Mirror Neurons and their Clinical Relevance

Abstract

One of the most exciting events in neurosciences over the past few years has been the discovery of a mechanism that unifies action perception and action execution. The essence of this 'mirror' mechanism is as follows: whenever individuals observe an action being done by someone else, a set of neurons that code for that action is activated in the observers' motor system. Since the observers are aware of the outcome of their motor acts, they also understand what the other individual is doing without the need for intermediate cognitive mediation. In this Review, after discussing the most pertinent data concerning the mirror mechanism, we examine the clinical relevance of this mechanism. We first discuss the relationship between mirror mechanism impairment and some core symptoms of autism. We then outline the theoretical principles of neurorehabilitation strategies based on the mirror mechanism. We conclude by examining the relationship between the mirror mechanism and some features of the environmental dependency syndromes.

Full text

24 NATURE CLINICAL PRACTICE NEUROLOGY JANUARY 2009 VOL 5 NO 1
www.nature.com/clinicalpractice/neuro
Mirror neurons and their clinical relevance
Giacomo Rizzolatti*, Maddalena Fabbri-Destro and Luigi Cattaneo
INTRODUCTION
Traditionally, it has been assumed that the under-
standing of actions performed by others depends
on inferential reasoning.1–3 Theoretically, when
we witness the actions of others, the information
could initially be subjected to sensory processing
and then be sent to higher order ‘association’ areas
where it is elaborated on by sophisticated cogni-
tive mechanisms and compared with previously
stored data. At the end of this process, we would
know what others are doing.4
It is possible that this cognitive operation
might indeed occur in some situations when
the behavior of the observed person is difficult
to interpret.5–7 However, the ease with which we
usually understand what others are doing sug-
gests that an alternative mechanism might be
involved in action perception. The essence of
this alternative system is that actions performed
by others, after being processed in the visual
system, are directly mapped onto observers’
motor representations of the same actions. The
observers are aware of the outcomes of their own
actions, so the occurrence of a neural pattern
similar to that present during their own volun-
tary motor acts will enable them to understand
the actions of others.
Evidence in favor of the existence of this direct
sensory–motor mapping mechanism came from
the discovery of a set of motor neurons, known
as mirror neurons, that fire both when a monkey
performs a given motor act and when it observes
another individual performing an identical or
similar motor act.8,9 In this article, we will first
review the basic properties of this mechanism,
which is known as the mirror mechanism. We
then examine the relevance of the mirror mecha-
nism for the interpretation of clinical syndromes
such as autism, and for the development of
motor rehabilitations strategies.
MIRROR NEURONS IN THE MONKEY
Mirror neurons were originally discovered in
the ventral premotor cortex (area F5) of the
macaque monkey.8,9 The defining characteristic
SUMMARY
One of the most exciting events in neurosciences over the past few years
has been the discovery of a mechanism that unifies action perception and
action execution. The essence of this ‘mirror’ mechanism is as follows:
whenever individuals observe an action being done by someone else, a set
of neurons that code for that action is activated in the observers’ motor
system. Since the observers are aware of the outcome of their motor acts,
they also understand what the other individual is doing without the need
for intermediate cognitive mediation. In this Review, after discussing the
most pertinent data concerning the mirror mechanism, we examine
the clinical relevance of this mechanism. We first discuss the relationship
between mirror mechanism impairment and some core symptoms of
autism. We then outline the theoretical principles of neurorehabilitation
strategies based on the mirror mechanism. We conclude by examining
the relationship between the mirror mechanism and some features of the
environmental dependency syndromes.
KEYWORDS autism, environmental dependency syndromes, mirror neurons,
neurorehabilitation, utilization behavior
G Rizzolatti is Professor of Human Physiology and chairs the Department
of Neuroscience of the University of Parma, Parma, M Fabbri-Destro is
a Psychologist at the University of Ferrara, Ferrara, and the University
of Parma, and L Cattaneo is a Neurologist at the Center for Mind–Brain
Sciences (CIMeC) in Rovereto, Italy.
Correspondence
*Department of Neuroscience, University of Parma, 39 via Volturno, 43100 Parma, Italy
giacomo.rizzolatti@unipr.it
Received 15 October 2008 Accepted 13 November 2008
www.nature.com/clinicalpractice
doi:10.1038/ncpneuro0990
REVIEW CRITERIA
PubMed was searched using Entrez for articles published up to September 2008.
The search term was “mirror neuron” OR “mirror neurons” OR “mirror neuron
system” OR “mirror system”. Owing to limitations on the number of references,
we cited only articles that we judged to be most important from a theoretical or
clinical point of view.
REVIEW REVIEW

JANUARY 2009 VOL 5 NO 1 RIZZOLATTI ET AL. NATURE CLINICAL PRACTICE NEUROLOGY 25
www.nature.com/clinicalpractice/neuro
of these neurons is that they discharge both when
the monkey performs a motor act and when the
monkey, at rest, observes another individual (a
human being or another monkey) performing
a similar motor act (Figure 1). The degree of
similarity that is required between executed
and observed motor acts in order to trigger a
given mirror neuron varies from one neuron to
another. For most mirror neurons, however, the
relationship between the effective observed and
executed motor acts is based on their common
goal (e.g. grasping), regardless of how this goal is
achieved (e.g. using a two-finger or a whole-hand
prehension). Importantly, mirror neurons do not
discharge in response to the presentation of food
or other interesting objects.
Mirror neurons have also been described in the
PFG and anterior intraparietal areas of the infe-
rior parietal lobule (IPL; Figure 1). The general
properties of parietal mirror neurons seem to be
similar to those of mirror neurons in the premotor
cortex. Like the latter neurons, the parietal
mirror neurons code for the goals of motor acts
rather than the movements from which they
are constructed.8,9
The PFG and anterior intraparietal areas are
both connected with the F5 area and the cortex
of the superior temporal sulcus. Neurons in the
superior temporial sulcus have complex visual
properties, and some respond to the observation
of motor acts done by others.10,11 However, they
lack the motor properties that are defining fea-
tures of mirror neurons, and cannot, therefore,
be considered to be part of the mirror system.
