Transcranial magnetic stimulation highlights the sensorimotor side of empathy for pain

Abstract

Pain is intimately linked with action systems that are involved in observational learning and imitation. Motor responses to one's own pain allow freezing or escape reactions and ultimately survival. Here we show that similar motor responses occur as a result of observation of painful events in others. We used transcranial magnetic stimulation to record changes in corticospinal motor representations of hand muscles of individuals observing needles penetrating hands or feet of a human model or noncorporeal objects. We found a reduction in amplitude of motor-evoked potentials that was specific to the muscle that subjects observed being pricked. This inhibition correlated with the observer's subjective rating of the sensory qualities of the pain attributed to the model and with sensory, but not emotional, state or trait empathy measures. The empathic inference about the sensory qualities of others' pain and their automatic embodiment in the observer's motor system may be crucial for the social learning of reactions to pain.

Full text

Transcranial magnetic stimulation highlights the
sensorimotor side of empathy for pain
Alessio Avenanti
1,2
, Domenica Bueti
1,2
, Gaspare Galati
2,3
& Salvatore M Aglioti
1,2
Pain is intimately linked with action systems that are involved in observational learning and imitation. Motor responses to one’s
own pain allow freezing or escape reactions and ultimately survival. Here we show that similar motor responses occur as a result
of observation of painful events in others. We used transcranial magnetic stimulation to record changes in corticospinal motor
representations of hand muscles of individuals observing needles penetrating hands or feet of a human model or noncorporeal
objects. We found a reduction in amplitude of motor-evoked potentials that was specific to the muscle that subjects observed being
pricked. This inhibition correlated with the observer’s subjective rating of the sensory qualities of the pain attributed to the model
and with sensory, but not emotional, state or trait empathy measures. The empathic inference about the sensory qualities of others’
pain and their automatic embodiment in the observer’s motor system may be crucial for the social learning of reactions to pain.
Empathy helps us to understand feelings and inner states of mind of
others and to share experiences, needs, beliefs and goals
1–3
.Current
neuroscientific models of empathy postulate that a given motor,
perceptual or emotional state of an individual activates corresponding
representations in another individual observing that state
1–3
. Single-cell
recording studies in monkeys show that premotor neurons become
active both during execution of a given action and during observation
of the same action performed by another human or monkey (mirror
neurons)
3–5
. In a similar vein, studies in humans demonstrate that
observation of other individuals acting, being touched or showing
facial emotions induces activity in neural networks that are also
activated when observers act
6–9
, are touched
10
or display the same
emotions
9,11,12
. Thus, empathy may be based on ‘mirror-matching’
simulation of others’ state
3
.
Various painful personal experiences, ranging from being pin-
pricked to feeling an aching phantom limb
13
or suffering from social
loss
14
, are represented in a complex neural network referred to as the
‘pain matrix’. Affectively distressing components (such as unpleasant-
ness) and sensory components (such as localization and intensity) of
the experience of pain are encoded in different nodes of the pain
matrix
15,16
. Although pain is an essentially private subjective experi-
ence
17
, the ability to understand and to experience indirectly the pain
of others is fundamental to social ties
18
. Thus, pain is an interesting
model for testing theories of empathy based on the notion of shared
representations. An empathic matching of others’ pain is suggested by
(i) the observation that a neuron in the human cingulate cortex
increases its firing rate both when pain is inflicted on the observing
subject and when it is inflicted on another person
19
and (ii) the
anecdotal report of a patient in whom genuine pain was evoked by
observation of potentially hurtful stimuli applied to his wife
20
. Recent
fMRI (functional magnetic resonance imaging) studies indicate that
only affective components of the pain matrix are crucial for the
empathic matching of others’ pain
21–23
, suggesting that only emotional
representations of pain are shared between self and others.
Neurophysiological and neuroimaging studies indicate that pain
systems are tightly linked to action systems that can be considered as
the part of the pain matrix
16,24–27
involved in the implementation of
appropriate reactions to actual or potential noxious stimuli. Trans-
cranial magnetic stimulation (TMS) studies, for example, have
demonstrated that actual painful stimuli delivered to the hand bring
about a massive inhibition of corticospinal excitability that affects
upper limb muscles
25,28–30
.
Despite this intimate relationship between pain and action systems,
knowledge about the possible motor mapping of others’ pain is lacking.
Here we explored whether pain and action systems are linked also at a
social level by looking for possible motor correlates of watching and
empathizing with others’ pain. We used single-pulse TMS in healthy
individuals to assess the functional modulation of the corticospinal
system during the observation of painful or non-noxious events shown
on the body of a model. During each observation condition, motor-
evoked potentials (MEPs) to focal TMS of the left motor cortex were
recorded simultaneously from two muscles of the observers’ right
hand: namely, the first dorsal interosseus (FDI) and the abductor
digiti minimi (ADM).
RESULTS
Specific motor mapping of others’ pain
In the first experiment, subjects observed different categories of stimuli:
(i) a needle penetrating the FDI muscle at the dorsal surface of the
right hand between the thumb and index finger (‘Needle in FDI’),
Published online 5 June 2005; doi:10.1038/nn1481
1
Dipartimento di Psicologia, Universita
`
degli studi di Roma La Sapienza, Via dei Marsi 78, 00185 Rome, Italy.
2
Centro Ricerche Neuropsicologia, IRCCS Fondazione Santa
Lucia, Via Ardeatina 306, 00179 Rome, Italy.
