Cognitive Brain Research 25 (2005) 226 239

www.elsevier.com/locate/cogbrainres

Research Report

On beyond mirror neurons: Internal representations subserving imitation

and recognition of skilled object-related actions in humans
Laurel J. Buxbaum a,b,*, Kathleen M. Kyle a, Rukmini Menon a
a

Moss Rehabilitation Research Institute, Korman 213, 1200 W. Tabor Road, Philadelphia, PA 19141, USA
b
Thomas Jefferson University, Philadelphia, PA 19107, USA
Accepted 27 May 2005
Available online 5 July 2005

Abstract
A considerable recent literature argues that the same representations, encoded by inferior prefrontal and parietal cells known as mirror
neurons, may be activated in both production and recognition of object-related actions. Here, we test several predictions derived from the
contemporary literature on the parity between production and recognition and the putative emergence of the mirror neuron system from a system
coding hand object interactions. Forty-four patients with left-hemisphere stroke, 21 of whom exhibited ideomotor apraxia, performed a number
of pantomime imitation and recognition tasks, and performance was scored with respect to hand posture, arm posture, amplitude, and timing.
Consistent with predictions, there were strong relationships between object-related pantomime imitation and object-related pantomime
recognition, and between imitation and recognition of the hand posture component of object-related actions. Skilled object-related gesture
representations are likely to be closely tied to evolutionarily more primitive systems controlling object grasping, to emerge from a mapping
between object and action information coded by ventral and dorsal streams, and to be lateralized to the left hemisphere in humans.
D 2005 Elsevier B.V. All rights reserved.
Theme: Neural basis of behavior
Topic: Cognition
Keywords: Apraxia; Action; Gesture; Imitation; Recognition

1. Introduction
According to the direct matching hypothesis, actions
performed by others are recognized by activating the same
spatiomotor representations used for performing the action
oneself. Numerous recent investigations in infant development (e.g., [37]) and adult cognitive psychology (e.g.,
[4,47]) suggest that there is a common coding between
perception and action. A possible physiological foundation
for at least some aspects of this common coding is provided
by the recent discovery of so-called mirror neurons in the
* Corresponding author. Moss Rehabilitation Research Institute, Korman
213, 1200 W. Tabor Road, Philadelphia, PA 19141, USA. Fax: +1 215 456
5926.
E-mail address: Lbuxbaum@einstein.edu (L.J. Buxbaum).
0926-6410/$ - see front matter D 2005 Elsevier B.V. All rights reserved.
doi:10.1016/j.cogbrainres.2005.05.014

inferior prefrontal cortex (area F5, the putative homologue

of human Brocas area) and inferior parietal lobule (area PF)
in the monkey. These cell units discharge both when the
monkey produces an object-related action and when a
comparable action is performed by an experimenter. The
neurons respond best to specific types of prehensile actions
upon objects (e.g., grasping), and are silent when a hand
alone or object alone is viewed [20,49].
Functional neuroimaging evidence suggests that inferior
prefrontal cortex may be involved in both action observation
and production in humans as well as in monkeys. Observation and execution of grasping movements or simple
finger movements has been shown to activate Brodmann
area (BA) 44 or 45 in a number of PET, fMRI, and MEG
studies [3,22,28,30,40]. These findings are consistent with
the possibility that human prefrontal cortex contains a

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

mirror neuron system similar to that described in the

monkey [20].
In addition to simple object prehension and simple finger
movements, humans produce and recognize a considerable
repertoire of skilled object-related (so-called transitive)
movements and pantomimes. Unlike simple grasping based
on object structure, recognition and production of learned
gestures and pantomimes entail a declarative semantic
component [32]. Skilled gestures also have strong requirements for decoding the particular spatial configuration of the
hand and fingers that distinguishes one skilled gesture from
another. Thus, the evidence that production and observation
of simple grasping and finger movements involve the same
or overlapping neural structures does not compel the
conclusion that imitation and recognition of complex,
skilled, meaningful behavior share the same underlying
cognitive representations. Additional support for the latter
hypothesis must be gleaned from studying the relationship
of imitation and recognition of complex skilled actions. An
important source for such data may be found in patients with
ideomotor apraxia (IM).
Individuals with IM are deficient in producing transitive
(familiar object-related) gesture in gesture pantomime (to
command or sight of object) and gesture imitation tasks.
Errors may be postural and/or may involved deficits in
amplitude and timing [44 46]. Deficits persist with actual
object use but are more subtle [44], presumably in part
because of the feedback from object structure helps to
constrain degrees of freedom of the movement (see [8]). IM
occurs in nearly 60% of left-hemisphere cerebral vascular
accident patients (LCVA) [1], and involves the non-paretic
left hand of approximately 50% of these patients [34]. In
contrast to their deficits in pantomiming or imitating familiar
object-related actions, IM patients are frequently relatively
unimpaired in the production or imitation of intransitive (nonobject related, symbolic) gestures such as waving goodbye,
signaling stop, or beckoning come here [38].
Evidence from several laboratories, including our own,
indicates that patients with IM may have particular
difficulties producing and recognizing the hand postures
appropriate for skilled object-related actions. In fact, there is
growing evidence that representations for skilled, objectrelated hand posture may be particularly vulnerable to loss
in IM, and encoded distinctly from on line programming
of hand posture for prehensile manipulations of objects.
Several reports indicate that apraxics impaired hand posture
may be in contrast to their substantially better arm posture
and trajectory for skilled actions [55], particularly in
pantomime tasks [44]. Hand posture errors are the most
frequent error type of IM patients on skilled sequencing
tasks [56] and pantomimed prehension tasks [33]. IM
patients are deficient in recognizing the appropriate hand
posture for interacting with familiar objects, but perform
normally in selecting hand postures for prehensile interactions with novel objects in response to their structure [7]
and relatively normally in reaching to and grasping geo-

227

metric three-dimensional shapes [8]. Patients exhibiting this

pattern often have lesions involving the left inferior parietal
lobe (IPL) and intraparietal sulcus (IPS). Critically, as will
be discussed further below, there is also evidence that the
hand posture component of skilled-object related pantomime is not deficient in these patients simply because it is
more difficult than other gesture components [9], or more
difficult than on-line reaching and grasping. For example,
the observed pattern of deficient skilled object-related
pantomime in the face of preserved object grasping doubly
dissociates from the pattern observed in patients with optic
ataxia due to superior parietal damage, who frequently
exhibit deficient grasping of visually-presented objects with
spared pantomime [43], and, in at least some cases, spared
ability to shape the hand when required to grasp familiar,
meaningful objects (cf. patient AT, [29]).
IM has been characterized variously as a disorder of
learned skilled movements [26], a difficulty making voluntary gestures [35], an impairment in gesturing to command
[58], and a deficit in the imitation of meaningless movements
[12]. We have suggested that these characterizations may not
capture some of the key features of the disorder [5]. The
marked disparity between performance of transitive and
intransitive actions across a range of production (e.g., to
command, to sight of object) and imitation tasks, and the
relative preservation of on line aspects of motor control
responsive to the structural features of objects, has influenced
our characterization of IM as a deficit in stored representations of the position and movements of the limb (and
particularly, the hand) subserving skilled object-related
actions. Further, these representations are likely to be
mediated by the left inferior parietal lobe [42].
The hypothesis that the inferior parietal lobe contains
gesture representations critical both for production and
recognition of actions with objects predicts that the two
should be impaired in parallel in patients with IM due to
parietal damage. Consistent with this [27], left parietal lesions
have been associated with both production and recognition
deficits, and frontal lesions only with production impairments. Note that this is potentially inconsistent with the
notion of a mirror neuron system in the frontal lobe that is
critical for gesture recognition. Several more recent studies
have suggested more generically that IM patients are
impaired in the recognition of gesture (e.g., [14,61]). On
the other hand, other recent studies have failed to find an
association between gesture production and recognition
[15,24,31]. It should be noted that these studies have used
small samples ranging from 1 to 14 patients, and it remains
possible that there was insufficient power to detect an
association.1
1
One exception is a study by Bell [2] with 38 subjects; however, the
pantomime recognition task used in that study required subjects to
understand the association between an observed gesture and an associated
object, and it is possible that this requires more extensive semantic
knowledge than that tapped by gesture recognition alone.

