Introduction: Compassion and the Brain
There is a growing literature examining the relationship between compassion and various aspects of nervous system function, especially the brain. This is part of a broader trend in research seeking to establish the links, if any, between brain function and various emotional and psychological states including empathy (e.g., de Vignemont and Singer, Reference De Vignemont and Singer2006) and compassion (e.g., Weng et al., Reference Weng, Fox and Shackman2013). This chapter explores this expanding field and seeks to place emerging research findings about the neuroscience of compassion in their broader context.
The chapter starts by exploring neuroimaging studies of compassion and then examines the topic of empathy and the brain, noting evidence that observing another person’s emotional state activates parts of the neuronal network that are also involved in processing that same state in oneself. This is followed by further discussion of evidence about the neuroscience of compassion and, especially, the effects of compassion training. In essence, evidence suggests that multiple areas within the brain are involved in compassion and compassion training, with some regions more strongly implicated than others. Finally, the chapter presents relevant conclusions and outlines potential directions for future work. The overall message is that neuroscientific findings to date underscore the importance of developing compassion as a skill and fundamental attribute for healthcare workers across all settings.
Neuroimaging Studies of Compassion
Functional magnetic resonance imaging (fMRI) studies are used to systematically examine the functional neural networks underlying human behaviour. Today, fMRI is an increasingly sophisticated technique for imaging brain function, albeit that it is not possible to infer a direct causal connection between, for example, a compassion-based task and any simultaneous or subsequent changes in neural activity. Even so, studies that seek to link neuroscience with emotional states such as empathy and compassion yield interesting results and hold further promise for the future (e.g., Hou et al., Reference Hou, Allen and Wei2017).
Kim and colleagues performed a particularly valuable study of ‘neurophysiological and behavioural markers of compassion’:
This study conducted an integrative, multi-method approach which first investigated two fundamental self-regulatory styles (self-criticism and self-reassurance) with fMRI, and second, measured participant’s HRV [heart-rate variability], a marker of parasympathetic response, during compassion training, pre- and post- a two-week self-directed training period.
These researchers reported significant links between compassion and neurophysiological markers:
We identified neural networks associated with threat are reduced when practicing compassion, and heightened when being self-critical. In addition, cultivating compassion was associated with increased parasympathetic response as measured by an increase in HRV, versus the resting-state. Critically, cultivating compassion was able to shift a subset of clinically-at risk participants to one of increased parasympathetic response. Further, those who began the trial with lower resting HRV also engaged more in the intervention, possibly as they derived more benefits, both self-report and physiologically, from engagement in compassion.
These are fascinating, valuable findings, but, looking at the literature more broadly, methodological heterogeneity presents a real challenge across this field as a whole. Difficulties identifying which areas of the brain are involved in compassion include varying definitions of compassion across studies, different tasks used to generate compassion, and factoring in the impact of the noise generated by fMRI scanning. Also, because compassion includes feeling motivated to act, it is particularly difficult to generate a task that will measure this aspect of compassion on a fMRI scan. In addition, many people consider empathy and compassion to be synonyms of each other, so, unless researchers explicitly explain the difference, these can be confused (Lamm et al., Reference Lamm, Rütgen and Wagner2019).
Notwithstanding these issues, an integrative approach to this topic, combining structural imaging and lesion studies with fMRI studies, for example, can enhance knowledge about brain structure and behaviour in relation to compassion. In 2018, Weng and colleagues published a study of ‘visual attention to suffering after compassion training’, which they found to be ‘associated with decreased amygdala responses’:
Increases in visual preference for suffering due to compassion training were associated with decreases in the amygdala, a brain region involved in negative valence, arousal, and physiological responses typical of fear and anxiety states. This pattern was specifically in the compassion group, and was not found in the reappraisal group. In addition, compassion training-related increases in visual preference for suffering were also associated with decreases in regions sensitive to valence and empathic distress, spanning the anterior insula and orbitofrontal cortex.
This study concludes that, ‘collectively, these findings suggest that compassion meditation may cultivate visual preference for suffering while attenuating neural responses in regions typically associated with aversive processing of negative stimuli, which may cultivate a more equanimous and nonreactive form of attention to stimuli of suffering’ (Weng et al., Reference Weng, Lapate, Stodola, Rogers and Davidson2018; p. 1).
The role of the insula was also highlighted by Novak and colleagues who published ‘an integrative systematic review’ of ‘neural correlates of compassion’ in 2022 (Novak et al., Reference Novak, Malinakova, Mikoska, van Dijk JP and Tavel2022). This paper included thirty-five studies ‘examining the relationship between brain structure or function and compassion’:
Data from 2922 participants revealed 98 neural locations associated with compassion. We found that compassion tendency has been most frequently associated with neural activity in the left insula among fMRI studies. Results from the structural studies indicated frequent neuroanatomical associations between compassion tendency and grey matter volume in left and right insula and right caudate nucleus.