3
Dipartimento di Scienze Cliniche e delle Bioimmagini, Universita
`
G. D’Annunzio, Via dei Vestini 31, 66010 Chieti, Italy.
Correspondence should be addressed to S.M.A. (salvatoremaria.aglioti@uniroma1.it).
NAT URE NEUROSCIENCE VOLUME 8
[
NUMBER 7
[
JULY 2005 955
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

(ii) a Q-tip gently moving over and pressing
the same region where the painful stimuli
were delivered (‘Q-tip on FDI’) and (iii) a
needle penetrating a tomato (‘Non-corpor-
eal’; Supplementary Video 1).
We found a significant main effect of
condition for MEPs recorded from the FDI
muscle (F
2,22
¼ 5.64, P ¼ 0.01; Fig. 1a,
Supplementary Fig. 1) underlying the region
where painful or touch stimuli were delivered
to the model. MEPs amplitudes recorded
from FDI were significantly lower in the
‘Needle in FDI’ condition than in the ‘Q-tip
on FDI’ (P ¼ 0.01), ‘Non-corporeal’ (P ¼
0.02) and baseline (t
11
¼3.17, P ¼ 0.009)
conditions, indicating a decrease of motor
excitability during the observation of pain.
By contrast, we found no modulation of
MEPs recorded from the ADM muscle
(F
2,22
¼ 0.56, P ¼ 0.58; Fig. 1b, Supplemen-
tary Fig. 1), which was not involved in the
pain or touch stimulation.
Somatotopic motor mapping of others’
pain
In the second experiment, we further investi-
gated the issue of muscle selectivity by record-
ing MEPs while participants observed a needle
entering the dorsum of a right foot (‘Needle in
foot’) or a Q-tip touching the same region (‘Q-tip on foot’). There was
no significant modulation of MEP amplitude recorded from FDI (F
1,7
¼ 0.04, P ¼ 0.85) or ADM muscles (F
1,7
¼ 0.01, P ¼ 0.91) when the
observers viewed foot stimulations (Fig. 2).
In a third experiment, we recorded MEPs during observation of
painful stimuli delivered to the ADM region of a right hand (‘Needle in
ADM’), and to the dorsum of a right foot (‘Needle in foot’). The results
are consistent with the topographic selectivity seen in experiment 1 and
2(Fig. 3). MEPs recorded from the ADM muscle during the observa-
tion of ‘Needle in ADM’ were significantly lower with respect to the
corresponding baseline (t
11
¼3.50, P ¼ 0.005) and to ‘Needle in foot’
(F
1,11
¼ 5.09, P ¼ 0.045). We did not find any modulation in MEPs
recorded from FDI (F
1,11
¼ 0.02, P ¼ 0.88) (Fig. 3).
Comparisons of subjective ratings in experiments 2 and 3 showed
that ‘painfulness’ of the observed stimuli did not differ between hand
and foot stimulations. Thus, the selectivity of the motor inhibition
cannot reflect differences in the perceived painfulness of the observed
events (Supplementary Fig. 2).
A fourth experiment suggested that the inhibition of hand repre-
sentations were likely to have a cortical origin: the observation of
painful stimuli delivered to the FDI region did not induce changes in
excitability at the muscle or peripheral nerve levels and spinal cord
segments controlling the same muscle (Supplementary Table 1).
Motor mapping of sensory qualities of others’ pain
To explore whether the observed reduction of corticospinal excitability
was related to an empathic mapping of different components of the
pain experience attributed to the model, we analyzed the subjective
judgments about the sensory and affective qualities of the pain ascribed
to the model during needle penetration. These judgments were
obtained in experiment 1 by means of Sensory and Affective subscales
of the McGill Pain Questionnaire
31,32
(MPQ) and two visual analogue
scales (VAS), one for pain intensity and the other for pain unpleasant-
ness. We found that amplitude changes of MEPs recorded from the FDI
muscle were negatively correlated with sensory aspects of the pain
purportedly felt by the model during the ‘Needle in FDI’ condition,
both for the Sensory scale of MPQ (r ¼0.76, P ¼ 0.004) and for pain
intensity VAS (r ¼0.71, P ¼ 0.01; Fig. 4a,c). In contrast, we found no
significant correlation with affective qualities of others’ pain (MPQ
Affective scale: r ¼0.26, P ¼ 0.42; VAS pain unpleasantness: r ¼ 0.22,
130
120
110
100
90
80
70
MEP amplitude (% of baseline)
130
120
110
100
90
80
MEP amplitude (% of baseline)
**
a
b
Figure 1 MEP amplitude with respect to the baseline during observation of ‘Needle in FDI’, ‘Q-tip on
FDI’ and ‘Non-corporeal’ conditions of experiment 1. (a) MEPs recorded from the FDI muscle (black
bars). (b) MEPs recorded from ADM (white bars). Error bars indicate s.e.m. Asterisks (*) indicate
significant post-hoc comparisons (P o 0.02).
130
120
110
100
90
80
MEP amplitude (% of baseline)
130
120
110
100
90
80
MEP amplitude (% of baseline)
a
b
Figure 2 MEPs amplitude with respect to the baseline during observation
of ‘Needle in foot’ and ‘Q-tip on foot’ conditions of experiment 2. (a)MEPs
recorded from the FDI muscle (black bars). (b) MEPs recorded from the ADM
muscle (white bars). Error bars indicate s.e.m.
