This article appeared in a journal published by Elsevier. The attached
copy is furnished to the author for internal non-commercial research
and education use, including for instruction at the authors institution
and sharing with colleagues.
Other uses, including reproduction and distribution, or selling or
licensing copies, or posting to personal, institutional or third party
websites are prohibited.
In most cases authors are permitted to post their version of the
article (e.g. in Word or Tex form) to their personal website or
institutional repository. Authors requiring further information
regarding Elsevier’s archiving and manuscript policies are
encouraged to visit:
Author's personal copy
Neurophysiologie Clinique/Clinical Neurophysiology (2008) 38, 189—195
Disponible en ligne sur www.sciencedirect.com
journal homepage: http://france.elsevier.com/direct/neucli
REVIEW/MISE AU POINT
What is ‘‘mirror’’ in the premotor cortex? A review
Propriétés miroir du cortex moteur: une revue
de la littérature
O. Morina, J. Grèzesb,∗
aInstitut Jean-Nicod, UMR 8129, EHESS, CNRS, département d’études cognitives, École normale supérieure,
29, rue d’Ulm, Paris, France
bLaboratoire de neurosciences cognitives, UMR 742 Inserm, département d’études cognitives, École normale
supérieure, 29, rue d’Ulm, Paris, France
Received 18 February 2008; accepted 18 February 2008
Available online 26 March 2008
cortex (Brodmann’s areas 6 and 44) during passive observation of actions. We found that such
activations regularly occurred. Looking for functional differentiation in the premotor cortex, we
found that one parameter was associated with systematic differences in location: this was the
presence or absence of targets. Observing biological actions with a physical target, compared to
a visual control showing no action at all, consistently activated the ventral premotor cortex (BA
6), and did so significantly more than observing target-less actions (with the same control). In
contrast, the activity in BA 44 (‘‘Broca’s area’’) was not modulated by the presence or absence
of targets. We propose that the ventral precentral gyrus, and not BA 44, shares the visual
properties of ‘‘mirror’’ neurons found in area F5 of the macaque brain.
© 2008 Elsevier Masson SAS. All rights reserved.
We review the findings of 24 fMRI studies examining activations in the premotor
Cortex prémoteur ;
Neurones miroirs ;
Aire de Broca ;
particulièrement sur le cortex prémoteur (aires de Brodmann 6 et 44) au cours de l’observation
passive d’actions. De nombreuses activations au sein des aires de Brodmann 6 et 44 sont
trouvées. La recherche d’une différenciation fonctionnelle entre ces deux aires du cortex pré-
moteur a permis d’identifier un paramètre qui semble avoir une influence systématique sur la
localisation des activations; ce paramètre est l’absence ou la présence d’un but à l’action.
L’observation de mouvements biologiques dirigés vers un but active de fac ¸on consistante l’aire
de Brodmann 6 du cortex prémoteur et ce significativement plus que pendant l’observation
de mouvements sans but. En revanche, la présence ou l’absence de but n’a pas d’influence
sur la présence d’activations au sein de l’aire 44 de Brodmann (aire dite de Broca). Nous
Cet article passe en revue les résultats de 24 études en IRMf, en se concentrant plus
E-mail address: firstname.lastname@example.org (J. Grèzes).
0987-7053/$ – see front matter © 2008 Elsevier Masson SAS. All rights reserved.
Author's personal copy
190O. Morin, J. Grèzes
suggérons que le gyrus précentral ventral (l’aire 6) et non pas l’aire 44 de Brodmann, partage
les propriétés visuelles des neurones «miroirs» enregistrés, chez le macaque, dans l’aire F5.
© 2008 Elsevier Masson SAS. All rights reserved.
Action and perception are not completely segregated in the
human brain. Thus, some overlap naturally exists between
systems for perceiving actions, and systems for producing
them. Since the premotor cortex is one of the key sites
of action organization in the human brain, premotor acti-
vations are expected to co-occur with action observation.
However, there is still considerable debate on: (1) the loca-
tion; and (2) the functional properties of premotor areas
where action meets action perception. The latter question
In this paper, we tackle the issue of location, and suggest
some clarifications concerning the problem of functional
Viewing actions (in what follows, we shall use that word
exclusively in the sense of: motion respecting biomechanical
constraints, as performed by a living being) is known to acti-
vate several specialized areas. Among these is the superior
temporal sulcus (STS) . In humans and monkeys, the STS is
consistently and specifically activated by the observation of
biological motion alone, independently of contexts or forms
. The STS is the visual entry to an important system that
was identified in macaques. It is located in the parietal and
premotor cortices, and has been called the ‘‘mirror system’’
after the discovery, in monkeys, of parietal and premotor
‘‘mirror’’ cells . These so-called ‘‘mirror neurons’’ were
originally found in area F5 of the macaque premotor cortex.
