The hypothalamic neuropeptide oxytocin (OXT) has increasingly been shown to influence aspects of social cognition, affiliative behavior and social bonds (Bartz et al., 2011; Bethlehem et al., 2013). Evidence also suggests that OXT affects high level cognitive appraisal processes, resulting in the more positive and less negative appraisals of socially-relevant information concerning both self and others (Guastella et al., 2009; Guastella and MacLeod, 2012). Some initial studies suggested that intranasal OXT administration may increase self-perception due to promoting positive self-referential processing (Cardoso et al., 2012) and the ability to discriminate between their own face (self) and that of an unfamiliar person (other; Colonnello et al., 2013). However, more recent studies have found that it increases taking another, but not a self-perspective, in the context of pain perception (Abu-Akel et al., 2015) and other- but not self-orientation in terms of subjects perceiving themselves as being more communal (Bartz et al., 2015). We have also shown that OXT appears to blur the distinction between self and others in terms of trait-judgments and reduced medial prefrontal cortex (mPFC) responses as well as reduced functional connectivity with other regions of the default mode network (DMN) involved in self-processing (Zhao et al., 2016). These latter studies indicating that OXT may actually reduce self-interest resonate more closely with its well established role in promoting affiliative behavior and social bonds (Bartz et al., 2011; Bethlehem et al., 2013). However, it is unclear whether this OXT-effect in reducing self-interest extends beyond the immediate social domain of enhanced responses to the social importance of others, and the emotional behaviors and attributes they exhibit, to a more non-social one such as possessions.

We always place greater value on items that we actually own and this is referred to as the endowment effect (Thaler, 1980; Kahneman et al., 1990, 1991). In other words, we assign higher value and desirability ratings for self-owned objects compared with same ones which are owned by others. The endowment effect is measured in terms of the difference between willingness to accept (WTA) a price for an object and willingness to pay for it (WTP), and WTA is always higher than WTP (Procaccia and Segal, 2003). Loss aversion has traditionally been used to explain the endowment effect which makes buyers frame goods as gains but sellers as losses, and people consider loss as greater than an equivalent gain (Morewedge et al., 2009; Morewedge and Giblin, 2015). However, a number of studies have reported that the endowment effect can also be viewed as an exemplar of a mere ownership effect whereby a target object is rated more favorably by an owner than a non-owner (Beggan, 1992; Gawronski et al., 2007) and a recent review of the field has also emphasized this aspect (Morewedge and Giblin, 2015). One’s sense of self can therefore extend beyond the sense of body ownership and agency (Aron et al., 1991, 2004), to include one’s possessions and these can therefore be considered as part of one’s extended self (Belk, 1988). Because of the intrinsic tendency to enhance one’s self, the association between possession and self (ownership) subsequently boosts the possession’s perceived value (Ebisch et al., 2011; Dommer and Swaminathan, 2013). Our tendency to emphasize our positive self-characteristics is also relevant in this respect and there is evidence that OXT can enhance this positivity (Cardoso et al., 2012). This concept of one’s sense of self being extended to include possessions is further supported in the context of the endowment effect by our recent study showing that in members of a collectivist culture (Chinese) where the sense of self is extended to include close others (mother), this also extends to placing greater value on their possessions (Feng et al., 2013). Based on previous experiments indicating that OXT reduces self-interest by increasing consideration of others we therefore hypothesized that in Chinese subjects it would either selectively reduce the endowment effect for self- and mother-owned objects or alternatively extend it to include objects owned by others not normally included in the extended self (i.e., colleagues or strangers).

The mPFC has been implicated in seemingly disparate cognitive functions, such as understanding the minds of other people, processing information about the self, mental states, physical characteristics (Mitchell et al., 2005; Cikara et al., 2014), self-reflection, person perception and making inferences about others’ thoughts (Amodio and Frith, 2006). In an imagined ownership paradigm, the mPFC showed greater activity for self-owned objects compared with other-owned ones. Additionally, mPFC activation is increased for self- vs. other-owned objects and recognition memory for self-owned objects is correlated with this (Kim and Johnson, 2010, 2013). Increased mPFC activation has also been found in association with the endowment effect (Knutson et al., 2008; Feng et al., 2013). Taken together, these findings provide evidence that the mPFC is involved in the incorporation of self-relevant objects into one’s sense of self. Thus we hypothesized that since OXT decreases the activity of both dorsal and ventral regions of the mPFC during self-processing involving trait-judgments (Zhao et al., 2016), it would have a similar impact in relation to the self-bias associated with possessions.

