BRAIN AND LANGUAGE 19, 204-215 (1983) Hemiplegic Writing in Severe Aphasia JASON W. BROWN New York University Medicul Center BARBARA J. LEADER New York Veterans Administration Hospital AND CYNTHIA S. BLUM New York University Medical Center Three patients with severe aphasia and right hemiplegia are described who could write to dictation with the right arm using a limb prosthesis though agraphic with the “intact” left hand. The phenomenon of “hemiplegic writing” is explained as an access to submerged or preprocessing levels in language and action structure, through the use of older proximal motor systems. This interpretation has implications for our understanding of language and brain function, as well as for approaches to the treatment of patients with severe language disorders Even in mild aphasics left-sided agraphia is not unusual. What is unusual is the occasional nonhemiplegic aphasic with left-sided agraphia and a moderate preservation of writing ability on the right side. Cases of this type are generally interpreted as disconnection syndromes, due to a lesion of the corpus callosum (Geschwind & Kaplan, 1962). The demonstration of a true left-sided agraphia depends on the ability of the patient to write with the right hand. Otherwise, this impairment could be explained as a reflection of the aphasia and not a result of a disconnection between the left language areas and the right hemisphere. When right-hand writing can be tested, most aphasics do not show a difference between the hands (Brown et al., 1983), so that left-sided Supported by grants from the Easter Seal Foundation and the World Rehabilitation Fund. Gratitude to Dr. Anton Leischner for assistance in this study. Send requests for reprints to Jason W. Brown, Department of Neurology, 530 First Avenue, New York University Medical Center, New York, NY 10016. 0093-934X/83 Copyright All rights $3.00 0 1983 by Academic Press, Inc. of reproduction in any form reserved. 204 HEMIPLEGIC WRITING 205 agraphia on a presumed disconnection basis (i.e., with good right-hand writing) would be an unusual occurrence. Since the majority of aphasics have some degree of right hemiparesis, it is ordinarily not possible to compare their left- and right-hand graphic performances. Some years ago, Leischner (1970, 1983) demonstrated hemiplegic writing in global aphasia. Using a plastic ball and penholder in the rigid right hand, the patient made thrusting movements in the forward or backward direction, with the hand repositioned by the examiner for the next stroke. The finding that some global aphasics could write words to dictation with the hemiplegic limb, though agraphic with the left hand, was interpreted as a disconnection of the right hemisphere/left hand combination from left-sided language areas. As part of a larger study (Brown et al., 1983) of manual asymmetries in graphic and motor performance in aphasia, we decided to look more closely at this phenomenon. For this study, we designed a skateboard apparatus which could be moved easily in all directions (Fig. I). The arm is strapped onto the device, and the arm and prosthesis can be moved en bloc by the residual proximal muscles without assistance on the part of the examiner. Leischner’s pen holder is placed in the patient’s hand. METHODS The subjects were administered a battery of subtests to assess their ability to copy letters, forms, and words, to write words and letters to dictation, and to assess limb praxis to command and imitation, and finger postures to imitation. with each hand tested separately. Four separate test protocols were devised in which the order of presentation of left hand (LH) and right hand (RH) tasks was counterbalanced within and across protocols. Stimuli FIG. 1. The penholder and skateboard prosthesis used in this study. 206 BROWN, LEADER, AND BLUM specific to each subtest were randomly assigned to the LH and RH conditions. with no stimulus appearing more than once in any condition. The letter-copying subtests incorporated all letters of the alphabet. Subjects were required to copy IO letters with the LH and IO other letters with the RH. Stimuli for the formcopying subtest were derived from the letters of the alphabet so that each form had the same number and type of component parts as its correlate letter. Subjects were asked to copy IO forms with the RH and IO with the LH. For the word-copying subtest highfrequency (20 per million or higher (Paivio, 1968)), picturable, and concrete (score of 47 on a 7-point scale of concreteness, Paivio, 1968) words were used. Words of 4, 5, and 6 letters in length were approximately equally represented. Different words were assigned to