Brain and Language 75, 1–16 (2000) doi:10.1006/brln.2000.2321, available online at http://www.idealibrary.com on Visuographemic Alexia: A New Form of a Peripheral Acquired Dyslexia J. Fernando Dalmás and Sergio Dansilio Laboratorio de Neuropsicologı́a, Instituto de Neurologı́a, Hospital de Clı́nicas, Facultad de Medicina, Montevideo, Uruguay We report a single-case study of peripherally acquired dyslexia that meets the clinical criteria of ‘‘alexia without agraphia.’’ The patient, AA, has a large infarct involving the left posterior cerebral artery. The most striking feature is a severe impairment in recognizing single visually presented letters that precludes explicit or implicit access to reading, even in a letter-by-letter fashion. AA can, however, differentiate letters from similar nonsense characters and digits, and he is also able to identify alphanumeric signs when the visual channel is bypassed (through somesthesic or kinesthesic presentation). Spelling tasks are also well performed. Since there is a breakdown in mapping a visually presented letter to its abstract graphemic representation, we propose the term ‘‘visuographemic alexia’’ for this kind of reading disorder. The pattern of deficits is interpreted following theoretical models previously developed in cognitive neuropsychology. An alexia for arabic numerals with preserved comprehension lends additional support for the crucial processing of different notational systems (e.g., phonographic vs logographic). More general perceptive disorders do not seem to account for these patterns; they are material-specific. Finally, we attempt to specify functional correlations with the implied neural networks.  2000 Academic Press INTRODUCTION To read is after all, to extract the meaning from the visual analysis of black letter strings over a white background, so there must be—beyond the current model adopted—an initial processing stage where the identification of letters and their sequential positions in a word are computed. There exists The authors acknowledge Jean-Luc Nespoulous and Gerard Deloche for their comments on earlier version of this article. We are grateful to Hugh Buckingham for his assistance on the English version. A special thanks for Miss Heather Lugar of the Editors Roster, Liaison Committee, International Neuropsychological Society, for editing the English text. J. Fernando Dalmás suddenly died shortly after the completion of this article. Address correspondence and reprint requests to Dr. Sergio Dansilio, Laboratorio de Neuropsicologı́a, Instituto de Neurologı́a, Hospital de Clı́nicas, Av. Italia s/n 11600, Montevideo, Uruguay. 1 0093-934X/00 $35.00 Copyright  2000 by Academic Press All rights of reproduction in any form reserved. 2 DALMÁS AND DANSILIO an essentially prelexical level of analysis. Shallice and Warrington (1980) have introduced the term ‘‘peripheral dyslexias’’ to account for those acquired reading disorders which are due to impairments in visual decoding processes, perceptive organization, or attentional distribution, all of which are sublexical phenomena. The term includes descriptions of (1) neglect dyslexia, spatial or positional (Ellis, Flude, & Young, 1987; Rapp & Caramazza, 1991; Ellis, Young & Flude, 1993); (2) attentional or ‘‘literal’’ dyslexia (Shallice & Warrington, 1977); and (3) ‘‘letter-by-letter’’ (LBL) reading (Warrington & Shallice, 1980; Patterson & Kay, 1982; Farah & Wallace, 1991; Kay & Hanley, 1991). With neglect dyslexia, the reading disorder is basically a unilateral attentional disruption secondary to right-hemispheric lesions; however, it has also been found with left lesions (Hillis & Caramazza, 1990; Nichelli, Veneri, Pentore, & Cubelli, 1993). The last description, letter by letter, implies that the subject must necessarily name or recognize, serially, each letter of the word as the only way to gain lexical access. Reading time is prolonged and there is a clear word-length effect. This kind of disorder corresponds globally to Dejerine’s original description (1892) of a ‘‘cécité verbale pure’’ so-called ‘‘alexia without agraphia’’ or ‘‘agnosic alexia’’ (Benson, 1985). It has been interpreted since then as a ‘‘disconnection syndrome’’ from visual association areas to language areas in the left hemisphere (Dejerine, 1892; Geschwind, 1965). Seeking a theoretically based explanation for peripheral alexias 1 from a representational–computational approach, Marr’s visual-object-perception model (1982), adapted for reading by Monk (1985) and introduced to cognitive neuropsychology by Caramazza and Hillis (1990), has been used. Further modifications to account for neglect dyslexias have been proposed by Rapp and Caramazza (1991) and Nichelli, Verneri, Pentore, and Cubelli (1993). According to these authors, there are three representational levels in written-word visual analysis. For each, there are two variables: stimulus identification and its location in a system of spatial coordinates (see Fig. 1): 1. A retinocentric visual features map, which computes lines, edges, orientation, points, and so on. Coordinates correspond to the topographically organized retinal surface. 