The Role of Oxytocin in Human Affect

The Social-Approach/Withdrawal Hypothesis

Recent research suggests that OT may facilitate social salience (Shamay-Tsoory et al., 2009) rather than facilitating the expression of prosocial emotions—that is, that it may regulate a much wider range of emotions, including aggression toward potential intruders and rivals. In the first experimental study in humans to examine this possibility (Shamay-Tsoory et al., 2009), participants who were either administered OT or not played a game of chance with another (fake) participant who won more money (envy manipulation), lost more money (gloating manipulation), or won and lost equal amounts of money; the authors reported that OT nasal spray administration increased rather than decreased the negative social emotions, envy and gloating. The authors argued that their findings reconcile the prevailing view that OT increases prosocial behavior with contradictory findings of increased aggression following administration of OT (e.g., Bosch, Meddle, Beiderbeck, Douglas, & Neumann, 2005). A second study (Rimmele et al., 2009) reported that OT administration improved the encoding of faces into memory, regardless of their emotional valence (but see Guastella, Mitchell, & Mathews, 2008—differences in experimental design may have contributed to the different findings). The authors concluded that OT may have a globally enhancing effect on the processing of face stimuli (but see Domes, Heinrichs, Gläscher, et al., 2007, who reported a global suppressing effect of OT on amygdala activity). More recently, OT was observed to improve empathic accuracy in less socially proficient individuals only (Bartz, Zaki, et al., 2010). Although empathy is a prosocial affective behavior, the authors interpreted their study as evidence that OT increases perceived salience of social cues, an interpretation consistent with that of Shamay-Tsoory and colleagues (Shamay-Tsoory et al., 2009). In another study, participants treated with OT nasal spray were found to behave altruistically toward members of their own group (i.e., cooperative motivation to benefit the ingroup without hurting the outgroup), as well as more aggressively toward outsiders, when their own group was threatened with financial loss (i.e., noncooperation toward the outgroup to the extent that aggression could hurt the ingroup; De Dreu et al., 2010). While these findings highlight that the impact of OT may affect a wider range of social-emotion-related behaviors than just positive prosocial ones, we propose another explanation than the social salience hypothesis.

Approach and withdrawal motivation are dimensions on which affective processes may be organised, and meta-analyses of neuroimaging studies have concluded that approach and withdrawal may be the best explanation for the way the brain processes emotional stimuli (Barrett & Wager, 2006). Approach motivation is appetitive in nature, generally associated with positive affective states, and may involve moving toward a desired goal (e.g., happiness, enthusiasm, trust). Withdrawal motivation is aversive in nature and may involve moving away from a source of negative stimulation (e.g., fear, disgust, anxiety). A distinction between positive emotions and approach-related motivation is supported by research in affective neuroscience (Carver & Harmon-Jones, 2009; Harmon-Jones, Peterson, & Harris, 2009). While anger and aggression are characterised by negative valence, they are approach-related emotions, and both elicit left-prefrontal cortical activity similar to that elicited by positive affective states (Carver & Harmon-Jones, 2009). For example, while anger is a negative emotion, it may involve approaching a particular desired outcome, which involves the creation of discomfort for someone else or of rectifying an injustice (Carver & Harmon-Jones, 2009). Similarly, jealousy, an experience consistent with envy, is correlated with anger (Harmon-Jones et al., 2009) and belongs to the “family” of approach-related emotions (Lazarus, 1991). Moreover, both jealousy (Harmon-Jones et al., 2009) and anger (Carver & Harmon-Jones, 2009) elicit left-frontal cortical activation, a pattern of activation consistent with approach-related motivation. Gloating is the malicious pleasure at another’s misfortune (Ortony, Clore, & Collins, 1990) and is therefore related to happiness, an approach-related motivation, albeit in a negative social context. We note that the psychology and neuroscience of aggression have also conceptualised aggressive behavior as being approach related (Harmon-Jones & Sigelman, 2001). While OT-induced enhanced aggression (De Dreu et al., 2010) may be interpreted as reflecting the “prickly” side of OT, we suggest that this defensive aggression may be considered an approach-related emotion. It is also consistent with findings from animal studies in which maternal aggression—involving protection of pups from intruders—is associated with intracerebral release of OT (Bosch et al., 2005).

