Empathy for Pain in the Chronic Pain Brain

(1)

UNIVERSITY OF AMSTERDAM

Empathy for Pain in the Chronic Pain

Brain

by

Lotte Warns

Literature Thesis

Research Master Brain and Cognitive Sciences Examiner 1: Dr. I. Timmers

Examiner 2: Dr. J.B. Engelmann

in the

Institute for Interdisciplinary Studies Research Master Brain and Cognitive Sciences

(2)

UNIVERSITY OF AMSTERDAM

Abstract

Institute for Interdisciplinary Studies Research Master Brain and Cognitive Sciences

Master of Science by Lotte Warns

Chronic pain – a highly prevalent health problem – does not only introduce elevated pain levels, but co-occurs with difficulties in emotion regulation, information processing bi-ases, and depression and anxiety symptoms. Interestingly, chronic pain is characterized by alterations in brain regions that are associated with the affective and motivational regulation of pain rather than the sensorimotor aspect of pain regulation. Moreover, chronic pain is paired with alterations in regions known as the core regions for empathy for pain. Therefore, in this review, the fields of chronic pain and empathy for pain are integrated to review if and how empathy for pain is modulated in patients with chronic primary pain. Additionally, potential underlying mechanisms that may exert a mediating influence on empathy for pain in patients with chronic pain are proposed. These include mechanisms related to emotion regulation and social support. The review offers multiple testable hypotheses for future empathy for pain research in chronic pain patients, which may advance our understanding of biopsychosocial modulation of pain in chronic pain and may inform transdiagnostic clinical approaches to ameliorate the quality of life of chronic pain patients.

(3)

Introduction

Empathy — the capacity to understand and share the affective states of others (Singer & Lamm,2009) — and specifically emotional empathy (i.e. empathic distress and concern) and empathy for another’s pain, play a key role in social interactions (Findlay et al.,

2006,Preston & De Waal,2002). Additionally, empathy for pain poses a critical survival mechanism in that it can signify immediate physical threat for both the experiencer and the observer of pain (Walters, 1994, Williams, 2002). Therefore, empathy for pain is crucial for social navigation.

Empathy for pain and the direct experience of pain have been linked to each other on a behavioral and neural level. For example, the study bySerbic et al.(2020) demon-strated that people in pain display higher levels of empathy for pain. In addition, an overlap emerges between the neural underpinnings of pain processing and empathy for pain – specifically its affective qualities – in the anterior insula (AI) and anterior cingu-late cortex (ACC; Lamm et al., 2011, Singer et al.,2004,Zaki et al., 2016), suggesting a shared representation of self and others when observing pain in others (Keysers & Gazzola, 2006). Interestingly, prior pain experiences seem to down-regulate activity in these brain regions when observing pain (Preis et al., 2013). Together, these studies strongly suggest that empathy for pain and one’s own pain experiences modulate each other.

However, what if these pain experiences are chronic? Chronic pain (i.e. pain that lasts for longer than 6 months) introduces alterations in brain activity, morphology, and connectivity (Apkarian et al.,2009,Davis & Moayedi,2013,L´opez-Sol`a et al.,2017,

May,2008,Thorp et al.,2018). Noteworthy, patients with chronic pain have gray matter abnormalities in regions identified as the core network of empathy for pain processing (i.e. insula and cingulate cortex; Davis & Moayedi, 2013, Lamm et al., 2011). In addition, patients display aberrant neural activity in the neurological pain signature on pain induction (L´opez-Sol`a et al., 2017, Wager et al., 2013) and show additional heightened activity in brain regions implicated in motivation and emotion regulation (Bushnell et al.,2013). Given these neural circuitry changes in brain regions implicated

(6)

Introduction 2 in pain and empathy for pain processing in patients with chronic pain, and the link between pain experience and empathy for pain on a behavioral level, this would suggest that brain alteration in pain processing regions due to chronic pain may affect empathy for pain processing in patients.

How exactly this empathic response to seeing someone in pain differs in patients with chronic pain, might be mediated by individual difference factors and social sup-port. Given that patients with chronic pain have alterations in brain regions related to the processing of the affective and motivation aspect of pain (Bushnell et al.,2013) and chronic pain co-occurs with difficulties in emotion regulation, information-processing bi-ases, and depression and anxiety symptoms (Koechlin et al., 2018, Noel et al., 2016), psychosocial factors (e.g. pain catastrophizing, hypervigilance, and alexithymia) may exert a mediating influence on empathy for pain processing in patients with chronic pain. Additionally, social support affects the pain experience of patients (C´anovas et al.,2018,

Penner et al., 2008), which in turn may modify the way patients deal with observing pain in others and thereby alter the empathic response to other’s pain. Therefore, psy-chosocial factors and social support could embody the underlying mechanism influencing empathy for pain processing in patients with chronic pain. That is, chronic pain may affect empathy for pain via altered emotion regulation and/or via the influence of social support on altered pain-related coping mechanisms.

However, despite the overlap between neural circuitry changes in patients with chronic pain and brain regions implicated in empathy for pain, the literature in both fields has been largely separate. Moreover, contradictory findings have emerged on whether patients with chronic pain do indeed differ in their empathy for pain ability and its directionality compared to healthy individuals (e.g. de Tommaso et al., 2019

and U¸ceyler et al.¨ , 2015). Therefore, in this literature review, I will draw both fields together by first discussing chronic pain and the neural circuitry changes in individuals with chronic pain, continue with the different facets of empathy for pain along with their neural correlates, and then demonstrate the neural overlap between brain alterations in patients with chronic pain and empathy for pain processing. I will then review studies that investigate empathy in patients with chronic pain to assess if and how empathy for pain is modulated in patients with chronic pain. Thereafter, I will discuss potential mechanisms that may exert a mediating influence on empathy for pain in patients with chronic pain and throughout provide testable hypotheses concerning these underlying mechanisms. These mechanisms may further inform the development of transdiagnostic clinical approaches to ameliorate our understanding and the quality of life of individuals with chronic pain.

(7)

Chronic pain

Chronic pain has been estimated to affect a large share of the population. For example, 19 % of European adults experience chronic pain (Breivik et al., 2006). Pain severity in patients has also been shown to increase with age and is more prominent in females than males (Nahin, 2015, Tsang et al., 2008). In addition, patients with chronic pain experience a variety of other symptoms besides pain, including fatigue and mental health disorder symptoms, such as anxiety and depression symptoms, all affecting quality of life (Noel et al.,2016). The symptomatology of chronic pain negatively contributes to school performances and results in increased unemployment rates, which have a long-term impact on socioeconomic status (Rosenbloom et al.,2017). According toVos et al.

(2016), the leading global cause of disability is lower back and neck pain. Chronic pain also affects patients’ social participation and their social support system, including their romantic relationships (Rosenbloom et al.,2017), and the well-being of parents of children with chronic pain (Hauser-Cram et al., 2001, Jordan et al.,2008). Moreover, chronic pain conditions have been estimated to cost ±$600 million to society (Institute of Medicine (US) Committee on Advancing Pain Research, Care, and Education,2011). These costs are not only driven by health care costs and treatment of additional symp-tomatology (Rayner et al.,2016), but also by work incapacity (Scholz-Odermatt et al.,

2019). This illustrates not only the immense impact chronic pain has on the patient itself, but also its tremendous impact on a patient’s environment and today’s society. I will now further define the type of chronic pain focused on in this literature review and continue with describing the psychological correlates of chronic pain and the brain signatures of chronic pain.

Defining chronic pain

Chronic pain can be subdivided into different classes, including chronic primary pain and chronic secondary pain. In this review, I will specifically focus on chronic pri-mary pain. Chronic pripri-mary pain conditions include chronic pripri-mary visceral pain (e.g. pelvic pain), chronic widespread pain (e.g. fibromyalgia), chronic primary musculoskele-tal pain (e.g. chronic lower back pain), chronic primary headache or orofacial pain (e.g.