2. A stimulus-centered, letter-shaped map where the letterform, as such, is constructed from previous features. This level is viewer-oriented, so left– right reference is a significant variable. 3. A word-centered, grapheme-description stage where a more abstract representational value is computed from the letterform: the graphemic identity of each letter. Since the coordinate frame is ‘‘object-centered,’’ canonical order of a word is crucial here. Following Zesiger and de Partz (1994) we use the term ‘‘grapheme’’ in its most common sense, with conventional writing models. That is, the graph1 We use the term ‘‘alexia’’ to mean ‘‘acquired dyslexia.’’ VISUOGRAPHEMIC ALEXIA FIG. 1. 3 Cognitive model for prelexical representational levels in written-word visual analysis. eme is the abstract representation of a letter irrespective of case, font, style, color, and so on. In the present article we report a single-case study of a patient with peripherally acquired reading disorder, resembling that of a ‘‘pure alexia without agraphia.’’ Practically all the cases of pure alexia published to date are LBL readers (Warrington & Shallice, 1980; Patterson & Kay, 1982; ReuterLorenz & Brunn, 1990; Farah & Wallace, 1991; Kay & Hanley, 1991; Arguin & Bub, 1993). They preserve some capacity for reading words, be it explicitly or implicitly (Coslett, Saffran, Greenbaum, & Schwartz, 1993). In our case, such an ability is virtually impossible, but other visual tasks are strikingly intact, including a literal-decision task. The impossibility of extracting the graphemic representation from a visually presented letter has been the salient feature in this patient. We propose to call this entity ‘‘visuographemic alexia.’’ We try to ‘‘localize’’ the impairment in the previously outlined cognitive model, make comparisons with the processing of arabic numerals, which is also compromised here, and finally, consider the neural networks most likely implicated in this special patterns of deficits. CASE REPORT AA is a 59-year-old, bilingual (Spanish–Gallecian) right-handed lawyer, born in Vigo (Spain), the town in which he completed elementary school. On March 21, 1994, he suddenly developed a right hemiparesis with tactile sensory loss, right homonymus hemianopia, and mild aphasia symptoms. The motor deficit improved rapidly but sensory loss remained. An MRI per- 4 DALMÁS AND DANSILIO FIG. 2. MRI of AA revealing an extensive left occipital infarction (see text). formed 24 h after the onset revealed an extensive left occipital infarction (see Fig. 2). The lesion involved the classic distribution of the left posterior cerebral artery: calcarine cortex, occipital white matter, lingual and fusiform gyrus, splenium of the corpus callosum, medial temporal lobe (hippocampus and parahippocampal gyrus), and posterior thalamus. Four days after onset he complained of a severe reading failure saying that he only saw ‘‘one or two letters’’ which he could not understand or put together. VISUOGRAPHEMIC ALEXIA 5 Neuropsychological assessment began 5 months after the onset of the initial symptoms. He was alert and oriented. Spontaneous speech was fluent, without articulatory, phonological, or syntactical impairments. Evaluated by the Spreen–Benton Battery, adapted to our language and culture by Mendilaharsu, Ulrikssen de Viñar, Santini de Souto et al. (1968b), we found no aphasia. The Token test (De Renzi & Vignolo, 1962) was not administered, since the patient had a color-recognition deficit affecting his naming (nonlanguage matching tests not altered) and comprehension of sequences. A test of geometrical simple and complex figure matching (see Appendix 1) was carried out in a multiple-choice procedure and was completed rapidly and efficiently by the patient (30/30). Visuoconstructive abilities were tested using the figure-copying test standardized by Mendilaharsu, Miglionico, Acevedo de Mendilaharsu et al. (1968a) and the Rey–Osterrieth complex figure (1959). Strokes were extremely clumsy due to his thalamic ataxia, but the topological, projective, and euclidian dimensions of space appeared preserved in the reproduction (27 points, 75th percentile for the last, type I). There was a moderate verbal memory loss, probably due to hippocampal–parahippocampal involvement as assessed with the Protocolo Montevideo (Fontan, 1993). EXPERIMENTAL INVESTIGATION Letters (A) Visual naming. All 27 letters of the alphabet are printed in uppercase and presented one at a time. He named correctly 3/27 (O, S, I, for ‘‘I’’ he said ‘‘a bar and a point make an ‘i’ ’’). He produced 8/27 erroneous responses without any clear visual similarity (D → /O/, C → /R/, M → /O/, N → / U/, R → /O/, T → /O/, P → /U/, G → /S/), but a recurrent perseveration (‘‘O’’). No response was obtained in the remaining 16. (B) Spoken to written letter matching. As in the previous test, the same number and type of letters as in the previous test were used. The target and three visually dissimilar character letters were presented, each time with a 3-cm separation among them. He performed only 2/27 correctly [(D) and (H)]. (C) Literal-decision task. This task was administered through two modalities (see Fig. 3): (1) two plates with 16 letters and 16 meaningless characters which differ from real letters in one or two strokes (same size, ink, etc.) and (2) two plates with 16 letters and 16 arabic digits—same procedure as above. We asked the patient to circle all those characters which were NOT letters. These tasks were done rapidly and perfectly (100% correct). Note that there were no hemifield effects. (D) Uppercase/lowercase matching. Target letters are presented in a printed uppercase form. The subject was to select among four printed lowercase choices (one correct and three distracters) with no time limits. All 27 letters were tested. AA gave 13/27 correct responses, all of which share an obvious visuoperceptual similarity (C, I, M, N, O, P, S, U, V, W, X, Y, Z). 6 DALMÁS AND DANSILIO FIG. 3. Literal decision task (plate 1). (E) Print/cursive format matching. Following same procedure as before, there were 5/27 correct responses, all of which the visuoperceptual similarity was again the determinant (C, I, N, O, V). (F) Different size, color, and stylus matching of the same letters. Eighteen letters were used in this experiment (target always in a printed uppercase form), varying in size, color, and style or design of choices on three successive runs for each separate variable. AA was asked to select the correct letter to match the target among three choices. Performance was good (18/18 correct). (G) Somesthesic and kinesthesic recognition of letters. With a sharp object, all 27 letters were delineated in the palm of AA’s left hand (repeating the stimulus, if no response was made, to avoid tactile memory factors). The patient obtained 24/27 correct responses, failing on somewhat ambiguous characters (W, Z, and V). Sensory deficits disrupted right-hand recognition. In a second test using large characters (8 ⫻ 4.5 cm), the patient was asked to identify the 27 letters by finger pointing without time limits. Performance was perfect (27/27). Discussion. It is obvious that AA cannot identify letters when they are presented through visual input, nor is he able to name them or to select a spoken one. Given the impossibility of the identification, it was expected that he would fail on cross-matching tasks whenever there was no visual similarity between different letter presentations (uppercase to lowercase or print to cursive). He did demonstrate, however, some kind of knowledge, as he was capable (explicitly or implicitly) of differentiating letters from very similar chimerical characters and from arabic numerals (single digits). It is surely not an anomia since they could be recognized and named once the visual channel was ‘‘unblocked’’ through somesthesic or kinesthesic modalities. Words (A) Oral reading. Oral reading of common words was virtually impossible (0/20 words of common objects, none of them exceeding four syllables, nine letters). The patient proceeded letter by letter indistinctly over the entire word and then gave up. We provide here two examples: VISUOGRAPHEMIC ALEXIA 7 CUCHARA (spoon) → ‘‘C-S-C-K-Q-A-H-A . . .’’ PELOTA (ball) → ‘‘H-S-P . . .’’ [stopped and said: ‘‘no se’’ (‘‘I don’t know’’); ‘‘no puedo’’ (‘‘I can’t,’’ etc.)] (B) Written-world to real object matching. This simple task was also impossible for AA to do (0/20 correct). (C) Lexical-decision task. A list made of 20 high-frequency words mixed with 20 pronounceable nonwords and similar number of letters, none of them exceeding two syllables (which is truly short in Spanish). AA produced only 4/40 correct responses. He hesitated often and gave an answer only if he was compelled to do so. Discussion. Nonrecognition of letters precludes AA’s reading of whole words. The subject tries unsuccessfully to name each character as a strategy that resembles an LBL reader. Failure on task (B) demonstrates that AA has no implicit knowledge of written words, unlike the case JWC of Coslett, Saffran, Greenbaum, and Schwartz (1993), since objects with high frequency of use were selected to be matched. Spelling (A) Oral spelling. Here, the patient must spell 15 words aloud from the Spreen–Benton Battery (Mendilaharsu et al., 1968b). Words are orally presented by the examiner with increasing length from 2 (e.g.: ‘‘TU,’’ ‘‘you’’) to 15 letters (e.g., ‘‘NACIONALIZACION,’’ ‘‘nationalization’’) and from one to six syllables.2 AA satisfactorily spelled the first 10 items (e.g., ‘‘CRISANTEMO,’’ ‘‘chysanthemum’’), but failed on one or more letters with words of 11 letters (five syllables) or more. In a