Authors:Laurel J. Buxbaum, PsyDKathleen Y. Haaland, PhdMark Hallett, MDLewis Wheaton, PhDKenneth M. Heilman, MDAmy Rodriguez, MA, CCC-SLPLeslie J. Gonzalez Rothi, PhD
Affiliations:From the Moss RehabilitationResearch Institute, Philadelphia,Pennsylvania (LJB); Thomas JeffersonUniversity, Philadelphia, Pennsylvania(LJB); Albuquerque Veterans AffairsMedical Center, Albuquerque, NewMexico (KYH); University of NewMexico, Albuquerque, New Mexico(KYH); Human Motor ControlSection, National Institute ofNeurological Disorders and Stroke,National Institutes of Health,Bethesda, Maryland (MH);Department of Veterans Affairs,Baltimore, Maryland (LW); BaltimoreGeriatric Research Education andClinical Center, Baltimore, Maryland(LW); North Florida/South GeorgiaVeterans Health System, Gainesville,Florida (KMH, AR, LJG); andUniversity of Florida, Gainesville,Florida (KMH, AR, LJG).
Correspondence:All correspondence and requests forreprints should be addressed toLaurel J. Buxbaum, PsyD, MossRehabilitation Research Institute,1200 West Tabor Rd., Philadelphia,PA 19141.
Disclosures:This paper is an outgrowth of aworkshop in plasticity/neurorehabilitation researchsponsored and supported by the VABrain Rehabilitation Research Centerof Excellence and the University ofFlorida Department of OccupationalTherapy, Gainesville, Florida. Workon the manuscript was supported inpart by NIH R01-NS036387 to Dr.Buxbaum.
Treatment of Limb ApraxiaMoving Forward to Improved Action
ABSTRACTBuxbaum LJ, Haal KY, Hallett M, Wheaton L, Heilman KM, Rodriguez A,Gonzalez Rothi LJ: Treatment of limb apraxia: moving forward to improvedaction. Am J Phys Med Rehabil 2008;87:149–161.
Limb apraxia is a common disorder of skilled, purposive movement that isfrequently associated with stroke and degenerative diseases such as Alzheimerdisease. Despite evidence that several types of limb apraxia significantly impactfunctional abilities, surprisingly few studies have focused on development oftreatment paradigms. Additionally, although the most disabling types of apraxiareflect damage to gesture and/or object memory systems, existing treatmentshave not fully taken advantage of principles of experience known to affect learningand neural plasticity. We review the current state of the art in the rehabilitation oflimb apraxia, indicate possible points of contact with the learning literature, andgenerate suggestions for how translational principles might be applied to thedevelopment of future research on treatment of this disabling disorder.
Apraxia is a common disorder of skilled, purposive movements. Praxis ismediated by a complex system that stores components of skilled movements,thus providing them a processing advantage (i.e., in terms of accuracy andresponse time) compared with less-practiced movements. Although severaltypes of apraxia have clear impact on functional abilities and are commonconsequences of stroke, Alzheimer disease, and corticobasal degeneration, fun-damental knowledge in a number of areas necessary to guide informed treat-ment is surprisingly lacking. There remains confusion about the definitions,distinctiveness, and mechanisms of various types of apraxia and, indeed,whether any have critical functional significance. In addition, although the mostdisabling types of apraxia reflect damage to systems involved in movement andgesture representation (i.e., memory), the nascent apraxia-treatment literaturehas not taken advantage of principles of experience known to affect skilllearning. The aim of this article is to review the current state of the rehabili-tation of limb apraxia and, on the basis of the learning and plasticity literature,generate suggestions for how translational principles might be applied to guidefuture treatment research.
DEFINITIONS OF APRAXIAThe term apraxia was introduced by Steinthal.1 Whereas this word is
derived from Greek and literally means without action, the term apraxia is used
February 2008 Treatment of Limb Apraxia 149
REVIEW & ANALYSIS
Apraxia
to describe a decrease or disorder in the ability toperform purposeful, skilled movements. The great-est advance in the description and understandingof these disorders is contained in a series of paperswritten between 1900 and 1920 by Hugo Liep-mann.2–4 Liepmann described three forms ofapraxia and, by virtue of his careful evaluations anddiscussions, brought about a paradigmatic shift inour understanding of motor control. These threetypes were limb kinetic apraxia (also called melo-kinetic apraxia or innervatory apraxia), ideomotorapraxia, and ideational apraxia. To this triad,Hanna-Pladdy and Rothi5 and Ochipa et al.6–7
added another type, termed conceptual apraxia,and DeRenzi et al.8 as well as Heilman et al.9
described a fifth type, now called dissociationapraxia.
In this manuscript, we will focus on ideomotorapraxia (hereafter, IMA), for two reasons. First, aswill be discussed, it is extremely common in strokeand degenerative disease (Alzheimer disease andcorticobasal degeneration). Second, it is increas-ingly recognized that IMA has important functionalconsequences, and the disorder is, thus, in need ofcontinued critical investigation, particularly in thearea of treatment.
IMA is usually diagnosed on the basis of spa-tiotemporal errors in the production of transitive(object-related) gesture pantomime to sight of ob-jects, to command, and on imitation of others.10–14
Kinematic analyses have revealed that IMA patientspantomime skilled tool-use movements with ab-normal joint angles and limb trajectories, and withuncoupling of the spatial and temporal aspects ofmovement.13 Spatiotemporal errors persist to alesser degree with actual tool use.15,16 The deficit isnot restricted to meaningful movements, and it hasalso been observed in meaningless postures17–19
and sequences.20,21 IMA is also associated with cog-nitive deficits in declarative knowledge of the ac-tion appropriate to objects,22 impairments in me-chanical problem solving,23 deficits in motorplanning,21,24–26 and difficulty learning new ges-tures.27,28 Testing for IMA frequently includes pan-tomiming to command of transitive (familiar ac-tions with objects, such as brushing teeth) andintransitive (symbolic movements without objects,such as the sign for “crazy”) movements, imitationof the examiner performing transitive, intransitive,and novel meaningless movements, and gesture inresponse to seeing and holding actual tools, as wellas the objects on which tools act.
Several investigators have distinguished be-tween IMA with impaired gesture recognition (rep-resentational IMA) and IMA with intact recognition(dynamic IMA).11,29,30 In representational IMA, aninability to discriminate correctly from incorrectlyperformed meaningful object-related hand move-
ments correlates strongly with an ability to pro-duce the same movements, suggesting that thesame representations are likely to underlie both.31
Additionally, representational (but not dynamic)IMA patients are significantly more impaired whenproducing object-related than symbolic, non–ob-ject-related movements.32 This, in turn, suggeststhat the damaged system underlying representa-tional IMA is specialized for movements related toskilled object use.
THE FUNCTIONAL IMPLICATIONS OFLIMB APRAXIA: DOES LIMB APRAXIAMATTER IN THE REAL WORLD?
Historically, most clinicians and researchersbelieved that limb apraxia had little or no real-world implications.4,10,33–35 This is emphasized byDeRenzi,35a who wrote that “apraxia rarely appears ineveryday situations and spontaneous motor behav-ior, predominantly emerging when gestures areproduced out of context as a purposeful response toan artificial request.” Although not specified, itseems that this view was particular to IMA andstemmed from the notion that apraxia was presentwhen pantomimes to command and imitation weretested but improved when the use of actual objectswere examined.