We have therefore investigated the effects of intranasal OXT effect on self-processing in the context of the endowment effect in two independent studies, the second of which was combined with functional magnetic resonance imaging (fMRI) to investigate the neural substrates involved. Performing two independent studies also allowed us establish the general reproducibility of any observed behavioral effects of OXT and additionally whether two different doses of OXT which have previously been shown to produce functional effects (24 and 40IU—see Striepens et al., 2011; Guastella et al., 2013) might have a different efficacy. In our previous study using Chinese subjects (Zhao et al., 2016) we found evidence using a trait-judgment paradigm both for an extended sense of self incorporating close relatives such as mother, and that OXT effects were moderated by levels of self-esteem. In the current study, we therefore also included ownership of objects by both close (mother/father) and more remote (classmate/stranger) others and investigated potential moderating effects of self-esteem.

Two independent, double-blind, between subjects, placebo (PLC) controlled design studies were performed on different groups of subjects. The first study (Experiment 1) involved a behavioral analysis of the influence of OXT on the endowment effect using a lower functional dose (24IU) while in a second study (Experiment 2) OXT effects on the endowment effect were investigated during simultaneous fMRI scanning and using a higher functional dose (40IU).

The present study investigated whether the effects of OXT on reducing the distinction between self and other which has previously been observed in the context of personal attributes (Zhao et al., 2016), pain empathy (Abu-Akel et al., 2015) and preference for communal situations (Bartz et al., 2015) extends to objects possessed by self and others using an endowment task. The behavioral results from two independent experiments showed a main effect of OXT whereby it significantly enhanced the strength of the endowment effect in individuals from a collectivist culture for both self- and all categories of other- (close: mother/father and remote: classmate/stranger) owned items. Neuroimaging results showed a differential pattern of OXT-evoked changes in the dmPFC and vmPFC with it decreasing activation in the vmPFC for self-owned items but increasing it for mother-owned ones in vmPFC and increasing it in the dmPFC for classmate and stranger-owned items. Activation changes in the dmPFC for self- and mother-owned items were associated with the magnitude of the endowment effect as were differences in activation between self- and remote other-owned items. Additionally, functional connectivity between the dmPFC and the VStr was reduced by OXT in both self and mother-conditions suggesting that self- and mother-owned items may have influenced reward processing. There was also reduced functional connectivity between dmPFC and the precuneus in the self-ownership condition which might further indicate a reduced importance of self-owned items.

The endowment effect can no longer solely be attributed to a traditional loss aversion account (Morewedge et al., 2009; Morewedge and Giblin, 2015), and in the current study it is considered more to reflect the tendency to value self over others (Beggan, 1992; Gawronski et al., 2007; Dommer and Swaminathan, 2013) and increasing the positivity or decreasing the negativity of the self is a means of maintaining this (Brown and Dutton, 1995). Belk (1988) has argued that possessions are an important component of sense of self with an association between a possession and self being created by virtue of ownership. On the other hand evidence suggests that “shared representations” between self and other lie at the core of the phenomena of affective sharing (Bartz et al., 2010). Our current results confirm those of our previous study (Feng et al., 2013) that in subjects from a Collectivist culture shared representations in the context of the endowment effect extend to close relatives, such as mother and father. The behavioral effect of OXT treatment in both experiments carried out was to generally enhance the value placed on both self and all other-owned items, which suggests that it is acting both to further promote the shared representation between self and close others and additionally between self and remote others. On the other hand, the bias towards placing a greater value on self- and mother-owned items compared to remote-other owned ones is still maintained under OXT. This is in general agreement with our previous findings of a main effect of OXT on decreasing response times for making trait-judgments for self as opposed to close or remote others, although in contrast to the current study OXT also abolished the normal self-bias in the accuracy of making trait-judgments (Zhao et al., 2016). Thus, together with findings showing that OXT promotes a more other-orientated social responses (Abu-Akel et al., 2015; Bartz et al., 2015), our current and previous results suggest that it generally facilitates a more extensive shared representation of self both in terms of the social attributes and feelings of others as well as their possessions. Such a shared representation of self might also lead to greater affiliation with others and thereby contribute to the well-established effects of OXT in animal models and humans in promoting the formation and maintenance of social bonds (see Striepens et al., 2011; Hurlemann and Scheele, 2016).