different session, the list of words used in the oral reading experiment (A) above was also presented. AA produced no errors here (20/20 correct). (B) Naming words spelled by the examiner (spelling identification). To avoid working memory variables as much as possible, only the previously mentioned 20-word list was used. Again, performance was perfect (20/20). Discussion. Provided that the item does not exceed his working memory span, AA maintains some type of written-word representation since he can access a lexical item from the name of its component letters or the reverse. These abilities intact, together with a noncentral agraphia (which is commented on below), imply intactness of the graphemic buffer (Hillis & Caramazza, 1995) or conservation of the ‘‘abstract graphemic storage’’ (Ellis, Young, & Anderson, 1988). 2 We must first keep in mind the transparent condition of Spanish writing and, second, the importance of syllabic dimension. 8 DALMÁS AND DANSILIO FIG. 4. Magnitude comparison task. Errors are marked. Arabic Numerals Reading and Comprehension Thirty-nine arabic numerals, varying in length form one to five digits (see Appendix 2), were presented for reading aloud. AA correctly transcoded only 5 of 39, with 9 clearly perseverative responses. Excepting three items (39 → /390/; 909 → /91/; 13 → /091/), the patient preserves the magnitude of each number. Errors could be considered to be syntactical in just 6 cases (e.g., 310 → /301/). Besides that, lexical errors largely predominate: TARGET: 73.000 (SETENTA Y TRES MIL) response: /setenta y cinco mil/ (75.000) 3 or TARGET: 19.897 (DIECINUEVE MIL OCHOCIENTOS NOVENTA Y SIETE) response: /veinticinco mil quinientos noventa y uno/ (25.591) To explore number comprehension, a magnitude comparison task was designed using the same numerals of the preceding list, presenting one pair each time (see Fig. 4), and having the patient select ‘‘the bigger’’ of the two. There were a total of 14 pairs. AA committed only two errors (pairs 938/ 909 and 70.064/73.000) and was able to correctly discriminate correctly complex pairs like 1.018/1.180 or 4.215/4.125. It is worth noting that AA was unable to correctly read numerals in which he could nevertheless compare very well. 3 AA’s production is transcribed between virgules. VISUOGRAPHEMIC ALEXIA 9 Discussion. AA’s relatively preserved ability to make magnitude comparisons with multidigit numbers implies—following McCloskey and Caramazza (1987) and McCloskey et al. (1990)—that number comprehension (arabic code) is intact and he accesses some sort of meaning. Although some authors (Poltrock & Schwartz, 1984) have shown that magnitude comparison between arabic numerals (multidigit) is computed—or at least could be— through a sequential analysis of component digits, we remind the reader that most of AA’s number-production errors are lexical. Furthermore, since he generally maintains the syntactic frame of the numbers presented, we may assume that errors arise on the accessing route to the output phonological lexicon for number names. Writing (A) Letter writing to dictation. Besides execution difficulties—affecting the strokes and quality of calligraphy—AA was able to write down all 27 letters in uppercase form and lowercase cursive. (B) Letter copying. The subject can only manage to do a ‘‘slavish’’ copy without a posteriori understanding of characters besides ‘‘O’’ and ‘‘I’’ (27 letters presented as printed uppercase and lowercase). (C) Passing printed uppercase (A) to lowercase (a) and printed uppercase (A) to cursive (a). This task was impossible if letters were dissimilar (e.g., R–r or L–l). (D) Word copying. The same results as in (B) were obtained. Beyond the slavish way of copying, no error was made (20/20 correct). (E) Writing to dictation. The same list of 20 concrete nouns used in previous tasks and taken from the Spreen–Benton Battery (Mendilaharsu et al., 1968b) was dictated to AA. He made just two omissions in two words (‘‘ANILLO,’’ ‘‘ring’’ and ‘‘LAPICERA,’’ ‘‘pen’’), all other 18 items being written without graphemic composition or ordering errors. (F) Written naming of visually presented real objects. A different set of 20 common objects was presented to the patient (to avoid learning effects). No graphemic or lexical level error was registered (20/20 correct using handwriting). (G) Other written tasks. Print-to-cursive transcoding was again impossible for words (0/20). Writing to dictation of nonwords (using the same list fixed for the lexical decision task) presented no difficulty to AA (20/20 correct). Finally, the patient was tested using a typewriter for a written naming task of 10 common objects. Since he just partially learned a ‘‘tactile’’ technique, graphemic composition of words was destabilized once a visually guided procedure was employed (2/10 correct). (H) Arabic numeral writing. AA could perform normally numbers up to 6 digits long (26/26 correct). Discussion. AA