228

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

We derive a number of predictions from the hypothesis

that IM reflects deficient inferior parietal representations
subserving the production, imitation, and recognition of
skilled object-related gesture pantomime and skilled hand
object interactions, six of which we test here. The first
prediction is that patients with IM should be disproportionately impaired in the imitation of transitive as compared to
intransitive gestures. The second is that they should be
particularly impaired in the hand posture component of
gesture imitation. The third is that we should observe an
association between performance of transitive imitation and
transitive recognition tasks. The fourth is that there should
be a similar and more specific association between
imitation and recognition of the hand posture component
of transitive actions. The fifth is that the association
between transitive recognition and intransitive gesture
imitation should be considerably weaker. The sixth is that
patients with deficient hand posture production and
recognition should have lesions that include the left inferior
parietal lobe. In the study that follows, we test these
predictions with a group of 44 left-hemisphere stroke
patients. We then discuss the implications of the data for
informing the likely characteristics of skilled object-related
action representations.

2. Subjects
Forty-four left-hemisphere stroke patients participated in
the study. All patients had suffered a single left-hemisphere
cerebral vascular accident; one subject had an additional
small asymptomatic right occipital infarct pre-dating the
left-hemisphere stroke. All subjects gave informed consent
to participate in accordance with guidelines of Albert
Einstein Healthcare Network and were paid for their
participation. Subjects were referred to the study from a
large database of potential research subjects in the Philadelphia area maintained by Moss Rehabilitation Research
Institute. Subjects were excluded if database records
indicated language comprehension deficits of sufficient
severity to preclude comprehension of task instructions.
Subjects over the age of 80 and/or with histories of comorbid neurologic disorders, alcohol or drug abuse, or
psychosis were also excluded. All subjects gave informed
consent to participate in accordance with the guidelines of
the IRB of Albert Einstein Healthcare Network, and were
paid for their participation.
We pursued two complimentary strategies in analyzing
the data. The first strategy was to classify participants as
apraxic or not, and assess whether performance of the two
groups differed on measures of interest. The second strategy
was to treat data as continuous variables to assess the
strength of relationships of scores on the measures of
interest.
Subjects were characterized as exhibiting ideomotor
apraxia (hereafter, IM) or not (hereafter, LCVA) on the

basis of the average of scores on (1) novel gesture imitation

and (2) gesture to sight of objects.2 Appendix A provides
details of the administration and scoring of the meaningless
gesture imitation and gesture to command tasks, as well as
transitive and intransitive gesture imitation tasks to be
described below.
Table 1 shows subject demographics as well as scores on
the IM apraxia composite score. Scores on the two tests
comprising the composite score (gesture to sight of object
and meaningless gesture imitation) were highly correlated
(r = 0.74, P < 0.0001). Subjects were characterized as
exhibiting IM if they obtained a composite score more than
2 standard deviations below the mean of age matched righthanded control subjects who performed the tasks with their
left hands (control n = 10; 5 females; mean age 64.7, range
43 77; mean education 14 years, range 10 18 years; mean
score = 92.5, SD = 4.5, range = 85 98). Subjects
characterized as IM (n = 21) and LCVA (n = 23) by this
criterion did not differ in age (t = 0.50, P = 0.70), education
(t = 0.65, P = 0.63), or months elapsed between lesion and
test (t = 0.82, P = 0.25).

3. Language comprehension
Participants performed the Comprehension subtest of the
Western Aphasia Battery (Kertesz Ref). Maximum possible
score was 200 points. Apraxics exhibited somewhat greater
deficits in comprehension (mean = 166, range = 117 196)
than did LCVA (mean = 187, range = 111 200), t(42) = 3.1,
P < 0.01.

4. Lesion analysis
Clinical T-1 or T-2-weighted MRI scans were available
for 36 of the 44 subjects (18 LCVA and 18 IM). Lesions
were segmented and interpreted by an experienced neurologist. Subtractions of the lesioned regions of the IM versus
LCVA groups were performed by one of the authors using
the MRIcro image analysis program developed by Dr. Chris
Rorden (see http://www.psychology.nottingham.ac.uk/staff/
cr1/mricro.html). Fig. 1 shows the result of this subtraction

The use of a composite of novel gesture imitation and gesture to sight of

objects to define apraxia was based on several considerations. First, the
study reported here is one of several studies ongoing in our laboratory
assessing various aspects of IM, and the designation of IM in these related
studies is inclusive, as it is designed to detect IM due either to loss of
knowledge of the gestures associated with objects and/or deficits in praxis
production. Second, although these aspects of praxis can dissociate, in the
great majority of patients they tend to co-occur (see text for evidence of
strong correlation in the present study). Third, as described earlier, there is
considerable disagreement in the literature regarding the most appropriate
test for IM. The combined score captures patients who would be defined as
having IM by many published criteria.

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

229

Table 1
Subject demographics and scores on gesture production and imitation
Subject Group
A1
A2
A3
A4
A5
A6
A7
A8
A9
A10
A11
A12
A13
A14
A15
A16
A17
A18
A19
A20
A21
L1
L2
L3
L4
L5
L6
L7
L8
L9
L10
L11
L12
L13
L14
L15
L16
L17
L18
L19
L20
L21
L22
L23

Gesture to sight of object Meaningless imitation Gesture composite Age Education Gender Handedness Lesion volume (cm3)

IM
64
IM
75
IM
60
IM
45
IM
75
IM
55
IM
38
IM
73
IM
63
IM
65
IM
63
IM
75
IM
70
IM
50
IM
65
IM
60
IM
55
IM
75
IM
50
IM
65
IM
58
LCVA 90
LCVA 85
LCVA 88
LCVA 88
LCVA 90
LCVA 90
LCVA 97
LCVA 85
LCVA 93
LCVA 88
LCVA 88
LCVA 90
LCVA 90
LCVA 88
LCVA 95
LCVA 100
LCVA 88
LCVA 93
LCVA 83
LCVA 93
LCVA 78
LCVA 85
LCVA 90

60
58
40
73
75
25
50
58
58
75
63
58
68
78
58
50
90
58
53
50
65
90
93
90
93
93
93
85
90
93
93
95
93
95
100
95
93
93
93
90
88
95
85
98

62
66
50
59
75
40
44
65
60
70
63
66
69
64
61
55
73
66
51
58
61
90
89
89
90
91
91
91
88
93
90
91
91
93
94
95
96
90
93
86
90
86
85
94

55
49
79
79
50
59
63
67
56
79
64
50
49
42
60
78
42
67
39
58
41
35
56
51
64
77
42
55
51
58
58
50
65
77
50
51
80
55
56
54
42
54
40
64

18
16
12
11
12
16
12
12
14
16
12
10
8
16
12
11
16
14
10
10
10
12
20
14
12
12
15
12
12
12
3
12
19
16
18
8
12
16
12
12
8
18
16
16

F
F
F
F
M
M
F
F
M
F
M
F
M
F
M
M
F
M
M
F
F
F
M
F
M
M
F
M
F
M
M
M
M
F
F
M
F
F
M
F
M
M
F
M

R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
R
L
R
R
R
R
R
L
R
R

208.4
110.5
58.2
8.5
64.3
68.7
253.5
7.2
44.9
NA
NA
151.9
49.0
143.6
161.5
51.1
131.3
119.7
180.8
45.3
46.1
41.8
96.8
56.7
77.4
NA
16.8
69.3
41.4
58.6
0.5
NA
61.6
13.6
31.2
28.0
0.4
18.4
22.4
NA
178.0
95.1
142.2
25.6

analysis. The lesion loci in the IM patients is consistent with

previous reports (e.g., [23]).

transitive gestures. In addition to the IM and LCVA patients,

also tested were the same 10 age-matched healthy control
subjects described earlier.

