Social media summary: Can psychopathic traits be adaptive? Evolutionary perspectives should consider developmental, social, and cultural factors.
1. Introduction
Psychopathy is an inherently fascinating and controversial topic. This fascination is particularly evident when looking at the myriad of movies and true crime documentaries which have a psychopathic protagonist (Hesse, Reference Hesse2009). In turn, popular culture portrayals have created a stereotyped and romanticised public perception of psychopaths as mad men, serial killers and unscrupulous geniuses (Skeem et al., Reference Skeem, Polaschek, Patrick and Lilienfeld2011). Beyond this public fascination, psychopathy has important legal and forensic implications as it is disproportionately represented in the criminal justice system. While studies that quantify the prevalence of psychopathy are limited, some studies estimate that individuals officially diagnosed with this condition comprise 20–30% of prison populations, 15% of forensic psychiatric patients and around 1–3% of the total population (Coid et al., Reference Coid, Yang, Ullrich, Roberts and Hare2009a; Hare et al., Reference Hare, Hart and Harpur1991; Johnson, Reference Johnson2019). Gender may also play a role in the expression of psychopathy as that prevalence is at least twice as high in men compared with in women (Coid et al., Reference Coid, Yang, Ullrich, Roberts and Hare2009a, Reference Coid, Yang, Ullrich, Roberts, Moran, Bebbington and Singleton2009b; Lilienfeld et al., Reference Lilienfeld, Latzman, Watts, Smith and Dutton2014; Johnson, Reference Johnson2019). Moreover, longitudinal studies also show that psychopathic traits are expressed early in life, and that they remain stable across the lifetime (Blonigen et al., Reference Blonigen, Hicks, Krueger, Patrick and Iacono2006; Fanti & Kimonis, Reference Fanti and Kimonis2017; Forsman et al., Reference Forsman, Lichtenstein, Andershed and Larsson2008; Hawes et al., Reference Hawes, Byrd, Waller, Lynam and Pardini2017, Reference Hawes, Byrd, Kelley, Gonzalez, Edens and Pardini2018; Hemphälä et al., Reference Hemphälä, Kosson, Westerman and Hodgins2015; Hyde et al., Reference Hyde, Shaw, Gardner, Cheong, Dishion and Wilson2013; Loney et al., Reference Loney, Taylor, Butler and Iacono2007; Lynam et al., Reference Lynam, Caspi, Moffitt, Loeber and Stouthamer-Loeber2007).
Among the prison population individuals with psychopathy account for a large proportion of severe and violent crimes (Brown & Forth, Reference Brown and Forth1997; Harris et al., Reference Harris, Skilling and Rice2001; Hart, Reference Hart1998). This is probably a consequence of the lack of empathy, guilt and remorse in psychopaths, which are necessary for the inhibition of antisocial and violent tendencies (Hare et al., Reference Hare, Neumann, Widiger and Widiger2012). Not only does psychopathy predispose to violence and crime, but it is also highly predictive of violent recidivism (Hare, Reference Hare, Gray, Laing and Noaks2002; Hemphill et al., Reference Hemphill, Templeman, Wong, Hare, Cooke, Forth and Hare1998; Salekin et al., Reference Salekin, Rogers and Sewell1996), and is particularly unresponsive to treatment (Harris & Rice, Reference Harris, Rice and Patrick2006; Salekin et al., Reference Salekin, Worley and Grimes2010). Despite the media attention and its legal and ethical implications, there is an overwhelming lack of understanding and consensus regarding the biological, developmental and evolutionary roots of psychopathy (Glenn et al., Reference Glenn, Kurzban and Raine2011; Gonzalez-Tapia et al., Reference Gonzalez-Tapia, Obsuth and Heeds2017), and hence treatment options remain limited. In this paper, we aim to fill this gap and critically review the theories on the adaptive function of psychopathy.
Psychopathy is a personality disorder broadly defined by shallow emotions, lack of empathy or remorse, deceptiveness, grandiosity, impulsivity, irresponsibility and chronic immoral and antisocial behaviour (Hare & Neumann, Reference Hare and Neumann2009). To measure this, the most widely used diagnostic instrument in forensic and clinical settings is the Psychopathy Checklist – Revised (PCL-R) (Hare, Reference Hare1991), which scores an individual's extent of psychopathy based on 20 traits grouped into two factors: an interpersonal-affective domain (factor 1) that encompasses core traits such as callousness and manipulativeness and an antisocial-lifestyle domain (factor 2) that entails disinhibition and chronic antisocial behaviour (Hare et al., Reference Hare, Hart and Harpur1991; Hare & Neumann, Reference Hare and Neumann2009; Table 1). Additionally, there are two items which contribute to the total score but do not load on any factors: promiscuous sexual behaviour (item 11) and many short-term marital relationships (item 17) (Hare et al., Reference Hare, Neumann, Widiger and Widiger2012). Although studies have shown that factors 1 and 2 are moderately correlated and share variance, suggesting some overlap in measurement, these factors are composed of unique traits with different predictive powers (Hunt et al., Reference Hunt, Bornovalova, Kimonis, Lilienfeld and Poythress2015). The PCL-R rating scale uses semi-structured interviews and case history information to rate each item on the scale and a score of 30 out of 40 is required for a diagnosis of psychopathy (sometimes 25/40 depending on the country; Hare, Reference Hare2020).