(8)

Chronic pain 4 temporomandibular disorder) and complex regional pain syndrome. Chronic primary pain (henceforth chronic pain) is defined as pain that persists for longer than 6 months in one or more anatomical regions and is paired with emotional distress or functional disability (Treede et al., 2015). Pain does not equal nociception (i.e. neural encoding of noxious stimuli), but is a brain output that depends on biopsychosocial modulation such as threat appraisal and attention (Apkarian, 2019, Davis et al., 2017). Whereas acute and chronic pain are both subjective, the transition from acute to chronic pain is paired with a shift in the salience of pain, in that pain becomes maladaptive and an ’internalized disease state’ instead of a threat signal (Apkarian et al.,2009). This shows that chronic primary pain is characterized by a complex interaction between biological, psychological, and social factors.

Psychological correlates

The International Association on the Study of Pain (IASP) defines pain as ”an unpleasant sensory and emotional experience associated with, or resembling that associ-ated with, actual or potential tissue damage” (Merskey & Bogduk,1994). Therefrom it follows that patients with chronic pain do not only suffer from the sensorimotor aspect of pain, but, in addition, are also affected by its psychological and cognitive conse-quences. Studies investigating non-pain related effects of chronic pain have shown that these patients frequently exhibit symptoms associated with high emotional distress such as symptoms seen in depression, anxiety, and alexithymia (Bushnell et al.,2013,Di Tella et al., 2015, Malfliet et al., 2017). Despite chronic primary pain encompassing several different conditions, it is thought that the affective aspects are most likely transdiagnos-tic.

Psychological symptoms co-occurring in patients include symptoms seen in depres-sion, anxiety disorders, personality disorders, and substance use disorders (Aaron et al.,

2020,Malfliet et al.,2017). These symptoms have been shown to develop after the onset of chronic pain (Fishbain et al.,1997), but in turn also contribute towards the develop-ment and maintenance of chronic pain (Kroenke et al.,2011). For example, a study by

Lerman et al.(2015) found that heightened levels of anxiety and depression may worsen pain. It has also been proposed that whereas in the initial stage of chronic pain, pain effectuates emotional distress leading to symptoms seen in anxiety and depression, in later stages, emotional distress is more prominent and modulates the pain experience (Lerman et al., 2015). In addition, not everyone with chronic pain develops anxiety or depression, which suggests that certain individual differences serve as risk factors. For instance, pain catastrophizing has been shown to greatly impact pain (Thibault et al., 2008). Together, these findings indicate a delicate interplay between pain and psychological symptoms upholding each other.

(9)

Chronic pain 5 In addition to the presence of these psychological symptoms in patients with chronic pain, cognitive impairments are also present. For instance, the fear-avoidance model de-lineates how negative states and difficulties with down-regulating negative affect (i.e. emotion regulation) results in a vicious cycle aggravating pain symptoms (Simons & Kaczynski,2012,Vlaeyen & Linton,2000). Furthermore, patients with chronic pain dis-play attentional and memory deficits due to pain interfering with attentional processes leading to altered attentional resource allocation (Dick & Rashiq,2007). Patients with chronic pain tend to allocate more attention to pain than healthy individuals by ampli-fying attention to sensorimotor and affective stimuli (Fallon et al., 2015b). A study by

L´opez-Sol`a et al.(2019) suggested that social touch might have an analgesic impact by shifting attentional resources to the touch providing a distraction from the pain. More-over, anxiety and depression also induce selective attention. That is, depression increases attention allocation to negative and positive stimuli, whereas anxiety to threat-related emotional stimuli. The latter may be due to difficulties with threat-safety discrimina-tion (Lau & Waters,2017). Interestingly, aberrations in threat-safety discrimination are also shown in adolescent and adult patients (Heathcote et al.,2020,Moseley & Vlaeyen,

2015,Vlaeyen & Linton,2000,2012). Given the transdiagnostic nature of chronic pain, depression, and anxiety, these information-processing biases in individuals with chronic pain can further perpetuate their condition by maintaining pain chronicity.

Taken together, this shows that patients with chronic pain do not solely suffer the physical pain associated with their condition, but also deal with its impact on their psychological experiences and their cognitive functioning and in turn, the contribution of these impairments towards the development and maintenance of chronic pain.

Brain signatures of chronic pain

Chronic pain can be differentiated from acute pain in terms of its pain-related neural activity. Acute pain activates a distributed network of brain areas - sometimes also called the ’pain matrix’ (Marsh,2018), which is not only made up of the traditional path of pain processing, but also involves areas implicated in emotion, attention, and motor responses in addition to the pain sensation itself (Davis & Moayedi, 2013). A recent meta-analysis has found a bilateral core network for acute nociceptive pain comprising of the secondary somatosensory cortex, midcingulate cortex, insula, and thalamus (Xu et al.,2020). In contrast, patients with chronic pain have aberrant or prolonged neuronal activity in this distributed network, which might be due to deviant descending inhibitory control systems, which in turn could effectuate the distinctive spontaneous pain seen in patients with chronic pain (Davis & Moayedi, 2013). For instance, various studies have shown that stimuli (even non-painful stimuli) are experienced as more painful

(10)

Chronic pain 6 by patients with chronic pain compared to healthy individuals which was paired with enhanced pain-related neural activations in the neurological pain signature (Giesecke et al., 2004, Gracely et al., 2002, Lawal et al., 2006, López-Solà et al., 2017, Pukall et al.,2005). This neurological pain signature (NPS) consists of brain areas implicated in sensory integration and emotion processing. The study by López-Solà et al. (2017) complemented this NPS with a fibromyalgia pain pattern (FPP). The FPP captured a brain signature characterizing fibromyalgia (FM) and consisted of regions that are not part of the NPS, but are implicated in motivational and emotion regulation. Therefore, chronic pain can be defined by not just its sensory components, but most importantly by its affective and motivational neural components (Bushnell et al., 2013). These studies indicate that chronic pain has distinct processing differences compared to acute pain.

Gray matter morphological alterations in patients with chronic

pain

In addition to this distinct neural signature of chronic pain from acute pain, sev-eral studies have reported on the gray matter morphological alterations in patients with chronic pain conditions. For example, a review byMay (2008) on the morphological al-terations of the brain in patients with chronic pain found that there is a common brain signature to be found. Specifically, gray matter (GM) decreases in the orbitofrontal cortex (OFC), cingulate cortex, insula, and dorsal pons were shown in different chronic pain conditions. In another systematic review by Davis & Moayedi (2013), the au-thors demonstrated that the most common brain regions showing GM abnormalities in patients with chronic pain compared to healthy individuals are the prefrontal cortex (PFC), insula, ACC, and midcingulate cortex (MCC). These morphological alterations in regions associated with processing emotions and salience of stimuli among other func-tions highlight the fact that chronic pain is characterized by differences in affective and motivational processing of pain compared to healthy individuals. Furthermore, a review on the functional connectivity alterations in patients with chronic pain byThorp et al.

(2018) has shown aberrant connectivity in brain regions that also show gray matter alterations. For instance, there is decreased functional connectivity between the peri-aqueductal gray (PAG) and insula and amygdala in chronic back pain (CBP) patients and PAG and ACC in FM patients, which could contribute to the aberrant pain mod-ulation in individuals with chronic pain. In addition, both in CBP and FM patients increased connectivity between the nucleus accumbens (NAcc) and prefrontal areas was found, brain regions involved in motivation and pain inhibition, respectively. Together, these findings demonstrate that chronic pain is paired with brain abnormalities in brain regions important in processing the affective and motivational aspects of pain.