It is now widely believed that IMA impairsreal-world functioning, but there are still remark-ably few studies demonstrating such a relationship.In addition, most studies to date have been fraughtwith problems. First, these studies usually have notruled out the influence of all other factors, such ashemiparesis. They commonly have compared theperformance of apraxic and nonapraxic patientswith left-hemisphere damage,36–41 but, relative tononapraxics, apraxics are often more impaired inother domains, such as language, sensory, and mo-tor skills. Therefore, it is difficult to know whetherlimb apraxia is the best predictor of functionalskills. Second, apraxics typically have larger lesionsthan do patients without apraxia, and those lesionsmore frequently damage the left-parietal and fron-tal regions,42 which are also important for manyother cognitive functions that could, again, con-found the findings. Regression approaches havebeen used to evaluate the unique impact of variousfactors, including limb apraxia, on activities ofdaily living (ADLs),41,43–45 in some cases after con-trolling statistically for factors such as lesion size,primary motor deficits, and/or other cognitive def-icits. However, these studies usually have sufferedfrom statistical problems related to a small numberof subjects relative to the number of predictorsexamined.
Another problem in efforts to understand theinfluence of apraxia on disability concerns theuse of a wide variety of functional measures,
150 Buxbaum et al. Am. J. Phys. Med. Rehabil. ● Vol. 87, No. 2
including object use,46,47 performance-basedmeasures of ADLs,36,37,39,41,43,48,49 and caregiveror patient report of daily functioning.39,44,45,50
These outcome measures vary in complexity fromisolated object use, such as brushing teeth,51 to sim-ulated ADLs, such as picking up a bean with aspoon,38,39,41 to instrumental ADLs, such as eating ameal,36 dressing,49,52 preparing food,43,48,53–55 orchanging batteries in a recorder.37 It is common inperformance-based studies to use instruments thatdo not have demonstrated reliability; thus, validity isfrequently demonstrated only in the context of thespecific study. In addition, there are significant prob-lems with obtaining reliable measures of these skillsbecause the tasks are usually quite complex and thenumber of possible errors is large. Furthermore, be-cause performance-based tasks are dependent on agreat number of cognitive abilities, patients may beimpaired for different reasons.56,57
Taken together, these problems in the litera-ture suggest that future studies must (1) examinethe relationship of different types of limb apraxia toreal-world functioning (ADLs and instrumental ac-tivities) of various kinds, and (2) use sufficientlylarge groups of patients to provide sufficient powerfor analysis. It is also reasonable to consider at leasttwo different approaches for subject recruitment.The first approach examines well-characterized pa-tients with unilateral focal lesions; the second ap-proach examines a broader range of patients withand without limb apraxia, without regard to lesionlocation. The latter approach may yield patientsmore broadly representative of the patients typi-cally seen in the clinic.
Finally, some of the most innovative work inthis area attempts to identify cognitive mecha-nisms that are associated with ideomotor limbapraxia and potentially with the resulting deficitsin real-world functioning (see Sunderland andShinner58 for a review). These cognitive processesinclude mechanical problem solving,46 sequenceplanning and organization,21 the ability to developand/or retrieve optimal motor programs,13 knowl-edge of how to manipulate an object,22,25,59 andknowledge of optimal hand position when real-world objects provide minimal cues.25,39
TREATMENT OF LIMB APRAXIAA recent review of the literature on the treat-
ment of limb apraxia yielded reports of ten treat-ment approaches, many of which were single-casestudies. Methods reported were varied and can besummarized as follows.
Multiple CuesThe multiple-cues treatment method was de-
veloped in 1991 by Maher et al.60 for a 55-yr-oldmale with chronic ideomotor apraxia and intact
gesture recognition. It focused on treatment ofgestures, using presentation of multiple cues, in-cluding tools, objects, visual models, and feedback.Errors were corrected using imitation and physicalmanipulation. As performance improved, cues weresystematically withdrawn. The individual partici-pated in daily, 1-hr sessions for 2 wks. The multi-ple-cues method resulted in positive effects, withtreated gestures showing some lasting improve-ment. Generalization to untreated gestures was notassessed.
Error ReductionIn an attempt to define the active components
of the multiple-cues method, Ochipa and col-leagues61,62 conducted a treatment study aimed attreating specific error types. Two males (44 and 66yrs old) with chronic Broca aphasia and ideomotorapraxia, but preserved gestural recognition, partic-ipated in the treatment. Treatment duration andintensity varied, with the 44-yr-old receiving treat-ment four times per week (n � 44 sessions) and the66-yr-old receiving treatment two times a day,twice a week (n � 24 sessions). The goals of treat-ment consisted of reduction of external configura-tion, movement, and internal configuration errors,depending on the error types exhibited by the in-dividual. Reduction of external configuration er-rors involved training the individual to correctlyorient his hand to objects, whereas reduction ofinternal configuration errors involved positioningof the hand and fingers to accommodate a tool.Movement errors were reduced through verbal de-scriptions to guide joint movement while gestur-ing. Only one error type was addressed at a time,and feedback was only provided about the errortype being trained. Error-reduction treatment re-sulted in a significant and lasting improvement ontreated gestures for both individuals. However, nogeneralization to untreated error types or gestureswas noted. Improvements were noted to continueat the 2-wk posttreatment follow-up, but later fol-low-ups were not performed.
Six-Stage Task HierarchyThe task hierarchy method was developed by
Code and Gaunt,63 who studied an individual withsevere chronic aphasia, ideomotor apraxia, and ide-ational apraxia. This six-stage task hierarchicaltreatment for limb apraxia was a modification of aneight-step continuum used to treat apraxia ofspeech.64 The Code–Gaunt method requires thepatient to produce target words and signs in vari-ous combinations and in concert with the therapistin response to a therapist model or picture elicita-tion. The patient participated in 45-min sessionsonce weekly for 8 mos. The six-stage task hierarchymethod resulted in acquisition of trained signs and
February 2008 Treatment of Limb Apraxia 151
a nonsignificant trend toward improvement in un-trained signs during treatment. Maintenance ofeffects was not formally tested, but the authorsprovide anecdotal reports of the patient’s contin-ued use of signs in group treatment sessions. Treat-ment did not impact limb apraxia.
Conductive EducationThe conductive education method was devel-
oped by Pilgrim and Humphreys65 for a patientwith head injury and chronic unimanual apraxia ofthe nondominant limb. Treatment focused on atask analysis of the movements and articulation ofgoal-directed tasks. The treatment began withphysical manipulation plus verbalization of taskelements (e.g., “reach the beaker, clasp the beaker,carry to my lips, drink, stop”), and those cues weresystematically withdrawn as performance im-proved. There were daily 15-min sessions for 3 wks.The conductive education method improved thispatient’s performance on treated items comparedwith untreated items. There was no generalizationto untreated objects. Maintenance of effects wasnot assessed.
Strategy TrainingThe strategy training method was developed as
a compensatory technique for individuals with ADLimpairment secondary to apraxia. This method wasfirst described in the literature in a study of 33individuals with apraxia secondary to left-hemi-sphere stroke.66 The patients were trained on threeADLs, and the specific method of treatment waschosen according to each individual’s performancein baseline testing of those tasks. A similar strategytraining method using five ADLs was studied inanother group of 56 individuals with left-hemi-sphere stroke and subsequent apraxia. Both strat-egy training approaches focused on the use of inter-nal compensatory strategies (i.e., self-verbalization)and external compensatory strategies (i.e., use of pic-tures) to maximize independence. The duration andintensity of treatments varied among individuals inboth studies. Strategy training resulted in positiveoutcomes across all domains measured (effect sizeswere 0.37 for the ADL tasks and 0.47 for the BarthelADL index; both were significantly greater than forpatients receiving usual occupational therapy treat-ment), but the improvements were not lasting.67,68 Inthe final study in this series, there was an additionalfinding of interest—namely, maintenance of gains intrained tasks at 5-mo follow-up.
