Journal ofNeurology, Neurosurgery, and Psychiatry 1994;57:1241-1244

Conduction apraxia

Cynthia Ochipa, Leslie J Gonzalez Rothi, Kenneth M Heilman

Department ofVeterans AffairsMedical Center,Tampa, Florida, USAC OchipaDepartment ofVeterans AffairsMedical Center,Gainesville, Florida,USAL J G RothiKM HeilmanCenter forNeuropsychologicalStudies andDepartment ofNeurology, Universityof Florida College ofMedicine, Florida,USAL G RothiK M HeilmanCorrespondence to:Dr Cynthia Ochipa,Audiology and SpeechPathology Service (126),James A Haley VeteransHospital, 13000 N 30th St,Tampa, Florida 33612,USA.

Received 12 October 1993and in revised form10 March 1994.Accepted 25 March 1994

Figure 1 Liepmann'sconceptualisation ofpraxisprocessing. Lesion A isproposed to account forpantomime agnosia.4 *Ithas been posited that themovementformulas arecoded in a threedimensional supramodalcode. These representationsmust be transcoded into aninnervatory pattern that inturn is fed to the motorcortex. The supplementarymotor area is thought toplay a part in thetranscoding of thespace-time representationsinto the innervatorypatterns. 10 14

AbstractA left hemisphere damaged patient withideomotor apraxia is described, whoseperformance on pantomime to verbalcommand was superior to pantomimeimitation. His reception of these samegestures (gesture naming) was spared.This syndrome has been named conduc-tion apraxia. To account for this selectiveimpaired performance on gesture imita-tion, a separation of the representationsfor gesture production and reception isproposed and a non-lexical gestureprocessing route for gesture imitation issuggested.

(7 Neurol Neurosurg Psychiatry 1994;57: 1241-1244)

Liepmann and Maas' proposed that the lefthemisphere contained "movement formulas"for the programming of skilled movements.Heilman, et a!2 proposed that these movementformulas or space-time movement representa-tions were stored in the left inferior parietallobule and lesions that ablate the parietal lobeshould be behaviourally dissociable frommore anterior lesions that leave these engramsintact but disconnect them from motor areas.Studies of gesture comprehension and dis-

Visual/gestural input

Right handgestural

production

Left handgestural

production

crimination have shown that apraxic patientswith parietal lobe lesions (supramarginal orangular gyrus) not only gesture and pan-tomime poorly to command and to imitationbut also have gesture comprehension deficits.Apraxic patients with anterior lesions (ante-rior to or deep to the supramarginal gyrus) areimpaired in producing skilled movement butare not impaired in comprehending or dis-criminating gestures.23 Failure to comprehendgestures in the absence of apraxia (pantomimeagnosia) has also been described. This disor-der was associated with lesions of the visualassociation areas. Rothi et al4 posited thatthese lesions deprived the space-time move-ment representations stored in the parietallobe of visual input (fig 1, lesion .A).

In the apraxic conditions described, gestureimitation performance is either comparablewith or somewhat better than gesture to ver-bal command. For example, patients withideomotor apraxia make the most errors onpantomime to command, whereas imitationperformance improves or remains equallyimpaired.56 We describe a left hemispheredamaged patient with ideomotor apraxiawhose imitation of transitive gestures wasmuch worse than his performance when pan-tomiming in response to verbal command.His poorer performance on gesture imitationcould not be explained on the basis of a dis-connection or dissociation of visual areas fromthe space-time movement representationsbecause gesture comprehension was pre-served. The behavioural dissociation betweengesture to command and gesture imitation wefound in this case provides support for amodel of gestural processing in which the rep-resentations used for praxis production andthose important for gesture recognition areindependent. Whereas these representationsare normally linked, in this case they mayhave been disconnected or dissociated.

Case reportA 69 year old, previously healthy, righthanded man was admitted to the hospitalafter having been found lying on the floor ofhis home for an undetermined amount of timein a confused state. Neurological examinationshowed a right foot drop but no other motorabnormalities. Sensory examination showeddecreased sensation to all modalities on theright hemibody. Aphasia and a right visualfield defect were also noted. Computedtomography conducted at the time of admis-sion showed an infarct involving the left

1241

on October 18, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.10.1241 on 1 O

ctober 1994. Dow

nloaded from

Ochipa, Rothi, Heilman

Figure 2 Computed tomography consistent with an infarct in the distribution of the leftmiddle cerebral artery involving the inferior parietal and superior temporal lobes.

inferior parietal and posterior superior tempo-ral lobe (fig 2).

