Feeling into other minds: The modulating effects of stimulus intensity on early somatosensory processing during observation of others’ pain

Feeling into other minds:

The modulating effects of stimulus intensity on

early somatosensory processing during

observation of others’ pain

Kim van Dijk, BSc

Master Brain and Cognitive Sciences, University of Amsterdam

Supervisor: Ir. C. C. de Vos

Institute for Biomedical Technology and Technical Medicine, University of Twente

UvA representative : Drs. M. Keestra

Institute for Interdisciplinary Science, University of Amsterdam

September, 2012

Abstract

Without direct access to another’s bodily experience we seem to “feel into their minds” without much effort. Simulation Theory proposes that is it simulation of a model’s experience that underlies this empathic skill. In recent years, neuroscientific data has indicated that the mirror neuron system is responsible for these simulations by mirroring both the affective and the sensory consequences of others’ feelings. Controversially, the sensory aspect of empathy is not always found and its existence is being questioned. However, there exist indications that stimulus intensity is an important modulator of empathic somatosensory processing. This study examined modulation of early somatosensory evoked potentials during observation of touch and different levels of pain. It was found that that a very early component of the sensory aspect of empathy (N20 modulation) is made possible by simulation of a model’s state, build upon activation of the MNS in the somatosensory cortex. Also, it is proposed that observation of pain in others might be used in discriminating pain intensities and facilitating appropriate action preparation.

Introduction

Empathy, the ability to understand and share feelings of another (Oxford dictionaries, 2012), is a fundamental component of our social interaction. Without direct access to another’s bodily experience we seem to “feel into their minds” without much effort. The question of how this “feeling into” is realized, given that one cannot directly experience another’s mental life, is vigorously debated in many fields of research. Theories of the mechanisms behind this mindreading are generally grouped into two opposing ideas; Theory-Theory (TT; Churchland, 1991) and Simulation Theory (ST; Gallese & Goldman, 1998 and Goldman, 2006). TT states that minds of others are understood by applying a set of rules that offers an explanation of observed behavior. ST proposes that rather than knowledge of a theoretical framework, it is simulation in one’s own mind that underlies our mindreading capacities. According to ST, an observer comes to share a model’s state of mind by means of automatic inner imagination of the experience of the observed model. It states that perceptual, motor, and introspective states that are gained by one’s own experience are re-enacted during simulation when one person observes another (Barsalou, 2008), rather than being inferred from a theory. Indeed, recent experimental data indicates that ‘as if body loops’ (Damasio, 1999) underlie empathy and play an important role in not only understanding emotional feelings of another but also mapping sensations of others unto our own sensory system (see Gallese, 2001 and Ogino et al., 2007) in an automatic manner (Avenanti et al., 2005 and Lamm et al., 2007). Hence, the definition of empathy, originally revolving around emotional contagion, has been expended to include the sharing of somatic sensations by mere observation (e.g. Preston & de Waal, 2002; Gallese, 2003; Decety & Jackson, 2004 and Avenanti & Aglioti, 2006).

Empathy is thought to have emerged from the tight coupling between action and perception in the brain. As explained by the common-coding principle (Prinz, 1997), perceptual contents and action plans are coded in a common representational system (also see Hommel et al., 2001). Empathy could then be understood as the automatic activation of autonomic and somatic responses that are associated with the perception of a model’s state (see Preston & de Waal, 2002). These simulations serve to predict, recognize and understand events that are likely to occur and are thus relevant not only for the model but also for the subject. From an evolutionary perspective, empathy could then be seen as the selective embodiment of others’ experiences that provides a mechanism for action preparation through simulated predictions that reside in the same system as perceptual processes (Barsalou, 2009). Just like emotions are thought to generate automatic action readiness (Frijda, Kuipers & ten Schure, 1989), recognizing emotions in others might make responses to

potential harmful situation faster and more efficient. Interestingly, the discovery of the mirror neuron system (MNS) has paved the way for empirical evidence for mental simulation of observed states in a model. Although MNS research has relevance for other subjects as well (e.g. imitation, language, theory of mind), this neuroscientific finding has tipped the balance in favour of ST in empathy research (see Barsalou, 2009) by providing a neuronal correlate for mindreading simulation.

Empathy and the mirror neuron system (MNS)

In ST, empathy is thought to be built on (partially) shared neuronal representations of a current state of mind between a model and an observing individual. Although bodily responses are usually inhibited in the observer, ST postulates that there should be an overlap between brain activity patterns in the model and the beholder (see Goldman, 2006). The MNS seems to be responsible for the simulation of others’ pain by means of mirroring the affective consequences as well as somatopical mapping of the sensory aspects of pain (e.g. Lamm et al., 2007 and Loggia, Mogil & Bushnell, 2008). Although first discovered in the domain of motor action (Rizzolatti et al., 2001), mirroring brain activity is also apparent when viewing facial expression depicting emotions (e.g. Carr et al., 2003), watching a model being touched (e.g. Martínez-Jauand et al., 2012; Bufalari et al., 2007 and Keysers et al., 2004), or during perception of others’ pain (e.g. Gallese, 2001; Singer et al., 2004 and Avenanti et al., 2005, 2006). The MNS can be seen as the physical realization of action-perception coupling (Prinz, 1997), previously proposed by psychologists to account for action coding through its perceptual effects (e.g. Buccino, Binkofski & Riggio, 2004). Because of the remarkable overlap between brain activity during firsthand experience and empathic mindreading, observation of pain and touch are excellent ways to study empathy.

Per definition, mirror neurons are used for both the firsthand experience of a given situation as well as for simulating that experience during mere observation. Therefore, empathic mirror activity has the potential to alter simultaneous processing of firsthand experience of somatic stimulation (e.g. see Martínez-Jauand et al., 2012; Bufalari et al., 2007 and Avenanti et al., 2006) and this paradigm is often used to study the underlying mechanism and the effects of empathy.