Transitive/Intransitive Gesture TrainingThe transitive/intransitive gesture training
method was investigated by Smania and colleagues69
in 22 individuals at least 2 mos after onset of aleft-hemisphere stroke with subsequent ideomotor
limb apraxia. Treatment focused on the training oftransitive and intransitive gestures. Transitive ges-ture training consisted of three phases in which theindividual was (1) shown use of common tools, (2)shown a static picture of a portion of the transitivegesture and asked to produce the pantomime, and (3)shown a picture of a common tool and asked toproduce the associated gesture. The intransitive ges-ture training also consisted of three phases in whichthe individual was (1) shown two pictures, one illus-trating a context and the other showing a relatedsymbolic gesture, and asked to reproduce the gesture;(2) shown the context picture alone and asked toreproduce the gesture; and (3) shown a picture of adifferent but related contextual situation and asked toreproduce the gesture. Fifty-minute treatment ses-sions were administered three times per week forapproximately 10 wks, with the number of total treat-ment sessions ranging from 30 to 35. A control groupwas administered aphasia treatment only for a similarintensity and duration. Results indicated a differencebetween the two groups after treatment, with thegesture training method resulting in improved per-formance on an IMA test (U � 69.00, P � 0.016), agesture comprehension test (U � 64.00, P � 0.018),and an ADL questionnaire (U � 53.50, P � 0.01).Importantly, patients and caregivers reported moreindependence in ADLs after treatment. Nine patientsshowed maintenance of gains at 2 mos after treat-ment.
“Rehabilitative Treatment”Smania and colleagues70 (p2052) reported a pos-
itive outcome with a so-called rehabilitative treat-ment. It was noted that the treatment was “devisedto treat a wide range of gestures and to reduceseveral types of praxic errors” and that it “useddifferent contextual cues to teach patients how toproduce the same gesture under different contex-tual situations.” Thus, although details were notprovided, the treatment seems substantially similarto the one previously reported by this group.69
Forty-one postacute left-hemisphere stroke pa-tients with limb apraxia (either ideational or IMA—not defined) were assigned randomly to treatmentor no-treatment groups. The no-treatment groupreceived aphasia therapy. Patients attended 30 �50-min sessions during the course of 10 wks. Al-though the groups were equivalent in ADL perfor-mance, apraxia scores, and ADL questionnairescores before treatment, they differed significantlyon these measures after treatment, both immedi-ately and after a 2-wk delay.
Errorless Completion � ExplorationTraining
The errorless completion/exploration trainingmethod was developed by Goldenberg and Hag-
152 Buxbaum et al. Am. J. Phys. Med. Rehabil. ● Vol. 87, No. 2
mann51 for 15 individuals with IMA (impairmenton gesture imitation and gesture to sight of ob-jects) who were, on average, 6.1 wks since onset ofa left-hemisphere stroke with subsequent aphasiaand severe limb apraxia. The errorless completionmethod used physical manipulation during ADLs,simultaneous demonstration of ADL by the exam-iner and imitation by the patient, and copy by thepatient after demonstration during performance ofthree ADLs. The exploration training method di-rected attention to functional significance of de-tails and critical features of action but did notincorporate direct practice of actions with actualobjects. These two methods were combined andtreatment was applied to one ADL at a time dailyfor 20–40 mins for 2–5 wks. Combined errorlesscompletion/exploration training resulted in posi-tive effects that were lasting for individuals whoremained active in ADLs at home. A subsequentstudy was conducted by Goldenberg et al.37 com-paring these two methods in six individuals withleft-hemisphere stroke and subsequent chronicaphasia and limb apraxia. Each treatment type wasapplied on a different pair of ADLs. The explorationtraining method had no effect. The errorless com-pletion method yielded a positive and lasting effect.When different objects were used to test ADL, how-ever, the rate of errors increased, and were com-parable with untrained gestures. Therefore, therewas no evidence of generalization.
SUMMARY OF TREATMENTLITERATURE
Table 1 provides a summary of the ten apraxiatreatment approaches discussed in the literature todate. Several trends are worth noting. First, apraxiatype is frequently poorly characterized. For exam-ple, although gesture recognition is clearly an im-portant index of the integrity of gesture represen-tations (which, in turn, may have importantimplications for rehabilitation strategies), recogni-tion testing is usually not performed. Second,whereas some studies provide data on treatmenteffects and generalization to untreated items, theyare more sparse with regard to treatment effects ondegree or nature of limb apraxia, maintenance oftreatment effect, and impact of treatment on ADLs.Third, the duration and intensity of treatment dif-fers within and across studies, making it difficult todetermine the active components of the treatment.Fourth, the length of time between termination oftreatment and follow-up differs across studies,which renders it difficult to compare the lastingeffects of treatment on limb apraxia or ADLs. Fi-nally, methods such as the nature of the feedbackor correction are commonly underspecified inthese reports if described at all, making replicationin additional subjects nearly impossible. Despite
these issues, the data consistently suggest thatintervention yields a treatment effect. Further-more, in the cases where it is reported, there isindication of maintenance of treatment effects, andimpact on nature/degree of limb apraxia as well ason ADL facility. Thus, it seems that the evidencebased on these ten Phase I studies suggests thatlimb apraxia is amenable to treatment. However,according to Robey and Schultz,71 the purpose ofPhase I research is to develop hypotheses, proto-cols, and methods; establish safety and activity;determine the best outcome measures; identify re-sponders vs. nonresponders; determine optimal in-tensity and duration; and determine why the treat-ment is producing an effect.71 Little of thisinformation is found in these ten reports, and,thus, we must continue to promote systematicinquiry until the objectives of Phase I research aresatisfied for limb apraxia.
Evidence suggests that nine of the ten treat-ments reported in the literature yielded a treatmenteffect. However, only four of these nine treatmentsresulted in generalization. Because the ultimate goalof rehabilitation is the use of acquired skill in theindividual’s natural environment, it is important toconsider why certain treatments resulted in gener-alization, whereas others did not.
Nadeau et al.72 recently have identified seventreatment attributes that may contribute to gener-alization in language rehabilitation: (1) intrinsic:application of knowledge acquired in therapy; (2)cross function: development of knowledge that canbe applied to multiple tasks; (3) extrinsic: acquisi-tion of a technique that can be applied outside oftreatment to rebuild function (requires motiva-tion); (4) mechanistic: training of key brain re-sources (i.e., working memory capacity, distributedconcept representations, intentional bias); (5) sub-strate mediated: development of a critical mass ofskill needed to further the therapeutic process—necessary for intrinsic/extrinsic mechanisms to op-erate; (6) contextual: learning environment resem-bles retrieval environment; and (7) sociallymediated: restoration of social context and changein perception regarding roles to promote activity inthe environment.
