Atypical Conduction Aphasia A Disconnection Syndrome Mario F. Mendez, MD, D. Frank Benson, MD \s=b\ Conduction aphasia was originally proposed to result from separation of the posterior language comprehension area and the anterior motor speech area of the left hemisphere. The arcuate fasciculus has been the most frequently suggested site of such a disconnection, but the syndrome has been reported in cases in which the abnormality involved the domi¬ nant Wernicke's area. This challenges the arcuate fasciculus theory, and it has been suggested that a cortical lesion, not a disconnection, is the crucial factor. Three new cases in which the lesion does not lie in the arcuate fasciculus are reported, two in left-handed patients with left temporoparietal lesions and one in a righthanded patient with a right temporoparietal infarct, a "crossed" aphasia. While atypical, these cases offer evidence that disconnection of the circuit linking lan¬ guage comprehension to motor speech output, not damage to a specific cortical region, underlies the syndrome of con¬ duction aphasia. (Arch Neurol 1985;42:886-891) T^he existence of conduction aphasia, a disorder characterized by dispro¬ portionate difficulty in the repetition of spoken language, was originally proposed by Wernicke1 on purely theo- Accepted for publication Aug 30, 1984. From the Department of Neurology, UCLA School of Medicine, Los Angeles. Dr Méndez is now with the Veterans Administration Center, Providence, RI. Medical Read in part before the 35th annual meeting of the American Academy of Neurology, San Diego, April 27, 1983. Reprint requests to Neurology Service, Veter¬ ans Administration Medical Center, Providence, RI 02908 (Dr Méndez). retical grounds. He postulated that any lesion disrupting the principal pathways connecting the site of audi¬ tory language comprehension and the site of motor expression of speech should result in a characteristic apha¬ sie disorder. The postulated language disturbance was subsequently de¬ scribed clinically,25 and although ini¬ tially considered rare,6 recent reports suggest that conduction aphasia may constitute up to 10% of classifiable aphasies.7 Wernicke's original expla¬ nation for the aphasie syndrome has not been universally accepted, howev¬ er.8 Several alternate explanations, eg, Goldstein's9 theory of a disturbance of a central language area or recent pre¬ sentations that suggest that damage to a specific language center10 or to the cortex of the parietal operculum,8 have been suggested for this syn¬ drome. One important step toward under¬ standing the mechanism underlying conduction aphasia is location of the neuroanatomic site or sites of lesions that produce the syndrome. Most cases reported in the literature show abnormality near the posterior rim of the left sylvian fissure, either tempo¬ ral or parietal.9" The usual lesion has been suprasylvian, involving the pari¬ etal opercular area (supramarginal gyrus) with many cases showing dis¬ turbance of the arcuate fasciculus, a fiber bundle coursing between tempo¬ ral and frontal association cortices. Arcuate fasciculus abnormality ap¬ pears consistent with Wernicke's1 dis¬ connection hypothesis. Based on this and the frequency of its occurrence with this disorder, interruption at the Downloaded From: http://archneur.jamanetwork.com/ by a New York University User on 05/15/2015 arcuate fasciculus can be considered "typical" model of conduction aphasia. the Other locations of neuroanatomic damage have resulted in this language disturbance. Destruction of the left temporal region, including Wernicke's area, a lesion that would characteris¬ tically produce impaired comprehen¬ sion, has been reported in a number of cases in which a defect in repetition was the major finding.1214 As one of the authors (D.F.B.) has previously pointed out," a lesion involving a sin¬ gle neuroanatomic location cannot explain all cases of conduction apha¬ sia. Three additional cases of patients with conduction aphasia who do not have a lesion in the arcuate fasciculus reported. They provide an oppor¬ tunity to reexamine the validity of the disconnection theory of conduction aphasia and to compare it with other hypotheses. are REPORT OF CASES Case 1.