REVIEWS

The mirror mechanism: a basic

principle of brain function
Giacomo Rizzolatti1,2 and Corrado Sinigaglia3,4
Abstract | The mirror mechanism is a basic brain mechanism that transforms sensory
representations of others’ behaviour into one’s own motor or visceromotor representations
concerning that behaviour. According to its location in the brain, it may fulfil a range of
cognitive functions, including action and emotion understanding. In each case, it may enable
a route to knowledge of others’ behaviour, which mainly depends on one’s own motor or
visceromotor representations.

Mirror neurons are a distinct class of neurons that dis- is that they may provide a route to knowledge of others,
Mirror neurons
Neurons that discharge both charge both when individuals perform a given motor one which can be taken just by capitalizing on one’s own
when an individual performs act and when individuals observe another person per- motor or visceromotor representations.
a given behaviour and when an forming a motor act with a similar goal. Mirror neu-
individual observes another rons were first discovered in the ventral premotor cortex Mirroring actions
person performing the same or
a similar behaviour; these
(PMv) of the macaque monkey (area F5)1–3. Neurons Action execution and observation
neurons are found in many with mirror properties have subsequently been found in After the discovery of mirror neurons in area F5 of the
brain cortical areas of monkeys many brain cortical areas of monkeys and other species, macaque monkey, it was found that a network of cortical
and other species, including including humans4,5. areas becomes active during action observation in both
humans, marmosets and birds.
The discovery that a large number of cortical areas monkeys and humans. Functional MRI (fMRI) studies
that are involved in the production of certain motor in monkeys6,7 showed that observing action activates both
behaviours selectively respond to those behaviours the dorsal and the ventral banks of the superior temporal
irrespective of whether they are being performed or sulcus (STS), which contains high-order visual areas that
observed indicates that the mirror mechanism, far from are typically involved in motion processing. The dorsal
being a specific characteristic of the premotor cortex, is bank projects to the parietal motor area PFG, whereas
a basic principle of brain functioning. This statement the ventral bank projects to the parietal motor area AIP
becomes less surprising once it is acknowledged that (anterior intraparietal area). Both PFG and AIP pro-
the brain acts, first and foremost, as a planning and ject to different subsections of area F5 (FIG. 1). Although
control system for organisms whose main job is explor- single-neuron studies recorded neurons with mirror
ing their surrounding world and facing its challenges properties from areas F5, AIP and PFG, neurons with
and that are able to catch positive opportunities and mirror properties have not been reported in STS areas8.
1
University of Parma, escape threats. More recently, pyramidal tract neurons (PTNs) orig-
Department of Neuroscience, The aim of this Review is to shed new light on the inating from area F5 were found to respond to action
via Volturno 39,
I-43100 Parma, Italy.
functional properties of the mirror mechanism and its observation9. Interestingly, a subset of PTNs, which
2
Brain Center for Motor and role in cognition. We focus on two main domains: action fired during action execution, showed a suppression of
Social Cognition, Italian and emotion. We show that observing others’ actions and their discharge during action observation. This inhibi-
Institute of Technology, emotions recruits different brain networks, each of which tory effect might have a role in preventing the occurrence
I-43100 Parma, Italy.
transforms the sensory information concerning others’ of movement during action observation. Similar results
3
University of Milan,
Department of Philosophy, actions and emotions into one’s own motor and viscer- were also reported by recording PTNs originating from
via Festa del Perdono 7, omotor representations of those actions and emotions. the primary motor cortex (area F1): most PTNs increased
I-20122 Milano, Italy. This sensorimotor transformation not only concerns their discharge during observation (‘facilitation-type’
4
Center for the Study of Social the content of actions and emotions but also their form mirror neurons), whereas others showed a discharge
Action, University of Milan,
I-20122 Milan, Italy.
or affective quality. Although the mechanism at work is suppression (‘suppression-type’ mirror neurons)10. A
Correspondence to G.R. the same — that is, the mirror mechanism — the cogni- comparison between F5 PTNs and F1 PTNs revealed
giacomo.rizzolatti@unipr.it tive functions it may fulfil vary according to the various that visual responses were much weaker in area F1 than
doi:10.1038/nrn.2016.135 brain networks that are involved. Finally, we argue that in area F5. This might explain why no F1 activation is
Published online 20 Oct 2016 what is common to all kinds of mirror-based processing typically reported in fMRI data.

NATURE REVIEWS | NEUROSCIENCE ADVANCE ONLINE PUBLICATION | 1

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

action goals, not just muscle contractions or joint dis-

F6 F1 placements21. This finding has since been supported by
a large number of studies in both monkeys and humans4.
Arm Before reviewing some of these studies, we need to
define the notion of action goal. This notion has been
VIP a potential source of confusion among authors discuss-
AIP ing mirror neurons22,23. We use the term ‘action goal’
IPS to describe an outcome to which the action is directed
PFG
(for example, the grasping of a piece of food or the plac-
F5p VPFc F5 ing of it) rather than an agent’s state (such as a mental
IAS representation of this outcome).
F5c STS How do we know that mirror neurons encode action
F5a goals rather than just bodily movements? One test is to
IT hold the goal constant while varying the movements,
and vice versa. Now, there is evidence that parietofrontal
Figure 1 | The monkey parietofrontal mirror network for hand-grasping actions. mirror neurons respond to the goal of an action such as
Lateral view of the macaque monkey brain localizing the main Nature Reviews
areas | Neuroscience
that are involved in the grasping of a piece of food even when this action is
action observation: the ventral premotor cortex (area F5), area PFG and anterior performed with different effectors2,24 (such as the hand
intraparietal area (AIP). The intraparietal sulcus (IPS) has been opened to show the areas or the mouth) or different tools25 (such as normal or
inside. The parietofrontal network receives high-order visual information from areas that
reverse pliers) requiring opposite sequences of move-
are located inside the superior temporal sulcus (STS) and inferior temporal lobule (IT).
Neither of these temporal regions has mirror properties. Area F5 is connected with area ments (closing or opening of the fingers). Conversely,
F1, which also contains mirror neurons. The parietofrontal circuit is under the control of mirror neuron activity has been shown to change when
the frontal lobe (area F6 or pre-supplementary motor area and ventral prefrontal cortex similar bodily movements (for example, reaching arm
(vPFC)). The inset shows an enlarged view of area F5, displaying its sectors buried inside movements) are directed to different action goals (for
the arcuate sulcus. The label ‘Arm’ indicates the area where the arm is represented. IAS, example, when grasping a piece of food to eat it or to
inferior limb of the arcuate sulcus; VIP, ventral intraparietal area. Figure from REF. 8, place it)26. A further test is to vary the sensory informa-
Nature Publishing Group. tion while holding the action goal constant. Now, there
is evidence that F5 mirror neurons selectively respond
Finally, neurons with mirror-like properties were to the goal of an action such as the cracking of the shell
described in both area F1 and the dorsal premotor cortex of a peanut regardless of whether this action is seen,
(PMd) when monkeys performed a highly learned reach- seen and heard, or just heard27. Similar results have been
ing task or observed the same task involving a visible reported in humans28,29.
hand or even just a cursor 11–13. Remarkably, mirror-like It is worth noting that mirror neurons do not differ in
neurons were also reported in the lateral prefrontal cor- their motor responses from other parietofrontal motor
tex 14. These neurons selectively responded not only to neurons, which are devoid of any mirror property. Since
present actions but also to past or future actions, thus the late 1980s, F5 motor neurons have been shown to be
facilitating social coordination. able to encode action goals rather than just mere joint
Substantial evidence that observing others’ actions displacements30. Similar goal-directed motor responses
recruits a pattern of cortical areas that are typically acti- have also been subsequently recorded from PFG and AIP
vated when executing those actions has also been reported neurons31. Thus, the mirror mechanism does nothing but
in humans. As several meta-analyses have shown15–17, the extend the same functional goal-related organization
main nodes of this pattern, in the case of grasping actions, from action execution to action observation.
are the inferior parietal lobule (IPL), PMv and caudal sec- Interestingly, a recent single-cell study investigated
tor of the inferior frontal gyrus (FIG. 2). It is worth noting motor and mirror responses when the monkeys had
Mirror mechanism
that most of the fMRI studies in these meta-analyses to refrain from achieving a specific action goal, such
A basic brain mechanism that
transforms sensory tested the activation of these areas during action obser- as the grasping of an object. The results showed that,
representations of others’ vation only. Nevertheless, the activation pattern was not whereas most F5 motor and mirror neurons were silent
behaviour into one’s own different when action execution was tested, not even in the refrain condition, some of them discharged when
motor or visceromotor when unsmoothed single-subject analyses were used18. the monkey or the experimenter grasped an object and
representations concerning
that behaviour and, depending
It has also been reported that the PMd and superior also when they refrained from doing so32. This indicates
on the location, can fulfil parietal lobule (SPL) are recruited by both action exe- that goal encoding might occur not only when an action
a range of cognitive functions, cution and action observation. Additional activations should be performed but also when the agent must
including action and emotion were found in the middle cingulate and somatosensory refrain from performing it.
understanding.
areas4,19. Finally, a single-cell study in humans recorded Analogous results have been reported in humans
Action goals neurons with mirror properties from the supplementary by brain-imaging and electrophysiological studies.
The outcomes to which actions motor cortex and hippocampus20. For instance, a recent fMRI study compared different
are directed. In non-technical observed actions (for example, climbing, walking and
contexts, the term is also used Goal encoding object manipulation). The main finding was that observ-
when people talk about the
goal of their struggles or assert
What do mirror neurons encode during action observa- ing an action such as climbing did not activate the pari-
that their goal is to work at tion? Early single-cell studies on parietofrontal mirror etal node of the hand grasping circuit. Instead, it recruited
their best. neurons showed that these neurons typically encode the rostral part of the dorsal SPL, even if the observed