These researchers also commented on ‘reasons for heterogeneity of neural findings’ across the literature:
We found a large divergence of neuroanatomical findings across both structural and fMRI studies. There might be several reasons for such a divergence: 1) the effect of social desirability, 2) empathizing with another person, instead of generating compassion, 3) random noise due to a small sample size, 4) differences in compassion inducing stimuli and 5) non-balanced proportion of females and males.
The issue of empathy merits particular consideration. The difference between empathy and compassion was explored in Chapter 3 of this book, which examined ‘What Compassion Is Not’. Fundamentally, empathy means feeling with someone, while compassion means feeling for someone (Klimecki et al., Reference Klimecki, Leiberg, Lamm and Singer2013). In addition, compassion includes the motivation to act to alleviate suffering. So, given that empathy and compassion differ somewhat as concepts, do they also differ at the neuroscientific level?
Empathy and the Brain
Psychological models of empathy are based on the idea that observing and imagining another person in a particular emotional state activates a similar state in others. There is now fMRI evidence that observing another person’s emotional state activates parts of the neuronal network that are also involved in processing that same state in oneself. Singer and colleagues studied brain activity while research subjects received a painful stimulus and while they perceived that a loved one was receiving a pain stimulus (Singer et al., Reference Singer, Seymour, O’Doherty, Kaube, Dolan and Frith2004). In both circumstances, the bilateral anterior insula, rostral anterior cingulate cortex, brainstem, and cerebellum were activated.
Consistent with this, Fan and colleagues performed an fMRI-based meta-analysis of evidence regarding a possible ‘core neural network in empathy’ (Fan et al., Reference Fan, Duncan, de Greck and Northoff2011). This research group included forty studies in their analysis and concluded that the dorsal anterior cingulate cortex-anterior midcingulate cortex-supplementary motor area and bilateral insula can be regarded as constituting a core network in empathy.
So, empathy can be linked with certain patterns of activation in particular areas of the brain, but how do these findings relate to brain activity in compassion? Do the patterns differ? Do they overlap?
Klimecki and colleagues studied the differential pattern of functional brain plasticity after empathy training and compassion training (Klimecki et al., Reference Klimecki, Leiberg, Ricard and Singer2014). This group found that empathy training increases negative affect and increases brain activation in anterior insula and anterior midcingulate cortex – brain regions which are associated with empathy for pain. In contrast, training in compassion could reverse the increase in negative effect, boost self-reports of positive affect, and increase activation in a non-overlapping brain network involving the ventral striatum, pregenual anterior cingulate cortex, and medial orbitofrontal cortex. The authors concluded that training in compassion might reflect a new coping strategy to surmount empathic distress and strengthen resilience.
Overall, it is clear from this work, and from brain lesion studies (Hogeveen et al., Reference Hogeveen, Salvi and Grafman2016), that a network of brain regions is involved in various emotional abilities and states. The roles of different brain regions in myriad emotional states are likely to be complex and overlapping. For example, while the insula is linked with empathy (Fan et al., Reference Fan, Duncan, de Greck and Northoff2011; Klimecki et al., Reference Klimecki, Leiberg, Ricard and Singer2014), Novak and colleagues also outline ‘the role of insula during compassion’:
Taken together, it is possible that while in functional studies the left insula activity can reflect (1) the integration processes of sensory input and (2) the awareness of participants of their experienced compassion towards other, processing of positive and/or negative emotions (3), in lesion studies insular damage did not allow participants to effectively integrate incoming sensory inputs, so the representation of compassion feeling in their awareness is likely to be impaired.
While empathy and compassion are related concepts, then, they also differ in certain respects (Klimecki et al., Reference Klimecki, Leiberg, Lamm and Singer2013), as already explored in Chapter 3 of this book (‘What Compassion Is Not’). In parallel with this, it is also clear that training in empathy and compassion affect the brain somewhat differently (Klimecki et al., Reference Klimecki, Leiberg, Ricard and Singer2014). In other words, empathy and compassion are linked with each other to a certain degree, but also differ in particular ways, both experientially and at the level of neuroscience.
Growing Understanding of the Neuroscience of Compassion
The neuroscience of compassion continues to attract considerable research interest, especially over the past two decades. A significant proportion of this work looks at not only compassion and the brain, but also the effects of compassion training at a neuroscientific level. Overall, evidence to date shows that multiple areas within the brain are involved in compassion and compassion training, with some brain areas more strongly implicated than others.