Firsthand experience of pain

The twofold pain network of the brain used for experiencing one’s own pain (Melzack, 1999) consists of an affective component represented by activity in the anterior cingulated cortex (ACC) and the anterior insula (AI), as well as a sensory aspect of pain perception found much earlier in the somatosensory cortices (Peyron et al., 2000). The anatomical dissociation between affective and sensory processing of pain correlates with the dissociation between ascribed pain unpleasantness

and pain intensity respectively (see Rainville, 2002). Whereas pain unpleasantness ratings correlate with activity in the caudal ACC (Tölle et al., 1999), fluctuations of pain intensity cause measurable changes in the primary somatosensory cortex (SI; Hofbauer et al., 2001). Although the contribution of sensory areas to pain processing in the brain has been questioned on various occasions (e.g. Rainville et al., 1997), accumulating evidence supports the notion that SI plays an eminent role in coding the sensory features of pain (see Bushnell et al., 1999; Kanda et al., 2000 and Hofbauer et al., 2001). Specifically, SI is thought to participate in the discriminative perception of pain intensity (Timmermann et al., 2001), and both SI and the secondary somatosensory cortex (SII) appear to be activated when attention is directed to localization of pain (Kulkarni et al., 2005).

Considering ST, it is expected that activations in both parts of the pain network can be found during observation of others’ pain as well. Since firsthand processing of affective consequences of pain as well as action preparation and sensory consequences of pain are built upon activity in the ACC/AI, sensorimotor, and somatosensory cortices respectively, empathy for pain is expected to make use of these same brain regions through the MNS.

Empathy for pain and touch

On some occasions, mirroring brain activity is indeed found for both affective and sensory processing, indicating that empathy is comprised of emotional (e.g. Singer et al., 2004 and Jackson, Meltzoff & Decety, 2005) and somatic (e.g. Keysers et al., 2004 and Avenanti et al., 2005) resonance. However, similar to what is seen in literature on the firsthand experience of pain, the role of SI in an observer of other’s pain is also being disputed. Although shared neuronal representations have been found in the MNS, there still is some controversy over which brain areas are involved in constituting empathy for pain. Brain imaging experiments using functional magnetic resonance imaging (fMRI) have shown activations of the ACC and AI when people are asked to imagine others’ pain, but have failed to show any mirror neuron activity in sensorimotor areas (e.g. Singer et al., 2004; Jackson, Meltzoff & Decety, 2005 and Morrison et al., 2004), casting doubt on theories of action-perception coupling and leading to the conclusion that empathy is exclusively based on emotional contagion (see Singer et al., 2004). However, findings from experiments using transcranial magnetic stimulation (TMS) and electroencephalography (EEG) indicate that observation of pain modulates corticospinal as well as cortical excitability, which is strongly correlated with subjective ratings of sensory intensity but not with any affective components (Avenanti et al., 2005 and Bufalari et al., 2007 resp.). It is proposed that next to processing of firsthand experience of painful stimuli, all the before mentioned brain areas also assist in processing anticipation of somatic events in oneself (Avenanti et al., 2005

and Wager et al., 2004). This is what would be expected based on the theory that observing endangerment in others results in mediation of a freeze or escape response (e.g. see Singer et al., 2004). Direct mapping of others’ pain occurring in sensorimotor areas could facilitate appropriate motor responses during anticipation of pain in oneself (Avenanti et al., 2005). Likewise, it is found that mere pain observation can alter the susceptibility of personal sensation and even firsthand experience of pain (Godinho et al., 2006), brought forth by changes in the thalamocortical gate (Steriade & Llinas, 1988). This action preparation can then create an adaptive advantage in an organism (see Gallese & Goldman, 1998).

The absence of somatic resonance during empathy is not the only explanation of why many imaging studies have not found mirroring activity in sensorimotor areas. Lately, a promising attempt to resolve the controversy around sensory processing of others’ pain has appeared in literature.

Stimulus-driven differences in empathy literature

A likely explanation for the discrepancy in the experimental data was offered by a TMS experiment showing that stimulus-related differences can cause subtle differential sensory modulation (Avenanti et al., 2006). Whereas most fMRI experiments had used indirect or static references to others’ pain (e.g. Singer et al., 2004), tests that did show sensorimotor involvement used videos of deep needle penetrations and directed attention to stimulus intensity (Avenanti et al., 2005, 2006). Indeed, in a TMS experiment where stimuli from one of the static imaging experiments (Jackson, Meltzoff & Decety, 2005) were used, no relation between stimulus intensity and SI modulation was found (Cheng et al., 2008). Considering the adaptive advantage of empathic skills, it makes sense that stimulus intensity modulates processes of somatic resonance. After all, observation of touch, mild pain, or extreme pain should lead to a respective increase of somatic anticipation since the predictive contents of these situation is likely to be different. Also, given SI’s response preference to sensory features such as pain intensity and localization, it is likely that indirect references to pain don’t get to be processed by somatosensory areas due to a lack of (vivid) visual input. For these reasons, it is very thinkable that this controversy in empathy literature (and perhaps also in general pain theories) is based on differences in stimulus material. However, not much data on how exactly stimulus intensity modulates early processing in somatosensory areas exists.

Experiment

This experiment will look at the effects of stimulus intensity on processing in SI during the observation of touch and pain anticipation, as well as during observation of painful and non-painful situations, using somatosensory evoked potentials (SSEPs) measured with EEG. When provided with

activation along the somatosensory pathway (e.g. by electrical stimulation of a peripheral nerve) measurable SSEP waves are generated. Evoked potentials can be recorded from SI during the first 100 ms after stimulation. Several characteristic peaks in the SSEP waves can be seen (see Allison, McCarthy, Wood, & Jones, 1991) and importantly, they are known to be modulated by empathy (e.g. Martínez-Jauand et al., 2012; Bufalari et al., 2007 and Avenanti et al., 2006). So far, short latency SSEP component P45, originating from the crown of SI around 45 ms post-stimulus (Allison et al., 1992) seems to be the only component that differentiates between observation of painful and non-painful stimuli, showing either a decrease or an increase in amplitude while observing others’ pain (de Vos, de Jonge & van Putten, in preparation and Bufalari et al., 2007 resp.) Contrary to this, it has recently been shown that SSEPs within this timeframe do not differentiate between painful and non-painful stimulation (Martínex-Jauand et al., 2012). By presenting both touch, mildly non-painful, and extremely painful situations this experiment could shed more light on these contradictory findings. Also, by looking at short latency SSEP component N20, support could be provided for the recently developed theory that not only P45 (e.g. Godinho et al., 2006) but also very early SSEPs are modulated by empathy (de Vos, de Jonge & van Putten, in preparation). Originating from the contra lateral SI (Allison et al., 1991) this negativity appears rather robust after peripheral stimulation at about 20 ms post-stimulus. In addition, SSEP component N60, thought to arise from SI as well as the supplementary motor area (Barba et al., 2002) is examined during observation of different pain intensities. Furthermore, in an effort to come up with a model that explains any relation between neural, behavioral, and psychological responses, SSEP modulation will be linked to trait empathy as well as state empathy scores.