956 VOLUME 8
[
NUMBER 7
[
JULY 2005 NATURE NEUROSCIENCE
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

P ¼ 0.49; Fig. 4b,d). We did not find any correlations between
amplitude changes of MEPs recorded from the ADM muscle and
qualities of the pain ascribed to the model (Supplementary Table 2).
Seeing painful or unpleasant stimuli may elicit arousal or aversion
(personal distress) reactions
18
. Subjects used VAS to rate arousal and
aversion induced by the different movies presented in experiment 1. We
found no correlation between these self-oriented emotional reactions
and MEP amplitude changes (Supplementary Fig. 3, Supplementary
Table 2).The selective motor mapping of sensory qualities of others’
pain is further suggested by experiment 3 where amplitude changes of
MEPs recorded from the pricked ADM muscle correlated with intensity
(r ¼0.58, P ¼ 0.046; Fig. 5) but not unpleasantness (r ¼0.07,
P ¼ 0.83) VAS scores of the pain attributed to the model. We did not
find any correlations between amplitude changes of MEPs recorded
from the FDI muscle and qualities of the pain ascribed to the model
(Supplementary Table 2).
State and trait empathy during observation of others’ pain
We carried out a fifth experiment to explore whether MEP inhibition
was related to inter-individual differences in specific aspects of empa-
thy. Subjects were delivered TMS pulses during observation of ‘Needle
in FDI’. After TMS sessions, four measures of state empathy, either
‘sensory’ or emotional, were acquired. In particular, subjects were asked
to evaluate along VAS (i) how much they simulated the pain of the
model in their mind (self-oriented), (ii) how intense the pain of the
model was (other-oriented), (iii) how much aversion they felt (self-
oriented) and (iv) how much compassion for the model they felt
(other-oriented). Moreover, we obtained two measures of trait empa-
thy by asking subjects to complete two subscales of the Interpersonal
Reactivity Index (IRI)
33,34
, namely Personal Distress (PD) and
Empathic Concern (EC), each of which corresponds to the notion of
self-oriented and other-oriented empathic emotional reactions.
Results confirmed the specific MEP inhibition (t
15
¼3.09, P ¼
0.008) contingent upon observation of pain found in experiment 1
(Supplementary Fig. 4). We found it important that scores on the two
measures of ‘sensory’ empathy were correlated with the amplitude
changes of MEPs recorded from the FDI muscle during the observation
of ‘Needle in FDI’ (VAS pain simulation: r ¼0.56, P ¼ 0.02; VAS pain
intensity: r ¼0.50, P ¼ 0.05; Fig. 6a,b; Supplementary Table 2).
We did not find any significant correlation between MEP amplitude
change and emotional state (VAS aversion: r ¼ 0.06, P ¼ 0.84; VAS
compassion: r ¼ 0.15, P ¼ 0.57; Fig. 6c,d) or trait (IRI Empathic
Concern: r ¼0.24, P ¼ 0.37; IRI Personal Distress: r ¼0.11,
P ¼ 0.67) empathy scores (Fig. 6e,f). Nor did we find any correlations
between amplitude changes of MEPs recorded from the ADM
muscle and ‘sensory’ or emotional empathy measures (Supple-
mentary Table 2).
DISCUSSION
Only three studies, all using fMRI, have so far explored the neural
underpinnings of empathy for pain in humans
21–23
. Despite several
differences in the experimental protocols, all the previous studies
indicate that only affective nodes in the pain network are concerned
with empathy for pain
21–23
. Here we highlight the sensorimotor side of
empathy for pain by showing a consistent reduction of excitability of
hand muscles during the mere observation of ‘flesh and bone’ painful
stimuli delivered to a model. The observational pain-related inhibition
was robust and conspicuous on several fronts. First, it was specific for
the observation of a needle entering the hand and absent during the
observation of a needle entering feet or non-corporeal objects. Second,
it was confined to the observation of pain and absent during the
observation of harmless tactile stimulation. Third, it was selective for
MEPs recorded from the hand muscle underlying the skin region
penetrated by the needle, and absent for MEPs recorded from a nearby
hand muscle. Fourth, the effect was clearly related to the observer’s
140
120
100
80
60
*
MEP amplitude (% of baseline)
ab
Figure 3 MEP amplitude recorded from the FDI (black bars) and the ADM
(white bars) muscles during the observation conditions of experiment 3.
(a) Observation of ‘Needle in foot’. (b) Observation of ‘Needle in ADM’. Each
painful condition was expressed with respect to the corresponding static
condition (baseline). Error bars indicate s.e.m. Asterisks (*) indicate
significant comparisons (P o 0.05).
30
0
r = –0.76
–30
–60
–2 –1 0 1 2
MEP amplitude
change (%)
MPQ sensory scale
30
0
r = –0.26
–30
–60
–2 –1 0 1 2
MPQ affective scale
30
0
r = –0.71
–30
–60
–2 –1 0 1 2
MEP amplitude
change (%)
VAS
p
ain intensit
y
30
0
r = –0.22
–30
–60
–2 –1 0 1 2
VAS
p
ain un
p
leasantnes
s
ab
cd
Figure 4 Amplitude changes of MEPs recorded from the FDI muscle and
subjective ratings (z-scores) of the pain ascribed to the model during the
‘Needle in FDI’ condition in experiment 1. (a,c) MEP amplitude changes were
negatively correlated with sensory ratings. (b,d) No significant correlation
between MEP amplitude change and affective scores was found. Details in
Supplementary Table 2.
