Nrurop,sychologia, Vol. 23. No Prmted ,n Great Britain. VISUAL I. pp I-12, 1985 AGNOSIA 002X 3932/85 $3(x)+00 C 19115Pergamon Preq\ L td CONTRASTED WITH DISCONNECTION* VISUAL-VERBAL GLENN J. LARk&BEE,t$ HARVEY S. LEVIN,t F. JACOB HUFF,$ MARILYN C. KAYO and FAUSTINO C. GUINTO, JR.? tuniversity of Texas Medical Branch. @Harvard Medical School at the Massachusetts ‘IMedical College of Wisconsin at Milwaukee General Hospital and (A<,cc/Itrd 30 MU? 1984) Abstract-Serial neuropsychologlcal findings are contrasted in two cases: one with a syndrome of visual agnosia, the other with a disorder resulting from visual verbal disconnection. Both patients were impaired in confrontation naming of objects and pictures. but the patient with visual-verbal disconnection was able to perform tasks of color object matching and pantomime recognition. whereas the patient with visual agnosia could not do so. demonstrating a failure to establish meaningful nonverbal visual-visual association. Additionally. the performance of the patlent with visual agnosia reflected an evolution from the apperceptive to associative forms of the disorder, suggesting that the various impairments of visual identification form a continuum of related disorders. INTRODUCTION VISUAL agnosia is a rare disorder characterized by a failure to recognize objects visually that is not associated with aphasia, primary visual disturbance or dementia (see RUBENS [25] for an extensive review). There has been controversy over whether visual agnosia is secondary to a more pervasive cognitive impairment [ 14,23,25]. BAY [Z, 31 attributed visual agnosia to a primary visual disturbance but others have demonstrated defects of primary visual functions in nonagnosic patients that are comparable to the defects in patients with agnosia [9,16]. GESCHWIND and F~SILLO [l l] contended that most cases of apparent visual agnosia represent a confabulatory visual anomia which interferes with otherwise intact gnostic capacities. According to this view, the receptive language area of the left hemisphere is deprived of visual input by splenio-left occipital lesions, resulting in a visual-verbal disconnection. Visually mediated deficits that frequently result from such a lesion include alexia and color anomia, in addition to visual object anomia [I 11. Since publication of the study by Geschwind and Fusillo, three cases of visual agnosia without alexia have been reported that cannot be attributed to a visual -verbal disconnection [l, 16, IS]. Two of the patients also had preserved color naming in addition to normal reading, features incompatible with a visual-verbal disconnection. All three cases had evidence of bilateral occipito-temporal damage, suggesting visual-limbic disconnection. *This study is based in part on a paper presented at the 5th International Neuropsychological Society European Conference, Deauville, France. June 19x1. Preparation of this paper was supported by the Center for the Study of Nervous System Injury Grant NS 07377-I I. IReprint requests should be addressed to: Glenn J. Larrabee, Psychology Service (116B), James A. Haley VA Hospital, 13000 North 30th Street. Tampa. FL 33612. U.S.A. Recent authors [ 14. 751 have adopted LISSAWR’S [ I73 original classitication of visual agnosia into apperceptive and associative forms. The apperceptive form, usually associated with bilateral posterior cerebral disease, is characterized by a failure of both recognition and integration of the features of-a visual stimulus into a unified pet-cept. Concomitant defects include inability to relate the stimulus to previousexperience, to \,isually match objects or to copy drawings of objects. Associative visual agnosia. which can occur with either bilateral posterior disease of dominant occipital lobe and splenial lesions, is characterized by a failure of recognition (evidenced by inability to describe or demonstrate the use of objects) in the context of preserved ability to match a&or copy stimuli [25]. RURENS [35] described two patients whose deficits evolved from associative visual agnosia into visual object anemia OI “optic aphasiaâ€. which could be explained on the basis of a visual verbal disconnection. Such cases suggest that there may be a continuum of disturbance in Csuul