2 | ADVANCE ONLINE PUBLICATION www.nature.com/nrn

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

PMd PF PMd
IPS IPS IFG and
PMv

IFG and PMv STS and STS and

pMTG pMTG
PF
Figure 2 | The human grasping-observation network. A meta-analysis of grasping-actionNature observation summarizing
Reviews all
| Neuroscience
effectors. All results are displayed on the left and right lateral-surface view of the MNI single-subject template. IFG,
inferior frontal gyrus; IPS, intraparietal sulcus; PMd, dorsal premotor cortex; pMTG, posterior middle temporal gyrus; PMv,
ventral premotor cortex; PF, parietal area F; STS, superior temporal sulcus. Adapted with permission from REF. 15, Elsevier.

climbing action involved a grasping hand33. Conversely, a Three different lines of evidence support this claim.
transcranial magnetic stimulation (TMS) adaptation study A first line concerns how deeply executing and observ-
investigated action-goal encoding, by comparing the rela- ing actions are interlinked. Several studies demonstrated
tive contribution of the PMv and IPL with that of a higher- that observing an action selectively affects the execution
order visual region such as the posterior STS (pSTS)34. of that action. It has been shown that observing an action
Participants were exposed to adapting movies of hand and can facilitate performing a congruent action and can
foot actions and were asked to match them to test pic- interfere with performing an incongruent action37–39.
tures showing the same or a different action, performed More interestingly for our purposes, there is compel-
with the same or a different effector. An adapting stimulus ling evidence that performing an action can affect the
induces habituation in a subset of cells encoding a particu- observer’s ability to judge what another individual
lar stimulus feature. TMS stimulation time locked to the is doing. For instance, repeated motor performance
cognitive task delivered over the cortical area containing of an action while blindfolded, such as pushing or pull-
the adapted neurons selectively improves the performance ing an object, induced in participants a strong percep-
in processing the adapted stimulus35. TMS pulses deliv- tual after-effect when judging whether other people were
ered over the PMv and IPL induced a selective shortening actually pushing or pulling an object 40. Crucially, the
of reaction times to adapted actions regardless of the effectors, after-effect vanished after delivering TMS over partici-
whereas this was not the case for the pSTS. pants’ PMv; this suggests a causal role for PMv in judging
This indicates that the PMv and IPL have a pivotal role others’ action goals.
in encoding others’ action goals. Interestingly, although A second line of evidence relies on the crucial role
both the PMv and IPL may encode action goals, the goal of motor expertise in identifying others’ action goals. A
encoding has been reported to be differently organized in large number of behavioural and brain-imaging studies
the two areas. Indeed, by comparing PMv and IPL activa- demonstrated that the richer people’s motor expertise is,
tion during the observation of similar actions performed the greater the sensitivity of their mirror mechanism to
with different effectors, a fMRI study showed that PMv others’ actions and the better their ability to identify the
goal encoding tended to be mainly mapped in a somato­ outcome to which those actions are directed will be. For
topic manner, whereas IPL goal encoding tended to be instance, viewing ballet steps has been found to recruit
mainly clustered according to the positive or negative premotor and parietal areas with mirror properties more
meaning of the observed action36. strongly in expert ballet dancers than in non-dancers or
in martial-art teachers41. Further studies systematically
Understanding action demonstrated that the recruitment of motor areas with
Because of their goal sensitivity, it has been claimed, mirror properties strongly correlates with motor rather
from the earliest studies, that mirror neurons have a than visual expertise42,43. In the same vein, it has been
role in understanding others’ actions1–3. Observing an shown that people can improve their ability to judge the
action elicits in the observer’s brain a motor representa- goal of an unusual action simply by practising that action
tion of the outcome to which the action is directed, themselves; this improvement occurs even when they
and this motor representation is similar to what would practise while blindfolded44.
occur if the observer himself were planning that action A third line of evidence comes from studies on
or even just imagining performing it. This would allow transient and permanent lesions of brain motor areas.
the observer to identify the goal of the observed action Transient lesions of PMv have been shown to signifi-
relying mainly on her or his own motor processes and cantly impair action-goal recognition. For instance,
representations8,21. disrupting participants’ PMv by repetitive TMS (rTMS)

