NOTE
OPTIC APHASIA WITH PURE ALEXIA: A MILD FORM
OF VISUAL ASSOCIATIVE AGNOSIA? A CASE STUDY
V. Chanoine1,3, C. Teixeira Ferreira2,4, J.F. Demonet1, J.L. Nespoulous1,3 and
M. Poncet2
(1INSERM U455, Services de Neurologie, CHU Purpan, Toulouse, France; 2INSERM CJF
9706, Laboratoire de Neurophysiologie et de Neuropsychologie, Faculté de Médecine,
Marseille, France; 3Laboratoire Jacques Lordat, Maison de la Recherche, Toulouse, France;
4Department of Neurology, HUCFF, Universidade Federal do Rio de Janeiro, Brasil)

ABSTRACT
A single-case study is reported of a naming disorder selective to the visual modality.
The patient showed intact access to structural knowledge of objects and letters, but impaired
access to complete semantic knowledge of objects and alphabetical knowledge of letters
from visual input. The impairment was most striking when the patient had to discriminate
between semantically similar objects or within a given symbolic repertoire, i.e. letters. The
co-occurrence of a partial deficit of visual recognition for objects and for letters indicated
features of optic aphasia and pure alesia. This symmetric performance between object and
letter processing may also constitute a mild form of visual associative agnosia.
Key words: optic aphasia, pure alexia, associative agnosia
INTRODUCTION
In 1889, Freund first reported a case of “optic aphasia” which resulted in the inability
to name visually presented objects whilst the patient could still name objects on tactile
presentation. Optic Aphasia, as defined by Freund, might relate to either a visual recognition
deficit (“visual agnosia”), or to a general naming impairment commonly observed in
aphasia. Following the detailed description of another case by Lhermitte and Beauvois
(1973), optic aphasia became a widely accepted syndrome. Their patient, JF, showed a
naming impairment confined to the visual presentation of objects, but his deficit was neither
agnosic, since he produced correct pantomimes of the objects he did not name, nor aphasic,
since he could name objects from tactile or verbal inputs. Thus, optic aphasia was defined,
in negative terms as a deficit in visual confrontation naming, without visual recognition or
perceptual and intellectual disorders.
However, the interpretation of optic aphasia is not easy. The main reason is probably
due to the fact that classical models of visual naming (e.g. Seymour, 1979; Warren and
Morton, 1982) cannot account for a modality-specific aphasia such as optic aphasia (see
Shallice, 1988; Farah, 1990; Bruyer, 1994). Based on Marr’s computational theory (e.g.
Marr, 1982), these models generally include four successive stages: (1) perceptual
processing; (2) access to structural representation; (3) access to semantic knowledge and
(4) access to the object’s name. The perceptual processes permit the detection of salient
visual features from which the observer is assumed to automatically construct an episodic
form representation of the perceived object. The episodic representation crucially depends
on contextual conditions, such as viewpoint, and it must be compared to an abstract
structural representation which stores visual attributes specific to an object in memory. At
the structural level, the same object can be recognised from different viewpoints (e.g.
canonical and non-canonical views). The semantic attributes of an object are identified in
Cortex, (1998) 34, 437-448

438

V. Chanoine and Others

the third stage. Finally, following access to semantic knowledge, object naming is
accomplished by accessing the lexical representation that matches the object. Most cognitive
models of object naming (e.g. Seymour, 1979; Warren and Morton, 1982) suppose that
access to the object’s name entails an activation of the entire semantic representation
attached to it. For example, the complete semantic representation of a hammer should
specify the categorical (e.g. tool), functional (e.g. used to drive in), and associative (e.g. a
nail) attributes. Some authors (e.g. Warren and Morton, 1982; Snodgrass, 1984) presume
that the semantic system is amodal, in that the access to semantic information is independent
of the input modality.
The error analysis of visual object naming in optic aphasic patients reveals a majority
of semantic errors (for an extensive review, see Iorio et al., 1992). These semantic
paraphasias observed on visual presentation only can hardly be explained by classical
models involving amodal semantic representation. Other models of object naming have
been proposed (e.g. Lhermitte and Beauvois, 1973; Ratcliff and Newcombe, 1982; Riddoch
and Humphreys, 1987a; Coslett and Saffran, 1989) in order to account for the behavioural
pattern of optic aphasics. These different models challenge (1) the idea of a sigle semantic
representation for objects (semantic representation being rather multiple, such as visual vs.
verbal semantics in Lhermitte and Beauvois, 1973, or left vs. right hemisphere semantics in
Coslett and Saffran, 1989; De Renzi and Saetti, 1997), (2) the uniqueness of the semantic
route in visual naming (e.g. Ratcliff and Newcombe, 1982), or (3) the necessity to access
complete semantic information in naming (e.g. Riddoch and Humphreys, 1987a, 1987b;
Hillis and Caramazza, 1995). The last two models dispute the integrity of visual object
recognition in optic aphasia. In other words, patients diagnosed optic aphasics should be
reinterpreted as cases of visual associative agnosia (Lissauer, 1890). For example, in a
recent case study reported by Hillis and Caramazza (1995), DHY, “labelled optic aphasic”,
failed to access detailed semantic knowledge in non-verbal association semantic tasks,
casting some doubt on the integrity of his visual recognition processes. Moreover, most
cases of optic aphasia (for a review, see Davidoff and De Bleser, 1993) and associative
agnosia (for a review, see Farah, 1991) presented a co-occurrence of visual anomia for
objects and letters. Two questions then arise, whether the object and letter visual naming
disorder shares a common mechanism, and whether optic aphasia has to be considered as a
mild form of visual associative agnosia.
In this paper, we report the detailed investigation of a patient, CN, who, following left
occipito-temporal infarction, showed the pattern of performance labelled as optic aphasia,
as well as a pure alexia.

