Oxytocin and vasopressin flatten dominance hierarchy and enhance behavioral synchrony in part via anterior cingulate cortex

The neuropeptides oxytocin (OT) and arginine vasopressin (AVP) influence social functions in many mammals. In humans and rhesus macaques, OT delivered intranasally can promote prosocial behavior in certain contexts. Yet the precise neural mechanisms mediating these behavioral effects remain unclear. Here we show that treating a group of male macaque monkeys intranasally with aerosolized OT relaxes their spontaneous social interactions with other monkeys. OT reduces differences in social behavior between dominant and subordinate monkeys, thereby flattening the status hierarchy. OT also increases behavioral synchrony within a pair. Intranasal delivery of aerosolized AVP reproduces the effects of OT with greater efficacy. Remarkably, all behavioral effects are replicated when OT or AVP is injected focally into the anterior cingulate gyrus (ACCg), a brain area linked to empathy and other-regarding behavior. ACCg lacks OT receptors but is rich in AVP receptors, suggesting exogenous OT may shape social behavior, in part, via nonspecific binding. Notably, OT and AVP alter behaviors of both the treated monkey and his untreated partner, consistent with enhanced feedback through reciprocal social interactions. These findings bear important implications for use of OT in both basic research and as a therapy for social impairments in neurodevelopmental disorders.


Intranasal oxytocin (OT) relaxes social interaction.
Most prior studies of nonapeptides in human and nonhuman primates have focused on examining task performance in the laboratory, thus limiting ecological validity and translational potential. For this reason, we decided to probe the effects of OT and AVP on spontaneous, naturally-occurring social behaviors (though still in a laboratory setting). We recorded a series of 5-minute long videos of pairs of adult male macaque monkeys facing each other in close proximity. They were free to interact without danger of physical contact (see Fig. 1A as well as Methods, as well as Supplementary video). Prior to each session, one monkey (M1) inhaled either saline or OT via a pediatric nebulizer 77 , whereas the other monkey (M2) did not receive any treatment. Both monkeys' behaviors were rated offline by 1-3 independent observers and subsequently converted to a set of ethograms (overall concordance across observers = 0.86, see Methods).  (Fig. 1C, from left to right), the duration of his stare systematically increased (F(5) = 2.79, P = 0.049; 1-way ANOVA). Following OT inhalation, M1 spent less time staring at other monkeys (saline = 83.09 ± 8.41 s; OT = 63.78 ± 6.86 s; df = 46, P = 0.024, Wilcoxon signed rank test) (Fig. 1D), and in turn his untreated partners also spent less time staring at him (saline = 35.86 ± 7.07 s; OT = 17.74 ± 3.09 s; df = 46, P = 0.025, Wilcoxon signed rank test) (Fig. 1E).
Across all possible monkey pairs (n = 7*6 = 42 pairs), M1 staring increased with dominance, M2 staring decreased correspondingly, and the difference between the two widened, reflecting the existing social hierarchy (Fig. 2C). OT narrowed this gap and diminished the difference in staring durations, suggesting a flatter dominance hierarchy   Figure 2C).
OT enhances behavioral synchrony. Recent social neuroscientific studies have linked behavioral and neural synchrony to stronger social bonds [78][79][80][81] , and nonapeptide signaling may contribute to such social synchrony [82][83][84] . To address this possibility, we examined the temporal correlation of behaviors within pairs of monkeys following saline or OT delivery. Figure    version). The autocorrelations of M1's staring and turning away were not influenced by OT (M1 staring at time ± 1 s, saline = 0.65 ± 0.02; OT = 0.61 ± 0.02; df = 59, P = 0.214, Wilcoxon rank sum test; M1 turning away at time ± 1 s, saline = 0.74 ± 0.04; OT = 0.75 ± 0.05; df = 59, P = 0.416, Wilcoxon rank sum test; empty chair sessions). Together, these findings demonstrate that the influence of OT on a treated monkey evoke behavioral changes in an untreated monkey. More specifically, under OT treatment, when faced with M2's stare, M1 either immediately stares back (to show dominance) or quickly turns away (to show subordination), which in turn leads to reduced stare by M2. Thus under OT both monkeys become more responsive to each other and more efficient in communicating dominance status, perhaps through an increase in attention to social cues 72 .