Lastly, several case studies of patients with an implanted spinal cord stimulator for treatment of chronic pain will be included. Similarly to the SI modulation found during pain observation, chronic pain patients show deviant SI activation in response to, for example, touch (Peyron et al., 2004). Moreover, pain syndromes cause changes in excitability of the somatosensory cortices (Pleger et al., 2004) and most importantly, these changes correlate with the intensity of the perceived pain, and fully disappear after successful treatment (Pleger et al., 2005 and Maihofner et al., 2004). Although it has not yet been tested, it is possible that these patients show different patterns of SI modulation during both baseline and pain or touch observation.

Methods

Sixteen healthy volunteers (9 males, 7 females) aged between 18 and 29 years (mean = 24, standard deviation (SD) = 2.9) participated in this study. Also, 3 patients (2 males, 1 female) with implanted spinal cord stimulators for treatment of chronic pain, aged between 56 and 64 years (mean = 61, SD = 4.6), as well as 2 aged matched controls (2 males; mean age = 69, SD = 3.5) were included. The patients were asked to set their spinal cord stimulators to burst stimulation 14 days prior to the experiment which, in these patients, led to low or even absent pain levels during the experiment. All other participants reported not to experience any pain at the time of the measurements. All participants were right handed and gave their written informed consent. The procedures were approved by the local medical ethics committee.

Stimuli

The videos depicted a total of 7 different observational conditions, all showing the right hand of a patient in the operating room, viewed from a first person perspective. The first 10 videos were shown 10 times each using 2 separate but comparable videos for each condition in order to avoid habituation effects. The videos for the first 5 conditions lasted 5 seconds each and consisted of a: 1) ventral or lateral view on a static hand (hand); 2) lateral view of a hand being disinfected with wet gauze (touch); 3) lateral view of a needle penetrating the ventral part of the wrist (needle1); 4) ventral view of a needle deeply penetrating the wrist (needle2) and 5) ventral view of an incision in either the palm of the hand or the ring finger (incision). The sixth condition included 3 recordings that depicted more intense pain and were therefore shown only 3 times each, at the end of the presentation: 6.1) lateral view of a deep incision being widened using scissors (extreme1); 6.2) dorsal view of an extraction of a piece of bone from the wrist (extreme2) and 6.3) ventral view of a metal tube penetrating the hand from the palm till the ring finger (extreme3). The 7th _{and last condition}

was comprised of a 5 second video showing a needle coming close to, but not touching the hand (anticipation) and was also shown 3 times.

Subjective reports

In order to assess the relevant psychological traits and states of the participants, they were presented with three different questionnaires. Firstly, participants were asked to fill out a Dutch version (Ponnet et al., 2005) of the Interpersonal Reactivity Index (IRI; Davis, 1983). This questionnaire is used as an empathy scale, supplying a trait empathy score by use of 4 subscales:

Perspective Taking (PT), Fantasy scale (FS), Empathic Concern (EC), and Personal Distress (PD). Secondly, participants rated all 14 previously viewed video clips separately using a 10-cm visual analogue scale (VAS). To examine state empathy the following questions were asked; 1) How intense is the pain that the model experiences? (sensory, model-oriented); 2) How intense is the pain that you experience? (sensory, self-oriented); 3) How much empathy do you feel for the model? (affective, model-oriented); 4) How unpleasant is this video to you? (affective, self-oriented) and 5) How unpleasant is this situation for the model? (affective, model-oriented, subjective). Lastly, participants filled out a Dutch version (Verkes, et al., 1989) of the McGill Pain Questionnaire (MPQ; Melzack, 1975). Participants were asked to indicate the pain ascribed to the model in 3 different videos, namely: one of the hand videos, one needle2 video, and a video from the extreme condition. By use of the MPQ participants were asked to indicate which pain words (describing the sensory, the affective, and the evaluative aspects of pain) fitted each situation best.

Procedure

The experiment lasted about 2 hours. Participants were seated approximately 70 cm away from a 15.6 inch computer screen that was used for presentation of the video clips. Firstly, participants filled out the IRI. Secondly, a 3-minute baseline SSEP recording was made while the room was semi-dark and no presentation was shown on the screen. Thirdly, the experimental video was presented while simultaneous SSEP stimulation was applied, also with the experimental room being semi-darkened. Participants were informed on the nature of the video clips; they depicted exclusively right hands of fully conscious and for most videos non-anaesthetized patients. Participants were instructed to imagine how the situation was experienced by the model and to focus on the video rather than the SSEP stimulation. Also, they were asked to blink during the blank screen in between videos. The total duration of the presentation was approximately 15 minutes. All videos were shown in random order and semi-counterbalanced over 3 different orders. The last category of videos was shown only at the end of each presentation and thus only randomized within that section. In between each video clip, a black screen followed by a fixation cross lasting 2 seconds and 1 second respectively, were shown. Lastly, participants filled out the VAS questions as well as the MPQ, both considering pain intensity and unpleasantness of the videos.

EEG recording

EEG recording was done via an elastic cap with 64 silver chloride electrodes (Electrocap International, USA) according to the extended 10-20 system. Impedance was kept below 5kΩ in order to reduce polarization effects. The recordings were made using a TMSi-64 REFA amplifier and Eemagine

acquisition software (ANT software, the Netherlands). All signals were digitized (at a rate of 5000 Hz) and stored for off-line averaging.