CASE REPORT
CN is a right-handed 67-year-old retired boiler maker with an eight grade of education.
He was admitted to the hospital in February 1995 for generalised seizure episodes probably
secondary to a stroke that occurred in 1979 in the territory of left posterior cerebral and left
anterior choroidial arteries. His past medical story was an arterial hypertension and an
ischemic heart disease. A general neurological examination revealed a right spastic
hemiplegia and hemianesthesia, as well as a dense right homonymous hemianopia. Magnetic
Resonance Images (MRI; see Figure 1) indicated a left posterior cerebral artery infarction,
with damage to the left lingual, fusiform and parahippocampal gyri. The lesion extended
into the posterior portion of the internal capsule and thalamus on the left side. In addition,
there was a paracallosal lesion. In the occipital portion, where this lesion was larger, it
seemed to completely interrupt the intra-hemispheric part of callosal fibers.
A general neuropsychological assessment (Boller and Hécaen, 1979) was given by the
time of the current experimental study (March 1996). Testing excluded global deterioration,
as demostrated by his performance on most tests of verbal and non-verbal intelligence,
attention and visuo-spatial capacities (see Table I). His spontaneous language did not reveal
syntactic and morphologic disorders, except for occasional word-finding pauses. Wordfinding difficulties were more obvious in visual naming.

Optic aphasia with pure alexia

439

Fig. 1 – Representative horizontal MRI cross-sections of patient CN.

TABLE I

CN’s Performance in General Neuropsychological Assessment

Intellectual and executive functions
WAIS similairities
Raven Progressive Matrices
Mini Mental State (MMS)
Simple and complex orders
Attention
Letters barring
Memory
Story recall (Wechsler)
Digit spain
Spatial span (Corsi)
Verbal evocation
Morphologic and semantic contraries
Automatic series: numbers and months
Gnosis
Body parts
Colours
Famous faces (photographs)
Complex pictures (photographs)
Visuo-spatial capacities
Space orientation
Left-right orientation
Constructive apraxia

Accuracy score

Judgement

9/12
34/36
24/30
17/20

–
–
++
–

33/33

–

11/24
4
3

++
–
+

5/6
4/4

–
–

4/5
2/10
8/10
5/5

–
+++
+
–

Note: normal performance: –; mild deficit: +; moderate deficit: ++; severe deficit: +++
(All the tests mentioned are described in Boller and Hécaen, 1979)

