A meta-analytic review of the relationship between empathy and oxytocin: Implications for application in psychopathy research
Introduction
The endogenous hormone and neurotransmitter oxytocin (OXT), has received significant attention due to its relationship to psychological and social constructs including social bonding (e.g., Carter, 2014), sensitivity to the fear of others (e.g., Marsh, 2018), altruism (e.g., Liu et al., 2017; Zak et al., 2007), trust (e.g., Kosfeld et al., 2005), and aggression (e.g., Hovey et al., 2016). Furthermore, OXT has been found to be associated with empathy (e.g., Barraza & Zak, 2009; Bartz et al., 2010; Daughters et al., 2017; Shamay-Tsoory et al., 2013) which may be of particular relevance to psychopathy, a syndrome characterized by callousness and/or indifference to the suffering of others and antisocial features (e.g., Cleckley, 1941; Hare, 2016; Skeem & Cooke, 2010).
Individuals with psychopathy have prominent affective deficits including lack empathy, guilt or remorse, shallow understanding or recognition of emotions and are callous (e.g., Marsh, 2018). Interpersonally, they are grandiose, arrogant, deceitful and manipulative. Behaviorally, individuals with psychopathy, often from an early age, engage in instrumental, planned acts of antisocial behaviors and aggression, but can also display impulsive and irresponsible behaviors (Blair, 2013). Studies assessing the construct of psychopathy have found that these symptoms coalesce around mostly two (interpersonal/affective and impulsive/antisocial e.g., Hare, 2003) or three (interpersonal, affective, lifestyle/antisocial e.g., Cooke & Michie, 2001; Patrick et al., 2009) large factors that are highly predictive of negative outcomes, including criminal recidivism. Offenders with psychopathy are more likely to commit versatile and violent offenses compared to general offenders, up to three times more likely to reoffend generally, and up to four times more likely to reoffend violently than other offenders (e.g., Barbaree, 2005; Hawes et al., 2013; Krstic et al., 2017; Neumann & Hare, 2008).
Despite debate within the field about how precisely to conceptualize psychopathy and which of these facets may be most predictive of various important life outcomes (e.g., Hare & Neumann, 2008; Patrick et al., 2009; Skeem & Cooke, 2010), lack of empathy emerges as a central feature of the disorder (e.g., Jones et al., 2010; Kiehl & Hoffman, 2011; Verschuere et al., 2018). Lack of empathy may also be germane to the understanding and future treatment of psychopathy because it likely contributes to the intensity and persistence of its antisocial tendencies (e.g., Hare, 1999; Marsh, 2018; Mullins-Nelson et al., 2006).
Currently, there is a need for effective treatments for psychopathy, especially for adults and persons with more severe psychopathic features (Caldwell et al., 2007; Harris & Rice, 2006; Kiehl & Hoffman, 2011; Salekin et al., 2010). Recent advances in neuropsychology, behavioral genetics, and neuroendocrinology have focused attention on brain anatomy and pathways, neuromodulator systems, and genetic markers related to psychopathy which may provide clues for its etiology and treatment (e.g., Larsson et al., 2006; Tuvblad et al., 2014; Viding et al., 2007). One brain region heavily implicated in psychopathy is the amygdala (Daversa, 2010; Fallon, 2006; Mitchell & Beech, 2011) which is also an integral part of the oxytocin (OXT) system (e.g., Keltner et al., 2014; Marsh, 2018). Reduced activity and volume in the amygdala has been shown to be related to psychopathy in general and to specific features of the disorder such as lack of compassion, empathy, and increased aggression or antisocial behaviors (e.g., Blair et al., 2006; Daversa, 2008; Fallon, 2006). Therefore, given OXT's relationship to emotions and neurobiological substrates linked to prosociality and antisociality, it is possible that further study of OXT administration might illuminate treatment avenues for psychopathy.
To date there have been many studies, including a handful of meta-analyses, that have examined the relationship between OXT administration and social constructs such as empathy or aggression in non-clinical samples (e.g., e.g., Hovey et al., 2016; Leppanen et al., 2017), but to our knowledge, there has not yet been a systematic review of these studies with a focus on elucidating its implications for the treatment of core psychopathy traits, particularly lack of empathy. Furthermore, there have been only three known studies which have examined the relationship between circulating OXT levels and measures of emotional empathy in samples of individuals with psychopathic tendencies (Dadds et al., 2014; Levy et al., 2015; Mitchell et al., 2013). Thus, the current study includes a meta-analysis of studies examining the influence of exogenous OXT administration on empathy across a variety of measures in non-clinical samples; however, we review the results of this meta-analysis in conjunction with existing literature on psychopathy, to highlight questions for possible future study in this area.