In 2012, Desbordes and colleagues ‘investigated how 8 weeks of training in meditation affects amygdala responses to emotional stimuli in subjects when in a non-meditative state’ (Desbordes et al., Reference Desbordes, Negi, Pace, Wallace, Raison and Schwartz2012; p. 1). For this study:
Healthy adults with no prior meditation experience took part in 8 weeks of either Mindful Attention Training (MAT), Cognitively-Based Compassion Training (CBCT; a program based on Tibetan Buddhist compassion meditation practices), or an active control intervention. Before and after the intervention, participants underwent an fMRI experiment during which they were presented images with positive, negative, and neutral emotional valences from the IAPS [International Affective Picture System] database while remaining in an ordinary, non-meditative state. Using a region-of-interest analysis, we found a longitudinal decrease in right amygdala activation in the Mindful Attention group in response to positive images, and in response to images of all valences overall. In the CBCT group, we found a trend increase in right amygdala response to negative images, which was significantly correlated with a decrease in depression score.
This work was ‘consistent with the hypothesis that meditation training may induce learning that is not stimulus- or task-specific, but process-specific, and thereby may result in enduring changes in mental function’ (Desbordes et al., Reference Desbordes, Negi, Pace, Wallace, Raison and Schwartz2012; p. 1).
In 2013, Weng and colleagues showed that greater altruistic behaviour following compassion training was associated with altered activation in brain areas involved in social cognition and emotion regulation, such as the inferior parietal cortex and dorsolateral prefrontal cortex (Weng et al., Reference Weng, Fox and Shackman2013). Nine years later, Lockwood and colleagues published an especially interesting fMRI study of ‘neural representations for prosocial and self-benefiting effort’, which helps inform this literature further:
During fMRI, participants completed a decision-making task where they chose in each trial whether to ‘work’ and exert force (30%–70% of maximum grip strength) or ‘rest’ (no effort) for rewards (2–10 credits). Crucially, on separate trials, they made these decisions either to benefit another person or themselves … Strikingly, we identified a unique neural signature of effort in the anterior cingulate gyrus (ACCg) for prosocial acts, both when choosing to help others and when exerting force to benefit them. This pattern was absent for self-benefiting behaviors. Moreover, stronger, specific representations of prosocial effort in the ACCg were linked to higher levels of empathy and higher subsequent exerted force to benefit others. In contrast, the ventral tegmental area and ventral insula represented value preferentially when choosing for oneself and not for prosocial acts.
Overall, Novak and colleagues, in their ‘integrative systematic review’ of the ‘neural correlates of compassion’ found ‘that the most frequent neural associations with compassion across all analysed studies can be found in the orbital part of the left inferior frontal gyrus, in the right cerebellum, the bilateral middle temporal gyrus, in the bilateral insula and the right caudate nucleus’ (Novak et al., Reference Novak, Malinakova, Mikoska, van Dijk JP and Tavel2022; p. 46). This is consistent with other research evidence supporting the role of the cerebellum in social cognition (Hoche et al., Reference Hoche, Guell, Sherman, Vangel and Schmahmann2016) and evidence that damage to the inferior frontal gyrus is associated with alexithymia, which is difficulty describing and identifying one’s own emotions (Hobson et al., Reference Hobson, Hogeveen and Brewer2018).
Given the depth and range of these neuroscientific findings, and notwithstanding their methodological heterogeneity, what are the practical implications of this accumulating field of research? Does it provide guidance about what to do in practice? Or, at the very least, does it help move towards a neuro-biological explanation of the benefits of compassion?
What Are the Practical Implications of These Findings?
The ReSource Project is one of the largest scientific studies on the mental trainability of such qualities as compassion, mindfulness, perspective-taking, and pro-social behaviour.Footnote 1 Tania Singer initiated this multi-disciplinary, multi-method study in 2008, based on a grant from the European Research Council. Following a testing period from 2013 to 2016, the project continues to flourish (Singer et al., Reference Singer, Kok, Bornemann, Zurborg, Bolz and Bochow2016). The key goal of the large-scale, nine-month training programme is the scientific evaluation of the effects of different kinds of mental practices on mental and physical health, brain plasticity, and pro-social behaviours. The skills taught include mindfulness and attention, social intelligence, compassion, empathy, emotion regulation, body awareness, coping with stress, cooperation, and altruism.