Unfortunately, in the realm of apraxia rehabil-itation, there is no clear relationship between theseputatively critical mechanisms and treatment gen-eralization. All four treatments that generalizedincluded cross function and extrinsic mechanisms,but some treatments that did not generalize in-cluded these mechanisms as well. Similarly, sometreatments that were mechanistic generalized,whereas others did not. Of the three treatmentsthat incorporated home practice (contextual mech-anism), none resulted in generalization. In addi-tion, on the basis of the available information,
February 2008 Treatment of Limb Apraxia 153
TAB
LE1
Sum
mar
yof
apra
xia
treat
men
tstu
dies
Apr
axia
Typ
e(s)
Tra
ined
Item
sD
urat
ion
Inte
nsit
yT
reat
men
tE
ffect
Gen
eral
izat
ion
Mai
nten
ance
Apr
axia
Impa
ctA
DL
Impa
ct
Mul
tipl
ecu
es(n
�1)
IMA
Ges
ture
s2
wks
1hr
daily
YY
Y—tr
eate
dit
ems
only
(2w
ks)
NA
NA
Err
orty
pere
duct
ion
(n�
2)IM
AG
estu
res
Vari
ed;
6–11
wks
Vari
ed;
once
daily
4da
ys/w
kor
twic
eda
ily2
days
/wk
YN
Y—tr
eate
der
ror
type
son
ly(2
wks
)
NN
A
Six-
stag
eta
skhi
erar
chy
(n�
1)IM
A�
IAG
estu
res
8m
os45
min
s;on
cew
eekl
yY
NN
AN
NA
Con
duct
ive
educ
atio
n(n
�1)
IMA
Ges
ture
s3
wks
Dai
lyY
NN
AN
AN
A
Stra
tegy
trai
ning
(n�
89)
IA?*
ADL
Vari
ed;
8–12
wks
Vari
ed;
25se
ssio
ns,
15hr
sto
tal
YY
N(5
mos
)Y
Y
Tran
siti
ve/in
tran
siti
vege
stur
etr
aini
ng(n
�13
)
IMA
Ges
ture
s10
–11
wks
35se
ssio
ns,
50m
ins
each
YY
NA
YN
A
Reh
abili
tati
vetr
eatm
ent(
n�
20)
IAor
IMA
Ges
ture
s10
wks
30se
ssio
ns,
50m
ins
each
YY
Y(2
wks
)Y
Y
Err
orle
ssco
mpl
etio
n�
expl
orat
ion
trai
ning
(n�
15)
NA
ADL
2–5
wks
5da
ys/w
kpl
us20
–40
min
sof
prac
tice
daily
YN
Y(6
–30
mos
)N
AN
A
Err
orle
ssco
mpl
etio
n(n
�6)
IMA
ADL
2w
ks6
sess
ions
,1
hrea
chY
NY
(3m
os)
NA
NA
Exp
lora
tion
trai
ning
(n�
6)IM
AAD
L2
wks
6se
ssio
ns,
1hr
each
NN
N(3
mos
)N
AN
A
IMA,
ideo
mot
orap
raxi
a;IA
,id
eati
onal
apra
xia;
Y,ye
s;N
,no
;N
A,no
tas
sess
ed/n
oin
form
atio
npr
ovid
ed.
*In
abili
tyto
carr
you
tpu
rpos
eful
acti
viti
es.
154 Buxbaum et al. Am. J. Phys. Med. Rehabil. ● Vol. 87, No. 2
there seems to be no consistent relationship be-tween duration/intensity/type of items trained andgeneralization of results. These equivocal resultssuggest that whereas limb apraxia may be amena-ble to treatment, systematic investigation of factorspromoting generalization is still essential.
MOTOR LEARNING AND MOTORPLASTICITY: AN OVERVIEW
The learning of skilled movements is calledprocedural learning, and its underlying mecha-nisms and neuroanatomical correlates differ fromdeclarative learning.73 In the following sections,we will provide a brief introduction to the litera-ture on motor learning and plasticity, with an eyetoward applying this literature to the study of IMA.
Some of the actions typically assessed in motorlearning studies differ in complexity and/or mean-ingfulness from the skilled actions that comprisepraxis. A number of motor learning studies, how-ever, have used complex, learned actions that arearguably akin to what we commonly term praxismovements. Other motor learning studies haveexamined complex spatiomotor transformationsthat may have relevance to spatial coding of com-plex action. Thus, it is important to carefully ex-amine the motor learning literature for points ofpossible convergence with the study of learning inapraxia.
Neuroanatomical ConsiderationsThe primary motor cortex in particular exhib-
its a great deal of plasticity as a function of motorlearning. Using transcranial magnetic stimulation,a number of investigations have mapped the degreeand extent of excitability of individual muscles onthe scalp surface. Body parts that are used morehave a larger representation, and this representa-tion shrinks if the body part is not used (see thestudy by Pascual-Leone et al.74). On the basis ofneuroimaging paradigms, a variety of brain regionshave been demonstrated to be active depending onthe task and the stage of motor learning; in nearlyall cases, however, there is activation of the pri-mary motor cortex.75
In most neuroimaging studies, cerebellar acti-vation is evident in the learning phase and declineswhen the movement is learned. This certainly in-dicates a role in learning and, in particular, sug-gests that the cerebellum may critical for develop-ing the movement representation but not storingit. The frontal and parietal lobes are also clearlyinvolved in motor learning, but the precise struc-tures involved in early vs. later stages of learningare unclear. For example, a frontal-to-parietal shiftin activation has been observed as a sequence taskis learned,76 and a prefrontal-to-premotor, poste-rior parietal, and cerebellar shift in activation has
been observed in force adaptation learning.77 Onthe other hand, several studies using motor se-quence tasks and at least one using a rotationallearning task have demonstrated that parietal acti-vation is associated with early stages of learning,with greater cerebellar and/or premotor involve-ment in later stages.78–81 At this juncture, we mayconclude that the parietal regions so frequentlylesioned in apraxic patients are clearly important inaspects of skill learning.
There is evidence that perilesional plasticitymay play a role in recovery of function after stroke.It has been shown, for example, that after fingertracking movements, paretic stroke patients im-proved in finger pointing accuracy and grasp andrelease capabilities.82 These functional gains were ac-companied by increased functional magnetic reso-nance imaging activations in sensorimotor areas ofthe lesioned hemisphere and diminished activationsin the intact hemisphere (see also Fridman et al.83).
At least one previous account has attributedpreserved function in apraxia to preservation ofnondominant (right)-hemisphere frontoparietal re-gions involved in praxis function.84 On the otherhand, nondominant-hemisphere plasticity changeshave been demonstrated to be maladaptive in re-covery from aphasia,85 and they may plausibly besimilarly counterproductive in apraxia recovery.Additional investigations are required to shed lighton this question.
Implicit and Explicit Skill LearningA considerable literature attests to important
differences between skill learning that is unavail-able to conscious experience (implicit learning)and that which is cognitively accessible. Ideally, thestudy of learning in apraxia could tap into thislarge body of evidence to support the framing ofhypotheses and predictions. However, one criticalconcern is that it is not clear whether to alignpraxis learning with explicit or implicit knowledge,or both. The types of complex skills that fall underthe rubric of praxis are not typically verbalized, yetthey can be made explicit under certain circum-stances. It is, perhaps, most reasonable to beginwith the hypothesis that praxis learning is moresimilar to implicit procedural learning than tolearning of declarative information. Specific inves-tigations that test predicted patterns of results ac-cording to this hypothesis need to be performed.
A typical exploration of skill learning entailsthe use of serial reaction-time tasks. Participantsare usually faster at performing sequences of keypresses that are repeated throughout an experi-ment, even though they are unaware of the repe-tition. This is an example of implicit learning. Withadditional practice, the sequence can frequently bespecified; in this case, the learned information has
February 2008 Treatment of Limb Apraxia 155
become declarative as well as procedural. Perfor-mance gets even better at this stage, but the sub-ject’s strategy can change because the stimuli canbe consciously anticipated.
