Tuesday, March 1, 2022

U.S. states with higher aggregate intelligence exhibit higher levels of alcohol consumption; regional alcohol consumption is also linked to better health, less health issues, and better well-being

Updated IQ and Well-Being Scores for the 50 U.S. States. Bryan J. Pesta. J. Intell. 2022, 10(1), 15; February 27 2022. https://doi.org/10.3390/jintelligence10010015

Abstract: At the level of the 50 U.S. states, an interconnected nexus of well-being variables exists. These variables strongly correlate with estimates of state IQ in interesting ways. However, the state IQ estimates are now more than 16 years old, and the state well-being estimates are over 12 years old. Updated state IQ and well-being estimates are therefore needed. Thus, I first created new state IQ estimates by analyzing scores from both the Program for the International Assessment of Adult Competency (for adults), and the National Assessment of Educational Progress (for fourth and eighth grade children) exams. I also created new global well-being scores by analyzing state variables from the following four well-being subdomains: crime, income, health, and education. When validating the nexus, several interesting correlations existed among the variables. For example, state IQ most strongly predicted FICO credit scores, alcohol consumption (directly), income inequality, and state temperature. Interestingly, state IQ derived here also correlated 0.58 with state IQ estimates from over 100 years ago. Global well-being likewise correlated with many old and new variables in the nexus, including a correlation of 0.80 with IQ. In sum, at the level of the U.S. state, a nexus of important, strongly correlated variables exists. These variables comprise well-being, and state IQ is a central node in this network.

Keywords: state IQ; aggregate IQ; well-being; g nexus; PIAAC; NAEP

Based on objective hormonal changes, these authors confirm that in this case, the woman was able to put herself into a relatively continuous orgasmic state after years of yoga and tantric training

A Case of Female Orgasm Without Genital Stimulation. James G. Pfaus, Karolin Tsarski. Sexual Medicine, Volume 10, Issue 2, April 2022, 100496. https://doi.org/10.1016/j.esxm.2022.100496

Introduction: Orgasm is a complex, multimodal reflex induced typically by genital stimulation. Genitally stimulated orgasms (GSOs) activate excitatory neurochemical pathways in the brain and spinal cord that ultimately stimulate sympathetic outflow and the inhibition of parasympathetic spinal circuits in the lower lumbar cord. However, some women claim to have orgasms spontaneously without genital stimulation.

Aims: To report the case of a 33-year-old woman who developed the ability to attain and control the duration of a subjective orgasmic state without genital stimulation after tantric training.

Methods: Blood was taken at weekly intervals before, during, and after spontaneously-induced orgasms that lasted 5 or 10 minutes, or after a 10-minutes period of book reading. Plasma was analyzed using ELISA for luteinizing hormone, follicle stimulating hormone, free testosterone, and prolactin. The woman also provided subjective scores for different types of orgasms using the Mah and Binik (2002) Orgasm Rating Scale (ORS).

Results: Prolactin levels post orgasm increased by 25% and 48%, respectively, after the 5- or 10-minutes non-genitally stimulated orgasm (NGSO), and were still elevated from baseline 30 minutes after orgasm. No changes were observed in FSH or free testosterone. The pattern of sensory, affective, and evaluative orgasm ratings after a 10-minutes NGSO was similar to orgasms induced by clitoral or anal stimulation. Book reading did not result in any change in prolactin.

Conclusion: Prolactin surges after orgasm are an objective marker of orgasm quality. The increase in prolactin after her NGSOs indicate that they induce the same physiological changes as GSOs and result from “top-down” processing in the brain.

Key Words: Spontaneous OrgasmHormoneProlactinOrgasm Rating Scale


Leeners et al6 proposed using the post-orgasmic prolactin surge as an objective marker of orgasm quality, given that the surge levels followed a lawful increase in women that experienced 1 or 2 orgasms during partnered intercourse, and whose subjective rating of the orgasm quality also increased progressively from 1 to 2 orgasms. Moreover, the surge is not likely to be affected by movement artifacts during orgasm. A similar increase in prolactin was observed from the 5-minutes to 10-minutes NGSO in the present case study. In fact, the prolactin surge following the 10-minutes NGSO was in the range reported by Leeners et al after 2 orgasms from partnered intercourse.