—A 60-year-old left-handed man with a strong family history of sinistrality presented with an acute onset of speaking disturbance and right arm and leg numb¬ ness of 12 hours' duration. There was a well-documented history of recurrent supraventricular tachycardia secondary to Wolff-Parkinson-White syndrome, a histo¬ ry of hypertension, and a left nepbrectomy performed 15 years previously for renal cell carcinoma. Blood pressure was 120/80 mm Hg, pulse rate was 80 beats per minute and regular, and respirations were 24/min with some respiratory distress secondary to conges¬ tive heart failure (CHF). There was an S, gallop, jugular venous distention, and moist rales heard at the lung bases. Neurologic examination revealed an alert man with a mild aphasia. The corneal reflex was decreased on the right and deep tendon reflexes were slightly increased on the right compared with the left. Primary sensory modalities were mildly impaired in the right upper extremity. Language examination demonstrated a fluent verbal output with word-finding pauses, circumlocutions, and increasing approximations in approaching the missed word. There was a strong tendency for phonemic substitutions (literal paraphasia), eg, "bundle" for "buckle" and "Mis¬ ter" for "crystal." The patient made many attempts to self-correct his paraphasic Spoken language comprehension was totally normal, even for complex multistep commands and the understanding of relational words. Repetition of spoken lan¬ Fig 1.—Case 1. Computed tomographic scans showing large left temporoparietal hypodense errors. was considerably abnormal with marked hesitations and multiple literal paraphasias. Number repetition revealed frequent semantic substitutions. Confron¬ tation naming was done well, except for paraphasic errors. Comprehension of writ¬ ten language was excellent, but both read¬ ing aloud and writing were contaminated by phonemic substitutions. All verbal com¬ mands were carried out without difficulty, the patient showing no evidence of ideomo¬ tor apraxia or body part substitution. The admitting diagnosis was embolie stroke. Initial laboratory data were unre¬ markable except for left ventricular hyper¬ trophy on electrocardiogram and evidence of mild congestive heart failure on chest guage roentgenogram. Computed tomographic (CT) scan was normal at the time of hospi¬ tal admission, but a repeated scan a few days later demonstrated a low-density area involving much of the left temporoparietal region (Fig 1). Heparin sodium anticoagulation and digoxin therapy were initiated. The hospital course was compli¬ cated by hypertensive episodes, runs of supraventricular tachycardia, and the dis¬ covery of mild hyperthyroidism. Reexamination at the time of discharge from the hospital and again as an outpa¬ tient one month after onset demonstrated improved language function with fewer word-finding difficulties and literal para¬ phasias on spontaneous speech, but contin¬ ued difficulty in repetition. Case 2.—A 26-year-old strongly lefthanded man noted the sudden onset of word-finding difficulties and frequent mis¬ pronunciation of words unaccompanied by any other symptoms. Three to four days later he noticed tingling paresthesias of the distal part of the right arm and leg and was subsequently admitted to the hospital for evaluation of his condition. Medical history proved unremarkable, except for a recent dental abscess. There was no family history of left-handedness. Results of general physical examination at the time of hospital admission were normal, but the neurologic examination revealed a right upper quadrantopsia and a flattened right nasolabial fold. Motor test¬ ing demonstrated mildly decreased fine motor movements of the right hand. Deep tendon reflexes were slightly more active on the right side, but the toes were Fig 2.—Case 2. Computed tomographic scans demonstrating hyperdense, enhancing area with surrounding area of edema, ultimately demonstrated as arteriovenous malformation. Left, parietal involvement; right, temporoparietal junction involvement. Fig 3.—Case 3. Computed tomographic scan demonstrating large right parietal hypodense area with extension to posterior temporal area. downgoing. No sensory deficits were found. Language was strikingly abnormal. Word-finding pauses and phonemic substi¬ tutions were present in conversational speech, confrontation naming, and reading aloud, but the greatest difficulty occurred when he was asked to repeat words or phrases. There was a strong tendency to self-correct paraphasic errors, especially by increasing phonemic approximations. Typical examples of his paraphasic errors included "wait" for "walk" and "rick" for "wrist." Comprehension of both oral and written language was tested carefully and was unimpaired. He understood complex sentences, grammatical words, and multi- Downloaded From: http://archneur.jamanetwork.com/ by a New York University User on 05/15/2015 step commands. His writing showed mis¬ spellings and paragraphic substitutions. A mild difficulty copying three-dimensional