40
0
r = –0.58
–80
–40
–160
–120
–2 –1 0 1 2
MEP amplitude
change (%)
VAS pain intensit
y
40
0
r = –0.07
–80
–40
–160
–120
–2 –1 0 1 2
VAS pain unpleasantness
ab
Figure 5 Amplitude changes of MEPs recorded from the ADM muscle and
subjective ratings (z-scores) of the pain attributed to the model during the
‘Needle in ADM’ condition in experiment 3. (a) Changes in MEP amplitude
negatively correlated with sensory ratings (b) but not with affective scores.
Details in Supplementary Table 2.
NAT URE NEUROSCIENCE VOLUME 8
[
NUMBER 7
[
JULY 2005 957
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

subjective empathetic rating of the sensory, but not affective, qualities
of the pain ascribed to the model. Fifth, the inhibition was related to
measures of state ‘sensory’ empathy scores but not to emotional state
empathy or trait empathy scores.
We suggest the effect may be due to activation of a pain resonance
system
3,20
(Supplementary Note) that extracts basic sensory aspects of
the model’s painful experience (such as source or intensity of a noxious
stimulus) and maps them onto the observer’s motor system according
to topographic rules. This hypothesis is further strengthened by the
high correlations found in experiments 1, 3 and 5 indicating that the
strongest motor inhibition was found in the observers who rated as
most intense the model’s pain.
The evidence presented here for a direct ‘mirror-matching’ simula-
tion of sensory but not of affective features of others’ pain may seem in
sharp contrast with a previous fMRI study. In that study, empathy for
pain was induced by means of arbitrary visual cues signaling an
impending painful stimulus to a loved one
21
. Empathy for pain
brought about an increase of the BOLD (blood oxygen level–depen-
dent) signal in anterior insula and anterior cingulate cortices, which are
part of the affective division of the pain matrix. We find it important
that there was a positive correlation between neural activity and
emotional empathy scores
21
. Neural activity in the affective pain
network was also reported in fMRI studies where subjects observed
pictures
22
or movies
23
in which potentially painful stimuli were
delivered to hands or other human body parts. Our study does not
indicate the absence of activity in the affective nodes of the pain matrix
during observation of ‘flesh and bone’ stimuli. Rather, it indicates that
empathy for pain may rely not only on affective-motivational
21–23
but
also on fine-grained somatic representations. Indeed, our results
suggest a link between the visual representation of others’ painful
experiences and the somatomotor representation of feeling the same
experience
35
. The results also suggest that the functional mechanism
that allows linkage of visual and somatomotor representations may rely
on the inner simulation of specific attributes of the observed stimulus.
Evidence for a somatic resonance system has been provided by a
fMRI study in which an increase of the BOLD signal in the secondary
somatosensory cortex (SII) was found both when the participants were
touched and when they observed someone else being touched
10
.
Moreover, additional structures that may be involved in somatic
processing such as the thalamus, brainstem, parietal cortex and
cerebellum are active when seeing or imagining others’ pain
21,22
.
Notably, affective and sensorimotor nodes of the pain matrix may
not be involved only in processing actual painful stimuli
15,16
but also in
anticipation of somatosensory and painful events administered to
the self
24,36,37
. According to shared representation models
1,2,22
,itis
entirely possible that the somatomotor contagion that may underlie
the corticospinal inhibition reported in our study implies pain antici-
pation in oneself.
As our subjects were informed that no painful stimulus would
actually be delivered at any time, we suggest that the anticipatory
quality of the sensorimotor mapping may be automatic. Moreover, we
posit that the selective embodiment of others’ pain, sensitively more
than emotionally denoted, may be crucial for the social learning of
reactions to painful stimuli in that it may help the observer’s corti-
cospinal system to implement escape or freezing reactions before
painful stimuli are actually experienced.
It may thus be possible to think of at least two forms of empathy
linked to one another in an evolutionary and developmental perspec-
tive. A comparatively simple form of empathy, based on somatic
resonance, may be primarily concerned with mapping external stimuli
onto one’s own body
3–10,38
. A more complex form of empathy, based
on affective resonance, may deal with emotional sharing
11,12,21–23
and
with the evaluation of social bonds and interpersonal relations
21,39
.All
in all, our results indicate that the motor system is an important node
in the complex neural network, recruited not only during the personal
experience of pain
16,21,24–30,36
but also during empathy for others’ pain.
We propose that a direct matching of specific sensory aspects of others’
pain occurs in sensorimotor structures of the pain matrix, whereas
emotional components of others’ painful experiences are coded in the
affective division of the network
21
. Hence, empathy for pain may take
different forms in different nodes of the complex neural network that
represent sensations, feelings and emotions linked to the experience of
pain. Philosophers have emphasized that our bodily sensations are
intrinsically private
17
. However, our findings suggest that, at least in
humans, the social dimension of pain extends even to the very basic,
sensorimotor levels of neural processing.
METHODS
Subjects. For experiments 1–5 there were 12, 8, 12, 8 and 16 participants,
respectively (6, 4, 6, 4 and 8 men). Ages ranged between 20–28, 21–27, 20–27,
20–29 and 19–30. All subjects were right-handed, according to a standard
handedness inventory
40
. Subjects gave written informed consent and were paid
for their participation. The protocol was approved by the ethics committee of
the Fondazione Santa Lucia, Rome and was carried out in accordance with the
ethical standards of the 1964 Declaration of Helsinki. None of the participants
had neurological, psychiatric or other medical problems or had any contra-
indication to TMS
41
.