(11)

Chronic pain 7 However, there also seem to be condition-specific brain alterations. For instance,

Baliki et al. (2011) developed a ’bar code’ for different chronic pain conditions. They showed that there is varying overlap in GM decreases in patients with chronic pain compared to healthy individuals between chronic pain conditions. For instance, while CBP patients have decreased GM in the bilateral posterior insula, secondary somatosen-sory cortex, and the pre and postcentral regions, patients with complex regional pain syndrome (CRPS) have diminished GM volume in the anterior insula (AI) and OFC, resulting in a lean overlap of 8.7%. Furthermore, between studies examining the same chronic pain condition differences are also found. For example, the review by Cagnie et al.(2014) on brain alterations in FM patients, showed – apart from the common GM decreases in ACC, PFC, and insula brain areas – decreases in the precuneus and parahip-pocampal gyri and did not conclude anything on GM increases. In contrast, the review bySawaddiruk et al. (2017) reported gray matter decreases in the amygdala, PAG, and putamen, which were not reported by Cagnie et al. (2014). In addition, Sawaddiruk et al. (2017) also reported on GM increases in for instance the cingulate cortex, insula, and OFC. This illustrates that while the field is advancing and the number of stud-ies published is rapidly growing, a consensus is yet to be reached on overlapping and discernible neural signatures of chronic pain.

Altogether, these studies show that despite all chronic pain conditions having some commonalities in their neural circuitry changes, each condition also has its own distinct neural modifications. These commonalities – together with altered affective aspects – could be used as a transdiagnostic approach when diagnosing chronic pain. It is impor-tant to mention that the studies discussed here are by no means a comprehensive list as not all imaging modalities are exhaustively discussed, but are merely used to illus-trate a general picture of neural circuitry changes in patients with chronic pain. Studies investigating different conditions also differ in their experimental design and patient demographics, which can make it difficult to compare between studies. A comprehen-sive meta-analysis of varying imaging modalities is needed to shed light on concrete distinguishable and overlapping brain alterations between chronic pain conditions.

(12)

Empathy

Before examining empathy for pain processing in patients with chronic pain, empathy terminology and its neural circuitry have to be defined. Therefore, I will now start by discussing the different facets of empathy and then continue by focusing on empathy for pain. Finally, I will delineate the neural correlates of empathy for pain and the emotional facet of empathy.

Empathy and its components

Empathy is the capacity to understand and vicariously experience the affective states of others, and encompasses both cognitive and emotional/affective components (Shamay-Tsoory et al., 2009,Singer & Lamm, 2009). Cognitive empathy is defined as the ability to understand another’s state, a term that is related to theory of mind and perspective taking (Decety & Jackson, 2004). On the other hand, emotional empathy can be understood as the capacity to vicariously experience another’s emotional state (Reniers et al., 2011). Emotional empathy can be further divided into empathic con-cern and empathic distress (Gleichgerrcht & Decety, 2013). Whereas empathic distress is associated with self-oriented feelings of discomfort in response to a distressed tar-get, empathic concern is related to other-oriented feelings of compassion and sympathy towards the distressed target (Davis et al.,1999).

Empathy can be evoked by a broad spectrum of emotions. One particular emotion that has been extensively studied in the context of empathy is pain (Craig, 2003, Fon-seca et al.,2017). Observing someone else in pain is assumed to automatically evoke a vicarious pain experience and feelings of empathy for the target’s pain in the observer (Krishnan et al.,2016,Morrison et al.,2004). Empathy for pain poses a critical survival mechanism in that it can signify immediate physical threat for both the experiencer and the observer of pain (Walters, 1994, Williams, 2002), but can also motivate prosocial behavior (Preston & De Waal, 2002). This latter, however, may depend on the evoca-tion of empathic distress or concern. That is, empathic distress is often the negative consequence of sharing other’s affective states and is considered an automatic, embodied

(13)

Empathy 9 feeling. As a result, distressed individuals engage in avoidance behavior from the target to reduce their own discomfort. On the other hand, empathic concern emerges when people down-regulate their own distress response to care about another’s need, resulting in approach behavior (Ashar et al., 2017, Grynberg & Maurage, 2014, Preston, 2013,

Yamada & Decety,2009). Empathy as a motivator for prosocial behavior is not exclu-sive to humans, but is also observed in animals. For example, rodents help distressed cagemates, and bonobos and monkeys aid their peers in need (Bartal et al.,2011,Meyza et al.,2017,Preston & de Waal,2002) This substantiates the evolutionary importance of empathy for social attachment and altruistic behavior (De Waal, 2008, Preston & De Waal,2002).

Neural underpinnings of empathy for pain

The body of literature on empathy for pain is growing boundlessly and various meta-analyses have tried to define a core network for empathy for pain processing. One of the earlier meta-analyses by Lamm et al. (2011), established a core network for em-pathy for pain in the posterior ACC/anterior MCC and AI. These brain regions have been implicated as the processors of the affective component of empathy for pain (Cox et al.,2012,Singer et al.,2004). Timmers et al. (2018) extended this finding by adding the medial frontal gyrus, supplementary motor area (SMA), and primary somatosensory cortex to the core network. More recently,Fallon et al. (2020) and Kogler et al.(2020) also provided a core network for empathy for pain, which included the MCC and AI in both reviews. However, the systematic review byFallon et al.(2020) found, additional to this core network, empathy for pain to be established in the ACC, SMA, supramarginal gyrus (SMG), right inferior frontal gyrus (IFG), and occipitotemporal cortex. Kogler et al.(2020) on the other hand, added the left SMG and the left dorsomedial prefrontal cortex (dmPFC) to the core network for empathy for pain. Kogler et al. (2020) ad-ditionally singled out cognitive empathy for pain and showed convergence in the left dmPFC with extensions into the MCC, SMG, and AI. Whereas not enough studies on affective empathy for pain have been published to meta-analyze, Kogler et al. (2020) did directly compare cognitive and affective empathy, showing activation in the bilateral dmPFG and left SMG for cognitive empathy and activation in the posterior part of the left dmPFG for affective empathy. In addition, they showed that studies on affective empathy consistently activated the IFG. Furthermore, the meta-analysis by Timmers et al. (2018) demonstrated that perceptual/affective empathy for pain paradigms ac-tivated more the right AI, whereas cognitive/evaluative paradigms recruited more the left MCC. In a previous meta-analysis, Fan et al. (2011) directly compared affective with cognitive empathy and found that affective empathy activated the right AI/IFG, whereas cognitive empathy elicited MCC activation. Conjointly, these meta-analyses

(14)

Empathy 10 suggest that the A/MCC and bilateral AI form the core network for empathy for pain processing and that they might be distinctively involved in cognitive and affective do-mains of empathy respectively.

In addition, empathic care and empathic distress have been shown to have disso-ciable brain systems. Other-oriented empathic care was preferentially associated with activity in the ventral striatum, ventromedial PFC, and medial OFC (Ashar et al.,2017), which are regions implicated in vicarious reward, a primary motivator for prosocial be-havior (Mobbs et al.,2009,Preston,2013). Self-oriented emphatic distress, on the other hand, was preferentially associated with premotor and somatosensory brain areas (Ashar et al.,2017), implicated in the mirror neuron system important for sharing or simulating experiences, which in turn evokes distress (Keysers et al., 2010). Moreover, the insula is also argued to be partitioned, in that empathic distress rather activates the poste-rior insula, whereas empathic concern activates the AI (Decety & Lamm,2006). These dissociable brain systems and their respective involvement in psychological constructs attest to their role in motivated behavior (Preston, 2013).

(15)

Empathy for pain in the context

of chronic pain

Empathy for pain and its neural correlates have been thoroughly investigated. However, little is know about how the neural alterations in patients with chronic pain discussed in the chapter on Chronic pain affect empathy for pain. One might suspect that, because these alterations are manifested in the core regions of empathy for pain and regions important for emotion processing, empathy for pain processing is altered in patients with chronic pain. To make this point, I will examine the overlap between chronic pain brain alterations and the neural processing of empathy for pain. Then, I will review studies investigating empathy in patients with chronic pain to determine if and how this overlap might affect empathy for pain in this patient group on a trait, behavioral and neural level. Finally, I will discuss related findings, considerations, and limitations.