Support for the Social-Approach/Withdrawal Hypothesis

The results from studies on trust lend support to the proposal that OT facilitates approach-related behaviors (i.e., emotional engagement) and reduces withdrawal-related behaviors (i.e., anxiety and fear). In one of the first human OT nasal- spray studies on social behavior (Kosfeld et al., 2005), OT-administered students participating in a trust game were more generous in giving money to others than were those receiving a placebo, highlighting an increase in prosocial and approach-related, trusting behavior. The authors argued that OT facilitated trust by decreasing the impact of betrayal adversity and increasing the desire for, and reward associated with, approach behavior, consistent with the conclusions drawn from animal studies. Another study (Zak et al., 2007) reported that OT increased generosity by 80% relative to a placebo in a one-shot decision on how to split a sum of money with a stranger that could be rejected. This is known as the ultimatum game in experimental economics, as it requires perspective taking on how a partner may respond. This study also reported that OT had no effect on a unilateral monetary transfer—known as the dictator game—designed to dissociate generosity from altruism, highlighting the role of OT during emotional engagement. A neuroimaging study (Baumgartner, Heinrichs, Vonlanthen, Fischbacher, & Fehr, 2008) utilizing a trust game reported that increases in trust following administration of OT were associated with reductions in activation in the amygdala, the midbrain regions, and the dorsal striatum, which may reflect reductions in the fear of social betrayal.

More recently, these findings on trust have been extended to patients with autism (Andari et al., 2010; see also Guastella et al., 2010). Andari and colleagues reported that autistic patients displayed stronger interactions with the most socially cooperative partner and heightened feelings of trust and preference after being treated with OT during a simulated ball game in which participants interacted with fictitious partners. A notable exception to the finding that OT may facilitate trust is a recent study on borderline personality disorder (BPD; Bartz, Simeon, et al., 2010), which reported that OT may actually hinder trust and cooperation. This study found that BPD patients trust their partners less and are more likely to defect in response to partner cooperation during the Assurance Game, a variant of the Prisoner’s Dilemma involving significant trust issues. The Prisoner’s Dilemma is a fundamental problem in game theory in which two people have two options whose outcome depends crucially on the simultaneous choice made by the other. It is interesting to note that while these findings appeared to be driven by anxiously attached, rejection-sensitive participants, OT actually promoted cooperative behavior in anxiously attached but low avoidant (i.e., intimacy seeking) individuals. The authors provided a number of explanations for the reported reductions in trust and cooperation. They suggested that, consistent with the social-approach/withdrawal hypothesis, OT may activate approach-type behaviors but remind BPD and anxiously attached individuals of prior experiences when affiliation has gone awry, eliciting chronic concerns about trust and closeness. However, the authors also suggest that OT might increase the salience of social cues, which would be consistent with the social-salience hypothesis, such that an increase in social salience would subsequently activate chronic concerns about trust and closeness and elicit preexisting maladaptive strategies, such as defection and punishment of partners. This study suggests that OT may have more nuanced effects on human affect and highlights the importance of collecting more detailed information on individual differences in attachment styles, as well as on approach and withdrawal behaviors more generally.