EMG and TMS recording. In experiments 1, 2, 3 and 5, MEPs were
recorded simultaneously from first right dorsal interosseus (FDI) and abductor
digiti minimi (ADM) by means of a Viking IV (Nicolet Biomedical)
50
0
r = –0.56
–100
–50
–2 –1 0 1 2
MEP amplitude
change (%)
VAS pain simulation
50
0
r = –0.50
–100
–50
–2 –1 0 1 2
VAS pain intensity
50
0
r = –0.15
–100
–50
–2 –1 0 1 2
MEP amplitude
change (%)
VAS aversion
50
0
r = –0.06
–100
–50
–2 –1 0 1 2
VAS compassion
50
0
r = –0.11
–100
–50
–2 –1 0 1 2
MEP amplitude
change (%)
IRI personal distress
50
0
r = –0.24
–100
–50
–2 –1 0 1 2
IRI empathic concern
ab
cd
ef
Figure 6 Amplitude changes of MEPs recorded from the FDI muscle during
‘Needle in FDI’ observation and state (a,b,c,d) and trait (e,f) empathy
measures (z-scores) in experiment 5. (a,b) MEPs amplitude changes were
negatively correlated with VAS pain simulation and VAS pain intensity.
Moreover, both ‘sensory’ state empathy measures independently predict the
inhibition (Supplementary Table 2). (c,d,e,f) No significant correlation
between MEP amplitude change and emotional state or trait empathy
scores was found. Details in Supplementary Table 2.
958 VOLUME 8
[
NUMBER 7
[
JULY 2005 NATURE NEUROSCIENCE
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

electromyograph. EMG signals were band-pass filtered (20 Hz–2.5 kHz,
sampling rate 10 kHz), digitized and stored on a computer for offline analysis.
Pairs of Ag/AgCl surface electrodes were placed over the muscle belly (active)
and over the associated joint or tendon (reference). A circular ground electrode
with a diameter of 30 mm was placed on the dorsal surface of the right wrist. A
figure-eight coil connected to a Magstim Super Rapid Transcranial Magnetic
Stimulator was placed over the left motor cortex. The intersection of the coil
was placed tangentially to the scalp with the handle pointing backward and
laterally at a 451 angle away from the midline. In this way, the current induced
in the neural tissue was directed approximately perpendicular to the line of the
central sulcus, optimal for trans-synaptic activation of the corticospinal path-
ways
42,43
. The coil was moved over the left hemisphere to determine the
optimal position from which maximal amplitude MEPs were elicited in FDI.
The intensity of magnetic pulses was set at 130% of the resting motor
threshold, defined as the minimal intensity of the stimulatory output that
produces MEPs with an amplitude of at least 50 mV with 50% probability
44
.
The complete muscle relaxation before TMS was verified by means of visual
and auditory monitoring of the EMG signal. F and M waves were elicited by
supramaximal electric square-wave pulses (duration, 0.3 ms) at 5-s intervals
and were recorded from the FDI muscle. The electric stimuli were delivered
to the right ulnar nerve at the wrist. Whereas the F wave is considered an index
of spinal cord excitability, the M wave is considered an index of nerve and
muscle excitability
45,46
.
Visual stimuli. In each experiment, different blocks of video clips were
presented on a 19-inch screen located 80 cm from the subjects. In experiment
1, the video clips showed (i) the dorsal view of a right hand (‘Static Hand’),
(ii) a needle penetrating the FDI muscle region (‘Needle in FDI’), a Q-tip
touching the same region (‘Q-tip on FDI’) and (iii) a needle penetrating a
tomato (‘Non-corporeal’; Supplementary Video 1). In experiment 2, video
clips of (i) the dorsal view of a right foot (‘Static foot’), (ii) a needle penetrating
a foot (‘Needle in foot’) and (iii) a Q-tip touching the same region (‘Q-tip on
foot’) were shown. In experiment 3, the video clips showed (i) a static view of
the ADM region of a right hand (‘Static ADM’), (ii) a static view of the dorsal
surface of a right foot (‘Static foot’), (iii) a needle penetrating the ADM muscle
of a right hand (‘Needle in ADM’) and (iv) a needle entering the dorsal surface
of a right foot (‘Needle in foot’). In experiments 4 and 5, the video clips showed
the ‘Needle in FDI’ and the ‘Static Hand’ conditions used in experiment 1.
Previous neurophysiological studies report that observation of moving body
parts brings about an increase of corticospinal excitability
6,47,48
and that
observation of a hand using tools elicits an activation of primary motor
cortex
49
. To avoid such effects in the present study, we checked that no
movements of hand, foot or tomato were evoked by pinprick or touch stimuli.
In a similar vein, we checked that in none of the videos was the holder of the
syringe or of the Q-tip visible.
Procedure. The experiments were programmed using Psychophysics Toolbox
(http://www.psychotoolbox.org) and Matlab (http://www.mathworks.com)
software to control sequence and duration of video clips and to trigger TMS
and EMG recording. Each type of video clip was presented in separate blocks.
