J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.66.3.393 on 1 March 1999. Downloaded from J Neurol Neurosurg Psychiatry 1999;66:393–396 393

SHORT REPORT

Conduction elicited by stimulation of the left posterior

Mark Quigg, Nathan B Fountain

Abstract Case report Objective—Disruption of fascicular tracts A 35 year old right handed woman had intrac- that connect Wernicke’s to Broca’s areas is table complex partial seizures consisting of the classic mechanism of conduction parosmias, tachycardia, and confusion of a aphasia. Later work has emphasised corti- year’s duration. Biopsy of the left temporal pole cal mechanisms. at another institution had disclosed a left fron- Methods—To determine the distribution totemporal oligodendroglioma. of language on dominant cortex, electrical The patient presented to the comprehensive cortical stimulation was performed using epilepsy program 4 months later with contin- implanted subdural electrodes during ued seizures. Neurological examination was mapping before epilepsy surgery. normal with findings of intact comprehension, Results—A transient, isolated deficit in repetition, and naming. Brain MRI disclosed a repetition was elicited with stimulation of large, non-enhancing, low intensity lesion in the posterior portion of the dominant the left frontotemporal region abutting the superior temporal gyrus. insula and extending to the Conclusion—This finding suggests that (figure (A–C)). The ap- cortical dysfunction, not just peared compromised anteriorly. Continuous disruption, can induce . EEG with video monitoring showed seizure (J Neurol Neurosurg Psychiatry 1999;66:393–396) onset in the left temporal region. was fluent on the preoperative Keywords: language; brain mapping; aphasia neuropsychological battery. On the Reitan- Indiana aphasia screening test the patient transposed syllables when repeating the word Conduction aphasia is thought to result from “episcopal” forming “ecispolal.” lesions of the arcuate fasciculus that disconnect For seizure localisation and cortical map- http://jnnp.bmj.com/ receptive from expressive language regions. It ping, intracranial monitoring with subdural consists of impaired repetition with preserved grids and occipitally inserted hippocampal comprehension, naming, and reading. Literal depth electrodes was performed before left are frequent, and ideomotor temporal lobectomy and partial subfrontal apraxias can also be present. tumour resection. Subsequent histopathology The lesions that most often produce conduc- confirmed the presence of oligodendroglioma. tion aphasia involve the white matter underly- Four months postoperatively the patient was 1–3

tested with the Boston aphasia severity rating on September 27, 2021 by guest. Protected copyright. Department of ing the dominant supramarginal gyrus. The Neurology, disconnection between expressive and recep- scale. Impaired repetition with mixed parapha- Comprehensive tive cortical regions was hypothesised by sic errors became evident only when tested Epilepsy Program, Wernicke and led to the first description of with complex, low probability vocabulary. University of Virginia, conduction aphasia. However, it has been Charlottesville, VA, reported in lesions that spare white matter, and USA M Quigg is unexpectedly absent in some cases involving Methods 4–7 N B Fountain the classic lesion. Some of these diYculties in Informed consent was obtained before proce- localisation have arisen because naturally dures were carried out. Subdural grids and Correspondence to: occurring lesions may indiscriminately involve Dr Mark Quigg, Department strips (intercontact distance=1 cm) were of Neurology, both white and grey matter. These inconsisten- placed on the left and temporal Comprehensive Epilepsy cies, in part, support an alternative hypothesis lobe as shown on the figure (D and E). We Program, Box 394, Health that conduction aphasia may be mediated by a stimulated adjacent pairs of contacts using a 50 Sciences Center, University 8 of Virginia, Charlottesville, specific region of cortex. Hz square wave signal starting at a duration of VA 22908, USA. Telephone We report a case of transient conduction 2 seconds and an intensity of 4.5 mA. We 001 804 924 5312; fax aphasia elicited by electrical cortical stimula- 001 804 982 1726; increased duration to 5 seconds and increased email [email protected] tion of the posterior superior temporal gyrus. current by 1 mA increments until (1) a The selective and reversible impairment of a response occurred, (2) an afterdischarge or Received 17 July 1998 specific region of cortex shows that conduction clinical seizure occurred, or (3) no response and in revised form 14 September 1998 aphasia may be induced by means other than was achieved at a stimulus of 12.5 mA/5 Accepted 16 September 1998 disconnection. seconds. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.66.3.393 on 1 March 1999. Downloaded from 394 Quigg, Fountain http://jnnp.bmj.com/

(A) Axial, (B) sagittal, and (C) coronal views of a T1 weighted, uncontrasted MRI taken before placement of subdural on September 27, 2021 by guest. Protected copyright. electrodes demonstrates the large, frontotemporal cystic mass of the dominant hemisphere. The arcuate fasciculus is intact posteriorly but is diYcult to identify anteriorly near Broca’s region. (D) Coronal view after electrode placement confirming location of electrode STS 4 upon the superior temporal gyrus (white arrow). (E) Three dimensional MRI reconstruction showing the array of subdural electrodes. Stimulation at pairs LG4, 5, and 10 caused either global or non-fluent aphasia, and stimulation at STS3,4 caused conduction aphasia. STS and ITS=superior and inferior temporal strip; UG and LG=upper and lower grid; SF=sylvian fissure.