Honda et al.86 examined the dynamic involve-ment of different brain regions in implicit andexplicit motor sequence learning using a serialreaction-time task and positron emission tomogra-phy. During the implicit learning phase, when thesubjects were not aware of the sequence, improve-ment of the reaction time was associated withincreased activity in the contralateral primary sen-sorimotor cortex. Explicit learning, reflected by apositive correlation with correct recall of the se-quence, was associated with increased activity inthe posterior parietal, precuneus, and premotorcortices bilaterally; in the supplementary motorarea, predominantly in the left-anterior part; in theleft thalamus; and in the right-dorsolateral pre-frontal cortex. In a study by Grafton et al.,87 therewas activation of the contralateral primary motorcortex, supplementary motor area, and putamen inan implicit learning task, and activation of ipsilat-eral dorsolateral prefrontal cortex and premotorcortex as well as bilateral parietal cortex duringexplicit learning.
In summary of the studies of motor learning inhealthy subjects, it seems that multiple structuresin the brain are involved, and differential involve-ment arises at different stages. The primary motorcortex and cerebellum (and, sometimes, the pari-etal cortex) are active early, and at least the formerseems to play a role in implicit learning. Premotorand parietal cortical areas are active later and seemto play a role in explicit learning, perhaps in part bystorage of the sequence. This concept is supportedby the observation that the premotor and parietalareas increase their activation in proportion to thelength of a sequence performed from memory.88
The relation is obvious to regions that, when dam-aged, cause apraxia.
PRINCIPLES OF MOTOR LEARNING ASTHEY MAY BE RELEVANT TO APRAXIAREHABILITATION
Several basic principles of motor learning havebeen explored in other aspects of motor controlrehabilitation, but they have received relatively lit-tle attention in the study of IMA.
Internal Models of MovementThe motor system in healthy participants is
adept at developing internal models that representthe kinematics (geometry and speed) and dynamics(forces) of a motor task. Forward models calculatethe movements resulting from a given pattern offorce (dynamics) or the limb positions resultingfrom a given pattern of joint rotation (kinematics).
Inverse models compute the muscle forces ormovements needed to reach a visual goal or goalposture.89 The learning (i.e., practice-dependentreduction of error) of kinematic and dynamic in-ternal models seems to be separable, and it may bedisrupted by different brain lesions.90
Several models of motor performance distin-guish a mode of action concerned with planning,learning, and motor prediction, and another spe-cialized for motor execution and control (seeKeele91). One influential account distinguishes se-mantic representations necessary for motor learn-ing and planning from pragmatic representationssubserving the control and execution of action.92
The planning mode has been proposed to generatemovement parameters by way of internal models.The execution mode, in contrast, emphasizes on-line control that is sensitive to current environ-mental conditions.
Recent investigations provide indirect evi-dence that patients with IMA may be impaired inlearning and/or accessing internal models of move-ment. Motor imagery has been proposed by severalinvestigators to serve as a proxy for motor planningin the absence of execution.93–97 Sirigu et al.98 andBuxbaum et al.25 have demonstrated that partici-pants with left-parietal lesions and IMA were im-paired in motor imagery. In contrast, these pa-tients perform well on tasks more reliant on onlinecontrol, such as reaching and grasping with visualfeedback.13,26 The nature and extent of putativedeficiencies in generating and accessing internalmodels are being explored in several of the authors’laboratories, using visuomotor and force-field ad-aptation paradigms borrowed from the motor con-trol literature. Such studies are an important stepin developing rehabilitation paradigms targeted atthe relearning of appropriate internal models.
Practice SchedulesIt is clear that practice benefits motor learn-
ing, but optimal types and schedules of trainingremain unclear and may vary across tasks. In mostmotor tasks, practice that is distributed over(rather than massed in) time seems to result inoptimized learning and retention.99 In learningnew sensorimotor transformations, rest breaks be-tween sessions are of benefit and may allow for theconsolidation of newly acquired internal models.100
It is also frequently beneficial to train a variety ofsimilar movements to encourage so-called contex-tual interference. Shea and Kohl,101 for example,found in a force-learning task that filling the in-tertest-trial interval with related motor tasks sig-nificantly improved retention. Ollis et al.102 havedemonstrated that learning a variety of knot-tyingmovements enhances learning, even for novicespracticing complex knots. It has been suggested,
156 Buxbaum et al. Am. J. Phys. Med. Rehabil. ● Vol. 87, No. 2
however, that the benefit of contextual interferencemay be task specific.103 Additionally, a contextualinterference manipulation in patients with Parkin-son disease did not enhance learning, suggestingthat successful learning strategies in healthy con-trols may not generalize to brain-damaged pa-tients.104 It is also of interest to note that thetraining of items that share many features withother items is disruptive and is not beneficial in thelexical–semantic domain.105,106 Because object-re-lated praxis movements are complex skills withclose ties to semantic knowledge,22 it remains un-clear whether training on shared or distinctivemotor features, semantic features, or both will beoptimal in praxis rehabilitation.
The Role of Feedback and ErrorCorrection
Feedback and knowledge of results frequentlyfacilitate motor skill acquisition. Recent investiga-tions have probed the types of feedback that may bemost optimal, and here, as in other areas of motorlearning, the answer is unclear. For example, vary-ing the movement component about which feed-back is provided may benefit simple skill learning,but it may also disrupt more complex motor skilllearning.107
In the domain of cognitive implicit learning,error may be disruptive. As a result, rehabilitationparadigms have evolved that emphasize errorlesslearning. Performance may be “shaped” by mini-mizing opportunities to make errors and by re-warding successful performance. In contrast, in thedomain of simple movements, such as reachingunder visual guidance, performance seems to be“tuned” by the opportunity to correct error (e.g.,Rossetti et al.108). The role of error in these differ-ent types of learning remains poorly understood;moreover, it is not clear whether and where praxismovements may fall on this continuum.
Hemiparetic stroke patients without IMA areable to adapt to forces applied perpendicularly tothe moving hemiparetic arm109 as well as tospringlike forces that act against movement110
when they receive feedback about error. This sug-gests that hemiparetic patients can use error toadjust internal models of movement to achieve anintended goal.109,111 It has also been suggested thatperception of gross errors may enhance the recov-ery process in stroke.112
Unfortunately, patients with apraxia frequentlyexhibit some degree of anosognosia, or unaware-ness of deficit. They may recognize that they areunable to move correctly, but they fail to recognizethe extent of deficit, or they may attribute it toclumsiness, memory loss, or intellectual de-cline.113 It may be necessary to provide augmentedfeedback about error. Fortunately, a number of
virtual-reality paradigms under recent develop-ment present promising opportunities to do justthis (see Holden114).
Paradigms using robot-assisted devices115,116
can launch correct actions based on electromyo-graphic activity that is associated with the inten-tion to act. Thus, preparatory activity is linked to acorrect response, and errors are prevented. Thiswould seem to be an extremely useful feature.However, given that IMA patients may fail at thelevel of planning and intention, it is not obviousthat robot-assisted therapies will be helpful in therehabilitation of IMA, unless the correct perfor-mance of an act can feed back to augment theputatively deficient internal model.
SUMMARY AND RECOMMENDATIONSThere are several different subtypes of apraxia,
resulting in some cases from damage to differingunderlying neural systems. Ideomotor, ideational,and conceptual apraxia all seem to impact real-world functioning. Development of appropriatetreatment paradigms is clearly needed. A review ofthe apraxia treatment literature to date reveals thatthe field is in the early stages of efforts to developeffective treatments and that most studies haverelied on individual-case, experimental designs. Ad-ditional problems include poor specification of pa-tient characteristics, including incidence of apha-sia; variable criteria for diagnosing apraxia; vaguedescription of treatments applied; unequal applica-tion of treatment, even within a given study; andabsence of information about treatment generali-zation. Most central to the aims of this review,principles from the existing motor learning litera-ture have not yet informed the development oftreatment studies.