SPEECH AND LANGUAGE EVALUATIONThe patient's speech and language were for-mally evaluated nine days after onset with thewestern aphasia battery.7 Results were consis-tent with conduction aphasia in that sponta-neous speech was fluent in the context ofrelatively spared auditory comprehension andpoor naming and repetition abilities. Verbaloutput was characterised by frequent literaland verbal paraphasic errors and occasionalneologisms. Auditory comprehension was

good for yes or no questions, auditory wordrecognition, and one and two step commands.

Comprehension of multistage, syntacticallycomplex commands was poor. Reading andwriting abilities were severely impaired.

PRAXIS EVALUATIONEvaluation of praxis was conducted threeweeks after onset with a modified version ofthe Florida apraxia screening test (table).'Pantomime in response to verbal command,pantomime imitation, actual tool or objectuse, and gestural comprehension were

assessed. The gestural comprehension subtestwas given before the production subtests.Gesture comprehension was tested by havingthe subject identify a gesture performed by theexaminer. For each item on the productionsubtests, he was required to produce a pan-

tomime in response to verbal command.Immediately after his performance, the exam-

iner performed the correct gesture, and hewas requested to imitate the examiner's pro-

duction. He was then given the actual tool or

object and was asked to demonstrate its use.

The patient's performance was videotapedand scored by a panel of three trained judgeswith the system for the qualitative analysis oferrors described by Rothi et al." (See appendixfor a description of the possible apraxic error

types.) As well as judging the types of errors,

imitation was also judged as to whether it wasinferior, superior, or equivalent to pantomimeto command. For comparison purposes, thepraxis performance of an additional five lefthemisphere damaged apraxic patients was

scored with the protocol described. All sub-jects in this group were right handed and hadhad a single, unilateral stroke of the left hemi-sphere. Hereafter they are referred to as thecomparison group.

ResultsThe patient exhibited a severe ideomotorapraxia to both verbal command and imita-tion. Table 1 gives the praxis test results.Gesture comprehension was spared (14/14correct). Although performance on pan-tomime to verbal command was poor (1/14correct), performance was much worse in theimitation condition on eight out of 14 items

Results ofpraxis testing: left hand

Praxis task

Pantomime content Pantomime to command Pantomime imitation Actual tool or object use

Scissors BPO BPO, M (decline) +Key S 5, 0, M (decline) SWatch A, EC A, EC (same) tToothbrush EC EC (decline) ECPencil M M, BPO (decline)Phone A, EC, M A, EC, M (decline)Screwdriver M M (same) eComb BPO BPO (same)Toothpaste EC EC, M (decline) XKnife M M, IC (decline) FGlass IC X (improved) XSpoon + A (decline) AFork M IC, EC (same) IC, ECDoorknob IC IC (same) F

BPO = body part as object; M = movement; S = sequencing; 0 = occurrence; A = amplitude; IC = internal configuration; BC =external configuration; + = correct. See appendix for error type description.

1 242

on October 18, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.10.1241 on 1 O

ctober 1994. Dow

nloaded from

Conduction apraxia

Figure 3 Modification ofLiepmann's model toaccountfor selectiveimpairment ofgestureimitation (lesion A) andselective sparing ofgestureimitation (route B).

(57%). For example, when asked to demon-strate the use of a key, he made a sequencingerror. When provided with a model for imita-tion, his performance deteriorated such thatthe incorrect gesture was classified as consist-ing of sequencing, occurrence, and movementerrors. On several test items, additional error

types were not seen in the imitation condition,but rather the same error type was noted to beworse in degree than in the pantomime tocommand condition. On only one occasionwas improvement noted with imitation.Considerable improvement was noted,however, in actual tool or object use (10/14correct).The imitation performance of all five of the

patients in the comparison group was judgedto be worse than pantomime to verbal com-mand on only two occasions (3%) for thecomparison group v 57% for our patient).Imitation performance on the remaining itemseither improved (38%) or was equivalent tothe pantomime to command task (59%).

DiscussionOur patient showed a severe ideomotorapraxia in response to both verbal commandand imitation. He did not show, however, thegestural comprehension deficit that has beenreported with other patients who have inferiorparietal lobe lesions. This suggests that otherareas in the left or possibly the right hemi-sphere may mediate his gesture comprehen-sion. These areas normally may have beenused to comprehend gesture or were able tocompensate for damaged areas.

Unlike other apraxic patients reported inthe medical literature56 and the comparisonpatients reported here, his performance on

gesture imitation, when compared with hisperformance to verbal command, deterioratedrather than improved. This particular patternof practic deficits, which has not been previ-ously described, calls for a modification ofprevious models of praxis processing systems.