In considering prior research of the relationship between circulating OXT and empathy, it is important to note that the literature suggests there are two related, but conceptually distinct, types of empathy – cognitive and emotional. Cognitive empathy is the ability to imagine or perceive the emotions of another person. Emotional empathy is the ability to feel and experience the emotions of another person. Cognitive empathy refers to a more rational understanding of someone else's emotional state while the latter is an emotional understanding (Duan & Hill, 1996). In support of this empathy dichotomy, Shamay-Tsoory et al. (2009), conducted a brain imaging experiment and found that the ventromedial prefrontal cortex and orbital frontal cortex are associated with cognitive empathy whereas the inferior frontal gyrus is associated with emotional empathy.
The bifurcation of empathy may be particularly relevant in understanding psychopathy. First, the brain regions identified by Shamay-Tsoory et al. (2009), as related to empathic ability, have also been identified as impaired brain regions in individuals with psychopathy. The orbital frontal cortex and the ventromedial prefrontal cortex, in addition to the amygdala, basal ganglia, and dorsolateral prefrontal cortex have all been found to be less active in the brains of individuals with psychopathy (Blair et al., 2006; Choi, 2018; Fallon, 2006 & 2014; Marsh, 2018; Perez, 2012). Furthermore, some research suggests that individuals with psychopathy may be characterized by deficits in emotional, rather than cognitive, empathy (Blair, 2005; Jones et al., 2010; Mullins-Nelson et al., 2006). However, some other authors have suggested that psychopathy may also be characterized by an inability to identify and understand others' fear (i.e., cognitive empathy; Marsh, 2018). Accordingly, studies examining the relationship between oxytocin and trait empathy typically distinguish between emotional and cognitive empathy measures. For example, some researchers have utilized the IRI (e.g., Daughters et al., 2017; Rodrigues et al., 2009), which contains separate subscales for emotional and cognitive empathy. Other researchers have used measurements that are explicitly geared towards measuring one of the two types of empathy (e.g., Reading in the Minds Eyes Task [RMET] for cognitive empathy and self-report questionnaires in response to an emotional stimulus for emotional empathy). Lastly, emotional and cognitive empathy may also be associated with genetic differences such that the former was associated with an OXT gene marker and the latter with a vasopression gene marker (Uzefovsky et al., 2015).
Circulating OXT is measured by analyzing the concentration of OXT protein levels in blood plasma, urine, or saliva or by administering intranasal OXT and observing the behavioral effects (Dadds et al., 2014; Feldman et al., 2012). An initial overview of studies examining the relationship between circulating OXT and empathy (see Appendix A for design characteristics of circulating OXT studies included in this review) reveals that there have been mixed results. Some researchers have found that OXT is positively related to empathy (Abu-Akel et al., 2015; Barraza & Zak, 2009; Daughters et al., 2017; Domes et al., 2007; Krueger et al., 2013; Procyshyn, 2017; Shamay-Tsoory et al., 2013), while others have found no significant effects (Mitchell et al., 2013; Tabak et al., 2015). Additionally, some researchers found different results depending on group characteristics. For example, some authors have found that OXT administration is related to empathy depending on age (Dadds et al., 2014), gender (Barraza & Zak, 2009; Hurlemann et al., 2010; Palgi et al., 2015; Theodoridou et al., 2013), or type of empathy assessed (i.e., cognitive vs. emotional, e.g., Hurlemann et al., 2010; Theodoridou et al., 2013). Furthermore, some of the studies have demonstrated that OXT may be related to empathy only among people with lower baseline social empathy (Bartz et al., 2010), lower ability to read social cues (Feeser et al., 2015), or higher conduct problems (Levy et al., 2015), These latter findings suggest that a relationship between circulating OXT and empathy may exist but only among persons with lower “social competency,”1 perhaps explaining some of the contradictory results of various studies and highlighting the need for a quantitative review that can account for such moderators.
Two previous meta-analyses have examined the effects of OXT administration on empathy or empathy-related constructs (Leppanen et al., 2017; Shahrestani et al., 2013). In the first of these meta-analyses, Shahrestani et al. (2013) examined the effects of OXT administration on emotion recognition (i.e., cognitive empathy) in a non-clinical sample. They found a significant positive relationship between empathy and administered OXT across primary studies (Hedges g = 0.29). In the second study, Leppanen et al. (2017) conducted a meta-analysis examining the effects of OXT administration on four constructs: 1) expression of emotions, 2) emotional theory of mind, 3) sensitivity to recognize basic emotions, and 4) recognition of basic emotions. Arguably, the empathy related constructs in Leppanen et al.' (2017) study can also be classified as cognitive empathy. Leppanen et al. (2017) examined 33 studies which included both non-clinical and clinical samples, such as individuals with PTSD, Borderline Personality Disorder, and Autism. They found that OXT administration significantly increased emotion recognition for some emotions among non-clinical samples but had no significant effects among the clinical populations studied. These meta-analytic reviews further suggest that circulating OXT and empathy are related, but differentially among varying populations.