Table 1. The Psychopathy Checklist – Revised (PCL-R) items, the most widely accepted clinical measure of psychopathy (after Hare & Neumann, Reference Hare and Neumann2009)
While the PCL-R is an important tool for the diagnosis of psychopathy in clinical and forensic settings, there is a lot of debate surrounding the classification of psychopathy and the validity of a strict classification of psychopaths and non-psychopaths in the general population (Edens et al., Reference Edens, Marcus, Lilienfeld and Poythress2006; Hare, Reference Hare2020; Wright, Reference Wright2009). The current view is that psychopathy is dimensional, rather than categorical: psychopathic traits exist on a continuum of severity in the population (Edens et al., Reference Edens, Marcus, Lilienfeld and Poythress2006; Guay et al., Reference Guay, Ruscio, Knight and Hare2007; Miller et al., Reference Miller, Lyman, Widiger and Leukefeld2001; Walters et al., Reference Walters, Duncan and Mitchell-Perez2007a, Reference Walters, Gray, Jackson, Sewell, Rogers, Taylor and Snowden2007b), and this has been argued for all personality disorders (Tyrer, Reference Tyrer2012; Zimmerman et al., Reference Zimmerman, Chelminski, Young, Dairympie and Martinez2011). Thus ‘psychopaths’ are at the extreme end of psychopathic personality which can be found at subclinical levels in the general population, and do not represent a separate class of individuals per se (Edens et al., Reference Edens, Marcus, Lilienfeld and Poythress2006). While the PCL-R cut-off point for psychopathy is useful for forensic, clinical and legal purposes where such categorisation may serve a function for recommendations and treatment, we should consider all people to be more or less psychopathic. The dimensional nature of psychopathy has important implications when looking at its adaptive potential, as we will discuss later on, because different psychopathic domains (interpersonal, affective, antisocial, lifestyle) can have different fitness outcomes. Thus, in the general population psychopathy is measured with self-report scales modelled after the PCL-R such as the Self-Report Psychopathy scale (Williams et al., Reference Williams, Nathanson and Paulhus2003), the Psychopathic Personality Inventory (Lilienfeld & Andrews, Reference Lilienfeld and Andrews1996) and the Levenson Self-Report Psychopathy Scale (Levenson et al., Reference Levenson, Kiehl and Fitzpatrick1995). Another useful area of research for psychopathy comes from studies in psychology on dark personality traits, which include psychopathic components. One of the most commonly used measures is the Dark Triad, which is a construct of three non-pathological, socially aversive personalities: Machiavellism, subclinical narcissism and subclinical psychopathy (Paulhus & Williams Reference Paulhus and Williams2002).
At a neurological level, psychopaths display impaired emotional processing, characterised by a reduced response to and recognition of fear, pain, shock and happiness (Aniskiewicz, Reference Aniskiewicz1979; Blair et al., Reference Blair, Jones, Clark and Smith1997; Dawel et al., Reference Dawel, O'Kearney, McKone and Palermo2012; Rothemund et al., Reference Rothemund, Ziegler, Hermann, Gruesser, Foell, Patrick and Flor2012). In addition, psychopathy has been linked to attentional deficits which make psychopaths less able to focus on emotional stimuli (Baskin-Sommers et al., Reference Baskin-Sommers, Curtin and Newman2011; Hiatt et al., Reference Hiatt, Schmitt and Newman2004; Newman et al., Reference Newman, Curtin, Bertsch and Baskin-Sommers2010). These atypical emotional and attentional processes are often, but not always, accompanied by functional ‘deficits’ in areas of the brain that are either located in the limbic system or adjacent to it (amygdala, prefrontal cortex, hippocampus, temporal cortex and anterior and posterior cingulate; Glenn & Raine, Reference Glenn and Raine2008; Harenski et al., Reference Harenski, Hare, Kiehl, Malatesti and McMillan2010; Vien & Beech, Reference Vien and Beech2006). Nonetheless, there are debates regarding whether the differences between psychopathic and non-psychopathic brains represent ‘dysfunctions’, and whether they are a causal factor in the disorder (Hare et al., Reference Hare, Neumann, Widiger and Widiger2012; Nadelhoffer & Sinnott-Armstrong, Reference Nadelhoffer, Sinnott-Armstrong and Vincent2013).
While traditionally conceived as a mental disorder, this idea has been challenged. Several researchers argue that psychopathy does not seem to fulfil the criteria for a ‘dysfunction’ in a biological sense (Barr &d Quinsey, Reference Barr, Quinsey, Crawford and Salmon2004; Jurjako, Reference Jurjako2019; Leedom & Almas, Reference Leedom and Almas2012; Lilienfeld & Marino, Reference Lilienfeld and Marino1995; Kendell, Reference Kendell2002; Krupp et al., Reference Krupp, Sewall, Lalumière, Sheriff and Harris2013). The debate is generally centred around divergent interpretations of Wakefield's (Reference Wakefield1992: 373) sociobiological definition of a harmful dysfunction, which states that ‘harmful is a value term based on social norms, and dysfunction is a scientific term referring to the failure of a mental mechanism to perform a natural function for which it was designed by evolution’. Similar approaches have been applied to explain other mental disorders and maladaptive traits where psychopathology is seen either as an adaptive response to the environment or as a by-product of adaptations involved in evolutionary trade-offs, and not as a biological dysfunction (Hunt & Jaeggi, Reference Hunt and Jaeggi2022; Nesse, Reference Nesse2004; Nesse & Stein, Reference Nesse and Stein2019; Raihani & Bell, Reference Raihani and Bell2019). Arguments can be broadly divided between those who argue that neurological markers and antisocial tendencies are indicative of a dysfunction (Leedom & Almas, Reference Leedom and Almas2012; Nadelhoffer & Sinnott-Armstrong, Reference Nadelhoffer, Sinnott-Armstrong and Vincent2013) and those who state that neurological differences do not necessarily imply a dysfunction and that the evidence that psychopathy is disadvantageous at an individual level is weak (Barr & Quinsey, Reference Barr, Quinsey, Crawford and Salmon2004; Jurjako, Reference Jurjako2019; Krupp et al., Reference Krupp, Sewall, Lalumière, Sheriff and Harris2013; Wiebe, Reference Wiebe2004). The self-centred, coercive and manipulative nature of psychopathy has led several evolutionary scientists to question the maladaptive nature of psychopathy (Barr & Quinsey Reference Barr, Quinsey, Crawford and Salmon2004; Brazil et al., Reference Brazil, Book, Volk, Shackelford and Weekes-Shackelford2019; da Silva et al., Reference da Silva, Rijo and Salekin2015; Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011; Mealey, Reference Mealey1995a; Wiebe, Reference Wiebe2004). They have argued that psychopathy could have evolved owing to selective pressures and might instead be an adaptation.
In this paper we will review the evolutionary explanations for psychopathy and address the question whether psychopathy is adaptive. We will go over the explanations based on frequency-dependent selection and life history theory and highlight the empirical support for and limitations of those explanations (Section 2). We will then review the studies examining the correlation between psychopathic traits and fitness proxies that concern reproduction and survival (Section 3). Our main argument shall be that certain environmental factors, such as stress experienced early in life, the nature of social interactions (e.g. the behaviour of the receiver) and social structure and cultural norms can mediate the relationship between psychopathy and fitness. Therefore, assessing the adaptiveness of psychopathic traits requires an interdisciplinary approach combining evolutionary, developmental, anthropological and psychological studies. We finish by proposing future research avenues based on the current gaps in the literature (Section 4).