Converging evidence from additional studies supports the proposal that OT facilitates social-approach-related behaviors. A study (Domes, Heinrichs, Michel, et al., 2007) that used the Reading the Mind With the Eyes Test (RMET)—a task that determines the ability to detect correct emotions from the eyes of others—reported that a single dose of OT was sufficient to increase the perception of subtle social cues from the eye region of other individuals and that this effect was more pronounced for more difficult-to-read faces than for easy ones. We have reported that OT may increase gaze specifically toward the eye region of human faces (Guastella, Mitchell, & Dadds, 2008), providing a potential mechanism by which OT increases emotion recognition and social approach. More recently, we (Guastella et al., 2010) administered the RMET to young people with autism spectrum disorders following administration of either OT or a placebo and reported improvements in emotion recognition in participants receiving OT (see also Andari et al., 2010). Another study (Marsh, Yu, Pine, & Blair, 2010) reported that OT improves the recognition of positive facial expressions but not anger, disgust, fear, sadness, or surprise, highlighting that OT facilitates a recognition bias to positive social cues (but see Fischer-Shofty, Shamay-Tsoory, Harari, & Levkovitz, 2010). Similarly, we have shown that healthy males administered OT, in comparison to a placebo, identified positive sex and relationship words more efficiently than threatening sex and relationship words (Unkelbach, Guastella, & Forgas, 2008). Intriguingly, Marsh and colleagues (Marsh et al., 2010) reported that oxytocin improved the recognition of happy facial expressions but not the recognition of other expressions and that these effects were more pronounced for subtle expressions of happiness that had been morphed with neutral faces. Others have reported that a single dose of OT reduced misclassification of positive emotions as negative and slowed reaction times for correctly identifying fearful facial expressions (Di Simplicio, Massey-Chase, Cowen, & Harmer, 2009). We have reported that OT-administered participants are more likely to remember familiar happy faces and less likely to remember angry and neutral human faces than are those administered a placebo (Guastella, Mitchell, & Mathews, 2008), suggesting that OT may have longer-term effects on emotional processing involving the processes involved in laying down memories. Other more recent research has reported that OT increases emotional but not cognitive empathy in response to both positively and negatively valent stimuli (Hurlemann et al., 2010). Still, empathy is an approach-related affective state regardless of whether the stimulus to which empathy is directed is negative or positive. Together, these studies highlight that OT facilitates emotion recognition, increases the recognition of social-reward-driven stimuli, decreases the experience of social threat, and encourages approach-related behaviors such as empathy.

There is also strong support for the hypothesis that OT may inhibit social-withdrawal-related behaviors. We have shown, for example, that OT improves self-appraisals (therefore reducing negative-self appraisals) of speech performance for patients with social anxiety disorder, entailing that OT may reduce key threat-associated cognitive-processing biases that characterize the disorder (Guastella, Howard, Dadds, Mitchell, & Carson, 2009). In another study (Petrovic, Kalisch, Singer, & Dolan, 2008), negative evaluations of faces paired with electric shock were abolished following administration of OT, and this effect was associated with an attenuation of activity in the anterior medial temporal and anterior cingulate cortices. This study clearly demonstrates the anxiety-reducing effects of OT on a particularly aversive stimulus. Consistent with these findings, Evans and colleagues reported more recently that OT decreases aversion to angry faces in an associative learning task (Evans, Shergill, & Averbeck, 2010). These authors highlighted that the impact of OT is specific to social and not financial information and that OT specifically decreased aversion to angry faces—a finding that was interpreted as decreasing the motivation to withdraw. In an earlier study, Heinrichs and colleagues (Heinrichs, Baumgartner, Kirschbaum, & Ehlert, 2003) administered OT or a placebo to participants before psychosocial stress; they reported that the combination of OT and social support resulted in the lowest cortisol concentrations as well as increased calmness and decreased anxiety during stress, and they concluded that OT plays an important role in a biological mechanism for stress-protective effects. The authors also reported a trend toward attenuated cortisol levels by OT even in the absence of social support (see also Ditzen et al., 2009). These findings suggest that OT plays an important role in attenuating withdrawal-related behaviors associated with physiologic stress reactivity, including reducing cortisol responses to stress and decreasing cardiac reactivity.

Neuroimaging studies provide further evidence that OT increases social approach and decreases withdrawal. One of the first neuroimaging studies on the impact of OT reported a reduction in amygdala activation and a coupling between amygdala and brainstem regions in response to angry or afraid faces and fearful/threatening scenes (Kirsch et al., 2005). The authors of this study were particularly interested in the amygdala region, given its role in fear, social avoidance, and phobia, and they concluded that OT affects regions implicated in autonomic and behavioral manifestations of fear. Importantly, these findings are contrary to what one would hypothesize on the basis of the social-salience hypothesis. In regard to social salience, an increase rather than a decrease in amygdala activation would be hypothesized, considering that the amygdala is involved in processing the salience of stimuli. Another neuroimaging study (Domes, Heinrichs, Gläscher, et al., 2007) reported that OT attenuates amygdala responses to emotional faces regardless of valence (but see Gamer, Zurowski, & Büchel, 2010). This study examined neural responses to fearful, angry, and happy facial expressions after OT administration, reporting that OT, relative to placebo, reduced (right-sided) amygdala responses to all three face categories. Interestingly, whole-brain analysis further revealed enhanced activations to these stimuli in several frontal, temporal, and brainstem areas in the placebo but not the OT condition. No significant clusters of activation were observed in the OT relative to the placebo condition. While this study does not support the social-salience hypothesis, it is also not what would be hypothesized from the perspective of the prosocial (i.e., increased activation in response to happy expressions) or social-approach/withdrawal (i.e., increased activation in response to happy expressions, as well as reduced or suppressed responses to negative facial expressions) hypotheses. However, it is interesting to note that the authors of this article concluded that OT may reduce uncertainty about the predictive value of social stimuli, which may then facilitate approach behaviors.