5. Experimental tasks

5.1.1. Methods
Participants watched videotapes of an examiner performing 10 transitive and 5 intransitive pantomimes with the
right hand, and were asked to imitate the gesture as precisely
as possible with the unimpaired left hand. Transitive
gestures were hammering, cutting with scissors, sawing,
using a screwdriver, writing with a pencil, using a comb,
winding a watch, brushing teeth, flipping a coin, and eating
with a fork. Intransitive gestures were saluting, waving
goodbye, hitch-hiking, signaling stop, and beckoning

5.1. Study 1: imitation of transitive and intransitive gesture

In the first study, we sought to replicate previously
reported findings that patients with IM are more likely to
have difficulties with imitation of transitive as compared to
intransitive gesture pantomimes. We also assessed the
prediction that IM patients would have disproportionate
difficulty in producing the hand posture component of

230

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

Fig. 1. Subtractions of lesioned regions of IM versus LCVA groups. These indicate the difference in the proportion of patients in the two groups having
involvement in a given region. Colors further to right on the color bar indicate relatively more lesion in IM group in increments of 20%. Left and right are
reversed. There were no regions uniquely damaged in the IM group; however, there were several regions in which lesions were more likely. These include
Brodmann areas 6 and 44 (dorsolateral frontal), 22 and 37 (temporal), and 39 and 40 (inferior parietal).

come here. Participants were permitted to begin imitation

while watching the videos. Gestures were scored according
to the detailed error taxonomy described in Buxbaum,
Giovannetti, and Libon [6] (and see [8,9]) and detailed in
Appendix A.3

Four Kruskal Wallis ANOVAs were performed, one with

each gesture component; therefore, a Bonferroni-corrected
P value of .0125 was required for significance. Again, we
tested for between-group differences in transitive versus

5.1.2. Results
Scores are reported here in terms of percent correct: IM
transitive mean = 63%, intransitive mean = 89%, mean
transitive intransitive difference = 25.6%; LCVA transitive mean = 93%, intransitive mean = 99%, mean
difference = 4.7%; CTL transitive mean = 93%, intransitive
mean = 100%, mean difference = 5.8%. In this and all
subsequent analyses, the difference between transitive and
intransitive performance was calculated for each subject,
and between-group comparisons of the difference scores
were performed with Kruskal Wallis nonparametric oneway ANOVAs. Post hoc testing was performed with Mann
Whitney tests. There was an effect of group, H = 28.36, P <
0.0001, indicating that the transitive intransitive difference
scores were greater for some group(s) than others. Post hoc
testing with the critical value of P set at .016 (i.e., 0.05/3)
confirmed that the difference scores were significantly
greater for the IM patients than for the other two groups
(Ps < 0.0001), who did not differ from one another (P >
0.36). These data replicate previous reports that objectrelated gestures are more difficult for IM patients than are
symbolic gestures.
In the next analysis, we assessed whether IM patients
were equally impaired in all components of transitive
gesture. Fig. 2 shows the data entered into these analyses.
3
To assess scoring reliability, gestures for 6 of the participants were
scored by 2 independent coders. Percent agreement between the coders
ranged from 78% to 100% across the 6 subjects (mean 88% agreement;
Cohens kappa = 0.60).

Fig. 2. Performance of IM, LCVA, and CTL groups in imitation of

intransitive and transitive gesture pantomimes, scored for hand posture
(HP), arm posture (AP), amplitude (AMP), and timing (TIM) components.

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

intransitive performance. There were significant betweengroup differences in hand posture, H = 35.2, P < 0.0001;
arm posture, H = 10.5, P = 0.003; and amplitude, H = 12.3,
P = 0.002. Post hoc testing of the 3 significant ANOVAs
with Mann Whitney tests using a Bonferroni-corrected P
value of .0055 (i.e., 0.05/9) indicated that there were
significant differences between the IM patients and the
other two groups for hand posture and amplitude (P <
0.0001), and a difference between IM and LCVA for arm
posture ( P = 0.004). None of the differences between LCVA
and CTL were significant (all Ps > 0.27). Thus, the
disproportionate disparity between transitive and intransitive gesture in the IM group was reliably observed in several
gesture components.
In a final analysis, we addressed the concern that the
disproportionate impairment in transitive as compared to
intransitive hand postures in the IM group may be related
to the greater complexity of the former. We examined the
data from the IM and LCVA groups for a subset of 5
transitive gestures having a simple, stable hand posture
denoting grasp of a tool (hammering, sawing, combing
hair, brushing teeth, and eating with a fork) and all 5 of the
intransitive gestures examined previously (waving goodbye, beckoning come here, hitch-hiking, signaling stop,
and saluting). Within-group comparisons were performed
with Wilcoxon Signed Ranks Tests (with a Bonferronicorrected P value of .008, i.e., .05/6, required for significance). For transitive movements, IM patients hand
posture (mean 50.4% correct) tended to be more deficient
than arm posture (66.7%; P = 0.02), amplitude (69%; P =
0.01), or timing (82%; P = 0.009). For intransitive
imitation, IM patients hand posture (mean 91% correct)
tended to be better than arm posture (84%; P = 0.01), and
was equal to amplitude (91%) and timing (89.5%). In
between-group comparisons, Mann Whitney U tests (with
a Bonferroni-corrected P value of 0.016 required for
significance) confirmed that the disparity between transitive
and intransitive hand postures was more pronounced for the
IM patients than the other two groups (Ps < 0.008), who
did not differ from one another (P = 0.5).
5.1.3. Discussion
In this study, we replicated previous findings indicating
that transitive gesture imitation is more impaired in patients
with IM than is intransitive imitation. Moreover, consistent
with the possibility that the system that is damaged in IM
patients is specialized for hand object interactions, the hand
posture component of gesture imitation for transitive
(object-related) gestures proved to be the most impaired
aspect of IM patients performance.
One possible objection to the proposed interpretation is
that transitive hand postures may simply be more difficult
than other components of transitive gesture, and more
difficult than intransitive hand postures, and thus more
sensitive to any type of impairment. There are two lines of
evidence against this interpretation. First, when we com-