The increase in luteinizing hormone during the 5-minutes NGSO was unexpected. Compared to her other pre baseline levels, the pre level before the 5-minutes orgasm was elevated, suggesting that she was in or around the ovulatory stage of her menstrual cycle. Indeed, as mentioned above, her pre levels of prolactin were increased before the 10-minutes orgasm condition a week later, suggesting that she was in her luteal phase at that time. Estradiol levels peak and fall before ovulation, which precedes the rise in progesterone secreted from the corpora lutea. During this time the ovaries release a pulse of testosterone27 which, combined with the genomic and non-genomic effects of the previous estradiol peak, increases women's responsivity to sex-related cues and their sexual desire.18,282930 An increase in luteinizing hormone-stimulated by orgasm during this phase could also enhance pregnancy, suggesting that female orgasm may have a reproductive function during this phase of the cycle.

Finally, yoga and tantra meditation training is reported to enhance sexual arousal, desire, and orgasm in both women31,32 and men,33 and can facilitate sexual satisfaction overall34 and intimacy in couple-based sex therapy.35 Brotto and colleagues have refined this into a mindfulness-based approach to treat sexual arousal, desire, and pain disorders in women.363738 Such training may well sensitize spinal and brain circuits for sexual climax and orgasm4,39 which should also feed forward to enhance neural mechanisms of sexual arousal and desire.18


The ability to be orgasmic sensitizes with sexual experience throughout the lifespan. Although orgasms are typically activated by genital stimulation in a “bottom-up” fashion, it is clear that they can be induced without genital stimulation in a “top-down” fashion that reflects both sensory and motor memory and may be activated by dreams, fantasy, and erotic imagery, especially in women. The woman examined in the present case study was able to put herself into a relatively continuous orgasmic state after years of yoga and tantric training. The increase in prolactin observed in the present study was an objective response to orgasm and followed a lawful pattern of doubling as the time spent in orgasm doubled, similar to the results of a prior study in which women had 1 or 2 orgasms during partnered sex.14 These data suggest strongly that NGSOs are not “faked” or partial orgasms, but rather reflect a top-down induction of a real subjective orgasmic state that includes objective hormonal changes.

Feelings of desire for an attractive alternative increase ambivalence toward the current partner, above & beyond how much people actually value their partner, & this has short- & long-term negative consequences for personal & relational well-being

Zoppolat, G., Faure, R., Alonso-Ferres, M., & Righetti, F. (2022). Mixed and conflicted: The role of ambivalence in romantic relationships in light of attractive alternatives. Emotion, 22(1), 81–99. FEb 2022. https://doi.org/10.1037/emo0001055

Abstract: People in romantic relationships tend to have positive feelings toward their partner and want their relationship to last. However, maintaining a romantic relationship over time is challenging, and people can often experience mixed and conflicting feelings (i.e., ambivalence) toward their significant other. While research has identified the serious consequences that ambivalence can have for personal and relational well-being, very little is known about the factors that can lead people to experience ambivalence in relationships. The present work examines how extradyadic desire (i.e., desire for someone other than the partner), a common difficulty people face in the context of monogamy, is a situation in which people feel more ambivalent toward their partner. In three studies (N = 1,178) using experimental, daily diary, and longitudinal approaches, we find that feelings of desire for an attractive alternative increase ambivalence toward the current partner, above and beyond how much people actually value their partner, and that this has short- and long-term negative consequences for personal and relational well-being. Furthermore, while most people could identify an attractive alternative in their life, desire for the alternative—rather than just their presence—seems to play a stronger role in increasing ambivalence. This work highlights the emotional processes through which attractive alternatives pose a threat to romantic relationships and the role that ambivalence plays in daily life and over time.

Endogenous oxytocin, cortisol, and testosterone in response to group singing: "We conclude that singing together can have biological and psychological effects associated with affiliation and social bonding"

Endogenous oxytocin, cortisol, and testosterone in response to group singing. D.L. Bowling et al. Hormones and Behavior, Volume 139, March 2022, 105105. https://doi.org/10.1016/j.yhbeh.2021.105105


• Singing and speaking were associated with decreases in salivary oxytocin concentrations, when performed together or alone.

• Salivary oxytocin concentrations decreased by less after singing together than after speaking together.

