
the selection of the correct allographic form.
His lesion was located in the subcortical
white matter under the middle frontal gyrus.
Frontal lesions have been associated with
pure agraphia.4 Our case was similar to that
reported by Albert and Obler.10 They
described a woman bilingual for Hebrew and
English who wrote Hebrew words containing
some English letters after bilateral frontal
damage caused by a car accident. They did
not state, however, if there was any correspondence between the English and Hebrew
characters.
This type of literal paragraphia occurred in
patients using two different types of alphabet.
By contrast, in Japanese patients or in
bilingual patients who use two completely
different systems for writing, such as syllabic
and ideographic characters, this type of
paragraphia does not occur. The grapheme
production is independent for the two
systems and dysgraphia involves predominantly one of the two systems.1' 12 It
would be interesting to study such patients
who use different writing systems (alphabetical systems of the Occidental languages,
alphabetical systems in which vowels and
consonants have different status such as the
Semitic languages, ideographic systems of the
Asiatic languages) to analyse the cerebral
connections between these different types of

writing.

Alphabetical paragraphia in a limited middle cerebral artery stroke
We thank Pete Jeffs, PhD, Nicholas K Gonatas, MD, and
Nathalie Kubis, MD for critical reading of the manuscript.
1 Bub D, Chertkow H. Agraphia. In: Boller F, Grafman J,
eds. Handbook of neuropsychology. Vol 1. Amsterdam:
Elsevier Science Publishers, 1988:393-414.
2 Ellis AW. Spelling and writing (and reading and speaking). In: Ellis AW, ed. Normality and pathology in cognitive functions. London: Academic Press, 1982:113-46.
3 Shallice T. From neuropsychology to mental structure.
Cambridge: Cambridge University Press 1988:130-57.
4 Anderson SW, Damasio AR, Damasio H. Troubled letters
but not numbers. Domain specific cognitive impairments following focal damage in frontal cortex. Brain

1990;113:749-66.
5 Cubelli R. A selective deficit for writing vowels in
acquired dysgraphia. Nature 1991;353:258-60.

651
6 Nieuwenhuys R, Voogd J, Van Huijzen C. The human
central nervous system. Berlin: Springer-Verlag, 1988.
7 Katz RB. Limited retention of information in the
graphemic buffer. Cortex 1991;27:111-9.
8 Kartsounis LD. Selective lower-case letter ideational dysgraphia. Cortex 1992;28:145-50.
9 Piccirilli M, Petrillo S, Poli R. Dysgraphia and selective
impairment of the graphemic buffer. Ital J7 Neurol Sci
1992;13:113-7.
10 Albert ML, Obler K. The bilingual brain. New York:
Academic Press, 1978.
11 Lyman RS, Kwan ST, Chao WH. Left occipito-parietal
brain tumor with observations in alexia and agraphia.
Chin MedJ3 (Engl) 1938;54:491-516.
12 Sasanuma S, Fujimara C. Selective impairment of processing phonetic and non-phonetic transcriptions of
words in aphasic patients: Kana and Kanji in visual
recognition and writing. Cortex 197 1;7:196-218.