The more recent study by Gamer and colleagues (Gamer et al., 2010) reported that OT suppressed activation in response to fearful faces but enhanced activity for happy expressions in lateral and dorsal regions of the anterior amygdala, indicating a shift toward the processing of positive social stimuli— a finding consistent with our social-approach/withdrawal hypothesis. This study had a larger sample size and higher-resolution functional neuroimaging than did the study by Domes and colleagues (Domes, Heinrichs, Gläscher, et al., 2007). It is important to note that while the amygdala has traditionally been associated with the emotion of fear, it is also engaged by positive stimuli, reward, and novelty (Barrett & Wager, 2006). This study (Gamer et al., 2010) also used an MRI-compatible eye-tracking camera and reported that OT increases the likelihood of gaze shifts toward the eye region—supporting our prior findings (Guastella, Mitchell, & Dadds, 2008)—and that this gazing pattern is associated with activity in the posterior amygdala and superior colliculi. These findings highlight that while the anterior amygdala displays a valence-specific response consistent with a bias to positive social cues, posterior amygdala activation may underlie OT-induced increases in gaze shifts to the eye region, which may contribute to improvements in emotion recognition. Finally, Labuschagne and colleagues (Labuschagne et al., 2010) reported that OT ameliorates the hyperactivity to fearful facial expressions in patients with generalized anxiety disorder, a finding that was interpreted as a normalization of activity. Unlike other studies on healthy individuals however, these authors also noted that OT did not affect amygdala activity during an emotional-face-matching task of fearful, angry, happy, and neutral faces in healthy control subjects, highlighting the need for further studies using mood-induction paradigms to maximally challenge underlying neural circuitry.

Another study (Kéri & Benedek, 2009) provides some insight into a potential mechanism by which OT may facilitate the motivation to approach versus withdraw from a stimulus. OT was observed to enhance the perception of biological motion (a walking character) over nonbiological motion (a rotating shape). The authors concluded that OT modulates the perception of socially relevant stimuli and speculated that OT may serve as a “moving limb” enhancer, which may facilitate the detection of living and moving organisms in the environment. We suggest that perception of biological motion may motivate reciprocal approach/withdrawal behavior such that the detection of biological motion in others (i.e., social perception) may be used to decide whether the perceiver him- or herself should approach or withdraw (i.e., motivation to act). That said, our own findings (Guastella, Carson, Dadds, Mitchell, & Cox, 2009) based on a different task—the visual-search paradigm, involving a series of schematic faces in which participants are asked to determine whether facial expressions were identical or whether there was a face with a different expression as quickly and as accurately as possible—indicated that OT did not influence the early detection of angry and happy faces. We concluded that OT may have greater influence at more conceptual and elaborate levels of processing rather than at the earlier perceptual stage of processing. Further studies are required to clarify the impact of OT on rapid versus elaborate processing of social stimuli.

The Psychological Science of OT

The literature suggests three possible hypotheses relating to the impact of OT on human social behavior. These include the prosocial, social-salience, and social-approach/withdrawal hypotheses. The prosocial hypothesis suggests that the OT system is primarily involved in positive prosocial behaviors; the social-salience hypothesis suggests that this system plays a key role in a wider range of social-emotion-related behaviors, while the social-approach/withdrawal hypothesis suggests that OT will facilitate approach-related social emotions (including negative emotions such as anger, aggression, envy, and gloating) and inhibit withdrawal-related social emotions (such as the perception of anger or experience of fear). We suggest that the findings (at least from research on humans) is more easily explained by the social-approach/withdrawal hypothesis than by either the prosocial or social-salience hypotheses. (See Table 1 for a summary of and critical comments on the studies that support each of these hypotheses). The social-approach/withdrawal hypothesis highlights an important role for OT in the social domain, in which OT may facilitate social-approach-related emotions—including the “prickly” states of jealousy, anger, gloating, and aggression. While researchers have yet to directly compare competing hypotheses for the role of OT in human affect, prior research provides increasing support for the social-approach/withdrawal hypothesis (Table 1).