Results charges were elicited. Stimulations at other Stimulation in adjacent contact pairs (lower sites had no clinical manifestations. grid electrodes 4, 5, and 10 in the figure (E)) induced anomia and speech arrest in Broca’s Discussion region. Stimulations of 5 seconds duration and Our finding of an isolated deficit of repetition 6 mA intensity at electrodes 3 and 4 of the elicited by electrical cortical stimulation is rel- superior temporal strip (figure (E)) induced a evant because it runs counter to the prevailing similar to conduction apha- view that disconnection of a white matter tract sia: (1) recitation was normal; (2) naming is necessary for conduction aphasia. The “thumb,” “index finger,” and “tie” was normal; precise localisation oVered by cortical stimula- (3) command following—“show me your tion suggests that a focal region of the posterior thumb”—was normal; and (4) repetition— superior temporal gyrus of the language domi- “Bill Clinton”, “Commonwealth of nant hemisphere can mediate language dys- Virginia”—was reproducibly impaired. No function most consistent with conduction paraphasic errors were found. No afterdis- aphasia. Because practical considerations limit J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.66.3.393 on 1 March 1999. Downloaded from Conduction aphasia by cortical stimulation 395

complex testing during intracranial electrical addition to its connecting white matter tracts is stimulation, we were only able to test the basic necessary to mediate repetition. Two findings abilities of naming, recitation, repetition, and support this conclusion. Firstly, the large fron- command following. We were not able to test totemporal tumour compromised the arcuate other findings that define the full syndrome of fasciculus but spared the overlying cortex (fig- conduction aphasia.2 ure). Despite the white matter lesion, the The arcuate fasiculus is a white matter tract patient had no preoperative evidence of a that runs from Wernicke’s area in the posterior significant conduction aphasia (although sub- superior temporal gyrus, arches around the tle findings were disclosed after partial resec- sylvian fissure, and runs anteriorly from the tion). Secondly, electrical stimulation of the inferior parietal lobe to the inferior frontal lobe superior temporal gyrus elicited transient con- of Broca’s region.3 Wernicke postulated that a duction aphasia, specifically demonstrating lesion of the arcuate fasciculus that discon- that cortical function is necessary for normal nected receptive from expressive centres would repetition. produce a deficit in repetition, or “conduction The large frontotemporal tumour could aphasia”. Others have proposed that a single confound our interpretation. It is theoretically cortical centre was responsible for integration possible that eloquent cortex was abnormally of receptive and expressive regions yet was represented as a consequence of a chronic mass independent of them.8 lesion. By this “remapping” process, stimula- Evidence from subjects with conduction tion of a makeshift language area may have aphasia usually supports the concept of discon- unexpectedly resulted in conduction aphasia, nection. In these studies (usually patients with even though this region might not normally ), disruption of the arcuate fasciculus is mediate this function in other subjects. This obligate with variable involvement of adjacent process seems unlikely as the tumour displaced regions of suprasylvian or subsylvian cortex.13 no other language functions, such as expressive Circumscribed lesions of the arcuate fasciculus language in Broca’s region. that spared overlying cortex have been oVered Another consideration is the possibility that as evidence of disconnection,9–11 but white electrical stimulation was not confined to the matter lesions are not able to diVerentiate cortex, and non-specific electrical disruption of between the relative importance of disruption underlying arcuate fasciculus may account for of the arcuate fasciculus versus disconnections our findings. Basic electrical properties and of overlying neurons along its course. previous studies suggest that neuronal activa- Physiological findings have also supported tion caused by electrical stimulation remains disconnection as the mechanism of conduction highly confined to the immediate vicinity of the aphasia. Regional blood flow determined by stimulating electrodes, especially at the low xenon CT was absent in Broca’s region in current required.14 15 Furthermore, there was patients with conduction aphasia, sug- no measurable afterdischarge that could spread gesting functional disconnection.12 However, to surrounding regions and confound results. studies of cortical strokes, in determinations of Electrical stimulation, unlike clinicopathologi- cortical versus subcortical mechanisms, can be cal correlations in stroke, more selectively misleading because regions of destruction separates cortical from white matter dysfunc-