They fire both when the monkey sees a goal-directed action,
and when the monkey performs some (but in many cases not
the same) goal-directed action. The very same action, in the
absence of a target object, does not activate the ‘‘mirror’’
neuron. Recently, ‘‘mirror’’ neurons have been shown to
react to abstract properties of actions, such as their place
in complex motor sequences , or the inferred presence or
absence of an invisible target . Because of this teleologi-
cal sensitivity, ‘‘mirror’’ neurons have been attributed a key
role in a monkey’s understanding of his conspecifics’ actions.
Ever since the discovery of ‘‘mirror’’ neurons, resear-
chers have been looking for a human homologue of F5 - that
is, one or several brain areas whose conditions of activa-
tions match those of ‘‘mirror’’ neurons found in F5. These
conditions are: (1) being active during the execution of
observation of some target-directed actions. As of today, the
human premotor cortex, which is thought to comprise the
precentral cortex (BA 6) and the pars opercularis of the infe-
rior frontal gyrus (BA 44), is considered one of the most likely
candidates . Human BA 44 and 6 clearly share many func-
tional and histological properties of the macaque premotor
cortex [7,8]. The cytoarchitectonic properties of BA 44 put
it clearly aside from BA 45, and move it closer to BA 6 as a
premotor area. Furthermore, almost everyone agrees that
the ventral part of BA 6 in humans bears much resemblance
with the macaque ventral premotor cortex, both anatomi-
cally  and functionally . Whether or not we can go
further in speculating homologies is a matter of conside-
rable controversy: some see BA 44 as the human homologue
of macaque area F5 , but others disagree . BA 44
has become the centre of attention in the research concer-
ning a putative mirror system in humans. BA 44 forms part
of Broca’s area, which is crucially (though not exclusively
nor exhaustively — see ) involved in language produc-
tion . The possibility that it could host mirror-neurons
is tantalizing. The visual and motor properties of BA 44, as
revealed by functional neuroimaging studies, are thought to
speak in favour of this view.
First, BA 44 has been found to play a role in the prepara-
tion  and execution [16,17] of target-directed actions,
and particularly of grasping [18,19] a type of gesture for
which many ‘‘mirror’’ neurons show a clear preference. BA
44 also seems to be involved in imitation of finger move-
ments , although it is unclear to what extent this is the
case [21—23]. If it was the case, some think that it would
count as a proof that BA 44 is a mirror-neuron area. However,
as far as we know, macaque ‘‘mirror’’ neurons are not invol-
ved in any kind of imitation, and the executed and observed
behaviours they react to are not necessarily similar.
Second, many authors reported activations in BA 44 when
subjects viewed a movement that was performed by ano-
ther human being. However, these activations are puzzling
for several reasons. First, they are found both for target-
expect from a mirror-neuron-like area. Second, they have
proven very difficult to replicate [24,25], and they are
quite difficult to locate precisely, because of the anatomi-
cal variability of BA 44, whose location varies a lot from one
individual to another . BA 44 is delicate to identify on
the basis of individual sulcal borders alone  — and even
these are not investigated by most neuroimaging studies. In
a meta-analysis of activations found in Broca’s area,  it
was shown that only the most dorsal fringe of BA 44 was acti-
vated during action observation. In their 2001 meta-analysis
of PET studies , Grèzes and Decety failed to report any
activation in BA 44 for action observation; nevertheless,
they found that the premotor cortex was activated by action
observation. This lead them to propose that the precentral
cortex (BA 6), and not Broca’s area, might be the cortical
area exhibiting ‘‘mirror’’ properties for action observation,
which in turn makes it a likely homologue of F5. Their meta-
analysis, however, took only PET studies into account, which
made the data vulnerable to technical limitations.
In this mini-review, we put Grèzes and Decety’s hypothe-
sis to the test, this time with fMRI studies, whose spatial
resolution is superior. We analyzed activations associated
with action observation in the inferior frontal gyrus (BA 44)
and the precentral cortex (BA 6).
Twenty-four fMRI studies, published from 1999 to 2007, were
analyzed [30—49,19,20,24,25]. Only studies performed with
Author's personal copy
What is ‘‘mirror’’ in the premotor cortex? A review 191
healthy volunteers were included (one study —  — also
involved autistic subjects, but only healthy subjects’ acti-
vations were taken into account). Since our goal was to
locate visuomotor activity in the premotor cortex as pre-
cisely as possible, we restricted our choice to those studies
that provided MNI or Talairach coordinates.