2. Evolutionary explanations of psychopathy
2.1. Frequency-dependent selection
The facts that psychopathy has been reported across several cultures (although we will discuss this further in Section 3.2), is moderately heritable and stable across the lifetime, and often serves the needs and interests of the individual have led many researchers to explore the possibility that psychopathic traits are the result of a functional evolutionary strategy (Wiebe, Reference Wiebe2004). In evolutionary research, psychopathy is most often discussed in relation to frequency-dependent selection (Glenn et al., Reference Glenn, Kurzban and Raine2011; Mealey, Reference Mealey1995a, Reference Mealey1995b; Book et al., Reference Book, Methot-Jones, Blais, Hosker-Field, Volk, Visser and D'Agata2019). Under the frequency-dependent selection model, the advantage of certain behaviours and traits present in a population depends on their frequency: there is a predominant cooperative strategy which is beneficial at high frequencies, accompanied by less common, non-cooperative strategies which are only beneficial when the individuals displaying it are low in number (Ayala & Campbell, Reference Ayala and Campbell1974). As a highly social species, our fitness depends on cooperation with other individuals and most humans are considered prosocial or cooperative strategists (Wiebe, Reference Wiebe2012). Therefore, psychopaths who display impaired social cognition and emotions are regarded as antisocial strategists who instead rely on exploiting others at their own advantage (Gervais et al., Reference Gervais, Kline, Ludmer, George and Manson2013). Their inability to relate to and recognise fear in others could make them impervious to threat which in turn can facilitate such exploitative strategies (Book et al., Reference Book, Stark, MacEachern, Forth, Visser, Wattam and Roters2020).
The assumption that psychopathy is a cheater strategy maintained under frequency-dependence generates several predictions on the behaviour and prevalence of psychopaths. Firstly, psychopathy is found at low levels in the population (approx. 1–3%, based on PCL-R diagnosis; Hare, Reference Hare1996; Johnson, Reference Johnson2019), and this seems to fit the predictions of the frequency-dependent selection. At a high frequency, cheaters become more easily detectable, and are more likely to encounter other cheaters, which undermines the effectiveness of this strategy (Brazil et al., Reference Brazil, Book, Volk, Shackelford and Weekes-Shackelford2019). Secondly, if psychopathy represents an antisocial strategy, then we would expect psychopaths to possess behavioural strategies aimed at exploiting others for their own benefit. Indeed, psychopathic traits are associated with both a desire to deceive others (Seto et al., Reference Seto, Khattar, Lalumière and Quinsey1997) and the ability to deceive others efficiently by appearing genuine during deceptive interactions (Book et al., Reference Book, Methot, Gauthier, Hosker-Field, Forth, Quinsey and Molnar2015; Porter et al., Reference Porter, ten Brinke, Baker and Wallace2011; Brazil et al., Reference Brazil, Dias and Forth2021). Psychopaths are also more likely to engage in uncooperative strategies in social dilemma studies in university (Curry et al., Reference Curry, Chesters and Viding2011; Gervais et al., Reference Gervais, Kline, Ludmer, George and Manson2013), as well as prison samples (Mokros et al., Reference Mokros, Menner, Eisenbarth, Alpers, Lange and Osterheider2008).
While the idea of psychopaths as cheaters has received some support, there are some limitations to this theory that we would like to address. Frequency-dependent selection is a model used to explain behavioural variance and the co-existence of different behavioural morphs as a result of stable, obligate, genetically influenced strategies employed by certain non-human species (Jones et al., Reference Jones, Afkhami, Akçay, Bronstein, Bshary, Frederickson and Porter2015; Lalumière et al., Reference Lalumière, Mishra, Harris, Duntley and Shackelford2008). One well-known example is the alternative male phenotypes in bluegill sunfish: larger, parental males who invest in growth and reproduction, and smaller ‘sneaker’ males who engage in quick and transient copulations (Gross & Charnov, Reference Gross and Charnov1980). While cheater phenotypes have been documented in species such as insects and fish, currently the presence of equivalent strategies in highly social species has not received enough empirical support (Riehl & Frederickson, Reference Riehl and Frederickson2016). The specific genetic mechanisms that maintain variance in personality traits in humans, and in particular the role of balancing selection (which refers to adaptive forces that maintain genetic variation in populations and includes frequency-dependent selection) have not been firmly established yet (Penke & Jokela, Reference Penke and Jokela2016). If psychopathy is maintained in populations under frequency-dependent selection, then we should expect it to have a strong genetic component (Fitzpatrick et al., Reference Fitzpatrick, Feder, Rowe and Sokolowski2007; Trotter & Spencer, Reference Trotter and Spencer2007). However, recent evidence shows that psychopathy is the result of multiple genetic and environmental factors which interact in dynamic ways (for a comprehensive review on this topic see Maung, Reference Maung2021). Psychopathy might have a genetic component, as shown by heritability studies that estimate that 40–60% of the variance in psychopathy can be explained by genetic factors (Bezdjian et al., Reference Bezdjian, Raine, Baker and Lynam2011; Blonigen et al., Reference Blonigen, Carlson, Krueger and Patrick2003, Reference Blonigen, Hicks, Krueger, Patrick and Iacono2005; Glenn & Raine, Reference Glenn and Raine2014: 23; Johansson et al., Reference Johansson, Santtila, Harlaar, von der Pahlen, Witting, Ålgars and Sandnabba2008; Larsson et al., Reference Larsson, Andershed and Lichtenstein2006; Taylor et al., Reference Taylor, Loney, Bobadilla, Iacono and McGue2003; Vernon et al., Reference Vernon, Villani, Vickers and Harris2008; Viding et al., Reference Viding, Blair, Moffitt and Plomin2005). However, genes most likely confer vulnerability to developing psychopathy, rather than causing it directly (Viding et al., Reference Viding, McCrory and Seara-Cardoso2014).