As previously mentioned, although past research has suggested that psychopathy is more closely characterized by deficits in emotional empathy rather than cognitive empathy (Blair, 2005; Jones et al., 2010; Mullins-Nelson et al., 2006), only three known studies have examined the relationship between circulating OXT levels and measures of emotional empathy in samples of individuals with psychopathic tendencies. In two of these studies, baseline endogenous OXT levels were compared with measures of Callous-Unemotional (CU) traits, which include lack of empathy, particularly affective or emotional empathy (Lethbridge et al., 2017; Lui et al., 2016), in samples of children and juveniles with conduct problems (Dadds et al., 2014; Levy et al., 2015). Callous Unemotional traits among children and juveniles with severe conduct problems are thought to represent a subset of individuals that are more susceptible to develop psychopathy as adults (Burke et al., 2007; Hawes & Dadds, 2005; Frogner et al., 2018; Pardini et al., 2012; Frick et al., 2014). Dadds et al. (2014), found that circulating, baseline OXT levels were significantly, negatively related to CU traits in male juveniles with DSM-IV diagnoses of either conduct disorder (CD) or oppositional-defiant disorder (ODD) who were aged 10 to 16-years-old, but not those aged four to nine-years-old. These finding suggest that endogenous plasma concentration of OXT may be negatively related to CU traits, but only in post-pubertal individuals. Therefore, it is possible that this age effect is driven by pubertal developmental factors and indicates the potential for early interventions. Similarly, Levy et al. (2015), found that endogenous salivary OXT levels were negatively correlated with CU traits among male juveniles, aged 15 to 19 years old, with a greater history of early-onset conduct problems (not limited to diagnoses of CD and ODD), but not among participants with a lower prevalence of conduct problems. These findings suggest that endogenous OXT may be lower in individuals with more callousness and lack of emotional empathy, similar to the findings by Daughters et al. (2017) in relation to cognitive empathy.
Lower circulating endogenous OXT levels in a juvenile sample with traits that may make them more susceptible to developing psychopathy later on could suggest that similar findings may be observed in a sample of adults with psychopathy. Mitchell et al. (2013), however, found that adult, forensic psychiatric patients who met criteria for psychopathy had higher urinary levels of OXT compared both to patients who did not meet criteria and to controls (staff members of the hospital). The authors also found that baseline circulating OXT levels across the patient sample were positively related to only facet four of the PCL-R, particularly to the items “early behavior problems” and “juvenile delinquency”, but were unrelated to the other facets, including facet two, which includes “lack of empathy”. One explanation for the apparent discrepancy between these results and Dadds et al.' (2014) and Levy et al.'s (2015) findings, could be due to the potential analgesic and anxiolytic effect of OXT (Neumann & Slattery, 2016), which may confer a higher risk of engaging in antisocial behaviors.
The current study had two aims. The first aim was to clarify the strength and direction of the relationship between administered OXT and empathy in non-clinical samples. As previously mentioned, many previous studies have found evidence to suggest a positive relationship between circulating OXT and both emotional and cognitive empathy (e.g., Abu-Akel et al., 2015; Barraza & Zak, 2009), yet there are also many others that have either found no relationship, or a negative relationship (e.g. Mitchell et al., 2013; Theodoridou et al., 2013). In addition, some studies have suggested that OXT may have a selective effect on empathy, moderated by participant characteristics such as social competency (e.g. Bartz et al., 2010), age (Dadds et al., 2014), or gender (Hurlemann et al., 2010; Theodoridou et al., 2013). Given these contradictory findings, an updated meta-analysis allows for the examination of the relationship between circulating OXT and empathy using more power and exploring potential moderators such as age or gender. Furthermore, the current study expanded on prior meta-analyses (i.e., Leppanen et al., 2017; Shahrestani et al., 2013) which examined administered OXT and empathy-related constructs by 1) including a greater number and more recently published studies, 2) including studies that measured both cognitive as well as emotional empathy, and 3) reviewing the results with consideration of psychopathy research and highlighting a void in this area of study.