Intranasal arginine vasopressin (AVP) reproduced the effects of OT with greater efficacy. In a new colony of 7 monkeys (M1 = 3 males, M2 = 3 males, 4 females), we systematically compared the effects of OT and AVP inhalation at the same concentration (25 IU delivered in 1 ml saline vehicle). We reconfirmed in this group that OT inhalation significantly reduced staring by the treated monkey, and found that AVP inhalation had a similar effect (saline = 42.48

Discussion
Oxytocin shapes social behavior in a variety of animals, from rodents like mice [85][86][87] and prairie voles 5,7,8 , to nonhuman primates like marmosets 9 , squirrel monkeys 88 , titi monkeys 89 , rhesus macaques 49,77,90,91 , and chimpanzees 92 . The overwhelming evidence supporting a functional role for OT in animal social behavior has greatly SCieNTifiC REPORTS | (2018) 8:8201 | DOI:10.1038/s41598-018-25607-1 motivated the surge of interest in OT function in humans 87,93 . The effects of acute OT administration in humans, however, remain highly controversial. Though early studies reported OT enhances prosocial behavior and improves social cognition, recent meta-analyses suggest that almost half of studies failed to find a significant effect of OT on human social behavior [13][14][15][16] . By contrast, using a dose (25 IU) comparable to those used in most human studies 11-14 , we found OT consistently and robustly relaxed social interactions, flattened the social hierarchy, and enhanced social communication in rhesus macaque monkeys.
Our results add to the long line of positive findings in the effectiveness of OT in rodents and nonhuman primates, but they also raise questions regarding the potential source of variability in OT research in humans. Firstly, OT often has task or stimulus specific effects in humans [94][95][96][97][98] , making it difficult to compare across studies 99,100 . OT also interacts with individual traits such as gender, personality, attachment styles, and psychopathology 34,[40][41][42]101 .
Here we focused solely on spontaneous interactions of adult male macaque monkeys living within a stable hierarchical social group, which provided us with a level of control that is rarely seen in human social studies, and may have contributed to the detectability of the effects of OT, as well as AVP.
Another source of variability in human studies is the method of delivery. In laboratory or clinical setting OT is typically applied via a nasal spray 102,103 . Intranasal OT spray in humans increases OT levels in the cerebrospinal fluid (CSF 104 ) and, similarly, intranasal spray of AVP increases AVP in CSF 105 . In primates, however, compared with nasal spray, aerosolized OT delivered via nebulizer (as in our study) either more consistently 106 or more effectively 107 increases OT in CSF. Moreover, most studies in humans rely on participants to self-administer OT spray 102,103,108,109 . Lack of standardization in self-administration may have led to variation in the effectiveness of OT delivery in humans. By contrast, aerosolized delivery through nebulizer eliminates the need for user  involvement, thus ensuring more consistent and thorough delivery. This may be critical for potential therapeutic use of OT in young children or psychiatric patients who have difficulty following instructions.
One other source of potential variability is the amount of OT used. Most human studies delivered 20-40 IUs of OT regardless of the subjects' age, gender, or body weight [11][12][13][14] 110 ). The average weight of a rhesus macaque monkey, however, is much lower than that of a human subject, leading to the possibility of a higher concentration of CSF OT in macaques when the same amount of exogenous OT is delivered. To facilitate the comparison of OT effects across future studies and especially across species, OT dosage may need to be calculated on an IU per kilogram base 111 .
