Socially related behaviors other than sexual behavior include social bonding and maternal behavior, each of which comprise recognition, memory, seeking close proximity to conspecifics, and other behavioral components. There are some well-described functions for both oxytocin and dopamine in the control of these social parameters. The fundamental evidence for oxytocin's role is that oxytocin- or oxytocin receptor-null mice exhibit deficits in social recognition and social memory [25,163–165] indicating the necessity of oxytocin in facilitating interaction between individuals. Oxytocin in the brain also has antistress and antianxiety roles , further showing its pleiotropic nature in handling social situations. Similarly, disruption of dopamine signalling in transgenic mice also leads to social abnormalities [167,168], e.g., dopamine transporter null mice, that have increased extracellular dopamine and decreased dopamine receptor expression, exhibit increased social reactivity. We will analyze the reciprocal role that oxytocin and dopamine play in such behaviors in rodents and humans and then discuss whether disorders of social behavior, such as autism, can be explained by oxytocin and/or dopamine dysfunction.
Social bonding includes elements such as time spent together in close proximity, choice of the known individual over another unknown individual (social preference), and social recognition. Social recognition is experimentally tested in the lab by measuring the time spent investigating a known individual compared to an unknown one and relies on olfactory and visual cues. Olfactory learning is a major component of recognition, involving noradrenaline-mediated disinhibition of mitral cells . However, olfactory and visual cues also both use oxytocin [170,171]. Unlike wild-type mice oxytocin null mice cannot remember conspecific mice that they have recently been exposed to. Oxytocin receptor is expressed in the olfactory bulb, and although the PVN is considered a major source of oxytocin mediating social behaviors, oxytocinergic fibers have not been observed to appose either mitral or glomerular cells so the origin of oxytocin must be from other release sites, e.g., magnocellular dendrites. Other centers for oxytocin action in social recognition include the lateral septum, medial amygdala and MPOA, which also express oxytocin receptor. Since oxytocin can act via noradrenaline or other classical neurotransmitters (glutamate, GABA) in some areas, it is thought to modulate olfactory learning rather than directly stimulate the olfactory substrate , so as dopamine also plays a role in the olfactory bulb in social recognition oxytocin may also modulate dopamine release in order to mediate its effects.
Dopamine modulates odour detection and discrimination, which is part of olfactory memory formation [170,172,173], although others argue that dopamine (D2) plays a more prominent role in consolidation of memory rather than recognition per se. So both oxytocin and dopamine play a role within the olfactory bulb in social recognition/memory. The Bruce Effect is an example of the consequences of lack of social recognition; in mated female mice, implantation failure occurs in response to pheromones from a stranger male; both oxytocin and dopamine action in the olfactory bulb are implicated in this phenomenon . However, due to lack of evidence for their interaction within the olfactory bulb, attempts to fit our proposed interaction framework to this behavior is difficult. The precise phenotypes of neurones expressing dopamine or oxytocin receptors are not reported and the source of dopamine or oxytocin in the olfactory bulbs is unclear, while pharmacological approaches have not yet targeted the olfactory bulb selectively. The evidence suggests that there is a local synergistic, if not interactive, effect of oxytocin and dopamine within the olfactory bulbs, but there may also be interaction upstream, in other dopamine or oxytocin source or target sites.
Associated with social recognition, bonding develops under certain circumstances. Animals can develop a close association that is measurable by the social preference test. Bonding behavior between adults has been closely studied over the past decade using prairie voles as a model species since they develop a well-defined partner preference. When male and female prairie voles spend time together partner preference increases. However, after copulation partner preference is strongly elicited, and the paired voles attain a monogamous-like affiliation. There is even an associated induction of parental behavior in male prairie voles which will be discussed later. This remarkable postcopulatory bonding between adults is attributed in part to both dopamine and oxytocin in both females and males , and the NA has been identified as the main forebrain center for the effect. As discussed below, this phenomenon represents another example of dopamine–oxytocin interaction in a behavioral context.