preserved the capacity for writing to dictation (of isolated letters, words, and nonwords) and for written naming of visually presented objects. This demonstrates intact graphemic or ortographic representation 10 DALMÁS AND DANSILIO for audiographic transcoding processes. Once the stimulus is presented visually, AA can only manage to write slavishly, being unable to change format or style except for those letters whose allographs are very similar. Clumsiness of strokes might be considered, following Ellis (1988). This would implicate an ‘‘execution’’ disorder of graphic-motor-pattern realization, given that the clumsiness also appears on figure copying. GENERAL DISCUSSION AA meets the clinical criteria to be considered as having an alexia without agraphia (pure or agnosic alexia) (Benson, 1985; Friedman, 1988). His lesion accords as well with classic descriptions (Alajouanine, Lhermitte, & De Ribaurcourt-Ducarne, 1960; Damasio & Damasio, 1983; Greenblatt, 1990). Benson (1985) reported cases of Hinshelwood—in 1900—with patients who had ‘‘letter blindness without a word blindness’’ and others with the reverse pattern. AA shows such a severe deficit in recognizing visually presented letters that he can be viewed as a ‘‘letter blind’’ subject, whose impairment affects word reading as well. A review of published cases (Warrington & Shallice, 1980; Patterson & Kay, 1982; Damasio & Damasio, 1983; Alajouanine, Lhermitte, & De Ribaurcourt-Ducarne, 1960; Greenblatt, 1990; Caramazza & Hillis, 1990; Reuter-Lorenz & Brunn, 1990; Farah & Wallace, 1991; Kay & Hanley, 1991; Rapp & Caramazza, 1991; Arguin & Bub; 1993; Coslett, Saffran, Greenbaum, & Schwartz, 1993) reveals very few observations of published cases of such profound deficits in visual letter identification. An exception could be case D.C. in Wilson (1994) on reading rehabilitation, although that patient, suffering from an extended war injury, had also significant intellectual and memory impairments. Leaving aside the discussion of whether pure alexia is a unitary or nonunitary syndrome from the neurocognitive point of view (Kay & Hanley, 1991; Arguin & Bub, 1993), we postulate a theoretical account for the case presented in this article, where evidence concerning the importance of letter identification in reading is provided (see McClelland & Rumelhart, 1981). The central issue that AA represents is the inability to identify letters when presented through a visual input. The essential question refers to establishing which level of information processing is impaired. Several alternatives may be proposed: 1. It may not be a deficit in letter naming, following Rapsak, Rubens, and Laguna’s case (1990), since naming is unimpaired when letter presentation bypasses the visual channel (through somesthesic or kinesthesic routes). AA can also recognize letters if they are given orally by the examiner to produce the word, developing a sequential (and effective) reconstruction of the string. By maintaining a set of letters in the graphemic buffer (Hillis & Caramazza, 1995)—being the reverse spelling task—the patient is able to mentally ‘‘read’’ the word (its ‘‘internal’’ representation). VISUOGRAPHEMIC ALEXIA 11 2. Nor is it pertinent to postulate a disruption of what Warrington and Shallice (1980) have called ‘‘word form system.’’ AA cannot even identify single letters. It also rules out the ‘‘simultanagnosic’’ hypothesis (Kinsbourne & Warrington, 1963): ‘‘nonlinguistic’’ visual tasks do not appear to be affected. 3. Graphemic representations are themselves intact, since direct and reverse spelling can be carried out satisfactorily by the patient. Although spelling mechanisms are a topic under discussion, we presume that, without graphemic representations, as previously stated by Ellis (1982); Ellis, Young, and Anderson (1988); and Rapsak, Rubens, and Laguna (1990), it would be impossible to name letters from a word or the reverse. Further evidence that abstract graphemic storage is intact can be derived from tactile and kinesthesic letter identification. As Hillis and Caramazza explained (1995), if processing abstract graphemic representation in the graphemic buffer is disrupted, all types of spelling tasks would be affected. Returning to the visual recognition model discussed above (see Introduction), we may be sure that the elementary visual features of letters (level 1) are correctly computed by AA. Furthermore, the first- and second-level representations are intact: excellent performance in literal-decision tasks is a decisive argument. The system here identifies letters accurately as such, discriminating them from other graphic characters that are similar (e.g., nonsense characters differing in one or two strokes from real letters) and digits. This implies that first there is a well-constructed lettershape map (its structural description) and it can be linked to some defined category: ‘‘a letter.’’ Unlike Farah and Wallace (1991), an impairment of visual perception nonspecific to orthographic material does not account for the pattern of deficits in AA. When AA must extract the graphemic representation from a wellperceived and categorized visually presented letter, he fails dramatically, demonstrating an inability to process level 3 abstract graphemic description. This is the reason why we proposed the term ‘‘visuographemic’’ for the kind of pure alexia reported here. There would be, then, a sort of functional ‘‘disconnection’’ between the structural/categorial description of a letter and the corresponding graphemic representation whenever the character is processed through the visual channel. Reuter-Lorenz and Brun (1990) have proposed for their case WL an inability to form abstract character representations automatically and efficiently. Arguin and Bub (1993) suggested, to explain a case of pure alexia, D.M., that this occurs on the alphanumeric level. But both cases are LBL readers, with some preserved capacity of accessing words—either explicitly or implicitly—and, obviously, accessing individual letters. In our case, such an alternative is virtually untenable, in light of the host of visual tasks that are intact (including literal decision). A different pattern has been found in arabic numeral reading, a finding previously demonstrated from different theoretical approaches (Seron & Deloche, 1984; Deloche & Seron, 1987; McCloskey & Caramazza, 1987; Mc- 12 DALMÁS AND DANSILIO Closkey, Sokol, Goodman-Schulman, & Caramazza, 1990) and which is due to the intrinsic nature of each notation system. The letter, as a graphic character, represents a grapheme, which in turn maps onto abstract sublexical units (phonemes, in the case of transparent writing as in Spanish). Arabic numerals are, however, truly logograms or ideograms representing, on the hand, certain concepts of magnitude (numerosity, cardinality, and seriality) and, on the other hand, those concepts related to a particular lexical item or morpheme. The status would be somehow parallel to the Japanese writing system: kana characters (syllabic), like letters (alphabetic) refer to ‘‘sounds,’’ so they are ‘‘phonograms.’’ Kanji (logographic), as the Arabic system, has an essentially double tie both to lexical units and concepts. Cases of alexia in Japanese-speaking subjects show comparable intrasystem dissociations (Yamadori, 1975; Hatta, 1977; Kawahata & Nagata, 1988). Oral reading of arabic numerals in AA is quite faulty (5/39 correct), failing even with single digits. Most errors are lexical and maintain the syntactic frame of the target number. Magnitude comparisons between simple and complex numbers can be done relatively well. Following McCloskey and Caramazza’s model (1987), we could assume that graphic representations of numbers map onto a central semantic system, providing good comprehension and preservation of syntactic frame. It would then be an impairment of ‘‘number production,’’ when the corresponding phonological representation must be mapped from a correctly processed arabic form. Dehaene and Cohen (1995), though, have developed a so-called ‘‘neurofunctional model,’’ with a specific reference to pure alexia patterns. According to their approach, the verbal system (lefthemisphere language areas) would fail to receive normal inputs from the visual word and number form (bilaterally represented). But pure alexics could select the larger of two arabic numerals that they failed to read aloud, since magnitude comparison is made from an intact visual number form processed by right occipitotemporal systems. Patient HAR, reported by McNeil and Warrington (1994), shows a similar pattern of deficits, including a singleletter identification impairment (restricted to naming tasks). No assumptions are established here as to whether multidigit comparisons imply parallel or serial processing (Poltrock & Schwartz, 1984). We must yet consider neural correlations of our patient’s deficits. AA has a large left occipitotemporal lesion, involving the classic distribution of the posterior cerebral artery (see Fig. 2). As the caudal portion of corpus callosum is clearly included, visual inputs are mainly processed by the right hemisphere. Its neural networks are then a sufficient basis for structural/categorial letter identification. There is enough experimental evidence which favors the right hemisphere’s competence in this special function. Wagner and Harris (1994), for instance, have demonstrated this in normal subjects using a tachistoscopic procedure. Patient DM of Arguin and Bub (1993) showed a priming effect on visually similar letters (e.g., B–B), but not when the format was changed (e.g., B–b), while normal subjects manifest priming either way (both hemispheres are functioning). Further suggestive evidence has been VISUOGRAPHEMIC ALEXIA 13 published by Reuter-Lorenz and Baynes (1992), whose patient JW, after callosotomy, showed a facilitation effect on