• Salivary cortisol concentrations decreased overall, but did not vary as a function of experimental conditions.

• Singing together improved self-perceived emotional status and social connectedness more than speaking together.

Abstract: Humans have sung together for thousands of years. Today, regular participation in group singing is associated with benefits across psychological and biological dimensions of human health. Here we examine the hypothesis that a portion of these benefits stem from changes in endocrine activity associated with affiliation and social bonding. Working with a young adult choir (n = 71), we measured changes salivary concentrations of oxytocin, cortisol, and testosterone from before and after four experimental conditions crossing two factors: vocal production mode (singing vs. speaking) and social context (together vs. alone). Salivary oxytocin and cortisol decreased from before to after the experimental manipulations. For oxytocin the magnitude of this decrease was significantly smaller after singing compared to speaking, resulting in concentrations that were significantly elevated after singing together compared to speaking together, after controlling for baseline differences. In contrast, the magnitude of the salivary cortisol decreases was the same across experimental manipulations, and although large, could not be separated from diurnal cycling. No significant effects were found in a low-powered exploratory evaluation of testosterone (tested only in males). At a psychological level, we found that singing stimulates greater positive shifts in self-perceived affect compared to speaking—particularly when performed together—and that singing together enhances feelings of social connection more than speaking together. Finally, measurements of heart rate made for a subset of participants provide preliminary evidence regarding physical exertion levels across conditions. These results are discussed in the context of a growing multidisciplinary literature on the endocrinological correlates of musical behavior. We conclude that singing together can have biological and psychological effects associated with affiliation and social bonding, and that these effects extend beyond comparable but non-musical group activities. However, we also note that these effects appear heavily influenced by broader contextual factors that shape social dynamics, such as stress levels, the intimacy of interactions, and the status of existing relationships.

Keywords: SpeechMusicSynchronyBondingAffectOxytocinCortisolTestosterone

4. Discussion

The hormone results described here indicate overall decreases in salivary concentrations of oxytocin and cortisol, and an absence of significant effects on salivary concentrations of testosterone in males. With respect to our experimental manipulations, vocal mode was determined to affect salivary oxytocin, with decreases being smaller after singing compared to speaking. In parallel, the survey results indicated that vocal mode and social context interacted to affect self-perceived affective status, with singing together producing the largest positive shift (see also Kreutz, 2014Pearce et al., 2015Schladt et al., 2017Weinstein et al., 2016). Singing together was additionally found to be more effective than speaking together at stimulating feelings of social connectedness (see also Pearce et al., 2015Weinstein et al., 2016). Finally, as might be expected, heart rates measured for ten participants suggested that physical exertion was highest during singing together and lowest during speaking alone. We now discuss each hormone result in turn (order-reversed from above), adding context from prior literature, exploring potential interpretations, and discussing connections with our other measures. We conclude by considering the connection between music and social bonding, as well as implications for future studies on the endocrinological correlates of musical behavior.

4.1. Testosterone

The absence of any significant effect on concentrations of testosterone in male saliva indicates that our experimental manipulations of vocal mode and social context were insufficient to systematically influence the acute regulation of testosterone in males. This is an important preliminary finding, but we emphasize caution in its interpreted because of the relatively low power of our testosterone model (observations per estimated effect =6.18). Nevertheless, we saw no indication that potential acute effects of music listening on male testosterone levels extend to singing (Fukui, 2001Fukui and Toyoshima, 2013Fukui and Yamashita, 2003). Future studies examining testosterone in the context of group singing or other musical behaviors should consider incorporating further factors or measurements potentially relevant to testosterone, such as the existence of interpersonal friendships, individual social status, singing skill, and competitive/cooperative group dynamics (Casto and Edwards, 2016Edwards et al., 2006Ponzi et al., 2016). It may also be useful to study testosterone and group singing over longer periods of time, as changes may appear gradually as participants move from forming new social bonds towards maintaining existing ones (Kornienko et al., 2016). Finally, it will be important for future studies to examine testosterone levels in females (Grant and France, 2001van Anders, 2013), especially given that most modern choirs are majority female (Elpus, 2015).