The organization of the cortical motor
system
To understand the functional role of mirror
neurons in the premotor cortex and IPL, it is
necessary to frame them within the modern
conceptualization of the organization of the
cortical motor system. Clear evidence exists that
most of the parietal and frontal motor areas code
for motor acts (i.e. movements with a specific
goal) rather than mere active displacement of
body parts.12–18 Even in the primary motor
cortex, approximately 40% of neurons code for
motor acts.15,18
Studies in which the properties of single neu-
rons were studied in a naturalistic context have
been particularly important for establishing this
new view on cortical motor organization.12 These
studies showed that many neurons discharge
when a motor act (e.g. grasping) is performed
with effectors as different as the right hand, the
left hand, or the mouth. Furthermore, for the
vast majority of neurons, the same type of move-
ment (e.g. an index finger flexion) that is effec-
tive at triggering a neuron during one particular
motor act (e.g. grasping) is not effective during
another motor act (e.g. scratching). By using
motor acts as classification criteria, premotor
neurons have been subdivided into various cate-
gories such as ‘grasping’, ‘reaching’, ‘holding’, and
‘tearing’ neurons.
Recently, evidence was provided that both
inferior parietal and premotor (area F5) neurons
are organized in motor chains.19,20 Grasping
neurons recorded from these areas were tested in
two main conditions (Figure 2). In one condition,
a monkey reached and grasped a piece of food
located in front of it and brought it to its mouth.
In the other condition, the monkey reached
and grasped an object and placed it into a con-
tainer. The results showed that the majority of
the recorded neurons discharged with a different
intensity according to the final goal of the action
(e.g. eating or placing) in which the grasping
motor act was embedded (‘action-constrained’
neurons). This ‘chained’ organization seems to be
particularly well adapted for providing fluidity
to action execution. Individual neurons not only
code for specific motor acts, but, by virtue of
being wired to neurons that code for the subse-
quent motor acts, they facilitate the activity of
these downstream neurons, thereby ensuring
smooth execution of the intended action.
The functional role of the mirror neurons
The existence of a class of motor neurons that
discharge during the observation of actions done
by others is not as bizarre as it might initially
seem. While it is true that an action done by others
could be recognized by inference on the basis of
previous visual experience without involving the
motor system, visual perception per se does not
provide the observer with the experiential aspects
of the action. Furthermore, the mirror system
provides a particularly efficient way to establish
links between the observed action and other
actions with which it is functionally related.21
Evidence in favor of the notion that mirror
neurons mediate action understanding came
from experiments in which monkeys were not
allowed to see the actions performed by others,
but were given clues for understanding them.
In one series of experiments, monkeys were
presented with noisy motor acts (e.g. peanuts
REVIEW REVIEW

26 NATURE CLINICAL PRACTICE NEUROLOGY RIZZOLATTI ET AL. JANUARY 2009 VOL 5 NO 1
www.nature.com/clinicalpractice/neuro
breaking, tearing a piece of paper), which they
could either both see and hear or only hear.22
The researchers found that many mirror neurons
in area F5 responded to the sound of the motor
act, even when it was not visible.
In another series, F5 ‘grasping’ and ‘holding’
mirror neurons were tested both when the
monkey observed the experimenter grasping a
piece of food and when the monkey was pre-
vented from seeing the experimenter’s hand
movements by use of a black screen.23 Despite
the fact that the monkey could not see the hand–
object interaction (the visual triggering feature
of the recorded neurons) in the latter condition,
many mirror neurons in F5 were active in this situ-
ation. The neurons typically began to discharge at
the beginning of the hand-reaching movement,
indicating that the monkey had a representation
of the action performed behind the screen, even
when it could not see the performed motor act.
The activity of mirror neurons per se describes
only what is happening in the precise moment
of occurrence of the observed actions. There is,
however, a broader function of mirror neurons.
This function is related to the recent discovery
that most action-constrained neurons (see above)
have mirror properties and selectively discharge
when the monkey observes motor acts embedded
in a specific action (e.g. grasping for eating but
not grasping for placing; see Figure 2).19 The
activation of action-constrained mirror neurons,
therefore, codes not only ‘grasping’, but ‘grasping
for eating’ or ‘grasping for placing’. This coding
implies that when the monkey observes grasp-
ing done by another, it is able to predict, on the
basis of contextual cues (e.g. repetition, presence
of specific objects), what will be the individual’s
next motor act. In other words, the monkey is
able to understand the intentions behind the
observed motor act.
ncpneuro_2008_075f1.eps
PF
PF
PFG
PFG
PG
PG
F1
F2
F2
F4
F4
F5
F5c
F5a
F5p
F7
F7
AS
AS
AI
P
PF
FE
EF
F
AI
CLu
Lu
LIP
MIP
L
IO
STS
PE PE
PEc PEc
A
C
B
PEip VIP
AIP
IP
500 ms
Figure 1 A cytoarchitectonic map of the monkey cortex and an example of a mirror neuron. The upper
part of the figure shows the activity of a mirror neuron recorded from area F5. The neuron discharges both
when the monkey grasps an object (A) and when it observes the experimenter grasping the object (B).
(C) The cytoarchitectonic parcellation of the agranular frontal cortex and the parietal lobe. PE, PEc, PEip,
PF, PFG and PG are parietal areas. An enlargement of the frontal region (inset on the left) shows the
parcellation of area F5 into three parts: F5c, F5p and F5a. The mirror neurons are typically found in F5c.
The inset on the right shows the areas buried within the intraparietal sulcus. Abbreviations: AI, inferior
arcuate sulcus; AIP, anterior intraparietal area; AS, superior arcuate sulcus; C, central sulcus; FEF, frontal
eye field; IO, inferior occipital sulcus; IP, inferior precentral sulcus; L, lateral sulcus; LIP, lateral intraparietal
area; Lu, lunate sulcus; MIP, medial intraparietal area; P, principal sulcus; STS, superior temporal sulcus;
VIP, ventral intraparietal area. Permission obtained from Elsevier Ltd © Rizzolatti G and Fabbri-Destro M
(2008) Curr Opin Neurobiol 18: 179–184.