drawings was demonstrated, but no cogni¬ were found; he performed flawlessly on all praxis tests. A CT scan on the day of admission disclosed a hyperdense and peripherally enhancing lesion with surrounding edema located in the left temporoparietal area (Fig 2). Because of the recent tooth infec¬ tion the patient was treated with intrave¬ nously administered penicillin and metronidazole until an angiogram was perform¬ ed. Angiography showed a left middle cere¬ tive deficits bral artery terminal branch arteriovenous malformation that had bled. After much discussion it was the decision of the patient and his family not to proceed with any therapeutic procedure but to continue with close clinical follow-up and serial radiologie studies. He was discharged from the hospital one month after onset having improved to the point of only a mild resid¬ ual language disturbance. Case 3.—A 58-year-old right-handed man, visiting at a friend's house, experi¬ enced the acute onset of a right temporal headache and an inability to speak proper¬ ly. His friends noted garbled speech and nonsensical phrases but believed this to be an effect of alcohol. The next morning he became frustrated when unable to speak coherently over the telephone and he drove himself 64 km to a hospital for evaluation. The patient was strongly right-handed and favored the right foot as well. There was no family history of any left-handedness. General examination revealed a welldeveloped man in no acute distress. No bruits were heard but the left carotid pulse was less forceful than the right. Peripheral pulses were absent in the right lower extremity. Neurologic examination showed a left hemi-inattention with extinction on double simultaneous stimulation and decreased swing of his left arm when walk¬ ing. Language output was fluent with normal artery with some enlargement of collateral vessels. The aphasia improved during the hospital course, with comprehension returning to normal, but the marked inability to repeat spoken words or phrases remained. COMMENT or These patients met established clin¬ ical criteria for conduction aphasia7; they showed fluent paraphasic con¬ versational speech, fairly normal oral comprehension, and a disproportion¬ ate disturbance in the ability to repeat spoken words. Spontaneous speech, confrontation naming, read¬ ing aloud, and particularly repetition were contaminated by literal para¬ phasias. In contrast to many reported cases, these three patients demon¬ strated considerable awareness of their errors, attempted self-correc¬ tion, and the conduit d'approche phe¬ of increasing phonemic nomenon approximations was notable. Repeti¬ tion of numbers was often failed because of semantic or whole-number substitutions. Reading comprehension was good, but writing had spelling and Reading comprehension was considerably errors and misspellings. There was a notable left-sided neglect including a "dressing disturbance" of the hemineglect variety, a failure to dress his left side. He was unable to copy threedimensional drawings nor could he locate areas on a map. He showed persistent difficulty in finger recognition and an acalculia. In contrast, there was no evidence suggesting visual agnosia, prosopagnosia, or amnesia. Results of all tests for ideomo¬ tor apraxia were normal. A CT scan demonstrated a large, nonenhancing, hypodense lesion in the right parietal area with extension to involve some portions of the posterior temporal region (Fig 3). Three serial CT scans did not show any lesion in the left hemisphere. Arteriography revealed absence of posteri¬ or branches of the right middle cerebral Establishment of the disconnection theory for conduction aphasia could be of importance. It would make con¬ duction aphasia one of the few lan¬ guage disorders with a truly under¬ aphasia. cal in location of the lesion and none could be explained by an interruption of the arcuate fasciculus pathway. The first two involved Wernicke's area of the left hemisphere, similar to the case reported by Kleist,14 while the third is a case of "crossed" conduction better than reading aloud, which was also contaminated by literal paraphasias, and his writing contained multiple paragraphic atypical with the pertinent motor areas.1718 While apraxia is frequently reported in conjunction with conduc¬ tion aphasia," dissociation of the two disorders is also reported," suggesting that separate routes may be available from language comprehension to sion of both oral and written material was nized his errors and tried to correct them. Examples included "botel-butel-buckle," "caid" for "pain," and "pirn" for "pills." sia