involve both cortex and the arcuate fasciculus. tion. http://jnnp.bmj.com/ Finally, electrical stimulation of eloquent In our patient, the posterior superior tempo- cortex produced both Broca’s and Wernicke’s ral gyrus was spared during partial resection of but not conduction aphasia,13 suggest- the tumour, in part because of our mapping of ing that conduction aphasia is not cortically conduction aphasia, but mainly because the mediated. Notably, in this series of patients tumour involved more anterior portions of the with implanted subdural electrodes, the testing . Our findings suggest that paradigm involved mainly reading aloud, and stimulation paradigms should test for conduc-

repetition may not have been adequately tion aphasia if the area of potential resection on September 27, 2021 by guest. Protected copyright. tested.13 Thus this finding may have been involves the posterior, perisylvian cortex. Risk missed in other patients and may not be unique and benefits of a resection amidst eloquent to our patient. cortex must be calculated on an individual Other studies suggest that disconnection basis. may not be the only mechanism of conduction In summary, these findings show that aphasia. Some cases of conduction aphasia dysfunction of the posterior superior temporal were caused by lesions that clearly spared the cortex can induce a syndrome similar to arcuate fasciculus.45 Similarly, lesions of the conduction aphasia. This study suggests that arcuate fasciculus have not always resulted in conduction aphasia is not simply a disconnec- conduction aphasia.6 Furthermore, physiologi- tion syndrome but can result from stimulation cal data provided by PET imaging does not of the posterior perisylvian cortex. clearly support the disconnection theory. In one study of stroke and conduction aphasia, We thank Dr H Robert Brashear for administration of the Bos- cerebral metabolic patterns had no clear corre- ton aphasia battery and for his helpful suggestions. lation to clinical findings,7 suggesting that functional disconnection is not necessary to 1 Benson D, Sheremata W, Bouchard R, et al. Conduction produce conduction aphasia. aphasia. Arch Neurol 1973;28:339–46. 2 Benson D. Aphasia. In: Heilman K, Valenstein E, eds. Clini- Our findings support the view that conduc- cal neurophysiology. New York: Oxford University Press, 1993:18–32. tion aphasia is a cortically based phenomenon. 3 Damasio H, Damasio A. The anatomic basis of conduction At the least, this case shows that cortex in aphasia. Brain 1980;103:337–50. J Neurol Neurosurg Psychiatry: first published as 10.1136/jnnp.66.3.393 on 1 March 1999. Downloaded from 396 Quigg, Fountain

4 Marshall R, Lazar R, Mohr J, et al. “Semantic” conduction 11 Arnett P, Rao SMH, Swanson S, et al. Conduction aphasia aphasia from a posterior insular cortex infarction. J in multiple sclerosis: a case report with MRI findings. Neu- Neuroimaging 1996;6:189–91. rology 1996;47:576–8. 5 Mendez M, Benson D. Atypical conduction aphasia: a 12 Demeurisse G, Capon A. Brain activation during a linguis- . Arch Neurol 1985;42:886–91. tic task in conduction aphasia. Cortex 1991;27:285–94. 6 Shuren J, ScheVtB,YehH,et al. Repetition and the arcuate 13 SchaZer L, Luders H, Beck G. Quantitative comparison of fasciculus. J Neurology 1995;242:596–8. language deficits produced by extraoperative electrical 7 Kempler D, Metter E, Jackson C, et al. Disconnection and stimulation of Broca’s, Wernicke’s, and basal temporal lan- cerebral metabolism. The case of conduction aphasia. Arch guage areas. Epilepsia 1996;37:463–75. Neurol 1988;45:275–9. 14 Ranck JJ. Which elements are excited in electrical 8 Goldstein K. Language and language disturbances. New York: stimulation of mammalian central nervous system: a Grune and Stratton, 1948. review. Brain Res 1975;98:417–40. 9 Tanabe H, Sawada T, Inoue N, et al. Conduction aphasia 15 Nathan S, Lesser R, Gordon B, et al. Electrical stimulation and arcuate fasciculus. Acta Neurol Scand 1987;76:422–7. of the human , theoretical approaches. In: 10 Aihara M, Oba H, Ohtomo K, et al. MRI of white matter Devinsky O, Beric A, Dogali M, eds. Advances in neurology. changes in the Sjogren-Larsson syndrome. Neuroradiology Vol 63. Electrical and magnetic stimulation of the brain and 1995;37:576–7. spinal cord. New York: Raven Press, 1993:61–85. http://jnnp.bmj.com/ on September 27, 2021 by guest. Protected copyright.