OPEN IN VIEWER

Table 1.

Key Studies Supporting One or More of Three Possible Hypotheses Relating to the Impact of Oxytocin (OT) on Human Social Behavior, and Critical Comments

Hypothesis supported	Study & finding	Critical comment
Prosocial	Zak, Stanton, & Ahmadi (2007): OT increased generosity by 80% relative to those given placebo.	This study distinguishes between generosity and altruism, highlighting the role of OT in emotional engagement. While generosity is prosocial behavior, the finding cannot particularly support or reject any of the three possible hypotheses.
	Guastella, Mitchell, & Dadds (2008): OT improved face perception and emotion recognition by increasing gaze to the eyes.	These findings suggest a potential mechanism for improved face perception and interpersonal communication. This study only examined gaze patterns to neutral faces.
	Guastella, Mitchell, & Matthews (2008): OT increased memory for happy faces relative to angry and neutral faces.	The study design does not allow for prosocial and social-approach hypotheses to be distinguished. But the findings do not support the social-salience hypothesis. The study also highlights that the effects of OT may be long lasting and have an effect on memory.
	Unkelbach, Guastella, & Forgas (2008): OT-administered subjects detected positive sex and relationship words more efficiently (as indicated by reduced response latency) than threatening sex and relationship words. OT had no effect on other positive and negative stimuli.	The findings highlight the specificity of the effect of OT on the recognition of words relating to sexuality, bonding, and social relationships. The findings do not support the social-salience hypothesis.
	Fischer-Shofty, Shamay-Tsoory, Harari, & Levkovitz (2010): OT improved subjects’ ability to recognize fear but not happiness, sadness, disgust, anger, or surprise. Authors argued for a specific role for OT in the recognition of fear.	Contradicts findings reported by Domes, Heinrichs, Gläscher, et al. (2007) and Marsh, Yu, Pine, & Blair (2010)—see below.
	Marsh, Yu, Pine, & Blair (2010): OT improved recognition of positive facial expressions but not anger, disgust, fear, or sadness.	The finding highlights that OT improves recognition of happiness specifically. The authors suggest that this finding may reflect increasing sensitivity to signs of trustworthiness.
Social-salience	Shamay-Tsoory et al. (2009): OT increased envy and gloating.	We (Kemp & Guastella, 2010) have proposed an alternative explanation of these findings: OT may instead serve to facilitate approach-related behaviors.
	Rimmele, Hediger, Heinrichs, & Klaver (2009): OT improved recognition memory for faces but not for nonsocial stimuli. The effect of OT was specific to “known” faces.	OT did not differentially impact on the recognition of positive vs. negative face stimuli, suggesting a globally enhancing effect of OT on memory for faces.
	Gamer, Zurowski, & Büchel (2010): OT increased activity in the posterior amygdala, and this activity was associated with the increased gaze shifts to the eye region irrespective of emotional expression type. However, OT also attenuated activation in the anterior amygdala for fearful faces but enhanced activity for happy faces.	The authors suggest OT may enhance emotion recognition (i.e., OT increases gaze shifts to the eye region regardless of whether valence is positive or negative), which then facilitates an increase in prosocial emotion. Decreased activity to fearful faces but increased activity to happy faces supports the social-approach/withdrawal hypothesis.
	Bartz et al. (2010): OT improved empathic accuracy for less socially proficient individuals assessed using the Autism Spectrum Quotient.	The authors conclude that OT increases the perceived salience of social cues, consistent with Shamay-Tsoory et al. (2009). The authors suggested that OT should benefit only those individuals less attuned to social information and that administration of OT may correct for social-cognitive inefficiencies.
	De Dreu et al. (2010): OT produced differential effects on group conflict including altruistic responses to the ingroup and defensive aggression to the outgroup.	Although others have interpreted these findings as evidence for the “prickly” side of OT, these findings may also be interpreted in the context of our social-approach/withdrawal hypothesis: Defensive aggression may be considered an approach-related emotion.
Social-approach/withdrawal	Heinrichs, Baumgartner, Kirschbaum, & Ehlert (2003): The combination of OT and social support was associated with the lowest cortisol concentrations, increased calmness, and decreased anxiety during stress.	The authors conclude that the findings concur with animal research highlighting a role for OT as a potential biological mechanism for stress-protective effects. We suggest that findings are also consistent with inhibition of withdrawal behaviors. Findings do not support the social-salience hypothesis.