The motor learning literature identifies severalprinciples that may benefit the rehabilitation ofapraxia, if appropriately applied. For example, dis-tributed practice of the target task seems to im-prove learning and retention. Creating contextualinterference by interleaving the target task withother similar tasks may aid117 or disrupt (c.f. Plautet al.106) generalization. Feedback of results shouldbe provided. Intensity of practice is also clearlyimportant.
One potential strategy in the development ofapraxia treatment studies is to systematically varyone treatment feature at a time (e.g., massed vs.distributed practice schedule; similarity or distinc-tiveness of items; presence or absence of feedback;shaping of easier to harder items to maximizesuccess, as opposed to allowing errors) while sys-tematically holding the others constant. This isclearly preferable, from the perspective of clarify-ing the features of the training that are critical. Onthe other hand, there is, unfortunately, very little
February 2008 Treatment of Limb Apraxia 157
to suggest how these motor learning principles arebest parameterized (e.g., in terms of strength, du-ration, or intensity) or applied to the treatment ofIMA. Another strategy, then, is to attempt to obtaina beneficial effect by “loading” the treatment on allof the motor learning features that may plausiblybe beneficial, and, if an effect is obtained, follow upwith studies designed to disentangle the critical vs.noncritical factors. Of course, if no training benefitis observed, then it would be unclear which fea-tures were applied incorrectly, and this, in turn,would necessitate a return to the “one feature at atime” strategy.
As an exercise, at least, we can imagine atreatment study based on the strategy of loadingthe treatment with principles derived from themotor learning literature. One might predict, forexample, that deficits in naturalistic action may bemost successfully treated by providing an intensebut distributed schedule of practice on a variety oftargeted naturalistic tasks, interleaved with othersimilar tasks. Principles of shaping might be pre-dicted to be beneficial, such that easy tasks are usedearly in training and harder tasks later in training,such that performance is successful. On the otherhand, opportunities to correct errors should beprovided, should they arise.
The apraxia literature also provides some hintsabout other factors that may impact rehabilitation.A recent learning study from the lab of one of theauthors118 has assessed the role of the affordancesof unfamiliar objects—in this case, the degree towhich the unfamiliar objects signal the actions asso-ciated with them by virtue of their shape—in learn-ing new, object-related gestures. Patients with IMA,but not age- and education-matched nonapraxic left-hemisphere stroke patients, were significantly betterat learning new gestures when the gestures werehighly afforded by their associated objects. This affor-dance benefit could clearly be exploited in the designof future treatment studies by focusing early treat-ment on high-affordance objects.
Tasks trained early in a shaping procedure maybe designed to be “easy” in a number of othercritical ways. Clearly, these early tasks should havefew steps. Arrays should be simple, with few visualelements, and no distracting (task-irrelevant) ob-jects. Spatial consistency of object placement fromtrial to trial is also critical.119 These task and objectfeatures may all be titrated gradually, such thattasks higher up in the shaping hierarchy are in-creasingly complex with respect to these features.
Treatments must be applied identically acrossall treated subjects. Treated and untreated patientsmust either be matched across a large number ofputatively important variables—including lesionsize, severity of cognitive and language deficits,apraxia type (and subtype) and severity, and motor
impairment—or sample sizes must be large andpatients randomly assigned to treated and un-treated groups. Efficacy of treatment should beassessed by applying pre- and posttreatment mea-sures of caregiver burden, performance of ADLs,and/or functional independence that are differentfrom the trained tasks.
REFERENCES1. Steinthal P: Abriss der Sprachwissenschaft. Berlin, Ferd,
8. DeRenzi E, Faglioni P, Sorgato P: Modality-specific andsupramodel mechanisms of apraxia. Brain 1982;105:301–12
9. Heilman KM, Gonzalez Rothi LJ: Apraxia, in Heilman KM,Valenstein E (eds): Clinical Neuropsychology, ed 3. NewYork, Oxford University Press, 1993, pp 141–50
10. DeRenzi E, Motti F, Nichelli P: Imitating gestures: a quan-titative approach to ideomotor apraxia. Arch Neurol 1980;37:6–10
11. Heilman KM, Rothi LJ, Valenstein E: Two forms of ideo-motor apraxia. Neurology 1982;32:342–6
12. Goldenberg G: Imitating getures and manipulating a man-nikin—the representation of the human body in ideomo-tor apraxia. Neuropsychologia 1995;33:63–72
13. Haaland KY, Harrington DL, Knight RT: Spatial deficits inideomotor limb apraxia: a kinematic analysis of aimingmovements. Brain 1999;122:1169–82
14. Foundas A, Maccauley B, Raymer A, Maher L, Rothi L,Heilman K: Ideomotor apraxia in Alzheimer’s disease andleft hemisphere stroke: limb transitive and intransitivemovements. Neuropsychiatry Neuropsychol Behav Neurol1999;12:161–6
15. Poizner H, Clark M, Merians A, Macauley B, GonzalezRothi L, Heilman K: Joint coordination deficits in limbapraxia. Brain 1995;118:227–42
16. Poizner H, Mack L, Verfaellie M, Rothi LJG, Heilman KM:Three-dimensional computergraphic analysis of apraxia:neural representations of learned movement. Brain 1990;113:85–101
17. Goldenberg G, Hagmann S: The meaning of meaninglessgetures: a study of visuo-imitative apraxia. Neuropsycho-logia 1997;35:333–41
18. Weiss PH, Dohle C, Binkofski F, Schnitzler A, Freund HJ,Hefter H: Motor impairment in patients with parietallesions: disturbances of meaningless arm movement se-quences. Neuropsychologia 2001;39:397–405
19. Rumiati R, Weiss P, Tessari A, et al: Common and differ-ential neural mechanisms supporting imitation of mean-ingful and meaningless actions. J Cogn Neurosci 2005;17:1420–31
20. DeRenzi E, Faglioni P, Lodesani M, Vecchi A: Performanceof left brain-damaged patients on imitation of single move-
158 Buxbaum et al. Am. J. Phys. Med. Rehabil. ● Vol. 87, No. 2
ments and motor sequences. Frontal and parietal-injuredpatients compared. Cortex 1983;19:333–43
21. Harrington DL, Haaland KY: Motor sequencing with lefthemisphere damage: are some cognitive deficits specific tolimb apraxia? Brain 1992;115:857–74
22. Buxbaum LJ, Saffran EM: Knowledge of object manipula-tion and object function: dissociations in apraxic and non-apraxic subjects. Brain Lang 2002;82:179–99