Importantly, the partner preference induction is dependent upon oxytocin receptor distribution in the brain. The NA and the caudate putamen have high densities of oxytocin receptors in animals where partner preference is induced but not in other species, such as the montane or meadow vole or mouse . Oxytocin release increases in the NA in female voles during pairing , although the source of oxytocin is unclear. Increased social contact is inducible by intra-cerebral oxytocin infusion and can be induced not only in voles but also in rats and squirrel monkeys; and partner preference is blocked by oxytocin antagonist infusion into the NA. After mating oxytocin release also increases in the PVN and extrahypothalamic regions such as the amygdala in male and female rats  (and see above), so conceivably these regions could also play a role in partner bonding. Interestingly mating-induced oxytocin release in rats has been shown to underlie postcopulation anxiolysis so if also true for other rodents this may be a component of the bonding behavior. Whether this occurs in the prairie vole model is unknown and pair bonding in rats is not typically induced by copulation.
The NA is also an important center for dopamine action during pair bond formation in prairie voles . Mating increases dopamine turnover in the NA in males and females- NA neurones are known to express dopamine receptors and dopamine acts via D2 receptors to facilitate partner preference. The phenotype of dopamine target neurones is not described but dopamine presumably acts by regulating the predominant NA GABA neurones, although some evidence indicates that it also modulates incoming glutamatergic afferents . As for oxytocin, dopamine D2 agonists facilitate and antagonists attenuate partner preference behavior.
There is a reasonable body of evidence revealing an oxytocin–dopamine interaction in partner preference. The NA is rich in both oxytocin and dopamine receptors; if coexpressed pre- or postsynaptically it would provide a clear basis for coregulation of the same target neurones, but this is unknown. As well as acting in the NA it appears that mating-dependent release of oxytocin activates a mesolimbic (VTA) dopamine circuit and induces dopamine release in the NA, indicating upstream regulation of a dopamine circuit. Therefore, in association with copulation oxytocin and dopamine link the state of sexual arousal with that of bonding . However, it has not been shown that oxytocin release increases in the VTA where it might stimulate dopaminergic projections at copulation or with pair bonding, although, neuroanatomical evidence has proved indicative (Melis et al., 2007. On the other hand, recent reports propose dopamine acts via oxytocin since concurrent reciprocal activation of dopamine and oxytocin receptors in the NA occurs [89,176]. Pharmacological studies are needed to tease out the precise roles of oxytocin and dopamine in the NA and VTA. One might also ask whether dopamine action on PVN (or SON) oxytocin neurones is necessary for induction of partner preference. Evidence in the prairie vole is lacking but we would speculate that it is part of the complex cascade leading to arousal and social interaction that precedes copulation and bonding.
It should, however, be noted that another hypothalamic peptide related to oxytocin, vasopressin, has also important interactions with dopamine in the NA in social recognition and the formation of pair bonding . There is substantial evidence of prominent vasopressin V1a receptor expression in the NA and ventral pallidum that facilitates the formation of pair bonds and this dependent upon a specific sequence in the 5’ flanking region of the gene that determines its brain expression pattern. Vasopressin neurones lie adjacent to oxytocin neurones in the PVN and SON, and also express D2-like receptors , so dopamine may control vasopressin release and action in parallel with oxytocin. Whether vasopressin also controls dopamine neurone activity or release (as oxytocin does, see above) is unclear at present. Interestingly though, viral vector insertion of the prairie vole V1A receptor into the mouse brain induces susceptibility to the formation of pair bonds similar to that seen in prairie voles .
So, it appears that robust oxytocin–dopamine and/or dopamine–oxytocin connections are involved in social interactions. In mating it may be that dopamine–oxytocin connections are initially involved in penile erection , (see above; Figure 3) but then oxytocin-driven dopamine effects mediate subsequent related behaviors such as bonding and reward. However, it is apparent that their interaction at multiple brain locations matches with our framework, and occurs in a social context with more than one potential outcome: copulation and bonding.
Studies of the roles of oxytocin and dopamine in social behaviors in humans are appearing more now with the application of functional magnetic resonance imaging (fMRI) and gene expression studies on post mortem tissue. For example, it can be observed that high intensity fMRI signals are observed in the VTA and SN upon viewing a loved partner , and these correlate with the distribution of human oxytocin receptor expression and an interaction with the dopamine system . Other strategies include analysis of peripheral oxytocin concentration [98,181]. Although these do not necessarily correlate with the release or action of brain oxytocin, some importance is attached to increasing levels that correlate to positive mood and prosocial behaviors and, along with studies investigating the effects of intra-nasal administration of oxytocin, have gained some credibility in recent years .