	Kosfeld, Heinrichs, Zak, Fischbacher, & Fehr (2005): OT increased prosocial, approach-related trust behavior. Baumgartner, Heinrichs, Vonlanthen, Fischbacher (2008): These effects may be associated with a reduction in amygdala activation.	Studies focusing only on trust cannot produce findings that particularly support or reject any of the three possible hypotheses. While Kosfeld, Heinrichs, Zak, Fischbacher, & Fehr (2005) hint of the possibility that OT promotes “blind” trust regardless of reciprocation, a recent study on autism (Andari et al., 2010) highlights the ability of OT to discriminate between good and bad outcomes. A notable exception to an increase in trust following OT is the study by Bartz, Simeon, et al. (2010).
	Domes, Heinrichs, Michel, Berger, & Herpertz (2007): OT increased perception of subtle social cues from the eye region of other individuals.	The authors interpret their findings as suggesting OT increases the motivation to engage. This effect was more pronounced for difficult vs. easy items.
	Domes, Heinrichs, Gläscher, et al. (2007): OT reduced neural responses to fearful, angry, and happy facial expressions.	The authors suggest that findings may reflect a reduction in uncertainty that may motivate the individual to initiate social-approach behavior. The findings are more consistent with the prosocial and social-approach/withdrawal hypotheses than with the social-salience hypothesis.
	Gamer & colleagues (2010): OT attenuated amygdala activation for fearful faces but enhanced activity for happy faces.	Differential activation to fear versus happy faces supports the social approach-withdrawal hypothesis, while the increased activity regardless of valence in posterior amygdala may underlie behavioral findings relating to increased gaze shifts to the eye region following OT.
	Di Simplicio, Massey-Chase, Cowen, & Harmer (2009): OT reduced misclassification of positive emotions as negative and slowed reaction times to correctly identifying fearful facial expressions.	Participants recalled more words describing positive characteristics than words describing negative characteristics. The authors conclude that OT reduces the emotional salience of potentially threatening social stimuli (i.e., fear). Findings provide direct support for the social-approach/withdrawal hypothesis and are in contrast to the social-salience hypothesis.
	Kéri & Benedek (2009): OT increased sensitivity to biological motion but not nonbiological motion.	The authors conclude that OT specifically modulates the perception of socially relevant stimuli. We suggest that these findings may also provide some insight into a potential mechanism by which OT may facilitate the motivation to approach versus withdraw from a stimulus. However, we note that OT did not facilitate other low-level perceptual abilities such as the early detection of emotional facial expressions (Guastella, Carson, Dadds, Mitchell, & Cox, 2009).
	Andari et al. (2010); Guastella et al. (2010): OT increased prosocial behaviours and affect.	Although difficult to distinguish between the different hypotheses, one of these studies (Andari et al., 2010) highlights the ability of OT to discriminate between good and bad outcomes, a finding that is not consistent with the social-salience hypothesis.
	Evans, Shergill, & Averbeck (2010): OT specifically decreased aversion to angry faces in an associative learning task.	The authors highlight that the impact of OT is specific to social and not financial information in a decision-making task. OT specifically decreased the motivation to withdraw from angry faces, whereas financial reward processing or choices related to happy vs. sad faces were not affected.
	Hurlemann et al. (2010): OT increased emotional but not cognitive empathy in response to both positive and negative stimuli.	Empathy is an approach-related emotion regardless of whether the stimulus to which empathy is directed is negative or positive. The authors conclude that OT facilitates a social-reinforcement advantage as operationalized by encouraging vs. discouraging feedback through presentation of smiling or angry facial expressions to guide behavior—a finding consistent with the social-approach/withdrawal hypothesis.
	Labuschagne et al. (2010): OT decreased hyperactivity to fearful facial expressions in patients with generalized anxiety disorder.	This study suggests that OT may help to normalize amygdala hyperactivity in response to social cues (i.e., decreased motivation to withdraw). The authors also report that OT did not have an effect on amygdala activity during matching of emotional facial expressions in healthy controls, highlighting a need for maximal challenge of underlying neural circuitry during processing of social stimuli.