Overlap between the neural correlates of empathy for pain

and brain alterations in patients with chronic pain

When integrating the core network of empathy for pain with the general neural circuitry changes in patients with chronic pain, an overlap seems to emerge (see figure1). As mentioned in the previous chapter, patients with chronic pain experience augmented pain, which is paired with enhanced activity in the NPS (L´opez-Sol`a et al.,2017). This NPS includes the ACC and AI (Wager et al., 2013), which were regions also found by the meta-analyses on empathy for pain (Fallon et al.,2020,Kogler et al., 2020,Lamm et al.,2011,Timmers et al.,2018). In addition, patients with chronic pain display gray matter abnormalities in the insula and A/MCC among others as shown by Davis & Moayedi (2013), and decreased functional connectivity with the insula as reviewed by

Thorp et al.(2018). Besides the overlap with the core network of empathy for pain, other brain regions with morphological or neural activity aberrations in patients with chronic pain overlap with important regions of empathy for pain processing, too. Timmers et al.

(2018) and Kogler et al. (2020) extended the core network of empathy for pain into 11

(16)

Empathy for pain in the context of chronic pain 12 the somatosensory and prefrontal cortex respectively, two brain areas that also show GM abnormalities in patients (Davis & Moayedi, 2013,L´opez-Sol`a et al., 2017,Thorp et al.,2018). These findings together display the great overlap between neural circuitry alterations in patients with chronic pain and the neural correlates of empathy for pain. This warrants the question of whether empathy for pain processing might be modulated by alterations in its neural signature in patients with chronic pain.

Empathy for pain in patients with chronic pain

Previous literature suggests that shared and prior pain experiences can shape em-pathy. A study by Preis et al. (2013) examining empathy for pain as modulated by prior pain experiences, found decreased activity in the aMCC and AI – previously iden-tified as the neural underpinnings of empathy for pain – when participants prior to the observation of someone in pain had experienced pain themselves. This decrease might be due to these brain areas signaling emotional significance of aversive stimuli instead of representing the shared experience (i.e. pain; Danziger et al. 2009). Additionally,

Figure 1: Overlap between brain alterations in chronic pain patients and the core empathy for pain brain regions. Gray indicates a brain region com-monly altered in chronic pain patients, whereas orange indicates the overlap between chronic pain brain alterations and the core network of empathy for pain processing. The orange-gray shaded brain regions are areas which are not part of the core net-work of empathy for pain, but do show up in some meta-analyses. Adapted from Bushnell et al.(2013). PFC = prefrontal cortex; ACC = anterior cingulate cortex; MCC = midcingulate cortex; S1/S2 = primary/secondary somatosensory cortex;

(17)

Empathy for pain in the context of chronic pain 13 they found increased activity in brain regions associated with emotion regulation, per-spective taking, and memory retrieval for participants with prior pain experiences. This may then suggest that prior experiences remove the signaling of the threat component when watching another person in pain, because their emotion regulation system appre-hends that observing pain in others does not pose a threat to themselves due to prior experiences, resulting in a more cognitive attitude towards other’s pain instead of shar-ing the experience. Chronic pain patients have an abundance of prior and shared pain experiences. Following the above logic, this suggests that this patient group may have decreased levels of empathy for pain. Indeed, shared negative experiences have also been shown to reduce emotion recognition accuracy when observing pain, partly explained by greater personal distress felt by the observer (Israelashvili et al., 2020b). However, a study by Serbic et al. (2020) demonstrated that people in pain (i.e. acute) display higher levels of empathy for pain. This discrepancy can be explained by the difference in dependent variable. Whereas Serbic et al. (2020) investigated empathy, Israelashvili et al. (2020b) investigated emotion recognition accuracy. It could be that while empa-thy increased due to shared experiences, which brings us closer to one another (Zaki,

2014), actual emotion recognition accuracy decreases because of internalization of the feeling (self-oriented) evoking personal distress, blinding us for how they actually feel (Davis, 1983). This is in line with the Affect-to-Cognition model by Israelashvili & Karniol (2018). This model poses that affective processes cause cognitive ones, which is facilitated by empathic concern, but inhibited by personal distress. More specifi-cally, Israelashvili & Karniol (2018) demonstrated that personal distress is negatively associated with perspective taking, which explains the decrease in emotion recognition accuracy as seen inIsraelashvili et al. (2020b) However, there was also evidence for in-creased emotion recognition accuracy in individuals with higher empathic concern levels (Israelashvili et al.,2020a). This would suggest that when people are able to inhibit the re-experience of the shared negative experience (i.e. avoid personal distress), empathic concern might actually help with emotion recognition.

In the previous section, I have shown the great overlap between the neural correlates of empathy for pain and the brain alterations in patients with chronic pain, including a GM decrease in the AI. Together with the findings described above on how prior pain experience might modulate empathy, I posit that there might be a relationship between the neural circuitry changes in patients with chronic pain in brain areas associated with empathy for pain and altered empathy for pain processing.

Various studies have investigated empathy for or the observation of pain by patients with chronic pain. de Tommaso et al.(2019),Di Tella et al.(2015),Fallon et al.(2015b),

Goldstein et al.(2019),Lee et al.(2013),U¸¨ceyler et al.(2015) andVandenbroucke et al.

(18)

Empathy for pain in the context of chronic pain 14 (2016) in patients with CRPS, Noll-Hussong et al. (2013) and Peng et al. (2019) in somatoform disorder (SD), andMa et al.(2020) in patients with chronic lower back pain (CLBP). These studies will now be reviewed in terms of their trait empathy, behavioral, and neural findings to determine if and how empathy for pain processing is altered in patients with chronic pain. In addition, related findings will be discussed.

Trait empathy findings

Contradictory results have emerged when investigating trait empathy in patients with chronic pain compared to healthy individuals (see table1). Some studies report no difference in empathy between chronic pain patients and healthy individuals (de Tom-maso et al.,2019,Di Tella et al.,2015,Vandenbroucke et al.,2014), whereas others report a decrease in empathy levels (Ma et al., 2020, Noll-Hussong et al., 2013, Sohn et al.,

2016) or an increase compared to healthy individuals (Peng et al., 2019,U¸¨ceyler et al.,

2015). However, all these studies use varying questionnaires to identify empathy levels in the population, such as the interpersonal reactivity index (IRI;Davis 1983), empathy for pain scale (EPS; Giummarra et al. 2015), and empathy quotient (EQ; Baron-Cohen & Wheelwright 2004) and study different chronic pain conditions. Most studies investi-gating empathy in chronic pain patients have been performed in patients with FM, three of which did not report differences in empathy levels between FM patients and controls (de Tommaso et al.,2019,Di Tella et al.,2015,Vandenbroucke et al.,2014). However,

¨

U¸ceyler et al. (2015) reported an increase in empathy levels in this patient group. More specifically, one can also distinguish between empathic concern (EC) and personal distress (PD), two facets of affective empathy (Gleichgerrcht & Decety,2013).

¨

U¸ceyler et al. (2015) found that both PD and EC increase in patients with FM. In contrast, the study by Sohn et al. (2016) reported decreased EC, but increased PD in patients with CRPS I compared to healthy individuals. A third study byNoll-Hussong et al.(2013) studying SD patients, showed lower EC levels in this patient group compared to the healthy control group, but did not find any differences in PD between groups. A study byPeng et al. (2019) studying the same chronic pain condition (i.e. somatoform disorder), did not find any difference in EC between patients and controls, but reported higher PD levels in patients than in controls. This hints towards a general trend across conditions, where chronic pain patients have decreased empathic concern levels, but increased personal distress levels compared to healthy individuals.