231

pared IM and LCVA patients performance in imitation of

intransitive hand postures and transitive postures having a
simple, stable grasp configuration, the intransitive hand
postures were still strikingly superior. Another line of
evidence comes from data we have reported from 4 patients
with corticobasal degeneration (CBD), a degenerative
disorder affecting primarily the superior parietal lobes (areas
5 and 7 bilaterally) in early stages of disease progression [9].
The disorder has been known as primary progressive
apraxia because of its devastating effects on action
production. Although the CBD patients were more impaired
than the IM patients overall, they were less impaired in the
hand posture component of transitive gesture imitation
(mean 88% correct) than arm posture (46%), amplitude
(54%), or timing (58%). The data are consistent across
subjects: hand posture was the least impaired gesture
component in all 4 CBD subjects. Additionally, unlike the
patients with IM due to stroke reported here, the transitive
hand postures of the CBD patients (mean 88% correct) were
slightly better than their intransitive hand postures (mean
80% correct). These data strongly suggest that transitive
hand posture is not simply more sensitive to brain damage
than other aspects of gesture imitation. The system damaged
in the IM patients appears to be particularly strongly
involved in coding information about the position of the
hand for object-related gestures.
It is also of note that Mozaz et al. [38] recently argued
that the frequently observed difference between transitive
and intransitive gestures cannot be reduced to differences in
movement complexity. Healthy subjects were asked to
produce both gesture types as well as to discriminate static
photographs of transitive and intransitive gestures. They
performed more poorly with transitive gestures, both on the
production and picture discrimination tasks. The investigators argued that the difficulty with the latter is not likely
attributable to the differential complexity of transitive versus
intransitive movements, but instead reflects differences in
the underlying representations of the two gesture types.
One additional point of interest in the present study is that
the deficit in transitive gestures, and in transitive hand posture
in particular, was observed on an imitation task. According to
2-route models of gesture production, gesture imitation can
be accomplished via a direct route that bypasses representational knowledge, but permits calculation of the current
position of the actors body parts in space, and transformation
of these coordinates into a body-centered system of coordinates appropriate for the observers action [5,21]. Presumably, such a route would be engaged regardless of whether a
gesture was meaningful or not, and transitive or not. The use
of such a direct route would not explain the difference
between transitive and intransitive gestures, unless again the
former were simply harder in terms of spatiomotor transcoding, and as we have discussed, the data from the CBD
patients speak against this possibility. Instead, these data
suggest that a representational (or indirect) system that is
sensitive to gesture transitivity is recruited for imitation

232

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

performance, and may be obligatorily involved in meaningful

imitation tasks (see [57]). We have previously reported data
from a large sample of left-hemisphere stroke patients
supporting this conclusion [54].
If IM patients are deficient in a system that represents
transitive actions and transitive hand postures, then we
would expect them to also be impaired in their recognition
of transitive gestures and, in particular, in recognition of the
hand posture component of transitive gestures. We assessed
this prediction in the next study.
5.2. Study 2: recognition of transitive gesture
5.2.1. Methods
In this study, participants performed a forced-choice
gesture recognition task modeled loosely after one reported
by Heilman et al. [27] requiring them to discriminate correct
from incorrect gesture clips presented on a computer
monitor. There were two conditions. In the Semantic
condition, the participant heard the name of a gesture (a
verb) repeated twice, and simultaneously saw the verb
presented on the computer monitor (e.g., hammering) for
a duration of 5 s. After a 2-s pause, they heard the letter A
spoken aloud, and then saw two repetitions of a videotaped
examiner performing a gesture. After an additional 2-s
pause, they heard the letter B spoken, followed by two
repetitions of a second gesture. One gesture of each pair was
the correct match to the verb, and the other was incorrect by
virtue of a semantic relationship between the target (e.g.,
hammering) and foil (e.g., sawing). Order of the target and
foil within the trial was randomized. On each trial, the
subject selected the correct gesture by verbalizing or
pointing to the appropriate letter (A or B) displayed
on a piece of paper on the table. There were 24 Semantic
trials.
In the Spatial condition, methods were identical, except
that the foil was incorrect by virtue of an error in the hand
posture, arm posture, or amplitude/timing components. For
example, the foil for the sawing trial consisted of a
correctly bent arm position and characteristic forward-andback arm movement paired with an incorrect clawed,
splay-fingered hand posture. There were 24 Spatial trials, 8
each with hand posture, arm posture, and amplitude/timing
foils.
5.2.2. Results
In the Semantic condition, IM patients performed more
poorly than the other groups (IM mean = 20, SD = 3.2;
LCVA mean = 23.1, SD = 0.95, CTL mean = 23.4, SD =
0.71). Data were subjected to a repeated measures ANOVA
with group as a between-subjects factor. There was a
significant effect of group, F(2,51) = 14.5, P < 0.0001. IM
patients were more impaired in discriminating semantic foils
from correct gestures than either LCVA or CTL groups (P <
0.0001 for both). The latter two groups did not differ from
one another (P = 0.7).

Data from the Spatial condition of the gesture recognition

task are shown in Fig. 3. It can be seen that IM patients are
more impaired in recognition of the hand posture component of gestures than in the other components.
The data from the Spatial recognition condition were
subjected to a repeated measures ANOVA with group as a
between-subjects factor and gesture component (hand
posture, arm posture, amplitude/timing) as a within-subjects
factor. There was a significant main effect of group,
F(2,51) = 28.1, P < 0.0001, and of gesture component,
F(2,102) = 18.1, P < 0.0001, and a significant interaction
of group  gesture component, F(4,102) = 6.6, P <
0.0001. Post hoc between-groups comparisons with Fishers PLSD tests revealed that apraxics were more impaired
than the other two groups for recognition of all 3 gesture
components (all Ps < 0.01). Within-group comparisons
revealed that for IM patients, hand posture recognition was
more impaired than arm posture recognition, t = 7.5, P <
0.0001, and more impaired than amplitude/timing recognition, t = 5.6, P < 0.0001. For the other 2 groups, none
of the comparisons between components were significant
(all Ps > 0.09).
Few previous studies of gesture recognition in aphasic
patients have controlled for level of aphasia severity. Given
that the gesture recognition task has a language comprehension component (i.e., targets were identified based on the
presentation of verbs), one concern is that the IM patients,
who have poorer comprehension than LCVA patients, may
have misunderstood the intended targets more frequently than
the other groups. To assess this, we performed an ANCOVA
on the Semantic Gesture Recognition data with WAB
comprehension as a covariate. There was a small but significant Group  Comprehension interaction, F(1,40) = 5.6,
P < 0.05, indicating that we cannot rule out a contribution
of impaired language comprehension to IM performance in
the Semantic Gesture Recognition condition.
A similar ANCOVA was performed on the Spatial
gesture recognition data from the IM and LCVA patients,
with group as a between-subjects factor, gesture score with

Fig. 3. Performance of IM, LCVA, and CTL groups in forced-choice

selection of transitive gestures when foil gesture was incorrect by virtue of
hand posture (HP), arm posture (AP), or amplitude/timing (AMP/TIM).

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

its three component levels (hand posture, arm posture, and

amplitude/timing) as a repeated within-subjects factor, and
WAB comprehension score entered as a covariate. As
expected, there was a group  WAB score interaction,
F(1,120) = 12.5, P < 0.001, deriving from the fact that the
IM patients had greater comprehension impairment than the
other groups. There was also a small but significant
interaction of gesture score and WAB score, F(2,120) =
3.3, P = 0.04, indicating that participants with poorer
language comprehension tended to perform more poorly on
the spatial gesture recognition task. If language comprehension influenced the observed disproportionate impairment of IM patients in hand posture recognition, then the
WAB covariate should interact with gesture component. It
did not [ F(2,120) = 2.0, P > 0.1]. Thus, taking into account
the poorer comprehension of the IM group does not explain
their particularly strong predilection toward errors in hand
posture recognition.
5.2.3. Discussion
Although previous literature led us to expect that IM
patients would be impaired in both semantic and spatial
aspects of gesture recognition ([19,27,59], the findings with
regard to semantic gesture recognition were not strongly
consistent with this expectation. Although IM patients were
indeed more impaired than non-IM patients in matching
auditory and spoken words to gestures presented with
semantic foils, that difference disappeared when we
controlled for comprehension severity. This may be an
artifact of the recognition test requirements, which had a
strong language comprehension component. Another possibility is that the same underlying semantic processes are
required both for the semantic aspects of gesture recognition
and language comprehension. We are currently constructing
a pre-test to assess comprehension of the test items that will
help to distinguish these possibilities.
The results from the spatial version of the gesture
recognition task indicate that not all aspects of IM patients
recognition problems are confounded with language impairment. IM patients were significantly worse than non-apraxics
in distinguishing correct gestures from spatially-similar foils,
and the difference between groups in recognition of the hand
posture component of the gesture persisted even when
comprehension was controlled. In the remaining analyses,
therefore, we focus on the spatial aspects of gesture recognition and in particular the hand posture component, and
the relationship of spatial gesture recognition to gesture
production.
5.3. Additional analyses of the relationship of production
and spatial gesture recognition
On the hypothesis that the same representations subserve
production and recognition of transitive (but not intransitive) gesture, we expect the former relationship to be
stronger than the latter.