NATURE REVIEWS | NEUROSCIENCE ADVANCE ONLINE PUBLICATION | 3

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

selectively impaired their ability to proactively gaze at centromedial nuclei of the amygdala51. The insula has a
the target of an observed hand action45. This effect was roughly triangular form and is subdivided by the ‘central
absent after rTMS disruption of STS areas. Even more sulcus’ into an anterior and a posterior part. Intracortical
strikingly, it has been shown that people are less accu- microstimulations in awake monkeys have individuated
rate in identifying hand action goals after receiving con- three major sectors: an anterior, a dorsocentral (with
tinuous theta-burst stimulation (cTBS) over the hand a caudal extension) and a ventral sector. The AI contains a
area than after receiving cTBS over the lip area of the mosaic of orofacial motor acts with a progressive dorsal to
left PMv, whereas they are less accurate in identifying ventral shift from ingestive responses to motor responses
mouth action goals after receiving cTBS over the lip with emotional content, such as disgust. The dorsocentral
area than over the hand area46. Given that the applica- insula contains a sensorimotor field, which functionally
tion of cTBS over motor areas diminished the excitability seems to be an extension of the parietal lobule. Finally, the
of cortical tissues, these findings suggest that PMv might stimulation of the ventral insula elicits affiliative mouth
have a causal role in understanding others’ actions. and face gestures and discomfort reactions52,53.
Patients with parietal or frontal brain lesions might A meta-analysis based on a large number of brain-
have selective deficits in both producing and recogniz- imaging studies reported a somewhat similar functional
ing actions. It is worth noting that the claim that the organization in humans54. According to this study, the
mirror mechanism has a role in action understanding human insula is formed by four distinct functional fields:
does not imply that every motor deficit should result a sensorimotor, an olfactory-gustatory, a socio-emotional
in an impairment of action understanding. Rather, this and a cognitive field. The sensorimotor field is located
claim implies that impairments of action understand- in the dorsocentral insula and corresponds to the sen-
ing might correlate with deficits in representing and sorimotor field of the monkey. The olfactory-gustatory
producing an action. And this is what has been found field fits well the ingestive sectors of the monkey insula,
since the 1980s47,48. More recently, it has been shown whereas the socio-emotional field is located in the
that patients with buccofacial ideomotor apraxia, who ventral part of the insula. Finally, the cognitive field is
are typically impaired in carrying out face and mouth functionally heterogeneous, given that it is activated by
actions, showed impairment in identifying the goals of short-term memory retrieval, language processing and
mouth-related actions, whereas patients with limb ideo- overt speech55,56. This field has not been described in the
motor apraxia, who are typically impaired in producing monkey brain.
hand actions, showed impairment in identifying the Disgust was elicited in monkeys by electrical stimula-
goals of hand-related actions49. This might seem to be tion of the ventral sector of the AI. Behavioural responses
at odds with the well-known finding that some patients consisted of typical disgust grimaces (such as the curling
with ideomotor apraxia have no obvious deficit in of the upper lip and the wrinkling of the nose), occa-
action understanding 50. However, the conflict is merely sionally followed by retching. Sometimes, more complex
apparent, because patients’ impairments in under- patterns have been observed, including food spitting or
standing others’ actions were not related to ideomotor even refusal of food intake52,53.
apraxia itself, as shown by the existence of patients with Electrical stimulation of the AI also elicits disgust in
ideomotor apraxia who do not have gesture-recognition humans. Disgust responses were reported in patients with
deficits. Instead, patients’ impairments were rather due epilepsy when the insula was stimulated before surgery to
to the extension of their lesions in parietal and frontal localize the epileptic focus57. These reports are in line with
areas, which are endowed with mirror properties (for a clinical data indicating that the occurrence of vomiting
more detailed discussion of this issue, see REF. 4). correlates in time with epileptic discharges affecting the
AI only 58,59.
Mirroring emotions Several studies indicated that the AI is also responsive
A link between action and emotion is already suggested to sensory stimuli. Single-cell recordings from the mon-
by etymology. However, the way in which emotions are key insula showed the presence of neurons responding to
usually conceived generally minimizes the role of motor gustatory stimuli and somatosensory stimuli such as tex-
and visceromotor processes and representations. In this ture, viscosity, and so on60,61. Similar results were reported
section, we show that the mirror mechanism operates in humans (reviewed in REF. 54).
in the emotion domain and plays a key part in process-
ing and understanding the emotions of others. We focus Is that my emotion or yours?
mainly on one emotion: disgust. This basic emotion has Critically for our purposes, the AI turns out to be involved
been extensively studied in monkeys and humans, and not only in feeling disgust but also in observing some-
there is strong consistency between neurophysiological, one else being disgusted. A seminal fMRI study showed
brain-imaging and lesion data. that the sight of disgusted facial expressions selectively
activated the AI, with the activation extending to some
The insula of disgust limbic cortico–striatal–thalamic structures when the
The main cortical node for disgust in both monkeys and observed facial expressions were strongly disgusted62.
humans is located in the anterior part of the insula (AI). These findings have been corroborated by other fMRI
The insula is connected with various cortical areas and studies that demonstrated a selective recruitment of the
sends descending connections to emotional core struc- AI in others’ disgust processing 63,64. In the same vein,
tures such as the hypothalamus, periaqueductal grey and intracortical recordings from the insula of patients with

4 | ADVANCE ONLINE PUBLICATION www.nature.com/nrn

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

epilepsy showed that the sight of others’ disgust selectively Evidence supporting this hypothesis comes from
activated the electrodes located in the AI; this effect was lesion studies. Particularly consistent are those find-
absent in the posterior insula65. ings related to selective impairments in experiencing
An important study 66 published in 2003 took a fur- and recognizing disgust. Of great interest is the case
ther step towards understanding the distinctive role of describing a patient who, after lesion of the left insula,
the AI in processing others’ disgust. For the first time, showed a consistent and largely selective impairment
this study demonstrated a clear overlap between insu- in recognizing disgust from others’ faces expressing
lar activation elicited by one’s own and others’ disgust. different emotions, non-verbal emotional sounds and
In some runs, participants were exposed to disgusting emotional prosody 80. Interestingly, the deficit for oth-
or pleasant odorants, whereas, in other runs, they were ers’ disgust was mirrored by an equivalent deficit in the
presented with short movie clips of other individuals patient’s first-person experience of the same emotion.
smelling the content of a glass and displaying a facial Another study 81 reported a similar case of deficit for
expression of disgust or pleasure. The main finding was disgust. The patient was described to have anatomical
that the observation of others’ disgust activated the same damage involving both left and right insular cortices.
neuronal substrates within the AI that were activated by When viewing facial expressions or hearing descriptions
the exposure to the disgusting odorants. This suggested a of emotionally related actions, he was able to retrieve
putative common insular ground for experiencing one’s knowledge about all basic emotions except disgust.
own disgust and for observing someone else being dis- Strikingly, when the experimenter acted out behaviours
gusted. Interestingly, a common insular ground has also that are typically associated with disgust (for example,
been demonstrated for vivid imagination of disgusting spitting out of food), the patient remained indifferent or
experiences. Indeed, it has been shown that experiencing even indicated that the food was “delicious”.
one’s own disgust, observing other’s disgust and imag- Last, voxel-based morphometry was used to inves-
ining being disgusted selectively activated a common tigate whether differences in grey-matter volume were
sector of the AI67. This sector was found to be embed- related to deficits in emotion processing in patients with
ded in distinct functional circuits, and this may account preclinical Huntington disease82. In these patients, sig-
for the differences between experiencing, observing and nificant correlations were found between the volume of
imagining a basic emotion such as disgust. grey matter in the left anteroventral insula and disgust
The existence of a putative common neuronal ground recognition. Remarkably, this effect did not generalize
for experiencing and observing emotions has also been to other emotions.
hypothesized for emotions other than disgust, such as fear Taken together, the above-mentioned findings suggest
or pleasure68–70. There have also been fMRI studies provid- that emotional brain centres that are putatively endowed
ing evidence for a putative common neuronal ground for with a mirror mechanism might have a distinctive role
nociceptive and vicarious pain, which would mainly com- in recognizing others’ emotions. Further research is
prise the AI and anterior cingulate cortex (ACC)71,72. That needed to investigate the possible involvement of the
said, several concerns have been raised about the pain mirror mechanism beyond basic emotions. However,
specificity of those areas73 and the functional significance there is evidence that observing the regrettable outcomes
of overlaps in their activation triggered by experiencing of others’ choices activates the same regions (ventrome-
one’s own pain and observing others’ pain74. dial prefrontal cortex, ACC and hippocampus) that are
Finally, a very recent study, which was performed on activated during a first-person experience of regret 83. An
patients who underwent surgery to implant intra­cerebral interesting issue is whether this also holds for other social
electrodes, showed that electrical stimulation of the ros- emotions such as envy or shame.
tral pregenual ACC elicited a burst of laughter, often
accompanied by a change in the emotional state, as indi- Mirroring vitality forms
cated by a sudden appearance of mirth75. Interestingly, Actions and emotions take many forms: a grip can be
preliminary data indicate that seeing someone else ‘vigorous’ or ‘delicate’, a caress can be ‘gentle’ or ‘rushed’,
laughing or exhibiting mirth and electrical stimulation- and even a basic emotion such as anger can ‘explode’ or
induced own laughter and mirth activate the same site be ‘cold’. Attention to the forms of gestures and emotional
within the ACC76. reactions can provide information about agents’ affective
states, moods, attitudes and even their character traits.
Understanding emotions These action and emotion forms have been variously
The existence of a putative common neuronal ground conceived by different researchers84–86. Stern coined the
for experiencing and observing emotions suggested that term “vitality affects”, or “vitality forms”, to highlight
cortical centres such as the AI or ACC might be endowed that these forms underlie our most basic social interac-
with a mirror mechanism77–79. Accordingly, observing an tions, thus deeply shaping our experience of ourselves
emotion such as disgust would elicit visceromotor pro- and of others85,86. According to Stern’s view, actions
cesses and representations in the observers’ brain that are and emotions cannot be experienced without one or
similar to those processes and representations that would another vitality form. Indeed, what we are doing or feel-
occur if the observers themselves were experiencing that ing gives content to our experience, whereas vitality
emotion. This could allow the observers to identify oth- forms provide these contents with a form, which char-
ers’ emotion by capitalizing on their own visceromotor acterizes how actions and emotions are displayed and
processes and representations concerning that emotion. experienced.