Studies also varied in the questionnaires used to assess empathy levels and the types of empathy. For instance, the study by de Tommaso et al. (2019) employed the EPS questionnaire – a measure of empathy for pain specifically – and summed all items to a total score, even though the EPS does offer different subscales representing EC and PD. If it is indeed the case that PD is higher and EC is lower in patients compared to controls,

(19)

Empathy for pain in the context of chronic pain 15 then this might explain why de Tommaso et al. (2019) did not find any differences in EPS scores between patients and controls as they cancel each other out when using a total score. Similarly, the results byDi Tella et al. (2015) can be explained. They used the empathy quotient (EQ) which measures affective empathy. Personal distress and empathic concern can be understood as two facets of affective empathy with opposing effects on prosocial behavior (Israelashvili et al.,2020a). Therefore,Di Tella et al.(2015) not finding any difference between FM patients and controls could be ascribed to these two facets of the EQ compensating each other. Although differences in the EC and PD subscales of the IRI have emerged, one scale of the IRI most studies seem to agree on is the fantasy scale (FS). This scale reflects the tendency of people to imaginatively experience the feelings and actions of fictitious characters (Davis, 1983). Peng et al.

(2019),Noll-Hussong et al.(2013), andVandenbroucke et al.(2014) all find lower fantasy scores in patients compared to controls. However, the study by Sohn et al. (2016) did not report a difference in the fantasy scale between groups. Instead of using empathy questionnaires,Di Tella et al. (2015) andShin et al.(2013) used the reading the mind’s eye test (RME). This is a measure to evaluate the ability to infer other people’s affective mental states. They found that fibromyalgia and CRPS patients, respectively, both display impairments in inferring other’s affective mental states.

Together, these studies show that within and across patient groups, a consensus has not yet been reached on the difference and directionality of empathy in patients with chronic pain in contrast to healthy individuals, nor on the different components of empathy. This demonstrates the need for further research. However, a careful trend seems to emerge in that empathic distress is heightened in patients compared to healthy individuals, whereas empathic concern is diminished.

Experimental behavioral findings

Besides investigating the trait empathy differences between patients and controls using questionnaires, several studies also examined how patients differ compared to con-trols when observing pain in others using experimental paradigms. Fallon et al.(2015a) demonstrated that patients with fibromyalgia attribute more pain and unpleasantness to pictures of others in pain than controls, whereas there was no difference between groups for non-pain pictures. In contrast, the study by Lee et al.(2013) did not show a difference in pain ratings for painful stimuli between FM patients and controls. Other studies byPeng et al.(2019) andMa et al.(2020) also found no difference in pain ratings for observing painful stimuli between SD patients and controls, and CLBP patients and controls, respectively. However, in contrast to the findings byFallon et al.(2015a),Peng et al.(2019) found lower unpleasantness ratings to observing painful stimuli in patients compared to controls. Moreover, the study byNoll-Hussong et al.(2013) even reported

(20)

Empathy for pain in the context of chronic pain 16 marginally lower pain ratings to observing painful stimuli by SD patients compared to controls. In addition, they showed a positive correlation between these pain ratings and empathic concern in the patient group.

These experimental results demonstrate that patients and controls have similar pain ratings of observing pain in others (Lee et al.,2013,Ma et al.,2020,Peng et al.,2019) and suggest an increase or decrease in unpleasantness in patients compared to controls (Fallon et al., 2015a and Peng et al., 2019, respectively). This could hint towards a modulated empathy for pain mechanism in patients, specifically in the affective domain. That is, patients with chronic pain might experience increased empathic distress for pain when observing others in pain resulting in an increase in unpleasantness compared to healthy individuals. However, as this was only demonstrated by one study (Fallon et al.,

2015a) andPeng et al. (2019) showed the opposite effect – albeit in a different chronic pain condition – further research is warranted.

Additionally, patients’ pain ratings may be biased by them attributing more pain to painful stimuli by internalizing the pain, resulting in no difference in pain ratings between patients and controls. Instead, patients may actually rate observed pain as lower than controls, but due to the internalization of pain, which increases their empathic distress, might score observed pain as higher, equating their pain ratings to controls. Indeed, an interesting study byVandenbroucke et al.(2014) examining vicarious somatosensory experiences found that FM patients had fewer neglect errors compared to controls when the stimulation site of pain in patients was not congruent with the observed stimulation site in others, which could possibly be attributed to patients being more hypervigilant for pain. Moreover, the study by U¸¨ceyler et al. (2015) found that FM patients had lower pain thresholds and assign higher pain intensities to their own pain compared to controls. However, further research is required to establish whether this bias could explain the lack of diverging pain ratings between patients and controls.

Neural findings

Some of these studies have also investigated the neural processing of empathy for pain in patients with chronic pain. Using fMRI, Lee et al. (2013) have shown that pa-tients with FM display lower signal changes in brain regions related to pain processing (e.g. the ACC and dorsolateral PFC among others) compared to healthy controls when observing pain in others. Another fMRI study on SD patients also demonstrated that compared to healthy controls, patients exhibited lower activation of the pACC when observing pain in others (Noll-Hussong et al.,2013). The authors suggest that this in-dicates both an inner and outer oriented emotional awareness alteration in patients in that pain is rather processed as a physical sensation, as this region provides ’emotional coloration’ (Lumley et al.,2011). In CLBP patients, Ma et al. (2020) showed that the

(21)

Empathy for pain in the context of chronic pain 17 decrease in cognitive empathy and emotional empathy in patients is related to impair-ments in the attention network and in the pain emotional memory, respectively. Using another imaging modality, namely EEG, de Tommaso et al. (2019) and Fallon et al.

(2015a) also demonstrated alterations in the attention processes in patients. de Tom-maso et al. (2019) showed that patients anticipated pain in someone else by activating their somatosensory circuits and reducing activity in the visuospatial circuits. On the other hand, healthy controls anticipated pain in others by involving visual sustained at-tention. This mechanism could possibly explain whyShin et al.(2013) andDi Tella et al.

(2015) found reduced abilities in chronic pain patients to infer the mental and emotional states of others. That is, their attention is oriented to their own pain (i.e. self-oriented) rather than other-oriented to the other’s mental states. In addition, the study byFallon et al.(2015a) showed augmented LPP amplitude, a marker for affective regulation, and P3 amplitude in patients with fibromyalgia when observing both pain and non-painful pictures. The authors suggested that this enhanced P3 activity could indicate a higher salience for somatic cues. Unfortunately, this study did not include a questionnaire to measure individual empathy levels. Therefore, they were unable to make any inferences about whether more empathic patients show augmented LPP responses compared to less empathic patients to index individual affective regulation levels. A similar explana-tion is suggested by Goldstein et al. (2019). They proposed that the lack of difference between neural responses measured with magnetoencephalography (MEG) for pain and non-pain stimuli in FM patients is not due to a lack of empathy for pain stimuli in this patient group, but rather because they perceive the non-pain stimuli as potentially painful. Indeed, Peng et al. (2019) found that the P3, LPP, and N2 components did not differ between pain and non-painful stimuli within patients, whereas there was a difference in P3, LPP, and N2 amplitude between these stimuli in controls. However, they also showed that the P3 and LPP amplitudes were generally lower in patients com-pared to in controls. This could indicate that, whereas amplitudes are lower in patients than in controls, it might indeed be the case that amplitudes are heightened for non-painful stimuli in patients evoking similar potentials for non-painful and non-non-painful stimuli. However, it could also be that painful stimuli cause lower amplitudes in patients. Both these two explanations would result in the same amplitude pattern between painful and non-painful stimuli in patients and warrants further investigation.

Everything considered, there seems to be a change in the neural processing of empathy for pain in patients with chronic pain. It could be that patients with chronic pain experience increased vicarious pain when observing someone in pain mediated by a neural circuitry alteration that emphasizes their own pain via enhanced somatosensory representation and reduced emotion regulatory mechanism, which in turn heightens their self-oriented empathic distress and reduces other-oriented empathic concern. Whereas

(22)

Empathy for pain in the context of chronic pain 18 this idea is supported by the findings byde Tommaso et al.(2019),Fallon et al.(2015a),

Goldstein et al.(2019), andPeng et al.(2019), studies do not find increased activation in somatosensory circuits in patients compared to controls when observing painful stimuli using MRI studies (Lee et al.,2013,Noll-Hussong et al.,2013). On the other hand, these MRI studies do show reduced activation of the ACC in patients compared to controls, which supports the idea that emotion regulation is altered when observing others in pain. It could be that patients do not activate areas implicated in emotion regulation (i.e. ACC) to process the emotional component of other’s pain or to regulate their own empathic distress, but rather only activate the somatosensory circuits, internalizing the feeling of pain resulting in heightened self-oriented empathic distress instead of other-oriented empathic concern.