REVIEW
REVIEW

JANUARY 2009 VOL 5 NO 1 RIZZOLATTI ET AL. NATURE CLINICAL PRACTICE NEUROLOGY 27
www.nature.com/clinicalpractice/neuro
Figure 2 Action-constrained neurons in the monkey IPL. (A) Apparatus and
paradigm used for a task designed to demonstrate action-constrained neurons.
The monkey starts from the same position in all trials, reaches for an object (1)
and brings it to the mouth (2a) or places it into a container (2b). (B) Activity of three
IPL neurons during the motor task in conditions 2a (grasp to place) and 2b (grasp
to eat). Raster histograms are synchronized with the moment when the monkey
touched the object to be grasped. Unit 67 fires during grasping to eat and not
during grasping to place. Unit 161 is selective for grasping to place. Unit 158
does not show any task preference. (C) Visual responses of IPL mirror neurons
during the observation of grasping to eat and grasping to place performed by
an experimenter. Unit 87 is selective for grasping to eat, unit 39 is selective for
grasping to place and unit 80 does not display any task preference. Abbreviation:
IPL, inferior parietal lobule. Permission obtained from American Association for
the Advancement of Science © Fogassi L et al. (2005) Science 308: 662–667.
THE MIRROR SYSTEM IN HUMANS
Understanding of goals and intentions
A large number of studies based on noninvasive
electrophysiological (e.g. EEG, magnetoencephalo-
graphy [MEG]) or brain imaging (e.g. PET,
functional MRI [fMRI]) techniques have demon-
strated the existence of the mirror mechanism in
humans.8,9 Brain imaging studies have enabled
the mirror areas to be located. These studies
showed that the observation of transitive actions
done by others results in an increase in blood
oxygen level-dependent (BOLD) signal not only
in visual areas, but also in the IPL and the ventral
premotor cortex, as well as the caudal part of the
inferior frontal gyrus (IFG). These latter three
areas have motor properties and closely corres-
pond to the areas that contain mirror neurons in
the monkey (Figure 3).
Both the premotor and the parietal areas of
the human mirror system show a somatotopic
organization.24 Observation of motor acts done
with the leg, hand or mouth activates the pre-
central gyrus and the pars opercularis of the IFG
in a medial-to-lateral direction, as in the classical
homunculus model of Penfield25 and Woolsey.26
In the IPL, mouth motor acts are represented ros-
trally, hand and arm motor acts are represented
caudally, and leg motor acts are represented even
more caudally and dorsally, extending into the
superior parietal lobule.
Most studies on the mirror mechanism in
humans have investigated transitive movements
such as grasping. In a recent fMRI study in which
volunteers were asked to observe video clips
showing a hand transport movement without
an effector–object interaction, activations were
found in the dorsal premotor cortex and also
in the superior parietal lobule, with the activa-
tion extending into the intraparietal sulcus.27
This finding indicates that the human brain is
endowed with a reaching mirror mechanism
that is anatomically separated from the mirror
mechanism that codes for the distal motor act.
As in the monkey, the parietal and frontal
mirror areas in humans code mostly for the
goals of motor acts. Gazzola et al.28 instructed
volunteers to observe either a human or a robot
arm grasping objects. In spite of differences in
shape and kinematics between the human and
robot arms, the parietofrontal mirror network
was activated in both conditions. Further evi-
dence in favor of goal coding was obtained in an
fMRI study based on repetition suppression29—a
technique that exploits the trial-by-trial reduction
of a physiological response to repeated stimuli.
The results showed that repeated presentation of
the same goal caused suppression of the hemo-
dynamic response in the left intraparietal sulcus,
but this region was not sensitive to the trajectory
of the agent’s hand.
ncpneuro_2008_075f2.eps
Unit 67
Grasp
to eat
spk/s
100
0
spk/s
100
1
s
0
Grasp
to place
Unit 161
Motor responses of mirror neurons
Unit 158
A
B
C
Unit 87
spk/s
100 100 150
1
s
2b 2a
Unit 39
Visual responses of mirror neurons
Unit 80
1
Grasp
to eat
Grasp
to place
REVIEW
REVIEW

28 NATURE CLINICAL PRACTICE NEUROLOGY RIZZOLATTI ET AL. JANUARY 2009 VOL 5 NO 1
www.nature.com/clinicalpractice/neuro
The study of aplasic individuals born without
arms and hands provided further evidence in
favor of a goal-coding mirror mechanism.30
During MRI scanning, two aplasic individuals
and a group of nonaplasic volunteers were
instructed to watch videos showing hand actions.
All participants also made actions with their feet,
mouths, and, in the case of the nonaplasic volun-
teers, hands. The results showed that in aplasic
individuals, the observation of hand motor acts,
which they had never themselves performed, acti-
vated the mirror areas. The communality of goals
between the never-executed hand motor acts and
those performed with the mouth and feet was the
most probable explanation for this activation.
Growing evidence exists that, in addition to goal
coding, the human mirror mechanism has a role
in the ability to understand the intentions behind
the actions of others. In an fMRI study, volunteers
observed motor acts (e.g. grasping a cup) embed-
ded in specific contexts (a condition in which the
agent’s intention could be easily understood)
or devoid of context (a condition in which the
agent’s intention was ambiguous).31 The results
showed that the mirror network was active in
both conditions. However, the understanding of
intention produced a stronger signal increase in
the caudal IFG of the right hemisphere.
The importance of the mirror system in under-
standing the intentions of others was confirmed
by a repetition-suppression fMRI experiment.32
Participants were asked to observe repeated
movies showing either the same movement or the
same action outcome regardless of the executed
movement. The result showed activity suppres-
sion in the right IPL and the right IFG when the
outcome was the same.