must account for the cases. In contrast to the typical clinical picture, these three cases were atypi¬ grammatical errors. partially compromised at. the time of hos¬ pital admission (he was unable to perform two-step commands), but this improved rapidly so that only minimal difficulty in the comprehension of complex material could be demonstrated within a few days. In contrast, repetition was disproportion¬ ately impaired and remained abnormal during hospitalization. He often failed even single-word repetitions. Attempts to name were characterized by pauses and literal paraphasias, although he recog¬ refute the disconnection model. Any explanation of conduction apha¬ standable theoretic mechanism. It would also lend support to the basic Wernicke1 model for the organization of language in the brain: two major language centers, one for verbal com¬ prehension and the other for verbal expression, linked by a direct pathway with additional circuits for the media¬ tion of semantic meaning. One additional neurobehavioral ob¬ servation, the presence or absence of accompanying ideomotor apraxia, de¬ serves attention in this discussion. The ability to carry out actions follow¬ ing verbal command has been said to depend on pathways connecting the grammar and prosody but contaminated by multiple literal paraphasias and fre¬ quent word-finding pauses. Comprehen¬ aphasia. Alternate postulations have been proposed on the basis of new information available concerning the pathophysiology of the disorder. Our three atypical cases, along with others reported in the literature, offer data that can be used to substitute, modify, Wernicke's original theory for the mechanism of conduction aphasia proposed a separation of the left-sided temporal auditory comprehension area (Wernicke's area) from the left frontal motor speech area (Broca's area).1 He suggested that this could result from damage to the white mat¬ ter tracts that form a direct pathway between the major language zones.1 The localization of lesions of cases reported in the literature has often been in accord with this formula¬ tion.10·15 A lesion involving these tracts in and of itself does not prove, howev¬ er, that the white matter lesion is causative because a lesion of the adjoining gray matter might be responsible. The theory that arcuate fasciculus lesions are one cause of conduction aphasia as proposed by Wernicke,1 Lichtheim,2 Kleist,14 and more recently, Geschwind,"' Benson et al," and Damasio and Damasio,15 has thus been contested. None of the recent investigations, however, have proposed arcuate fasciculus damage as the only explanation for conduction Downloaded From: http://archneur.jamanetwork.com/ by a New York University User on 05/15/2015 auditory language comprehension area motor areas. Cases 1 and 2 of this report closely resemble Kleist's early description.19 Both subjects were left-handed and suffered acute vascular lesions that involved the posterotemporal cortex, including Wernicke's area plus ad¬ joining parietal cortex of the left hemisphere. Comprehension of spo¬ ken language was apparently per¬ formed through the right auditory association area, while motor speech was performed by the left frontal cortex. Repetition of spoken language thus demanded an interhemispheric connection that was damaged. In con¬ trast, the motor activities performed to verbal command (praxis) apparent¬ ly could be performed via a right temporal to right motor cortex and then anterior callosal connection. In other studies good correlations have been reported between the types of aphasia and localization of the lesion by CT scan.2021 The areas of pathologic involvement demonstrated by x-ray CT scans in these two cases and the pathways theorized for both repeti- Fig 4.—Diagram based on disconnection theory for cases 1 and 2, demonstrating interruption of language pathways from lesion in left temporoparietal area, including Wernicke's area. Language pathways are indicated by solid line. Praxis pathways are indicated by dotted line. M, indicates left motor area; M2, homologous area on right; B, Broca's area or left motor speech area; W,, location of Wernicke's classic area; W2, homologous area for language comprehension in right hemi¬ sphere. tion and ideomotor apraxia are out¬ lined in Fig 4. These two cases replicate neuropathologic observations that Wer¬ nicke's area may be destroyed without production of a severe comprehension deficit.4·11-22 Kleist14 promoted the explanation that conduction aphasia could result from a disconnection of a locus for comprehension located in the right temporal area from the more consistently located