Key publications from this work include research on attention, compassion, and theory of mind (Trautwein et al., Reference Trautwein, Kanske, Böckler and Singer2020), brain plasticity (Valk et al., Reference Valk, Bernhardt and Trautwein2017), experiences of contemplation and meditation (Kok and Singer, Reference Kok and Singer2017; Przyrembel and Singer, Reference Przyrembel and Singer2018), regulation of heart rate variability and altruistic behaviour (Bornemann et al., Reference Bornemann, Kok, Böckler and Singer2016), and the impact of mental training on cortisol stress reactivity (Engert et al., Reference Engert, Kok, Papassotiriou, Chrousos and Singer2017), body awareness (Bornemann and Singer, Reference Bornemann and Singer2017), and pro-social behaviour (Böckler et al., Reference Böckler, Tusche, Schmidt and Singer2018). This project provides enormously valuable insights into the nature and effects of mental training and helps to operationalise some of the emerging research findings about compassion, training, and the brain.
For example, Trautwein and colleagues examined ‘differential benefits of mental training types for attention, compassion, and theory of mind’:
Here we tested three consecutive three-month training modules aimed at cultivating either attention, socio-affective qualities (such as compassion), or socio-cognitive skills (such as theory of mind), in three training cohorts and a retest control cohort (N=332). While attentional performance improved most consistently after attention training, compassion increased most after socio-affective training and theory of mind partially improved after socio-cognitive training. These results show that specific mental training practices are needed to induce plasticity in different domains of mental functioning, providing a foundation for evidence-based development of more targeted interventions adapted to the needs of different education, labor, and health settings.
Consistent with these findings, Böckler and colleagues reported that ‘distinct mental trainings differentially affect altruistically motivated, norm motivated, and self-reported prosocial behaviour’:
we investigated the malleability of prosociality by three distinct mental trainings cultivating attention, socio-affective, or socio-cognitive skills. We assessed numerous established measures of prosociality that capture three core facets: Altruistically motivated behaviours, norm motivated behaviours, and self-reported prosociality … linear mixed effects models reveal differential effects of mental trainings on the subcomponents of prosociality: Only training care and compassion effectively boosted altruistically motivated behaviour. No effects were revealed for norm-based behaviour. Self-reported prosociality increased with all training modules; this increase was, however, unrelated to changes in task-based measures of altruistic behaviour.
Böckler and colleagues go on to note that ‘these findings corroborate our motivation-based framework of prosociality, challenge economic views of fixed preferences by showing that socio-affective training boosts altruism, and inform policy makers and society about how to increase global cooperation’ (Böckler et al., Reference Böckler, Tusche, Schmidt and Singer2018; p. 1).
There is a recognised neuroscientific dimension to this work. In 2017, Valk and colleagues found structural changes in the brain following different kinds of mental training, with training of present-moment focused attention mostly associated with increases in cortical thickness in prefrontal regions, socio-affective training inducing plasticity in fronto-insular regions, and socio-cognitive training including change in inferior frontal and lateral temporal cortices (Valk et al., Reference Valk, Bernhardt and Trautwein2017). The brain, then, shows considerable plasticity in response to different types of mind training, consistent with outcomes from more experientially oriented studies of these techniques.
Conclusions: The Neuroscience of Compassion
Overall, research to date confirms not only a neuroscientific basis to compassion, but also a neuroscientific basis to compassion training. This is consistent with the broader literature about various forms of contemplative practice which also affect the brain (e.g., Lutz et al., Reference Lutz, Slagter, Dunne and Davidson2008). This growing literature about the neuroscience of compassion is clearly exciting, but it still needs to be interpreted with caution, humility, and an awareness of its limitations. Future research could usefully seek to minimise methodological heterogeneity across this field and continue to focus on links between compassion training and benefits in practice, especially in healthcare, as well as the neuroscientific underpinnings of both compassion and compassion training. There is a particular need for further work on fears, blocks, and resistances to various aspects of compassion, along with any possible consequences, including potential negative effects (Kirby et al., Reference Kirby, Day and Sagar2019).
Despite these caveats, and research yet to be done, evidence about the neuroscience of compassion to date supports the idea that compassion can be deliberately cultivated through training, and that this affects the brain. There is evidence that activities such as compassion training and meditation can increase positive affect, boost resilience, facilitate altruistic behaviour, and possibly even assist with equanimity, and these effects are underpinned by growing neuroscientific evidence of impact on the physical brain (Klimecki et al., Reference Klimecki, Leiberg, Lamm and Singer2013; Klimecki et al., Reference Klimecki, Leiberg, Ricard and Singer2014; Weng et al., Reference Weng, Fox and Shackman2013; Weng et al., Reference Weng, Lapate, Stodola, Rogers and Davidson2018). These valuable findings underscore the importance of developing compassion as a skill and fundamental attribute for healthcare workers across all settings.
Open access