each hand. There were also two dictation tasks. Subjects were asked to write to dictation 10 letters and 20 words with the LH, and 10 letters and 20 words with the RH. Half the words were low frequency (less than IO/million) and half were abstract (3 or less on a 7point scale). In addition to these tests, the larger population study employed tests of praxis, object naming, spelling, letter and word repetition, and word comprehension. In the three subjects described in this paper, additional “disconnection” tasks were introduced. In one test, 10 small three dimensional objects were placed in the subject’s left hand, which was screened from view. The subject was required to write the name of the object held in the left hand using the right hand and prosthesis. Items consisted of: pencil, cup, razor, ball, scissors, key, matches, toothbrush, fork, and comb. Additionally, IO pictures from the Boston Naming Test were presented to the subject, who was required to write the name of the object depicted using the right hand and prosthesis. Pictures included: tree. pencil, house, comb, flower, bed, toothbrush, scissors, saw, and broom. In a second trial, subjects were asked to write the names to the same pictures and objects placed in the left hand using the refr hand. In a third trial, subjects were asked to verbally name the objects and pictures. Left- and right-hand trials with objects and pictures were alternated, and the order of items within the trials was random. The verbal naming was placed last so as not to influence the written productions. Case Reports Case I (CA). This 23-year-old right-handed male student underwent surgery for a left trigeminal sarcoma in December 1981. In the immediate postoperative period he developed sudden aphasia and right hemiplegia, and again underwent surgery, this time to remove a left frontal intracerebral hematoma. The CT shows an extensive left-hemisphere lesion (Fig. 2). A left anterior temporal lobectomy was also carried out. At the time of evaluation 3 months later, he still showed a severe global aphasia which had changed little over the intervening months. There was a moderate spastic right hemiplegia and a left facial palsy. Of interest is the fact that CA was a “hooker” in writing with the right hand, and had a left-handed brother. Speech was nonfluent, agrammatic, and well-articulated, with single nouns and occasional cliche expressions. Often nouns appeared semantically related to targets. Naming was poor with occasional semantic paraphasias. Comprehension was severely impaired, as was oral spelling. He could repeat most single words presented to him. There was moderate apraxia of the left arm. Scores on movements to oral command were l/l0 correct; to imitation, 7/10 correct: and reproducing finger postures, S/IO correct. CA took to the prosthesis from the beginning, and was able to write smoothly and effectively with it. Performance on copying letters, shapes, and words was similar in both hands (Fig. 3), indicating that failures on word dictation on the left side were not the result of mechanical difficulties. When he was asked to write letters to dictation, there was letter substitution with the left hand, but consistently accurate writing with the right hand and prosthesis. This difference was quite pronounced on writing both high- and low-frequency 207 HEMIPLEGIC WRITING FIG. 2. CT Scan in Case I (CA). operative subgaleal fluid collection. There is a large left-hemisphere lesion, FORK MONKEY FINGER WHIP RAZOR BELL LEFT with a post- RIGHT FIG. 3. Word copy in Case 1 (CA). The LH is accurate on copy, indicating that the difficulty on dictation is not due to purely motor factors. The orthography with the RH + prosthesis is excellent. 208 BROWN, LEADER, AND BLUM words to dictation. In Fig. 4, the marked paragraphia with the intact left hand stands in sharp contrast to the accurate spelling of words with the hemiplegic right arm. CA was unable to write with the left hand the names of objects palpated by that hand. When an object was held in the left hand, eyes closed, he could often (40%) achieve the correct written name using the prosthesis in the right hand. His ability to name objects in the left hand was comparable to his visual naming ability (about 40% correct), and if he could name the object he could as often write the object name with the right arm but never with the left. CA made several attempts, all unsuccessful, to write with the prosthesis on the left arm, but could not be induced to use it for careful testing. Case 2 (TV). A St-year-old man with a left CVA, right hemiplegia, and global aphasia, onset in December 1981. He was initially admitted