Lack of social recognition and inability to form social bonds are characteristics of a host of psychiatric disturbances, the most profound perhaps being autism. Autism and autism-spectrum-disorders represent a profound disturbance in the ability to form social bonds in humans and are most commonly found in males. Although underlying causes are multiple, and include a variety of potential genetic mechanisms, naturally associations between oxytocin and dopamine in autism have been sought to explain the condition.
The DSM-IV definition of autism is the inability to socially interact due to a deficit in verbal and nonverbal communication, social awareness and interactions and imaginative play. Autism-spectrum-disorders are complex neurodevelopmental disorders characterized by social withdrawal  and no adequate animal models have yet been reported. A role for oxytocin has been implied since plasma oxytocin concentrations (which can be a marker for social behaviors in humans) [98,181] are low in autistic boys [184,185] and oxytocin infusions or intranasal administration improve emotion recognition and facilitate trust analyzed using fMRI studies in humans, particularly revealing the amygdala as a target [186–188]. Additionally there is a proposed link between oxytocin receptor polymorphisms and autism in some families [182,189–191], which might adversely alter expression patterns and densities in a way that contributes to the altered social behavior. However, like bonding, autism is also associated with polymorphisms in the vasopressin receptor 1A gene, particularly in the amygdale . Overall the evidence for oxytocin is contradictory, as discussed in more detail in an excellent review by Hammock and Young (2008) . For example, in one study autistic children with circulating oxytocin levels in the normal range were the most profoundly asocial, so peripheral hormone levels may not reflect central oxytocin release or action in this condition. In contrast to oxytocin, stronger evidence links dopamine dysfunction with autism-like disorders. Reports are mixed but variations in the dopamine system such as dopamine transporter and D4 receptor genes and activity are implicated. Attention deficit/hyperactivity disorder, one of the range of autism-spectrum-disorders, also appears to exhibit similar dysfunction of dopamine systems [193–197]. Whether there are parallel changes in oxytocin release and dopamine activity in relevant brain regions, such as the NA, VTA or ventral pallidum in people with autism-related disorders is uninvestigated, but a joint role for oxytocin and dopamine seems possible.
Oxytocin and dopamine are also key neuromodulators of maternal and paternal behavior. It has been recognized for many decades that oxytocin in the brain plays a key role in maternal behaviors , and is mainly facilitatory rather than essential, unlike for other social behaviors as discussed above [163,165,202]. More recently, reports of its role in paternal behaviors are emerging, including in prairie voles where mating not only induces social bonding but also paternal care behaviors in males [203–206]. Similarities between mechanisms in mothers and fathers include enhanced oxytocin expression in the PVN in paternal compared to nonpaternal males and increased vasopressin V1a receptors in the prefrontal cortex in the male marmoset . However, maternal behavior is much better investigated and understood, especially the role of oxytocin in sheep and rodents, so this section will concentrate on evidence from females.
Initially oxytocin was considered to be important for onset of maternal care perinatally but is now recognized to also maintain the suite of behaviors involved. Thus, central oxytocin facilitates care behaviors (pup licking, grooming and nesting behavior), delivery of nutrition to offspring (particularly the milk provision via the ejection reflex for which oxytocin is essential , and arched back nursing [kyphosis]), offspring protection (against predators but also including maternal aggression against conspecifics) and, importantly, bonding [26,201,208,209]. Some of these social-like behaviors are comparable to the adult-adult bonding as discussed above, and involve recognition mechanisms.