Somatosensory evoked potentials (SSEPs)

SSEPs were obtained by non-painful electrical stimulation delivered to the median nerve at the wrist of the right hand via surface electrodes fixed with Velcro tape. Square wave pulses of 0.2 ms duration (constant current stimulator model DS7A, Digitimer Limited, UK) were administered at a constant frequency of 1.8 Hz. Per subject the amplitude was set to elicit a noticeable twitch, typically 5 - 12 mA.

SSEPs were extracted off-line by averaging over 20 ms before and 200 ms after median nerve stimulation. This was done for EEG recordings in CP3 since this position is thought to overly the (left) SI which is the origin of short latencies N20 and P45 (Allison et al., 1991). Although FC1 also provides information from SI, this location is thought to be more effective in highlighting N60 (Bufalari et al., 2007) and this position was therefore used to extract amplitude values for N60.

Data analysis

All processing was done by MATLAB (The Mathworks, USA). A second order Butterworth bandpass filter as well as a notch filter was applied (0.5-150 Hz and 50 Hz respectively). SSEP epochs at positions CP3 (for N20 and P45) and FC1 (for N60) were referenced against M1M2 and a mean amplitude value of 20 ms before stimulation was subtracted for each epoch. For each condition and per participant a SSEP value around 0 ms was chosen manually to be the individual SSEP baseline and this baseline correction was used to reduce any artifact interference for the following peak values. SSEP peaks at 20 ms (N20), 45 ms (P45), and 60 ms (N60) were determined manually, and the corresponding amplitude values were calculated by subtracting the SSEP specific baseline value. A multivariate analysis of variance (MANOVA) with condition as the fixed factor (hand, touch, needle1, needle2, incision, extreme, and anticipation) was performed for all of the previously mentioned latencies.

A forward linear regression analyses was carried out to test whether any or all of the trait empathy subscales (PT, FS, EC, and PD) were good predictors of SSEP amplitude changes while looking at painful situations. Also, a bivariate correlation analysis was used to examine any correlation between IRI scores and SSEP amplitude modulation. Next, a MANOVA with condition as fixed factor was carried out on the VAS scores to test for the effects of the videos on the subjective ratings. For several of the conditions independent t-tests were performed to check for differences between ratings of the videos within one condition. Also independent t-tests were used to compare

subjective VAS ratings between the anticipation video and other conditions. A bivariate correlation analysis between VAS scores and all 3 SSEP latencies was conducted in order to examine whether SSEP modulation due to observation of others’ pain was related to the subjective rating of the models’ experience as well as one’s own perception in both sensory and affective areas. For all MPQ scores the same analyses as for the VAS scores were performed, but only for conditions hand, needle2 and extreme, using the Number of Words Chosen (NWC) for each subsection (sensory, affective, and evaluative).

Where needed, post hoc tests were carried out using Bonferroni adjusted multiple pairwise mean comparisons.

Results

Healthy participants

SSEPs

Averaged SSEP amplitudes for all conditions and latencies are listed in table 1. Grand average SSEPs for all conditions measured in electrode CP3 are depicted in figure 1.

Latency Electrode Condition Grand average SSEP amplitude Standard deviation N Baseline 1.5220 .85125 Hand 1.3227 .88353 Touch 1.2983 .96162 N20 CP3 Needle1 1.3407 1.04631 16 Needle2 1.3837 .73309 Incision 1.3387 .82947 Extreme 2.7296 1.64636 Anticipation 1.6803 1.39295 Baseline 1.5956 1.41308 Hand 1.8026 1.61692 Touch 1.8693 1.56536 P45 CP3 Needle1 1.6173 1.37436 16 Needle2 1.5406 .84167 Incision 1.6869 1.04088 Extreme 2.4494 2.84350 Anticipation 2.5000 1.79604 Baseline 2.7960 1.40436 Hand 1.7361 1.45947 Touch 1.8225 1.45391 N60 FC1 Needle1 1.5526 1.19037 16 Needle2 1.6620 1.34620 Incision 1.8896 1.13335 Extreme 2.9560 2.49767 Anticipation 2.1621 2.09333

The MANOVA used to examine the effect of the different conditions on all three amplitudes showed a significant effect of the videos on these SSEP components, Wilks’ λ = .741, F (21,339.382) = 1.775, p = .020. Univariate testing showed that this effect was due to a significant influence of condition on N20 amplitudes, F (7, 120) = 3.179, p = .004 (see fig. 2). Importantly, it should be noted that the assumption of equality of variances was not met as the Levene’s test gave significant results for latencies N20, P45, and N60. However, when using log values for the amplitudes, the Levene’s test for N20 proved to be non-significant, while still leaving a (marginal) effect for condition on N20 amplitudes intact, F (7, 120) = 2.089, p = .050. Our univariate conclusions are thus thought to be valid. Also, due to the small number of participants the box’s test gave significant results as well. The data should therefore still be interpreted with extra care.

Table 1; Grand averaged somatosensory evoked potential (SSEP) amplitudes and standard deviations (SD) for all conditions and all examined latencies, also noted are the used electrodes and the number of participants (N) included.

Figure 2; Grand average SSEPs for all conditions as measured by electrodes CP3 (top) and FC1 (down). Modulations of

SSEP amplitudes (N20 and P45 in the top figure and N60 in the lower figure) by the videos can be seen; this modulation is only significant for N20 (p<0.01).

IRI & SSEPs

Mean scores on all subscales of the IRI are shown in table 2. A forward linear regression model showed that scores on the four subscales of the IRI (i.e. empathic concern, perspective taking, personal distress, and fantasy scale) provided no accurate predictive model of SSEP differences between conditions of observing a hand versus extreme pain for any of the above mentioned latencies. Also, no correlation was found between subscale scores and SSEP amplitudes. Table X displays means, standard deviations, and ranges for all subscales of the IRI.

VAS

Averaged VAS scores are displayed in figure 3. The multivariate result was significant for the effect of condition on VAS scores, Wilks’ λ = .281, F = 5.062,

df = (30,406), p = .000,

indicating a difference in the subjective ratings of the used videos. Some caution is in place since a

significant result of the Box’s test implies an inequality of covariances. The univariate F tests however, showed there was a significant effect of condition on all 5 VAS question, even when log values of these VAS scores were used to prevent any violation of the equal error variances as displayed by the Levene’s test. The effect of the condition was highly significant for VAS questions 1 to 5; F = 17.547, df = (6,105), p = .000; F = 7.225, df = (6,105), p = .000; F = 6.771, df = (6,105), p = . 000; F = 10.112, df = (6,105), p = .000 and F = 21.226, df = (6,105), p = .000 respectively.