233

To assess the prediction the transitive gesture production,

but not intransitive gesture production, should predict
performance of transitive gesture recognition, we performed
a multiple regression with Spatial Recognition as the
dependent measure, and transitive and intransitive imitation
as the two predictor variables. Transitive imitation contributed strongly to the prediction of recognition (standard
coefficient = 0.62, P < 0.0001), whereas intransitive
imitation did not (standard coefficient = 0.18, P > 0.1).
Thus, taking into account the shared variance between the
two imitation tasks enables us to clarify that only transitive
imitation predicts transitive recognition.
To assess the prediction that components of transitive
gesture recognition (recognition of hand posture, arm
posture, amplitude/timing) should correlate more strongly
with production of these same gesture components in
transitive than in intransitive gesture, we performed nonparametric (Spearman) correlational analyses of the components of gesture recognition (hand posture, arm posture and
amplitude/timing) and these same components in the
transitive and intransitive imitation tasks. As 18 correlations
were performed, a Bonferroni-corrected P value of <0.0028
was required for significance. As shown in Table 2, the
strongest correlations were between the components of
transitive imitation and transitive recognition, whereas the
correlations between intransitive imitation and transitive
recognition were much weaker, and in many cases nonsignificant. Of particular interest is the quite strong relationship of transitive hand posture imitation and transitive hand
posture recognition.
Another way of looking at the relationship between
transitive and intransitive gesture imitation, on the one hand,
and transitive gesture recognition, on the other, is to assess
whether patients who are relatively impaired in gesture
recognition show a larger disparity between transitive and
intransitive gesture than patients for whom gesture recognition is relatively intact. To this end, we used the median
score on the spatial gesture recognition task (median = 91.7)
to divide the group of 44 patients into two groups (below
median = Low Recognition, n = 20; above median = High
Recognition, n = 24). We calculated the difference between
transitive and intransitive gestures and compared these
Table 2
Correlations of components of gesture recognition and gesture imitation
tasks

Imit.
Imit.
Imit.
Imit.
Imit.
Imit.

trans hand post.

trans arm post.
trans amp/time
intrans hand post.
intrans arm post.
intrans amp/time

T P < 0.0028.
TT P < 0.001.
TTT P < 0.0001.

Recog. trans
hand posture

Recog. trans
arm posture

Recog. trans
amp/time

0.80TTT
0.77TTT
0.58TT
0.38
0.57TT
0.49T

0.48T
0.52TT
0.25
0.40
0.40
0.21

0.37
0.33
0.29
0.35
0.43
0.30

234

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

scores for the Low Recognition and High Recognition

Groups using a Mann Whitney Test. As shown in Fig. 4,
there was greater relative impairment of the Low Recognition group in transitive as compared to intransitive
gestures (U = 62.0, P < 0.0001).

during observation of grasping actions have been identified

in monkey F5, the putative homologue of BA 45 in humans,
as well as in the inferior parietal lobe. In the present study,
we approached lesion analyses in two ways to be described
below.

5.3.1. Discussion
The data from three analyses of the relationship between
gesture production and recognition tell a consistent story.
The first analysis showed that transitive gesture imitation is
a strong and unique predictor of transitive gesture recognition. The second analysis showed that individual components of transitive gesture recognition (hand posture, arm
posture, and amplitude/timing) are more strongly related to
these same components instantiated in transitive as compared to intransitive gesture imitation. The third analysis
showed that patients who fare poorly in transitive gesture
recognition are disproportionately impaired in transitive (as
compared to intransitive) imitation, whereas patients who
are better at transitive gesture recognition show less
disparity between the two types of imitation. Together,
these data argue that the same representations subserve the
recognition and imitation of transitive gesture. Furthermore,
they suggest that these representations may be componential, with hand posture representations for transitive actions
particularly vulnerable to disruption in left-hemisphere
stroke (but not in CBD, as noted earlier). In the General
discussion section, we will present a model that addresses
what may be special about these representations for
transitive gesture and how they may differ from intransitive
gesture representations.

5.4.1. Methods
Patients lesions were segmented by a neurologist (H. B.
Coslett) and drawn into MRIcro software by an experienced
physician (R. Menon). Brodmann areas were identified by
the first author and Dr. Coslett, who were blinded to
subjects identities, using templates from Damasio and
Damasio [11] and Mai and Assheuer [36].
Patients were ranked in terms of their performance on
the Spatial gesture recognition task, the hand posture score
from the gesture recognition task, the total score from the
transitive gesture imitation task, and the hand posture score
from the transitive gesture imitation task (4 separate
rankings). For each measure, we divided the patient group
into thirds (insofar as permitted due to ties) and discarded
the patients whose performance was in the central third of
the distribution. The lesion data from the high and low
performing groups on the recognition task are shown in
Fig. 5.
We used MRIcro software to identify whether a lesion
was present or absent in 5 Brodmann areas of interest: areas
44 and 45 (inferior prefrontal) and areas 39, 40, and 7
(posterior parietal lobe and intraparietal sulcus) in the high
versus low performing group. Lesions in BA 39 (angular
gyrus) and the inferior portion of area 7, including the
superior bank of the intraparietal sulcus, were more
frequently associated with low Spatial gesture recognition
scores than high Spatial recognition scores v 2 > 4.6, P <
0.02 for both regions), with low hand posture recognition
scores than high hand posture recognition scores (v 2 > 4.3,
P < 0.02 for both), and with low total gesture imitation
scores than high total gesture imitation scores (v 2 > 5.0, P <
0.02 for both). Lesions in the inferior portion of area 7,
including the intraparietal sulcus, were also significantly
more strongly associated with low hand posture imitation
scores than with high hand posture imitation scores (v 2 =
5.6, P = 0.01). None of the comparisons involving areas 44
or 45 even approached significance (all Ps > 0.2).

5.4. Neuroanatomy
A final question regards the neuroanatomic basis of
transitive gesture representations, and particularly the
component of those representations that is specialized for
hand object interactions. As noted, mirror neurons active

Fig. 4. Transitive and intransitive gesture imitation performance of all

patients, divided into High Gesture Recognition and Low Gesture
Recognition groups on the basis of their performance of the Spatial
Gesture Recognition task.

5.4.2. Discussion
Lesion analyses suggest that in stroke patients, the
lesion(s) significantly associated with deficits in the
recognition of transitive gesture, and the hand posture
component of transitive gesture, are located in the inferior
parietal lobe and intraparietal sulcus. For prefrontal cortex,
in contrast, the association with deficits in gesture recognition or imitation did not even approach significance. This
is at least partially consistent with earlier findings of Varney
and Damasio [60] indicating that patients with deficits in
matching pantomimes to associated objects were likely to
have lesions in area 40 (supramarginal gyrus), areas 22 and

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

235

Fig. 5. Subtracted lesioned regions in high versus low performing groups on the Spatial Recognition task (top left and right) and Hand Posture score of the
Spatial Recognition task (bottom left and right). See Fig. 1 for additional explanation.

37 (posterior superior temporal lobe), and the basal ganglia,

but not the prefrontal cortex. They are also consistent with
data from Ferro, Martins, Mariano, and Caldas [18],
indicating that parietal lobe involvement was frequent in
post-acute (>3 months post-stroke) and chronic patients
with gesture recognition impairments. Together with the
present data, this suggests that in humans, Brocas area may
play a smaller role than the inferior parietal lobe in
recognizing complex familiar actions and hand postures.