Moreover, explanations of psychopathy based on frequency-dependent selection assume that psychopaths are a distinct class of individuals that can be separated from non-psychopaths. However, the dimensional nature of psychopathic traits questions the existence of a specific exploitative strategy of this kind. A critique of the view of specific strategies is that humans have been shaped by natural selection to respond more flexibly to the environment, rather than to rely on fixed strategies (Buss & Greiling, Reference Buss and Greiling1999; Glenn et al., Reference Glenn, Kurzban and Raine2011). As psychopathic traits vary on a continuum of severity in populations, and psychopathic traits are a result of an interplay between multiple factors in development, we believe that a more nuanced approach is required to explain whether and how psychopathy is adaptive. This would require a consideration of how psychopathic traits develop, which can ultimately elucidate how the fitness benefits of psychopathy are delivered. In the next section, we will review a different evolutionary explanation for psychopathy that takes into account the variation in early life environment and is based on life history theory.
2.2. Life history theory
2.2.1. Life history theory in a nutshell
Life history traits are traits of an organism that concern growth, reproduction and survival: the age and size at maturation, number and size of offspring, age at first reproduction, the overall lifespan, etc. Life history theory (LHT) explains inter-species variation in those traits as a result of environmental variation (in mortality risk) and predicts trade-offs between different fitness components (e.g. reproduction and somatic maintenance, quality and quantity of offspring; Flatt & Heyland, Reference Flatt, Heyland, Flatt and Heyland2011; Stearns, Reference Stearns1976). Accordingly, species can be classified into fast, ‘r-selected’ strategists (high growth rate, fecundity and mortality) and slow, ‘K-selected’ (slow growth rate, fewer offspring, longer lifespan) strategists (MacArthur & Wilson, Reference MacArthur and Wilson1967). Some scholars have argued that individuals can vary on this slow–fast continuum within a species, and this premise has been applied to certain human traits and behaviours (for critical reviews see Nettle & Frankenhuis, Reference Nettle and Frankenhuis2020; Sear, Reference Sear2020). Specifically, a harsher and more unpredictable environment (i.e. higher extrinsic mortality) early in life is predicted to favour behavioural traits associated with fast life strategies (Brumbach et al., Reference Brumbach, Figueredo and Ellis2009). Following from this view, psychopathy represents a fast life history strategy focused on investment in mating effort over parenting effort (Barr & Quinsey, Reference Barr, Quinsey, Crawford and Salmon2004; Jonason et al., Reference Jonason, Li, Webster and Schmitt2009, Reference Jonason, Koenig and Tost2010).
2.2.2. Early life environment, stress management and psychopathy
Our species possesses a high degree of developmental plasticity, which enables us to react to external influences from our environment, especially in early life, leading to changes in our brain, anatomy and physiology (Frankenhuis & Amir, Reference Frankenhuis and Amir2022). This is evolutionarily advantageous because it means that we can respond to the environment flexibly and adapt to adversities. At a molecular level this occurs through epigenetic processes which alter genetic expression without changing genetic sequences, and this in turn irreversibly alters the development of an individual (Lea et al., Reference Lea, Tung, Archie and Alberts2017). Recent evidence shows that genes and the environment are interdependent in predicting individual susceptibility to psychopathy (Maung, Reference Maung2021). Accordingly, individuals with an inherited risk might develop the condition only in the presence of environmental triggers, and vice-versa.
Longitudinal studies have linked psychopathy in adulthood to several detrimental factors in childhood, such as low neighbourhood socioeconomic status, parental criminal record, mother's low level of education and mother's young age of pregnancy, factors thought to contribute to non-optimal parenting (Bamvita et al., Reference Bamvita, Larm, Checknita, Vitaro, Tremblay, Côté and Hodgins2017; Farrington et al., Reference Farrington, Coid, Harnett, Jolliffe, Soteriou, Turner and West2006; Piquero et al., Reference Piquero, Farrington, Fontaine, Vincent, Coid and Ullrich2012). This correlation between psychopathic traits and early stress seems to be moderated by genetic and developmental factors. For instance, low socioeconomic status was associated with callous–unemotional traits only in youth with a homozygous genotype for the serotonin transporter protein gene (SLC6A4), which is associated with reduced threat and stress responsivity (Sadeh et al., Reference Sadeh, Javdani, Jackson, Reynolds, Potenza, Gelernter and Verona2010). In a longitudinal study, Beaver et al. (Reference Beaver, Hartman and Belsky2015) found that infants with an easy temperament were the most at risk of presenting affective psychopathic traits in adolescence if they experienced low parental sensitivity, while being the least at risk if they experienced high parental sensitivity. The researchers further argued that individuals might have differential susceptibility to developing psychopathy, whereby infants who are more sensitive to the environment might be especially susceptible to external conditions for better or worse (Ellis et al., Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and Van IJzendoorn2011). Callous–unemotional (CU) traits are a constellation of traits related to lack of empathy, guilt, low empathy and interpersonal callousness, which are a hallmark of psychopathy in youth (Forsman et al., Reference Forsman, Lichtenstein, Andershed and Larsson2008). Willoughby et al. (Reference Willoughby, Mills-Koonce, Propper and Waschbusch2013) found that harsh-intrusive parenting in infancy predicted callous–unemotional traits at age 3, and that this relationship was stronger for children with a polymorphism of the brain-derived neurotropic factor (BDNF) gene, which is involved in processing fear and learning from punishment. Another study is reflective of how the environment can affect biological predisposition: in an adoption cohort of 561 families, the biological mother's self-reported antisocial behaviour predicted children's CU traits at 27 months despite not having contact with the children (pointing to a genetic/inherited bias), but the study also showed that the adoptive mother's positive reinforcement buffered the prenatal risk and protected against CU traits (Hyde et al., Reference Hyde, Waller, Trentacosta, Shaw, Neiderhiser, Ganiban and Leve2016). Other studies have further validated the role of both heritable risk and protective factors: the biological mother's fearlessness predicted CU traits in children, while the adoptive mother's positive parenting buffered against this inherited bias (Waller et al., Reference Waller, Shaw and Hyde2017).