–
–
–

440

V. Chanoine and Others

In picture naming of different objects, the accuracy score was about 67% (10/15) and
error analysis mainly indicated confusions in the semantic domain (e.g. mask → a cowl;
sugar tongs → nutcracker). However, errors in picture naming were not present for very
frequent objects (accuracy scores: 5/5 high frequency objects; 3/5 middle frequency objects;
2/5 low frequency objects). Picture recognition of the same objects among distractors on
spoken-name presentation was perfect (15/15).
On visual confrontation with common use objects, his responses, althought roughly
correct (accuracy score: 9/10), revealed a problem in lexical access (e.g. knife → “I can
say it... It’s something, one tooth, teeth... One thing made for this”). Naming of the same
real objects from palpation was perfect (10/10).
As far as written language was concerned, CN matched the diagnosis of pure alexia.
He was totally unable to read words, non-words and sentences. His reading deficit also
affected written language comprehension. However, spontaneous writing or writing from
dictation were relatively preserved except for some perseverative phenomena (e.g.
/courtyard/: /cour/ → cowr). Writing from copy was servile: he copied each letter, featureby-feature, as if he were drawing a picture. Furthermore, CN was unable to read what he
had correctly written, as is typical forms of pure alesia.
CN’s spelling was relatively spared, except for some irregular words (e.g. /oursin/ →
O.U.C.I.N) and his repetition was normal.
Abnormalities in colour processing were noticed: (1) in colour naming from visual
presentation: 2/5, (2) in colour designation from verbal order (e.g. “Show me the red...”):
1/5 and (3) in colour designation from name of objects (e.g. “Show me the colour of
blood...”: 1/5. However, CN correctly named the colour of the same objects from verbal
description (e.g. What is the colour of blood?): 5/5.
Although CN was not prosopagnosic, he was impaired in naming photographs of famous
faces (accuracy score = 2/5) though knowing their profession (e.g. Jean Gabin → an actor)
and in matching them with their spoken name (accuracy score: 2/5).
EXPERIMENTAL STUDY
An experimental study was conducted to identify the nature of CN’s visual naming
disorder and to determine whether object and letter processing deficits shared a common
mechanism.
Object Processing Assessment
In this section, we focus on CN’s performance on tasks designed to identify the
cognitive mechanism underlying naming from visual stimuli. This analysis involved 4 types
of tasks: (1) naming tasks; (2) assessment of elementary visual perception; (3) access to
structural representation from vision (or “recognition of visual attributes”) and (4) access to
semantics from vision (or “recognition of semantic attributes”). Detailed analysis of object
naming errors were provided in a separate paper (see Teixeira Ferreira et al., 1997).
Naming Tasks
Naming assessment included: (1) object naming from the visual modality; (2) object
naming from the tactile modality; (3) object naming from a verbal description of their
function (by the experimenter) and (4) pantomime of object use (by the patient). Although
pantomime is not a linguistic task, it was grouped together with naming tasks, since it may
be viewed as a gestural counterpart of the lexical label of a given manipulable object. The
same 36 objects were used for all tasks, except for pantomimes, since some objects lacked
gestural specificity. As shown in Table II, the deficit observed in visual naming was
significantly more marked than that in tactile naming (Chi-square = 6.25, p. <.02), in
naming from verbal description (Chi-square = 4.96, p. <.03), and in miming (Chi-square =
9.94, p. <.002). Errors made in visual naming included perseverations (59%),

Optic aphasia with pure alexia

441

TABLE II

CN’s Performance in Object Processing Assessment

Object naming and gesturing
Visual input
Tactile input
Verbal description
Pantomime
Elementary visual perception
Colours vision (Ishihara; Farnsworth)
Line orientation judgment (Benton et al.)
Identical shapes matching (PEGV; Agniel et al.)
Overlapping shapes matching (PEGV; Agniel et al.)
Access to structural representation
Different-views object matching
Different-views face matching (Benton et al.)
Access to semantic representation
Functional matching (PEGV; Agniel et al.)
Categorical matching (PEGV; Agniel et al.)
Semantic association (from Howard et al.)
Visual input
Verbal input

Accuracy score

Judgement or %

19/36
29/36
28/36
18/19

53%
81%
78%
95%

13/13; 15/15
26/30
10/10
12/12

–
–
–
–

25/26
47/54

–
–

10/10
10/10

–
–

6/12
10/12

++
–

Note: normal performance: –; slight deficit: +; moderate deficit: ++; severe deficit: +++.

circumlocutions (29%) and semantic confusions (12%) (for more details, see Teixeira
Ferreira et al., 1997).
In conclusion, CN’s profile does not fit the conventional definition of anomia as his
naming of tactually and auditory presented objects is better than that of visually presented
objects. Also, CN cannot be classified strictly as an agnosic patient since he often gives
correct gestures for objects he cannot name. Thus, CN has been shown to have a modalityspecific naming impairment in conjunction with a relatively spared ability to make the
correct gesture for a given object, and according to Lhermitte and Beauvois (1973), he
should be labelled as an optic aphasic.
Elementary Visual Perception
To exclude a possible deficit in elementary visual perception that might explain CN’s
visual-specific naming deficit, the following tasks were used: (1) two colour vision tests
(Ishihara, 1982; Farnsworth, 1957); (2) line orientation judgement test (Benton et al., 1983);
(3) identical shapes matching test (Protocole d’Evaluation des Gnosies Visuelles or PEGV;
Agniel et al., 1993) and (4) overlapping shapes matching test (PEGV; Agniel et al., 1993).
Colour Vision. Ishihara test consists of colour classification with isochromatic cards.
Only isochromatic cards without digits were proposed because of CN’s alexia. His
performance on that subtest was perfect (13/13). Farnsworth test is similar but a finer
discrimination of colour shades is requires. CN correctly ranged the 15 colours of than test.
Line Orientation Judgement. CN was asked to select orientation of the target line from
among fan-shaped lines. CN’s accuracy score (26/30) was slightly above the normal mean
(24.3).
Identical Shapes Matching. The aim of this tests was to determine the capacities to
discriminate the identical shape target from among visually similar geomtric shapes. CN’s
shape discrimination was perfect (10/10).
Overlapping Shapes Matching. The test assesses shape discrimination abilities and
capacities to extract some elements included in a complex pattern of 3 overlapping shapes.
CN was asked to select one target from among 8 shapes belonging to the same category.
He did not make any error: 12/12.
Comment. Perfect performance on these tasks indicate that perceptual processes were
spared.