We picked up all the contrasts comparing activations
elicited by the observation of an action made by a living
being with another visual stimulus. Stimuli used in the stu-
dies were all films (not photographs) depicting real actions.
Contrasts that lacked a visual control — even a rudimen-
tary one, such as a grey background — were not taken into
account (the contrasts we had to cast aside for this reason
supported our main finding concerning the difference bet-
ween BA 44 and BA 6 regarding the preference for goals).
Only contrasts measured over the whole brain were analy-
zed, except two . This left us with 40 contrasts in 19
In all studies, subjects were asked to watch the action
and the control visual stimulus; in six studies, subjects
had to perform some low-level attentional task (like n-
back matching). In all other studies subjects simply watched
passively. Studies in which subjects were asked to ima-
gine themselves performing the action, and contrasts
in which the subject observed the action in order to
plan another (or the same) action, were not taken into
Contrasts were sorted into three main groups. One first
group gathered five contrasts that compared an observed
biological action with another action that was not percei-
ved as biological (according to subjects), either because
it was unrecognizable, or because it was artificially distor-
ted. Contrasts in which biological action was not recognized
were discarded. This was the ‘‘biological versus non biologi-
cal’’ group. The second group gathered all contrasts (n=12)
that compared an observed action with another observed
action. This group was the ‘‘action versus action’’ group.
The last group gathered all contrasts (23) that compared an
observed action with an action-less visual stimulus (either
a neutral background or a still picture of the observed
action video). This group was the ‘‘action versus no action’’
In each of these three groups, a difference was made
between contrasts that used a goal-directed action as their
main stimulus, and contrasts that used an action without a
target as their main stimulus. A ‘‘goal’’, here and in the fol-
lowing, is an independent physical target that the effector
reaches at the end of the action.
For each contrast, the MNI coordinates of all fron-
tal activations were analyzed using SPM anatomy toolbox
(Statistical parametric mapping, http://www.fil.ion.ucl.
ac.uk/spm). Talairach coordinates were transformed before
analysis. Only activations found in the premotor cortex that
is, in the precentral cortex (BA 6) and in the inferior frontal
gyrus (IFG, BA 44), are given and discussed in this paper. IFG
activations found in BA 45 are mentioned only when rele-
vant. Activations are signalled as falling in BA 6 or BA 44
when SPM anatomy toolbox located them in one of these
areas with a probability of 30% or more. Activations reco-
gnized by Anatomy Toolbox as falling in the pars opercularis
of the inferior frontal gyrus, but not in BA 44, are mentioned
as belonging to the inferior frontal gyrus, but not in BA 44.
Results and discussion
Overall, more than half of our contrasts (27/40) found acti-
vations in the premotor cortex, whether in BA 6, BA 44,
or in both. Activations were not found to be especially
lateralized, both generally and for individual categories of
contrasts. BA 6 and BA 44 did not differ in this respect. Some
aspect of action observation seems to recruit the premotor
cortex in a very diverse array of conditions. This common-
place finding is coherent with a more general observation:
perception, even of nonbiological visual stimuli, very often
has a motor component, which can be observed in the pre-
motor cortex . Motor involvement in visual perception
is not limited to action perception, and explanations based
on mirror neurons are not the only available option. In the
detailed analysis that follows, however, we can try to find
what caused these cases of premotor involvement in action
Biological versus nonbiological actions
Five studies [45—47,36,38] contrasted the observation of
biological actions to the observation of non-biological
actions. None of the actions used in the contrasts were goal-
directed; four of them used point-light biological motion
stimuli, as designed by Johansson . Point-light stimuli
to biological motion, since it presents in a reliable and reco-
gnizable way all the dynamic information that characterizes
living beings in motion, minus all kinds of configural informa-
a gesture without paying attention to its dynamic proper-
ties. The other one  used a video of a moving finger
contrasted with a video of a moving pair of scissors. None
of these contrasts revealed any activation in the precentral
cortex (BA 6). The study by Costantini et al. , that did
not make use of point-light biological motion, found acti-
vation in the IFG (BA 44). Only one out of four studies that
employed point-light biological motion stimuli found acti-
vation in our region of interest. PET studies (for example
) also failed to show motor activity following biological
motion observation. This may be explained by a lack of sen-
sitivity of the apparatus used: the scanner in Saygin et al.
 was much more powerful (4T).
However, another explanation may be that the premotor
cortex is not sensitive to biological motion per se. Several
fMRI and PET studies found activations in the inferior fron-
tal gyrus (BA 44) when contrasting actions that were not
perceived as biological with actions that were perceived as
biological. Three studies [38,53,36] found activations in BA
44 for biomechanically impossible hand actions contrasted
with mechanically impossible, nonbiological movements.