6. General discussion
The hypothesis that IM reflects a deficit in the
representations underlying skilled object-related gestures,
with particular degradation of the hand posture component
of skilled object-related gestures, enabled us to generate a
number of predictions that were tested in the present study.
We demonstrated that patients with IM were disproportionately impaired in the imitation of transitive as compared to
intransitive gestures, and were particularly impaired in the

hand posture component of transitive gesture imitation. This

pattern persisted even when we examined only the transitive
gestures having the simplest hand postures. We also
demonstrated a strong association between performance of
transitive imitation and recognition tasks (compared to a
much weaker association between transitive recognition and
intransitive gesture imitation), as well as a particular
association between imitation and recognition of transitive
hand postures. Importantly, the deficits in object-related
hand posture recognition were not associated with aphasia
severity. Finally, we showed that deficiencies in the spatial
aspects of object-related gesture recognition and hand
posture recognition were associated with lesions to the left
inferior parietal lobe and intraparietal sulcus.
These data suggest that the same representations subserve the imitation and recognition of complex, skilled,
object-related movements, and thus provide support for the
likelihood that the direct matching hypothesis applies to
these complex, meaningful movements as well as to the
simple grasping and finger movements previously examined
in monkey and fMRI studies. In addition, the present data

236

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

suggest that the representations for skilled object-related

arm and hand actions may differ in important ways from the
representations subserving symbolic, non-object-related
gesture. In the following section, we discuss possible
reasons for this disparity.
One possible basis for the distinction between transitive
and intransitive gesture representations is that the former is
more closely related to an evolutionarily pre-existing system
specialized for manual grasping of objects. Fagg and Arbib
[17] take into account previous findings by Sakata and
colleagues (e.g., [52]) and Rizzolatti et al. (e.g., [50]) to
develop a model of the interactions between the anterior
intraparietal sulcus (AIP) and prefrontal F5 in programming
grasping movements. They suggested that AIP represents
the grasps afforded by objects, whereas F5 selects and
drives grasp execution. Thus, F5 aids in selecting from
among the multiple affordances that may be present in an
object based on the relevance of these affordances given
task constraints. AIP is hypothesized to provide an active
memory of the affordance selected by F5, and to
participate in updating this memory to reflect the grasp that
is actually executed. This knowledge can then be used to
inform subsequent interactions with objects (see also
[41,51,63,64]). A related conceptualization is offered by
Schettino, Adamovich, and Poizner [53], who suggest that
hand preshaping in early phases of reaching movements
reflects selection of a family of grasps based on the basic
geometry of the target object as well as developmental
experiences with hand object interactions. They contrast
this with adjustments occurring later in the reach that
modulate the basic grasp based on current information about
the object and hand.
Oztop and Arbib [41] develop this area of work further to
address the relationship between action execution and action
recognition. They offer a model that incorporates another
subclass of neurons in F5, called canonical neurons [39],
that discharge when a suitably graspable object is viewed. In
brief, the model illustrates how the mirror neurons may have
evolved to augment canonical neurons by providing visual
feedback on hand-state, i.e., the relationship of the shape
of the hand to the shape of an object. The hand-state
representation forms the basis for the ability to generalize
from ones hand to anothers hand, which in turn is
hypothesized to undergird the understanding of others
actions.
A number of investigators, then, hypothesize that the brain
computes and stores representations of hand posture based on
previous experiences with prehensile hand object interactions. These representations of hand posture are postulated to
inform (1) subsequent interactions with objects, and (2)
recognition of object-related prehensile actions, and (3)
arguably, to be present in both monkeys and humans. How
does the mirror system that stores memories of prehensile
actions differ from the skilled transitive gesture system of
humans? The extensive range of complex, skilled postures
and movements exhibited by humans (consider typing,

fingering a stringed instrument, grasping and re-grasping a

screwdriver as it is turned) requires knowledge regarding
functional uses of objects, the functional portions of the
object that are to be grasped, and the specific positions of the
fingers, hand, and arm for particular objects. In some
instances, the functional hand posture may be at odds with,
or an elaboration of, the prehensile hand posture called for by
object structure. The fact that the object-specific representations can be evoked even in pantomime tasks, without an
object present, suggests that they differ from the representations encoded by mirror neurons.
In humans, such an indirect route is likely to have
evolved to map between stored representations of objects in
the ventral stream, and representations of the body and
objects in space mediated by the dorsal stream. Even in the
monkey, the IPL is richly interconnected with both dorsal
and ventral stream structures. Regions within monkey IPL
project to distinct subdivisions of the dorsolateral frontal
cortex [10]. Reciprocal connections also exist between the
parietal lobe (area LIP) and the inferior temporal cortex
(areas TE and TEO), an area known to be involved in both
humans and monkeys in object recognition [13,62]. But
when overall brain volume is controlled, human IPL is
significantly larger than that of the rhesus monkey or
chimpanzee [16]. It has been suggested that human left IPL
has undergone evolutionary expansion paralleling the
development of language [16].
We suggest that left IPL/IPS system in humans mediates
between representations of object identity in the ventral
pathway and object/action structure in the dorsal pathway.
By conveying information about object identity to the dorsal
stream, we propose that the left IPL allows movements to be
selected that are appropriate to an objects category
membership, rather than only to its structural attributes.
Furthermore, we conjecture that the representations are
shaped through learning to map between identity and
structure information, taking the form of abstract movement
and posture representations that capture only those aspects
of the target action that are invariant with respect to the
initial body posture and the shape and location of any target
object. We can speculate that these transitive skilled action
representations are uniquely human, and are an elaboration
of the representations of hand object interactions encoded
by mirror neurons, both in terms of their complexity and
their requirements for long-term memory capacity. Their
abstract nature i.e., their independence from the details of
current constraints of objects and the environment renders
them potentially useful for recognizing transitive actions of
others even when the actions are pantomimes performed
without objects.
Unlike object-related actions, intransitive, symbolic
actions are not likely to retain such close ties to evolutionarily more primitive systems for object grasping, and are
not likely to be as strongly lateralized to the left hemisphere.
Rapcsak et al. [48] reported a right-handed man who,
consequent to a stroke resulting in virtually complete

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

destruction of the left hemisphere, was severely impaired in

transitive pantomime but relatively unimpaired with intransitive actions. Consistent with these data, left and right stroke
patients demonstrate equivalent impairments in intransitive
gesture pantomime and imitation [25]. Finally, as noted,
normal subjects perform more accurately in discriminating
intransitive gestures than transitive gestures, suggesting that
the former representations may be more widely distributed
and/or more readily activated than the latter [38].
In summary, we have presented evidence that the same
representations, mediated by the left inferior parietal lobe
and intraparietal sulcus, are likely to be evoked in production and recognition of pantomimed object-related
actions. The skilled gesture representations are active even
without the physical presence of objects, and unlike mirror
neurons, they appear to encode body and hand postures that
are specific to the functional use of particular objects.
Despite their differences, however, the evidence that the
same representations underlie perception and action for
abstracted (pantomimed) versions of complex functional
actions, just as for prehensile interactions with physical
objects, suggests that the mirror property may be a basic
organizing principle of the brain.

237

posture, arm posture, amplitude, and timing, for a maximum

score per gesture of 4 points. Normative data for the
meaningless imitation test are presented in Buxbaum,
Johnson-Frey, and Bartlett-Williams [8].
A.2. Gesture to sight of objects
Participants were instructed to show how they would
hold and use a common object (e.g., hammer, scissors,
screwdriver, saw, toothbrush) displayed on the tabletop,
pretending they had it in their hand. They were not
permitted to touch the object. The first instance of body
part as object error (e.g., for toothbrushing, forefinger
extended and moved over teeth) was corrected with a
repetition of the task instructions and an additional
instruction to show me how you would hold the object
as if it were in your hand. The second instance of such
errors was not corrected. There were 10 trials.
A.3. Error scoring
Both the apraxia screening tests and the gesture imitation
tests were scored according to the taxonomy reported in
Buxbaum, Giovannetti, and Libon [6] and reproduced
below.