Five (18 trials), four (18 trials), four (15 trials) and two (18 trials) blocks were
performed in experiments 1, 2, 3 and 5, respectively. On each trial, a magnetic
pulse was randomly delivered between 200 and 600 ms before the end of the
movie to avoid any priming effects that could affect MEP size. A blank screen
was shown for 7.2 s in the intertrial intervals. The choice of long intertrial
intervals was based on a study demonstrating that TMS delivered for 1 h at
0.1 Hz frequency did not induce any change of excitability
50
. In experiments 1
and 2, the first and the last block served as baseline and consisted of video clips
showing ‘Static Hand’ and ‘Static foot’, respectively. The order of the other
blocks was counterbalanced. In experiment 3, ‘Static ADM’ and ‘Static foot’
served as baseline for ‘Needle in ADM’ and ‘Needle in foot’, respectively. In
experiment 4, 20 F and M waves were recorded in two blocks. In experiment 5,
‘Static Hand’ served as baseline for ‘Needle in FDI’. The order of the different
blocks in experiments 3–5 was counterbalanced. In all the experiments, a
central cross (1,000 ms) indicated the begin of a trial and initiated EMG
recording. The duration of each video was 1,800 ms. In all observation
conditions, participants were asked to watch carefully and pay attention to
the events shown in the video clips. Moreover, in the conditions involving
observation of body parts, participants were instructed to focus on what the
stimulated individual may have felt.
Subjective reports. After TMS sessions of experiment 1, subjects were
presented with all movies and asked to judge the arousal and aversion (personal
distress) induced by each movie by marking a vertical, 10-cm visual analogue
scale (VAS) with 0 cm indicating ‘no effect’ and 10 cm ‘maximal effect
imaginable’. In experiments 1, 2, 3 and 5, VAS were used to rate the intensity
and unpleasantness of the pain purportedly experienced by the model when
being injected or touched on the hand or the foot. In experiment 1, qualities of
the pain ascribed to the model were also measured by means of the Italian
version
32
of the McGill Pain Questionnaire (MPQ)
33
. Sensory and Affective
subscales of MPQ were used.
Measures of state and trait empathy. After TMS sessions of experiment 5, four
measures of state empathy (sensory or emotional, self-oriented or other-
oriented) were acquired. Subjects were shown the ‘Needle in FDI’ movies
and asked to evaluate along a VAS (i) their inner mental simulation of the
model’s pain (sensory, self-oriented), (ii) the intensity of the pain they
attributed to the model (sensory, other-oriented), (iii) the aversion they felt
(emotional, self-oriented) and (iv) the compassion they felt for the model
(emotional, other-oriented). Two measures of emotional trait empathy were
obtained by asking subjects to complete two subscales of the Italian version
33
of
the Interpersonal Reactivity Index (IRI)
34
, namely Empathic Concern and
Personal Distress. The Empathic Concern subscale assesses the tendency to
experience feelings of sympathy and compassion for others in need, and the
Personal Distress subscale assesses the tendency to experience distress and
discomfort in response to extreme distress in others.
Neurophysiological measures analysis. Data were processed offline. Trials with
EMG activity before TMS (less than 5%) were discarded from analysis. In
experiments 1, 2, 3 and 5, mean MEP amplitude in each condition was
measured peak-to-peak (in mV). In experiment 1, MEP amplitude in the first
and last block was comparable for both FDI (t
11
¼ 0.51, P ¼ 0.62) and ADM
(t
11
¼0.18, P ¼ 0.86) muscles. The same result was obtained in experiment 2
(FDI: t
7
¼ 1.02, P ¼ 0.34; ADM: t
7
¼ 1.40, P ¼ 0.20). Therefore, in both
experiments, the baseline was obtained by averaging MEP values from the first
and the last block. In experiments 1, 2, 3 and 5, mean MEP amplitude in each
dynamic observation condition was expressed as percentage of the correspond-
ing baseline (‘Static Hand’ in experiment 1 and 5, ‘Static foot’ in experiment 2,
and ‘Static ADM’ and ‘Static foot’ in experiment 3). For each muscle,
normalized MEP values were analyzed by means of repeated-measure one-
way ANOVAs with Condition as the main factor with three levels (‘Needle in
FDI’, ‘Q-tip on FDI’, ‘Non-corporeal’) in experiment 1, two levels (‘Needle in
foot’, ‘Q-tip on foot’) in experiment 2 and two levels (‘Needle in foot’, ‘Needle
in ADM’) in experiment 3. Post-hoc comparisons were carried out by means of
the Newman-Keuls test. In experiment 4, mean amplitudes of F and M waves
(in mV) were measured peak-to-peak. F and M waves recorded during ‘Needle
in FDI’ condition were normalized using ‘Static Hand’ condition. Normalized
MEP (experiments 1, 2, 3 and 5) and F and M wave (experiment 4) values were
compared against the value of 1 (baseline) by means of one-sample t-tests.
Subjective measures analysis. In experiment 1, mean VAS ratings for arousal
and aversion induced by the different types of video clip were analyzed by
means of two repeated-measure one-way ANOVAs with type of movie as main
factor (‘Needle in FDI’, ‘Q-tip on FDI’, ‘Non-corporeal’; Supplementary
Fig. 3). We compared the VAS ratings of pain qualities ascribed to the model
by means of two one-way ANOVAs with type of movie as main factor (‘Needle
in FDI’, ‘Q-tip on FDI’ in experiment 1; ‘Needle in foot’, ‘Q-tip on foot’, in
experiment 2; ‘Needle in ADM’, ‘Needle in foot’ in experiment 3), one for pain
intensity and one for pain unpleasantness (Supplementary Fig. 2). Post-hoc
comparisons were carried out by means of the Newman-Keuls test.