Mother-infant recognition may be similar to adult-adult recognition in that oxytocin mediates olfactory memory for offspring. This is very well investigated in sheep, where olfactory memory is a crucial, postbirth link between ewe and lamb and is necessary before any further behaviors, including allowing lamb suckling, are performed [26,170]. Such olfactory memory is part of mother-infant bonding in sheep that occurs at birth and can be induced by vagino-cervical stimulation under appropriate steroid conditions. Bonding between ewe and lamb does not occur in the absence of vagino-cervical stimulation, and even in women it has been reported that after caesarean birth, bonding with the baby takes significantly longer than after vaginal delivery . Bonding combined with olfactory memory of the young initiates the other maternal caring and nurturing behaviors . Initially, birth and vagino-cervical stimulation increase oxytocin release in various brain regions, including the PVN, SON, MPOA, SN, septum and olfactory bulb in both rats and sheep [26,170], which are known to be key regions mediating maternal interaction from a variety of activity and lesioning studies. Mimicking this endogenous release pattern with central administration of oxytocin induces, and oxytocin antagonist inhibits, maternal behaviors in rodents and sheep. However, as for prairie vole bonding, oxytocin receptor distribution is most important in quality of behavior performed. Not only does oxytocin receptor expression increase perinatally and into lactation , but ‘good’ maternal behavior, commonly defined as high licking, grooming and arched back nursing, correlates with wider and higher density of distribution of oxytocin receptor in a rat model . Furthermore, such good behavior can be epi-genetically inherited by daughters, as care received neonatally determines their brain oxytocin receptor distribution and maternal behavior quality in adulthood . Therefore receptor patterns rather than strictly quantity of neuropeptide release may be particularly important in determining quality across the range of social behaviors. Maternal experience is important too: stress exposure perinatally alters oxytocin receptor expression patterns correlated with poorer behavior , and this may also extend to stress emanating from prolonged conditions relating to a difficult (e.g., psychological disturbances) or ill child. The best evidence links oxytocin to maternal behavior in women in a correlative way, e.g., fMRI imaging of the maternal brain while viewing photos of their baby reveals activation of oxytocin target regions (e.g., SN) , and increased CSF or plasma oxytocin levels [26,215] are evident compared to women who are not mothers. Further indication of a role for oxytocin in primates comes from central administration of oxytocin to rhesus monkeys which increases maternal behaviors , and oxytocin is widely believed to be evolutionarily important in maternal care across a range of species.
Proof of the importance of dopamine activity in mothers’ brains was shown in dopamine transporter knockout mice which exhibited impaired maternal behavior [217,218]. Like oxytocin, dopamine release is also elevated in the PVN, SON, MPOA, SN, septum and olfactory bulb in rats and sheep [26,170]. Unlike oxytocin, where receptor distribution is primarily important, levels of dopamine in the NA equate with quality of maternal behavior . Lesioning the VTA, using 6-hydroxydopamine to selectively destroy monoamine cells, also blocks maternal behavior  indicating that as a potentially important source of dopamine in mothers’ performance. Evidence indicates that extensive interaction between oxytocin and dopamine plays a role.
Oxytocin modifies targets directly and/or presynaptically, at least partly via monoamines, including dopamine, and GABA , so could potentially gate dopamine effects. At birth oxytocin receptor increases in many classical dopamine target regions, including NA, olfactory bulb, and prefrontal cortex, increasing the potential for oxytocin regulation of the dopamine release at onset of maternal behavior . However, oxytocin receptor expression evidently decreases in the SN, showing that control of the nigrostriatal dopamine neurones and their emanating pathway(s) is differentially altered from other dopamine sources. Oxytocin released after birth or vagino-cervical stimulation modulates dopamine activity in the NA, VTA and SN, and dopamine reciprocates with action within the PVN . It has been proposed that oxytocin action on dopamine neurone activity in the SN promotes immobility to facilitate offspring suckling, but the coregulation is also associated with behavioral drive as well as feelings of reward associated with the neonate . How this equates with decreased oxytocin receptor in the SN is unknown, but may indicate a shift in oxytocin control from one mediator to another within the SN. Interestingly, mothers showing better maternal behavior (high licking and grooming) have increased oxytocinergic projections to the VTA and more dopamine release within the NA, powerfully revealing the importance of the oxytocin–dopamine interaction located in the mesolimbic pathway . So there are similarities between maternal bonding and adult bonding in the neurochemical oxytocin–dopamine interaction but corresponding experiments have not all been performed to clearly compare them or to match to our framework. Interestingly, recognition of the mother by the offspring (rat pups) also involves monoamines , and others have speculated that oxytocin action in the offspring brain might parallel that of the mother .
Disorders of parental behavior are not yet well described, let alone understood physiologically, but have profound consequences on offspring. One topical and emerging field is that of offspring adverse programming arising from parental and/or postnatal stress. Adverse neonatal programming can also be produced by poor parental care, whether or not stress is manifest. Such adverse long term consequences on offspring psychological and physical development arising from parental stress or poor parental care in the neonatal period include susceptibility to anxiety and depression disorders, obesity, and cardiovascular disease . In fact, perinatal stress and level of maternal attachment also has an impact on dopamine systems in the offspring, detrimentally altering its release and activity parameters and reproductive development [223,224].