recognition. In a11 effort to explain the range of observed deficits. RATCLIFF~. and NISYXIMBE [3] have proposed a model of ob.ject recognition, based on MARR’S computational theory of vision [IY]. In Marr’s theory. visual representation of objects proceeds through successi\.e lel,els of representation from a coding ofintensity changes in the retinal image to a three-dimensional (physical and functional representation which has access to a “semantic system†association) and a “name retrieval†system. According to Ratcliffe and Newcombe’s model. the various clinical syndromes (apperceptive and associative visual ngnosia, visual i,erbal disconnection) result from disruption at different levels of representation. The purpose of this communication is to contrast a case of resolvin g visual agnosia of the appercepti\,e type with a case ofvisual verbal disconnection. Despite the apparent similarit> of their misnaming of objects presented visually, these subjects differed in nonverbal visual recognition and matching of objects. .4lthough both patients were alexic, the cast of visual verbal disconnection had splenio-left occipital lesions, whereas the resolving case of visual apnosia had bilateral lesions. Neither patient had evidence of aphasia or dementia and their visual acuity was sufficient for the neuropsychological tasks. CASE I C.E.. a 57-qr-old. right-handed woman with hyperlcnsion and diabetes mellitus. but no previous neul-oln~ic~tl symptoms. was observed to be disoriented 111her own home on 26 June IYXI. She was unable to local fan~har objects such as light switches. and tended to bump inlo objects situated in her right \lsual field. She complained ufa headache and was temporally disoriented Two days aftcr the onset of these symptoms. she was treated in a hospital emergency room with intramuscular apresoline and suffered a brief hypotenslve syncopal episode. Subsequently. her visual symptoms persisted and her gait was unstead!. Several days later. while sllting 111a hospital waiting room. she suffered a seizure beginnmg with clonic movement of the right leg followed by Ios\ of consciousness. Blood pressure and cardiac rhythm were normal. Blood glucose was 3YO mgâ€,,. Several hour5 later. upon admission to the neurology servxe. she was alert and oriented to per‘;on. but nor to time or place. Her speech was ilucnt. with occaGonnl perseverations but without paraphasias Repetition and comprehension of speech were normal. She could not idenNy objects presented visuall) or tactually nor could she demonstrate or describe their use. She could not identify colors and did not recognve the faces ofli~mil> members. She ~asalex~c. and was unable II) manipulate a pencil appropriately in order to write (she confused the ends of the pencil). Motor responses to commands were apraxic bilaterally. more so m the right extremities. A dense right homonymous hemlanopsla was present. She could not visually track objects. but produced full saccades on command. Strength was normal. Deep tendon reflcxeh were symmetric; plantar responses were flexor. Sensation was intact except for intermittent eutlnctlon ofthe right sitlc on simultaneous stimulatton. Her gait was unsteady. wlrh a tendency IO list IO the right O+cr the lirst several hospital days visual tracking, praxis and gait became normaI. Computed tomography (CT). which will, performed 7 daq after the onset of illness.demonstrated decreased attenuation in the arcas ofthe Icft occlpltal lobe and the anterior limb of the internal capsule on the right. Figure I (upper row) deplcls II CT scan ohtalned 20. I months post-onset (GE XX00 scanner, wlthout contrast) which was compatible with cerebral infarc:ions. fhe clectroenceph~~lo~rarn (EEG) was biremporally slowed. with maximal slowins in the Icft mldtemporal arc<i. VISUAL A<iNOSIA FIG. 1. Top. Case C.E. Multiple contiguous CT slices show areas of low attenuation in the anterior limb of the right internal capsule and the left occipital lobe. consistent with infarction. Bottom. Case M.W. Contiguous CT cuts showing a large surgical defect through the left half of the splenium extending mto ipsilateral occipital lobe. Residual calcification is just anterior to splenium. F-I<; 2. (‘ax M.W. Coronal CT reveals a large defect