442

V. Chanoine and Others

Access to Structural Representation
Structural representation is assumed to be stored in long term memory (LMT) and to
involve visual features that are independent of viewpoint. This supposes storage of canonical
description of objects, along with mapping procedures so that object’s representation can be
accessed from several viewpoints. One component process in “visual object recognition”
then, is the ability to perceive that an object’s representation is constant across changes in
viewpoint (see Warrington and Taylor, 1978). Object constancy was tested in CN using a
different-views matching test for objects and faces.
Different-view Object Matching. CN was given three pictures: one of an object taken
from a canonical view, one of the same object taken from a non-canonical view and one of
a morphological distractor. He was required to choose which of two pictures represented
the same object. His performance was almost perfect (25/26).
Different-view Face Matching (Benton et al. 1983). CN was shown photographs of
faces and asked to select the two depicting the same face from different views and lighting.
CN’s correct responses (47/54) fell within the normal range (41-54).
Comment. CN’s spared ability to identify two different views of the same object
provides evidence that his difficulties in visual naming of objects occur in spite of correct
access to the structural representations of these objects in LTM. Faulty processes are
therefore likely to concern later stages, namely, the semantic representation of objects.
Access to Semantic Representation
CN’s ability to access semantic knowledge was tested using three tasks: (1) a functional
matching test (PEGV, 1993); (2) a categorical matching test (PEGV, 1993) and (3) a
semantic association test (drawn from “Pyramid/Palm-tree association”, Howard and
Orchard-Lisle, 1984; see also Goldblum et al., 1994).
Functional Matching. CN was shown three object pictures (e.g. stamp, cube, stairs) and
had to select one of them as presenting a functional relationship with an other object picture
(e.g. envelope). The two functionally associated objects could not be matched on the basis
of visual information because they were selected so as to be visually dissimilar. Access to
semantic knowledge (or, at least, to some semantic attributes) is necessary to pass the test.
CN did not make any errors (accuracy score = 10/10).
Categorical Matching. The principle of the categorical matching task is the same as
that of functional matching, except that CN was required to select categorically related
objects (e.g. carrot, salad, hammer, aeroplane). His performance was perfect (10/10).
Semantic Association. This test was given in a visual version (i.e. object pictures) and a
verbal version (i.e. spoken names) using identical objects. To avoid learning effects, one week
interval separated the two versions. CN was presented with three objects or words (e.g. palmtree, pyramid, fir) and asked to select the two that were most closely related (ssociative or
functional relationship). Such semantic association is normally found to be stronger than the
semantic relationship that refers to co-ordinates in a category (e.g. two trees). Compared to
prior tests, this association test required the extraction of more detailed semantic information.
In the visual version, CN’s accuracy score was at a random level (50%), while his performance
was well above chance (83%) in the verbal version. For comparison purposes, a non-braindamaged, age-matched control scored 11/12 (92%) correct in both the visual and verbal versions
of the same test. The across-subject difference was significant in the visual version (Chi-square
= 5.04, p <.03) but not in the verbal version (Chi-square = 0.38, p = .5371). Results indicate that
the access to detailed semantic information from visual input was disrupted in CN whereas it
seemed to be preserved from verbal input.
Comment. CN performed well on visual tasks in which functional or categorical
matching was required, but was poor when more detailed semantic information was required
(i.e. the pyramid/palm-tree association). More to the point is the fact that he was particularly
disrupted in processing detailed semantic information in the visual version as opposed to
the verbal version of the task. This result is consistent with the argument that CN had
impaired access to detailed semantic knowledge from vision, accessing only general
categorical information.

