Introduction
It is difficult to overstate the role of trust in facilitating the smooth functioning of social and market institutions in modern societies. Trust can be seen to provide the basis for reducing social complexity [1], enhancing social order [2] and social capital [3], as well as overcoming the inherent risk involved in economic exchange and social interaction [4]. In experimental economics, Berg, Dickhaut, and McCabe (1995) invented an economic game, called the Trust Game (TG) in which the first mover is endowed with certain amount of money, and can send any part of it to the second player, called the trustee, which is endowed with no money. The amount received by the trustee is typically tripled the amount sent. The trustee has the option to send any proportion of the tripled amount to the first mover, and keep the rest. Notice that the amount sent by the first mover can be a measure of the degree of trust while the amount sent by the trustee back to the first mover can be a measure of trustworthiness. The TG provides invaluable insights into many basic concepts in human relationships and demonstrates that “reciprocity exists as a basic element of human behavior which is accounted for in the trust extended to an anonymous counterpart” [5]. Since its inception, the incentivized TG has served as the mainstay for the study of trust in the controlled laboratory setting. More recently, the burgeoning field of neuroeconomics has begun to use this game to examine the biological underpinnings of trust [5].
Remarkably, using the TG in the laboratory has enabled the identification of the nonapeptide hormone, oxytocin (OT) as a promoter of trust related behavior. A series of experiments initiated by the seminal study of Kosfeld et al [6] showed that intranasal administration of OT enhances trust but not trustworthiness in the TG. Altogether, a growing body of work has now demonstrated the robust effect of intranasal administration of OT on trust related behaviors. Notably, the effects of sniffing OT on face recognition and in-group trust are significant in recent meta-analysis [7]. Similarly, a comprehensive literature review of the effects of sniffing OT showed that release of this peptide correlates with behavioral changes [8].
In the brain, the main source of OT is the magnocellular neurons of the paraventricular (PVN) and supraoptic (SON) nuclei of the hypothalamus. From these nuclei this hormone reaches the posterior pituitary by axonal transport and is released into the peripheral circulation where it regulates a number of critical physiological processes including parturition and lactation [9]. Importantly, OT is also released from neuronal dendrites and acts at distant brain targets [10]. In the last decade, accumulating evidence shows that this neuropeptide is important in shaping human social cognition and affiliative behaviors [11]. Towards revealing the role of OT in humans, intranasal administration, aka ‘sniffing’, has been a widely used strategy in understanding the action of this hormone in both normal [6], [12] and abnormal [13], [14] social behaviors. Neurogenetic strategies have also been very effectively used in unravelling the role of OT in autistic disorders as well as prosocial behavior in non-clinical subjects [15], [16], with some exceptions [17], [18].
Another widely used strategy in characterizing the role of OT in human social behaviors is the determination of OT levels in the peripheral circulation, albeit the contours of the relationship between plasma OT and CNS oxytocin remain unclear [19]. Indeed, plasma OT is also likely to partially reflect peripheral release of this neuropeptide, however, with no less relevance we suggest to the social brain [8]. Notably, plasma OT level has been shown to be remarkably stable over time. For example, OT levels at early pregnancy and the postpartum period are highly correlated at more than 90% [20]. Although there are technical issues relating to the measurement of plasma OT that remain to be resolved [21], many investigations have reported intriguing correlations between plasma OT and a wide range of behaviors including parent-infant bonding, adult pair-bonding and social relationships among others (reviewed by [8]).
It is the importance of trust and reciprocity in human social exchanges coupled with the considerable evidence that plasma OT levels are related, however non-linearly, to human social behavior, which prompts the current investigation. We are aware of only two previous studies that examined plasma OT levels in relation to trust and trustworthiness using the TG in a laboratory-based setting. Zak et al [22] used a sequential anonymous TG with monetary payoffs with 156 subjects. They find that OT levels rise in subjects who receive a monetary transfer that reflects an intention of trust relative to an unintentional monetary transfer of the same amount. Keri and Kiss [23] used a non-conventional trust paradigm mimicking everyday intimate transactions with 60 subjects and showed that OT was elevated in the trust-related condition relative to a neutral baseline. They also observed a significant positive relationship between trust-related oxytocin level and habituation of autonomic arousal.
Here, we hypothesized that base-line plasma OT, which is correlated with a range of human behaviors, is a biomarker for trust and trustworthiness, beliefs that underpin most human exchanges. To test this hypothesis we measured base-line plasma OT levels in a very large sample of 1,158 undergraduate Han Chinese students at the National University of Singapore who participated in one-shot TG. Notably, this investigation represents the largest sample of subjects yet examined for plasma OT levels in a single study.
10.1371/journal.pone.0051095.g001 Figure 1  Plasma oxytocin and trust.
(A) Scatter Plot on the relationship between plasma oxytocin and trust. (B) Histogram on the relationship between plasma oxytocin and trust.

Ethics Statement
The study is approved by the Internal Review Board at the National University of Singapore. Each participant provides written informed consent (as outlined in the PLoS consent form) to participate in the experiment at the beginning of the study.