Discussion
We have examined whether baseline plasma OT is a biomarker which partially predicts the levels of trust and trustworthiness observed in the TG in a large Han Chinese undergraduate sample. Notably, we observed a U-shape relation between the levels of trust and marginally trustworthiness and subjects' baseline plasma OT level. Specifically, subjects with more extreme levels of plasma OT were more likely to be trusting as well as trustworthy than those with moderate levels of plasma OT. U-shape relationship in hormone and neurotransmitter action is a frequent observation and documented in [32], [40]. In the brain the U-shaped dose-response relationship is particularly characteristic of peptide hormones often presented as an inverted U [41], [42] and sometimes as a simple U shape [32], [33] as observed in the current investigation.
We suggest the notion that the current findings, viz., the extremes of plasma OT predict trust and trustworthiness, reflect the complexities of the mechanisms underlying the role of OT in social cognition and especially, but not exclusively, the trust phenotype. As Bartz et al underscore in a cogent and recent review [43], the effects of OT are constrained by features of situations – simply put ‘context matters’. Indeed, the role of OT on social cognition and prosociality like trust/trustworthiness is nuanced and subtle. The current findings are consistent with this view and, moreover, help refine predictions about plasma OT levels and how this measure can inform the understanding of social cognition. Although a superficial view of OT actions might at first suggest a situation-invariant effect of this hormone on behavior, closer scrutiny suggests that the effects of OT are often moderated by contextual factors, and perhaps equally importantly, by trait characteristics of the subjects themselves. This scenario is not unique to OT. A good example is provided by the paradoxical effect of the stimulant methylphenidate in children with attention deficit; in these hyperactive children an amphetamine (“speed”) like drug has a calming effect [44]. Similarly, paradoxical effects have been observed for positive modulators of the GABA-A receptor (benzodiazepines, barbiturates, alcohol, GABA steroids) which generally induce inhibitory (e.g. anesthetic, sedative, anticonvulsant, anxiolytic) effects but some individuals have adverse effects (seizures, increased pain, anxiety, irritability, aggression) upon exposure [45].
Evidence specifically supports such a non-linear role of OT tone on the complex trust phenotype. For example, a recent investigation shows that administered OT enhances cooperation and reduces betrayal aversion contingent on other personality factors [46]. OT has a non-linear effect on trust, cooperation and betrayal aversion contingent upon an individual's background personality trait of Attachment Avoidance. Similarly, such non-linear effects of OT on trust also characterize borderline personality disorder (BPD) [47]. Results showed that intranasal OT produced opposite actions in BPD (compared to the trust-enhancing effect of OT in normal subject), decreasing trust and the likelihood of cooperative responses. Moreover, U-shaped relationships between OT and behavior are not restricted to humans but have also been observed in animal studies. An especially relevant example has been reported for the role of OT in memory storage and consolidation in mice [48] and rats [49].
Summing up, the U shaped relationship herein observed between plasma OT and trust/trustworthiness is another example, we suggest, of how hormones overall, and OT specifically, may have paradoxically opposite actions contingent on individual differences. We suggest that the quadratic relationship between plasma OT and trust/trustworthiness captures the concept put forward by Bartz et al that ‘context and person matters’ in the action of this nonapeptide hormone [43]. In some individuals, low central OT tone reflected in low plasma OT levels, is associated with trust whereas in other individuals high plasma OT, presumably reflecting high central OT tone, is associated with trust. Bartz et al have suggested in their recent review that endogenous OT reflected in plasma measurements could be a biomarker of sensitivity to social cues and/or social motivation. Low plasma OT, which has been reported in autism [50], would reflect social insensitivity and motivation whereas high plasma OT could reflect increased social sensitivity and motivation. Hence, both low and high social sensitivity may drive trust/trustworthiness as observed in the current report. Low social sensitivity may make such individuals less betrayal averse and less fearful of exploitation and hence more likely to trust in the TG. Similarly, subjects with high plasma OT indicating increased social sensitivity and motivation would also show enhanced trust/trustworthiness driven by the warm glow of positive social interactions incurred in participating in the TG.