4.2. Cortisol

The experimental manipulations did not differentially impact concentrations of cortisol in saliva, which decreased markedly from before to after participation regardless of condition. Whether the magnitude of the decrease that we observed was specifically related to our experiment, as opposed to natural diurnal cycling, cannot be determined with certainty. A recent meta-analysis that aggregated cortisol data from more than 18,000 individuals found an average diurnal decrease during the time at which our experiment took place of approximately 12% per 30 min (range: 3% to 20% assuming wake-times from 7:00 and 10:00; Miller et al., 2016). On the basis of this data, it may be argued that the magnitude of the cortisol decrease that we observed (30%) was too large to be explained by diurnal cycling alone, which would in turn suggest that recreational singing or speaking, together or alone, comprise particularly effective ways to relax. That said, we emphasize that testing this hypothesis would have required inclusion of a “no-treatment” control condition in which hormone levels were assessed at the same times but in the absence of any experimental manipulations. Although such control conditions are not typically included in non-clinical research, it is important that future studies aiming to identify experimental effects of group singing on cortisol levels prioritize their inclusion, particularly if measurements are made in the evening, as they were here, and as they have been in previous studies (Beck et al., 2000Fancourt et al., 2015Fancourt et al., 2016Kreutz et al., 2004Schladt et al., 2017).

With respect to previous studies, our cortisol results mostly conformed to expectations. In relatively low-stress recreational contexts like that examined here, four previous studies have reported significant cortisol decreases after group singing (ranging in magnitude from approximately 18% to 27%; Beck et al., 2000Fancourt et al., 2015Fancourt et al., 2016Schladt et al., 2017), and two others have reported non-significant changes (in opposite directions; Kreutz et al., 2004Kreutz, 2014). Although all of these studies have been conducted in the evening, some have nonetheless found experimental effects on salivary cortisol. Of particular relevance here, Schladt et al. also compared singing together with singing alone. In contrast with our results, which indicated similar decreases in salivary cortisol after singing together (29%) and singing alone (31%), Schladt et al. found a significantly greater decrease after singing together (~32%2) than after singing alone (~20%). This difference in the effects of singing alone between our study and Schladt et al. is particularly notable given that the relevant conditions were highly similar in terms of design and execution (see below). Together, these findings indicate that recreational group singing is typically associated with reduced salivary cortisol concentrations, but that the effects of singing alone are more variable.

4.3. Oxytocin

Focusing on group singing first, salivary oxytocin concentrations decreased from before to after participation in our singing-together condition (Fig. 2B). Interpreting the generality of this effect is complicated by the literature. Four previous studies have assessed changes in oxytocin from before to after group singing, all in relatively low-stress contexts. Two of these studies report decreases similar in magnitude to that found here (Fancourt et al., 2016Schladt et al., 2017), and two report increases (Keeler et al., 2015Kreutz, 2014). Contrasting these “decrease studies” and “increase studies” provides insight into contextual factors that may modulate the relationship between group singing and oxytocin. The decrease studies include the present one (15% reduction after a 20-min choir rehearsal, n = 37), Schladt et al. (~22% reduction after a 20-min choir rehearsal, n = 38), and Fancourt et al. (24% reduction after a 70-min choir rehearsal, n = 193, across 5 choirs). The increase studies include Kreutz (39% increase after a 30-min choir rehearsal, n = 21) and Keeler et al. (~19% increase after a 6 min bout of improvised singing in a jazz quartet, n = 4). An additional study, Grape et al. (2003), is also relevant despite not measuring group singing per se (25% increase after a 45-min private singing lesson, n = 16). We consider each of the increase studies in detail below, attempting to determine factors underlying their differential effects, relative to our study and other decrease studies.

Starting with the study most different from our own, Grape et al. examined singing lessons between adult students and a teacher with whom they had practiced for a minimum of 6 months. This context differs from all other studies considered here, not only because participants sang alone rather than in a group, but because private lessons involve more focused social interaction than choir rehearsals. This point is particularly important because the intimacy of an interaction appears to be of greater relevance for understanding oxytocin responses to social context than whether an activity is technically performed alone or together with others (Bartz et al., 2011). Additional support for this idea comes from Keeler et al.'s study of improvised singing in a jazz quartet. Although their sample of just four individuals was too small to appropriately evaluate statistical significance, the average oxytocin increase that they observed (~19%) appeared specific to improvised singing—a separate condition where the quartet sang a precomposed jazz standard was instead associated with a small oxytocin decrease (~3%). Taken together, these results suggest that focused collaborative interaction, characteristic of teaching and social improvisation, is likely an important determinant of oxytocin activity in the context of group singing.3