Action versus action
Thirteen contrasts in six studies [39—44] contrasted an
action (or a sequence of movements) with another action
(or sequence of actions) that differed from the first because
it was put in a different context (two contrasts), because
it was viewed from another perspective (two contrasts),
because the motor sequence was not the same although
Author's personal copy
192O. Morin, J. Grèzes
the effector remained the same (five contrasts) or, lastly,
because the effector was different although the motor
sequence did not vary much (three contrasts). This latter
category will be examined in the ‘‘somatotopy’’ section, in
the next part.
According to an influential hypothesis, the involvement
of the premotor cortex in action observation is due to the
role the mirror-neuron system plays in action understanding,
which means the extraction of things like goals, agency, or
a specific type of action, from visual information. If that
was the case, one should expect premotor activations not
only when action observation is compared with observa-
tion of ‘‘no action’’ stimuli, but also when comparing the
observation of different actions (highlighting some speci-
fic type of action), or of the same gesture within different
contexts (suggesting different higher-order objectives), or
of the same action viewed from different perspectives
(which might suggest different kinds of agency).
Same action, different context
If the premotor cortex (or part of it) was the human homo-
logue of a mirror-neuron system, it should be sensitive to
the context of an action, just as mirror-neurons are [4,5].
The data we surveyed here speak against this view. Two stu-
dies [39,41] compared the condition of seeing a gesture in a
certain context with the condition of seeing it outside that
context. In Iacoboni et al. , two different contexts were
tested (one was a table prepared for breakfast, the other
was the same table once breakfast had taken place); in
Koski et al. 2002 , symbolic target-like dots were added
in videos representing short intransitive finger movements.
None of these two studies found any activation in our region
of interest. Iacoboni et al.  found several activations in
the inferior frontal gyrus, most of them falling squarely in
BA 45, none in BA 44.
Same action, different perspective
One contrast, in Jackson et al.’s study , compared videos
of a gesture seen from first-person perspective, with videos
of the same gesture seen from third-person perspective. A
second contrast made the reverse comparison. No activa-
tion was found in the inferior frontal gyrus (BA 44). Only
the first contrast (first person versus third person activation)
revealed activation in the precentral gyrus (BA 6).
Same effector, same object, different action sequence
Five contrasts in two studies [42,43] compared videos sho-
wing complex sequences of goal-directed movements with
other complex sequences of movements that differed in
the way they were organized; this could be because the
objects used were different or used in different ways, or
in a different order. These contrasts were devised to reveal
sensitivity to the more abstract properties of actions. Two of
these contrasts found activations both in the premotor cor-
tex and the inferior frontal gyrus (BA 44); two others found
activations in the precentral cortex (BA 6) only. One found
no activation at all.
Taken together, these results are inconclusive. We cannot
rule out some kind of premotor involvement in judging or
classifying actions — for example, in Manthey et al. .
But a systematic association of the premotor cortex with
our understanding of goals and agency is difficult to support
with the present data.
Action versus no action
Twenty-three contrasts in nine studies [31—34,37,19,
24,40,48] compared activations associated with making sub-
jects observe actions with activations provoked by visual
stimuli that presented no movement. A first set of studies
used, as a control, a still image of the action (or simply
of the effector) that was shown as a dynamic stimulus in
the experimental condition. All other studies contrasted an
action with a rudimentary visual baseline, such as a grey
background or a fixation cross. The results for these two
sets are presented together. Nine contrasts out of 22 used a
goal-directed action as the main action stimulus: these will
be discussed along with the others.
Somatotopy along BA 6 and BA 44
In the ‘‘action versus action’’ group, Sakreida et al. 
contrasted actions made with distal effectors (mouth and
fingers) with actions made with more proximal effectors
(such as ankles or knees), and with even more proximal
(‘‘axial’’) effectors, such as shoulders or waist muscles.