left-hemifield presentation only if the previous stimulus was a visually similar letter (e.g., B–B), but not when format differed (e.g., B–b). Right-hemifield (left-hemisphere) presentations, though, benefit from both kinds of facilitation, assuming that access to graphemic representation is crucial in relating different allographs. From these data and case A.A., we conclude that neural networks of the right occipitotemporal lobe are capable of constructing a lettershape map, but the process of relating it to abstract graphemic representation requires intact left-hemisphere linguistic function. The right hemisphere may also ‘‘compute’’ logographic signs such as arabic numerals and even extract some sort of semantics about them, at least enough to make magnitude comparisons between multidigit numbers. Mapping of the phonological correlates (production stages) needs left-hemisphere language areas, as suggested by Holender and Peereman (1987). APPENDIX 1 Test of Geometrical Simple and Complex Figure Matching: Example of a Trial 14 DALMÁS AND DANSILIO APPENDIX 2 Arabic Numerals Reading: Verbal Responses Are Transcribed to Arabic Code 6 → (9) 4 → (9) 2 → (8) 5 → (9) 3 → (6) 7 → (9) 8 → correct 9 → correct 12 → (19) 13 → (0–9) 17 → (19) 15 → correct 38 → (19 . . . 39 . . . 38) 35 → (39) 39 → (390) 72 → (39) 76 → (36) 70 → (30) 50 → (30) 28 → (20 . . . 28) 310 → (301) 301 → correct 909 → (‘‘9 and 91’’) 983 → (9 . . . 8 . . . 3 . . . 983) 1180 → (9980) 1018 → (9081) 3005 → (33051) 2756 → (3956) 4125 → (5982) 4215 → (5415) 8079 → (3989) 8102 → (8903) 20003 → (80031) 19897 → (25591) 73000 → (25000) 70064 → (51095) 73000 → (75000) 30021 → (40021) 32001 → (25100) REFERENCES Alajouanine, T., Lhermitte, F., & De Ribaurcourt-Ducarne, B. (1960). Les alexies agnosiques et aphasiques. In T. Alajouanine (Ed.), Les Grandes Activités du Lobe Occipital. Paris: Masson. Arguin, M., & Bub, D. N. (1993). Single character processing in a case of pure alexia. Neuropsychologia, 31, 435–458. Benson, D. F. (1985). Alexia. In P. J. Vinken, G. W. Bruyn, & H. L. Klaawans (Eds.), Handbook of clinical neurology. Amsterdam: Elsevier. Vol. 45. Caramazza, H., & Hillis, H. E. (1990). Levels of representation, coordinates flames and unilateral neglect. Cognitive Neuropsychology, 7, 391–445. Coslett, H. B., Saffran, E. M., Greenbaum, S., & Schwartz, H. (1993). Reading in pure alexia. Brain, 116, 21–37. Damasio, H. R., & Damasio, H. (1983). The anatomic basis of pure alexia. Neurology, 33, 1573–1583. Dehane, S., & Cohen, L. (1995). Towards an anatomical and functional model of number processing. Mathematical Cognition, 1, 83–120. Dejerine, J. (1892). Contribution à l’étude anatomopathologique et clinique des différentes variétés de cécité verbale. Mémoires de la Société de Biologie, 4, 61–90. Deloche, G., & Seron, X. (1987). Numerical transcoding: A general production model. In G. Deloche & X. Seron (Ed.), Mathematical disabilities: A cognitive neuropsychological perspective. Hillsdale/London: Erlbaum. De Renzi, E., & Vignolo, L. A. (1962). The Token Test. A sensitive test to detect receptive disturbances in aphasies. Brain, 85, 665–678. VISUOGRAPHEMIC ALEXIA 15 Ellis, A. W. (1982). Spelling and writing (and reading and speaking) In A. W. Ellis (Ed.), Normality and pathology in cognitive functions. London: Academic Press. Ellis, A. W. (1988). Normal writing processes and peripheral acquired dysgraphias. Language and Cognitive Processes, 3, 99–127. Ellis, A. W., Flude, B. M., & Young, A. W. (1987). Neglect dyslexia and the early visual processing of letters in words and nonwords. Cognitive Neuropsychology, 4, 439–464. Ellis, A. W., Young H. W., & Anderson, C. (1988). Modes of word recognition in the left and right cerebral hemispheres, Brain and Language, 35, 254–273. Ellis, A. W., Young H. W., & Flude, B. M. (1993). Neglect and visual language. In I. M. Robertson & J. C. Marshall (Eds.), Unilateral neglect: Clinical and experimental studies. Hove: Erlbaum. Farah, M. J., & Wallace, M. A. (1991). Pure alexia as a visual impairment: A reconsideration. Cognitive Neuropsychology, 8(3/4), 313–334. Fontan, L. (1993). Semiologı́a de la memoria. In F. Dalmás (Ed.), La memoria desde la Neuropsicologı́a. Montevideo: Roca Viva. Friedman, R. (1988). Acquired alexia. In F. Boller & J. Grafman (Eds.), Handbook of neuropsychology. Amsterdam: Elsevier. Vol. 1. Geschwind, N. (1965). Disconnection syndromes in animals and man. Brain, 88, 585–645. Greenblatt, S. H. (1990). Left occipital lobectomy and the preangular anatomy of reading. Brain and Language, 38, 578–595. Hatta, T. (1977). Recognition of Japanese Kanji in the left and right visual fields. Neuropsychologia, 15, 685–688. Hillis, M. E., & Caramazza, A. (1990). The effects of attentional deficits on reading and spelling. In A. Caramazza (Ed.), Cognitive neuropsychology and neurolinguistics. Hillsdale, NJ: Erlbaum. Hillis, M. E., & Caramazza, A. (1995). Spatially specific deficits