Movement, emotions and language
As we have discussed, the mirror mechanism
located in the parietal and frontal areas codes
mostly for the goals of observed motor acts.
However, studies that involved transcranial
magnetic stimulation (TMS) have shown that
the human motor system also responds to the
observation of movements devoid of a goal.33,34
This ‘movement mirror mechanism’ seems to
be extremely sensitive to movement kinematics.
Dayan et al.35 studied brain responses to the
observation of curved hand movements that either
obeyed or disobeyed the law—known as the 2/3-
power law—that describes the coupling between
movement curvature and velocity. Mirror hand
areas were more active during the observation of
movements that obeyed this law than during other
types of motion.
The mirror mechanism is located not only in
centers that mediate voluntary movement, but
also in cortical areas that mediate visceromotor
emotion-related behaviors.36,37 Brain imaging
studies showed that when an individual feels or
observes emotions in others caused by disgusting
stimuli or stimuli representing pain, there is
activation in two structures: the cingulate cortex
and the insula. Interestingly, the same voxels are
activated in these two structures in both ‘feeling’
and ‘observing’ conditions. This finding strongly
suggests that feeling emotions and recognizing
them in others are mediated by the same neural
substrate.
It should be made clear that the anterior insula,
where the aforementioned activations were found,
has a dysgranular–agranular structure,38 and is,
therefore, cytoarchitectonically similar to motor
areas. Electrical stimulation of the insula in the
monkey produces movements of various body
ncpneuro_2008_075f3.eps
37 19 18
17
18
19
7b
7a
64
8
9
3125
10
IF
45
FEF
SP
C
PMv (F4)
PMv (F5c)
PMv (F5p)
IP
PrePMd
(F7)
PMd
(F2)
44
4
SF
(F5a)
11
38
20
21
22
39
42
41
52
43
47
45
46 40
44
44
Figure 3 The parietofrontal mirror system in humans. Lateral view of the human
cerebral cortex showing Brodmann cytoarchitectonic subdivision. The areas
in yellow correspond to areas that respond to the observation and execution
of hand motor acts. The left-hand panel shows an enlarged view of the frontal
lobe. The possible homology between monkey and human premotor cortex is
indicated by arrows. Note that in monkeys area F5 consists of three subareas:
F5c, F5p and F5a. Area 44 is considered to be the most likely human homolog of
area F5. Abbreviations: C, central sulcus; FEF, frontal eye field; IF, inferior frontal
sulcus; IP, inferior precentral sulcus; PMd, dorsal premotor cortex; PMv, ventral
premotor cortex; PrePMd, pre-dorsal premotor cortex; SF, superior frontal
sulcus; SP, upper part of the superior precentral sulcus. Permission obtained
from Elsevier Ltd © Rizzolatti G and Fabbri-Destro M (2008) Curr Opin Neurobiol
18: 179–184.
REVIEW
REVIEW

JANUARY 2009 VOL 5 NO 1 RIZZOLATTI ET AL. NATURE CLINICAL PRACTICE NEUROLOGY 29
www.nature.com/clinicalpractice/neuro
parts, accompanied by a variety of visceromotor
responses.39–40 Similar effects have also been
described in humans.41,42 It is, therefore, appro-
priate to define these structures as ‘mirror areas’
in which the motor response includes a visceral
component.
In humans, the mirror mechanism is also
located in Broca’s area, which is involved in lan-
guage processing and speech production. Evidence
for a mechanism that translates heard phonemes
into the motor programs necessary to produce
them has been provided by TMS experiments.43
The mouth motor field was stimulated in volun-
teers while they heard words containing pho-
nemes requiring tongue movements (e.g. “birra”)
or not requiring tongue movements (e.g. “baffo”).
Motor evoked potentials recorded from the
tongue muscles increased with the presentation
of verbal material containing a double ‘r’ relative
to those containing a double ‘f’.
THE MIRROR SYSTEM IN NEUROLOGY
The mirror system and autism
Autism spectrum disorder (ASD) is a hetero-
geneous developmental syndrome characterized
by a marked impairment in social interaction
and communication.44 Communication deficits
include disturbances in most domains of language
and are not limited to its pragmatic aspects.45
Impairment in the domains of affective links
and emotion recognition is another important
component of ASD.46 A restricted repertoire of
activity and interests, repetitive motion, and hyper-
sensitivity to certain sounds are other symptoms
that are often present in ASD.
Autism affects a variety of nervous structures,
from the cerebral cortex to the cerebellum and
brainstem.47 However, in a context of a broader
neurodevelopmental deficit, a set of ASD symp-
toms (impairment in communication, language
and emotion, as well as in the capacity to under-
stand others) seems to match the functions medi-
ated by the mirror mechanism. A hypothesis has,
therefore, been advanced that this set of deficits
might depend on an impairment of the mirror
mechanism,48,49 and there is growing evidence
to support this view.50–53
One classical EEG observation is that mu
rhythm (an EEG rhythm recorded from the motor
cortical areas) is blocked when a person makes a
voluntary movement. This rhythm is also sup-
pressed when a person observes another person
performing a movement. Oberman et al.50 used
this phenomenon to test the mirror mechanism
in children with ASD. The results showed that
although individuals with ASD exhibited a sup-
pression of mu rhythm during voluntary move-
ments, this suppression was absent when they
watched some one else performing the move-
ment (Figure 4). Martineau et al.54 have reported
similar observations.
Oberman et al.55 recently reported an inter-
esting observation concerning the mirror system
of children with ASD. The authors investigated
how familiarity between an observing indivi-
dual and a person performing a movement
modulates the entity of mu rhythm suppression.
Typically developing children and children with
ASD viewed video clips showing the hand of a
stranger performing a grasping action, the hand
of a child’s guardian or sibling performing the
same action, and the participant’s own hand per-
forming the action. The study revealed that mu
suppression depended on the familiarity of the
observer with the agent, and that children with
ASD showed mu suppression when a familiar
person performed the action but not when it was
performed by an unfamiliar person.