Note: Although studies included in the table may support multiple hypotheses, we categorized studies according to the way in which authors of the studies have interpreted their findings.

Table 1 provides a useful framework for interpreting the results of recent human studies that have examined the impact of OT on social behavior. The bottom line is that studies that differentiate the three competing hypotheses remain to be conducted. Studies focusing on trust could be argued to support any of the three hypotheses. Most studies have employed stimuli (e.g., faces) that do not allow the impact of OT on negatively valenced, approach-related behaviors (such as anger or aggression) to be examined. Studies have reported that OT attenuates amygdala activity during presentation of both angry and fearful faces (e.g., Domes, Heinrichs, Gläscher, et al., 2007; Gamer et al., 2010; Kirsch et al., 2005), findings that would be expected during inhibition of withdrawal behaviors. We suggest that the impact of OT on the experience of anger would be quite different from OT’s impact on the perception of it, as the experience of anger involves approach-related motivation (Carver & Harmon-Jones, 2009; Harmon-Jones & Sigelman, 2001). Thus, our hypothesis would be that OT will inhibit withdrawal-type behaviors during the perception of anger (e.g., an angry facial expression) but facilitate approach-type behaviors during the experience of anger. Studies that directly test such hypotheses remain to be conducted. While we do not propose that evidence to support our social-approach/withdrawal hypothesis can only be obtained with paradigms involving emotion experience—indeed few such studies have been conducted—such studies may be helpful in distinguishing between these hypotheses, as emotional experience may elicit approach or withdrawal more reliably.

Conclusion

In conclusion, we propose a novel, alternative hypothesis for the impact of OT on social behavior in humans: that social approach—including negative, approach-related social emotions (e.g., anger, aggression, jealousy, envy, and gloating)—is increased while social withdrawal is inhibited. This hypothesis is suggested by convergent findings from a number of clinical (e.g., social-anxiety disorder; Guastella, Howard, et al., 2009; autism, Andari et al. 2010; Guastella, et al., 2010; though borderline personality disorder is an exception, Bartz, Simeon, et al., 2010), behavioral (e.g., valence-specific effects; Unkelbach et al., 2008; Marsh et al., 2010; but see Fischer-Shofty et al., 2010), biological (e.g., cortisol; Heinrichs et al., 2003) and neuroimaging (e.g., attenuated amygdala activation, a finding that would not be expected under the social salience hypothesis; Domes, Heinrichs, Gläscher, et al., 2007; valence-specific amygdala activation; Gamer et al., 2010) studies. The social-approach hypothesis highlights an important role for OT in facilitating approach-related social behaviors that may include the prickly states of anger and aggression. A number of questions remain to be answered. For example, does OT enhance amygdala activation during the experience of approach-related negative (e.g., anger, aggression) versus withdrawal-related (e.g., fear, anxiety) affective states? And does OT modulate biological markers of approach/withdrawal and social engagement such as frontal alpha asymmetry (Mathersul, Williams, Hopkinson, & Kemp, 2008) and heart-rate variability (Kemp et al., 2010)? Intriguingly, preliminary findings (Norman et al., 2010) suggest that OT may increase autonomic cardiac control, including parasympathetic and sympathetic nervous system activity consistent with a role for heart rate variability in social approach. The authors also reported that loneliness ratings are associated with reduced parasympathetic response to OT, suggesting that loneliness may shift a salubrious response to OT involving autonomic coactivation to a deleterious state involving selective sympathetic activation. Future research should attempt to further clarify the role of OT in human affect by using paradigms that involve emotion experience and that directly compare the three hypotheses we have discussed.