of left \plenium lobe. and medial aspect of occtpttal VISUAL AGNOSIA 5 Evaluation of visual acuity was difficult because of alexia and problems in visual recognition. Within 2 weeks of her infarction, acuity was estimated to be at least 20/400 by optokinetic nystagmus (OKN), and at least 20/100 by her response to eye chart stimuli (although she could not read the stimuli, she could indicate when they began to blur). At 2 months OKN (using a smaller tape) revealed at least 20/100. At 5 months status post, resolution 01 apperceptive deficits allowed training her to discriminate circles. triangles and squares, which was achieved at 2Oi70. Tactile sensitivity (pressure aesthesiometer) was relatively preserved. CASE 2 W.M., a 62-yr-old, right-handed man presented with a total right homonymous hemianopsia. total alexia without agraphia, object and color naming disturbance and visuoconstructive difficulties following surgical division of the splenium and the posterior 2.5 cm of the body of the corpus callosum to remove a left intraventricular meningioma. Prolonged retraction during surgery produced necrosis in the left occipital tip, which was amputated 3.5 cm forward from the occipital pole. W.M.‘s visual cognitive deficits have remained stable over the 4 yr since his nearly complete visualLverbal disconnection developed after surgery 1151. Details of the patient’s preoperative and early postsurgical neuropsychological and neurological findings have been reported 1151. A CT scan with contrast (GE 8800) obtained 4 yr after surgery (Figs 1 lower row and 2) demonstrated the dominant occipital and extensive splenial damage. Visual acuity, assessed by having him specify the orientation of the open side of a rotated letter “Eâ€. was estimated at 20/50 for O.D. Acuity in OS.. assessed by OKN, was at least 20:70. GENERAL TESTING PROCEDURES The test data for subjects C.E. and W.M. are presented in Tables 1 and 2, respectively. Data on object and picture naming are presented separately in Fig. 3. and data on pantomime recognition and colorPobject matching are presented in Fig. 4. References for published tests are cited in Tables 1 and 2 and Figs. 3 and 4. The object naming test for confrontation visual naming, tactile naming and for the condition of visual confrontation naming while holding the object was from the Neurosensory Center Comprehensive Examination for Aphasia 1271. as were the writing-to-dictation and from-copy tests. Visual naming of pictures and reading comprehension of words and phrases were from the Multilingual Aphasia Examination [S]. The visual visual matching procedure, which was a series of pictures from the discrimination section of the Continuous Visual Recognition Memory Test [I 31. consisted of discriminating a picture which matched a target stimulus (e.g. a beetle) from five visually similar foil stimuli (other insects). The auditory naming task [ 151 was a verbal analog of the visual naming of pictures test (e.g. “What is the name for a body of land surrounded by water‘?“). The tests of facial recognition and three-dimensional block construction were developed by fkNToN et rd. 16. 73. The verbal memory for colors test required the subject to state the colors of IO common objects (e.g. banana. snow, etc.) which the examiner named out loud. Color sorting and color naming were conducted using two sets ofeight jumbo crayons for matching-to-sample and confrontation naming. The various reading tasks (letters. numbers. words vs nonwords) were constructed by hand printing the stimuli on X by 5 in. sheets of paper. All performances are presented as percentage correct, since some of the performances, such as confrontation naming were so impaired that they fell far below the first percentile of published normative data. Defective performances were defined as those scores falling below the fifth percentile of the score distribution based on normal controls. RESULTS Visuul reco~nitim of djects C.E.