left frontal motor speech area. Interhemispheric trans¬ fer across the corpus callosum to the homologous structure in the left hemisphere would be interrupted by damage to the left hemisphere Wer¬ nicke's area. Such an explanation is clearly compatible with the discon¬ nection theory. Our first patient had a strong familial history of left-handedness, while the second had no family histo¬ ry of sinistrality. Patient 2 may have shifted hand dominance because of the arteriovenous malformations, with a partial shift of language com¬ prehension to the right hemisphere as well. Pathologic left-handedness of this nature is well recorded,23 and a hemispheric shift of language with left temporal and/or parietal develop¬ mental abnormality has also been demonstrated.24 The third case is a rare crossed aphasia, a language disturbance resulting from a right hemisphere lesion in a dextral. Both clinical and CT scan evidence indicated significant Fig 5.—Diagram for case 3 based on disconnection theory. Language pathways are interrupted by lesion in right temporoparietal area as they cross posteriorly to homologous area on left. Praxis pathways that travel anteriorly in right hemisphere are spared. Abbreviations are same as in Fig 4. right hemisphere temporoparietal damage. Comprehension appears to be subserved by the right temporal audi¬ tory area as substantiated by the ini¬ tial comprehension difficulty corre¬ lated with the presence of a limited degree of damage in the right posteri¬ or temporal cortex. The possibility that a small second lesion in the left hemisphere underlies the aphasia seems highly unlikely, given the tem¬ poral relationship of his right hemi¬ sphere stroke with the aphasia and the total absence of any left hemi¬ sphere abnormality on repeated neu¬ rological examinations and three seri¬ al CT scans. Figure 5 presents a theoretical explanation of case 3. Right temporo¬ parietal damage disconnects the right temporal area (comprehension) from the left temporal area and, hence, the pathway to the left frontal motor area. Right temporal to right frontal connections are again used to carry out motor activities in this patient who also lacked any evidence of ideo¬ motor apraxia. The model is identical to that of the first two cases, even though the causative abnormality is in the opposite hemisphere. A less likely schema would be inter¬ ruption of a pathway for motor speech running forward in a fully languagedominant right hemisphere. With this explanation the pathway for carrying out verbally commanded motor activi¬ ties must travel separately, either in the same hemisphere or through a Downloaded From: http://archneur.jamanetwork.com/ by a New York University User on 05/15/2015 posterior connection. A third possibil¬ ity would be an exact reversal of the Kleist14 diagram, with comprehension in the left hemisphere and motor speech in the right, an even less likely possibility given the initial compre¬ hension deficit. Other theories have been put for¬ ward to account for the clinical find¬ ings in conduction aphasia. One prom¬ inent rival model is "central aphasia." Goldstein9 proposed the existence of a central cortical language area to relate "inner speech" to external speech. With disruption of this pro¬ cess a language disturbance charac¬ terized by a disproportionate problem in correctly repeating spoken lan¬ guage can be theorized. Regardless of some strengths of the Goldstein mod¬ el (eg, the ability to account for more variability in the clinical presenta¬ tion), it requires a "central" area con¬ nected to a speech decoding site and a speech encoding site and is really only a modification of the Wernicke' model featuring a "language center" rather than a pathway. Location of this lan¬ guage center is difficult; conduction aphasia has been reported with lesions occupying a semicircle around the posterior rim of the left sylvian fissure, including abnormalities indi¬ vidually or in combinations in the temporal auditory complex, supramarginal gyrus, insula, the white matter deep to these regions, and pos¬ sibly even the adjacent posterior fron¬ tal cortex.2528 Of 25 cases with neuro- anatomie information reviewed by Green and Howes,10 13 had mixed lesions extending from temporal gyrus to supramarginal gyrus, nine had supramarginal gyrus involvement with sparing of the temporal area, and three involved the temporal audi¬ tory cortex without abnormality in the supramarginal gyrus. Our third case suggests that even right hemisphere lesions can be asso¬ ciated with this language disorder. While it is possible that Goldstein's central area could have such a vari¬ able location, a disconnection between the major centers appears more plau¬ sible. Green and Howes proposed a specific center to synthesize semantic meaning with linguistic form, a hypothesis poorly compatible with the number of reported neuropathologic sites. More recently, Levine and Calvanio8 proposed another hypothesis for con¬ duction aphasia. They postulated a gradient of overlapping auditory speech and motor speech cortices extending from Broca's area to Wer¬ nicke's area and suggested that abnormality anywhere along this course could produce a repetition defi¬ cit. They further suggested that the varying concentrations of neurons for speech perception and speech produc¬ tion determine the clinical features of the syndrome. Thus, the most posteri¬ or lesions result in a clinical picture approaching Wernicke's aphasia, while the more anterior lesions result in a more nonfluent conduction apha¬ sia. They also suggested that the dis¬ proportionate repetition deficit is actually a milder form of auditory perceptual deficit in comparison with Wernicke's aphasia. This model has obvious difficulty accounting for the occurrence of conduction aphasia with large temporoparietal lesions that encompass Wernicke's area such as in our first two patients and fails to account for the almost perfect com¬ prehension reported in some patients with conduction aphasia. Many pa¬ tients with conduction aphasia have been shown to have correct memory of the phrases they fail to repeat, again suggesting adequate to excellent audi¬ tory comprehension. The Levine-Calvanio8 hypothesis of cortical damage in the supramarginal gyrus does not differ markedly from the arcuate fasciculus theory. In fact, Geschwind16(p629) clearly points out that "since there is considerable evidence that the arcuate fasciculus consists in great part of short fibers ...," it is likely that purely cortical lesions of the supramarginal gyrus would destroy part of the pathway linking Wernicke's and Broca's areas. Various neurolinguists have inter¬ preted the language difficulties of conduction aphasia as either a speech encoding or a speech decoding disor¬ der. Dubois et al29 proposed that the essential problem involved the pro¬ duction of phonemes. Conduction aphasia became an expressive aphasia with a marked tendency to transpose and substitute phonemes resulting in prominent phonemic substitutions in all aspects of language. This theory would be compatible with the often reported ability of conduction apha¬ sies to monitor and self-correct. Repe¬ tition would be expected to show more paraphasias than spontaneous speech, because paraphasias have a wider selection of alternative words. Other investigators have defined the prob¬ lem as one of decoding. Stengel and Lodge-Patch30 regarded conduction aphasia as an incomplete form of receptive aphasia with a decreased ability to focus attention on words. Luria and Hutton31 hypothesized a failure to attend to the phonologic structure of words sufficiently to ana¬ lyze their structures. Warrington and Shallice32 suggested that selective impairment of auditory verbal shortterm memory produced the repetition disturbance, and Tzotzis and Albert33 proposed a specific impairment of memory for sequences. Strub and Gardner34 suggest that the repetition deficit actually occurs between these major processes, sever¬ al steps after phonologic processing but before phonemic encoding. This formulation appears neurolinguistically more satisfactory10 and is consis¬ tent with an interruption of transmis¬ sion of correctly decoded speech from the posterior comprehension area to the anterior speech encoding region. One assumption critical for use of the disconnection theory in our three atypical cases is the presence of audi¬ tory language comprehension ability in the right hemisphere. This is in accord with neuropathologic observa¬ tions of conduction aphasia with severe involvement of Wernicke's area, Heschl's gyrus, and surrounding structures in the left hemisphere.4·22 In the left-handed there is considerable evidence for a right hemisphere role in language, particularly in auditory processing.3536 Most patients whose cases were reported in the literature conforming to Kleist's14 formulation were either left-handed or had a fam¬ ily history of sinistrality. There is further evidence for a right hemisphere role in comprehension. Downloaded From: http://archneur.jamanetwork.com/ by a New York University User on 05/15/2015 Adult patients who have undergone left hemispherectomy show better preservation of comprehension than fluency,37 and a similar pattern has been reported in patients with