Correlation analysis. In each experiment, a correlation analysis between
neurophysiological and subjective measures (experiments 1, 2, 3 and 5) and
state and trait empathy scores (experiment 5) was performed for observation
conditions in which MEP amplitude was significantly different from the
NAT URE NEUROSCIENCE VOLUME 8
[
NUMBER 7
[
JULY 2005 959
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience

baseline. An index of MEP amplitude change with respect to the baseline was
computed. The only two conditions significantly different from the baseline
were ‘Needle in FDI’ in experiments 1 (t
11
¼3.17, P ¼ 0.009) and 5 (t
15
¼
3.09, P ¼ 0.008) and ‘Needle in ADM’ in experiment 3 (t
11
¼3.50, P ¼
0.005). Indices of MEP amplitude change were computed as follows: amplitude
during observation of the pain condition minus amplitude during observation
of the static hand condition divided by the average of the same two conditions.
Pearson correlation coefficients between indices of amplitude change of MEPs
recorded from each muscle and subjective reports were computed in each
experiment. In experiment 5, we carried out a standard regression analysis to
test whether sensory state empathy scores were independent predictors of
amplitude change of MEPs recorded from FDI during the observation of
‘Needle in FDI’.
Note: Supplementary information is available on the Nature Neuroscience website.
ACKNOWLEDGMENTS
This research was supported by grants from the Ministero Istruzione Universita
`
e
Ricerca and Finanziamento Italiano Ricerca di Base, Italy, both awarded to S.M.A.
COMPETING INTERESTS STATEMENT
The authors declare that they have no competing financial interests.
Received 12 April; accepted 17 May 2005
Published online at h ttp://www.nature.com/natureneuroscience/
1. Preston, S.D. & de Waal, F.B.M. Empathy: its ultimate and proximate bases. Behav.
Brain Sci. 25,1–71(2002).
2. Decety, J. & Jackson, P.L. The functional architecture of human empathy. Behav. Cogn.
Neurosci. Rev. 3, 71–100 (2004).
3. Gallese, V. The manifold nature of interpersonal relations: the quest for a common
mechanism. Philos. Trans. R. Soc. Lond. B Biol. Sci. 358, 517–528 (2003).
4. Gallese, V., Fadiga, L., Fogassi, L. & Rizzolatti, G. Action recognition in the premotor
cortex. Brain 119, 593–609 (1996).
5. Rizzolatti, G., Fogassi, L. & Gallese, V. Neurophysiological mechanisms underlying the
understanding and imitation of action. Nat. Rev. Neurosci. 2, 661–670 (2001).
6. Fadiga, L., Fogassi, L., Pavesi, G. & Rizzolatti, G. Motor facilitation during action
observation: a magnetic stimulation study. J. Neurophysiol. 73, 2608–2611
(1995).
7. Iacoboni, M. et al. Cortical mechanisms of human imitation. Science 286, 2526–2528
(1999).
8. Buccino, G. et al. Action observation activates premotor and parietal areas in a
somatotopic manner: an fMRI study. Eur . J. Neurosci. 13, 400–404 (2001).
9. Leslie, K.R., Johnson-Frey, S.H. & Grafton, S.T. Functional imaging of face and
hand imitation: towards a motor theory of empathy. Neuroimage 21,601–607
(2004).
10. Keysers, C. et al. A touching sight: SII/PV activation during the observation and
experience of touch. Neuron 42, 335–346 (2004).
11. Carr, L. et al. Neural mechanisms of empathy in humans: a relay from neural systems for
imitation to limbic areas. Proc. Natl. Acad. Sci. USA 100, 5497–5502 (2003).
12. Wicker, B. et al. Both of us disgusted in My insula: the common neural basis of seeing
and feeling disgust. Neuron 40, 655–664 (2003).
13. Flor, H. Phantom-limb pain: characteristics, causes, and treatment. Lancet Neurol. 1,
182–189 (2002).
14. Eisenberger, N.I., Lieberman, M.D. & Williams, K.D. Does rejection hurt? An fMRI study
of social exclusion. Science 302, 290–292 (2003).
15. Rainville, P. Brain mechanisms of pain affect and pain modulation. Curr. Opin.
Neurobiol. 12, 195–204 (2002).
16. Ingvar, M. Pain and functional imaging. Philos. Trans. R. Soc. Lond. B Biol. Sci. 354,
1347–1358 (1999).
17. Wittgenstein, L. Philosophical Investigations (Blackwell, Oxford, 1963).
18. Williams, A.C. Facial expression of pain: an evolutionary account. Behav. Brain Sci. 25,
439–488 (2002).
19. Hutchison, W.D. et al. Pain-related neurons in the human cingulate cortex. Nat.
Neurosci. 2, 403–405 (1999).
20. Bradshaw, J.L. & Mattingley, J.B. Allodynia: a sensory analogue of motor mirror neurons
in a hyperaesthetic patient reporting instantaneous discomfort to another’s perceived
sudden minor injury? J. Neurol. Neurosurg. Psychiatry 70, 135–136 (2001).
21. Singer, T. et al. Empathy for pain involves the affective but not sensory components of
pain. Science 303, 1157–1162 (2004).
22. Jackson, P.L., Meltzoff, A.N. & Decety, J. How do we perceive the pain of others? A
window into the neural processes involved in empathy. Neuroimage 24, 771–779
(2005).
23. Morrison, I., Lloyd, D., di Pellegrino, G. & Roberts, N. Vicarious responses to pain in
anterior cingulate cortex: is empathy a multisensory issue? Cogn. Affect. Behav.