Figure 3; Mean scores on the visual analogue scale (VAS) for healthy participants per condition. The VAS questions were: 1) How intense is the pain that the model

experiences? (sensory, model-oriented); 2) How intense is the pain that you experience? (sensory, self-oriented); 3) How much empathy do you feel for the model? (affective, model-oriented); 4) How unpleasant is this video to you? (affective, self-oriented) and 5) How unpleasant is this situation for the model? (affective, model-oriented, subjective).

Table 2; Mean scores, standard deviations (SD), and ranges for all four subscales of the Interpersonal Reactivity Index (IRI).

IRI subscale Mean (SD) Range

Empathic concern 19.9375 (3.25512) 15-25 Perspective taking 22.1875 (4.77799) 13-30 Personal distress 22.1250 (4.33397) 14-29 Fantasy scale 23.2500 (3.95811) 15-30

For each condition, participants were shown two comparable videos, and it was expected that those videos would be rated in similar ways. Despite our expectations several people (i.e. 7 out of 21) rated videos within one condition different on the VAS scales. In particular the videos from the hand and needle2 conditions were often rated differently. However, these differences yielded no significant difference between total VAS scores of any two videos within one condition. Although these videos were still taken together, it is interesting to note that for all latencies except P45, the corresponding SSEP amplitudes for hand were higher than for touch. Although not significant, this difference corresponds with subjective ratings of the hand videos being more unpleasant as well as more intense than the touch videos.

Furthermore, all three extreme videos were indeed rated similarly. Since no significant difference on any of the VAS scores was found between videos 6.1, 6.2 and 6.3, these scores were averaged in all analyses.

Lastly, anticipation was shown to yield no different subjective ratings on any VAS question compared to either the observation of a static hand or touch. The high model-oriented scores for the hand condition correspond with participants’ verbal descriptions that the hands used in these videos seemed to be in pain, even though they were not being harmed in the video itself.

VAS & SSEPs

Correlations (two-tailed) were calculated between VAS scores and SEPP amplitudes elicited during the different videos. The bivariate correlation analysis indicated that for the N20 latency there exist highly significant correlations between SSEP amplitudes and the sensory model-oriented ratings, the sensory self-oriented, the affective model-oriented and the affective self-oriented ratings, r (15) = . 261, p = .005; r (15) = .366, p = .000; r (15) = .283, p = .002 and r (15) = .323, p = .001 respectively. Also, a significant correlation between N20 amplitudes and the affective model-oriented subjective questions was found, r (15) = .215, p = .023. Furthermore, a significant correlation between N60 amplitudes and sensory self-oriented ratings was found, r (15) = .242, p = .010.

MPQ & SSEPs

Also, bivariate correlation analysis showed that some correlations between MPQ scores and SSEPs exist. N20 amplitudes are highly correlated with the Number of Words Chosen (NWC) in the evaluative section, r (15) = .429, p = .002. Also, N20 amplitudes are correlated with the NWC for the sensory and the affective subsections of the MPQ, r (15) = .308, p = .033 and r (15) = .352, p = .014 respectively. Table 3 displays the means, standard deviations, and ranges for all MPQ-NWC scores.

Table 3; Mean scores, standard deviations (SD), and ranges for scales of the McGill’s Pain Questionnaire (MPQ) per video.

MPQ measure (video) Mean (SD) Range NWC (hand) Sensory Affective Evaluative 2.0000 (2.70801) .8125 (1.37689) .9375 (1.28938) 0-9 0-5 0-3 NWC (needle2) Sensory Affective Evaluative 4.1875 (2.88035) 1.0625 (1.52616) 2.4375 (.81394) 0-11 0-5 0-3 NWC (extreme) Sensory Affective Evaluative 4.8750 (3.89658) 1.9375 (1.61116) 2.8125 (.54391) 0-11 0-5 0-3

Patients and matched controls

SSEPs

Although no statistical analyses were performed due to the small number of patients tested, this data is still used to give an indication of possible difference in pain observation processing compared to healthy aged matched controls. Averaged SSEP amplitudes for healthy participants, patients, and aged matched controls, separated for each latency, are depicted in figure 4.

Notably, next to overall higher amplitudes, patients show less variation of P45 amplitudes in the different conditions.

VAS

This divergent pattern of brain activity for chronic pain patients was similar to the observed patterns in the subjective ratings of the videos (see fig. 5). Next to higher overall ratings on all VAS scores, patients seem to dissociate less between the conditions.

Figure 4; Grand average SSEP amplitudes for all latencies examined (N20, P45, and N60) for patients and aged matched controls (left, blue and red resp.) and healthy young participants (right) per condition. Patients have noticeably higher P45

Figure 5; Mean scores on the visual analogue scale (VAS) for patients (left) and age matched controls (right) per condition. The VAS questions were: 1) How intense is the pain that the model experiences? (sensory, model-oriented); 2)

How intense is the pain that you experience? (sensory, self-oriented); 3) How much empathy do you feel for the model? (affective, model-oriented); 4) How unpleasant is this video to you? (affective, self-oriented) and 5) How unpleasant is this situation for the model? (affective, model-oriented, subjective). Patients show overall higher scores and seem to dissociate less between some of the videos (needle1, needle2, incision, and extreme).

MPQ

Likewise, the Pain Rating Index (PRI) of the MPQ demonstrated an overall higher subjective rating of the videos from the hand, needle2 and extreme conditions for patients as compared to matched controls. And again, patients seemed to be less influenced by the intensity of the depicted pain than healthy participants (see fig. 6).

Figure 6; Mean Number of Words Chosen (NWC) scores on the McGill’s Pain Questionnaire (MPQ) for patients (blue) and age matched controls (red) in sensory, affective, and evaluative domains, for three different videos (hand, needle2, and extreme).