23. Spatt J, Bak T, Bozeat S, Patterson K, Hodges J: Apraxia,mechanical problem solving and semantic knowledge:contibutions to object usage in corticobasal degeneration.J Neurol 2002;249:601–8
24. Clark M, Merians AS, Kothari A, et al: Spatial planningdeficits in limb apraxia. Brain 1994;117:1093–1106
25. Buxbaum LJ, Johnson-Frey SH, Bartlett-Williams M: De-ficient internal models for planning hand-object interac-tions in apraxia. Neuropsychologia 2005;43:917–29
26. Jax SA, Buxbaum LJ, Moll AD: Deficits in movement plan-ning and intrinsic coordinate control in ideomotorapraxia. J Cogn Neurosci 2006;18:2063–76
27. Gonzalez Rothi LJ, Heilman KM: Ideomotor apraxia: ges-tural learning and memory, in Roy EA (ed): Neuropsycho-logical Studies in Apraxia and Related Disorders. NewYork, Oxford University Press, 1985, pp 65–74
28. Faglioni P, Basso A, Botti C, Aglioti S, Saetti C: Gesturallearning in apraxia, in Jeannerod M (ed): Attention andPerformance XIII. Hillsdale, Lawrence Erlbaum, 1990, pp837–56
29. Buxbaum LJ: Ideomotor apraxia: a call to action. Neuro-case 2001;7:445–58
30. Buxbaum LJ, Giovannetti T, Libon D: The role of thedynamic body schema in praxis: evidence from primaryprogressive apraxia. Brain Cogn 2000;44:166–91
31. Buxbaum LJ, Kyle KM, Menon R: On beyond mirrorneurons: internal representations subserving imitationand recognition of skilled object-related actions in hu-mans. Brain Res Cogn Brain Res 2005;25:226–39
32. Buxbaum LJ, Kyle K, Grossman M, Coslett HB: Left infe-rior parietal representations for skilled hand-objectinteractions: evidence from stroke and corticobasal degen-eration. Cortex 2007;43:411–23
34. Kimura D: Translations from Liepmann’s Essays onApraxia. London, University of Western Ontario, 1980
35. Poeck K: The clinical examination for motor apraxia. Neu-ropsychologia 1986;24:129–34
35a. De Renzi E: Methods of limb apraxia examination and theirbearing on the interpretation of the disorder, in Roy EA(ed): Neuropsychological Studies of Apraxia and RelatedDisorders, North-Holland, Elsevier Science Publishers,1985, p 45
36. Foundas AL, Macauley BL, Raymer AM, Maher LM, Hei-lman KM, Roth LJG: Ecological implications of limbapraxia: evidence from mealtime behavior. J Int Neuropsy-chol Soc 1995;1:62–6
37. Goldenberg G, Daumuller M, Hagmann S: Assessment andtherapy of complex activities of daily living in apraxia.Neuropsychol Rehabil 2001;11:147–69
38. Sunderland A, Bowers MP, Sluman SM, Wilcock DJ,Ardron ME: Impaired dexterity of the ipsilateral hand afterstroke and the relationship to cognitive deficit. Stroke1999;30:949–55
39. Sunderland A, Sluman SM: Ideomotor apraxia, visuomotorcontrol and the explicit representation of posture. Neuro-psychologia 2000;38:923–34
40. van Heugten CM, Dekker J, Deelman BG, Stehmann-SarisJC, Kinebanian A: Rehabilitation of stroke patients withapraxia: the role of additional cognitive and motor impair-ments. Disabil Rehabil 2000;22:547–54
41. Wetter S, Poole JL, Haaland KY: Functional implicationsof ipsilesional motor deficits after unilateral stroke. ArchPhys Med Rehabil 2005;86:776–81
43. Donkervoort M, Dekker J, Deelman B: Sensitivity of dif-ferent ADL measures to apraxia and motor impairments.Clin Rehabil 2002;16:299–305
44. Hanna-Pladdy B, Heilman KM, Foundas AL: Ecologicalimplications of ideomotor apraxia: evidence from physicalactivities of daily living. Neurology 2003;60:487–90
45. Sundet K, Finset A, Reinvang I: Neuropsychological pre-dictors in stroke rehabilitation. J Clin Exp Neuropsychol1988;10:363–79
46. Goldenberg G, Hagmann S: Tool use and mechanical prob-lem solving in apraxia. Neuropsychologia 1998;36:581–9
47. Laimgruber K, Goldenberg G, Hermsdorfer J: Manual andhemispheric asymmetries in the execution of actual andpantomimed prehension. Neuropsychologia 2005;43:682–92
48. van Heugten CM, Dekker J, Deelman BG, van Dijk AJ,Stehmann-Saris FC, Kinebanian A: Measuring disabilitiesin stroke patients with apraxia: a validation study of anobservational method. Neuropsychol Rehabil 2000;10:401–14
49. Walker CM, Sunderland A, Sharma J, Walker MF: Theimpact of cognitive impairment on upper body dressingdifficulties after stroke: a video analysis of patterns ofrecovery. J Neurol Neurosurg Psychiatry 2004;75:43–8
50. Sunderland A: Recovery of ipsilateral dexterity after stroke.Stroke 2000;31:430–3
51. Goldenberg G, Hagmann S: Therapy of activities of dailyliving in patients with apraxia. Neuropsychol Rehabil1998;8:123–41
52. Walker MF, Lincoln NB: Factors influencing dressing per-formance after stroke. J Neurol Neurosurg Psychiatry1991;54:699–701
54. Schwartz MF, Montgomery MW, Buxbaum LJ, et al: Nat-uralistic action impairment in closed head injury. Neuro-psychology 1998;12:13–28
55. Schwartz MF, Buxbaum L, Montgomery MW, et al: Natu-ralistic action production following right hemispherestroke. Neuropsychologia 1999;37:51–66
56. Hartmann K, Goldenberg G, Daumuller M, Hermsdorfer J:It takes the whole brain to make a cup of coffee: theneuropsychology of naturalistic actions involving techni-cal devices. Neuropsychologia 2005;43:625–37
57. Mayer NH, Reed ES, Schwartz MF, Montgomery MW,Palmer C: Buttering a hot cup of hot coffee: an approachto the study of errors of action after brain damage, inCicerone K, Tupper D (eds): The Neuropsychology of Ev-eryday Life: Assessment and Basic Competencies. Boston,Kluwer Academic Publishers, 1990, pp 259–83
58. Sunderland A, Shinner C: Ideomotor apraxia and func-tional ability. Cortex 2007;43:359–67
59. Buxbaum LJ, Veramonti T, Schwartz MF: Function andmanipulation tool knowledge in apraxia: Knowing ‘whatfor’ but not ‘how’. Neurocase 2000;6:83–97
60. Maher L, Rothi L, Greenwald M: Treatment of gestureimpairment: a single case. Am Speech Hear Assoc 1991;33:195
61. Maher LM, Ochipa C: Management and treatment of limbapraxia, in Gonzalez-Rothi LJ, Heilman KM (eds): Apraxia:The Neuropsychology of Action. Hove, Psychology Press,1997, pp 75–91
63. Code C, Gaunt C: Treating severe speech and limb apraxiain a case of aphasia. Br J Disord Commun 1986;21:11–20
64. Deal JL, Florance CL: Modification of the eight-step con-tinuum for treatment of apraxia of speech in adults.J Speech Hear Disord 1978;43:89–95
65. Pilgrim E, Humphreys GW: Rehabilitation of a case ofideomotor apraxia, in Riddoch MJ, Humphrey GW (eds):Cognitive Neuropsychology and Cognitive Rehabilitation.Hillsdale, Lawrence Erlbaum Associates, 1994, pp 271–85
66. van Heugten CM, Dekker J, Deelman BG, van Dijk AJ,Stehmann-Saris JC, Kinebanian A: Outcome of strategytraining in stroke patients with apraxia: a phase II study.Clin Rehabil 1998;12:294–303