So, abnormal or inappropriate parental behavior in humans has far-reaching consequences for children and, for offspring, perinatal experience received has a huge impact on their later development. As indicated earlier, there is also a reciprocal effect, where disruptive or socially compromised children inflict stress and long-term consequences on the parents and their behaviors. This is only just being recognized as a problem for families, and will have costs for the health and social services in the future. Understanding of the causes of poor parental behavior is only just emerging, but includes parents receiving poor parental care when they were young and peri-natal stress exposure, setting up a vicious cycle that is difficult to investigate in humans and challenging to treat. However, some progress is being made to enlighten the underlying maternal brain dysfunction, and recent evidence indicates that mothers’ low responsiveness to their toddler correlates with less efficient oxytocin receptor gene variants . Breaking the repeated cycle of transgenerational poor parental care would have long lasting desirable emotional and economical benefits for individuals, families and governments. Simply targeting the dopamine/oxytocin systems is not a viable option. Since problems occur at multiple cellular, organismal and social levels, addressing each parameter on its own is not likely to impact on the problem in any meaningful way, and a multidimensional approach will be required.
A dysregulation of oxytocin and/or dopamine activity coupled with an impairment of social interactions has also been observed in a variety of diseases and disorders ranging from anorexia to Parkinson's disease . Furthermore there has been an expansion in the research of an oxytocin or dopamine basis (and those neural circuitries upstream/downstream) in multiple behavioral syndromes. This review does not aim to cover the all diseases and disorders exhibiting inherent abnormalities in behavior but will now consider the role of these neuromodulators in selected behavioral disorders such as drug addiction and anorexia.
The DSM-IV recently classified drug addiction as an individual who persists in the using of alcohol or other drugs despite problems related to use of the substance, resulting in a significant impairment in functioning. Due to the devastating impact on addicts’ lives and the socioeconomic burden associated with addiction, the neural basis of drug use disorders is being increasingly explored. Hurdles such as the heterogeneous nature of biological and genetic determinants add to the complexities of understanding drug seeking behavior. Additionally, drug addiction models used in experimental studies do not consider social environment and so do not closely mirror drug-seeking behavior in humans. Some elegant preclinical studies however, have revealed interesting information regarding social consequences of drug use [226,227] and those neural correlates, such as oxytocin, subserving addiction [228,229].
The role of the mesolimbic dopamine system in the rewarding effects of drugs of abuse has been well documented [230–233], with MDMA, cocaine and opiates known to influence dopaminergic neurotransmission in these motivational pathways [229,234] and produce marked impairment in prosocial behavior [235,236]. The mesolimbic system has been established as a key component of rewarding properties of natural rewards such as sex and food and maladaptive rewards such as drugs of abuse and is believed to drive chronic drug use. As mentioned previously, dopaminergic projections originating in the VTA and extending to the NA, amygdala, olfactory tubercle and prefrontal cortex (see basic framework diagram in Figure 2, not all nuclei shown) comprise the mesolimbic pathway. In humans, fMRI studies revealed distinct brain activation in the VTA during cocaine self-administration in cocaine addicts which is indicative of mesolimbic dopamine activation . Animal studies using in vivo microdialysis, where small changes in neurotransmitter concentrations can be detected, have shown that psychostimulants such as alcohol and cocaine activate mesolimbic reward circuits and increase dopamine levels in the NA [238,239]. Specific dopamine receptors mediating drug reinforcement pathways have yet to be elucidated; however, dopamine D1-like and D2-like receptors appear to be implicated in regulating the acute reinforcing properties of cocaine as evidenced from rodent studies where dopamine agonists and antagonists acting on all dopamine receptors reinstated or inhibited drug-seeking behavior, respectively [240,241]. Furthermore, disruption to the mesolimbic dopamine via selective ablation of the NA results in decreased self-administration of cocaine without affecting feeding behavior in rats . Taken together, these findings emphasize the importance of the mesolimbic dopamine reward pathway in the processing of maladaptive rewards in particular, which drives continued drug administration leading to chronic drug abuse.
Because drug addiction has such a profound effect on a wide spectrum of social behaviors including social bonding, maternal and sexual behavior, it is not surprising that oxytocin has also become a key area of focus in addiction research. A role for oxytocin in drug addiction is not altogether surprising since oxytocin fibers innervate and receptors are expressed in known dopamine-containing nuclei important in reward assessment including the VTA and amygdala [79,80,82,243].