An fMRI study has provided strong evidence
in favor of a deficit of the mirror mechanism in
ASD. High-functioning children with ASD and
matched controls were scanned while they imi-
tated and observed emotional expressions. The
results showed a markedly weaker activation
in the IFG in children with ASD than in typi-
cally developing children. Most interestingly, the
degree of activation was inversely related to
symptom severity.53
Impaired motor facilitation during action
observation has been reported in individuals with
ASD by use of TMS.52 Furthermore, unlike typi-
cally developing individuals, children with ASD
tend not to imitate other individuals in a mirror
fashion when viewing them face-to-face.56 This
imitation peculiarity is probably attributable to
a deficit in the ability of the mirror mechanism
to superimpose another person’s movements on
one’s own.
Deficits in the mirror mechanism in ASD have
also been addressed from another perspective.57
Typically developing children and children with
ASD were tested while they observed an experi-
menter either grasping a piece of food for eating
or grasping a piece of paper to place it into a con-
tainer (Figure 5). The EMG activity of the mylo-
hyoid muscle, which is involved in opening of
the mouth, was recorded. The results showed that
observation of food grasping produced activation
REVIEW
REVIEW

30 NATURE CLINICAL PRACTICE NEUROLOGY RIZZOLATTI ET AL. JANUARY 2009 VOL 5 NO 1
www.nature.com/clinicalpractice/neuro
of the mylohyoid muscle in typically developing
children, but not in children with ASD. In other
words, whereas the observation of an action done
by another individual intruded into the motor
system of a typically developing observer, this
intrusion was lacking in children with ASD. This
finding indicates that, in this disorder, the mirror
system is silent during action observation, and
that the immediate, experiential understanding
of the intentions of others is absent.
Both children with ASD and typically develop-
ing children were also asked to perform the two
actions described above (grasp to eat and grasp to
place) while the EMG activity of the mylohyoid
muscle was recorded.57 In typically developing
children, the muscle became active as soon they
moved the arm to reach the food. By contrast,
no mylohyoid muscle activation was observed
during food reaching and grasping in children
with ASD; activation of the muscle was evident
only when these children brought the food to
their mouths. These data indicate that children
with ASD are not only unable to organize their
own motor acts into a unitary action charac-
terized by a specific intention, but that they
also show a deficit in the mirror mechanism, as
reflected in the absence of motor activation of
the muscles involved in an observed action.
These findings show an apparent contradiction
between the cognitive capacities of children with
ASD to report the purpose of an experimenter’s
action and their lack of motor resonance with the
action. To clarify this incongruity, a further experi-
ment was performed in which typically develop-
ing children and children with ASD observed an
actor performing goal-directed motor acts and
were asked to report what the actor was doing
and why he was doing it (Rizzolatti G et al.,
unpublished data). These tasks test two different
abilities: the ability to recognize a motor act (e.g.
grasping an object) and the ability to understand
the intention behind it (e.g. grasping to eat). The
results showed that both typically developing
children and children with ASD were able to
recognize what the actor was doing, but children
with ASD failed to recognize why the act was
being performed. Children with ASD systema-
tically attributed to the actor the intention that
could be derived by the semantics of the object—
for example, an intention to cut when scissors
were shown—regardless of how the object was
grasped. This finding indicates that children
ncpneuro_2008_075f4.eps
0.5 –
0 –
–0.5 –
–1 –
–1.5 –
Cz
Ball
C4C3 Cz
Hand
C4C3 Cz
Move
C4C3
Condition
*** *** **
*** ** ***
B Autism spectrum disordersA Controls
Log (condition baseline)
Cz
Ball
C4C3 Cz
Hand
C4C3 Cz
Move
C4C3
Condition
** ***
Figure 4 Absence of mirror EEG responses in autism. The charts show suppression of the mu rhythm
in controls (A) and patients with autism spectrum disorder (B) during observation of movement of an
inanimate object (ball, pale green) or movements made with a hand (hand, green), and during active
hand movements made by the individual from whom recordings were being taken (move, red). The bars
represent the amount of mu activity in central scalp locations; C3, Cz and C4 refer to scalp coordinates
of the 10/20 EEG system. Significant suppression of this activity, indicated by asterisks, is present for
the hand observation condition only in controls, showing that patients with autism spectrum disorder
fail to respond in a standard way to the observation of other people’s actions. Permission obtained from
Elsevier Ltd © Oberman LM et al. (2005) Brain Res Cogn Brain Res 24: 190–198.
REVIEW
REVIEW

JANUARY 2009 VOL 5 NO 1 RIZZOLATTI ET AL. NATURE CLINICAL PRACTICE NEUROLOGY 31
www.nature.com/clinicalpractice/neuro
with ASD interpret the behavior of others on the
basis of the standard use of objects rather than
the actual behavior of a person performing a
task. Children with ASD, therefore, seem to lack
the ability to read the intentions of others on the
basis of behavior.
ncpneuro_2008_075f5.eps
A
B Typically developing children
C Children with ASD
0.08 –
0.07 –
0.06 –
0.05 –
0.04 –
0.03 –
0.02 –
0.01 –
Eat
Place
Rectified mylohydroid EMG
–
–
–
–
–
–
–
–
Execution Observation
Execution Observation
0.08 –
0.07 –
0.06 –
0.05 –
0.04 –
0.03 –
0.02 –
0.01 –
–2.0 –1.5 –1.0 –0.5 0.0 0.5
Time (s)
Rectified mylohydroid EMG
1.0 1.5 2.0
–
–
–
–
–
–
–
–
–2.0 –1.5 –1.0 –0.5 0.0 0.5
Time (s)
1.0 1.5 2.0
Figure 5 Motor behavior in typically developing children and children with ASD. This experiment was
designed to assess whether an action-constrained motor organization is present in typically developing
children and children with ASD.57 (A) Schematic representation of the tasks. The individual reaches for an
item on a plate and either brings it to their mouth or puts it into a container placed on their shoulder. Time
course for typically developing children (B) and children with ASD (C) of the rectified electromyographic
activity of mouth-opening muscles during the execution (left side) and observation (right side) of the
‘bringing-to-the-mouth’ action (red line) and of the ‘placing’ action (blue line). All curves are aligned with
the moment of object lifting from the touch-sensitive plate (time = 0). The results demonstrate a lack of
anticipatory motor activity during execution and a lack of mirror motor activation during observation of a
given action in children with ASD. Abbreviations: ASD, autism spectrum disorder; EMG, electromyography.