A recently proposed model could explain this role of the environment in facilitating psychopathic traits, namely the Adaptive Calibration Model (ACM; Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011). The ACM uses an evolutionary-developmental framework to explain individual differences in stress responsivity and how they affect behaviour. According to the ACM, the calibration of an individual's stress-response system, composed of the autonomic nervous system and the hypothalamic–pituitary–adrenal axis, influences the development of life history-related traits such as mating, risk-taking and parenting (Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011). The pattern an individual adopts depends on both genetic predisposition and environmental conditions early in life, which can affect the body's response to stress. Harsh and threating environments can cause a frequent activation of the stress-response system, which can alter gene expression, which in turn affects hormone levels and behavioural outcomes (Figure 1). Over time this can lead to the adoption of an unemotional pattern characterised by low stress responsivity via low stress-response system activity. Arguably, this can be advantageous because it protects individuals from stress, while also facilitating a fast life history strategy which is beneficial in ‘dangerous’ environments (Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011; Ellis & Del Giudice, Reference Ellis and Del Giudice2019). This is achieved because individuals become insensitive to danger and negative social feedback (disproval, shame, etc.), which enables competitive risk taking, antisocial behaviour and impulsivity. Moreover, this pattern is predicted to be male biased because a fast strategy focused on risk-taking and mating effort may be more beneficial for males (Del Giudice et al., Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and Van IJzendoorn2011), based on the sex differences in the consequences of mating and parenting efforts (Trivers, Reference Trivers1972; Kokko and Jennions, Reference Kokko and Jennions2008).
Figure 1. Structure of the stress response system and its regulatory processes. The stress response system encodes information from the local environment which feeds back into the development of life history traits (after Del Giudice et al., Reference Del Giudice, Ellis and Shirtcliff2011).
Psychopathic traits are not only correlated with stressful conditions in early life, but also seem to protect against stress later in life. There is evidence that psychopathic traits protect against posttraumatic stress through emotional numbing (Anestis et al., Reference Anestis, Harrop, Green and Anestis2017; Hicks et al., Reference Hicks, Vaidyanathan and Patrick2010; Kerig et al., Reference Kerig, Bennett, Thompson and Becker2012; Willemsen et al., Reference Willemsen, De Ganck and Verhaeghe2012). In a student sample, affective psychopathic traits were negatively correlated with indicators of emotional distress (depression, anxiety and stress; Međedović et al., Reference Međedović, Wertag and Sokić2018). Similarly, in male inmates interpersonal-affective traits were correlated with lower rates of anxiety (Sandvik et al., Reference Sandvik, Hansen, Hystad, Johnsen and Bartone2015). These findings show that psychopathy can act as a coping mechanism for emotional distress, while also reducing mental illness associated with it. These observations fit the predictions of the ACM that the unemotional and remorseless aspect of psychopathic traits should be advantageous in adverse environments because it buffers against the detrimental effect of long-term stress (Del Giudice et al., Reference Ellis, Boyce, Belsky, Bakermans-Kranenburg and Van IJzendoorn2011).
The behavioural traits associated with life history strategies involve risk-taking, parenting and mating behaviours, and some scholars suggested that psychopaths adopt a fast life history strategy focused on investment in mating effort over parenting effort (Barr & Quinsey, Reference Barr, Quinsey, Crawford and Salmon2004; Jonason et al., Reference Jonason, Li, Webster and Schmitt2009, Reference Jonason, Koenig and Tost2010). This prediction has received some empirical support. Jonason et al. (Reference Jonason, Koenig and Tost2010) used the Mini-K, a 20-item measure of life history strategy, to show that psychopathic traits are correlated with life history characteristics such as risk-taking, future discounting, and social, familial and sexual relationships that are representative of a fast strategy. It is important to note here that there have been some criticisms of the application of LHT to psychology. In their reviews both Sear (Reference Sear2020) and Nettle and Frankenhuis (Reference Nettle and Frankenhuis2020) have argued that LHT applied to psychology significantly differs from the life history research in evolutionary biology. While the latter relies on testable mathematical models quantifying variations in strategies according to ecological differences at population and species level, LHT as applied to human psychology lacks theoretical formalisation and relies on intuitive reasoning as to why certain traits might facilitate a specific strategy (Nettle & Frankenhuis, Reference Nettle and Frankenhuis2020). Moreover, psychological research using LHT is often limited to Western, educated populations, which limits our understanding of the variation in life history strategies under different environments (Sear, Reference Sear2020). Resolving these issues and integrating evolutionary and psychological models of LHT is required to appropriately assess whether psychopathy represents a fast life history strategy in a strict sense.
3. Is psychopathy adaptive?
3.1. Psychopathic traits are involved in adaptive trade-offs
The most common definition of evolutionary fitness is based on the probability of an individual passing their genes to the next generation, which is referred to as ‘reproductive fitness’ (Hamilton, Reference Hamilton1964), although current fitness definitions remain contentious (Kuzenkov & Morozov, Reference Kuzenkov and Morozov2019; Orr, Reference Orr2009). More accurately, we can extend the definition of fitness to represent the survival of a genetic lineage which depends not only on individual reproduction and survival, but also on the survival of those carrying one's genes, termed ‘inclusive fitness’ (Akçay & Cleve, Reference Akçay and Van Cleve2016; West & Gardner, Reference West and Gardner2013). Therefore, considerations of how individuals interact with their close social group and kin are important for assessments of evolutionary fitness.
Both the frequency-dependent selection based and LHT-based evolutionary explanations of psychopathy predict it to be fitness-maximising under certain conditions, ultimately enhancing survival and reproduction. The frequency-dependent selection model predicts psychopathic traits to be adaptive when they are expressed in low frequencies, while LHT-based explanations predict those traits to be adaptive under certain developmental contexts. What evidence do we have about psychopathic traits and their fitness correlates? In this section, we will review the primary research on psychopathic traits and fitness proxies, behavioural tendencies that can influence reproduction and survival, as well as some measurable fitness outcomes. Table 2 summarises the reviewed studies and their main findings. It is worth mentioning that these fitness proxies may not directly correspond to actual fitness, a limitation which we will discuss below.