Discussion

This study examined the effects of observing different pain intensities on short latency SSEPs brought forth by non-painful median nerve stimulation. It was shown that N20 amplitudes decrease while watching a hand that receives no stimulation, when it’s being touched, when it receives painful stimulation by either light or deep needle pricks, or when an incision is being made. However, when watching extreme pain being inflicted on a model’s hand or during pain anticipation induced by a needle hovering over a hand, N20 amplitudes increase. Also, trait empathy as measured by the IRI was shown to not only hold no predictive value over modulation of short latency SSEP components during empathy, but also showed not to be correlated with any of the examined SSEP amplitudes (i.e. N20, P45, and N60) during observation others’ pain. Furthermore, subjective ratings showed that videos depicting pain were rated more painful as well as more unpleasant for both the model as well as the observer. Interestingly, both intensity and unpleasantness ratings attributed to the model and experienced by the observer are strongly related to changes in N20 amplitudes elicited by observation of different pain intensities. Also, a relationship between N60 amplitudes and indications of the pain intensity experienced by the observer was found, whereby amplitudes are decreased during observation of no stimulation, touch, and (mild) pain, while increased amplitudes are seen during videos that are rated as most painful for the observer. Subjective pain ratings as measured by the MPQ (in sensory, affective, and evaluative domains) also showed a strong relationship with modulation of N20 during observation of different pain intensities.

The results of this study showed that observation of others’ bodily sensations can cause modulation of very early (i.e. at 20 ms) sensory processing of somatic stimulation on one’s own body. This finding is inconsistent with other studies indicating that the earliest modulatory effect of pain or touch observation is found in P45 (e.g. Bufalari et al., 2007) but it is in line with earlier data from our own lab (de Vos, de Jonge & van Putten, in preparation). In expansion, it was shown here that N20 amplitudes decrease during observation of touch or pain, but increase while witnessing extreme pain. Because of the observation that early SSEP components don’t seem to differentiate between watching touch or pain, it has been suggested that this modulatory effect is due to an attentional mechanism (Martínez-Jauand et al., 2012). However, this study showed that, at least for N20 amplitudes, this is not the case. An attentional mechanism would not provide a decrease as well as an increase of amplitudes for watching touch and mild pain, or extreme pain respectively. We feel therefore that it is more likely that even in such an early stage of processing, action preparation is specified by the intensity of the observed situation. Likewise, Martínez-Jauand et al. (2012) proposed that the use of electrical stimulation is more likely to become a painful situation than pneumatic

stimulation like they used, and it is thus thought to change the outcome of early sensory processing more extremely. Indeed, studies that used electrical stimulation showed that early sensory processing can differentiate between observation of painful and non-painful situations (Bufalari et al., 2007 and Cheng et al., 2008).

The modulatory effects of empathy on N20 amplitudes were shown to be related to ascribed as well as experienced pain intensity and unpleasantness. This finding conflicts with other studies (Avenanti et al., 2005 and Bufalari et al., 2007) claiming that early somatosensory processing during observation of others’ pain is not correlated with unpleasantness ratings. Although it seems indeed unlikely that such a sensory component would correlate with affective ratings of a stimulus, it should be noted that N20 modulation was not found in these studies and therefore cannot directly be compared to these findings. However, it remains odd that N20 amplitudes correlate so strongly with unpleasantness ratings of the videos. The relation between N60 amplitudes and pain intensity experienced by the observer comes closer to what we would expect based on the discussed literature since the neuronal origin of this potential is involved in processing stimulus intensity and not emotion. Also, contrary to some studies (Martínez-Jauand et al., 2012 and Lamm et al., 2007) we found no correlation between SSEP amplitudes and scores on any of the subscales of the IRI.

Although no statistical tests were performed on the data from the three patients included in this study, results indicate that these patients show overall higher P45 amplitudes. Interestingly, they also show less modulation of that SSEP component during the different conditions, leading to relatively high but stable amplitudes. Similarly, in their subjective ratings of the videos they rate videos more painful and more unpleasant but seem to differentiate less between the different pain videos compared to healthy participants. This pattern could be expected since patients with chronic pain are more sensitive to stimulation due to an increased excitability in sensory areas (Pleger et al., 2004) and they therefore respond more extreme to for example, touch (Peyron et al., 2004). It is thinkable that based on their own experience their simulation of an event also shows this sensitivity. However, the patients in this experiment were not experiencing any pain during their participation due to the spinal cord stimulator, and it is known that the higher excitability in SI disappears with treatment success (Pleger et al., 2005 and Maihofner et al., 2004). Also, an important confound could not be eliminated; these patients have been exposed to medical procedures for years and it is therefore thinkable that there has been some desensitization in visual observation of these procedures. However, although they verbally indicated to not be so affected by these types of images anymore, they still gave much higher pain and unpleasantness ratings to all videos. A more thorough examination of sensory based empathy in patients with chronic pain should be conducted.

A few limitations and shortcomings of this experiment should be highlighted. Although modulation of P45 amplitudes has been a rather robust finding in empathy literature, we failed to find any significant increase or decrease of amplitudes at this latency. A possible explanation for the absence of modulation in P45 amplitudes in this study cannot involve the perspective instructions that were given. While it is known that adopting an egocentric perspective while observing stimulation in others increases somatosensory activity (e.g. Jackson et al., 2006), studies that have used an allocentric perspective like we did (Bufalari et al., 2007) or even implemented a distraction task (Cheng et al., 2008) did find significant early somatosensory modulation due to pain or touch observation. However, in contemplating our findings such as the absence of P45 modulation or the correlation between N20 amplitudes and affective subjective ratings, there are some other methodological considerations to be made. Firstly, we cannot know whether the data was subject to motor mirror neuron interference since almost all videos contain not only the model’s hand, but also the surgeon’s hand performing the procedure. Based on this experiment it is not possible to conclude whether any of the found differences are due to activation of the MNS based on action observation (e.g. Gazzola & Keysers, 2009) rather than somatic resonance. Secondly, based on the subjective ratings of the videos using VAS questions, it was shown that the used videos were not optimal. The hand videos were not perceived as neutral which resulted in higher pain and unpleasantness ratings than for the touch videos. Also, the intended intensity difference between light and deep needle penetrations was not perceived as such, leading to equal ratings for these videos. Therefore, the real comparison that we were left with was the difference between static hand/touch, mild pain, and extreme pain. Lastly, participants were informed that some videos contained recordings of anesthetized hands since it has been found that especially sensory processing of others’ pain is automatic and is not disturbed by cognitive appraisal (Lamm et al., 2007). However, this knowledge led to much lower subjective ratings of both pain intensity and pain unpleasantness in some but not all participants. A dissociation was found between participants that did take anaesthesia into consideration and the ones who did not. Therefore, mean VAS scores used in the analysis do not reflect a very homogeneous group. It is advised other experiments look further into early somatosensory processing during empathy while paying attention to eliminating action interference, creating bigger subjective differences between the levels of pain intensity, and using videos that outside of a medical setting that require no anaesthesia.