‘s visual identification of objects markedly improved and changed qualitatively over the course of the study (Fig. 3). Initially (IO days post-infarction), she could not name common objects presented visually (e.g. bottle) nor could she describe or demonstrate their use. Imagined use of objects and symbolic gestures to oral command were intact bilaterally, excluding ideomotor apraxia as a basis for her inability to use actual objects. The anemic errors were frequently unrelated to the object, either semantically or by similarity of visual form. She was unable to copy drawings or visually match-to-sample common objects such as a brush. These features were consistent with apperceptive visual agnosia. By 4X days postinfarction C.E. could match pairs of common objects on the basis of visual cues while verbal identification by naming or description remained impaired. Thus, her visual agnosia apparently evolved from an apperceptive to an associative type. Naming to auditory description (e.g. “What is a large grey animal with floppy ear5 and a long nose?†.Ans\vcr: “Elephantâ€) was markedly better than visual identification. Tactile naming of common objects was impaired bilaterally initially (despite normal sensory thresholds to nylon filaments), later changing to a left-hand anemia and finally. preserved naming in both hands. At nearly 1 yr post-infarction, she continued to misname objects presented visually but could improve her naming performance if allowed to hold the object during visual inspection. Naming objects under the condition of combined tactile and visual cues. however. was notably slowed with occasional errors. Pointing to objects named by the examiner improved considerably on the examination nearly 1 yr post-infarction. At almost 1 yr post-onset. C.E. was able to name objects. but still performed at a defective level OII confrontation naming of pictures. As can be seen from Fig. 3. W.M.‘s naming of objects on the basis of visual cues has remained impaired. In contrast. verbal identification of objects which he held during visual inspection was rapid and accurate. Most of W.M.‘s anomicerrors OII confrontation naming VISUAL Table 2. Visual. cognitive and related 1 AGNOSIA neuropsychological functions Months post-surgery 20.9 49 success rate unless otherwise stated) I .4 Test (Figures Tactile naming [27] Right hand Left hand Visual naming of objects, allowed to hold and manipulate 1271 Visual visual match of pictures [13] Auditory naming [IS] Facial recognition 16. 71 3-D block construction [6] Dvorine (tracing) [X] Farnsworth Munsell [lo] Verbal memory for colors Color sort Color name Reading camp.. words and phrases Reading: letters numbers words words vs nonwords Writing: Dictation [27] From copy indicate 9.9 percent 75t 371- I 00 62t x7.5: 100: 100: x5: _~ 71 X9.7.t 71 100 100: 100: 20: 15t 10: [S] 100: 33: 43: .~ 0: 0: Preserved Defective lOOgo60_ IO0 671IO0 100: 79.6 Preserved Defective *Tests without references were devised in the course of clinical examination. tDefective performance relative to normal controls. $No data on normal controls. but most literate subjects would be expected ~ for W.M. ---- Preserved Defective to approximate 69? I 00 197 errorst 100: 100: 25: 16.7t 16.7: 16.7: O$ 100: Preserved Defective a perfect score. 0 VNP. Visual Naming, Pictures @.VNO. Visual Naming. Objects CE, Visual Agnosla VNO WM. Visual-Verbal Dlsconnectlon * Defectwe Performance P) 7om : 6Om 0llSZt 1.6 31 Months FIG. 3. Confrontation 109 Post-Onset naming 20 9 performance. were semantically related to the test object, other objects presented earlier during the examination or to his previous responses. W.M.‘s persistent inability to point to objects named by the examiner has paralleled his residual anomia for visual presentation. Visual--visual matching-to-sample of common objects (e.g. razor) has been consistently preserved. Description and demonstration of the use of objects presented visually were frequently misdirected by his anemic errors. Naming objects described by the examiner (e.g. “What is a body of land surrounded by water called‘?†Answer: “Islandâ€) in the absence of visual presentation was preserved. Naming objects palpated by the right hand in the absence of visual cues was intact, whereas there was a mild anomia for objects placed in the left hand. The onset of C.E.