There is a large body of literature in evolutionary biology on gender differences regarding reproduction and investment cost towards offspring [51]. As females usually incur higher costs for reproduction including giving birth and raising offspring, it is suggeseted that females will increase their fitness (number of viable offspring) by investing heavily in fewer children but insuring their the survival. On the other hand, males would increase their fitness by accessing to the maximum number of females they can mate with since the male investment in child bearing and rearing is less than the female. The behavioral consequences of gender difference in reproduction are unclear. It is, however, possible that these different investment strategies could lead to different psychological mechanisms towards trust and trustworthiness but also different biological mechanisms underpinning these two constructs for each of the sexes. In our findings, although the relationship between plasma OT and trust related behavior is significant in the pooled sample, separate analysis by gender revealed that the effects are more specific to the male side. This is in agreement with a large number of animal studies suggesting that OT differentially affects females and males, likely reflecting the interactions between OT and gonadal hormones [52], [53]. Indeed, the effects of neuropeptides across different species show sexual dimorphic effects on behavior. For example, the parenting behavior of female prairie voles is more dependent on OT, whereas male parenting behavior is more associated with AVP [54]. Furthermore, there are gender differences in OT receptor binding depending on the brain area and the species. For instance, female prairie voles show higher binding of OT receptors in the medial prefrontal cortex compared to males [55]. Higher OT receptor binding was found in the hypothalamus of female rats compared to male rats [56]. Altogether then, it is not surprising that in the current investigation we observe a gender-specific relationship between plasma OT levels and the level of trust and marginally trustworthiness observed in the trust game.
In “All's Well That Ends Well” Shakespeare advised “Love all, trust a few, do wrong to none”. This sage advice recognizes the inherent risk of betrayal embodied in bestowing trust [57], [58], [59]. The TG is designed to measure trust but its measurement is confounded by the element of risk since we must also consider that trust is not reciprocated [60]. To minimize the confounding aspect of risk inherent in the TG, we also assessed subjects' risk attitude using a portfolio choice design [25]. The use of this design allowed us to decompose trust and achieve a more refined estimate of an individual's trusting behavior by minimizing potential confound of risk proneness or risk aversion. Our experimental design shows that plasma OT is reflecting trust and not risk attitude, as there appears to be no significant relationship between this hormone measure and risk. However, we are aware of the potential confound of trust and trustworthiness with other prosocial behaviors including altruism and inequity aversion [61]. As a consequence, our study needs to be interpreted with caution since the results we report might not necessarily reflect only trust and trustworthiness, but other facets of prosocial behaviors that are related to the role of plasma OT.
There are two caveats in place regarding measurement of plasma OT used in our study. The first question has to do with the laboratory method, for which we argue that our procedure is indeed reliable and robust for measuring plasma OT as detailed in the Methods section. Notwithstanding, we would also like to point out to the reader the report by Szeto et al [21], which raises the possibility that the procedure adopted by us and studies of others could be subjective to criticism and could be a potential limitation of the current investigation. However, the strengths of this study need also to be underscored viz., the careful measurement of the phenotype as well as the very large number of subjects examined. Secondly, whether plasma OT indeed is an informative measurement for CNS oxytocin remains unclear and needs to be fully resolved [19]. Many questions remain regarding how robustly and by what pathways (peripheral and central release) this biological marker reflects human social behavior. In the current report we interpret base-line plasma OT as a partial indicator or biomarker for neuropeptide ‘tone’ that reflects long-term chronic oxytocin activity. An alternative hypothesis proposed by Porges is that peripheral OT levels partially indexed in plasma levels could also be a measure in part of the vagal regulated ‘social engagement system’ [62]. Porges has suggested in an extensive series of publications that “the mammalian autonomic nervous system provides the neurophysiological substrates for the emotional experiences and affective processes that are major components of social behavior”. The role of OT in parasympathetic modulation, especially as a break on sympathetic heart activation, may facilitate prosocial behavior by establishing a calmer, less-threatening environment. Indeed, vagal tone predicts positive emotions and social connectedness [63]. Altogether, regardless of the source of plasma OT, peripheral or central release, there is good reason to believe that plasma OT levels is related albeit indirectly to social brain/social engagement. Nevertheless, there remain methodological issues surrounding the measurement of oxytocin and hence until these questions are resolved results using plasma measurements of this hormone, need to be interpreted cautiously.
Trust pervades human society and is a critical element in facilitating social interaction and exchange between individuals, groups, businesses, governments and nation states. It is therefore not unexpected that trust is the subject of intense inquiry by scholars across academic disciplines. Over the past decade, by examining trust through the lens of experimental economics, it has been possible to begin to unveil its neurobiological and neuroendocrinological underpinnings. Of special interest is the identification of OT, underpinned by a rich tradition of translational research in animal models [9], with trust in humans. The current report strengthens the link between OT and trust and most importantly, indicates that basal plasma levels of OT may serve as a provisional biomarker for trust and trustworthiness. Zak and Knack [64] have characterised the social, economic and institutional environments in which trust will be high, and show that low trust environments reduce the rate of investment. By exploring the neurobiological roots of trust and specifically showing that plasma OT may be a potential biomarker for human trust, we suggest the notion that a new window of investigation has opened regarding individual, cross cultural and cross group differences in trust related behavior.