In conclusion, there seems to be a modulating effect of stimulus intensity on early somatosensory processing during observation of others’ pain, like previously proposed by Avenanti et al. (2006). This study is in line with their theory stating that somatosensory modulation can only be found by using

graphic and direct observation of pain so that it can be localized and assessed based on intensity and thus accurately mapped onto one’s own body. This then explains why so many imaging studies did not find any proof for somatic resonance. Based on the results we found for N20 amplitudes, it is concluded that the sensory aspect of empathy is made possible by simulation of a model’s state, build upon activation of the MNS in the somatosensory cortex. Considering the tight relation and close proximity of somatosensory and motor areas in the brain, we speculate that not only do our own emotions serve to create action readiness, but observation of pain in others might also be used in discriminating pain intensities and facilitating appropriate action preparation through simulation in the MNS.

Literature

Allison, T., McCarthy, G. & Wood, C. C. (1992). The relationship between human long-latency somatosensory evoked potentials recorded from the cortical surface and from the scalp,

Electroencephalography Clinical Neurophysiology, 84, 301-314.

Allison, T., McCarthy, G., Wood, C. C. & Jones, S. J. (1991). Potentials evoked in human and monkey cerebral cortex by stimulation of the median nerve: a review of scalp and intracranial recordings, Brain, 114 (6), 2465-2503.

Avenanti, A., Bueti, D., Galati, G. & Aglioti, S. M. (2005). Transcranial magnetic stimulation highlights the sensorimotor side of empathy for pain, Nature Neuroscience, 8 (7), 955-960.

Avenanti, A., Paluello, I.M., Bufalari, I. & Aglioti, S.M. (2006). Stimulus-driven modulation of motor evoked potentials during observation of others’ pain. NeuroImage, 32, 316-324.

Barba C. Frot, M., Valeriani, M., Tonali, P. & Mauguiere, F. (2002). Distinct fronto-central N60 and supra-sylvian N70 middle-latency components of the median nerve SEPs as assessed by scalp topogrpgic analysis, dipolar source modeling and depth recordings, Clinical Neurophysiology, 113, 981-992.

Barsalou, L. W. (2008). Grounded cognition, Annual Review of Psychology, 59, 617-645.

Barsalou, L. W. (2009). Simulation, situated conceptualization, and prediction, Philosophical

Transactions of the Royal Society B, 364, 1281-1289.

Buccino, G., Binkofski, F. & Riggio, L. (2004). The mirror neuron system and action recognition, Brain

and Language, 89 (2), 370-376.

Bufalari, I., Aprile, T., Avenanti, A., Di Russo, F. & Aglioti, S.M. (2007). Empathy for pain and touch in the human somatosensory cortex. Cerebral Cortex, 17 (11):2553-2561.

Bushnell, M. C., Duncan, G. H., Hofbauer, R. K., Ha, B., Chen, J.-I. & Carrier, B. (1999). Pain

perception: Is there a role for primary somatosensory cortex?, Proceedings of the National

Academy of Sciences, 96, 7705-7709.

Carr, L., Iacoboni, M., Dubeau, M. C., Mazzioto, J. C. & Lenzi, G. L. (2003). Neural mechanisms of empathy in humans: a relay form neural systems for imitation to limbic areas, Proceedings of

the National Academy of Sciences, 100, 5497-5502.

Cheng, Y., Yang, C.Y., Lin, C.P., Lee, P.L. & Decety, J. (2008). The perception of pain in other

suppresses somatosensory oscillations: A magnetoencephalography study, NeuroImage, 40, 1833-1840.

Churchland, P. M. (1991). Folk psychology and the explanation of human behavior. In: J. D.

Greenwood (Ed.). The future of folk psychology. Cambridge: Cambridge University Press. 51– 69.

Damasio, A. R. (1999). The Feeling of What Happens: Body and Emotion in the Making of

Consciousness, New York: Harcourt Brace.

Davis, M. (1983). Measuring individual differences in empathy: Evidence for a multidimensional approach, Journal of Personality and Social Psychology, 44, 113-126.

Frijda, N. H., Kuipers, P. & ter Schure, E. (1989). Relations among emotion, appraisal, and emotional action readiness, Journal of Personality and Social Psychology, 57(2), 212-228.

Gazzola, V. & Keysers, C. (2009). The observation and execution of actions share motor and somatosensory voxels in all tested subjects: single-subject analyses of unsmoothed fMRI data, Cerebral Cortex, 19(6), 1239-1255.

Gallese, V. (2001). The ‘Shared Manifold’ Hypothesis: From Mirror Neurons to Empathy, Journal of

Consciousness Studies, 8, 33-50.

Gallese, V. & Goldman, A. (1998). Mirror neurons and the simulation theory of mind-reading, Cell

Godinho, F., Magnin, M., Frot, M., Perchet, C. & Garcia-Larrea, L. (2006). Emotional modulation of pain: Is it the sensation or what we recall? Journal of Neuroscience, 26 (44), 11454-11461.

Goldman, A. (2006). Simulating Minds: The Philosophy, Psychology, and Neuroscience of

Mindreading, Oxford: Oxford University Press.

Hofbauer, R. K., Rainville, P., Duncan, G. H. & Bushnell, M. C. (2001). Cortical representation of the sensory dimension of pain, Journal of Neurophysiology, 86, 402-411.