to another hospital with a “thyroid storm”, atria1 fibrillation, and pulmonary edema. A CT scan on December 24, 1981 was normal. The next day, he had an acute embolic infarct with global aphasia and right hemiplegia. Repeat CT in August 1982 showed a large left central lesion (Fig. 5). He was subsequently transferred to the VA Hospital, where further testing was carried out. The PICA was in the 12th percentile of aphasics, the RCPM score was 24/36. Speech was fluent with stereotypic jargon, dense comprehension impairment, poor naming with semantic paraphasia, and poor single word repetition. There was apraxia of the left arm. Scores to oral command were 4/10 correct; to imitation, 4/10 correct; and to finger postures, 7/10 correct. With the prosthesis, performance on copying words and shapes was comparable with both left and right limbs, with better orthography on the left side (Fig. 6). To dictation, there was marked letter substitution with the LH, resulting in a complete agraphia on that side, whereas writing both low- and high-frequency words was remarkably well preserved with the hemiplegic right arm and prosthesis (Fig. 7). Moreover, LH writing to dictation was unimproved with the prosthesis on the /eft arm. TV was unable to name objects placed in the left hand or presented visually, but with the RH could generally write the name of an object before him or one held in the LH. Thus, on one occasion, he wrote (RH) the CRISIS c 0 P LEFT RIGHT FIG. 4. Writing to dictation in Case I (CA). The target words are shown for the LH, the RH performing accurately on all 10 low- and high-frequency words. FIG. 5. CT Scan in Case HEMIPLEGIC WRITING 2 (TV), a large showing 209 infarct in the left hemisphere. PENCIL BELL Left FIG. argues Hand 6. Word copy in Case 2 (TV). against an agraphia on a purely Right Hand The ability to copy motor basis. words correctly with the LH 210 BROWN, LEADER, CRUTCH CN FORK WOE DRILL 'r\ Left ' AND BLUM TABLE c CARROT HORSE "?py Hand Right Hand FIG. 7. Writing to dictation in Case 2 (TV). With the LH, there is correct orthography but marked agraphia, whereas the RH + prosthesis produces distorted but correctly spelled words. The targets are shown at the side of each column. names of 9 of 10 objects held in the LH, and 10 before him, but could not name a single object, nor write its name, with the LH, whether the object was seen or held in the LH. Case 3 (RC). A 53-year-old man tested 1 year postonset left CVA, with dense right hemiplegia and global aphasia. The CT scan demonstrated a large left central lesion (Fig. 8). Speech consisted of well-articulated but nonfluent stereotypic utterances, such as “Well . . oh yes .” and swearing. Repetition of single words was about 80% correct. Naming was nil, comprehension severely impaired. RC was able to copy words correctly with the RH and prosthesis, but not with the LH (Fig. 9), and in this respect differed from the other two patients. He performed well to dictation with the right hand and prosthesis though spelling errors were present, whereas LH writing to dictation was markedly agraphic (Fig. 10). This was also true for the LH and prosthesis. Thus, the LH score for letters to dictation was 3/10 correct (words O/IO) without the prosthesis, and 2/10 correct (? l/10 words) with it. There was marked apraxia of the LH, scoring to oral command, l/IO correct; to imitation, l/IO; and reproducing finger postures, 6/10 correct. RC was tested for “callosal” transfer, by placing a series of IO objects in the left hand and requiring him to write the names of the objects with the right hand and prosthesis. He wrote four names correctly, three others were almost correct (razsor for razor; sissors for scissors; brush for toothbrush), and three others with various types of spelling errors (piny for pencil; ker for key; czmb for comb). In effect, the target item was clear in all 10 written names. His performance was about the same on writing the name of a picture before him with the right arm and prosthesis. He was able to name 4 of 10 objects in the left hand, and 6 of IO object pictures. In contrast, he was unable to write the name of a single object held in the left hand, using the left hand. DISCUSSION Hemiplegic writing, first described by Leischner (1970), is documented in this study. The phenomenon appears to occur in severe, generally global, aphasia rather than in partial cases with hemiplegia. In fact, hemiplegic writing in globals is often superior to right-hand writing in milder aphasics without hemiplegia (Brown et al., 1983). Moreover, it does not seem to be a rare occurrence. The majority of global aphasics with sufficient limb rigidity to grip the pen holder, and sufficient proximal motility to drive the prosthesis, may well show a manual asymmetry favoring the hemiplegic limb. HEMIPLEGIC FIG. 8. CT Scan in Case 3 CRC). 