With a parallel drawn from models oflanguage processing, Rothi et al'0 suggestthat the term "action lexicon" correspondsto the visuokinesthetic motor engrams orLiepmann's "movement formulas." Similarto the lexicon of the linguistic system, theaction lexicon provides a processing advan-tage to gestures for which the user has hadprior experience (learned skilled motor acts).The model of praxis processing presented infig 1 suggests that imitation failure wouldoccur with disorders of action lexicon accessor egress and therefore occur concomitantwith either failures of gesture reception orproduction or from the cumulative effect ofboth reception and production difficulties.The patient's increased difficulty with imita-tion of known gestures, however, cannot beaccounted for solely by a production deficitbecause imitation was worse than perfor-mance to verbal command. It can also not beaccounted for by a failure to gain access to theaction lexicon, a receptive defect, or thecumulative effect of receptive and productiondeficits because the patient had no difficultywith gesture comprehension. The finding thatpantomime to command was superior to pan-tomime imitation in this patient with intactgesture comprehension, supports the postu-late that there may be a division of the actionlexicon into an action input lexicon and anaction output lexicon (fig 3), similar to thatfound in models of word recognition and pro-duction." Intact gesture comprehension inthe context of particular difficulty with imita-tion suggests dysfunction at some point afteraccess to or processing by the action inputlexicon. Because pantomime to commandrequires access to the action output lexicon,defective pantomime to command may resultfrom impairment at or after access to theaction output lexicon. Particular difficultywith pantomime imitation may be accountedfor by increased production difficulty result-ing from additional problems arising beforeaccess to the action output lexicon (fig 3,lesion A).

Again, drawing a parallel from the languagesystem, Strub and Gardner'2 suggest that theverbal repetition deficit in conduction aphasiaoccurs after phonemic analysis of auditoryinput and before the actual encoding ofphonemes for production. Our patient'sapraxic deficit may be similar in that webelieve that there is a disruption at the stagebetween visual analysis of gestures and themotor encoding of known gestures for pro-duction. We therefore propose that "conduc-tion apraxia" may be a suitable descriptiveterm for this disorder. This conductionapraxic patient's behaviour is almost oppositeto that of patients with pantomime agnosia4whose gestural comprehension is impairedwhile pantomime production and imitationare spared. Rothi et al'0 suggest that this con-stellation of symptoms cannot be accountedfor by the action input lexicon to action out-put lexicon route shown in fig 3, whichimplies that pantomime imitation alwaysrequires the same input processing as gesture

1 243

on October 18, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.10.1241 on 1 O

ctober 1994. Dow

nloaded from

Ochipa, Rothi, Heilman

comprehension. Rothi et al 10 comment thatsparing of pantomime imitation in these casessuggests a direct connection between visualanalysis of gesture and the innervatory pat-terns for movement implementation. That is,these patients may be using what Rothi et al 10

term a "non-lexical" action processing routefor imitation (fig 3, route B). Such a routemay also be used by normal subjects who areable to imitate meaningless novel gestures.Mehler'3 described two cases who wereunable to imitate intransitive, non-symbolichand or arm movements in the context of nor-mal pantomime to command and gesturecomprehension. Mehler" did not, however,report testing the imitation of known transi-tive gestures. Rothi et al° suggest that it maybe this "non-lexical" system that was selec-tively impaired in the cases described byMehler"3 whose deficits were limited to theimitation of non-familiar limb movements,those for which there were no "lexicaladdresses." Rothi et al'I also suggest that theexistence of a non-lexical action processormay account for the improvement seen inthose apraxic patients who are less impairedon pantomime imitation. That is, these maybe the patients who selectively spare the non-lexical route. By contrast, those patients whodo not improve in the imitation condition maybe those who impair both the lexical and non-lexical routes of praxis imitation.

This work was supported by the Rehabilitation R and DService and the Medical Service of the Department ofVeterans Affairs.

Appendix9: types of errorsCONTENTP = perseverative: The subject produces a responsethat includes all or part of a previously produced pan-tomime.R = related: The pantomime is an accurately pro-duced pantomime associated in content to the target.For example, the subject might pantomime playing atrombone for a target of a bugle.N = non-related: The pantomime is an accuratelyproduced pantomime not associated in content to thetarget. For example, the subject might pantomimeplaying a trombone for a target shaving.

TEMPORALS = sequencing: Some pantomimes require multiplepositionings that are performed in a characteristicsequence. Sequencing errors involve any perturbationof this sequence including addition, deletion, or trans-position of movement elements as long as the overallmovement structure remains recognisable.T = timing: Any alterations from the typical timing orspeed of a pantomime and may include abnormallyincreased, decreased, or irregular rate of production.O = occurrence: Pantomimes may involve either sin-gle (for example, unlocking a door with a key) orrepetitive (for example, screwing in a screw with ascrewdriver) movement cycles. This error type reflectsany multiplication of single cycles or reduction of arepetitive cycle to a single event.D = delay: Delay in the initiation of a movement.