NATURE REVIEWS | NEUROSCIENCE ADVANCE ONLINE PUBLICATION | 5

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

a Rude Gentle Control

RH LH RH LH RH LH

x 38 x –36 x 38 x –38 x 40 x –38

b
1.4 * * 1.4 * * 1.4 * *
Signal change (a.u.)

1.2 1.2 1.2

1.0 1.0 1.0
0.8 0.8 0.8
0.6 0.6 0.6
0.4 0.4 0.4
0.2 0.2 0.2
0.0 0.0 0.0
RH insula (38 6 –6) LH insula (–36 4 2) RH insula (38 6 –8) LH insula (–38 4 2) RH insula (40 6 –4) LH insula (–38 2 6)
Rude Gentle Control

Figure 3 | Insular region encoding vitality forms. a | Parasagittal sections showing the insularNatureactivations in the right
Reviews | Neuroscience
hemisphere (RH) and left hemisphere (LH) when participants observed actions that are performed with two different
vitality forms (rude and gentle) and a control action (placing a ball). In the panels, the stereotaxic parameter x is indicated.
b | Signal changes in six regions of interest centred on the central insula. The regions of interest were defined
centring the sphere (radius 10 mm) on the maxima of the functional maps resulting from the conjunction analysis of
observation, imagination and execution tasks. In the panels, the stereotaxic parameters x, y, and z are indicated. a.u.,
arbitrary unit. *Significant differences at P < 0.05 (Bonferroni corrected). Adapted with permission from REF. 87, PNAS.

One’s own and others’ vitality forms Anatomical data in monkeys indicate that the DCI
In spite of their pervasiveness in our life and their cru- is connected with areas AIP, PFG 88, F5 (REF. 89), 46v and
cial role in interpersonal relations, vitality forms have 12r 90 (FIG. 4). These areas form the circuit that is involved
rarely been investigated from a neuroscientific point of in the organization and control of arm movements in
view. Little or nothing is known about their neuronal monkeys and humans (see above). Intracortical electri-
substrates. How and where are vitality forms encoded cal stimulation of the middle and posterior short gyri
at the brain level? What happens when actions are per- of the human insula — a sector basically corresponding
formed with different vitality forms? And what happens to the monkey DCI – elicited evoked potential in the pre-
when actions with different vitality forms are observed central gyrus and IPL91. Taken together, these findings
rather than performed? suggest that the DCI plays a key part in the production
In a fMRI study investigating the neuronal bases of and control of different vitality forms, by modulating the
vitality forms87, participants were required to execute, cortical circuit controlling movements. It is worth noting
imagine and observe actions that were performed with that this modulation differs radically from that related
two different vitality forms. For instance, in the execu- to the activation of the AI, which is typically involved in
tion condition, the participants had to move an object, driving emotionally laden behaviour.
such a packet of crackers, as if to offer it to another indi-
vidual in a gentle or rude way. In the imagination condi- Understanding vitality forms
tion, they had to imagine themselves passing an object to As described above, the DCI becomes active not only
someone else in a gentle or rude way. In the observation when performing an action with a given vitality form but
condition, participants observed an actor performing also when observing someone else performing the action
gently or rudely a passing action. with that vitality form. This finding is not per se conclusive
The results showed, in the execution condition, a bilat- evidence for a common neuronal ground for vitality-form
eral activation of the premotor and parietal cortices, plus execution and observation. However, taken together with
a strong activation of the left somatosensory and motor the above-mentioned data about the anatomical and func-
cortices. In the observation condition, there was a bilat- tional properties of this sector of the insula and its involve-
eral activation of the occipital lobe, PMd and SPL extend- ment in motor imagery, this finding makes plausible the
ing anteriorly into the IPL. The imagination condition hypothesis that the DCI might be endowed with a mirror
resulted in a similar pattern but with a strongly reduced mechanism that transforms the sensory information con-
activation of the occipital areas. A conjunction analysis of cerning others’ vitality forms into motor representations of
Vitality form
the activations found in all the three conditions revealed those forms. Accordingly, observing an action performed
Agents’ affective states, moods
and attitudes that characterize a bilateral signal increase in the putative human middle with a given vitality form would elicit in the observers’
how their actions and emotions temporal area MT/V5+, posterior parietal cortex and insula a representation of this vitality form that is similar
are displayed and experienced; PMd, as well as in the dorsal part of the cerebellum. Most to that which occurs when the observers themselves are
it has been hypothesized that interestingly, this analysis demonstrated a selective activa- performing an action with this vitality form. This would
the identification of others’
vitality forms, which is crucial
tion of the dorsocentral part of the insula (DCI) when an allow the observers to identify others’ vitality forms by cap-
for social interaction, could action was performed, imagined to perform and observed italizing on their own motor processes and representations
involve the mirror mechanism. to be performed with a given vitality form (FIG. 3). concerning these vitality forms.