The mPFC is involved in tasks requiring self-evaluation and reflection (Kelley et al., 2002; Moran et al., 2006; Feng et al., 2013). Our current fMRI findings both confirm and extend our previous results showing that the mPFC is the key region where OXT alters activity in association with its behavioral effects in influencing both self and other processing (Zhao et al., 2016). Indeed, the whole brain analysis only revealed both main effects of OXT and ownership in medial prefrontal regions and there were no effects in other brain regions at p < 0.05 FDR corrected. Overall, in the current study OXT increased mPFC activation across conditions whereas for the trait-judgment task it produced an overall decrease (Zhao et al., 2016). In line with our previous finding using a trait-judgment paradigm OXT altered activity in both dmPFC and vmPFC, although while similar reductions in activation were seen for self-owned items in vmPFC a notable difference in the current study was that vmPFC activation was increased for mother-owned items and dmPFC activation was increased for classmate and stranger-owned ones. Thus in the context of the endowment effect OXT increased mPFC activity in all three other-ownership conditions whereas for trait-judgments it decreased it. The dmPFC, activation for both self- and mother-owned items was positively associated with the size of the endowment effect in the OXT group. However, the absolute activation difference between self- and remote other-owned items was also positively associated with the size of the difference in the endowment effect between them (i.e., the smaller the dmPFC activation difference between self and remote others the smaller the difference in the size of their respective endowment effects). Thus OXT’s facilitation of the endowment effect in both self and others appears to be associated with a reduced difference between dmPFC activation in self- and other-owned conditions. In our previous study we also found that mPFC activation during the endowment task was positively correlated with selling price (Feng et al., 2013). On the other hand, for the vmPFC there is a positive association with the size of the extended endowment effect observed for mother-owned items, with OXT effectively promoting a larger difference in activation between self and close others. The vmPFC is activated during tasks involving self-knowledge and person perception (Amodio and Frith, 2006) and has recently been proposed to be particularly associated with the personal importance assigned to self-representations and with aspects of self which have high personal value such as possessions and close others (D’Argembeau, 2013). Thus OXT may have additionally influenced the size of the endowment effect for the possessions of close others (i.e., mother-owned items) by acting on the vmPFC.

Overall therefore the main effect of OXT on increasing the size of the endowment effect for both self and close and remote others may have been contributed to by its differential actions on the dmPFC and vmPFC. Thus OXT may act primarily via the vmPFC to increase the size of the endowment effect in close others (mother) whereas for remote others (classmate and stranger) its effect on the dmPFC appears to be more involved. The relative contributions of the dmPFC and vmPFC to the increase in the size of the endowment effect in the self-condition under OXT appears to be more complex however. While the decreased vmPFC response might indicate a reduction in the size of the endowment effect in the self-condition, there is at the same time a reduced difference between the vmPFC response in the self-condition and that in all three of the other-conditions compared to the PLC group. Thus OXT may overall have acted to increase the size of the endowment effect in the self-condition via the vmPFC despite its overall effect being to reduce vmPFC activity. On the other hand, in the dmPFC the magnitude of the response in the self-condition is positively associated with the size of the endowment effect and at the same time OXT is again acting to decrease the difference between self and remote-others. Thus OXT may also be acting at the level of the dmPFC to increase the size of the endowment effect both in the self- and remote other-conditions.