REVIEW
REVIEW

32 NATURE CLINICAL PRACTICE NEUROLOGY RIZZOLATTI ET AL. JANUARY 2009 VOL 5 NO 1
www.nature.com/clinicalpractice/neuro
The mirror mechanism and motor
rehabilitation
As well as having a role in action understanding,
the mirror mechanism also modulates the motor
behavior of the observer. This function forms
the basis for the imitation of simple motor acts58
and for learning through imitation.59 Particularly
interesting from a clinical point of view was the
demonstration that the mirror mechanism is
involved in the building of motor memories. The
most convincing evidence for such a role came
from studies by Stefan et al.60,61 that involved
TMS. The authors showed that when partici-
pants simultaneously performed and observed
congruent movements, the learning of these
movements was potentiated with respect to
learning through motor training alone. These
findings indicate that the coupling of observation
and execution strongly facilitates the formation
of motor memories.
Could this mechanism be exploited for motor
rehabilitation? Many current behavioral neuro-
rehabilitation techniques use strategies that
induce long-term plasticity in the motor cortex
either by depressing activity on the unaffected side
or by potentiating activity on the affected
side.62 The possibility that plasticity might be
induced in the motor cortex by coupling action
observation and execution represents the theo-
retical basis of a recent study that examined
the effect of an 18-day cycle of active motor
training with the paretic limb in two groups
of patients with chronic stabilized stroke in
the middle cerebral artery territory.63 The test
group was required to perform hand motor acts
prompted by movies showing similar motor
acts, whereas the control group performed the
same motor training without any visual cues.
Functional assessment of the upper limb showed
a significant improvement in the test group
relative to the control group.
The mirror mechanism probably also forms
the neurophysiological basis for ‘mirror therapy’
(the word ‘mirror’ being used here in its literal
sense), which has been shown to improve upper-
limb function in patients with stroke.64,65
In mirror-therapy protocols, the patients are
required to perform movements with their
nonparetic hand while watching the hand and
its reflection in a parasagittal mirror. This proce-
dure gives a visual illusion of movement of the
paretic hand. The generation of cortical plas-
ticity and the consequent rehabilitative results
strongly suggest a role in patient improvement
for a mechanism that matches seen and executed
actions, thereby implicating the mirror mecha-
nism in this process.66
Deficits in the control of mirror mechanisms
Clinical observations have shown that frontal
lesions can cause a series of disturbances charac-
terized by the appearance of forced motor behavior
triggered by external stimuli.67 Among these
manifestations, imitation behavior is particularly
interesting in relation to the mirror mechanism.
The main feature of this syndrome is the sponta-
neous imitation of motor acts done by others, and
it is considered to be part of the so-called ‘environ-
mental dependency syndrome’.68 The condi-
tion arises from unilateral, or, more frequently,
bilateral
prefrontal lesions.68,69 Imitation
behavior is generally attributed to an imbalance
between exogenously and endogenously deter-
mined behaviors. The observation of actions done
by others leads to the coding of potential motor
acts in the parietal and premotor mirror areas by
means of the mirror mechanism. These poten-
tial motor acts typically do not determine overt
movements in the healthy adult brain because the
manifestation of these acts is suppressed by the
frontal lobe. Damage to this lobe would destroy
this control mechanism, thereby transforming the
potential motor acts into actual motor behavior.
In view of the temporal latency between observa-
tion and imitation that patients often show, an
additional mechanism could be also involved, but
the essence of the phenomenon seems to depend
on a release of potential motor acts.
Echopraxia is a term that describes forced and
uncritical imitation of behaviors. The exogenously
triggered behavior is sustained through endo-
genous mechanisms, resulting in its perseveration.
In view of the simplicity of the imitated behaviors,
combined with the total lack of criticism of the
patient to the imitated behavior, echopraxia is
perceived as a distinct disorder from imitation
behavior. Echopraxia can arise in the context of
basal ganglia dysfunction, as well as after frontal
lobe damage. It is probable, however, that in both
cases the mechanism that underlies echopraxia is
a disinhibition of the mirror areas through loss
of suppression by the frontal lobe.70
CONCLUSIONS AND FUTURE PROSPECTS
The discovery of the mirror mechanism radi-
cally changed our views on how individuals
understand actions, intentions and emotions.
The identification of this mechanism has had
REVIEW
REVIEW

JANUARY 2009 VOL 5 NO 1 RIZZOLATTI ET AL. NATURE CLINICAL PRACTICE NEUROLOGY 33
www.nature.com/clinicalpractice/neuro
a profound impact on a variety of disciplines,
ranging from cognitive neurosciences to soci-
ology and philosophy. Until recently, this
discovery had influenced clinical research to a
much lesser degree. However, it has now provided
deeper insights into the interpretation of certain
neurological syndromes, such as the environ-
mental dependency syndrome, and has provided
a new theoretical basis for establishing rehabili-
tation techniques in patients with motor deficits
following stroke.
Autism is one condition in which the discov-
ery of the mirror neuron mechanism could have
important practical implications in the future.
Recent experimental data suggest that indivi-
duals with ASD have a deficit in representing
goal-directed actions, both when the actions are
performed and when they are observed. Children
with ASD, therefore, show impairments in
organizing their own motor acts according to an
action goal, as well as in using this motor mecha-
nism to understand the intentions of others.
This new view on ASD could be used to establish
new rehabilitation strategies based on a motor
approach. The rationale of such an approach is
that if the motor knowledge of individuals with
ASD is improved, their social knowledge and
behavior would also be enhanced.