6 | ADVANCE ONLINE PUBLICATION www.nature.com/nrn

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

with ASD and healthy controls were presented with

pairs of videos showing two actors executing the same
or a different action with the same or a different vitality
form. As in the fMRI study mentioned before, there were
SAS
two tasks: a what task related to the goal of the observed
IPS CS PS actions and a how task concerning their vitality forms.
46v The results showed a clear dissociation between the two
AIP 12r
tasks. Whereas the performance of individuals with ASD
PFG
F5 in recognizing others’ action goals was similar to that of
IAS controls (see also REFS 95,96), individuals with ASD were
clearly impaired in recognizing others’ vitality forms.
LF
These findings are in line with a series of studies
DCI demonstrating a clear deficit in individuals with ASD
in imitating movements that are executed with different
performance profiles or vitality forms97,98. In particular,
STS individuals with and without ASD were similar to one
another in their ability to imitate goal-directed actions
(but see REFS 99,100), but those with ASD were signif-
Figure 4 | Insular connections withReviews
the parietofrontal
Nature | Neuroscience icantly impaired in imitating the style of the observed
grasping circuit in the monkey. The parietofrontal circuit
underlying motor organization is depicted in blue. The
actions, especially when the style was not necessary for
insular connections with the parietofrontal areas are goal achievement.
depicted in red. AIP, anterior intraparietal area; CS, central
sulcus; DCI, dorsocentral part of the insula; IAS, inferior Concluding remarks
limb of the arcuate sulcus; IPS, intraparietal sulcus; LF, The functional properties of the mirror mechanism indi-
lateral fissure; PS, parietal sulcus; SAS, superior arcuate cate that the motor processes and representations that are
sulcus; STS, superior temporal sulcus. primarily involved in generating and controlling a given
behaviour can also be recruited in an individual who is
Evidence for a specific role of the DCI in vitality-form observing someone else displaying that behaviour. By
recognition comes from a fMRI study 92. Participants were means of this recruitment, the individual may take advan-
presented with pairs of consecutive videos showing two tage of his or her own processes and representations to
actors executing the same or a different action with the understand others’ actions and emotions, as well as their
same or a different vitality form. There were two different corresponding vitality forms.
tasks. The ‘what task’ required the participants to judge Elsewhere, we have defined mirror-based understand-
whether the observed actions had the same or a differ- ing as an understanding from the inside8,101,102 for two main
ent goal, regardless of their vitality form. The ‘how task’ reasons. First, this definition highlights that the mirror
required the participants to judge whether the observed mechanism may provide a route to knowledge of others,
vitality forms were similar or different, regardless of which can be mainly exploited just in ways that depend
the action goal. The results showed an activation of the on one’s own motor and visceromotor processes and
parieto­frontal circuit in both tasks. Interestingly, the con- representations. Second, it suggests that there are plau-
trast between how and what tasks revealed a significant sible aspects of phenomenal character that are common
activation of the sensorimotor field located in the DCI, to experiences of one’s own and others’ actions, emo-
thus suggesting that this sector of the insula is selectively tions and vitality forms, given that both experiences are
involved not only in producing one’s own vitality forms but shaped by the same motor or visceromotor processes
also in recognizing vitality forms when observed in others. and representations103. In this respect, action, emotion
One might object that DCI activation did not reflect and vitality-form understanding would not substan-
anything else but kinematic profiles. After all, rude actions tially differ from motor imagery, at least as conceived by
are typically faster or jerki er than gentle ones. A very Jeannerod. In his view, motor imagery is functionally so
recent multivoxel analysis paper tackled this issue, scru- close to action preparation and execution that the experi-
tinizing whether there were DCI voxels discriminating ence of imagining acting shares a core phenomenal aspect
vitality forms from mere kinematic profiles93. The results with the experience of actually performing the imagined
Understanding from the showed that the DCI contained voxels specific for vitality- action104. The suggestion is that how people experience
inside form processing. Although further research is needed, their own actions, emotions and vitality forms shares a
A notion that describes how these results suggest that vitality forms are processed in the core phenomenal aspect with how they experience the
the mirror mechanism might
DCI differently from kinematics profiles, even from those same actions, emotions and vitality forms when observing
provide a route to knowledge
of others by capitalizing on that are, taken individually, the bases of vitality forms. others displaying them.
one’s own motor or Strikingly, there is evidence not only that ‘what’ and The claim that the mirror mechanism might allow oth-
visceromotor representations. ‘how’ are processed differently at the brain level but also ers’ actions, emotions or vitality forms to be understood
It also reflects the idea that that action goal and vitality-form recognition might be from the inside does not imply that the mirror-based pro-
there are phenomenal aspects
common to experiences of
dissociated. Indeed, a behavioural study showed a selec- cessing is enough to fully understand them (as previously
one’s own and other’s actions, tive impairment in vitality-form recognition in individu- emphasized, see REF. 8). Fully understanding an action, an
emotions and vitality forms. als with autism spectrum disorder (ASD)94. Individuals emotion or a vitality form is a multilevel process. The first

NATURE REVIEWS | NEUROSCIENCE ADVANCE ONLINE PUBLICATION | 7

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
REVIEWS

level involves identifying which outcome is the goal of In spite of some attempts to theoretically integrate the
the observed action and which emotion or vitality form different levels of action and emotion processing 22,109–111,
other individuals are displaying. Further levels may many questions are still open. A natural question con-
involve representing others’ mental states (for example, cerns whether and how the mirror-based action and
beliefs, desires, intentions, and so on). These mental emotion processing might contribute, in part at least, to
states might provide reasons explaining why the action, reasoning about others’ minds. Another question con-
the emotion or the vitality form happened and how those cerns the development of abilities to understand others’
reasons are linked to others’ minds and their behaviour. actions and minds. There is evidence that infants succeed
There is evidence that brain areas with mirror properties in ascribing goals by capitalizing on their own motor pro-
are usually recruited at the first level of action, emotion cesses and representations from very early in life112–114.
and vitality-form processing, which concerns goal iden- However, further research is needed to assess whether
tification and emotion and vitality-form recognition, infants similarly recognize others’ emotions and vitality
whereas reasoning about others’ mental states mainly forms. Finally, it is far from clear how the putative ‘men-
activates a putative ‘mind-reading network’ that is formed talizing network’ actually works and whether it could
by the mesial frontal cortex, ACC and the so‑called modulate the mirror-based processing of action, emotion
temporoparietal junction105–108. and vitality form.