Acknowledgments
Supported by NIH RO1-NS36387 and NIDRR H133G0
30169 to the first author, and by the Pennsylvania Department of Health. The Department of Health specifically
disclaims responsibility for any analyses, interpretations, or
conclusions. We are grateful to Branch Coslett for performing lesion analyses and Michael Arbib for his thoughtful
comments on an earlier draft of the manuscript.

Appendix A. Details of praxis screening tasks and

scoring
A.1. Meaningless movement imitation
Participants imitated 10 meaningless movements that
were spatial and temporal analogues of transitive gestures in
terms of plane of gesture (horizontal or vertical) and the
joints around which the movement occurred. The hand
posture was also modified to be unlike the meaningful
gestures (see Buxbaum, Giovannetti, and Libon [6] for
details). For example, the meaningless movement analogous
to hammering was a vertical movement performed by the
side of the body, with movements of the elbow and shoulder
joints, and the hand in a claw shape. Participants were
permitted to begin imitation (with the unimpaired left hand)
while still observing the target gesture. Performance was
videotaped and later scored by an experienced coder in
accordance with guidelines published in Buxbaum, Giovanetti, and Libon [6] with respect to the components hand

1. Hand Posture
Score as 0 if hand posture/grasp is unrecognizable,
flagrantly incorrect, or only transiently correct (small
fragment of total gesture with correct posture or
grasp). Score 0 for Fbody part as object_ (BPO)
errors.
Score as 1 if posture is correct or subtly incorrect
(e.g., hand aperture slightly too big or small; wrist
angle slightly incorrect).
2. Arm Posture/Trajectory
Score as 0 if arm posture and/or trajectory (e.g.,
joint angles, plane of movement relative to body/
environment (e.g., side to side instead of back and
forth), shape of movement (e.g., circular instead of
linear) are flagrantly incorrect or only transiently
correct (small fragment of total gesture with correct
posture).
Score as 1 if both arm posture and trajectory are
correct; or if arm posture and/or trajectory are subtly
incorrect (e.g., elbow slightly too bent, trajectory at
slight angle relative to what is appropriate, shape of
movement slightly distorted).
3. Amplitude
Score as 0 if size of movement is clearly too large or
too small (e.g., sawing with small scratching
movement), or if size is only transiently correct (small
fragment of total gesture with correct amplitude).
Score as 1 if size is correct or subtly too large or too
small (e.g., slight Fovershoot_ or Fundershoot_ in
movement amplitude).

238

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

4. Timing/Frequency
Score as 0 if speed of movement is flagrantly too
fast or slow; and/or if number of cycles of movement
is flagrantly too few or many (e.g., Fflipping_ coin 4
times in succession; Fscissoring_ only once).
Score as 1 if speed of movement is subtly too fast or
slow; and/or if frequency is subtly inappropriate (e.g.,
flipping coin twice; scissoring only twice).

References
[1] C. Barbieri, E. DeRenzi, The executive and ideational components of
apraxia, Cortex 24 (4) (1988) 535 543.
[2] B. Bell, Pantomime recognition impairment in aphasia: an analysis of
error types, Brain Lang. 47 (1994) 269 278.
[3] G. Buccino, F. Binkofski, G.R. Fink, L. Fadiga, L. Fogassi, V. Gallese,
R.J. Seitz, K. Zilles, G. Rizzolatti, H.J. Freund, Action observation
activates premotor and parietal areas in a somatotopic manner: an
fMRI study, Eur. J. Neurosci. 13 (2001) 400 404.
[4] G. Butterworth, On reconceptualizing sensori-motor coordination in
dynamic system terms, in: H.B.B.I. Bertenthal (Ed.), Sensory Motor
Organization and Development in Infancy and Early Childhood,
Kluwer Academic Press, The Netherlands, 1990, pp. 57 73.
[5] L.J. Buxbaum, Ideomotor apraxia: a call to action, Neurocase 7 (2001)
445 458.
[6] L.J. Buxbaum, T. Giovannetti, D. Libon, The role of the dynamic body
schema in praxis: evidence from primary progressive apraxia, Brain
Cogn. 44 (2000) 166 191.
[7] L.J. Buxbaum, A. Sirigu, M.F. Schwartz, R.L. Klatzky, Cognitive
representations of hand posture in ideomotor apraxia, Neuropsychologia 41 (2003) 1091 1113.
[8] L.J. Buxbaum, S.H. Johnson-Frey, M. Bartlett-Williams, Deficient
internal models for planning hand object interactions in apraxia,
Neuropsychologia 43 (2005) 917 929.
[9] L.J. Buxbaum, K. Kyle, M. Grossman, H.B. Coslett, Left inferior
parietal representations for skilled hand object interactions: evidence
from stroke and corticobasal degeneration, Cortex (in press).
[10] C. Cavada, The visual parietal areas in the macaque monkey: current
structural knowledge and ignorance, NeuroImage 14 (2001) S21 S26.
[11] H. Damasio, A.R. Damasio, Lesion Analysis in Neuropsychology,
Oxford Univ. Press, New York, NY, 1989.
[12] E. DeRenzi, Methods of limb apraxia examination and their bearing
on the interpretation of the disorder, in: E.A. Roy (Ed.), Advances in
Psychology: Neuropsychological Studies of Apraxia and Related
Disorders, vol. 23, North Holland, Amsterdam, 1985.
[13] C. Distler, D. Boussaoud, R. Desimone, L.G. Ungerleider, Corical
connections of inferior temporal area TEO in macaque monkeys,
J. Comp. Neurol. 334 (1993) 125 150.
[14] J.R. Duffy, R.J. Duffy, Three studies of deficits in pantomimic
expression and pantomimic recognition in aphasia, J. Speech Hear.
Res. (1981) 70 84.
[15] C. Dumont, B. Ska, Pantomime recognition impairment in Alzheimers
disease, Brain Cogn. 43 (2000) 177 181.
[16] D. Eidelburg, A.M. Galaburda, Inferior parietal lobule: divergent
architectonic asymmetries in the human brain, Arch. Neurol. 41
(1984) 843 852.
[17] A. Fagg, M.A. Arbib, Modeling parietal premotor interactions in
primate control of grasping, Neural Netw. 11 (1998) 1277 1303.
[18] J.M. Ferro, I.P. Martins, G. Mariano, A.C. Caldas, CT scan correlates
of gesture recognition, J. Neurol., Neurosurg. Psychiatry 46 (1983)
943 952.
[19] G. Gainotti, M. Lemmo, Comprehension of symbolic gestures in
aphasia, Brain Lang. 3 (1976).