in processing graphemic representations in reading and writing. Brain and Language, 48, 263–308. Holender, D., & Peereman, R. (1987). Differential processing of phonographic and logographic single-digit numbers by the two hemispheres. In G. Deloche & X. Seron (Eds.), Mathematical disabilities: A cognitive neuropsychological perspective. Hillsdale/London: Erlbaum. Kawahata, N., & Nagata, K. (1988). Alexia with agraphia due to the left posterior inferior temporal lobe lesion: Neuropsychological analysis and its pathogenetic mechanisms. Brain and Language, 33, 296–310. Kay, T., & Hanley, R. (1991). Simultaneous form perception and serial letter recognition in a case of letter-by-letter reading. Cognitive Neuropsychology, 8, 249–273. Kinsbourne, M., & Warrington, E. K. (1963). The localizing significance of limited simultaneous visual form perception. Brain, 86, 697–702. Marr, D. (1982). Vision. New York: W. H. Freeman. McClelland, J. L., & Rumelhart, D. E. (1981). An interactive activation model of context effects in letter perception: Part I. An account of basic findings. Psychological Review, 88, 357–407. McCloskey, M., & Caramazza, A. (1987). Cognitive mechanisms in normal and impaired number processing. In G. Deloche & X. Seron (Eds.), Mathematical disabilities: A cognitive neuropsychological perspective. Hillsdale/London: Erlbaum. McCloskey, M., Sokol, S. M., Goodman-Shulmman, R. A., & Caramazza, A. (1990). Cognitive representations and processes in number production: Evidence frome cases of ac- 16 DALMÁS AND DANSILIO quired dyscalculia. In A. Caramazza (Ed.), Cognitive neuropsychology and neurolinguistics. Hillsdale/NJ: Erlbaum. McNeil, J. E., & Warrington, E. K. (1994). A dissociation between addition and substraction with written calculation. Neuropsychologia, 32, 717–728. Mendilharsu, C., Miglionico, A., Acevedo de Mendilharsu, S., Budelli, R., & Santini de Souto, H. (1968a). A propos d’une épreuve d’etude de la apraxie constructive pour diferencier les lésions de l’hemisphère droit et gauche. Acta Neurológica Latinoamericana, 14, 130– 154. Mendilharsu, C., Ulrikssen de Viñar, M., Santini de Souto, H., Budelli, R., Miglionico, A. & Acevedo de Mendilharsu, S. (1968b). La baterı́a de pruebas de Benton y Spreen para el estudio de la afasia. Adaptación para el Uruguay y resultados. Acta Neurológica Latinoamericana, 14, 208–220. Monk, A. F. (1985). Theoretical note: Coordinate systems in visual word recognition. Quarterly Journal of Experimental Psychology A, 37, 613–625. Nichelli, P., Venneri, H., Pentore, R., & Cubelli, R. (1993). Horizontal and vertical neglect dyslexia. Brain and Language, 44, 264–283. Patterson, K. E., & Kay, J. (1982). Letter-by-letter reading: Psychological descriptions of neurological syndrome. Quarterly Journal of Experimental Psychology A, 34, 411–441. Poltrock, S. E., & Schwartz, D. R. (1984). Comparative judgements of multidigit numbers. Journal of Experimental Psychology: Learning, Memory and Cognition, 10, 32–45. Rapp, B., & Carramazza, H. (1991). Spatially determined deficits in letter and word processing. Cognitive Neuropsychology, 8, 275–311. Rapsack, S. Z., Rubens, H. B., & Laguna, J. F. (1990). From letters to words: Procedures for word recognition in letter-by-letter reading. Brain and Language, 38, 504–514. Reuter-Lorentz, P. A., & Baynes, K. (1992). Modes of lexical access in the callosotomized brain. Journal of Cognitive Neuroscience, 4, 155–164. Reuter-Lorenz, P. A., & Brunn, J. L. (1990). A prelexical basis for letter-by-letter reading: A case study. Cognitive Neuropsychology, 7, 1–20. Rey, A. (1959). Manuel de copie et de reproduction de memoire de figures géométriques complexes. Paris: Centre de Psychologie Apliquée. Seron, X., & Deloche, G. (1984). From 2 to two: An analysis of transcoding process by means of neuropsychological evidence. Journal of Psycholinguistic Research, 13, 215–236. Shallice, T., & Warrington, E. K. (1977). The possible role of selective attention in acquired dyslexia. Neuropsychologia, 15, 31–41. Shallice, T., & Warrington, E. K. (1980). Single and multiple component central dyslexia syndromes. In M. Coltheart, K. E. Patterson, & J. C. Marshall (Eds.), Deep dyslexia, London: Routledge and Kegan Paul. Wagner, N. M., & Harris, J. L. (1994). Effects of typeface characteristics on visual field asymmetries for letter identification in children and adults. Brain and Language, 46, 41– 58. Warrington, E. K., & Shallice, T. (1980). Word form dyslexia. Brain, 30, 99–112. Wilson, B. (1994). Syndromes of adcquired dylexia and patterns of recovery: A 6- to 10-year follow up study of seven brain-injured people. Journal of Clinical and Experimental Neuropsychology, 3, 354–371. Yamadori, A. (1975). Ideogram reading in alexia. Brain, 98, 231–238. Zesiger, P., & de Partz, M. P. (1994). Perturbations du langage écrit: les dyslexies et les dysgraphies. In X. Seron & M. Jeannerod (Eds.), Neuropsychologie Humaine, Liège: Mardaga.