Table 2. Summary of the literature on the fitness correlates of psychopathic traits and their potential relevance for reproduction and/or survival
Although there are a few studies investigating the correlation between psychopathic traits and direct measures of fitness (e.g. lifetime number of children and number of grandchildren), most studies use fitness proxies such as mating and parenting behaviour (Table 2). Psychopathic traits have been associated with desires, motivations and behaviours that reflect heightened sexuality, which could influence reproductive success. For example, they are linked to promiscuous sexual fantasies and behaviours (Jonason et al., Reference Jonason, Li, Webster and Schmitt2009; Visser et al., Reference Visser, Pozzebon, Bogaert and Ashton2010, Reference Visser, DeBow, Pozzebon, Bogaert and Book2015), an early sex life (Harris et al., Reference Harris, Rice, Hilton, Lalumiere and Quinsey2007; Međedović, Reference Međedović2019b; Visser et al., Reference Visser, Pozzebon, Bogaert and Ashton2010), sexual risk-taking and sensation-seeking (Fulton et al., Reference Fulton, Marcus and Payne2010; Kastner & Sellbom, Reference Kastner and Sellbom2012), and unrestricted sexual relationships (Jonason et al., Reference Jonason, Li, Webster and Schmitt2009). Psychopathic individuals not only seek out short-term relations, but also potentially possess the ability to attract mates through favourable impressions (Brazil & Forth, Reference Brazil and Forth2020; Holtzman & Strube, Reference Holtzman and Strube2013; Monteiro et al., Reference Monteiro, Lopes, Nascimento, Gouveia, Shackelford and Zeigler-Hill2017; Visser et al., Reference Visser, Pozzebon, Bogaert and Ashton2010). These observations suggest that psychopathic traits not only influence the desire for promiscuity and short-term relationships, but potentially also reduce the inhibition of those behaviours.
To test the predictions from the LHT, several studies investigated the associations between early life environment, psychopathic traits and fitness proxies. In a community sample, individuals higher in the affective component of psychopathy which captures a lack of empathy had an especially strong relationship between childhood environmental harshness (family dysfunction, poverty) and high mate seeking, short partner relationships and lower parental effort (Međedović, Reference Međedović2019b). Likewise, the same study group found that in male convicts, affective traits promoted reproductive success in harsh environments: individuals with affective traits had a higher number of children if raised in families with parental substance abuse than if they were raised in families without substance abuse (Međedović et al., Reference Međedović, Petrović, Želeskov-Đorić and Savić2017). Thus, a harsh environment can mediate the relationship between psychopathy and fitness and serve as an adaptive response to stress.
Meanwhile, psychopathy could have detrimental effects on long-term relationships as it is often correlated with a lack of investment in partners (Međedović, Reference Međedović2019b; Smith et al. Reference Smith, Øverup and Webster2019) and abuse in relationships (Kirkman, Reference Kirkman2005). Parents with higher scores of psychopathy show lower parental investment and more negative parenting in self-reported (Beaver et al., Reference Beaver, da Silva Costa, Poersch, Freddi, Stelmach, Connolly and Schwartz2014b; Međedović, Reference Međedović2019b) and offspring-rated questionnaires (Međedović & Petrović, Reference Međedović and Petrović2019), as well as in qualitative studies (Kirkman, Reference Kirkman2005; Leedom et al., Reference Leedom, Bass and Almas2013). Additionally, psychopathy might negatively influence whether individuals support their children financially (Dion et al., Reference Dion, Braver, Wolchik and Sandler1997), while also making them more likely to favour partners that also display low parental care (Lyons et al., Reference Lyons, Brewer and Carter2020). The findings summarised in Table 2 highlight potential adaptive trade-offs between fertility and mortality, as well as offspring quantity and quality. In terms of fertility, findings are inconsistent with studies showing both positive (Međedović & Petrović, Reference Međedović and Petrović2019; Neumann et al., Reference Neumann, Schmitt, Carter, Embley and Hare2012) and negative correlations (Carter et al., Reference Carter, Lyons and Brewer2018), while others suggest that only interpersonal traits of psychopathy positively predict fertility (Međedović, Reference Međedović2019a; Međedović et al., Reference Međedović, Petrović, Želeskov-Đorić and Savić2017). However, the use of different samples (men vs. women; reproductive vs. post-reproductive individuals) could be partially responsible for these inconsistent findings. On the other hand, psychopathic scores, especially in the lifestyle-antisocial domain, predict higher individual and offspring mortality and morbidity (Beaver et al., Reference Beaver, Nedelec, da Silva Costa, Poersch, Stelmach, Freddi and Boccio2014a; Hudek-Knežević et al., Reference Hudek-Knežević, Kardum and Mehić2016; Jonason et al., Reference Jonason, Baughman, Carter and Parker2015; Međedović & Kujačić, Reference Međedović and Kujačić2020; Neumann et al., Reference Neumann, Schmitt, Carter, Embley and Hare2012), as well a potential decrease in fecundity (Jonason & Lavertu, Reference Jonason and Lavertu2017). This is most likely due to irresponsible, reckless and impulsive behavioural tendencies which can result in negative health outcomes and decreased longevity.
Nevertheless, trade-offs between offspring quantity and quality, and fertility and mortality do not necessarily lead to straightforward predictions on fitness. At present most research is limited to proxies of fitness rather than measurable fitness outcomes, which limits our ability to test the evolutionary predictions. Another point is that research in this area rarely explains why the use of such fitness proxies (e.g. mating and parenting behaviours) is adequate to validate theories. Even if psychopaths engage in multiple sexual relationships, for example, that does not mean that they will have increased fitness as per number of children, compared with individuals who engage in few relationships with long-term commitments. Similarly, a psychopathic parent with lowered parental investment will not have reduced offspring fitness if the other partner compensates or there is adequate resource availability. Moreover, while psychopathic individuals might have abusive parenting styles, there is no considerable evidence that they will endanger the lives of their children. There is only one study to date that looks at nepotistic behaviours in psychopaths. Data from the case files of 289 violent males revealed that psychopathy predicts nepotistic inhibition of violent behaviour: as the PCL-R score increased, the relatedness between the offender and its victim decreased (Krupp et al., Reference Krupp, Sewall, Lalumière, Sheriff and Harris2012). This reflects an inhibition on harming relatives and therefore might increase the inclusive fitness of psychopathic traits.