The last increase study—Kreutz (2014)—is also the most similar to our own (and the other decrease studies). There are, however, a number of potentially important differences. A first difference is that whereas we and others have examined choirs that existed prior to the initiation of study, Kreutz examined a choir that was specifically formed for their study. Whereas many of the members of our choir had pre-existing friendships and previous experience singing together (often over multiple semesters), the members of Kreutz’s choir presumably had little such relations. This difference in social context is directly relevant to the value of affiliative behavior and potential for social bonding. A second difference is that the choir in our study was made-up entirely of members of the same age cohort (mean = 22 years, range = 17–28), whereas the choir in Kreutz included a greater mix of ages (mean ~ 49, range ~ 18–65). Schladt et al.'s choir was similar to ours in participant age (mean = 23 years, range = 19–29); Fancourt et al.'s was older (mean = 59 years) but still relatively narrow in age range (SD ~ 12). Given that extra-familial social bonds are biased towards members of the same age group (Nahemow and Lawton, 1975), the difference in age diversity of the choir studied by Kreutz may have similarly led to differences in affiliative value and bonding potential. Kreutz (2014) is also unique for its connection to a local television station, which ran advertisements to recruit participants, filmed rehearsals for a short documentary feature, and televised a final concert. This televised aspect may have led to a systemic difference in the personality traits of participants, potentially favoring group-level differences in individual oxytocin biology (Li et al., 2015). Finally, the music sung by Kreutz’s choir (e.g., the 1960s pop song “California Dreaming”) was more popular than the relatively obscure choral music used in our study and other decrease studies. Effects of music type on changes in salivary oxytocin concentrations after listening have been reported, and may be attributable to acoustic parameters conveying different types of affect (Ooishi et al., 2017).

A summary of these findings is that group singing in the context of choir rehearsals is primarily associated with decreases in salivary oxytocin concentrations, but that this effect can be modulated to the point of reversal by a variety of factors related to the broader social context in which singing occurs. The intimacy of the social interactions that take place during singing appears to be among the most important of these factors, with more intimate, direct, collaborative interactions potentially driving higher oxytocin activity than more anonymous, diffuse, independent interactions. Related to this, the nature of existing social relationships between group members also appears important, as it sets the stage for differences in the relative value of affiliative behavior and potential for social bonding. Finally, cultural and affective connotations of the specific music being sung may contribute further to differences in the effects of group singing on oxytocin.

Turning to our experimental manipulation of musical versus non-musical vocal production, we found evidence of a smaller decrease in salivary oxytocin concentration after speaking compared to singing (35% vs. 15% respectively; Fig. 2C). This novel comparison suggests that singing may sustain relatively higher levels of oxytocin activity than speaking, a similar vocal behavior that is also used for communication and that also involves high levels of interpersonal coordination. Psychologically, vocal production mode also influenced self-perceived affective status, with singing stimulating greater positive shifts in affect, particularly when performed together with others, as well as greater feelings of social connectedness. What is the basis for these differences in the biological and psychological effects of singing and speaking?

Despite broad similarities, singing and speaking are obviously characterized by a variety of neural, behavioral, and acoustic differences, any of which could be investigated as a potential basis for the differential effects that we observed. One straight-forward approach to understanding the different of effects singing versus speaking on salivary oxytocin is focused on potential differences in physical activity between them. Physical exertion can increase salivary oxytocin concentrations as well as feelings of affiliation (Anshel and Kipper, 1988de Jong et al., 2015Tarr et al., 2015). Just 10 min of jogging, for example, was found to increase salivary oxytocin by ~320% on average (de Jong et al., 2015). Accordingly, singing may increase salivary oxytocin concentrations more than speaking because it requires higher levels of physical exertion. Albeit preliminary, our heart rate data do not generally support this possibility. The largest and smallest estimated changes in salivary oxytocin occurred in conditions with approximately equal mean heart rates: salivary oxytocin decreased by 42% after speaking together, where mean heart rate was 94.8 bpm (SD = 11.3), but only decreased by 15% after singing alone, despite a similar mean heart rate of 92.8 bpm (SD = 14.0; cf. Figs. 2B and 7B). Nevertheless, further experiments are required to fully rule out physical exertion as a potential driver of differential effects of singing and speaking on salivary oxytocin.