Hommel, B., Müsseler, J., Aschersleben, G. & Prinz, W. (2001). The theory of event coding (TEC): a framework for perception and action planning, Behavioral and Brain Sciences, 24, 849-937.

Jackson, P. L., Brunet, E., Meltzoff, A. N. & Decety, J. (2006). Empathy examined through the neural mechanisms involved in imaging how I feel versus how you feel pain, Neuropsychologia, 44(5), 752-761.

Jackson, P.L., Meltzoff, A.N. & Decety, J. (2005). How do we perceive the pain of others? A window into the neural processes involved in empathy, NeuroImage, 24, 771–779.

Kanda, M., Nagamine, T., Ikeda, A., Ohara, S., Kunieda, T., Fujiwara, N., Yazawa, S., Sawamoto, N., Matsumoto, R. Taki, W. & Shibasaki, H. (2000). Primary somatosensory cortex is actively involved in pain processing in human, Brain Research, 853, 282-289.

Keysers, C. & Gazzola, V. (2006) Towards a unifying neural theory of social cognition, Progress in

BrainResearch, 156, 379-401.

Kulkarni, B., Bentley, D. E., Elliott, R., Youell, P., Watson, A., Derbyshire, S. W. G., Frackowiak, R. S. J., Friston, K. J. & Jones, A. K. P. (2005). Attention to pain localization and unpleasantness discriminates the functions of the medial and lateral pain systems, European Journal of

Lamm, C., Nusbaum, H. C., Meltzoff, A. N. & Decety, J. (2007). What are you feeling? Using functional magnetic resonance imaging to assess the modulation of sensory and affective responses during empathy for pain, PLoSONE, 12, e1292.

Loggia, M. L., Mogil, J. S. & Bushnell, M. C. (2008). Empathy hurts: Compassion for another increases both sensory and affective components of pain perception, Pain, 136, 168-176.

Martínez-Jauand, M., Gonzáles-Roldán, A. M., Muñoz, M. A., Sitges, C., Cifre, I. & Montoya, P. (2012). Somatosensory activity modulation during observation of other’s pain and touch, Brain

Research, 1467, 48-55.

Melzack, R. (1975). The McGill Pain Questionnaire: major properties and scoring methods, Pain, 1, 277–299.

Melzack, R. (1999). From the gate to the neuromatrix, Pain. 6(Suppl), 121-126.

Morrison, I., Lloyd, D., di Pellegrino, G. & Roberts, N. (2004). Vicarious responses to pain anterior cingulated cortex: is empathy a multisensory issue?, Cognitive, Affective & Behavioral

Neuroscience, 4, 270-278.

Ogino, Y., Nemoto, H., Inui, K., Saito, S., Kakigi, R. & Goto, F. (2007). Inner experience of pain: Imagination of pain while viewing images showing painful events forms subjective pain representation in human brain, Cerebral Cortex, 17, 1139-1146.

Oxford Dictionaries (2012). oxforddictionaries.com. Oxford University Press. Retrieved July 25, 2012, from

http://oxforddictionaries.com/definition/english/empathy

Perry, A. & Bentin, S. (2009). Mirror activity in the human brain while observing hand movements: A comparison between EEG desynchronization in the μ-range and previous fMRI results, Brain

Research, 1282, 126-132.

Peyron, R., Laurent, B. & Garcia-Larrea, L. (2000). Functional imaging of brain responses to pain. A review and meta-analysis, Clinical Neurophysiology, 30, 263-288.

Pineda, J. A. (2008). Sensorimotor cortex as a critical component of an ‘extended’ mirror neuron system: Does it solve the development, correspondence, and control problems in mirroring?,

Behavioral and Brain Functions, 4, 47.

Ponnet, K., Buysse, A., Roeyers, H., & De Corte, K. (2005). Empathic accuracy in adults with a pervasive developmental disorder during an unstructured conversation with a typically developing stranger, Journal of Autism and Developmental Disorders, 35, 585–600.

Preston, S. D. & de Waal, F. B. M. (2002). Empathy: its ultimate and proximate bases, Behavioral and

Brain Sciences, 25, 1-72.

Prinz, W. (1997). Perception and action planning, European Journal of Cognitive Psychology, 9(2), 129-154.

Rainville, P. (2002). Brain mechanisms of pain affect and pain modulation, Current Opinion in

Neurobiology, 12, 195-204.

Rainville, P., Duncan, G. H., Price, D. D., Carrier, B. & Bushnell, M.C. (1997). Pain affect encoded in human anterior cingulate but not somatosensory cortex, Science, 277, 968–971.

Rizzolatti, G., Fogassi, L. & Gallese, V. (2001). Neurophysiologic mechanisms underlying the understanding and imitation of action, Nature Review Neuroscience. 2, 661-670.

Singer, T. & Frith, C. (2005). The painful side of empathy, Nature Neuroscience, 8(7), 845-846.

Singer, T., Seymour, B., O’Doherty, J., Kaube, H., Dolan, R.J. & Frith, C.D. (2004). Empathy for pain involves the affective but not sensory components of pain, Science, 303, 1157-1162.

Steriade, M. & Llinas, R.R. (1988). The functional states of the thalamus and the associated neuronal interplay, Physiological Review, 68, 649–742.

Timmermann, L., Ploner, M., Haucke, K., Schmitz, F., Baltissen, R. & Schnitzler, A. (2001). Differential coding of pain intensity in the human primary and secondary somatosensory cortex, Journal

Tölle, T. R., Kaufmann, T., Siessmeier, T., Lautenbacher, S., Berthele, A., Munz, F., Zieglgänsberger, W., Willoch, F., Schwaiger, M., Conrad, V. & Bartenstein, P. (1999). Region-specific encoding of sensory and affective components of pain in the human brain: a positron emission tomography correlation analysis, Annals of Neurology, 45, 40-47.

Verkes, R. J., Vanderiet, K., Vertommen, H., Kloot, W. A. & van der Meij, J. (1989). De MPQ-DLV: een standaard Nederlandstalige versie van de McGill Pain Questionnaire voor België en

Nederland. Leiden: Swets & Zeitlinger.

Wager et al. (2004). Placebo-induced changes in FMRI in the anticipation and experience of pain,