Optic aphasia with pure alexia

443

In summary, object processing in CN may be featured as follows: (1) the object naming
deficit is selective to the visual modality; (2) elementary visual perception is intact; (3)
“structural recognition” on visually presented objects is spared and (4) “semantic
recognition” is preserved when tested with miming and tests tapping functional and
categorical knowledge, but fails when more detailed semantic knowledge is required. This
analysis confirms the conventional definition of optic aphasia according to Lhermitte and
Beauvois (1973). Furthermore, results seem to lend support to a breakdown in visual access
to semantic knowledge (Riddoch and Humphreys, 1987) as the underlying mechanism of
CN’s visual naming deficit.
In the following section, CN’s ability in visual letter processing will be considered. We
will examine whether visual letter naming deficit might be due to a partial deficit in
accessing alphabetical information from vision. In other words, may the visual naming
deficit relative to objects and letters be explained by a common underlying mechanism,
namely a partial deficit in visual recognition?
Letter Processing Assessment
This section aims at identifying the nature of CN’s naming deficit on visually presented
letters. Visual letter naming is often assumed to involve processing stages similar to those
of visual object naming, namely: (1) elementary perception; (2) access to structural
representation (i.e. stored representation of visual salient features that are independent of
viewpoint); (3) access to alphabetical representation (i.e. stored representation that is related
to our knowledge of letters) and (4) retrieval of letter names.
Tabel III presents CN’s performance on tests designed to assess letter processing. These
tasks included: (1) letter naming in severalsensorial modalities; (2) assessment of access to
letter structural representations from visual input and (3) assessment of access to
alphabetical representations from visual or verbal inputs. Evaluation of elementary visual
perception is not described here (see supra Table II).
Naming Tasks
By analogy to CN’s performance with objects (see supra), we expected his letter naming
disorder to be selective to the visual modality. This prediction was tested by comparing the
patient’s letter naming performance in various sensory modalities.
Letter Naming from Visual Input. CN was successively presented the 26 letters of the
French alphabet and was required to name each letter. As indicated in Table III, CN presents
a severe deficit in visual letter naming (accuracy rate = 46%). Errors consisted of
substitutions within the alphabet, i.e. no number name or symbol name was producted.
TABLE III

CN’s Performance in Letter Processing Assessment

Letter naming
Visual input
Somesthesic input
Alphabet series
Kinesthesic assistance
Access to structural representation
Letter decision task
Access to alphabetical representation
Letter verbo-visual matching
Across-category task
Within-category task
Allographic letter matching
Across-category task
Within-category task

Accuracy score

Accuracy rate (%)

12/26
11/26
24/26
19/26

46
42
92
73

94/104

90

20/20
12/20

100
60

20/20
11/20

100
55

444

V. Chanoine and Others

Letter Naming from Somesthesic Input. The experimenter drew with a pencil a letter in
the left palm of the subject. The patient, his eyes closed, was asked to name each letter.
CN’s performance for somesthesic letter naming was as bad as in the visual letter naming
test (from visual input: accuracy rate = 46%; from somesthesic input: accuracy rate = 42%;
Chi-square = .08; p =.7801, non significant). However, no consistency was observed in the
errors, different letters being involved in erroneous responses in the visual and in the
somesthesic modality.
Visual Letter Naming with Kinesthesic Assistance. The patient, his eyes opened, was
asked to let the experimenter move his left index and draw by passive movements the shape
of a letter that he had to name. As shown in Table III, CN made significantly less errors in
kinesthesic naming (accuracy rate = 73%) than in visual (Chi-square = 3.91, p < .05) and
somesthesic conditions (Chi-square = 5.04, p < .03).
Naming of Letters from Alphabet Series. CN was asked to recite all the letters of the
French alphabet. Although his recitation was slow, CN succeeded in producing all the
letters, except for two of them (“N” and “V”). He did not seem to be conscious of his
two omissions and made significantly less errors in this test (accuracy rate = 92%) than in
visual (Chi-square = 13.00, p < .0003) and somesthesic conditions (Chi-square = 14.77,
p < .0001).
Comment. CN’s behaviour in letter naming tasks can be summarised as follows.
First, spared letter naming from alphabet series rules out an impairment of alphabetical
representations or a disruption of letter names. Another source of evidence for the sparing
of the phonological representations of letters comes from CN’s normal performance on
spelling (see CN’s general neurophsychological assessment).
Second, the passive writing of a letter facilitatd the naming performance compared to
the letter naming on strictly visual presentation. Such facilitation of visual naming by a
kinesthesic assistance suggests a disruption of a mechanism specific to the visual modality.
Third, CN made as many errors when he had to name letters from a somesthesic input
as in visual naming. In the somesthesic condition, CN’s errors could be explained by the
unusual nature of the task. Also, it is likely that letter naming from somesthesic input
requires information transcoding to spatial and visual representations as well as visuospatial memory processing.
Fourth, the error analysis of the patient in overall visual naming tasks indicates an
absence of “consistency” across tasks, which seems to evoke and impaired access to
alphabetical representations from visual input rather than a loss of alphabetical
representations themselves (Shallice, 1988).
In conclusion, the analysis of CN’s performance on letter naming tasks reveals that his
difficulties affected preferentially letter naming from vision. The following section aims at
determining which level of visual letter processing was impaired in CN. The lower
perceptual processing stages were not formally assessed because, as mentioned in prior data
about objects, CN did not appear impaired at this level.
Access to Structural Representation
Letter Decision Task. CN was asked to discriminate between line drawings of letters
and pseudo-letters. Pseudo-letters were constructed by substituting a part of a letter with
that of another letter (Figure 2). This task can be performed on the basis of stimulus

Fig. 2 – Examples of stimuli used in letter decision task.