Another approach to explaining the observed differential effects is focused on oxytocin's role in mitigating the effects of HPA axis activation (Gibbs, 1986Neumann, 2002Schladt et al., 2017). The idea here is that oxytocin activity is upregulated in response to stress, being released alongside cortisol and acting as an anxiolytic (Brown et al., 2016). Accordingly, if singing stimulated greater HPA axis activity than speaking, proportional oxytocin activity may underlie the differential effects that we observed. This interpretation is not supported by our salivary cortisol results, which showed a stable decrease in concentrations across conditions rather than an increase (or smaller decrease) after singing. This suggests that the singing and speaking conditions had similar effects on HPA axis activation. Furthermore, although salivary oxytocin and cortisol concentrations were significantly correlated across all of our measurements, we did not find evidence that these factors changed together over the course of our study. It is therefore unlikely that differential effects of singing versus speaking on salivary oxytocin are explained by differences in activation of the HPA-axis between conditions.

A final issue to consider in accounting for the differential effects of singing versus speaking on salivary oxytocin is focused on baseline differences. Careful inspection of Fig. 2 shows that salivary oxytocin concentrations measured before the speaking conditions were relatively elevated compared to those measured before the singing conditions. Thus, the interaction between time and vocal mode that we observed is confounded with baseline differences in salivary oxytocin between conditions. To test whether or not baseline differences (rather than experimental effects) are responsible for the observed interaction, we fit an alternate oxytocin model that incorporated baseline salivary oxytocin concentrations as a covariate in predicting post-experiment salivary oxytocin concentrations. This “baseline” oxytocin model was a significantly better predictor of post-experiment oxytocin concentrations than a parallel null model (χ2 = 7.85, d.f. =3, p = 0.049), and post-hoc tests indicated that the two-way interaction between vocal mode and social context was significant (estimate ± s.e. =0.38 ± 0.16; χ2 = 98.7, d.f. =1, p = 0.016). Specifically, oxytocin concentrations were estimated to be significantly higher after singing together compared to speaking together (72.7 pg/ml [c.i. =64.1–83.1] vs. 56.0 pg/ml [48.4–64.9], t(100.5) = −2.497, p = 0.014), but not after singing alone compared to speaking alone (62.8 [52.5-74.5] vs. 71.0 [59.2-85.6], t(100.5) =0.919, p = 0.360). Thus, post-experiment salivary oxytocin concentrations were found to differ as a function of vocal mode after specifically accounting for the influence of baseline differences. This reflects marked differences in the magnitudes of salivary oxytocin decreases across conditions in our original model, and indicates that the corresponding experimental effect is not an artifact of unexplained differences in baseline salivary oxytocin concentrations (see Supplementary Text 1 for further details).

Last, we come to our manipulation of social context. Whereas our original oxytocin model indicates that the effect of vocal mode on salivary oxytocin was not impacted by social context (Fig. 2B), the baseline oxytocin model just described indicates that post-experiment salivary oxytocin concentrations were affected by social context, which interacted with vocal mode. In fact, the pattern of results for singing versus speaking performed together was reversed for singing and speaking performed alone (although the difference was not significant). Nevertheless, in both models, singing together was similarly associated with relatively higher levels of oxytocin activity than speaking together. This is reflected by smaller decreases in salivary oxytocin concentrations from before to after singing versus speaking together in our original model, and higher salivary oxytocin concentrations after singing versus speaking together when controlling for baseline differences in the baseline model. Insight into the disruption of this pattern for singing versus speaking alone in the baseline model can be found by comparing our results with Schladt et al. (2017), who also examined singing together versus singing alone. Although Schladt et al.'s results were similar to ours for singing together (we found an average decrease of 15%, compared to their average decrease of ~22%), their results were markedly different for singing alone (we again found an average decrease of 15%, but they found an average increase of ~10%, Fig. 2B). As noted above in our discussion of discrepant cortisol results between alone conditions in our study and Schladt et al., this suggests that the endocrine effects of singing together are more stable than those of singing alone. This appears to be true despite a high level of procedural similarity between the alone conditions in our study and Schladt et al. In both studies, amateur young-adult chorists were provided with written instructions leading them through a short vocal warm-up followed by practice of their choir's repertoire, for a total duration of 20 min. Additionally, both studies found that singing alone produced positive shifts in self-perceived affect. Hormone changes in response to singing alone thus appear quite sensitive to more subtle differences between individuals and/or contexts (e.g., differences in stress and anxiety, or external performance pressures). Conversely, singing together with others seems to be capable of overriding individual or contextual differences to induce hormonal responses that are more similar across individuals.