1. di Pellegrino, G., Fadiga, L., Fogassi, L., Gallese, V. & 15. Caspers, S., Zilles, K., Laird, A. R. & Eickhoff, S. B. ALE 32. Bonini, L., Maranesi, M., Livi, A., Fogassi, L. &
Rizzolatti, G. Understanding motor events: meta-analysis of action observation and imitation in the Rizzolatti, G. Ventral premotor neurons encoding
a neurophysiological study. Exp. Brain Res. 91, human brain. Neuroimage 50, 1148–1167 (2010). representations of action during self and others’
176–180 (1992). This review summarizes a large number of inaction. Curr. Biol. 24, 1611–1614 (2014).
2. Gallese, V., Fadiga, L., Fogassi, L. & Rizzolatti, G. Action experiments on action observation and imitation, This single-neuron study shows that the monkey
recognition in the premotor cortex. Brain 119, and individuates a basic circuit that is involved in PMv encodes action goals not only when
593–609 (1996). goal ascription in humans. performing and observing actions to be performed
3. Rizzolatti, G., Fadiga, L., Gallese, V. & Fogassi, L. 16. Grosbras, M. H., Beaton, S. & Eickhoff, S. B. Brain but also when refraining from acting and observing
Premotor cortex and the recognition of motor actions. regions involved in human movement perception: actions to be refrained from.
Cogn. Brain Res. 3, 131–141 (1996). a quantitative voxel-based meta-analysis. Hum. Brain 33. Abdollahi, R. O., Jastorff, J. & Orban, G. A. Common
4. Rizzolatti, G., Cattaneo, L., Fabbri-Destro, M. & Mapp. 33, 431–454 (2012). and segregated processing of observed actions in
Rozzi, S. Cortical mechanisms underlying the 17. Molenberghs, P., Cunnington, R. & Mattingley, J. B. human SPL. Cereb. Cortex 11, 2734–2753 (2012).
organization of goal-directed actions and mirror Brain regions with mirror properties: a meta-analysis 34. Cattaneo, L., Sandrini, M. & Schwarzbach, J. State-
neuron-based action understanding. Physiol. Rev. 94, of 125 human fMRI studies. Neurosci. Biobehav. Rev. dependent TMS reveals a hierarchical representation of
655–706 (2014). 36, 341–349 (2012). observed acts in the temporal, parietal, and premotor
This rich, exhaustive review describes the functional 18. Gazzola, V. & Keysers, C. The observation and cortices. Cereb. Cortex 20, 2252–2258 (2010).
organization of the cortical motor system and the execution of actions share motor and somatosensory 35. Silvanto, J., Muggleton, N. & Walsh, V. State
goal-directed mirror mechanism in monkeys and voxels in all tested subjects: single-subject analyses of dependency in brain stimulation studies of perception
humans. unsmoothed fMRI data. Cereb. Cortex 19, and cognition. Trends Cogn. Sci. 12, 447–454 (2008).
5. Keysers, C. & Gazzola, V. Expanding the mirror: 1239–1255 (2009). 36. Jastorff, J., Begliomini, C., Fabbri-Destro, M.,
vicarious activity for action, emotion and sensation. 19. Keysers, C., Kaas, J. H. & Gazzola, V. Somatosensation Rizzolatti, G. & Orban, G. A. Coding observed motor
Curr. Opin. Neurobiol. 19, 666–671 (2009). in social perception. Nat. Rev. Neurosci. 11, 417–428 acts: different organizational principles in the parietal
6. Nelissen, K., Luppino, G., Vanduffel, W., Rizzolatti, G. & (2010). and premotor cortex of humans. J. Neurophysiol.
Orban, G. A. Observing others: multiple action 20. Mukamel, R., Ekstrom, A. D., Kaplan, J., Iacoboni, M. 104, 128–140 (2010).
representation in the frontal lobe. Science 310, & Fried, I. Single-neuron responses in humans during 37. Brass, M., Bekkering, H., Wohlschläger, A. & Prinz, W.
332–336 (2005). execution and observation of actions. Curr. Biol. 20, Compatibility between observed and executed finger
7. Nelissen, K. et al. Action observation circuits in the 750–756 (2010). movements: comparing symbolic, spatial, and
macaque monkey cortex. J. Neurosci. 31, 21. Rizzolatti, G., Fogassi, L. & Gallese, V. imitative cues. Brain Cogn. 44, 124–143 (2000).
3743– 3756 (2011). Neurophysiological mechanisms underlying the 38. Craighero, L., Bello, A., Fadiga, L. & Rizzolatti, G.
This detailed connectional study combines fMRI and understanding and imitation of action. Nat. Rev. Hand action preparation influences the responses to
neural tracer techniques, and defines the functional Neurosci. 2, 661–670 (2001). hand pictures. Neuropsychologia 40, 492–502
paths by which visual information concerning others’ 22. Kilner, J. M. More than one pathway to action (2002).
actions reaches the premotor cortex. understanding. Trends Cogn. Sci. 15, 352–357 39. Kilner, J. M., Paulignan, Y. & Blakemore, S. J. An
8. Rizzolatti, G. & Sinigaglia, C. The functional role of (2011). interference effect of observed biological movement
parieto-frontal mirror circuit: interpretations and 23. Cook, R. & Bird, G. Do mirror neurons really mirror on action. Curr. Biol. 13, 522–525 (2003).
misinterpretations. Nat. Rev. Neurosci. 11, 264–274 and do they really code for action goals? Cortex 49, 40. Cattaneo, L. et al. One’s motor performance
(2010). 2944–2945 (2013). predictably modulates the understanding of others’
9. Kraskov, A., Dancause, N., Quallo, M. M., Shepherd, S. 24. Ferrari, P. F., Rozzi, S. & Fogassi, L. Mirror neurons actions through adaptation of premotor visuo-motor
& Lemon, R. N. Corticospinal neurons in macaque responding to observation of actions made with tools neurons. Soc. Cogn. Affect. Neurosci. 6, 301–310
ventral premotor cortex with mirror properties: in monkey ventral premotor cortex. J. Cogn. Neurosci. (2011).
a potential mechanism for action suppression? Neuron 17, 212–226 (2005). 41. Calvo-Merino, B., Glaser, D. E., Grezes, J.,
64, 922–930 (2009). 25. Rochat, M. J. et al. Responses of mirror neurons in Passingham, R. E. & Haggard, P. Action observation
10. Vigneswaran, G., Philipp, R., Lemon, R. N. & area F5 to hand and tool grasping observation. Exp. and acquired motor skills: an fMRI study with expert
Kraskov, A. M1 corticospinal mirror neurons and their Brain Res. 204, 605–616 (2010). dancers. Cereb. Cortex 15, 1243–1249 (2005).
role in movement suppression during action 26. Fogassi, L. et al. Parietal lobe: from action This article provides the most elegant
observation. Curr. Biol. 23, 236–243 (2013). organization to intention understanding. Science 302, demonstration that acquired motor skills change the
This single-neuron study provides the first 662–667 (2005). responsiveness of the mirror cortical areas during
demonstration that corticospinal neurons of the 27. Kohler, E. et al. Hearing sounds, understanding the observation of actions exploiting those skills.
monkey primary motor cortex are endowed with mirror actions: action representation in mirror neurons. 42. Calvo-Merino, B., Grezes, J., Glaser, D. E.,
properties and suggests a mechanism for suppressing Science 297, 846–848 (2002). Passingham, R. E. & Haggard, P. Seeing or doing?
involuntary movements during action observation. 28. Gazzola, V., Aziz-Zadeh, L. & Keysers, C. Empathy and Influence of visual and motor familiarity in action
11. Cisek, P. & Kalaska, J. F. Neural correlates of mental the somatotopic auditory mirror system in humans. observation. Curr. Biol. 16, 1905–1910 (2006).
rehearsal in dorsal premotor cortex. Nature 431, Curr. Biol. 16, 1824–1829 (2006). 43. Cross, E. S., Hamilton, A. F. & Grafton, S. T. Building
993–996 (2004). 29. Lewis, J. W., Brefczynski, J. A., Phinney, R. E., a motor simulation de novo: observation of dance by
12. Tkach, D., Reimer, J. & Hatsopoulos, N. G. Congruent Janik, J. J. & De Yoe, E. A. Distinct cortical pathways dancers. Neuroimage 31, 1257–1267 (2006).
activity during action and action observation in motor for processing tool versus animal sounds. J. Neurosci. 44. Casile, A. & Giese, M. A. Non visual motor training
cortex. J. Neurosci. 27, 13241–13250 (2007). 25, 5148–5158 (2005). influences biological motion perception. Curr. Biol. 16,
13. Dushanova, J. & Donoghue, J. Neurons in primary 30. Rizzolatti, G. et al. Functional organization of inferior 69–74 (2006).
motor cortex engaged during action observation. Eur. area 6 in the macaque monkey. II. area F5 and the 45. Costantini, M., Ambrosini, E., Cardellicchio, P. &
J. Neurosci. 31, 386–398 (2010). control of distal movements. Exp. Brain Res. 71, Sinigaglia, C. How your hand drives my eyes. Soc.
14. Falcone, R., Brunamonti, E., Ferraina, S. & 491–507 (1988). Cogn. Affect. Neurosci. 9, 705–711 (2014).
Genovesio, A. Neural encoding of self and another 31. Bonini, L. et al. Selectivity for grip type and action goal 46. Michael, J. et al. Continuous theta-burst stimulation
agent’s goal in the primate prefrontal cortex: human– in macaque inferior parietal and ventral premotor demonstrates a causal role of premotor homunculus in
monkey interactions. Cereb Cortex http://dx.doi. grasping neurons. J. Neurophysiol. 108, 1607–1619 action understanding. Psychol. Sci. 25, 963–972
org./10.1093/cercor/bhv224 (2015). (2012). (2014).