KEY POINTS
■ The mirror mechanism is a neural system that
unifies action perception and action execution
■ The mirror mechanism is organized into two
main cortical networks, the first being formed
by the parietal lobe and premotor cortices,
and the second by the insula and anterior
cingulate cortex
■ The role of the mirror mechanism is to provide
a direct understanding of the actions and
emotions of others without higher order
cognitive mediation
■ Limited development of the mirror mechanism
seems to determine some of the core aspects
of autism spectrum disorders
■ The recently demonstrated link between limited
development of the mirror mechanism and that
of some aspects of the motor system suggests
that rehabilitation in children with austism
spectrum disorder should take into account
both motor and cognitive strategies
■ The use of action-observation-based protocols
could represent a new rehabilitation strategy to
treat motor deficits after stroke
References
1 Jacob P and Jeannerod M (2005) The motor theory
of social cognition: a critique. Trends Cogn Sci 9:
21–25
2 Csibra G and Gergely G (2007) ‘Obsessed with
goals’: functions and mechanisms of teleological
interpretation of actions in humans. Acta Psychol
(Amst) 124: 60–78
3 Frith CD (2007) The social brain? Philos Trans R Soc
Lond B Biol Sci 362: 671–678
4 Damasio A and Meyer K (2008) Behind the looking-
glass. Nature 10: 167–168
5 Brass M et al. (2007) Investigating action
understanding: inferential processes versus action
simulation. Curr Biol 17: 2117–2121
6 Keysers C and Gazzola V (2007) Integrating simulation
and theory of mind: from self to social cognition.
Trends Cogn Sci 11: 194–196
7 Liepelt R et al. (2008) How do we infer others’ goals
from non-stereotypic actions? The outcome of
context-sensitive inferential processing in right inferior
parietal and posterior temporal cortex. Neuroimage 43:
784–792
8 Rizzolatti G and Craighero L (2004) The mirror-neuron
system. Annu Rev Neurosci 27: 169–192
9 Fabbri-Destro M and Rizzolatti G (2008) The mirror
system in monkeys and humans. Physiology 23:
171–179
10 Perrett DI et al. (1989) Frameworks of analysis for the
neural representation of animate objects and actions.
J Exp Biol 146: 87–113
11 Jellema T et al. (2000) Neural representation for the
perception of the intentionality of action. Brain Cogn
442: 280–302
12 Rizzolatti G et al. (1988) Functional organization of
inferior area 6 in the macaque monkey: II: area F5 and
the control of distal movements. Exp Brain Res 71:
491–507
13 Crutcher MD and Alexander GE (1990) Movement-
related neuronal activity selectively coding either
direction or muscle pattern in three motor areas of the
monkey. J Neurophysiol 64: 151–163
14 Alexander GE and Crutcher MD (1990) Neural
representations of the target (goal) of visually guided
arm movements in three motor areas of the monkey.
J Neurophysiol 64: 164–178
15 Kakei S et al. (1999) Muscle and movement
representations in the primary motor cortex. Science
285: 2136–2139
16 Kakei S et al. (2001) Direction of action is represented
in the ventral premotor cortex. Nat Neurosci 4:
1020–1025
17 Hoshi E and Tanji J (2000) Integration of target and
body-part information in the premotor cortex when
planning action. Nature 408: 466–470
18 Umiltà MA et al. (2008) When pliers become fingers in
the monkey motor system. Proc Natl Acad Sci USA
105: 2209–2213
19 Fogassi L et al. (2005) Parietal lobe: from action
organization to intention understanding. Science 308:
662–667
20 Fogassi L et al. (2007) Time course of neuronal activity
reflecting the final goal of observed and executed
action sequences in monkey parietal and premotor
cortex [abstract #636.4]. Presented at the 37th Annual
Meeting of the Society for Neuroscience: 2007
November 3–7, San Diego, CA, USA
21 Rizzolatti G et al. (2001) Neurophysiological
mechanisms underlying the understanding and
imitation of action. Nat Rev Neurosci 2: 661–670
22 Kohler E et al. (2002) Hearing sounds, understanding
actions: action representation in mirror neurons.
Science 297: 846–848
REVIEW
REVIEW

34 NATURE CLINICAL PRACTICE NEUROLOGY RIZZOLATTI ET AL. JANUARY 2009 VOL 5 NO 1
www.nature.com/clinicalpractice/neuro
23 Umiltà MA et al. (2001) “I know what you are doing”:
a neurophysiological study. Neuron 32: 91–101
24 Buccino G et al. (2001) Action observation activates
premotor and parietal areas in a somatotopic manner:
an fMRI study. Eur J Neurosci 13: 400–404
25 Penfield W and Rasmussen T (1950) The Cerebral
Cortex of Man: a Clinical Study of Localization of
Function. New York: Macmillan
26 Woolsey CN et al. (1979) Localization of somatic
sensory and motor areas of human sensory cortex as
determined by direct recording of evoked potentials
and electrical stimulation. J Neurosurg 51: 476–506
27 Filimon F et al. (2007) Human cortical representations
for reaching: mirror neurons for execution, observation,
and imagery. Neuroimage 37: 1315–1328
28 Gazzola V et al. (2007) The anthropomorphic brain:
the mirror neuron system responds to human and
robotic actions. Neuroimage 35: 1674–1684
29 Hamilton AF and Grafton ST (2006) Goal
representation in human anterior intraparietal sulcus.
J Neurosci 26: 1133–1137
30 Gazzola V et al. (2007) Aplasics born without hands
mirror the goal of hand actions with their feet. Curr Biol
17: 1235–1240
31 Iacoboni M et al. (2005) Grasping the intentions of
others with one’s own mirror neuron system. PLoS Biol
3: 529–535
32 Hamilton AF and Grafton ST (2008) Action outcomes
are represented in human inferior frontoparietal cortex.
Cereb Cortex 18: 1160–1168
33 Fadiga L et al. (1995) Motor facilitation during action
observation: a magnetic stimulation study.