Optic aphasia with pure alexia

445

familiarity only and does not require access to stored knowledge about letter attributes (e.g.
vowel vs. consonant), not letter names.
The stimuli (26 letters and 26 pseudo-letters) were presented individually on a computer
screen in random order. CN was instructed to decide whether or not the stimulus was a real
letter. Two blocks were given; they differed only in stimuli order presentation. CN’s
accuracy rate was about 90% on the letter decision task: he accepted 51 stimuli among the
52 letter (98%) and dismissed 43 among the 52 pseudo-letters (83%). For comparison
purposes, the same test was given to a non-brain-damaged, age matched control who scored
51/52 (98%) correct on letters and 47/52 (90%) correct on pseudo-letters. The across-subject
difference did not reach significance (for letters: Chi-square = 0, p = 1; for pseudo-letters:
Chi-square = 1.32, p = .2505).
Comment. Despite his massive problem on visual letter naming, CN performed correctly
in discrimination between letters and pseudo-letters, suggesting that he could access
structural representations of visually presented letters. Similarly to object processing, CN
suffers neither from disorders of elementary perceptual processing, nor from deficit of
access to structural representation from visual input.
Access to Alphabetical Representation
Two tasks were proposed to evaluate visual access to alphabetical representations: (1)
letter verbo-visual matching and (2) allographic letter matching.
Letter Verbo-Visual Matching. CN was asked to point among alternatives to the letter
named by the examiner. Two versions of the task were given.
In the across-category task, distractors were digits (e.g. “6”) and geometrical shapes
(e.g. “♦”). Each trial comprised: a target (e.g. “G”); a digit visually close to the target (e.g.
“6”) and a shape visually distant from the target (e.g. “♦”). Out of 20 trials, CN did not
make any error in this task (accuracy rate = 100%).
In the within-category task, all alternatives were letters. Each trial comprised: a target
(e.g. “G”); a distractor visually close to the target (e.g. “C”) and a distractor visually distant
from the target (e.g. “T”). On 20 trials, CN’s accuracy rate was 60%. This performance
was poorer than that obtained in across-category task (Chi-square - 10.00, p < .002).
Allographic Letter Matching. CN was asked to indicate which of 4 visual stimuli were
allographic forms of the same letter upper-case vs. lower-case letter; e.g. “G” and “g”).
In the across category task, the target letters (e.g. “G” and “g”) and the two distractors
belonged to different symbolic categories: one was a digit distractor (e.g. “6”) and the other
was a geometrical distractor (e.g. “♦”). Moreover, one of the two distractors (e.g. “6”) and
the upper-case letter target (e.g. “G”) were visually similar whilst the other distractor was
visually dissimilar (e.g. “♦”). In this way, the success of the choice could not rely on
strictly visual information. On 20 trials, CN did not make any error (accuracy rate = 100%).
In the within-category task, all the 4 visual stimuli were letter. Two were 2 allographic
forms of the same letter. The distractors included an upper-case letter (e.g. “T”) and a
lower-case letter (e.g. “c”); one of 2 distractors was selected to be visually similar to one
of the 2 allographic targets. On 20 trials, CN’s accuracy rate was 55%. This performance
was poorer than that obtained in across-category task (Chi-square = 11.61, p < .0007).
Comment. CN was selectively impaired in both tests when the targets were all letters,
but was not impaired when he had to discriminate letter from pseudo-letters. These results
indicate that he could not access detailed alphabetical representation from visual input.
In sum, the pattern of CN’s letter naming deficit is consistent with a partial deficit of
visual recognition as already mentioned in the analysis of object processing.
DISCUSSION
CN presents a visual-specific object naming impairment that matches the classical
definition of optic aphasia, since he can perform accurate gestures to the visually presented
objects he names incorrectly. The detailed investigation of CN’s naming impairment yelds
to main findings: (1) CN suffers neither from a deficit in elementary perceptual processing,