Related findings

In contrast to patients with chronic pain, patients with congenital insensitivity to pain (CIP) have a dramatic reduction of pain perception. Therefore, they have limited prior pain experiences. Understanding how empathy for pain emerges in CIP patients can give a unique perspective on how shared and prior pain experience contribute to empathy for pain. Their reduction in pain perception would suggest that they do not experience additional personal distress when observing someone in pain, because pain is not a threat to them in contrast to patients with chronic pain. Indeed, two studies on trait empathy in patients with CIP did not report any difference between CIP patients and healthy controls concerning personal distress, nor are there differences in empathic concern (Danziger et al., 2009, 2006). However, they did report decreased pain levels when judging pain in others, suggesting that these patients do discount pain in others. In addition,Danziger et al.(2006) reported that balanced emotional empathy scale (BEES) scores positively correlated with pain judgments and pain discrimination ability from facial expressions, whereas these relationships were not present in controls. This indi-cates that CIP patients rely on their emotional empathy abilities to judge pain in others (i.e. identifying emotions). The implication was confirmed in a later study (Danziger et al., 2009), which showed a positive correlation between ventromedial PFC activity and empathy levels in CIP patients. This brain area is shown to be implicated in inte-grating information to become aware of other’s and own’s emotions (Olsson & Ochsner,

2008). CIP patients also demonstrated normal aMCC and AI activity to observed pain, despite their lack of prior experiences. This challenges the common conception of mir-ror matching which is based on activating one’s own pain circuit when observing pain in others. The authors suggest that activation in these areas despite the lack of prior experiences may depend on processing emotional significance rather than these shared representations of pain. This would then suggest that empathy for pain modulation in

(23)

Empathy for pain in the context of chronic pain 19 patients with chronic pain does not solely depend on the extensive prior and shared ex-periences of this patient group, but on the emotional significance of observing someone in pain as well. Observing pain in others has emotional evolutionary significance in that it signals threat. Therefore, patients with chronic pain may experience more distress when observing another in pain due to the observation signifying possible threat shaped as pain in addition to their own chronic pain (Peng et al.,2019,Sohn et al.,2016,U¸¨ceyler et al.,2015). Compared to healthy individuals, chronic pain patients may catastrophize this threat due to their shared experiences enhancing distress.

Chronic pain and empathy for pain: conclusion and

consid-erations

In summary, although a consensus has not yet been reached on empathy for pain in patients with chronic pain, the findings indicate a careful trend towards a decrease in empathic concern and an increase in empathic distress in patients with chronic pain compared to healthy individuals. Neural findings substantiate this inclination in that they show a shift from attention to somatosensory circuits and an alteration in emotion regulation processing brain areas in patients. However, as there were not enough studies per chronic pain condition and the underlying brain alterations in patients with different types of chronic pain varied, further research is needed to determine if empathy for pain processing is differentially affected in different types of chronic pain or whether all conditions follow this global trend.

However, it remains unclear how this global trend of increased empathic distress and decreased empathic concern affects patients’ behavior when observing individuals in pain. From behavioral measures, it appears that pain ratings of observing someone in pain are not altered in patients with chronic pain, but unpleasantness ratings might be. The pain versus gain task could be employed to examine this and could even take it one step further by assessing its impact on motivated altruistic behavior (FeldmanHall et al.,

2015,2013). In the pain versus gain task, people determine how much money they are willing to give up to reduce the shock intensity administered to another individual. This task offers a unique way to evaluate the impact of altered empathy for pain processing in patients with chronic pain on altruistic behavior and to assess the difference between a patient’s own unpleasantness and pain ratings, and a patient’s perceived pain and un-pleasantness in the pain receiver. Additionally, visual feedback of their choice (i.e. seeing the level of pain inflicted to the receiver) can be removed to determine whether this re-duces empathic distress, alters pain and unpleasantness ratings, and influences altruistic behavior in patients. Moreover, due to prior pain experiences, patients with chronic pain might experience additional empathic distress when observing others in pain. This

(24)

Empathy for pain in the context of chronic pain 20 could be established by studying patients on pain and pain-free days. However, this lacks the long-term effects of chronic pain. To also examine long-term effects of prior pain experiences due to chronic pain, studies could distinguish between patients with a short and long chronic pain history on pain and pain-free days, and control for pain severity. Other unexplored questions include whether the site of pain infliction determines the level of empathy for pain in patients and whether this level depends on the nature of vi-sual feedback of pain infliction. For example, observing pain infliction to the lower back or neck may evoke more empathic distress and/or concern in patients with CLBP and chronic neck pain, respectively, compared to pain infliction to sites that are incongruent with a patient’s own pain. Additionally, emotional engagement may affect empathy lev-els. That is, observing pain in facial expressions or only the site of infliction by varying the visual feedback may result in different levels of emotional engagement and, thereby, differentially affect empathy for pain. For example, Greene et al.(2001) suggested that moral judgments may depend on the degree of emotional engagement in that more “up-close-and-personal” dilemmas (e.g. footbridge dilemma) reduce the willingness to harm, whereas more impersonal dilemmas (e.g. trolley dilemma) are less emotionally engaging and therefore influence moral judgments differently. It would be interesting to deter-mine how emotional engagement manipulation of stimuli (e.g. visual feedback) interacts with empathy for pain components in patients with chronic pain and affects pain and unpleasantness ratings of observing the infliction of pain. Moreover, given the limited number of studies per chronic pain condition and the disagreement within conditions, more research is needed to validate whether the suggested trend is generalizable across conditions or whether each condition has its own relation with empathy for pain pro-cessing. The latter appears favorable given the differences in brain alterations between chronic pain conditions (May,2008). However, the overlap that emerges in the affective component of the pain matrix between conditions would endorse generalizability of the proposed trend given this component’s involvement in empathy for pain (Singer et al.,

2004). Further research should be undertaken to explore these questions.

However, there are also some common limitations in studies on patients with chronic pain. For instance, the patients investigated in these studies usually take some form of medication to alleviate pain. A study by Mischkowski et al. (2016) found that ac-etaminophen (i.e. paracetamol) reduces empathy for other’s pain. Moreover, R¨utgen et al.(2015) suggested that reduced empathic concern might be an unwanted side effect of analgesics by showing that placebo analgesics reduced pain and unpleasantness ratings when observing others in pain. This shows that medication is a mediating factor and should be included in analyses when empathy for pain is investigated in patients with chronic pain. For instance, the study bySohn et al.(2016) included participants taking acetaminophen, but did not consider medicinal effects. However, acetaminophen’s effect

(25)

Empathy for pain in the context of chronic pain 21 on empathy for pain was tested in healthy individuals (Mischkowski et al., 2016) and might have a different effect on patients with chronic pain. A study by DeWall et al.

(2010) showed that acetaminophen reduces neural activity in the AI and ACC (during social pain), which are brain areas affected by chronic pain, and therefore acetaminophen might work differently. Indeed, a review by Ennis et al. (2016) has shown that there is little evidence to support an efficacious treatment of chronic pain conditions with acetaminophen. However, this does not yet tell us whether it does or does not affect empathy in patients with chronic pain. Therefore, the effect of medication on cogni-tion in patients with chronic pain should be investigated further. In addicogni-tion, studies differed in their use of empathy measures to measure general empathy or empathy for pain or did not include one at all, such as Fallon et al.(2015a),Goldstein et al. (2019), and Shin et al.(2013). As pain differs from other affective states which evoke empathy and the neural correlates for empathy and empathy for pain diverge (Timmers et al.,

2018), future studies should use specific empathy for pain questionnaires such as the EPS (Giummarra et al.,2015) to more specifically assess empathic distress and concern for pain. Other studies did not include a subjective measure of own and other pain or discomfort when observing pain in others (de Tommaso et al., 2019, Goldstein et al.,

2019). These measures together could provide better insights into the relationship be-tween trait empathy and its effect on empathy for pain when observing others in pain. Therefore, the absence of these measures complicates deducing conclusions on empathy for pain alterations in patients with chronic pain.