The functional connectivity analysis showed that OXT treatment reduced the strength of the connectivity between dmPFC and VStr for both self- and mother-owned items, although in neither case was there a significant association with the size of the endowment effect in the two conditions, and the corresponding dmPFC BOLD response was also not significantly altered. The VStr processes information on expected reward magnitude to guide adaptive action preparation and reward learning, and the computation of anticipated gains or expected values (Knutson et al., 2007; Brown and Cai, 2010). Possibly a reduced functional connectivity from the dmPFC to the VStr might therefore reflect an action of OXT in decreasing the reward or valuation of self- and close-other items, thereby enhancing that for remote other-owned items. However, since behaviorally OXT acted to increase the size of the endowment effect in both self- and mother-owned conditions, suggestive of enhanced reward or value, the reduced functional connectivity observed could instead reflect an increase in their reward or value as a result of reduced negative feedback from the VStr to the dmPFC, since we cannot identify the specific direction of the observed functional connectivity change between these two regions. On the other hand the reduced functional connectivity between the vmPFC and the precuneus in the self-owned condition occurred in association with a corresponding reduced vmPFC BOLD response. The functional connectivity change in this case may therefore primarily be driven by the reduction in the BOLD response with both activation and functional connectivity decreases combining to produce a reduced self-bias, the precuneus along with the vmPFC being known to play a role in self-processing (Cavanna and Trimble, 2006).

In contrast to our previous study investigating the effects of OXT on trait-judgments, we found no correlation between its behavioral or neural effects and levels of self-esteem in the endowment task. Another previous study has also reported that activation in the anterior cingulate in response to processing of self-referential items was negatively associated with trait levels of self-esteem (Yang et al., 2012). Since the magnitude of self-esteem tends to have a greater influence in contexts where social rejection is a potential factor (Yang et al., 2012), it is possible that this is of less relevance when judging the value of possessions owned by self or others than in the more personal context of attributing positive or negative behavioral traits to self and others. We also found no evidence for a modulatory influence of SCS scores on behavioral or neural effects of OXT and so it would appear that the relative strength of collectivism compared to independent traits within our cohort of Chinese subjects was not of great importance. It would clearly be of interest however to confirm in future studies whether similar effects of OXT on the endowment effect occur in independent cultures in order to establish whether there are modulatory effects of an independent vs. collectivist orientation.

The present study has several limitations. First, we found no evidence for dose-dependent effects of OXT in the endowment task. Since we chose to use two OXT doses which have both previously been reported to have significant functional effects (Striepens et al., 2011; Guastella et al., 2013) it is possible that this impaired our chances of demonstrating clear dose-dependent effects which might require inclusion of lower doses with weaker functional effects. Additionally, small timing and other differences in task presentation and scanner noise in the second experiment may have also contributed. Second, we only included male subjects in the study to avoid potential issues for controlling for menstrual cycle effects. However, there is growing evidence for sex-dependent behavioral and neural effects of OXT (see Gao et al., 2016) and thus we cannot exclude the possibility that there may also be sex-dependent effects of OXT in relation to self-processing and the endowment effect. Thirdly, in the current study the anxiety/mood assessments were only made prior to treatment and so we cannot exclude the possibility that a non-specific OXT effect on mood/anxiety might have influenced subjects’ behavior. However, a number of our previous studies and those by other groups have reported no measureable influence of intransal OXT treatment on either state anxiety or positive or negative mood (see Scheele et al., 2015; Tabak et al., 2016; Xu et al., 2017).

In summary, the current study has shown in two separate experiments that OXT enhances the value placed on possessions owned by both self and others in the context of an endowment effect task. Thus OXT effects on self- and-other processing are not limited to personal attributes. The behavioral effects of OXT on the value placed on possessions also appear to involve differential actions on both the dmPFC and vmPFC and modulating their functional connections with reward and other self-processing regions. Taken together these findings provide further support for a role of OXT in reducing a self-other distinction.

WZ and KMK designed the experiments; WZ, YG and LL conducted the experiments; WZ, ZZ, XM, LX and SY analyzed the data; WZ, YG, LL, ZZ, XM and LX drafted the manuscript, WZ and KMK revised the article. All authors reviewed the manuscript and agree to be accountable for the content of the work.

This work was supported by National Natural Science Foundation of China (NSFC) grant number: 31530032.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.