446

V. Chanoine and Others

nor from a deficit in accessing to stored structural knowledge of objects from visual input.
(2) CN is selectively impaired on tasks requiring the retrieval of detailed semantic
information on visually presented objects. In fact, CN’s deficit consists of an impairment in
discriminating semantically close objects on visual input.
The finding that semantic knowledge is impoverished in optic aphasics has been
reported in others studies (Riddoch and Humphreys, 1987; Hillis and Caramazza, 1995; De
Renzi and Saetti, 1997). Like JB (Riddoch and Humphreys’ patient, 1987), DHY (Hillis
and Caramazza’s patient, 1995) and Luciano (De Renzi and Saetti’s patient, 1997), CN
meets all the criteria for a diagnosis of optic aphasia when assessed with classical tasks,
but performs poorly on tests that require an access to detailed semantic knowledge from
visual input (within-category tasks).
Two alternative accounts for the faulty recognition (from visual input) evidenced by
patients with so called optic aphasia are given in the recent literature. The first is proposed
by Hillis and Caramazza (1995; for a somewhat similar account, see also Riddoch and
Humphreys, 1987). These authors explain the pattern of performance of optic aphasics by
an impaired access to complete, amodal semantic representation from a structural
representation. The syndrome would thus result from a partial disconnection between presemantic (i.e. perceptual and structural) processes carried out by the right hemisphere, and
semantic processes carried out by the left hemisphere. Semantic representation is assumed
to be exclusively located in the left hemisphere and access to this representation from visual
input is only possible via the corpus callosum as visual input cannot originate from the
damaged left occipital cortex. The second alternative is proposed by De Renzi and Saetti
(1997). In this account, the right hemisphere has a certain degree of semantic competence
and optic aphasia could be explained by a disconnection between right-hemisphere
processing and refined semantic processing of the left hemisphere. According to these
authors, inter-subject variations in the semantic capacities of the right hemisphere explain
why similar lesions may induce either optic aphasia or associative agnosia. In this
conception, associative agnosia would correspond to cases with particularly limited semantic
capacities in the right hemisphere. In the absence of convincing anatomical evidence from
lesion-based studies (see De Renzi and Saetti’s argumentation, 1997), it is difficult to
determine which account is the most appropriate. Functional neuroimaging techniques have
recently provided a new line of evidence concerning the neural substrates of object-based
semantic processes. Among these studies, some described an involvement of right
hemispheric regions (Martin et al., 1996) and others did not (Vandenberghe et al., 1996).
Further studies are needed to disentangle the question of the participation of the right
hemisphere in semantic tasks.
However, we concur with De Renzi and Saetti’s view that there is a continuum between
optic aphasia and associative agnosia. Optic aphasia, manifested by CN, could be considered
as a mild form of associative agnosia. Three pieces of evidence seem in favour of the
existence of an agnosic deficit in CN.
First, CN performs poorly on visual tasks that require fine-grained discrimination (withincategory semantic association). The better performance in gesturing than in naming visually
presented objects could be explained by the fact that less specific semantic characteristics are
required in gesturing (for a similar argument, see Ratcliff and Newcombe, 1982).
Second, his letter processing shows problems that are relatively similar to those
observed with objects: impairment on visual tasks that require within-category
discrimination, despite spared perceptual and structural abilities. In particular, CN is able to
visually access rough alphabetical representation (i.e. to discriminate a letter from a pseudoletter) whilst unable to access more detailed alphabetical information (i.e. to discriminate a
letter from other letters).
Finally, while a semantic disorder deficit should result in asymmetric performance
between visual object and letter naming (see Caramazza et al., 1990), the absence of an
asymmetric in CN’s performance seems to point to a pre-semantic, i.e. agnosic, deficit.
In conclusion, these findings suggest that visual object as well as letter recognition are
not completely preserved in CN and that optic aphasia with pure alexia can be considered
as the manifestation of a mild form of associative agnosia. It is noteworthy that most cases
of optic aphasia (for a review, see Davidoff and De Bleser, 1993) and visual associative