8 | ADVANCE ONLINE PUBLICATION www.nature.com/nrn

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.
 REVIEWS

47. Heilman, K. M., Rothi, L. J. & Valenstein, E. Two forms of 68. Sprengelmeyer, R., Rausch, M., Eysel, U. T. & 93. Di Cesare, G. et al. Vitality forms processing in the
ideomotor apraxia. Neurology 32, 342–346 (1982). Przuntek, H. Neural structures associated with insula during action observation: a multivoxel pattern
48. Rothi, L. J., Heilman, K. M. & Watson, R. T. Pantomime recognition of facial expressions of basic emotions. analysis. Front. Hum. Neurosci. http://dx.doi.
comprehension and ideomotor apraxia. J. Neurol. Proc. Biol. Sci. 265, 1927–1931 (1998). org/10.3389/fnhum.2016.00267 (2016).
Neurosurg. Psychiatry 48, 207–210 (1985). 69. Hennenlotter, A. et al. A common neural basis for 94. Rochat, M. J. et al. Impaired vitality form recognition
49. Pazzaglia, M., Pizzamiglio, L., Pes, E. & Aglioti, S. M. receptive and expressive communication of pleasant in autism. Neuropsychologia 51, 1918–1924
The sound of actions in apraxia. Curr. Biol. 18, facial affect. Neuroimage 26, 581–591 (2005). (2013).
1766–1772 (2008). 70. Jabbi, M., Swart, M. & Keysers, C. Empathy for 95. Boria, S. et al. Intention understanding in autism.
This neuropsychological study convincingly positive and negative emotions in the gustatory PLoS ONE 4, e5596 (2009).
demonstrates a causative link between execution of cortex. Neuroimage 34, 1744–1753 (2007). 96. Hamilton, A. F., Brindley, R. M. & Frith, U. Imitation
actions and their recognition. 71. Singer, T. et al. Empathy for pain involves the affective and action understanding in autistic spectrum
50. De Renzi, E. & Faglioni, P. in Handbook of Clinical and but not sensory components of pain. Science 303, disorders: how valid is the hypothesis of a deficit in the
Experimental Neuropsychology (eds Denes, G. & 1157–1162 (2004). mirror neuron system? Neuropsychologia 45,
Pizzamiglio, L.) 421–440 (Psychology Press, 1999). 72. Lamm, C., Decety, J. & Singer, T. Meta-analytic 1859–1868 (2007).
51. Gothard, K. & Hoffman, K. in Primate Neuroethology evidence for common and distinct neural networks 97. Hobson, R. P. & Lee, A. Imitation and identification in
(eds Platt, M. & Ghazanfar, A.) 292–315 (Oxford associated with directly experienced pain and autism. J. Child Psychol. Psychiatry 40, 649–659
Univ. Press, 2010). empathy for pain. Neuroimage 54, 2492–2502 (1999).
52. Caruana, F., Jezzini, A., Sbriscia-Fioretti, B., Rizzolatti, G. (2011). 98. Hobson, R. P. & Hobson, J. A. Dissociable aspects of
& Gallese, V. Emotional and social behaviors elicited by 73. Iannetti, G. D., Salomons, T. V., Moayedi, M., imitation: a study in autism. J. Exp. Child Psychol.
electrical stimulation of the insula in the macaque Mouraux, A. & Davis, K. D. Beyond metaphor: 101, 170–185 (2008).
monkey. Curr. Biol. 21, 195–199 (2011). contrasting mechanisms of social and physical pain. 99. Gizzonio, V. et al. Failure in pantomime execution
53. Jezzini, A., Caruana, F., Stoianov, I., Gallese, V. & Trends Cogn. Sci. 17, 371–378 (2013). correlates with the severity of social behavior deficits
Rizzolatti, G. Functional organization of the insula and 74. Zaki, J., Wager, T. D., Singer, T., Keysers, C. & in children with autism: a praxis study. J. Autism Dev.
inner perisylvian regions. Proc. Natl Acad. Sci. USA Gazzola, V. The anatomy of suffering: understanding Disord. 45, 3085–3097 (2015).
109, 10077–10082 (2012). the relationship between nociceptive and empathic 100. MacNeil, L. K. & Mostofsky, S. H. Specificity of
This detailed study of the physiological organization pain. Trends Cogn. Sci. 20, 249–259 (2016). dyspraxia in children with autism. Neuropsychology
of the monkey insula shows the profound functional 75. Caruana, F. et al. Mirth and laughter elicited by 26, 165–171 (2012).
heterogeneity of different sectors of the insula. electrical stimulation of the human anterior cingulate 101. Rizzolatti, G. & Sinigaglia, C. in Action Science:
54. Kurth, F., Zilles, K., Fox, P. T., Laird, A. R. & cortex. Cortex 71, 323–331 (2015). Foundations of an Emerging Discipline (eds Prinz, W.,
Eickhoff, S. B. A link between the systems: functional 76. Caruana, F. et al. A mirror mechanism for laughter in Beisert, M. & Herwig, A.) 201–227 (MIT Press,
differentiation and integration within the human insula the anterior cingulate cortex. Emotion (in the press). 2013).
revealed by meta-analysis. Brain Struct. Funct. 214, 77. Gallese, V., Keysers, C. & Rizzolatti, G. A unifying view 102. Sinigaglia, C. & Rizzolatti, G. Through the looking
519–534 (2010). of the basis of social cognition. Trends Cogn. Sci. 8, glass: self and others. Consci. Cogn. 20, 64–74
This detailed meta-analysis of fMRI studies of the 396–403 (2004). (2011).
human insula highlights its functional 78. Goldman, A. & Sripada, C. S. Simulationist models of 103. Sinigaglia, C. & Butterfill, S. A puzzle about the
heterogeneity. face-based emotion recognition. Cognition 94, relations between thought, experience, and the
55. Ojemann, G. A. & Whitaker, H. A. Language localization 193–213 (2005). motoric. Synthese 192, 1923–1936 (2015).
and variability. Brain Lang. 6, 239–260 (1978). 79. Niedenthal, P. M. Embodying emotion. Science 316, 104. Jeannerod, M. The representing brain: neural
56. Vignolo, L. A., Boccardi, E. & Caverni, L. Unexpected 1002–1005 (2007). correlates of motor intention and imagery. Behav.
CT‑scan findings in global aphasia. Cortex 22, 55–69 80. Calder, A. J., Keane, J., Manes, F., Antoun, N. & Brain Sci. 17, 187–202 (1994).
(1986). Young, A. W. Impaired recognition and experience of 105. de Lange, F. P., Spronk, M., Willems, R. M., Toni, I. &
57. Penfield, W. & Faulk, M. E. The insula: further disgust following brain injury. Nat. Neurosci. 3, Bekkering, H. Complementary systems for
observations on its function. Brain 78, 445–470 (1955). 1077–1078 (2000). understanding action intentions. Curr. Biol. 18,