Neurosci. 4, 270–278 (2004).
24. Wager, T.D. et al. Placebo-induced changes in FMRI in the anticipation and experience
of pain. Science 303, 1162–1167 (2004).
25. Farina, S., Tinazzi, M., Le Pera, D. & Valeriani, M. Pain-related modulation of the human
motor cortex. Neurol. Res. 25, 130–142 (2003).
26. Juottonen, K. et al. Altered central sensorimotor processing in patients with complex
regional pain syndrome. Pain 98, 315–323 (2002).
27. Saitoh, Y., Shibata, M., Sanada, Y. & Mashimo, T. Motor cortex stimulation for phantom
limb pain. Lancet 353, 212 (1999).
28. Farina, S. et al. Transient inhibition of the human motor cortex by capsaicin-induced
pain. A study with transcranial magnetic stimulation. Neurosci. Lett. 314,97–101
(2001).
29. Svensson, P., Miles, T.S., McKay, D. & Ridding, M.C. Suppression of motor evoked
potentials in a hand muscle following prolonged painful stimulation. Eur. J. Pain 7,
55–62 (2003).
30. Urban, P.P. et al. Different short-term modulation of cortical motor output to distal and
proximal upper-limb muscles during painful sensory nerve stimulation. Muscle Nerve
29, 663–669 (2004).
31. Maiani, G. & Sanavio, E. Semantics of pain in Italy: the Italian version of the McGill Pain
Questionnaire. Pain 22, 399–405 (1985).
32. Melzack, R. The McGill Pain Questionnaire: major properties and scoring methods. Pain
1, 277–299 (1975).
33. Bonino, S., Lo Coco, A. & Tani, F. Empatia. I Processi di Condivisione delle Emozioni
(Giunti, Florence, Italy, 1998).
34. Davis, M.H. Empathy: A Social Psychological Approach (Westview Press, Madison,
Wisconsin, 1996).
35. Adolphs, R. Neural systems for recognizing emotion. Curr. Opin. Neurobiol. 12,169–
177 (2002).
36. Porro, C.A., Cettolo, V., Francescato, M.P. & Baraldi, P. Functional activity mapping of
the mesial hemispheric wall during anticipation of pain. Neuroimage 19, 1738–1747
(2003).
37. Ploghaus, A., Becerra, L., Borras, C. & Borsook, D. Neural circuitry underlying
pain modulation: expectation, hypnosis, placebo. Trends Cogn. Sci. 7,197–200
(2003).
38. Meltzoff, A.N. & Decety, J. What imitation tells us about social cognition: a rapproche-
ment between developmental psychology and cognitive neuroscience. Philos. Trans. R.
Soc. Lond. B Biol. Sci. 358, 491–500 (2003).
39. Leibenluft, E., Gobbini, M.I., Harrison, T & Haxby, J.V. Mothers’ neural activation in
response to pictures of their children and other children. Biol. Psychiatry 56, 225–232
(2004).
40. Oldfield, R.C. The assessment and analysis of handedness: the Edinburgh inventory.
Neuropsychologia 9, 97–113 (1971).
41. Wasserman, E.M. Risk and safety of repetitive transcranial magnetic stimulation: report
and suggested guidelines from the International Workshop on the Safety of Repetitive
Transcranial Magnetic Stimulation, June 5–7, 1996. Electroenceph. Clin. Neuro-
physiol. 108,1–16(1998).
42. Brasil-Neto, J.P. et al. Optimal focal transcranial magnetic activation of the human
motor cortex: effects of coil orientation, shape of the induced current pulse, and
stimulus intensity. J. Clin. Neurophysiol. 9, 132–136 (1992).
43. Mills, K.R., Boniface, S.J. & Schubert, M. Magnetic brain stimulation with a double coil:
the importance of coil orientation. Electroencephalogr. Clin. Neurophysiol. 85,17–21
(1992).
44. Rossini, P.M. et al. Non-invasive electrical and magnetic stimulation of the brain,
spinal cord and roots: basic principles and procedures for routine clinical application.
Report of an IFCN committee. Electroencephalogr. Clin. Neurophysiol. 91,79–92
(1994).
45. Eisen, A. Electromyography in disorders of muscle tone. Can. J. Neurol. Sci. 14,
501–505 (1987).
46. Kimura, J. Electrodiagnosis in Diseases of Nerve and Muscle (F.A. Davis Company,
Philadelphia, 1993).
47. Maeda, F., Kleiner-Fisman, G. & Pascual-Leone, A. Motor facilitation while observing
hand actions: specificity of the effect and role of observer’s orientation. J. Neurophysiol.
87, 1329–1335 (2002).
48. Gangitano, M., Mottaghy, F.M. & Pascual-Leone, A. Modulation of premotor mirror
neuron activity during observation of unpredictable grasping movements. Eur. J.
Neurosci. 20, 2193–2202 (2004).
49. Ja
¨
rvela
¨
inen, J., Schu
¨
rmann, M. & Hari, R. Activation of the human primary motor cortex
during observation of tool use. Neuroimage 23, 187–192 (2004).
50. Chen, R. et al. Depression of motor cortex excitability by low-frequency transcranial
magnetic stimulation. Neurology 48, 1398–1403 (1997).
960 VOLUME 8
[
NUMBER 7
[
JULY 2005 NATURE NEUROSCIENCE
ARTICLES
© 2005 Nature Publishing Group http://www.nature.com/natureneuroscience