Other factors that could have possibly affected the results of the studies mentioned in table1 are small sample sizes, socio-economic status, education level, and menstrual cycle – fibromyalgia is more prevalent in females than in males. A study by Gupta et al.(2017) suggested that mixed-sex studies could create biased data, while single-sex studies limit generalizability due to sex differences in brain alterations. Hence, both are necessary to paint a broader picture. These limitations might explain contradictory find-ings and may prevent drawing concrete conclusions about empathy for pain in patients with chronic pain.

(26)

Empathy for pain in the context of chronic pain 22 T able 1: Characteristics and main findings of e mpath y studies in patien ts with chronic pain. Study Chronic pain patien ts Metho ds Main findings Danziger et al. (2006) P atien ts with CIP: 7♀ ; 5♂ (29.8 y ears; range 16-50) T rait emp athy: BEES-T ot; Behavior al task: P ain, v alence and arousal rat-ings when observing pain-inducing videos, an d STEP test T rait emp athy: No difference in BEES-T ot scores b et w een pa-tien ts and HC. Behavior al: P ain judgmen ts and discrimination abilit y from facial expressions p ositiv ely correlated with BEES-T ot scores. CIP patien ts demonstrated lo w er v alence and arousal to pain videos than HC. Danziger et al. (2009) P atien ts with CIP: 7♀ ; 6♂ (32 ± 12 y ears) T rait emp athy: IRI-EC, IRI-FS, IRI-PD, IRI-PT, BEES-T ot; Behavior al task: P ain and arousal ratin gs when observing pictures of b o dy parts and vid e os of facial expressions; Imaging: fMRI T rait emp athy : No difference in IRI subscales and BEES -T ot scores b et w eem CIP p atie n ts and HC. Behavior al: Lo w er pain rat-ings in CIP patien ts than in HC. No difference b et w een groups in pain d is cri m in ation abilit y from facial expressions and arousal ratings. Neur al: BEES-T ot and IRI-EC p ositiv ely correlated with vmPF C and pgA CC activit y more in pati e n ts than in HC when observing pain in others. de T ommaso e t al. (2019) P atien ts with fibrom y algia: 21 ♀ (31.8 ± 7.9 y ears) T rait emp athy: EPS-T ot; Behavior al task: P ain rat-ings when observing o wn or other electrical stim ula-tion, op en or blind ; Imaging: EEG, LEPs T rait emp athy: No difference in EPS-T ot scores b et w een patien ts and HC. Behavior al: Higher pain rati ngs in patien ts than HC. Neur al: Desync hronization of b eta rh ythm o v er fron to-cen tral re-gions in more empathic patien ts. Con trols u sed visuospatial atten-tion, whereas patien ts activ ated somatosensory c ircuits en reduced visuospatial circuits when an ticipating others’ stim ulation. Di T ella et al. (2015) P atien ts with fibrom y algia: 40 ♀ (51.8 ± 7.8 y ears) T rait emp athy: EQ-T ot; Be-havior al task: RME test, Ekman 60 faces test T rait emp athy: No difference in EQ-T ot scores b et w een patien ts and HC. Behavior al: Lo w e r RME scores and lo w er abilit y to recognise em oti ons (esp ecially anger and disgust) in FM patien ts.

(27)

Empathy for pain in the context of chronic pain 23 T able 1 con tin ued from previous page Study Chronic pain patien ts Metho ds Main findings F allon et al. (2015) P atien ts with fibrom y algia: 19 ♀ (40.0 ± 8.0 y ears) Behavior al task : P ain and unpleasan tness ratin gs when observing painful pictures; Imaging: EEG Behavior al: P atien ts attribute more pain and unpleasan tness to painful pictures than con trols. Neur al: P1 amplitude w as reduced in patien ts compared to c on trols, whereas N2, P3 and LPP ampli-tudes w ere all increased. LPP amplitude w as higher for pain than for non-pain sti m uli in patien ts. Goldstein et al. (2019) P atien ts with fibrom y algia: 17 ♀ ; 2♂ (28.3 ± 10.6 y ears) Behavior al task: Observ ation of painful pictures; Imaging: MEG Neur al: Con trols demonstrated decreased alpha activit y for pain compared to non-pain stim uli, whereas patien ts did not. Lee et al. (2013) P atien ts with fibrom y algia: 23 ♀ (38.0 ± 7.3 y ears) Behavior al task: P ain rat-ings when observing painful pictures; Imaging: fMRI Behavior al: No difference in sub je ctiv e pain ratings to observing painful stim uli b et w een patien ts and con trols. Neur al: Observing painful compared to non-painful stim uli res u lted in greater activit y in the A CC, dlPF C, pre/p ost cen tral gyrus, IPL, SMA, insula, putamen, thalam us, hipp o campus, SN and cereb ellum in con trols compared to patien ts. In patien ts, pain ratings correlated with R dlPF C, L IPL, L SMA, L insula and R cereb ellum activit y .

(28)

Empathy for pain in the context of chronic pain 24 T able 1 con tin ued from previous page Study Chronic pain patien ts Metho ds Main findings Ma et al. (2020) P atien ts with CLBP: 15 ♀ ; 9♂ (39.2 ± 8.4 y e ars) T rait emp athy: BES-A-ECon t, B E S-A-CE , BES-A-ED, BES-A-T ot; Behav-ior al task: P ain ratings when observing painful pictures; Imaging: fMRI, DTI T rait emp athy: No difference in BES-A-ECon t, lo w er BES-A-CE, BES-A-ED and BES-A-T ot in patien ts. Behavior al: No difference in pain ratings of observing painful stim uli b et w een patien ts and con trols. Neur al: F C b et w een AI an d R PHP gyrus, b et w een AI and L dlPF C and b et w een L SPL and R precuneus c or relate d p os-itiv ely with discomfort ratings, cognitiv e empath y and emotional disconnection, resp ectiv ely . In patien ts increased F C is foun d b e-tw e en the AI an d L dlPF C, L FF G, L SPL, R precuneus, and bilateral cereb ellum, decreas ed F C b et w een AI and L caudate, increased F C b et w een L SPL and L FF G, L dlPF C and R pre-cuneus, increased F C b et w een L FF G and L dlPF C . Noll-Hussong et al. (2013) P atien ts with somatoform disorder: 17 ♀ ; 4♂ (46.6 ± 12.5 y ears) T rait emp athy: IRI-EC, IRI-FS, IRI-PD, IRI-PT; Behav-ior al task: P ain ratings when observing painful pictures; Imaging: fMRI T rait emp athy: Lo w er IRI-FS en IR I-EC scores in patien ts, did not hold when con trolled for BDI scores. Behavior al: P atien ts at-tribute marginally lo w er pain ratings to observing painful stim uli than con trols. Neur al: P atien ts had less ac tivit y in the L p osterior A CC when observing painful compared to non-painful stim uli. P eng e t al. (2019) P atien ts with somatoform disorder: 6♀ ; 12 ♂ (35.9 ± 1.8 y ears) T rait emp athy: IRI-EC, IRI-FS, IRI-PD, IRI-PT; Be-havior al task: P ain and un-pleasan tness ratings when observing painful pictures; Imaging: EEG T rait emp athy: Lo w er IRI-FS and higher IRI-PD scores in pa-tien ts. Behavior al: No difference in pain ratings for observing painful stim uli b et w een patien ts and con trols, but patien ts assign lo w e r unpleasan tness scores to observing painful stim uli than con-trols. Neur al: P3 and LP P amplitudes w ere lo w er in patien ts than con trols and P3, N2 and LPP amplitude differences b et w een pain and non-pain st im uli w ere greater in con trols than in patien ts .