The pattern of activations they found showed activations
in BA 44 alone in the ‘‘distal versus proximal and axial’’
contrast, and BA 6 in the other two, suggesting that action
observation in the premotor cortex is indeed somatotopi-
cally organized in a dorsoventral arrangement, the distal
movements being represented more dorsally and the proxi-
mal movements (such as those involving fingers and lips)
In the 23 contrasts of this group, three studied finger
movements, eight studied mouth movements, and four stu-
died hand-arm actions, two studied feet actions, one, leg
actions, and two, whole-body actions. Three contrasts, that
did not compare a specific part of the body with another,
were not taken into account. The localization of activations
in BA 6, BA 44, the precentral cortex or the inferior frontal
gyrus, was not found to be influenced by the kind of effec-
tors shown. Activations associated with viewing fingers or
mouth actions did not involve the inferior frontal gyrus (BA
44) more than activations associated with viewing less distal
effectors. In particular, no activation whatsoever was found
in BA 44 for the observation of finger movements (tested in
[34,48,24]). If any, activations associated with observing leg
and whole-body actions are slightly more often located in
the inferior frontal gyrus (BA 44), than those provoked by
mouth or finger actions (see for example ). Our data do
not seem to follow the familiar somatotopic pattern. This
is all the more surprising since one study of this group 
found exactly the same kind of somatotopy as was evidenced
though undeniably present, are highly task-dependent, idio-
syncratic, and consequently likely to be destroyed when
pooling together results from different protocols.
Activations in the inferior frontal gyrus (BA 44)
Eight out of 23 contrasts found activation in Brodmann’s area
44 associated with seeing biological actions, as opposed to
visual stimuli presenting no action. BA 44 activations were
Author's personal copy
What is ‘‘mirror’’ in the premotor cortex? A review193
or the absence of a goal, in the precentral area 6 and in the inferior frontal area BA 44. a: the activations are presented according
to their Y; and b: their Z-axis coordinates. The location of BA 6 activations (blue and green) is posterior to BA 44 (yellow and purple)
on the Y-axis, and dorsal to BA 44 on the Z-axis. In the two left bar charts (a), the number of activations (effectif) found in BA 6
for goal-directed actions (blue) and for non goal-directed actions (green) is presented as a function of the Y coordinate. In the two
right bar charts (a), the number of activations (effectif) found in BA 44 for goal-directed actions (yellow) and for non goal-directed
actions (purple) is presented as a function of the Y-coordinate. The same applies for the bar chart in (b). One can see how the
presence of goals during action perception influences the number of activations in the premotor cortex (BA 6), not in BA 44.
Figure 1 Les diagrammes à barres représentant le nombre d’activation ont trouvé pendant l’observation des actions en fonction
de la présence ou de l’absence d’un but, dans le secteur precentral 6 et dans le BA inférieur 44 de secteur frontal. Les activation
sont présentées selon leur Y (a) et leur axe de Z coordonne (b). L’endroit du BA 6 activation (bleues et vertes) est postérieur au BA
44 (jaune et pourpre) sur l’axe de Y, et la dérive dorsale au BA 44 sur l’axe de Z. Dans les deux diagrammes à barres laissés (a), le
nombre d’activation (effectif) trouvées en BA 6 pour des actions but-dirigées (bleues) et pour des actions non but-dirigées (vert)
est présenté en fonction de y. Dans les deux bons diagrammes à barres (a), le nombre d’activation (effectif) trouvées en BA 44 pour
des actions but-dirigées (jaune) et pour des actions non but-dirigées (pourpres) est présenté en fonction de y. Le même s’applique
pour le diagramme à barres en (b). On peut voir comment la présence des buts pendant la perception d’action influence le nombre
d’activation dans le cortex de premotor (BA 6), pas en BA 44.
Bar charts representing the number of activations found during the observation of actions as a function of the presence
accompanied by activations in the precentral gyrus (BA 6)
in all cases except one. The contrast that revealed BA 44
alone presented the subjects with videos of moving feet
compared to photographs of still feet. A disputable case of
BA 44 activation is found in Beauchamp et al. : they
presented their subjects with images of human whole-body
actions, and animated tools (such as hammers rocking back
and forth). The authors  could not dissociate the activa-
tions provoked by these two kinds of stimuli in BA 44.
Five (out of 14) non goal-directed stimuli provoked acti-
vations in BA 44; three (out of 9) goal-directed action stimuli
provoked activations in BA 44. Therefore, goals do not
appear to have a significant effect on BA 44 activations (chi
square test: 0.014, d.f. 1, P-value 0.907 — Pearson, uncor-
Activations in the precentral gyrus (BA 6)
Fourteen out of 23 contrasts found activations in Brodmann’s
area 6 for seeing action. Eight of these contrasts revealed
BA 6 in isolation, that is, unaccompanied by activations in BA
44. Thus, while activations in BA 44 are strongly associated
with activations in BA 6, the reverse association does not
BA 6 activations for the whole group were equally present
in the right and left hemisphere, and located on average at
y=0, z=41, that is probably in the ventral part of the pre-
central gyrus, at the border with the dorsal part . Out of
14 non goal-directed stimuli, only five provoked activations
in Brodmann’s area 6. Likewise, in PET studies, activations
associated with observing actions without targets, still in
the premotor cortex, vary a lot in localization, and prove
very difficult to obtain [55—57].