3.2. The importance of cultural context
We should also expect the adaptiveness of psychopathy to depend on the cultural context as different social norms and values will dictate how tolerant people are of psychopathic individuals. Some researchers argue that modern individualistic societies facilitate psychopathic behaviour by encouraging competition and impersonal relationships as opposed to collectivist societies that emphasise altruism and social responsibility (Cooke & Michie, Reference Cooke and Michie1999; Jakobwitz & Egan, Reference Jakobwitz and Egan2006). Indeed, a recent cross-cultural study showed an association between cultural values associated with individualism and the presentation of psychopathy: psychopathic traits were positively correlated to a value-set of selfishness, a preference for social inequality and a lack of social cohesion, and negatively correlated with collectivist values of group responsibility and integrity (Shou et al., Reference Shou, Lay, De Silva, Xyrakis and Sellbom2021). Empirical observations also show that individuals who score higher on self-report measures of psychopathy place more value on acquiring power, financial success and material possessions, but not on achieving these goals through ambition and achievement (Glenn et al., Reference Glenn, Efferson, Iyer and Graham2017). Moreover, there is some research that points to the propensity of psychopathic individuals to occupy leadership and high-risk professional roles (Boddy, Reference Boddy2010; Lilienfeld et al. Reference Lilienfeld, Latzman, Watts, Smith and Dutton2014; Howe et al., Reference Howe, Falkenbach and Massey2014; Patton et al., Reference Patton, Smith and Lilienfeld2018). Another insightful area of research is that on ‘corporate psychopaths’, i.e. individuals who score high on psychopathy and tend to occupy power roles in companies (Wellons, Reference Wellons2012). For example, one study showed that in a sample of 203 corporate professionals, 3.9% of the sample presented PCL-R scores of 30 or more (which meets the criteria for a psychopathy diagnosis), which is higher than the 1% seen in the general population (Babiak et al., Reference Babiak, Neumann and Hare2010). These observations are suggestive of the adaptive potential of psychopathic traits in modern societies where occupational prestige can act as a proxy for social ranking. Status-seeking traits (dominance, prestige, etc.) have relevance for fitness in non-egalitarian human societies as they influence resource allocation, mating, cooperation and conflict patterns (Cheng et al., Reference Cheng, Tracy and Henrich2010).
In small-scale societies where social norms promote egalitarianism, prestige is not determined by one's individualistic achievements but through prosocial behaviours (Glowacki and Lew-Levy, Reference Glowacki and Lew-Levy2022). For most of human evolutionary history, our species lived by hunting and gathering, hence the key differences between hunter–gatherer and industrialised societies are likely to have profound implications for mental health (for a recent review see Chaudhary & Salali, Reference Chaudhary, Salali, Abed and St John-Smith2022). One of these key differences lay in the social organisation. Immediate-return hunter–gatherers (who do not store food and are highly mobile) live in campsites of 30–40 people that comprise multiple households (Marlowe, Reference Marlowe2005), and the social organisation is characterised by egalitarianism (Woodburn Reference Woodburn1982). Decisions are consensus-based, there is no leader or formal social ranking, and there is a heavy emphasis on individual autonomy and rejection of dominant behaviour (Woodburn, Reference Woodburn1982). The ultimate driver of this social organisation is the hunting–gathering niche: because successful acquisition of meat is unpredictable, the shortfall in food is compensated for by food-sharing networks (Kaplan et al., Reference Kaplan, Hill, Lancaster and Hurtado2000; Dyble et al., Reference Dyble, Thompson, Smith, Salali, Chaudhary, Page and Migliano2016). Social networks in hunter–gatherers are not only crucial in food sharing but also in childrearing (Page et al., Reference Page, Chaudhary, Viguier, Dyble, Thompson, Smith and Migliano2017) and the accumulation of medicinal plant knowledge (Salali et al., Reference Salali, Chaudhary, Thompson, Grace, van der Burgt, Dyble and Migliano2016), which are relevant for individual fitness. Indeed, among the BaYaka hunter–gatherers in Congo, individuals with larger social networks are found to have a higher body mass index and greater fertility (Chaudhary et al., Reference Chaudhary, Salali, Thompson, Rey, Gerbault, Stevenson and Migliano2016). Individuals who do not cooperate are ostracised (Wiessner, Reference Wiessner2005). Thus, for our ancestors, social networks were a matter of life and death, egalitarian group living was the standard and social isolation was rare, with fatal consequences (Chaudhary & Salali, Reference Chaudhary, Salali, Abed and St John-Smith2022). The strong social networks and egalitarianism of hunter–gatherers could undermine the benefits of antisocial behaviours employed by psychopaths. Psychopaths would also be more likely to acquire an unfavourable reputation in small-scale societies owing to their interconnectedness.
Nonetheless, while we might expect psychopathy to be influenced by cultural factors, empirical observations are yet to establish whether psychopathy rates vary between cultures. There are several studies which confirmed the broad validity of psychopathy and its syndromes across cultures using both PCL-R (Cooke at al., Reference Cooke, Michie, Hart and Clark2005; Fanti et al. Reference Fanti, Lordos, Sullivan, Kosson and Patrick2018; Reale et al., Reference Reale, Lee and Sea2022; Shariat et al., Reference Shariat, Assadi, Noroozian, Pakravannejad, Yahyazadeh, Aghayan and Cooke2010; Sullivan et al., Reference Sullivan, Abramowitz, Lopez and Kosson2006; Wilson et al., Reference Wilson, Abramowitz, Vasilev, Bozgunov and Vassileva2014) and self-report measures (Latzman et al., Reference Latzman, Megreya, Hecht, Miller, Winiarski and Lilienfeld2015; Shou et al., Reference Shou, Sellbom and Han2017) with some subtle differences in total scores and the presentation of specific psychopathic domains. For instance, in a comparison of North American and Iranian psychopathic offenders and patients, while a deficient emotional response was a core characteristic of Iranian psychopaths, interpersonal attributes such as deceitfulness and superficiality did not hold as much predictive power to differentiate between Iranian psychopaths and non-psychopaths (Shariat et al., Reference Shariat, Assadi, Noroozian, Pakravannejad, Yahyazadeh, Aghayan and Cooke2010). The authors suggested that this could be due to the collectivism of Iranian society, whereby individuals are expected to obey societal norms regardless of personal aspirations, and therefore it is more difficult to assess the genuineness of intentions. However, it is important to note that the studies mentioned above were carried out in large-scale, industrialised societies, no studies to date have assessed psychopathy in small-scale societies and few attempt to conduct measurement invariance studies across populations (Mokros et al., Reference Mokros, Neumann, Stadtland, Osterheider, Nedopil and Hare2011).
Currently, the lack of cross-cultural studies undermines the predictive value of adaptive models of psychopathy. There are theoretical expectations as to why the expression of psychopathy should differ in individualistic and impersonal societies vs. collectivist societies according to contrasting societal values (individual success vs. cooperation). Moreover, cultural variations in the expression of specific psychopathic domains of psychopathy (affective, interpersonal, etc.) across cultures could imply that certain psychopathic traits are culture specific. However, such cultural differences could be suggestive of different rates of psychopathy or that psychopathic individuals inhibit their behaviours to comply with societal norms and to avoid ostracism.