Comparison of our behavioral and injection results lends support to the hypothesis that inhaled OT influences social behavior at least partially through nonspecific binding with AVP receptors in medial frontal cortex. Because of similarity in molecular structure, OT can bind to AVP receptors with high affinity [58][59][60] . Indeed, in rodents OT and AVP can act on the 'opposite' receptor to influence behavior [112][113][114] . In nonhuman primates, the distribution of post-synaptic OT receptors (OXTR) is extremely sparse in frontal cortex 76 . By contrast, AVP receptors (V1a) are much more abundant in cortical areas implicated in social behavior, including prefrontal, cingulate, and entorhinal cortex, as well as in amygdala, hypothalamus, and brainstem 115,116 . Our study offers one of the first pieces of evidence strongly implicating the AVP system in what was traditionally considered to be oxytocinergic effects on primate social behavior. These results highlight the need to further investigate the interaction between AVP and OT systems through selective agonist or antagonist administration. For example, if targeted ACCg injections of AVP antagonist could partially abolish the behavioral effects of intranasally delivered OT, we will be able to rule out the possibility that local injections of OT impact behavior via different pathways than OT inhalation, and conclude with greater confidence that inhaled OT affects behavior through cortical AVP receptors. More generally, our results also emphasize the importance of further examination on the AVP system independent of OT, as we observed in many aspects AVP is more effective than OT in positively modulating social behaviors. Indeed, an ongoing clinical trial (Hoffmann-La Roche) has demonstrated the possibility of tackling the AVP system alone to rescue social deficients in autism disorder.
In our study, focal injections of both OT and AVP into ACCg recapitulated most of the behavioral effects of intranasal delivery. ACC has long been linked to social behavior in mammals [117][118][119] . ACCg is a specialized sub-region within ACC that is interconnected with various other brain areas in the 'social brain network' , including the temporoparietal junction (TPJ), dorsolateral prefrontal cortex (dmPFC), ventromedial prefrontal cortex (vmPFC), and amygdala [120][121][122][123] . Resting-state connectivity of ACCg with these areas increases with the size of an individual's social network 124,125 , and abnormal cytoarchitecture and functional connectivity of ACCg are prevalent in autism spectrum disorder (ASD 126,127 ). Both single-unit recordings and neuroimaging studies suggest ACCg processes 'other-oriented" information including the decisions made by others and the rewards and punishments they receive [68][69][70][71][72][73][74][75] . Our findings further endorse the idea that ACCg plays a fundamental role in natural social interactions 128 .
Finally, we discovered that altering the behavior of a monkey with either OT or AVP has a contagious effect on his untreated social partners, and thus the social environment in general. This significant modulation of others' behavior is mediated through not only changes in behavioral frequency or duration, but also improved temporal synchrony across subjects. Thus, the prosocial effects of OT and AVP can be interpreted as an improvement in the efficiency of bidirectional social communication. These results directly support previous reports of OT improving behavioral synchrony 83,84 and mutual gaze [129][130][131] in neurotypical individuals as well as ASD patients, and thus bear important implications for use of social peptides in both basic research and as a therapy for social impairments in neurodevelopmental disorders.

Methods
Animals. Data  A video camera (Logitech, 60 fps) was positioned on the right side of M1/ the left side of M2, and simultaneously recorded both monkeys' behaviors into an MP4 file. On each day, one M1 faced six different M2s sequentially. In addition, the same M1 also faced an empty chair for 5 minutes. The order in which M1 faced the other monkeys together with the empty chair was determined randomly each day. Pharmacological Manipulation. The procedure for intranasal OT delivery in macaque monkeys has been described in detail previously 72,86,87 . Briefly, monkeys were trained to accept a pediatric nebulizer mask (Pari Labs) over the nose and mouth. Through the nebulizer 1 ml of OT solution (25 IU/ml in saline; Agrilabs/Sigma Aldrich) or saline was delivered at a constant rate (0.2 ml/min) over a total of 5 minutes. Behavioral testing began 30 minutes after intranasal delivery and continued for 1-2 hours. This OT dosage and timing protocol was similar to that typically used in humans [11][12][13][14] and other non-human primate studies 72,87,103,104 . We followed the same procedure for intranasal AVP delivery (25 IU in 1 ml saline; Sigma Aldrich). The same amount of neuropeptide (25 IU) was delivered to all monkeys regardless of their weights (ranging from 10-16 kg at the time of the experiment), which resulted in a dosage of ~1.5-2.5 IU/kg. Neuropeptide and saline treatments were delivered on alternating days, with each monkey receiving no more than 5 treatments per week. The same monkey never received OT and AVP treatment within the same week. The order of treatments was counterbalanced across monkeys as well as within monkeys between weeks to mitigate any possible order effects. Furthermore, to rule out the possibility that the particular sequence with which different monkey pairs were tested (after saline or neuropeptide delivery) had any impact on behavior, 1-way ANOVAs were performed on different behaviors. The testing order here was treated as a nominal variable with 6 levels (with 1 corresponding to the first pair tested and 6 corresponding to the last pair tested on the same day). This analysis did not reveal any significant order effect (for example, for staring duration under saline-OT-AVP conditions, 1-way ANOVA, F(5,264) = 1.03, P = 0.400). In addition, general linear models were constructed for different behaviors as well, with the testing order being treated as a continuous variable and measured as the estimated time passed from drug administration to behavioral testing for each pair. This analysis revealed no significant order effect either.