In contrast, all of nine goal-directed action stimuli were
associated with ventral activations in BA 6. The presence of
goals does seem to have an effect upon activations in BA 6
(chi square: 9.505, d.f.1, P<0.002 — Pearson, uncorrected).
BA 6 and BA 44 significantly differ from each other in this
respect (chi square: 9, d.f. 1, P<0.003 — Pearson, uncorrec-
ted). This effect remains very significant if we consider the
inferior frontal gyrus at large (including BA 45). Similarly,
if we take into account activations situated in the precen-
tral gyrus that do not fall squarely within BA 6, both the
effect of goals, and the difference between the IFG and
the precentral gyrus concerning this effect, remain signi-
ficant. Our calculation gives the same weight to each of
the 23 tests, but we could find the same result by looking
at the number of activations found for observing goal and
non-goal-directed actions in BA 6 and BA 44 (see Fig. 1).
This finding is coherent with existing data from PET stu-
dies [29,58] that find the precentral gyrus, and not BA 44,
Author's personal copy
194O. Morin, J. Grèzes
consistently active for goal-directed action observation. Fol-
lowing Grèzes and Decety , we suggest that the ventral
precentral gyrus shares the visual properties of ‘‘mirror’’
neurons found in F5 in macaques, whereas BA 44 does
Observing actions, as opposed to still pictures or neutral sti-
muli, activates the premotor cortex, ventral or dorsal, in a
majority of studies. One factor makes a systematic diffe-
rence in the location of these premotor activations, that is,
the presence or absence of targets. As could be expected
from monkey data, the human ventral premotor cortex is
consistently and specifically activated when subjects view
goal-directed actions. More surprisingly, this kind of visuo-
motor activity can be precisely located in ventral BA 6
(around z=41), quite beyond the boundaries of BA 44. Visuo-
motor activations in this area are reliable and consistent
as far as target-directed gestures are concerned, whilst in
the inferior frontal gyrus (BA 44), visuomotor activations are
much more erratic, and show no sensitivity to the presence
of targets. We tentatively conclude that the ventral precen-
tral cortex has the visual properties of F5 ‘‘mirror’’ neurons,
contrary to BA 44. However, this conclusion must remain
tentative for the moment. Indeed, it can only be proven
by further fMRI research comparing the observation of gras-
ping in the macaque and human brains. Relevant fMRI data
in macaques have already been provided by Nelissen et al.
If our conclusion proved to be correct, theories specu-
lating that a mirror-neuron-like system gave rise to our
language faculty  may be found to rest on fragile
grounds. Still, we do not wish to minimize the important
motor properties of BA 44. These properties are fascina-
ting, as they may represent the missing link between human
language and executive functions . It would indeed be
interesting to find if either BA 6 or BA 44 are necessary for
(and not merely involved in) action recognition; but this is
precisely what neuroimagery cannot tell us — studies using
lesions, whether in brain-damaged patients or produced by
TMS, may provide the answer.
 Perrett DI, et al. Frameworks of analysis for the neural
representation of animate objects and actions. J Exp Biol
 Allison T, Puce A, McCarthy G. Social perception from
visual cues: the role of the STS region. Trends Cogn Sci
 di Pellegrino G, Fadiga L, Fogassi L, Gallese V, Rizzolatti G.
Understanding motor events: a neurophysiological study. Exp
Brain Res 1992;91:176—80.
 Fogassi L, et al. Parietal lobe: from action organization to
intention understanding. Science 2005;308(5722):662—7.
 Umiltà M, Kohler E, Gallese V, Fogassi L, Fadiga L, Rizzolatti G.
I know what you are doing: a neurophysiological study. Neuron
 Rizzolatti G, Craighero L. The mirror-neuron system. Annu Rev
 Geyer S, Matelli M, Luppino G, Zilles K. Functional neuroana-
tomy of the primate isocortical motor system. Anat Embryol
 Schubotz RI, von Cramon DY. Functional-anatomical concepts
of human premotor cortex: evidence from fMRI and PET stu-
dies. Neuroimage 2003;20:120—31.
 Picard N, Strick PL. Imaging the premotor areas. Curr Opin
 Fink GR, Frackowiak RS, Pietrzyk U, Passingham RE. Multiple
nonprimary motor areas in the human cortex. J Neurophysiol
 Rizzolatti G, Arbib MA. Language within our grasp. Trends Neu-
 Passingham RE. Broca’s area and the origins of human vocal
skill. Philos Trans R Soc Lond B Biol Sci 1981;292(1057):167—
 Dronkers NF, Plaisant O, Iba-Zizen MT, Cabanis EA. Paul Broca’s
historic cases: high resolution MR imaging of the brains of
Leborgne and Lelong. Brain 2007;130:1432—41.