Drugs of abuse known to target the mesolimbic system, such as cocaine markedly reduce the levels of oxytocin in the hippocampus, hypothalamus, NA and plasma when taken repeatedly over time . These long-term disruptions to central oxytocin systems due to chronic abuse of psychostimulants, presumably contributes to the impaired social and emotional capabilities often observed in drug addicts. Whilst it is apparent that dopamine can influence oxytocin release in a chronic drug use context, these two neuromodulators may also function in a bi-directional manner during the development of tolerance and dependence to drugs of abuse.
It has been suggested that oxytocin may possess antipsychotic properties due to its ability to block cocaine-induced dopamine release in the NA  and to attenuate characteristic locomotor activities associated with cocaine addiction [229,246–248]. Furthermore, microinjection of physiological doses of oxytocin into the NA, amygdala and the hippocampus attenuate morphine tolerance and dependence and cocaine-induced hyperactive locomotor activity . So, the data suggest a potential role for oxytocin in the regulation of chronic drug abuse by influencing dopaminergic activity in key limbic brain sites and altering behavioral responses associated with addiction.
In addition to having a neuromodulatory role influencing neuroadaptive processes responsible for tolerance and dependence , recent findings have also alluded to a role for oxytocin in drug withdrawal [234,250]. Opiate drugs strongly inhibit oxytocin neurones since they coexpress opioid receptors. Endogenous opioid systems control oxytocin during physiological conditions, e.g., pregnancy and birth . However, oxytocin neurones develop tolerance and dependence for opiates, which means that they are able to maintain their physiological roles even under pathologic conditions which may include drug addiction . On the flip side, it also means that upon withdrawal and similar to dopamine systems [252,253], the whole oxytocin system is dysfunctional (temporarily or could be permanently) and this has long term physiological and social implications for those trying to withdraw and avoid drug taking situations or triggers.
The data suggest oxytocin action in key limbic brain regions has a regulatory role in attenuating drug tolerance and dependence and promoting drug withdrawal, presumably via its actions on mesolimbic dopamine reward pathways. So, referring to our original framework (Figure 2), there appears to be a growing body of evidence to suggest dopamine and oxytocin pathways may be two potential neural correlates mediating drug addiction. Central oxytocin sites are one area of addiction neurobiology that has yet to be fully explored and may serve as a potential neural substrate that could be potentially exploited for pharmacotherapeutic benefit in the treatment of drug use disorders and withdrawal.
The DSM-IV classification for eating disorders such as anorexia nervosa includes the refusal to maintain normal body weight, having an intense fear of gaining weight and a disturbance in the perception of body weight or shape. Anorexia nervosa and bulimia nervosa are complex psychological disorders associated with dysfunctional control of appetite for food and body weight. Anorexia involves fear-like and obsession-like behaviors, and affects about 0.3% of the population, mostly teenage girls. It is a DSM-IV classified mental disorder with serious comorbid consequences, including clinical depression, drug abuse and obsessive-compulsive disorder, as well as other related conditions like lack of reproductive ability. Bulimia is similar, though more common than anorexia, but also involves binge eating  and sufferers are particularly susceptible to addictive behaviors. Control of appetite is multidimensional, involving many interacting hypothalamic neuropeptides, peripheral energy signaling peptides and other neurotransmitters mediating hunger, satiety and reward, and evidently several of these elements are disturbed in anorexia and bulimia. The arcuate nucleus is a major integrating center for the control of appetite receiving input from higher centers and the gut as well as being able to sense circulating factors due to its leaky blood brain barrier. Major arcuate neuropeptides primarily include the orexigenic factors neuropeptide Y (NPY) and agouti-related protein (AgRP), and anorexigenic peptides such as alpha melanocyte stimulating hormone (αMSH), and cocaine-and-amphetamine-related-transcript (CART). The arcuate communicates with the PVN and lateral hypothalamus in particular to control appetite, but also with other nuclei such as the SON, ventromedial nucleus and brainstem.
In fact the SON and PVN are emerging as regions that respond particularly to eating, neurones becoming activated strongly at the start of a meal . Oxytocin is one of the nonarcuate anorexigenic peptides involved in signaling satiety, and so administration of oxytocin intracerebrally inhibits appetite, and oxytocin antagonist prolongs meal duration. Oxytocin is thought to be recruited downstream of the major appetite peptides from the arcuate nucleus . For example, αMSH stimulates SON oxytocin neurones via melanocortin receptor 4 (MCR4). Interestingly, MCR4 agonists developed by the pharmaceutical industry to treat obesity also precipitated side effects such as prolonged penile erection in clinical trials, and we now know that is due to activation of central oxytocin systems , so there is a potential neuropeptide link between eating and sex. As well as the arcuate-oxytocin drive there is a bi-directional link between the gut and oxytocin neurones, where eating (gastric distension) stimulates oxytocin activity in the SON and PVN  and oxytocin projections from the PVN to the brainstem regulate gastric reflexes . However, apart from the brainstem, how and where oxytocin acts to mediate satiety is unclear.