4. Future directions
So far, we have reviewed evolutionary explanations of psychopathy, the evidence for and shortcomings of those explanations and the studies investigating fitness proxies of psychopathy. Throughout our review we have identified some limitations to the existing data, which in turn have led to the potential future directions below.
Firstly, current evolutionary models of psychopathy often look at proxies of fitness such as increased mating effort or reduced parental investment. Nevertheless, such proxies may not always correspond to reproductive success. As we discussed earlier, the context in which interactions with psychopathic individuals take place can alter the fitness consequences of those interactions (e.g. lack of parental investment can be compensated by cooperative childcare). More studies should investigate fitness outcomes using measures of fertility (total number of children), fecundity (e.g. through measures of reproductive health) and individual and offspring mortality, as these have higher predictive value for reproduction and survival than behavioural proxies only. Additionally, the sample choice is important: post-reproductive individuals might provide the most accurate measure of fitness by assessing lifetime reproductive success (Jones et al., Reference Jones, Afkhami, Akçay, Bronstein, Bshary, Frederickson and Porter2015; Orr, Reference Orr2009). There are also virtually no studies that assess the interactions with kin and nepotistic tendencies of psychopaths. At the moment the lack of data in this area prevents us from making predictions on the inclusive fitness of psychopathic traits, which is a crucial component of evolutionary fitness. Furthermore, an interdisciplinary approach which considers the environmental context and integrates observations from evolutionary, psychological, developmental and anthropological studies is the best way forward. Potentially, this approach would provide insight into the specific factors that confer fitness advantages to psychopathic traits, as well as how they can be maintained by natural selection.
Secondly, theories that conceptualise psychopathy as an adaptive strategy rely on the assumption that psychopathic tendencies to deceive and manipulate automatically translate to actual manipulation taking place. Yet we do not currently have enough data to understand the receiving end of this manipulation, and whether it is indeed effective, or whether there are any social factors that determine who is most susceptible to manipulation by psychopaths. While we understand that genuine interactions of this sort cannot be feasibly assessed, this limitation should be acknowledged in discussions about the advantages of cheating behaviours of psychopaths. Potentially, more interviews or surveys with family members and peers of psychopaths could be insightful on this topic as it might give us an idea about the frequency and efficiency of manipulative tactics employed by psychopathic individuals. In addition, incorporating this data would also help minimise the overreliance on self-report measures, which are predominant in psychopathic literature, especially in studies carried out outside clinical and forensic settings.
Although not discussed in detail in our paper, narrow psychopathy dimensions are differentially related to functionally adaptive traits. There is some evidence to suggest that factor 1 traits (callous affect and manipulative interpersonal style) have a higher degree of adaptive functionality than factor 2 (antisocial behaviour and reckless lifestyle). For instance, factor 1 traits are associated with higher intellectual abilities and executive functioning than factor 2 traits (Ben-Yaacov & Glicksohn, Reference Ben-Yaacov and Glicksohn2018; Neumann & Hare, Reference Neumann and Hare2008; Pasion et al., Reference Pasion, Cruz and Barbosa2018; Ross et al., Reference Ross, Benning and Adams2007; Vitacco et al., Reference Vitacco, Neumann and Jackson2005, Reference Vitacco, Neumann and Wodushek2008). Moreover, factor 1 traits can potentially protect against other mental disorders (Benning et al., Reference Benning, Patrick, Salekin and Leistico2005; Međedović et al., Reference Međedović, Wertag and Sokić2018; Ragsdale & Bedwell, Reference Ragsdale and Bedwell2013). Some of the studies we reviewed in Section 3 also hint at the possibility that factors 1 and 2 could have different outcomes in terms of fitness with only interpersonal traits positively predicting fertility and lifestyle–antisocial traits predicting higher morbidity. Future studies that assess how independent psychopathic facets, rather than total psychopathy scores, relate to fitness could lead to more accuracy in adaptationist models, while also acknowledging that psychopathy is multidimensional and not a uniform and categorical construct.
Finally, research on psychopathy, and particularly that concerned with evolutionary phenomena, is almost exclusively carried out on WEIRD (Western, educated, industrialised, rich and democratic) samples (Henrich et al., Reference Henrich, Heine and Norenzayan2010). As a recent paper in this special issue has argued, more studies are needed to assess individual psychological differences and their fitness consequences in hunter–gatherer and other non-industrialised societies (Hunt & Jaeggi, Reference Hunt and Jaeggi2022). Our prediction is that psychopathic traits would correlate with fitness in different ways across cultures. For example, in hunter–gatherer groups where individuals with dominant and anti-social behaviour are ostracised, we predict psychopathic traits to be negatively correlated with fitness. It is also worth noting that such traits may be expressed in different ways in those societies, which would require balancing emic (culture-specific) with etic (universal) approaches for measuring those traits (Chaudhary & Salali, Reference Chaudhary, Salali, Abed and St John-Smith2022).
5. Conclusion
Here we reviewed the evidence for and limitations of evolutionary explanations of psychopathy. Although there is some evidence for psychopathic traits presenting adaptive trade-offs between mate-seeking and fertility on the one hand, and mortality, morbidity and decreased parental investment on the other, they do not necessarily lead to straightforward consequences on fitness. At present most research is limited to proxies of fitness rather than measurable fitness outcomes, which limits our ability to test the evolutionary predictions. A more fruitful approach would require a move away from studies that only look at the correlation between psychopathy and fitness proxies, to a more integrated approach that considers proximate mediating factors and developmental triggers. The adaptive potential of psychopathic traits should be considered and assessed in relation to the social factors such as cultural norms, which might affect whether people tolerate, ostracise or are easily deceived by psychopaths.
Acknowledgments
We thank Louise Barret for including our paper in this special issue and her and the two anonymous reviewers’ very helpful comments on the manuscript.
Authors’ contributions
IE conceived the project and reviewed the literature under the supervision of GDS and KKW. IE led the writing with ongoing feedback and contribution of GDS and KKW. All authors commented on the drafts.
Financial support
GDS was funded by the British Academy Postdoctoral Research Fellowship. KKW is currently funded by UKRI Research England. This research received no specific grant from any funding agency, commercial or not-for-profit sectors.
Competing interests
We declare we have no competing interests.
Open access