‘s cerebral vascular insult was followed by a period of resolving prosopagnosia. Although she could recognize her family when examined IO days postinfarction. she could not identify facial photographs of famous persons by naming them 01 selecting the target face from a multiple-choice array. By 3 months. she could identify famous persons in ;I multiple-choice format. although confrontation naming of these persons remained impaired. This did not change on examination almost 2 yr following her infarction. In contrast, although W.M. had difficulty in confrontation naming of familial faces, at no time did he have difficulty identifying familiar facts in a multiple choice format. This dissociation extended to unfamiliar taces. W.M. was able to match photographs of unramiliar laces [7], while C.E. could not. Both W.M. and C.E. were impaired in constructing designs from three-dimensional models [S] throughout the serial examinations. However. W.M. performed better relative to C.E. Disturbances in color recognition can occur in visual agnosia [?I] and in alexia without agraphia 141. Both C.E. and W.M. displayed preserved color perception [X] and verbal color memory (e.g. “What color is a banana?“) despite markedly impaired confrontation naming of colors. In contrast to C.E.‘s marked impairments in ordering hues according to their saturation [IO] and matching colors to line drawings ofcommon objects with characteristic colors [7Y], W.M.‘s Farnsworth Munsell performance, though impaired, was considerably better relative to C.E., and he achieved a normal level of color-object matching when instructed to suppress his misleading verbalizations of the objects. For example when he first attempted to match colors with line drawings of objects, he would try to name the color and the object. When instructed to perform the task without verbal response (by pointing) performance was normalized. Both patients’ color object matching performance is displayed in Fig. 4. Keudirty und pantomime recognition Both C.E. and W.M. had alexia without agraphia. W.M.‘s total alexia extended to letters. \vords and numbers. and improved minimally during the 4-yr study period. C.E.‘s alexia was also Initially total, but evolved within 3 months to a verbal alexia which persisted during the 20-month follow-up period. At 20 months, she was able to figure out short words (e.g. “ballâ€) by spelling them aloud in a letter by letter fashion. True reading of words did not COM, Color~Ob~ect Matching PR. Panromlme Recognltlon 100, 904 G ; 801 70 6 60 i E ,†50-1 40 2 20 10 30 I - CE. “,$“a, Agnom ___-WM. Visual Verbal D~sconnectmn VISUAL AGNOSIA 9 return, however. Interestingly, follow-up examinations disclosed that both C.E. and W.M. were able to correctly discriminate a word from two accompanying nonword anagrams (e.g. “in†vs “dsâ€, “mgâ€) by pointing, although neither could read the real word. The number of correct responses varied from 6 out of 10 to 10 out of 10 depending on the type of task. That is, W.M.‘s discrimination was maximal when the nonwords did not contain vowels. He could also distinguish words from nonword homophones (e.g. “coat†vs “toteâ€) at a level of 9 out of 10 correct, but he could not perform synonym or antonym matching tasks nor could he match simple words with objects (e.g. “baby†with a toy doll). C.E. demonstrated the same pattern of performance in discriminating words from nonword homophones (8 out of IO correct). Her word-nonword discrimination was also maximal when the nonwords did not contain vowels. Matching pictures of common objects while the examiner pantomimed their use (e.g. aiming a rifle) [28], was normal for W.M. but impaired for C.E. Pantomime recognition performance for both patients is displayed in Fig. 4. DISCUSSION The similarity in the inability of both patients to identify visually presented objects verbally and in their initial levels of alexia contrasts with the divergence in their nonverbal visual recognition performance. W.M. could not connect a name or description with its visual referent, whereas he appropriately used common objects and he could match them with their characteristic color and pantomimed use. In comparison, C.E. could not initially demonstrate the use of objects nor could she match identical objects on the basis of their visual cues, interpret pantomimed object use of perform color-object matching. This constellation of features suggests the C.E. was unable to link the contents of her visual perception with previous experience. The failure of object matching suggests, further, that the perceptual aspect of recognition was impaired. These inferences are supported by clinical observations of C.E.