large left hemisphere strokes.38 Kanshepolsky et al39 reported the case of an individual who developed conduction aphasia following a left hemisphere lesion and then pure-word deafness after a subsequent lesion in the right temporal area. Damasio and Dama¬ sio40 have shown right ear extinction with dichotic listening in conduction aphasia. Studies on patients who have undergone corpus callosum commisurotomy also demonstrate some degree of comprehension in the right hemi¬ sphere.41 Rasmussen and Milner's42 data, using amobarbital sodium injected unilaterally into one carotid artery to selectively anesthetize one hemisphere (Wada test), revealed that among 262 patients (right-hand¬ ed, left-handed, or mixed) without evi¬ dence of early damage to the left hemispheres, 16% had evidence of right hemisphere or bilateral lan¬ guage representation. In those cases with bilateral representation, lan¬ guage testing demonstrated a dissoci¬ ation of naming from repetition in half with one language function in one hemisphere and the other in the oppo¬ site hemisphere. Kinsbourne,43 per¬ forming Wada testing of a righthanded patient with conduction apha¬ sia, induced complete interruption of language with a right carotid injec¬ tion. It could be suggested that the atypical conduction aphasia of patient 3 represented a disconnection of interhemispheric pathways from a right temporal language reception area to the more routine left frontal motor speech area. At least six other cases of crossed conduction aphasia are recorded in the literature.4447 It is remarkable that, although most patients with crossed aphasia are nonfluent with initial mutism and agrammatism,48 the predominant type of fluent crossed aphasia has been the conduc¬ tion type.49 We are aware of only one neuropathologically proved case of crossed Wernicke aphasia.49 These findings are not likely to be simply due to the proportional occurrence of the aphasie syndromes in dextrals with total homologous language rep¬ resentation in the right hemisphere. There seems to be a specific tendency to develop crossed conduction aphasia in cases of fluent aphasia from righthemisphere lesions as predicted by disconnection theory. "Atypical" conduction aphasia, then, is a syndrome with fluent para¬ phasic verbal output, relatively good auditory comprehension, and signifi¬ cantly disturbed ability to repeat with variable sites of abnormality inconsis¬ tent with the arcuate fasciculus inter¬ ruption model. Atypical conduction aphasia apparently occurs when lan¬ guage comprehension is at least partially subserved by the right hemi¬ sphere. The variety of sites of abnor¬ mality for atypical conduction apha- sia, both as reported in the literature and as demonstrated by our three cases, the neurolinguistic data, and the evidence for right-hemisphere auditory language comprehension, are all harmonious with the original mechanism postulated by Wernicke, a separation of auditory language com¬ prehension and motor language out¬ put areas. Finally, the absence of an apraxia in our three cases, showing that the ability to carry out verbally commanded motor acts can be suc¬ cessful even though the patient fails to repeat spoken language, supports the existence of more than one route from auditory comprehension to ante¬ rior motor areas. Atypical conduction aphasia can be explained by the dis¬ connection concept, which remains the most tenable model for this lan¬ guage disturbance. References 1. Wernicke C: Der Aphasische Symptomen¬ komplex. Breslau, Poland, Cohn & Weigert, 1874. 2. Lichtheim L: On aphasia. Brain 1885;7:434484. 3. Pershing HT: A case of Wernicke's conduc¬ tion aphasia with autopsv. J Nerv Ment Dis 1900; 27:369-374. 4. Leipmann H, Pappenheim M: Über einem Fall von sogenannter Leitungsaphasie mit anato¬ Gesamte Neurol Psychiatr mischen Befund. 1914;27:1-41. 5. Pötzl 0: Ober die parietal bedingte Aphasie und ihren Einfluss auf das Sprechen mehrerer Sprachen. Gesamte Neurol Psychiatr 1925; 96: 100-124. 6. Luria A: Higher Cortical Functions in Man. New York, Basic Books, 1966. 7. Benson DF: Aphasia, Alexia, and Agraphia. New York, Churchill Livingstone Ine, 1979 8. Levine DN, Calvanio R: Conduction aphasia, in Kirshner HS, Freemon FR (eds): The Neurol¬ of Aphasia: Neurolinguistics. Amsterdam, Swets & Zeitlinger, 1982, vol 12, pp 79-112. 9. Goldstein K: Language and Language Dis¬ ogy turbances. New York, Grune & Stratton Ine, 1948. 10. Green E, Howes DH: The nature of conduc¬ tion aphasia: A study of anatomic and clinical features and of underlying mechanisms, in Whi¬ taker H, Whitaker HA (eds): Studies in Neuro¬ linguistics. New York, Academic Press Ine, 1978 vol 3, pp 123-156. 