The second aim was to interpret our findings in the context of prior research related to psychopathy to provide potential directions for future research in psychopathy treatment. Findings that OXT administration has been positively related to empathy only among participants with diminished social competency such as those with greater severity of conduct problems (Levy et al., 2015), higher than average, yet non-clinical, symptoms of autism (Bartz et al., 2010), or male participants who were less empathic at baseline (Feeser et al., 2015; Hurlemann et al., 2010) suggest that a similar effect could be hypothesized to be observed in other similar groups. Moreover, because psychopathic traits are dimensionally distributed (Patrick, 2018), a greater understanding of the relationship between administered OXT and empathy, even in non-clinical samples, is particularly relevant to our understanding of psychopathy. Examining circulating OXT, and more specifically administered OXT, is most germane to our second aim because it may be hypothesized that biological treatments, such as administration or manipulation of OXT, could increase social functioning specifically among individuals with psychopathy, a disorder characterized by a lack of empathy (e.g., Verschuere et al., 2018) underlain by a large genetic and neurobiological component (Frazier et al., 2019; Larsson et al., 2006; Perez, 2012; Tiihonen et al., 2019). To the best of our knowledge there has been only one clinical study to date that examines the relationship between circulating OXT and psychopathic features in participants with psychopathy (Mitchell et al., 2013). This indicates the need and merit for further exploration into this relationship.
Because most studies have found a significant, positive relationship between both emotional and cognitive empathy and OXT, we hypothesized that the average effect size across primary studies would demonstrate that individuals who were administered OXT would have greater empathy scores on measurements of both cognitive and emotional empathy, on average, than individuals who were administered placebo. Moderator variables we considered were mean age, amount of OXT administered, amount of time elapsed between OXT administration and empathy measurement administration, empathy type measured, type of empathy measurement used, how much estimation was used to compute effect size, country, and gender.
Section snippets
Selection of primary studies
Studies were eligible if they: (i) examined the relationship between administered OXT and empathy; (ii) included at least one quantitative measure of either emotional or cognitive empathy; (iii) examined circulating, exogenous OXT by administering intra-nasal synthetic OXT (or placebo) to participants; (iv) collected OXT and empathy data at baseline and then post intervention and/or obtained OXT and empathy data from an experimental group and OXT and empathy data from a control comparison
Key features of primary studies
The final analysis included 17 primary studies conducted between 2008 and 2018 (see Appendix B for list of primary studies). All primary studies were randomized, placebo-controlled experiments in which scores on measures of emotional empathy and/or cognitive empathy were compared between non-clinical, healthy, adults administered OXT vs. controls (placebo). Measures of empathy were administered during two time points: once prior to OXT/placebo administration and once following. All studies were
Discussion
The purpose of this meta-analysis and review was to explore the relationship between administered OXT and performance on empathy tasks across studies using non-clinical samples in order to speculate whether further exploration of OXT's effects on individuals with psychopathy may be warranted, given that to date there is a dearth of studies examining the effects of administered OXT in psychopathic samples. The current meta-analysis also expands upon prior meta-analyses of similar topics (
Conclusion
The purpose of this meta-analysis was to systematically review the relationship between administered OXT and performance on empathy tasks in non-clinical samples and contextualize them to offer possible directions for future research examining the treatment utility of OXT administration for individuals with psychopathy. The results of this study suggest that administering intranasal OXT to a non-clinical sample may increase both cognitive and emotional empathy abilities slightly as assessed by
CRediT authorship contribution statement
Nicole Stark: Conceptualization, Methodology, Validation, Formal analysis, Investigation, Visualization, Data curation, Writing – original draft, Writing – review & editing, Supervision. Leonardo Bobadilla: Conceptualization, Visualization, Methodology, Project administration, Writing – original draft, Writing – review & editing. Paul Michael: Methodology, Writing – review & editing. Sarina Saturn: Writing – review & editing. Matt Portner: Investigation, Validation.
Declaration of competing interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.
Acknowledgements
None.
Funding
This research did not receive any specific grant from funding agencies in the public, commercial, or not-for-profit sectors.
References (116)
- et al.
Influence of oxytocin on emotion recognition from body language: A randomized placebo-controlled trial
Psychoneuroendocrinology
(2016) Responding to the emotions of others: Dissociating forms of empathy through the study of typical and psychiatric populations
Consciousness and Cognition
(2005)- et al.
Oxytocin improves "mind-reading" in humans
Biological Psychiatry
(2007) - et al.
Intranasal oxytocin: Myths and delusions
Biological Psychiatry
(2016) - et al.
Meta-analysis of the effects of intranasal oxytocin on interpretation and expression of emotions
Neuroscience and Biobehavioral Reviews
(2017) - et al.
Intranasal oxytocin enhances emotion recognition from dynamic facial expressions and leaves eye-gaze unaffected
Psychoneuroendocrinology
(2012) - et al.
A field study of the association between CD38 gene and altruistic behavior: Empathic response as a mediator
Psychoneuroendocrinology
(2017) - et al.
Personality-dependent effects of oxytocin: Greater social benefits for high alexithymia scorers
Biological Psychology
(2011) - et al.
Deficits in facial affect recognition among antisocial populations: A meta-analysis
Neuroscience and Biobehavioral Reviews
(2008) - et al.
Towards a neurobiological model of offending
Clinical Psychology Review
(2011)