We close this section on oxytocin with a brief discussion of our finding that male participants had higher salivary oxytocin concentrations than female participants (by 1.7 times on average). Importantly, this results does not reflect simple methodological error: our experimental procedures were the same for females and males, and their saliva samples were equally distributed across assays. We also emphasize that although oxytocin has often been associated with female-specific behaviors and effects—e.g., in the context of mothering, parturition, and lactation—it clearly functions in social behavior and cognition (as well as metabolism, cardiovascular function, and bone regeneration) across both females and males, with evidence of overlapping effects, as well as female- and male-specificity (Caldwell, 2018Quintana and Guastella, 2020). Simple rules about oxytocin and sex differences remain “very difficult” to state (Caldwell, 2018), and evidence that oxytocin responses and reactions can be more pronounced in males is not unusual (Dumais et al., 2016Feldman et al., 2010Herzmann et al., 2013Lynn et al., 2014Rilling et al., 2014Theodoridou et al., 2013). Nevertheless, our finding of higher salivary oxytocin concentrations in males contrasts with at least one other result in the human oxytocin literature. In their study of male and female medical staff at an Italian hospital (n = 90, 45 female), Marazziti et al. (2019) found that blood oxytocin concentrations were ~ 2.8 times higher in females compared to males. In considering this sex difference together with our own, it is apparent that neither Marazziti et al. nor the present study examined random samples. This leads to two hypotheses about the sex difference that we observed. One is that male chorists are a “special” group that exhibits higher levels of oxytocin activity (reflected in salivary concentrations). Modern choirs tend to be female-biased, and relative to males who are not in choirs, males in choirs tend to be more accepting of atypical gender behavior and identify less with gender stereotypes like “guys are physical” and “girls are feminine” (Nannen, 2017), potentially reflecting differences in trait empathy. A second hypothesis is that male oxytocin activity is specifically sensitive to variation in opportunities to affiliate with females. Our choir was 63% female (n = 71, 45 female) and young (mean age = 23). These circumstances may have led males to feel relatively safe and sociable in support of affiliation and courtship (some may have even have been “in love”). Although most previous endocrine studies of group singing have not included sex as predictor, some insight into evaluating these two hypotheses comes from Schladt et al. (2017). Despite looking, Schladt et al. did not find a significant sex difference in salivary oxytocin in their choir, which was more balanced in terms of participant sex at 55% female (n = 38, 21 female). This is a strike against the ‘male chorists are special’ hypothesis, but not necessarily the ‘affiliative opportunity’ hypothesis. If male chorists generally have higher levels of oxytocin activity, this should have been apparent also in Schladt et al. (2017), despite the more balanced choir. In contrast, if male oxytocin activity partially reflects male-female ratios that favor opportunities for males to affiliate with females, the absence of a sex difference in Schladt et al. (2017) would be predicted on the basis of their more balanced choir. Accordingly, we favor increased opportunities for males to affiliate with females as a preliminary explanation of the sex difference that we observed. However, we emphasize that the two hypotheses described above are not mutually exclusive, and that more research on salivary and blood oxytocin concentrations in the general population is needed before the sex differences found here and elsewhere can be interpreted in broader context.