‘s visual disability during the subacute stage after her cerebral infarction: she required assistance in using familiar implements and behaved as if “blindâ€. C.E.‘s initial visual disturbance, which included prosopagnosia and topographic disorientation, was compatible with the apperceptive type of agnosia [ 171. While C.E. was initially more visually impaired in daily activities than W.M., her alexia partially resolved during the early months of her convalescence. C.E.‘s clinical course was characterized by marked improvements in visual matching performance, her ability to utilize combined visual and tactile cues for object naming at 11 months post-onset and eventually, her accurate object naming at 20 months post-onset. Picture confrontation naming remained impaired when examined 20 months after the onset of her symptoms. C.E.‘s residual visual disturbance at 11 months post-onset was compatible with the associative type of visual agnosia described by Lissauer. This inference is supported by her recovery of visual matching despite residual defects in meaningful visual-visual associations, i.e. inability to match pictures of objects to their corresponding color and pantomimed use. These sequelae would not be predicted by a visual-verbal disconnection [1 I]. At 20 months post-onset, although object naming had recovered, she still experienced difficulties in picture naming, matching of unfamiliar faces, color-object matching and pantomime recognition. Thus, although her object naming performance was far superior to W.M.‘s, he performed a variety of complex visual cognitive tasks at a much higher level than C.E. We infer that different mechanisms were responsible for each patient’s performances, i.e. C’.E.‘s deficits were due to an impairment in visual recognition, whereas W.M.‘s deficits lvere due to ;I disconnection of generally prescrvcd visual rccogniti0n functions from language functions. W.M.‘s alexia without agraphia is characteristic of previously reported cases 01‘ this clisorcler productd by splellio-occipitnl lesions [I 21. As might be predicted from the surgical division of his splenium. W.M.‘s aleuia was more complete and permanent than the reading disturbance typically produced by left posterior cerebral artery disease. Although W.M shuwd minimal recover! d spontaneous reading. he could distinguish between words and random combinations of letters. In view of the surgical division of W..M.‘s splenium and his unresolved dense right how can his residual capacity to diacrimmate \~ords l‘rom homonymous hemianopsia, nonworda be espluined‘! Consistent with previous studies of patients who have undergone section of the corpus callosum to relieve epilepsy 1761. wc postulate that W.M.‘s residual lexic capability is subserved bq his right hemisphere. In fact, UC obser\cd that 111s spontaneous verbalizations frequently disrupt the silent linguistic operations performed by his ncmdominant hemisphere. When asked to point to the picture corrcspondin~ t~l an object named by the examiner, W.M. was frequently led astray by his anemic errors. This view is supported by W.M.‘s report of misinterpreting street signs in his neighborhood. He would correctly comprehend the signs initially. only to reJect his interpretation in favor ol verbalized paralexic errors. The reinterpretations generally had negative or restrictive effects (e.g. misinterpreting ‘I sign as reading “Keep off’) and controverted his past experience. The contrastin: clinical courses in C.E. and W.M. suggest that diffcrcnt mechanisms ot recovery were involved. In accord with the prr\ailing concepls of prosopagnosia, C.‘.E.‘s acute visual disturbance ma\; bc attributed to bilateral disconnection of \:isual input to the infcro-medial temporal lobes [X]. While prcsumptivc damage to C_‘.F. ‘s right OCCI~I~Otemporal connections could not be conti~-med h\ computed tomography. it is possible that she suffered transient ischemia or it snnaII infarct of the postcrivi- right hcmisphet-c. citlli‘r 4‘ which may bc undetectable by CT scanning. liecovcry from impairment in the postwar circulation of her right hemi<pherc may account for the evolution of her \ iaual :ugnosia from the appercepti\e In the associati\c type. C‘.E.