Optic aphasia with pure alexia

447

agnosia without prosopagnosia (for a reivew, see Farah, 1991) seem to present alexic
difficulties. The question, then, remains whether a common mechanism actually explains
visual object naming and reading deficits in optic aphasia. Further investigations should
assess the possibility of such a parallelism between visual-specific naming and reading
impairment in optic aphasia.
Acknowledgements. We thank Prof. Ennio De Renzi and Dominique Cardebat for the
helpful comments. We are very grateful to the patient for his collaboration.
REFERENCES
AGNIEL, A., JOANETTE, Y., DOYON, B. and DUCHEIN, C. Protocoles d’Évaluation des Gnosies Visuelles
(PEGV). In J.L. Nespoulous, Y. Joanette and A.R. Lecours (Eds.), Les Protocoles MontrealToulouse Neurolinguistique et Neuropsychologique. Isbergues: L’Ortho Edition, 1993.
BENTON, A.L., HAMSHER, K., VARNEY, N.R., and SPREEN, O. Contributions to Neuropsychological
Assesment. New York: Oxford University Press, 1983.
BOLLER, F., and HÉCAEN, H. L’évaluation des fonctions neuropsychologiques: examen standard de
l’Unité de recherches neuropsychologiques et neurolinguistiques (U-111), INSERM. Revue de
Psychologie Appliquée, 29: 247-266, 1979.
BRUYER, R. Partie 2: Les agnosies visuelles. In X. Seron et M. Jeannerod (Eds.), Neuropsychologie
Humaine. Liège: Pierre Margada, 1994.
CARAMAZZA, A., HILLIS, A.E., RAPP, B.C., and ROMANI, C. The multiple semantics hypothesis: multiple
confusions? Cognitive Neuropsychology, 7: 161-189, 1990.
COSLETT, H.B., and SAFFRAN, E.M. Preserved object recognition and reading comprehension in optic
aphasia. Brain, 112: 1091-1110, 1989.
DAVIDOFF, J., and DE BLESER, R. Optic aphasia: a review of past studies and reappraisal. Aphasiology, 7:
135-154, 1993.
DE RENZI, E., and SAETTI, C. Associative agnosia and optic aphasia: qualitative or quantitative
difference? Cortex, 33: 115-130, 1997.
FARAH, M.J. Visual Agnosia: Disorders of Object Recognition and What They Tell Us about Normal
Vision. Cambridge: The M.I.T. Press, 1990.
FARAH, M.J. Patterns of co-occurrence among the associative agnosias: Implications for visual object
representation. Cognitive Neuropsychology, 8: 1-19, 1991.
FARNSWORTH, D. Manual The Farnsworth-Munsell 100-Hue Test. Baltimore: Munsell Color Co. Inc.,
1957.
FREUND, D.C., 1889. Über optische Aphasie and Seelenblindheit. Archiv für Psychiatrie und
Nervenkrankheinten, 20: 276-297, quoted by Davidoff, J., and De Bleser, R., 1993.
GOLDBLUM , M.C., CHANOINE, V., ROSENTHAL, V., and PONCET, M. Les noms et le choses: étude
comparative de deux cas d’anomie progressive. Revue de Neuropsychologie, 4: 187-221, 1994.
HILLIS, A.E., and CARAMAZZA, A. Cognitive and neural mechanisms underlying visual and semantic
processing: implications from “Optic Aphaisa”. Journal of Cognitive Neuroscience, 7: 457-478,
1995.
HOWARD, D., and ORCHARD-LISLE, V.M. On the origin of the semantic errors in naming: Evidence from
the case of a global aphasic. Cognitive Neuropsychology, 1: 163-190, 1984.
IORIO, L., FALANGA, A., FRAGASSI, N.A., and GROSSI, D. Visual associative agnosia and optic aphasia. A
single case study and a review of the syndromes. Cortex, 28: 23-37, 1992.
ISHIHARA, S. Ishihara’s test for Color Blindness. Tokyo: Kanehara, 1982.
LECOURS, A.R., and LHERMITTE, F. Phonemic paraphasias: Linguistic structures and tentative hypotheses.
Cortex, 5: 193-228, 1969.
LHERMITTE, F., and BEAUVOIS, M.F. A visual-speech disconnexion syndrome: report of a case with optic
aphasia, agnosic alexia and colour anomia. Brain, 96: 695-714, 1973.
LISSAUER, H. Ein Fall von Seelenblindheit nebst einem Beitrage zur Theorie derselben. Archiv für
Psychiatrie und Nervenkrankheiten, 21: 22-270, 1980. (English translation by M. Jackson,
Cognitive Neuropsychology, 5: 157-192, 1988).
MARR, D. Vision. San Francisco: Freeman, 1982.
MARTIN, A., WIGGS, C.L., UNGERLEIDER, L.G., and HAXBY, J.V. Neural correlates of category-specific
knowledge. Nature, 379: 649-652, 1996.
RATCLIFF, G., and NEWCOMBE, F. Object recognition: Some deductions from the clinical evidence. In
A.W. Ellis (Ed.), Normality and Pathology in Cognitive Functions. New York: Academic Press,
1982.
RIDDOCH, M.J., and HUMPHREYS, G.W. Visual object processing in optic aphasia: a case of semantic
access agnosia. Cognitive Neuropsychology, 4: 131-185, 1987a.

448

V. Chanoine and Others

RIDDOCH, M.J., and HUMPHREYS, G.W. A case of integrative visual agnosia. Brain, 110: 1431-1462,
1987b.
SEYMOUR, P.H.K. Human Visual Cognition. New York: St Martin’s Press, 1979.
SHALLICE, T. From Neuropsychology to Mental Structure. Cambridge: Cambridge University Press,
1988.
SNODGRASS, J.G. Concepts and their surface representation. Journal of Verbal Learning and Verbal
Behavior, 23: 3-22, 1984.
TEIXEIRA FERREIRA, C., GIUSIANO, B., CECCALDI, M., and PONCET, M. Optic Aphaisa: evidence of the
contribution of different neural systems to object and action naming. Cortex, 33: 499-513, 1997.
VANDENBERGHE, R., PRICE, C., WISE, R., JOSEPHS, O., and FRACKOWIACK, R.S.J. Functional anatomy of
a common semantic system for words and pictures. Nature, 383: 254-256, 1996.
WARREN, C., and MORTON, J. The effects of priming on picture recognition. British Journal of
Psychology, 73: 117-129, 1982.
WARRINGTON, E.K., and TAYLOR, A.M. Two categorical stages of object recognition. Perception, 7: 695705, 1978.
Valerie Chanoine, INSERM U455, Services de Neurologie, CHU Purpan, 31059 Toulouse Cedex, France.
E-mail: chanoine@purpan.inserm.fr

(Received 25 September 1997; accepted 21 January 1998)