4.4. Caveats

The most important caveat here concerns our evolving understanding of the oxytocin system and the aspects of its function that may be assessed in saliva. The principal neural sources of oxytocin—magnocellular and parvocellular neurons in the hypothalamus—exert their effects through at least three different release mechanisms. These include neurosecretory release from magnocellular neurons into peripheral circulation via projections to the neurohypophysis, somato-dendritic release into central extracellular and/or cerebrospinal fluid from magnocellular neurons for paracrine action in the central nervous system, and release from magnocellular and parvocellular projections onto specific central targets in the forebrain, hypothalamus, brainstem, and spinal cord (Jirikowski, 2019Jurek and Neumann, 2018Knobloch et al., 2012Lewis et al., 2020Ludwig and Leng, 2006Menon et al., 2018). These mechanisms may act alone or in concert with the others to exert a diversity of influences on social function in different contexts (Jurek and Neumann, 2018Landgraf and Neumann, 2004). Recent evidence from mice indicates that central synaptic release of oxytocin onto dopaminergic neurons in the reward system is particularly important for understanding affiliation in “consociate” (non-reproductive) relationships (Dölen et al., 2013Gunaydin et al., 2014Hung et al., 2017), which may thus provide an appropriate model for relationships between chorists in our study. If comparable central oxytocin projections are similarly key to human consociate affiliation, detection of their activity in saliva seems dubious. Related to this issue, it should be noted that oxytocin's mode of entry into saliva is poorly understood (MacLean et al., 2019Martin et al., 2018). Like other hydrophilic peptides, oxytocin does not easily cross the blood brain barrier. This can create differences in oxytocin concentrations between the central nervous system and the periphery (McEwen, 2004Neumann and Landgraf, 2012). Salivary oxytocin is typically interpreted as an indicator of peripheral levels—presumably on the basis of correlations with oxytocin concentrations in blood (Grewen et al., 2010White-Traut et al., 2009), and because saliva also lies outside the blood brain barrier—but the matter is contested. In particular, two studies conducted in critically ill patients have reached different conclusions, one finding moderate to strong correlations between oxytocin concentrations measured in saliva and cerebrospinal fluid (Martin et al., 2018), and the other indicating very weak correlations (except in the morning; Kagerbauer et al., 2019). The above points about oxytocin and the oxytocin system emphasize that our experimental effects on salivary oxytocin concentrations should be interpreted with caution, as their relation to central oxytocin mechanisms remains unclear. A second caveat concerns the absolute salivary oxytocin concentrations observed in our study (mean = 112 pg/ml across all 200 measurements; see Table 1). Although these concentrations are consistent with those reported by previous studies that have used the same Arbor Assays oxytocin ELISA (e.g., Akimoto et al., 2018Erickson et al., 2020Leeds et al., 2020), they are considerable higher than most results in the literature. For example, the absolute oxytocin concentrations in saliva measured using another popular oxytocin ELISA (originally produced by Assay Designs Inc., now Enzo Life Sciences Inc.) are more typically in the 5–20 pg/ml range (e.g., Blagrove et al., 2012Feldman et al., 2010Grewen et al., 2010Holt-Lunstad et al., 2011). The reason for the order of magnitude difference between concentrations reported using these two oxytocin ELISAs is unclear. They use different antibodies and protocols. Both now recommend and specify procedures for extracting samples, though it should be noted that these procedures are designed for blood rather than saliva. In a recent review of the challenges faced in measuring oxytocin, MacLean et al. (2019) propose that discrepant results in the literature are due to different methods of sample preparation and measurement being differentially sensitive to “diverse conformational states of the oxytocin molecule” (see also Brandtzaeg et al., 2016Gnanadesikan et al., 2021). This would indicate that differences in the absolute concentrations of oxytocin reported using different methods reflect real differences in oxytocin biochemistry, rather than the relative validity or invalidity of particular assays. In accord with this perspective, and in defense of our results, we provide the following rationale in support of our oxytocin data having value for examining relative concentration changes across our experimental conditions, despite ongoing debate over absolute values. First, we used the same procedure for all oxytocin measurements; second, the correlation between different measurements of the same participant was r = 0.83; and third, the patterns of change that we report are quite similar to those reported by several other recent studies of group singing of similar design (Fancourt et al., 2016Schladt et al., 2017). A final caveat is that singing and speaking may result in differences in salivary flow rate that could potentially influence measures of salivary oxytocin. Although we are not aware of any evidence that singing and speaking differ in this way, future endocrine studies of group singing that use salivary assays should record the volumes of collected samples so that flow rates may be determined.