 Hillis AE, Work M, Barker PB, Jacobs MA, Breese EL, Maurer K.
Re-examining the brain regions crucial for orchestrating speech
articulation. Brain 2004;127:1479—87.
execution and suppression of copied movements in the human
brain. Exp Brain Res 1998;120:386—98.
 Binkofski F, Buccino G, Posse S, Seitz RJ, Rizzolatti G, Freund
HJ. A fronto-parietal circuit for object manipulation in man:
evidence from an fMRI study. Eur J Neurosci 1999;2:3276—86.
 Binkofski F, Buccino G. Motor functions of the Broca’s region.
Brain Lang 2004;89:362—9.
 Ehrsson HH, Fagergren A, Jonsson T, Westling G, Johansson RS,
Forssberg H. Cortical activity in precision — versus power-grip
tasks: an fMRI study. J Neurophysiol 2000;83:528—36.
 Grèzes J, Armony JL, Rowe J, Passingham RE. Activations rela-
ted to ‘‘mirror’’ and canonical neurons in the human brain: an
fMRI study. Neuroimage 2003;18:928—37.
 Iacoboni M, Woods RP, Brass M, Bekkering H, Mazziotta JC, Riz-
zolatti G. Cortical mechanisms of human imitation. Science
 Makuuchi M. Is Broca’s area crucial for imitation? Cereb Cortex
 Tanaka S, Inui T. Cortical involvement for action imitation of
hand/arm postures versus finger configurations: an fMRI study.
 Tanaka S, Inui T, Iwaki S, Konishi J, Nakai T. Neural substrates
involved in imitating finger configurations; an fMRI study. Neu-
 Jonas M, Siebner HR, Biermann-Ruben K, Kessler K, Baumer T,
Buchel C, et al. Do simple intransitive finger movements consis-
tently activate fronto-parietal mirror neuron areas in humans?
 Leslie KR, Johnson-Frey SH, Grafton ST. Functional imaging of
face and hand imitation: towards a motor theory of empathy.
 Amunts K, Zilles K. Advances in cytoarchitectonic mapping of
the human cerebral cortex. Anat Bases Funct Magn Reson Ima-
 Tomaiuolo F, MacDonald JD, Caramanos Z, Posner G, Chiavaras
M, Evansz AC, et al. Morphology, morphometry an probability
mapping of the pars opercularis of the inferior frontal gyrus:
an in vivo MRI analysis. Eur J Neurosci 1999;11:3033—46.
 Molnar Szakacs I, Iacoboni M, Koski L, Maziotta JC. Functional
evidence from fMRI studies of imitation and observation. Cereb
 Grèzes J, Decety J. Functional anatomy of execution, mental
analysis. Hum Brain Mapp 2001;12:1—19.
Author's personal copy
What is ‘‘mirror’’ in the premotor cortex? A review195
 Aziz-Zadeh L, Koski L, Zaidel E, Mazziotta J, Iacoboni M. Late-
ralization of the human mirror neuron system. J Neurosci
 Beauchamp MS, Lee KE, Haxby JV, Martin A. Parallel visual
motion processing streams for manipulable objects and human
movements. Neuron 2002;34:149—59.
 Buccino G, Lui F, Canessa N, Patteri I, Lagravinese G, Benuzzi
F, et al. Neural circuits involved in the recognition of actions
performed by nonconspecifics: an fMRI study. J Cogn Neurosci
 Buccino G, Vogt S, Ritzi A, Fink AR, Zilles K, Freund HJ, et al.
Neural circuits underlying imitation-learning of hand actions:
an event-related fMRI study. Neuron 2004;42:323—34.
 Buccino G, Fink L, Fadiga L, Fogassi L, Gallese V, Seitz RJ,
et al. Action observation activates premotor and parietal
areas in a somatotopic manner: an fMRI study. Eur J Neurosci
 Carr L, Iacoboni M, Dubeau MC, Mazziotta JC, Lenzi GL.
Neural mechanisms of empathy in humans: a relay from neu-
ral systems for imitation to limbic areas. PNAS 2003;100:
 Costantini M, Galati G, Ferretti A, Caulo M, Tartaro A,
Romani GL, et al. Neural systems underlying observation of
humanly impossible movements: an fMRI study. Cereb Cortex
 Grèzes J, Pichon S, de Gelder B. Perceiving fear in dynamic
body expressions. Neuroimage 2007;35(2):959—67.