58. Isnard, J., Guenot, M., Sindou, M. & Mauguiére, F. 81. Adolphs, R., Tranel, D. & Damasio, A. R. Dissociable 454–457 (2008).
Clinical manifestation of insular lobe seizures: neural systems for recognizing emotions Brain Cogn. 106. Brass, M., Schmitt, R. M., Spengler, S. & Gergely, G.
a stereo-electroencephalo-graphic study. Epilepsia 45, 52, 61–69 (2003). Investigating action understanding: inferential
1079–1090 (2004). 82. Kipps, M., Duggins, A. J., McCusker, E. A. & processes versus action simulation. Curr. Biol. 17,
59. Catenoix, H. et al. The role of the anterior insular Calder, A. J. Disgust and happiness recognition 2117–2121 (2007).
cortex in ictal vomiting: a stereotactic correlate with anteroventral insula and amygdala 107. Liepelt, R., Von Cramon, D. Y. & Brass, M. How do we
electroencephalography study. Epilepsy Behav. 13, volume respectively in preclinical Huntington’s infer other’s goals from non stereotypic actions? The
560–563 (2008). disease. J. Cogn. Neurosci. 19, 1206–1217 outcome of context-sensitive inferential processing in
60. Yaxley, S., Rolls, E. T. & Sienkiewicz, Z. J. Gustatory (2007). right inferior parietal and posterior temporal cortex.
responses of single neurons in the insula of the 83. Canessa, N. et al. Understanding others’ regret: Neuroimage 43, 784–792 (2008).
macaque monkey. J. Neurophysiol. 63, 689–700 a fMRI study. PLoS ONE 4, e7402 (2009). 108. Corradi-Dell’Acqua, C., Hofstetter, C. & Vuilleumier, P.
(1990). 84. Trevarthen, C. in Intersubjective Communication and Cognitive and affective theory of mind share the same
61. Scott, T. R., Plata-Salaman, C. R., Smith, V. L. & Emotion in Early Ontogeny (ed. Bråten, S.) 15–46 local patterns of activity in posterior temporal but not
Giza, B. K. Gustatory neural coding in the monkey (Cambridge Univ. Press, 1998). medial prefrontal cortex. Soc. Cogn. Affect. Neurosci.
cortex: stimulus intensity. J. Neurophysiol. 65, 76–86 85. Stern, D. N. The Interpersonal World of the Infant 9, 1175–1184 (2014).
(1991). (Basic Books, 1985). 109. Kilner, J. M. & Frith, C. Action observation: inferring
62. Phillips, M. L. et al. A specific neural substrate for 86. Stern, D. N. Forms of Vitality: Exploring Dynamic intentions without mirror neurons. Curr. Biol. 18,
perceiving facial expressions of disgust. Nature 389, Experience in Psychology, Arts, Psychotherapy, and R32–R33 (2008).
495–498 (1997). Development (Oxford Univ. Press, 2010). 110. Keysers, C. & Gazzola, V. Integration simulation and
63. Phillips, M. L. et al. Neural responses to facial and 87. Di Cesare, G., Di Dio, C., Marchi, M. & Rizzolatti, G. theory of mind: from self to social cognition. Trends
vocal expressions of fear and disgust. Proc. Biol. Sci. Expressing our internal states and understanding Cogn. Sci. 11, 194–196 (2007).
265, 1809–1817 (1998). those of others. Proc. Natl Acad. Sci. USA 112, 111. Engen, H. G. & Singer, T. Empathy circuits. Curr. Opin.
64. Hennenlotter, A. & Schroeder, U. Partly dissociable 10331–10335 (2015). Neurobiol. 23, 275–282 (2012).
neural substrates for recognizing basic emotions: This fMRI study was the first to provide evidence 112. Southgate, V., Johnson, M. H., El Karoui, I. &
a critical review. Prog. Brain Res. 156, 443–456 for the presence of the mirror mechanism for Csibra, G. Motor system activation reveals infants’
(2006). vitality forms in the dorsocentral insula. on‑line prediction of others’ goals. Psychol. Sci. 21,
65. Krolak-Salmon, P. et al. An attention modulated 88. Borra, E. et al. Cortical connections of the macaque 355–359 (2010).
response to disgust in human ventral anterior insula. anterior intraparietal (AIP) area. Cereb. Cortex 18, 113. Gerson, S. A. & Woodward, A. L. The joint role of
Ann. Neurol. 53, 446–453 (2003). 1094–1111 (2008). trained, untrained, and observed actions at the origins
66. Wicker, B. et al. Both of us disgusted in my insula: the 89. Gerbella, M., Belmalih, A., Borra, E., Rozzi, S. & of goal recognition. Infant Behav. Dev. 37, 94–104
common neural basis of seeing and feeling disgust. Luppino, G. Cortical connections of the anterior (F5a) (2014).
Neuron 40, 655–664 (2003). subdivision of the macaque ventral premotor area F5. 114. Woodward, A. L. & Gerson, S. A. Mirroring and the
This study was the first to demonstrate that feeling Brain Struct. Funct. 216, 43–65 (2011). development of action understanding. Phil. Trans.
an emotion and observing an emotion in another 90. Borra, E., Gerbella, M., Rozzi, S. & Luppino, G. R. Soc. B 369, 20130181 (2014).
individual activate the same voxels in the AI. Anatomical evidence for the involvement of the
67. Jabbi, M., Bastiaansen, J. & Keysers, C. A common macaque ventrolateral prefrontal area 12r in Acknowledgements
anterior insula representation of disgust observation, controlling goal-directed actions. J. Neurosci. 31, The authors thank S. Butterfill for his helpful comments on
experience and imagination shows divergent functional 12351–12363 (2011). the manuscript. This works was supported by a grant from
connectivity pathways. PLoS ONE 3, e2939 (2008). 91. Almashaikhi, T. et al. Functional connectivity of insular Fondazione Cariparma, by Inter University Attraction
This paper is an important extension of the efferences. Hum. Brain Mapp. 35, 5279–5294 Program (IUAP) and an European Research Council (ERC)
previous study (reference 61) showing that not only (2014). Advanced Grant to G.R.
feeling one’s own disgust and observing others’ 92. Di Cesare, G. et al. The neural correlates of ‘vitality
disgust but also the internal imagery of disgust form’ recognition: an fMRI study. Soc. Cogn. Affect. Competing interests statement
have a common representation in the AI. Neurosci. 9, 951–960 (2014). The authors declare no competing interests.

NATURE REVIEWS | NEUROSCIENCE ADVANCE ONLINE PUBLICATION | 9

©
2
0
1
6
M
a
c
m
i
l
l
a
n
P
u
b
l
i
s
h
e
r
s
L
i
m
i
t
e
d
,
p
a
r
t
o
f
S
p
r
i
n
g
e
r
N
a
t
u
r
e
.
A
l
l
r
i
g
h
t
s
r
e
s
e
r
v
e
d
.