[20] V. Gallese, L. Fadiga, L. Fogassi, G. Rizzolatti, Action recognition in

the premotor cortex, Brain 119 (1996) 593 609.
[21] L.J. Gonzalez Rothi, C. Ochipa, K.M. Heilman, A cognitive neuropsychological model of limb praxis, Cogn. Neuropsychol. 8 (1991)
443 458.
[22] S. Grafton, M.A. Arbib, L. Fadiga, G. Rizzolatti, Localization of grasp
representations in humans by positron emission tomography: 2.
Observation compared with imagination, Exp. Brain Res. 112 (1996)
103 111.
[23] K.Y. Haaland, D.L. Harrington, R.T. Knight, Neural representations of
skilled movement, Brain 123 (2000) 2306 2313.
[24] U. Halsband, J. Schmitt, M. Weyers, F. Binkofski, G. Grutzner, H.J.
Freund, Recognition and imitation of pantomimed motor acts after
unilateral parietal and premotor lesions: a perspective on apraxia,
Neuropsychologia 39 (2001) 200 216.
[25] M. Heath, E.A. Roy, S.E. Black, D.A. Westwood, Intransitive limb
gestures and apraxia following unilateral stroke, J. Clin. Exp. Neuropsychol. 23 (2001) 628 642.
[26] K.M. Heilman, L.J. Gonzalez Rothi, Apraxia, in: K.M. Heilman, E.
Valenstein (Eds.), Clinical Neuropsychology, Oxford Univ. Press,
New York, 1993, pp. 141 150.
[27] K.M. Heilman, L.J. Rothi, E. Valenstein, Two forms of ideomotor
apraxia, Neurology 32 (1982) 342 346.
[28] M. Iacoboni, R.P. Woods, M. Brass, H. Bekkering, J.C. Mazziotta, G.
Rizzolatti, Cortical mechanisms of human imitation, Science 286
(1999) 2526 2528.
[29] M. Jeannerod, J. Decety, F. Michel, Impairment of grasping movements following a bilateral posterior parietal lesion, Neuropsychologia
32 (1994) 369 380.
[30] S.H. Johnson-Frey, F.R. Maloof, R. Newman-Norland, C. Farrer,
Actions or hand object interactions? Human inferior frontal cortex
and action observation, Neuron 39 (2003) 1053 1058.
[31] D. Kareken, F. Unverzagt, K. Caldemeyer, M. Farlow, G.
Hutchins, Functional brain imaging in apraxia, Arch. Neurol. 55
(1998) 107 113.
[32] R. Klatzky, J. Pellegrino, B. McCloskey, S. Lederman, Cognitive
representations of functional interactions with objects, Mem. Cogn. 21
(1993) 294 303.
[33] K. Laimgruber, G. Goldenberg, J. Hermsdorfer, Manual and hemispheric asymmetries in the execution of actual and pantomimed
prehension, Neuropsychologia 43 (2005) 682 692.
[34] H. Liepmann, The Left Hemisphere and Action, University of Western
Ontario, London, 1905.
[35] F. Lucchelli, O.L. Lopez, P. Faglioni, F. Boller, Ideomotor and
ideational apraxia in Alzheimers disease, Int. J. Geriatr. Psychiatry 8
(1993) 413 417.
[36] J. Mai, J. Assheuer, Atlas of the Human Brain, 2nd edR, Academic
Press, 2003.
[37] A.N. Meltzoff, M.K. Moore, Imitation of facial and manual gestures
by human neonates, Science 198 (1977) 75 78.
[38] M. Mozaz, L. Rothi, J. Anderson, G.P. Crucian, K. Heilman, Postural
knowledge of transitive pantomimes and intransitive gestures, J. Int.
Neuropsychol. Soc. 8 (2002) 958 962.
[39] A. Murata, L. Fadiga, L. Fogassi, V. Gallese, V. Raos, G. Rizzolatti,
Object representation in the ventral premotor cortex (area F5) of the
monkey, J. Neurophysiol. 78 (1997) 2226 2230.
[40] N. Nishitani, R. Hari, Temporal dynamics of cortical representation for
action, Proc. Natl. Acad. Sci. U. S. A. 97 (2000) 913 918.
[41] E. Oztop, M.A. Arbib, Schema design and implementation of
the grasp-related mirror neuron system, Biol. Cybern. 87 (2002)
116 140.
[42] B. Pelgrims, M. Andres, X. Seron, J.-R. Duhamel, A. Sirigu, E.
Olivier, Role of the Left Supramarginalis Gyrus in Coding Hand
Gestures: Implication for Ideomotor Apraxia, Cognitive Neuroscience
Society, New York, 2005.
[43] M.T. Perenin, A. Vighetto, Optic ataxia: a specific disruption in
visuomotor mechanisms, Brain 111 (1988) 643 674.

L.J. Buxbaum et al. / Cognitive Brain Research 25 (2005) 226 239

[44] H. Poizner, L. Mack, M. Verfaellie, L.J.G. Rothi, K.M.
Heilman, Three-dimensional computergraphic analysis of apraxia:
neural representations of learned movement, Brain 113 (1990)
85 101.
[45] H. Poizner, M. Clark, A. Merians, B. Macauley, L. Gonzalez Rothi, K.
Heilman, Joint coordination deficits in limb apraxia, Brain 118 (1995)
227 242.
[46] H. Poizner, A.S. Merians, M.A. Clark, B. Macauley, L.J. Rothi, K.M.
Heilman, Left hemispheric specialization for learned skilled, and
purposeful action, Neuropsychology 12 (1998) 163 182.
[47] W. Prinz, Perception and action planning, Eur. J. Cogn. Psychol. 9
(1997) 129 154.
[48] S.Z. Rapcsak, C. Ochipa, P.M. Beeson, A.B. Rubens, Praxis and the
right hemisphere, Brain Cogn. 23 (1993) 181 202.
[49] G. Rizzolatti, G. Luppino, The cortical motor system, Neuron 31
(2001) 889 901.
[50] G. Rizzolatti, L. Fadiga, M. Matelli, V. Bettinardi, E. Paulesu, D.
Perani, F. Fazio, Localization of grasp representations in human by
PET: 1. Observation vs. execution, Exp. Brain Res. 111 (1996)
246 252.
[51] D. Rosenbaum, S. Engelbrecht, M. Bushe, L. Loukopoulos, Knowledge model for selecting and producing reaching movements, J. Mot.
Behav. 25 (1993) 217 227.
[52] H. Sakata, M. Taira, A. Murata, S. Mine, Neural mechanisms of visual
guidance of hand action in the parietal cortex of the monkey, Cereb.
Cortex 5 (1995) 429 438.
[53] L. Schettino, S. Adamovich, H. Poizner, Effects of object shape and
visual feedback on hand configuration during grasping, Exp. Brain
Res. 151 (2003) 158 166.

239

[54] J. Schwoebel, L. Buxbaum, H.B. Coslett, Representations of the

human body in the production and imitation of complex movements,
Cogn. Neuropsychol. 21 (2004) 285 299.
[55] A. Sirigu, L. Cohen, J.R. Duhamel, B. Pillon, B. Dubois, Y. Agid, A
selective impairment of hand posture for object utilization in apraxia,
Cortex 31 (1995) 41 55.
[56] A. Sunderland, S.M. Sluman, Ideomotor apraxia, visuomotor control
and the explicit representation of posture, Neuropsychologia 38 (2000)
923 934.
[57] A. Tessari, R.I. Rumiati, The strategic control of multiple routes in
imitation of actions, J. Exp. Psychol. Hum. Percept. Perform. 30 (6)
(2004) 1107 1116.
[58] A. Travinicek-Marterer, W. Danielczyk, M. Simanyi, P. Fischer,
Ideomotor apraxia in Alzheimers disease, Acta Neurol. Scand. 88
(1993) 1 4.
[59] N.R. Varney, A.L. Benton, Qualitative aspects of pantomime recognition defect in aphasia, Brain Cogn. 1 (1982) 132 139.
[60] N.R. Varney, H. Damasio, Locus of lesion in impaired pantomime
recognition, Cortex 23 (1987) 699 703.
[61] L. Wang, H. Goodglass, Pantomime, praxis, and aphasia, Brain Lang.
42 (1992) 402 418.
[62] M. Webster, J. Bachevalier, L.G. Ungerleider, Connection of inferior
temporal areas TEO and TE with parietal and frontal cortex in
macaque monkeys, Cereb. Cortex 4 (1994) 470 483.
[63] A. Witney, P. Vetter, D.M. Wolpert, The influence of previous
experience on predictive motor control, NeuroReport 12 (2001)
649 653.
[64] D.M. Wolpert, Z. Ghahramani, J. Flanagan, Perspectives and problems
in motor learning, Trends Cogn. Sci. 5 (2001) 487 494.