11. Benson DF, Sheremata WA, Bouchard R, et al: Conduction aphasia: A clinicopathological study. Arch Neurol 1973;28:339-346. 12. Goldstein K, Marmor J: A case of aphasia with special reference to the problems of repeti¬ tion and word-finding. J Neurol Neurosurg Psy¬ chiatry 1938;1:329-339. 13. Hoeft H: Klinisch-anatomischer Beitrag zur Kenntnis der Nachsprechaphasie (Leitungsaphasie). Dtsch Nervenheilk 1957;175:560-594. 14. Kleist K: Sensory Aphasia and Amusia. New York, Pergamon Press, 1962. 15. Damasio , Damasio AR: The anatomical basis of conduction aphasia. Brain 1980;103:337350. 16. Geschwind : Disconnexion syndromes in animals and man. Brain 1965;88:237-294, 585644. 17. Geschwind : The apraxias: Neural mecha¬ nisms of disorders of learned movement. Am Sci 1975;63:188-195. 18. Heilman KM, Coyle JM, Gonyea EF, et al: Apraxia and agraphia in a left-hander. Brain 1973;96:21-28. : Gehirnpathologie. Leipzig, 19. Kleist Germany, Barth, 1934. 20. Damasio , Damasio AR, Hamsher K, et al: CT scan correlates of aphasia and allied disorders. Neurology 1979;29:572. 21. Kertesz A, Harlock W, Coates R: Computer tomographic localization, lesion size and progno¬ sis in aphasia and nonverbal impairment. Brain Lang 1979;8:34-50. 22. Boiler F: Destruction of Wernicke's area without language disturbance: A fresh look at crossed aphasia. Neuropsychologia 1973;11:243246. 23. Satz : Pathological left-handedness: An explanatory model. Cortex 1972;8:121-127. 24. Rausch R, Crandall PH: Psychological sta¬ tus related to surgical control of temporal lobe seizures. Epilepsia 1982;23:191-202. 25. Hécaen , Dell MB, Roger A: L'aphasie de conduction. Encéphale 1955;2:170-195. 26. De Ajuriaguerra J, Hécaen H: Examen anatomique d'un cas d'aphasie de conduction. Rev Neurol 1956;94:434-435. 27. Coenen W: Klinischer und anatomischer Beitrage zur Frage der Leitungsaphasie. Arch Psychiatr Nervenkr 1941;112:664-678. 28. Sheremata WA: Conduction aphasia from a frontal lobe lesion. Trans Am Neurol Assoc 1974;99:249-252. 29. Dubois J, Hécaen H, Angelergues A, et al: Etude neurolinguistique de l'aphasie de conduc¬ tion. Neuropsychologia 1964;2:9-44. 30. Stengel E, Lodge-Patch IC: 'Central' aphasia associated with parietal symptoms. Brain 1955;78:401-416. 31. Luria AR, Hutton JT: A modern assess¬ ment of the basic forms of aphasia. Brain Lang 1977;4:129-151. 32. Warrington EK, Shallice T: The selective impairment of auditory verbal short-term mem¬ ory. Brain 1969;92:885-896. 33. Tzotzis C, Albert ML: Impairment of mem¬ ory for sequences in conduction aphasia. Neuro¬ psychologia 1974;12:355-366. 34. Strub RL, Gardner H: The repetition Downloaded From: http://archneur.jamanetwork.com/ by a New York University User on 05/15/2015 defect in conduction aphasia: Mnestic or linguis¬ tic? Brain Lang 1974;1:241-255. 35. Zangwill OL: Speech and the minor hemi¬ sphere. Acta Neurol Psychiatr 1967;67:10131020. 36. Angelergues R, Hécaen H, Djindjian R, et al: Un cas d'aphasie croisé. Rev Neurol 1962; 107:543-545. 37. Smith A: Speech and other functions after left (dominant) hemispherectomy. J Neurol Neu¬ rosurg Psychiatry 1966;29:467-471. 38. Kinsbourne M: The minor cerebral hemi¬ sphere as a source of aphasie speech. Arch Neu¬ rol 1971;25:302-306. 39. Kanshepolsky J, Kelly JJ, Waggener JD: A cortical auditory disorder: Clinical audiologic and pathologic aspects. Neurology 1973;23:699705. 40. Damasio H, Damasio AR: Dichotic listen¬ ing pattern in conduction aphasia. Brain Lang 1980;10:281-286. 41. Zaidel E: Unilateral auditory language comprehension on the token test following cere¬ bral commisurotomy and hemispherectomy. Neuropsychologia 1977;15:1-18. 42. Rasmussen T, Milner B: The role of early left-brain injury in determing lateralization of cerebral speech functions. Ann NY Acad Sci 1977;299:355-369. 43. Kinsbourne M: The mechanism of conduc¬ tion aphasia. Neurology 1971;21:429. 44. Brown JW, Wilson FR: Crossed aphasia in a dextral. Neurology 1973;23:907-911. 45. Hécaen H, Mazars G, Ramier AM, et al: Aphasie croisé chez un sujet droitier bilingue. Rev Neurol 1971;124:319-323. 46. Henderson VW, Oken B, Alexander MP: Fluent crossed aphasia with crossed Gerstmann syndrome in a right-handed man. Ann Neurol 1981;10:102. 47. Yarnell PR: Crossed dextral aphasia: A clinical radiological correlation. Brain Lang 1981;12:128-139. 48. Brown JW, Hécaen H: Lateralization and language representation: Observations on apha¬ sia in children, left-handers, and 'anomalous' dextrals. Neurology 1976;26:183-189. 49. Sweet EWS, Pañis W, Levine DN: Crossed Wernicke's aphasia. Neurology 1984;34:475-479.