J Neurophysiol 73: 2608–2611
34 Strafella AP and Paus T (2000) Modulation of cortical
excitability during action observation: a transcranial
magnetic stimulation study. Neuroreport 11: 2289–2292
35 Dayan E et al. (2007) Neural representations of
kinematic laws of motion: evidence for action-
perception coupling. Proc Natl Acad Sci USA 104:
20582–20587
36 Gallese V et al. (2004) A unifying view of the basis of
social cognition. Trends Cogn Sci 8: 396–403
37 Singer T (2006) The neuronal basis and ontogeny of
empathy and mind reading: review of literature and
implications for future research. Neurosci Biobehav
Rev 30: 855–863
38 Mesulam MM and Mufson EJ (1982) Insula of the old
world monkey: I: architectonics in the insulo-orbito-
temporal component of the paralimbic brain. J Comp
Neurol 212: 1–22
39 Frontera JG (1956) Some results obtained by electrical
stimulation of the cortex of the island of Reil in the
brain of the monkey (Macaca mulatta). J Comp Neurol
105: 365–394
40 Showers MJ and Laurer EW (1961) Somatovisceral
motor patterns in the insula. J Comp Neurol 117:
107–115
41 Penfield W and Faulk ME (1955) The insula; further
observations on its function. Brain 78: 445–470
42 Krolak-Salmon P et al. (2003) An attention modulated
response to disgust in human ventral anterior insula.
Ann Neurol 53: 446–453
43 Fadiga L et al. (2002) Speech listening specifically
modulates the excitability of tongue muscles: a TMS
study. Eur J Neurosci 15: 399–402
44 Kanner L (1943) Autistic disturbances of affective
contact. Nerv Child 2: 217–250
45 Tuchman R and Rapin I (2002) Epilepsy in autism.
Lancet Neurol 1: 352–358
46 Hobson RP (1993) Autism and the Development
of Mind. Hillsdale, NJ: Erlbaum
47 Minshew NJ and Williams DL (2007) The new
neurobiology of autism: cortex, connectivity, and
neuronal organization. Arch Neurol 64: 945–950
48 Altschuler EL et al. (1997) Mu wave blocking by
observation of movement and its possible use as a
tool to study theory of other minds [abstract #67.23].
Presented at the 30th Annual Meeting of the Society
for Neuroscience: 2000 November 4–9, New Orleans,
LA, USA
49 Williams JHG et al. (2001) Imitation, mirror neurons and
autism. Neurosci Biobehav Rev 25: 287–295
50 Oberman LM et al. (2005) EEG evidence for mirror
neuron dysfunction in autism spectrum disorders.
Brain Res Cogn Brain Res 24: 190–198
51 Nishitani N et al. (2004) Abnormal imitation-related
cortical activation sequences in Asperger’s syndrome.
Ann Neurol 55: 558–562
52 Theoret H et al. (2005) Impaired motor facilitation
during action observation in individuals with autism
spectrum disorder. Curr Biol 15: R84–R85
53 Dapretto M et al. (2006) Understanding emotions in
others: mirror neuron dysfunction in children with
autism spectrum disorders. Nat Neurosci 9: 28–30
54 Martineau J et al. (2008) Impaired cortical activation in
autistic children: is the mirror neuron system involved?
Int J Psychophysiol 68: 35–40
55 Oberman LM et al. (2008) Modulation of mu
suppression in children with autism spectrum
disorders in response to familiar or unfamiliar stimuli:
the mirror neuron hypothesis. Neuropsychologia 46:
1558–1565
56 Avikainen S et al. (2003) Impaired mirror-image
imitation in Asperger and high-funtioning autistic
subjects. Curr Biol 13: 339–341
57 Cattaneo L et al. (2007) Impairment of actions chains in
autism and its possible role in intention understanding.
Proc Natl Acad Sci USA 104: 17825–17830
58 Iacoboni M et al. (1999) Cortical mechanisms of human
imitation. Science 286: 2526–2528
59 Buccino G et al. (2004) Neural circuits underlying
imitation learning of hand actions: an event-related
fMRI study. Neuron 42: 323–334
60 Stefan K et al. (2005) Formation of a motor memory by
action observation. J Neurosci 25: 9339–9346
61 Stefan K et al. (2008) Concurrent action observation
modulates practice-induced motor memory formation.
Eur J Neurosci 27: 730–738
62 Dobkin BH (2008) Training and exercise to drive
poststroke recovery. Nat Clin Pract Neurol 4: 76–85
63 Ertelt D et al. (2008) Action observation has a positive
impact on rehabilitation of motor deficits after stroke.
Neuroimage 36 (Suppl 2): T164–T173
64 Altschuler EL et al. (1999) Rehabilitation of hemiparesis
after stroke with a mirror. Lancet 353: 2035–2036
65 Yavuzer G et al. (2008) Mirror therapy improves hand
function in subacute stroke: a randomized controlled
trial. Arch Phys Med Rehabil 89: 393–398
66 Moseley GL et al. (2008) Is mirror therapy all it is
cracked up to be? Current evidence and future
directions. Pain 138: 7–10
67 Ropper AH and Brown RH (2005) Adams and Victor’s
Principles of Neurology. New York: McGraw-Hill
68 Lhermitte F (1986) Human autonomy and the frontal
lobes: part II: patient behavior in complex and social
situations: the “environmental dependency syndrome”.
Ann Neurol 19: 335–343
69 De Renzi E et al. (1996) Imitation and utilization
behaviour. J Neurol Neurosurg Psychiatry 61: 396–400
70 Ghika J et al. (1995) Environment-driven responses
in progressive supranuclear palsy. J Neurol Sci 130:
104–111
Acknowledgments
This study was supported
by European Union Contract
012738, Neurocom, by
PRIN 2006 to GR and by
Fondazione Monte Parma.
MF-D was supported by
Fondazione Cassa
di Risparmio di Ferrara.
Competing interests
The authors declared no
competing interests.
REVIEW