SPATIALA = amplitude: Any amplification, reduction, or irreg-ularity of the characteristic amplitude of a target pan-tomine.

IC = internal configuration: When pantomiming, thefingers and hand must be in a specific spatial relationto one another to reflect recognition and respect forthe imagined tool. This error type reflects any abnor-mality of the required finger or hand posture and itsrelation to the target tool. For example, when asked topretend to brush teeth, the subject's hand may closetightly into a fist with no space allowed for the imagedtoothbrush handle.BPO = body part as object: The subject uses his orher finger, hand, or arm as the imagined tool of thepantomime. For example, when asked to smoke a ciga-rette, the subject might puff on his index finger.ECO = external configuration orientation: When pan-tomiming, the fingers/hand/arm and the imagined toolmust be in a specific relation to the "object" receivingthe action. Errors of this type involve difficulties ori-enting to the "object" or in placing the "object" inspace. For example, the subject might pantomimebrushing teeth by holding his hand next to his mouthwithout reflecting the distance necessary to accommo-date an imagined toothbrush. Another example wouldbe when asked to hammer a nail, the subject mighthammer in a differing location in space reflecting diffi-culty placing the imagined nail in a stable orientation.M = movement: When acting on an object with atool, a movement characteristic of the action and nec-essary for accomplishing the goal is required. Any dis-turbance of the characteristic movement reflects amovement error. For example, a subject, when askedto pantomime using a screwdriver, may orientate theimagined screwdriver correctly to the imagined screwbut instead of stabilising the shoulder and wrist andtwisting at the elbow, the subject stabilises the elbowand twists at the wrist or shoulder.

OTHERNR = no response.UR = unrecognisable response: A response that is notrecognisable and shares no temporal or spatial featuresof the target.

1 Liepmann H, Maas 0. Concerning the function of the cor-pus callosum and the relationship of aphasia and apraxiato intelligence. Medizinische Klinik 1907;25-26. (Trans-lated by D Kimura, 1980. Translations fromLiepmann's Essays on Apraxia. In: Research Bulletin No506, Department of Psychology, the University of WesternOntario.)

2 Heilman KM, Rothi U, Valenstein E. Two forms of ideo-motor apraxia. Neurology 1982;32:342-6.

3 Rothi LJG, Heilman KM, Watson RT. Pantomime com-prehension and ideomotor apraxia. J Neurol NeurosurgPsychiatry 1985;48:207-10.

4 Rothi LJG, Mack L, Heilman KM. Pantomine agnosia. JNeurol Neurosurg Psychiatry 1986;49:451-4.

5 Heilman KM, Rothi LJG. Apraxia. In: Heilman KM,Valenstein E, eds., Clinical Neuropsychology. New York:Oxford University Press 1985:131-50.

6 Heilman KM, Watson RT, Rothi LJG. Limb apraxia.Current Neurology 1989;9: 179-90.

7 Kertesz A. The western aphasia battery. New York: Gruneand Stratton, 1982.

8 Rothi LJG, Heilman KM. Acquisition and retention ofgesture in apraxic patients. Brain Cogn 1984;3:426-32.

9 Rothi LJG, Mack L, Verfaellie M, Brown P, Heilman KM.Ideomotor apraxia: Error pattern analysis. Aphasiology1988;2:381-8.

10 Rothi LJG, Ochipa C, Heilman KM. A cognitive neu-ropsychological model of limb praxis. CognitiveNeuropsychology 1991;8:443-58.

11 Coltheart M. Functional architecture of the language-processing system. In: M Coltheart, G Sartori andR Job, eds. The cognitive neuropsychology of language.London: Lawrence Erlbaum, 1987.

12 Strub R, Gardner H. The repetition deficit in conductionaphasia: Amnestic or linguistic? Brain Lang 1974;1:241-57.

13 Mehler MF. Visuo-imitative apraxia. Neurology 1987;37:129.

14 Watson RT, Fleet S, Rothi LJG, Heilman KM. Apraxiaand the supplementary motor area. Arch Neurol1986;43:787-92.

1244

on October 18, 2020 by guest. P

rotected by copyright.http://jnnp.bm

j.com/

J Neurol N

eurosurg Psychiatry: first published as 10.1136/jnnp.57.10.1241 on 1 O

ctober 1994. Dow

nloaded from