‘s uI~resnlved dcnsc I-i&t honion~mous hemianopsia suggests that clinicxl improvement \\a5 unrelated 10 changes in her lei‘t posterior circulation. A\sociatiLe visual agnosia hux generally been attributed to bilateral postcrtor cercbrx1 hemisphere Iesions [I. 16. 1311,However, kR7I:SL [ 141 has reported a ca\e in which \ ihua1 agnosia persisted for over 10 yr following a head injury, and a CT scan showed lesions in the left occipital lobe and deep in the right frontal lobe. probably representing coup and contrecoup inJuries. This patient \vas similar to (‘YE. in many respects. I-fer ability to identil‘h objects presented visually was severely impaired, although she performed well in discriminnting si/e and shape differences 111objects and drawings. Because the patient exhibited prosopagnosia anti ;I be\cre memor!~ deficit. Kertesz >u,,~~w~ted that bilateral parlctooccipital Icsions ma4 have been present, cvcn though no right pal-icto-occipital Iesions \vcre detected on C’T scanning. Ho\+~evcr. since in both Kertes/‘s patient and our patient C’.E. ;I visual agnosia is correlated with 1cft occipital and deep right frcmtal Icsiona. thr possibility must be considered that this combination of lesions may be sufficient to produce visual agno5ia. A possible mechanism for the contribution of deep frontal lesions to agnosia i\ suggested bj, the recent discor/crh that projections from basal for-ebrain nuclei (principally. the nucleus basnlis of Meyncrt) prcrvidc most of the cholinerglz input to the cerebral cortex [2 I. 241. 7 1~ VISUAL AtiNOSIA II nucleus basalis is situated ventral to the internal capsule, and in the rhesus monkey some of its neurons are often found among the fibers of the internal capsule [31]. Although the projection pathway from nucleus basalis to cortex in man is not known, it is possible that lesions in deep frontal white matter may interrupt cholinergic fibers projecting to posterior cerebral cortex. Cholinergic deficiency in the cerebral cortex has been well documented in Alzheimer’s disease, in which visual agnosia is a common late and occasional early symptom [22,24]. Hence, a deficiency of cortical acetylcholine in posterior right cerebral cortex may contribute to the visual agnosia in patients such as Kertesz’s and our case C.E. This hypothesis is testable by means of postmortem analysis of regional cholinergic enzyme activity in cerebral cortex. In conclusion, our findings clearly distinguish between visual--verbal disconnection and visual agnosia. Our results in W.M. show that complete deprivation of visual input to the language dominant hemisphere does not impair nonverbal identification of objects and persons, although the “blind†speaking hemisphere can disrupt the lexical performance of the nondominant hemisphere. Secondly, our results suggest that visual agnosic deficits fA along a continuum rather than into discrete clinical entities and that considerable evolution of agnosic symptoms can occur, ranging from appcrceptive visual agnosia to associative visual agnosia to recovery of identification of real objects in spite of persistent agnosia for pictorial materials. This evolution highlights the importance of longitudinal examinations in the evaluation of visual agnosia. ilc.X-,~oIl./~,r/~/mnc~n~r~~the authors Marsel Mesulam. The assistance gratefully acknouledgc the helpful comments of David N. Levine and Marekof Christina A. Meyers. who helped test the patients, is also appreciated. REFERENCES AI.REKT. M. L., SOWR, D.. SILVFRRFKG. R. and RWHES. A. The anatomlc basis ofvisual agnosia. Nruroloy~~29, X76 X7Y. 1979. ~rnt//ir,~X-rio,rs~~~crncl~~/. Springer. Berlin. 1050. Cited by E ITI.I~~C;EK. G. Sensory defects in visual 2. BAY, E. .4$qrro.vrc~ agnosia. J. Neural. Neurosurg. Psychiat. 19, 297~307, 1956. 3. BAY. E. Disturbances of visual perception and their examination. Br<rirr 76, 515 ~550, 1953. iisuospatial and visuoconstructive disorders. In Cliuicul Ncuro~~s?ch~,/o~/~,. K. 4. BENTOK. A. L. Vlsuoperceptive. M. HEII.MAN and E. VAI.P~‘S’IEIN(Editors). New York. Oxford University Press. 1979. of Neurology. 