211 WRITING There is a large left-hemisphere infarct. ytw= MONKEY SNAKE ,a CROWN Left FIG. 9. to copying Hand Right Word copy in Case 3 CRC). In this patient words, though form copy was adequately Hand there is marked performed. agraphia on the left 212 BROWN, LEADER, Left Hand AND BLUM Right Hand 10. Writing to dictation in Case 3 (RC). The target words shown to the right are the same for the LH and the RH. FIG. The explanation of this phenomenon has ordinarily focused on the defective writing with the left hand rather than the superior writing with the right. Accordingly, the left-sided agraphia is presumed to result from destruction of callosal fibers leading from left language centers to right hemisphere. Yet there are many problems with this interpretation. First, there was no CT evidence of callosal pathology in these stroke cases. The argument that any hemispheric lesion disrupts callosal efferents at their origin does not explain the association with severe aphasia, nor does it explain the occurrence of hemiplegic writing. After all, the disconnection of regions controlling left-hand writing in right hemisphere should be accompanied by a destruction of homologous regions in left hemisphere, and give an agraphia in both hands. In other words, if callosal fibers are disrupted at their origin, this should involve left hemisphere regions controlling writing with the right hand as well. Second, the results of a separate population study on manual asymmetries in aphasia (Brown et al., 1983), indicate that nonhemiplegic aphasics overall are slightly more apraxic on the right side, the opposite of that predicted by a disconnection. In fact, while all three patients showed left-sided apraxia, there was no relationship between the severity of the apraxia and the agraphia. Moreover, the interpretation of left-sided (“sympathetic”) apraxia in stroke cases assumes a right-sided apraxia buried in the hemiplegia. If left-sided apraxia is taken to explain the agraphia, how then account for the correct writing on the hemiplegic side? Even in cases of callosal agraphia there are difficulties with the disconnection account. Cases of left-sided agraphia with callosal lesions have been described without left-sided apraxia (Sugishita et al., 1980; Gersh & Damasio, 1981). The reverse also occurs, left-sided apraxia without agraphia (Klein & Ingram, 1958). In patients with callosal section, left-sided apraxia is often transient while agraphia persists. This lack of HEMIPLEGIC WRITING 213 correspondence between left-sided agraphia and apraxia is a problem for a “theory” which posits an interruption between a command to the left hemisphere language areas (e.g., “write: chair”; “make a fist”) and a right hemisphere control mechanism implementing the target movements in the left hand. Third, and most importantly, all three cases showed transfer from left to right hand on a tactile and written naming task. Subjects were asked to hold an object in the left hand with the eyes closed, and write its name with the hemiplegic right arm and prosthesis. Their performance on this task was comparable to, or better than, their written naming with the right arm of objects presented visually. In one patient, the names of 9 of 10 objects held in the left hand could be written with the hemiplegic limb, though none of the object names could be written with the left hand. This evidence of cross-manual transfer directly controverts the disconnection interpretation of this phenomenon. In this respect, it is of interest that patients with partial commissurotomy (callosal trunk) do not show a permanent left tactile anomia, whereas those with complete commissural section name objects in the left hand about 20% as often as in the right (Bogen et al., 1981). Right-hand naming in such cases should be comparable to visual confrontation naming. In our cases, tactile naming in the left hand was about the same as visual confrontation naming. The failure to find a left tactile anomia greater than the visual naming impairment also argues against a disconnection. If the disconnection account of this phenomenon is inapplicable, what other interpretation can be offered? One possibility is that the prosthesis enables the patient to use the proximal limb musculature to access submerged levels in language in the damaged (or intact?) hemisphere. There is evidence for the existence of an older motor system subserving axial and proximal movement, and a more recent