A role for oxytocin in anorexia or bulimia is not well investigated. One might expect that if oxytocin dysregulation is involved in the low appetite element of anorexia that its expression and or activity would be enhanced, but the evidence available tends to suggest that oxytocin is not an underlying cause . Another anorexigenic peptide, nesfatin-1, is associated with anorexia. Intriguingly, nesfatin-1 is coexpressed in PVN oxytocin neurones  and stimulates intra-PVN oxytocin release. Since oxytocin mediates nesfatin-induced inhibition of appetite , this perhaps indicates a previously unrecognized role of oxytocin in anorexia, though further detailed studies are required. Evidence also seems to point towards mis-control of the orexigenic peptide AgRP in anorexia , and this might act at least partly via oxytocin too- AgRP acts as a MCR4 antagonist and inhibits oxytocin neurone responses to ingestive behavior . On the other hand, lack of oxytocin may underlie decreased satiety in the Prader Willi syndrome , which is a rare genetic condition characterized by hyperphagia, obesity, reproduction failure and mental retardation. Clearly further research is needed to clarify a role for oxytocin in eating disorders.
Dopamine's function in signalling both food-seeking behavior and reward plays an important role in appetite and satiety [266,276]. First, dopamine deficient mice become completely aphagic and need dopamine replacement for any feeding to occur . Secondly, mesocortical dopamine (from the VTA acting in the striatum and prefrontal cortex), but not mesolimbic dopamine (from VTA projecting to NA) pathways have a propensity to mediate food reward and anticipation of food [242,278]. Dopamine activity responds to olfactory and visual food-related stimuli and so in parallel with social recognition and sexual arousal responds to valid external cues. Dopamine is reported to be recruited by many appetite neuropeptides, including NPY, αMSH and AGRP, especially modulating NA dopamine release and activity . The relevant dopaminergic regions, the VTA and NA, respond to gut signals such as ghrelin, which acts partly directly in the NA and VTA and partly indirectly via the arcuate, particularly AgRP  or αMSH, both of which also regulate oxytocin responses to appetite signals as indicated above. In addition, catecholamine lesion of the PVN disrupts AgRP expression and signalling, so upstream dopamine modulation of neuroendocrine systems may also be recruited in control of the arcuate peptides .
Much evidence links dopamine with anorexia: there is altered dopamine function in the striatum and caudate putamen, along with other monoamines, particularly serotonin [266,281]. There is thought to be decreased dopamine turnover, as indicated from measuring CSF levels of dopamine metabolites in anorectic patients. This seems to reflect dysfunction of the dopaminergic system, which is associated with symptoms such as the inability to experience motivation and reward. Low dopamine is also associated with motor hyperactivity, a core symptom of anorexia nervosa that increases with the level of starvation . PET imaging implicates altered dopamine activity in anorexic patients , while binge eating is associated with a polymorphism in the dopamine transporter gene in women , also indicating that altered dopamine concentrations at relevant target regions play an important role. Dysregulation of other monoamines is also apparent, with serotonin inhibiting food intake. However, dieting reduces tryptophan, the precursor for serotonin, perhaps explaining why patients have unexpectedly low serotonin in their CSF . Serotonin and noradrenaline transmitter systems are both dysfunctional in bulimic patients , again based on CSF analysis so precision in changes at monoamine sources or targets are unknown.
With the lack of a relevant animal model for anorexia or bulimia, it is hard to find evidence in the literature for whether dopamine or oxytocin play a primary role in these eating disorders, let alone whether their interaction explains part of the phenotype. There have been limited opportunities to experimentally study any potential interaction between dopamine and oxytocin in this context, perhaps because the study of oxytocin in anorexia or obesity has been relatively limited up to now. However, if as indicated above there is a robust coregulation of reward by oxytocin and dopamine it might be expected that disturbances in food reward that are associated with anorexia/bulimia could be explained by disruption in oxytocin–dopamine pathways.