5. BENTON. A. L. and HAMSH~K. K. deS. ,Mul/ili+/rrtr/ APIIN.FI’O~.\~rrninn/iorl. Department University of Iowa Hospitals and Clinics, Iowa City. 197X. 6. BENTON, A. L.. HAMSIJER. K. deS., VARNFY. N. R. and SPREEN. 0. Contrihutiorl.s 10 Ric,urop.s~,cllo/oc/Ical .4ssc~.ssrn~~~7/. .4 C‘lirlicul hfanutrl. Oxford Umversity Press, New York. 19X3. Recognition, form SL. 7. BENTOV, A. L., VAN ALI.EN, M. W., HAMSHER, K. deS. and L~:~IN, H. S. Test of’ Facial Department of Neurology. University of Iowa Hospitals. Iowa City, 1975. Plates. 2nd edn. Harcourt, Brace and World, New York, 1953. 8. DVORINE, I. Duorine Pseudo-isochromatic 9. ETTLIYGER, G. Sensory defects in visual agnosia. J. Neural. Neurosurg. Psychiat. 19, 297 307, 1956. IO&Hue Test Manual. Munsell Color, Baltimore, 1957. 10. FARNSWORTH, D. The Farnsworrh-Munsell Il. GESCHWIND, N. and FUSILLO, M. Color-naming defects in association with alexia. Arch. Neural. 15, 137~-146, 1966. 12. GREENBLATT. S. H. Neurosurgery and the anatomy of reading: a practical review. Neurosurqer~ 1.6 15. 1977. 13. HANNAY. H. J.. LEVIN. H. S. and GROSSMAN. R. G. Impaired recognition memory after head injury. C’ortu 15, 269- 283. lY79. 14. KERTESL. A. Visual agnosia: the dual deficit of perception and recognition. Cortex 15, 403%419, 1979. in a musician after transcallosal removal of a left 15. LEVIN. H. S. and ROSE, J. E. Alexia without agraphia intraventricular meningioma. Neurosurgery 4, 168-174, 1979. and visual object agnosia: a behavioral study. Brui,~ Lanq. 5, 341-365. 1978. 16. LEVINE. D. N. Prosopagnosia nebst Beitrage zur theorie derselben. Arch. Ps,chiat. Newen. 21, 17. LISSAUER. H. Ein fall von Seelenblindheit 222-270. 1889. Cited by KF.RTESZ,A. Visual agnosia: the dual deficit of perception and recognition. Cortr.u 15, 403 419. 1979. I. 12 GI I NU J. LAKKABII rt ul. IX. MA(.R, J. L. and &II I.TR. F. Associative vlaual agnos~a and its related deficits: the role ofthe mmor hemqhere in assigning meaning to visual perceptions. .~~~u~op.c-,~c~/io/r~c/rc/ IS. 34.5 349. 1077. 19. M8\~a, D. Visual information processing: the structure and creation of visual representations. Phrl. 7‘~m.s. R. SW. Land. Series B 290, 199 218, 1980. Cited by R,\ I(I IFSI. G. and N~,\rc.oz~n~, F Object recognition: some deductions from the clinical evidence. In .l’<~rv~u/i/\, ~nll Porlroloql~ 111Coqnitir (’ Fu,~~~/~~~II. .4 W. EI 1.1s(EdItor). Academic Press. New York, IYX?. 20. Mr hoows, J. C. The anatomical basis of prosopagnosia. J. ‘L’eud. K;ruwsury. Psyhut. 37, 189 501. 1974. ?I. MI,SULAM. M. M.. MUFSON. E. J.. Lr VI y. .4. I. and WAIyt:K. B. H. Cholinergic innervation ofcortex by basal forebrain: cytochemistry and cortical connections of the septal area, diagonal band nuclei. nucleur hasahz (substantia innominata). and hypothalamus in the rhesus monkey. ./. cor~p. :VCUI.O/.214, 170 197. IYX.1 21. Nlssl:u. M. J.. CORKI\, S.. <~Ro~~xn\, .I.. WKAI’, S and BAYER. J. Spallal contrast sensitivity in Alzhelmer‘h disease. Presented at the Association for Research in Vision and Ophthalmology meeting. Sarasota. Florlcla. May 19x3. 23. RAT(.LIF~. G and NriwcUMat:. F Object recognition: some deductions fl-om the ctinlcal c\ Idence. In .~~~~vtrl~rt. tr~ri F’crrlrob~y,, in Coq,~i/ir-cp F~r,c/&, A. W. EI 1.1s (Editor). Academic Press, New York. IYX?. 24. ROSVIR, M. N., G~RKEI I, N. J.. Jo~~Nso~, A. L.. Mu,\\ I-IOY.C. Q., Ro-rtr. M. and I\ t KSO\,L. L. A post-mortem study of the cholinergic and GABA systems 111senile dementia. Brc~r~ 105, 313 330. 19X2. 25. Rr nt~\, A. B. t\gnosla. In Clinicd Xeuropsdwloyy, K. M. HI II.VAN and E. V‘\I I,\\IF IN (Editors). Oxford Umversity Press. New York. 197’). 70. SIIII IS. .I J., Vor.l>t.. B. T.. WILSON. L). H. RAYPOKI. M. and <;A(LA~I(;A. M. S. Vurlabiht! in right hcmispherc language function after callosal section: evldcnce for ;I continuum of generative capactty .1. l<~~r~~.r~~i.I. 323 331, 19x1. ~‘ll/ll/~rc~llc’lr.\,I <’ fI‘vi/lr!i/rtrti(Jl, /(I,_ .lpirr:wr. 27. SPRI+N. 0. and Bt,h’ToY. A. L. :V~~rro.\c,,v~,,:i, (‘?,lil,l Neuropsychology Lahoratorq. Department of Psychology. Ilniverslt) of Victoria, Victoria. B.C. 1969. ?X ~‘ARNI3’. N. R. Linguistic correlate\ of pantomime recognition in aphasic patlents. .I. .V~,UU~/ %‘cir~o.crrni P.v,d1itrr.41, 564 56X. lY7i