(29)

Empathy for pain in the context of chronic pain 25 T able 1 con tin ued from previous page Study Chronic pain patien ts Metho ds Main findings Shin et al (2013) P atien ts with CRPS I or II: 17 ♀ ; 26 ♂ (38.8 ± 11.9 y ears) Behavior al task: RME test Behavior al: Lo w er RME scores CRPS patien ts. Sohn et al. (2016) P atien ts with CRPS I: 13 ♀ ; 19 ♂ (36.9 ± 7.8 y ears) T rait emp athy: IRI-EC, IRI-FS, IRI-PD, IRI-PT T rait emp athy: No difference in IRI-FS scores, but lo w er IRI-PT and IRI-EC scores and higher IRI-PD scores in patien ts. ¨ Uceyler et al. (2015) P atien ts with fibrom y algia: 23 ♀ ; 2♂ (59.1 ± 5.4 y ears) T rait emp athy: IRI-EC, IRI-FS, IRI-PD, IRI-PT; Behav-ior al task: Rate pain in PPT T rait emp athy: Higher IRI-PD and IRI-EC scores in patien ts. Be-havior al: Lo w er pain threshold in patien ts. V anden brouc k e et al. (2014) P atien ts with fibrom y al-gia: 37 ♀ ; 2 ♂ (39.7 ± 11.2 y ears) T rait emp athy: IRI-EC, IRI-FS, IRI-PD, IRI-PT; Be-havior al task: Detect tactile stim ulation wh ile observing videos with congruen t/incon-gruen t painful stim ulation T rait emp athy: No difference in IRI-EC, IRI-PD and IRI-PT scores, but lo w er IRI-FS scores in patien ts. Behavior al: P atien ts ha v e few er neglect errors than con trols. A bbr eviations: A CC = an terior cingulate cortex; AI = an terior insula; BDI = Bec k depress ion in v en tory; BEES = balanced emotional empath y scale; BES-A = basic empath y scale in adults; CE = cognitiv e empath y (BES-A); CIP = congenital insensitivit y to pain ; CRPS = complex regional pain syndrome; dlPF C = dorsolateral prefron tal cortex; DTI = diffusion te n sor imaging; EC = empathic conce rn (IRI); ECon t = emotional con tagion (BES-A); ED = e motional disconnection (BES -A); EEG = electro encephalograph y; EPS = empath y for pain scale; EQ = em p ath y quotien t; F C = fu nctional connectivit y; FF G = fusiform gyrus; FM = fibrom y algia; fMRI = functional magnetic resonance imaging; FS = fan tasy sc al e (IRI); HC = health y con trols; IPL = inferior parietal lobule; IRI = in terp ersonal reactivit y index; LEPs = laser ev ok ed p oten tials; LPP = late p ositiv e p oten tial; MEG = mag n e to encephalograph y; PD = p ersonal distress (IRI); pgA CC = p rege n ual an terior cingulate cortex; PHP = parahipp o cam p al; PPT = pain pressure test; PT = p ersp ec tiv e taking (IRI); RME = reading the mind’s ey e test; SMA = supplemen tary motor area; SPL = sup erior p arietal lobule; STEP = sensitivit y to expressions of pain; T ot = total score of all questionnaire items; vmPF C = v en tromedial prefron tal cortex;

(30)

Potential underlying mechanisms

affecting empathy for pain in

patients with chronic pain

Alterations in empathy for pain in patients with chronic primary pain could be mediated by other cognitive modifications correlated with chronic pain presence. For instance, the brain regions implicated to be altered in patients with chronic pain, are also associated with emotion regulation and pain catastrophizing. In addition, depression and anxiety symptoms are often found in this patient group, which are both also related to deficits in emotion regulation. Therefore, I will now discuss psychosocial factors as hypothesized underlying factors modulating empathy for pain in patients with chronic pain.

Psychosocial factors

Emotion regulation

Impairments in emotion regulation – the ability to modulate emotional expressions and states towards a desired state in the person experiencing them – form an important risk factor for the development as well as the maintenance of chronic pain (Aaron et al.,

2020). The review by Koechlin et al. (2018) has shown a direct association between maladaptive emotion regulation and chronic pain and the review byAaron et al.(2020) describes the difficulties patients with chronic pain experience with down-regulation of negative affect (e.g. pain). Studies have also shown aberrant neural activity in brain regions related to pain regulation in adults and children with chronic pain (Baliki et al.,

2006, Bhatt et al., 2019, Jensen et al., 2009). For instance, the study by Fallon et al.

(2015a) suggests that enhanced LPP amplitudes seen in patients compared to controls in the presence of pain cues can be attributed to patients up-regulating their coping strategies for pain to dull the pain. In addition, patients with chronic pain have been shown to have trouble describing their own emotional state (i.e. alexithymia; Di Tella & Castelli, 2016), which has been linked to deficits in emotion regulation (Di Tella

(31)

Potential underlying mechanisms 27

& Castelli, 2013) and enhanced pain experience (Aaron et al., 2019). In contrast, a decrease in alexithymia levels has been shown to improve the outcome of emotion regu-lation (Connelly & Denney,2007). Correctly identifying and describing emotional states is a prerequisite for empathy (Coll et al., 2017, Zaki, 2020). As a result, impairments in identifying and describing emotions will likely affect empathy. Indeed, research has demonstrated that impairments in emotion regulation and alexithymia are linked to empathy deficits (Guttman & Laporte, 2002, Moriguchi et al., 2007, Zaki, 2020). For instance, the study byMoriguchi et al. (2007) showed that individuals with heightened alexithymia scores, activated the pain matrix to a lesser extent when observing painful stimuli. In addition, they scored lower on the empathic concern and higher on personal distress subscales of the IRI. Furthermore, the study by Peng et al. (2019) found that alexithymia mediated the empathic N2 response – reflecting the reduced empathic re-sponse of patients towards others’ pain – and that impaired empathy abilities in patients could therefore be explained by their hardship with identifying and describing emotions. Additionally, anomalies in empathy in patients with chronic pain might be due to altered functional connectivity between the prefrontal lobe and other brain regions in patients, which disturbs top-down emotion regulation of empathy (Ma et al., 2020). Whereas proper regulation of one’s emotions can result in prosocial motives by evoking empathic concern, not regulation one’s emotion when sharing the pain experience – and there-with evoking empathic distress in the observer – motivates avoidance behavior from the target (Zaki, 2020). Therefore, the altered emotion regulation in patients with chronic pain might contribute towards modified empathy for pain processing and might exert a mediating influence on motivated empathic behavior (Zaki,2014).

Pain catastrophizing

Difficulties with regulating negative affect are associated with pain catastrophizing. Pain catastrophizing – the inability to shift away the focus from threatening and harmful situations, and denoting more harm to neutral stimuli – has been linked to chronic pain (Vlaeyen et al., 1995). More specifically, catastrophizing patients with chronic pain are more likely to interpret pain as a threat, report greater pain (i.e. nocebo effect), and show spontaneous pain and enhanced brain activity in response to painful stimuli (Baliki et al., 2006, Fallon et al., 2015b). This activity is linked to brain regions not only implicated in pain processing, but also in emotion and attention to pain (Malfliet et al.,2017). The study byFallon et al. (2015b) suggested that people that are prone to catastrophizing are more likely to evoke previous pain memories when observing others in pain or might require greater neural resources for emotion regulation. In turn, this would complicate the down-regulation of negative affect such as pain (Aaron et al.,2020). Consecutively, this then enhances feelings of distress interfering with empathic feelings to motivate prosocial behavior (Preston, 2013). Pain catastrophizing might therefore

Empathy for Pain in the Chronic Pain Brain