The procedure for OT injection has also been described previously 49 . For the three monkeys receiving OT injections, the location of their ACCg was first identified by a series of structural magnetic resonance images (MRI; 3 T, 1-mm slices) and then confirmed with a set of preliminary electrophysiological recordings. On each experiment day, 2 μL of OT solution (0.5 IU/μL of OT in saline; Agrilabs/Sigma Aldrich) or saline was injected into the ACCg via a Hamilton microsyringe (Hamilton). To minimize tissue damage, all injections were delivered at a rate of 0.2 μL/ min and restricted unilaterally in each animal (C, O: right; D, S: left). The procedure for AVP injection (1 IU in 2 μL saline; Sigma Aldrich) was the same as OT injection. Neuropeptide and saline injections were delivered on alternating days, with each monkey receiving no more than 1 injection every other day. The same monkey never received OT and AVP injections within the same week. The order of treatments was counterbalanced across monkeys as well as within monkeys between weeks to mitigate any possible order effects (for example, for staring duration under saline-OT-AVP conditions,1-way ANOVA, F(5,174) = 1.68, P = 0.142). For the same monkey, inhalation and injection treatments were always delivered in distinct time periods (at least 2 weeks apart). Data Analysis. Each behavioral video was rated offline by 1 to 3 independent viewers, all of whom were blind to treatment conditions. Viewers used a Python GTK based, custom GUI to play and pause the video, adjust its speed, and code monkey behaviors by pressing a set of keys on the keyboard. The string of keyboard inputs and their corresponding time stamps were imported into MATLAB (Mathworks) and converted into a pair of behavioral ethograms via custom MATLAB scripts. When more than one viewer rated the same video (~50% of all the videos), their ratings were averaged to generate the ethograms. For the same videos rating consistency was very high across different viewers (for fixation duration: r = 0.58, df = 184, P = 0.000; for number of fixations: r = 0.24, df = 202, P = 0.001; for turning away duration: r = 0.65, df = 184, P = 0.000; for number of turning aways: r = 0.41, df = 184, P = 0.000; for number of yawns: r = 0.80, df = 184, P = 0.000; for number of threats: r = 0.91, df = 184, P = 0.000). The overall concordance across observers was 0.86 (0.88 between DX and YJ, 0.79 between DX and JF, 0.91 between YJ and EZ, 0.82 between YJ and SRM). All subsequent data analyses were accomplished with custom MATLAB scripts.
All statistical tests were two-tailed. For hypothesis testing between two samples, a non-parametric Wilcoxon signed rank test (for paired samples) or Wilcoxon rank sum test (for un-paired samples) was used. For comparison among more than two samples, an ANOVA was used together with multiple comparisons (Tukey's HSD test) when appropriate. Correlation coefficients were estimated with Pearson's r. All behavior histograms were fitted with Gamma distributions: where Γ is the Gamma function (Γ = − n n ( ) ( 1)!), a is a shape parameter, and b is a scale parameter. Cross correlations and auto-correlations were calculated in non-overlapping 100 ms windows.
Data Availability. The datasets generated during and/or analyzed during the current study are available from the corresponding author upon request.