 Grèzes J, Fonlupt P, Bertenthal B, Delon-Martin C, Segebarth C,
Decety J. Does perception of biological motion rely on specific
brain regions? Neuroimage 2001;13(5):775—85.
 Iacoboni M, Molnar Szakacs I, Gallese V, Buccino G, Mazziotta
JC, Rizzolatti G. Grasping the intentions of others with one’s
own mirror-neuron system. PloS Biol 2005;3:e79.
 Jackson PL, Meltzoff AN, Decety J. Neural circuits invol-
ved in imitationand perspective-taking.
 Koski L, Wolschlager A, Bekkering H, Woods RP, Dubeau MC,
Mazziotta JC, et al. Modulation of motor and premotor acti-
vity during imitation of target-directed actions. Cereb Cortex
 Molnar-Szakacs I, Kaplan I, Greenfield PM, Iacoboni M.
Observing complex action
fronto-parietal mirror neuron system. Neuroimage 2006;33:
 Manthey S, Schubotz R, von Cramon DY. Premotor cortex in
observing erroneous action: an fMRI study. Cogn Brain Res
 Sakreida RI, Schubotz R, Wolfensteller I, von Cramon DY.
Motion class dependency in observers’ motor areas revea-
led by functional magnetic resonance imaging. J Neurosci
 Saygin AP, Wilson SM, Hagler DJ, Bates E, Sereno M. Point-
light biological motion perception activates human premotor
cortex. J Neurosci 2004;24(27):6181—8.
sequences:the roleof the
 Servos P, Osu R, Santi A, Kawato M. The neural substrates
of biological motion perception: an fMRI study. Cereb Cortex
 Vaina LM, Solomon J, Chowdhury S, Sinha P, Belliveau JW.
Functional neuroanatomy of biological motion perception in
humans. PNAS 2001;98:11656—61.
 Wheaton KJ, Thompson JC, Syngeniotis A, Abbott DF, Puce A.
Viewing the motion of human body parts activates different
regions of premotor, temporal, and parietal cortex. Neuroi-
 Williams JG, Waiter GD, Gilchrist A, Perrett DI, Murray AD, Whi-
ten A. Neural mechanisms of imitation and ‘‘mirror neuron’’
functioning in autistic spectrum disorder. Neuropsychologia
 Schubotz, RI. Prediction of external events with our motor sys-
tem: towards a new framework. Trends in cognitive sciences,
 Johansson G. 1973 visual perception of biological motion and
a model for its analysis. Percept Psychophys 1973;14:201—
 Bonda E, Petrides M, Ostry D, Evans A. Specific involvement of
human parietal systems and the amygdala in the perception of
biological motion. J Neurosci 1996;16(11):3737—44.
 Stevens JA, Fonlupt P, Shiffrar M, Decety J. New aspects
of motion perception: selective neural encoding of apparent
human movements. Neuroreport 2000;11(1):109—15.
 Tomassini V, Jbabdi S, Klein JC, Behrens TE, Pozzilli C,
Matthews PM, et al. Diffusion-weighted imaging tractography-
based parcellation of the human lateral premotor cortex
identifies dorsal and ventral subregions with anatomical
and functional specializations. J Neurosci 2007;27(38):10259—
 Grèzes J, Costes N, Decety J. The effects of learning and
intention on the neural network involved in the perception of
meaningless actions. Brain 1999;122:1875—87.
 Decety J, Grezes J, Costes N, Perani D, Procyk E, Grassi F, et al.
Brain activity during observation of actions. Influence of action
content and subject’s strategy. Brain 1997;120:1763—77.
 Hermsdörfer J, Goldenberg G, Wachsmuth C, Conrad B, Cebal-
los Baumann AO, Bartenstein P, et al. Cortical correlates of
gesture processing: clues to the cerebral mechanisms under-
lying apraxia during the imitation of meaningless gestures.
 Perani D, Fazio F, Borghese NA, Tettamanti M, Ferrari S, Decety
J, et al. Different brain correlates for watching real and virtual
hand actions. Neuroimage 2001;14:749—58.
 Nelissen K, Luppino G, Vanduffel W, Rizzolatti G, Orban GA.
Observing others: multiple representations in the frontal lobe.
 Rizzolatti G, Fogassi L, Gallese V. Neurophysiological mecha-
nisms underlying the understanding and imitation of action.
Nat Rev Neurosci 2001;2:661—70.
 Koechlin E, Jubault T. Broca’s area and the hierarchical orga-
nization of human behavior. Neuron 2006;6:963—74.