one for fine distal movements (Kuypers, 1973). These systems in motility can be viewed as levels in the structure of an action which are linked to corresponding levels in language and cognition (Brown, 1977). On this view, the association between hemiplegic writing and severe or global aphasia occurs because the phenomenon requires the destruction of a “surface” level in left-hemisphere language, a level which is coextensive with, and drives or controls, a motor organization for the distal innervation. This gives an agraphia for distal writing on the left side as well as on the right. However, in hemiplegics this distal agraphia is obscured by the loss of the distal innervation on the right side. The resultant loss of surface processing in both action and language representation permits access to submerged preprocessing stages in both of these systems (Brown, 1983). In two patients tested (TV, RC) left-limb writing was persistently agraphic using the prosthesis on the left side. The third case (CA) made unsuccessful 214 BROWN, LEADER, AND BLUM attempts to use the prosthesis on the LH, and would not continue. The failure to find improvement in writing with the LH and prosthesis is not surprising, since the occurrence of proximal muscle writing seems to require disruption of the distal limb innervation. This disruption, it is argued, allows the submerged proximal system to become the endpoint of a truncated action development. When distal movements are intact, the proximal system cannot serve as a microgenetic endpoint, but is derived to the stage of fine distal motility. This is why the agraphic writing with the LH reflects a damaged surface level in language processing in left hemisphere, regardless of whether or not the prosthesis is used. Apart from the theoretical basis of hemiplegic writing, this study has demonstrated that severe aphasics, otherwise unable to communicate in speech or with the left hand, may be able to write with the hemiplegic limb with the aid of a prosthesis. This demonstration opens a new approach to treatment in aphasia. Not only do severe aphasics retain the ability to write with the hemiplegic limb, but the possibility that preliminary, subconscious, or preprocessing stages in language can be tapped as a means of facilitating communication in aphasia provides an impetus to further studies of restitution from this point of view. Now added in proof. A prototype of this prosthesis has been constructed in molded plastic, and is available to the public and to those working with aphasic patients at a modest cost. Information regarding the prosthesis can be obtained by writing to the Institute for Research in Behavioral Neuroscience, a nonprofit Medical Foundation, located at 530 First Avenue, New York, New York 10016. REFERENCES Bogen, J., Campbell, A., & Thompson, A. 1981. Long-term persistence of unilateral anomia following complete cerebral commissurotomy. Proceedings of the Society of Neuroscience, I, 945. Brown, J. W. 1983. Emergence and time in microgenetic theory. Journul Academy of Psychoanalysis, J. W. 1977. Mind, brain, of the American 11, 35-54. and consciousness. Brown, New York: Academic Press. Brown, J. W., Leader, B., Blum, C., & Piasetsky, G. 1983. Manual asymmetries in praxis and drawing in aphasia, in preparation. Gersh, F., & Damasio, A. 1981. Praxis and writing of the left hand may be served by different callosal pathways. Archives of Neurology, 38, 634-636. Geschwind, N., & Kaplan, E. 1962. A human cerebral deconnection syndrome. Neurology, 12, 675-685. Klein, R., & Ingram, I. 1958. Functional disorganization of the left limb in a tumour of the corpus callosum infiltrating the hemispheres. Journal of Mental Science, 104, 732742. Kuypers, H. 1973. The anatomical organization of the descending pathways and their contributions to motor control, especially in primates. In J. Desmedt (Ed.), New developments in electromyography and clinical neurophysiology. Karger: Basel. Vol. 3. Pp. 38-68. Leischner, A. 1970. Seitendifferente Schreibleistungen bei Agraphien. In Die Rehabilitation der Aphasie in den romanischen Landern. Stuttgart: Thieme Verlag. Pp. 70-83. HEMIPLEGIC WRITING 215 Leischner. A. 1983. Side differences in writing to dictation of aphasics with agraphia: A graphic disconnection. Brain and Languuge. 18, I-19. Paivio, A., Yuille, J., & Madigan, S. 1968. Concreteness, imagery and meaningfulness values for 925 nouns. Journal of Experimental Psychology, 76 (Suppl.), l-25, Sugishita, M., Toyokura, Y., Yoshioka, M., & Yamada, R. 1980. Unilateral agraphia after section of the posterior half of the truncus of the corpus callosum. Bruin and Languagr, 9, 215-225.