67. Donkervoort M, Dekker J, Stehmann-Saris FC, DeelmanBG: Efficacy of strategy training in left hemisphere strokepatients with apraxia: a randomized clinical trial. Neuro-psychol Rehabil 2001;11:549–66
68. Geusgens C, van Heugten C, Donkervoort M, van den EndeE, Jolles J van den Heuvel W: Transfer of training effects instroke patients with apraxia: an exploratory study. Neuro-psychol Rehabil 2006;16:213–29
69. Smania N, Girardi F, Domenicali C, Lora E, Aglioti S: Therehabilitation of limb apraxia: a study in left-brain-dam-aged patients. Arch Phys Med Rehabil 2000;81:379–88
70. Smania N, Aglioti SM, Girardi F, et al: Rehabilitation oflimb apraxia improves daily life activities in patients withstroke. Neurology 2006;67:2050–2
71. Robey RR, Schultz MC: A model for conducting clinical-outcome research: an adaptation of the standard protocolfor use in aphasiology. Aphasiology 1998;12:787–810
72. Nadeau SE, Rothi LJG, Rosenbek JC: Language rehabili-tation from a neural perspective, in Chapey R (ed): Lan-guage Intervention Strategies in Adult Aphasia. Balti-more, Lippincott Williams and Wilkins (in press)
73. Hallett M: Motor learning, in Freund HJ, Jeannerod M,Hallett M, Leiguarda R (eds): Higher-Order MotorDisorders: From Neuroanatomy and Neurobiology toClinical Neurology. Oxford, Oxford University Press, 2005,pp 123–40
74. Pascual-Leone A, Dang N, Cohen LG, Brasil-Neto JP, Cam-marota A, Hallett M: Modulation of muscle responsesevoked by transcranial magnetic stimulation during theacquisition of new fine motor skills. J Neurophysiol 1995;74:1037–45
75. Karni A, Meyer G, Jezzard P, Adams MM, Turner R, Un-gerleider LG: Functional MRI evidence for adult motorcortex plasticity during motor skill learning. Nature 1995;377:155–8
76. Hikosaka O, Nakamura K, Sakai K, Nakahara H: Centralmechanisms of motor skill learning. Curr Opin Neurobiol2002;12:217–22
78. Bapi RS, Miyapuram KP, Graydon FX, Doya K: fMRI in-vestigation of cortical and subcortical networks in thelearning of abstract and effector-specific representations ofmotor sequences. Neuroimage 2006;32:714–27
79. Graydon FX, Friston KJ, Thomas CG, Brooks VB, MenonRS: Learning-related fMRI activation associated with arotational visuo-motor transformation. Brain Res CognBrain Res 2005;22:373–83
80. Muller RA, Kleinhans N, Pierce K, Kemmotsu N,Courchesne E: Functional MRI of motor sequenceacquisition: effects of learning stage and performance.Brain Res Cogn Brain Res 2002;14:277–93
81. Toni I, Krams M, Turner R, Passingham RE: The timecourse of changes during motor sequence learning: awhole-brain fMRI study. Neuroimage 1998;8:50–61
82. Carey DP: Multisensory integration: attending to seen andfelt hands. Curr Biol 2000;10:R863–5
83. Fridman EA, Hanakawa T, Chung M, Hummel F, Lei-guarda RC, Cohen LG: Reorganization of the human ip-silesional premotor cortex after stroke. Brain. 2004;127:747–58
85. Winhuisen L, Thiel A, Schumacher B, et al: Role of thecontralateral inferior frontal gyrus in recovery of languagefunction in poststroke aphasia: a combined repetitivetranscranial magnetic stimulation and positron emissiontomography study. Stroke 2005;36:1759–63
86. Honda M, Deiber MP, Ibanez V, Pascual-Leone A, ZhuangP, Hallett M: Dynamic cortical involvement in implicit andexplicit motor sequence learning. A PET study. Brain1998;121:2159–73
87. Grafton ST, Hazeltine E, Irvy R: Functional mapping ofsequence learning in normal humans. J Cogn Neurosci1995;7:497–510
88. Catalan MJ, Honda M, Weeks RA, Cohen LG, Hallett M:The functional neuroanatomy of simple and complex se-quential finger movements: a PET study. Brain 1998;121:253–64
89. Wise SP, Shadmehr R: Motor control, in RamachandranVS (ed): Encyclopedia of the Human Brain, vol 3. SanDiego, Academic Press, 2002, pp 137–57
90. Krakauer JW, Ghilardi MF, Ghez C: Independent learningof internal models for kinematic and dynamic control ofreaching. Nat Neurosci 1999;2:1026–31
91. Keele SW: Motor control, in Brooks WB (ed): Handbook inPhysiology, Section 1: The Nervous System, vol 2. Balti-more, Williams & Williams, 1981, pp 1391–414
100. Bock O, Thomas M, Grigorova V: The effect of rest breakson human sensorimotor adaptation. Exp Brain Res 2005;163:258–60
101. Shea CH, Kohl RM: Composition of practice: influence onthe retention of motor skills. Res Q Exerc Sport 1991;62:187–95
102. Ollis S, Button C, Fairweather M: The influence of profes-sional expertise and task complexity upon the potency ofthe contextual interference effect. Acta Psychol (Amst)2005;118:229–44
160 Buxbaum et al. Am. J. Phys. Med. Rehabil. ● Vol. 87, No. 2
105. Park H, Arndt J, Reder LM: A contextual interferenceaccount of distinctiveness effects in recognition. MemCognit 2006;34:743–51
106. Plaut DC, McClelland JL, Seidenberg MS, Patterson K:Understanding normal and impaired word reading: com-putational principles in quasi-regular domains. PsycholRev 1996;103:56–115
107. Wulf G, Horger M, Shea CH: Benefits of blocked over serialfeedback on complex motor skill learning. J Mot Behav1999;31:95–103
108. Rossetti Y, Rode G, Pisella L, et al: Prism adaptation to arightward optical deviation rehabilitates left hemispatialneglect. Nature 1998;166–9
109. Patton JL, Stoykov ME, Kovic M, Mussa-Ivaldi FA: Evalu-ation of robotic training forces that either enhance orreduce error in chronic hemiparetic stroke survivors. ExpBrain Res 2006;168:368–83
110. Dancause N, Ptito A, Levin MF: Error correction strategiesfor motor behavior after unilateral brain damage: short-term motor learning processes. Neuropsychologia 2002;40:1313–23
111. Gandolfo F, Mussa-Ivaldi FA, Bizzi E: Motor learning byfield approximation. Proc Natl Acad Sci U S A 1996;93:3843–6
112. Rossetti Y: Implicit short-lived motor representations ofspace in brain damaged and healthy subjects. ConsciousCogn 1998;7:520–58
113. Heilman KM, Gonzalez-Rothi LJ: Apraxia, in Heilman KM,Valenstein E (eds): Clinical Neurophysiology. New York, Ox-ford University Press, 2003, pp 215–35
114. Holden MK: Virtual environments for motor rehabilitation:review. Cyberpsychol Behav 2005;8:187–211
115. Kahn L, Rymer W, Reinkensmeyer D: Adaptive assistancefor guided force training in chronic stroke. Conf ProcIEEE Eng Med Biol Soc 2004;4:2722–5
116. Masiero S, Celia A, Rosati G, Armani M: Robotic-assistedrehabilitation of the upper limb after acute stroke. ArchPhys Med Rehabil 2007;88:142–9
117. Hanlon RE: Motor learning following unilateral stroke.Arch Phys Med Rehabil 1996;77:811–5
118. Barde L, Buxbaum L, Moll A: Abnormal reliance on objectstructure in apraxics’ learning of novel object-related ac-tions. J Int Neuropsychol Soc 2007;13:997–1008
119. Giovannetti T, Schwartz MF, Buxbaum LJ: The effect ofdivided attention on action slips: a new error elicitationparadigm. J Clin Exp Neuropsychol 2007;29:690–705
