In Oxytocin We Trust

Oh oxytocin, you cuddly hug drug, you fine upstanding moral molecule, why are you so maligned by critics? That's because you're overrated, and misunderstood by those who look to you as a beacon of empathy, trust, love, peace, and prosperity. Sure, you're all about pair bonding in monogamous prairie voles — we have no beef with rigorous animal studies — but in humans, you're downright complicated. Yes, you can be magnanimous and romantic some of the time. But you're not always a moral molecule. You can promote antisocial behaviors such as envy and schadenfreude and aggressive tendencies. And even in voles, too much of a good thing can backfire.

Four new papers on oxytocin in humans have been published this week.¹ Collectively, there's something for nearly everyone to complain about:

• Intranasal oxytocin administration – assumptions about mechanism of action
• Ridiculously large numbers of variables – ripe for p-hacking
• Small n candidate gene studies – not replicable
• Studies of candidate gene × environment interactions – also not replicable
• Epigenetics – now nearly a woo word (The Trouble with Epigenetics, Part 1 and Part 2)
• Sex differences – studies often overly essentialist (although in this case, for good reason)

Oxytocin and Epigenetics

Of the four, the one that has garnered the most media attention is on epigenetics and sociability (Haas et al., 2016, published in PNAS). DNA methylation in the promoter region of the oxytocin gene (OXT) was quantified as an indicator of OXT expression. Lower methylation is presumably associated with higher OXT expression, and all sorts of sociable characteristics such as “more secure attachment styles, improved ability to recognize emotional facial expressions, greater superior temporal sulcus activity during two social-cognitive functional MRI tasks, and larger fusiform gyrus gray matter volume.”

Are these findings plausible from a mechanistic standpoint? Assuming that OXT expression was higher in the sociable sorts, which in turn assumes that methylation in saliva is a good proxy for expression in brain, how did “more“ oxytocin have all these effects? And on what time scale?

The Daily Mail was predictably credulous and hyperbolic,² using terms like “breakthrough”, “a ‘chatty’ gene which makes people sociable”, and [the kicker] “new treatments for autism”. The autism reference comes from the paper itself (and from the university press release):

"Participants with greater methylation of the OXT gene were less accurate in describing the emotional states of the people they saw in pictures," [first author Brian W. Haas] said. "That's a typical characteristic associated with autism, for example."

But it's not that simple. Let's look at the relationship between emotion recognition and OXT methylation. The task was to view 10 sec video clips of human faces morphing from neutral expressions to happy, sad, fearful and angry, and to identify the emotion as soon as it was detected. This led to 10 different dependent measures: reaction time and accuracy for each of the individual expressions, and for the mean of all expressions. The (conservative) Bonferroni corrected significance level is α = .05/10 = .005 [but the authors said it should be .025]. Overall accuracy is shown below.

Not all that impressive, eh? Another scatterplot was based on self-report questionnaires. The association between an anxious and insecure attachment style and OXT methylation fared better (p=.005), but the association between OXT methylation and avoidant attachment style was not significant. Why?

Silencing a gene promotes human sociability https://t.co/cUKaATnAnv "magical" omics problem strikes again @PNASNews pic.twitter.com/yXRxJCxRzg

— Eric Topol (@EricTopol) June 20, 2016

Oxytocin and Spirituality

The next paper, on oxytocin and spirituality (Van Cappellen et al., 2016), has gained traction on Twitter. Dan Quintana has already written an inspired blog post about it (Spray and pray: Does intranasal oxytocin increase spirituality?), so you should go and check it out. Dan has published at least 7 papers on oxytocin, so his critique is more informed than mine. I'll highlight his main points and then add a few of my own.

Good news:

First, the authors should be congratulated for posting the data for the paper on Open Science Framework (OSF). It’s great to see this dataset online considering the hype surrounding oxytocin...

Here’s a few other things I liked about this paper (or a list of things that oxytocin papers often don’t do): i) Effect sizes and confidence intervals are reported, ii) the alpha for the main outcome was adjusted for multiple tests, and iii) the placebo spray was a “true” placebo that contained all the same ingredients as the oxytocin spray, except the actual oxytocin (i.e, not just saline spray). It’s much easier to taste the difference between oxytocin and saline so this is an important point.

Bad news, genetics:

The authors included an “exploratory analysis” (their words in the intro and a section of the results) of three oxytocin pathway polymorphisms (rs53576, rs6449182, and rs3796863). There are about 10–15 candidate oxytocin pathway SNPs the could be analysed so it’s not clear why these three were chosen rather than others. Sure, rs53576 has been studied a lot, but so have many other oxytocin pathway SNPs (especially rs2254298).

Bad news, religious affiliation:

...it appears that there was a main effect of condition on both spirituality scales. However, a close read of table 1 reveals that this was after correcting for religious affiliation. Now this is reasonable when you consider that someone who’s an atheist is likely to report that spiritually is “not at all” important in my life. In fact, the data bears this out as the average spiritual rating (which can range from 0 to 7) for the atheist/agnostic group was 1.97 during the experimental visit and 1.88 a week later, whereas the average rating for the religiously affiliated group was 4.8 during the first visit and 4.9 during the second visit (I was able to calculate this from their posted dataset — isn’t open data great!).

It’s plausible for someone who identifies as agnostic or atheist to report “not at all” on both occasions — and many did. In fact, when you look at the agnostics/atheist group alone, there’s a statistically significant increase in spirituality after oxytocin compared to placebo both during the lab visit and 1 week later... However, there was no significant difference when assessing the religiously affiliated group.

@sarcastic_f the apostles were just brimming with oxytocin @edyong209 @WiringTheBrain @sTeamTraen @CoyneoftheRealm

— My Cousin Amygdala (@CousinAmygdala) June 20, 2016

(or not)

 Dr. Quintana concluded his post by advocating pre-registration and replication.


Oxytocin, Meditation, Positive Emotions, Negative Emotions, Oxytocin Receptor Gene (OXTR rs53576), CD38 (rs6449182 and rs3796863), Religious Affiliation, and Spirituality After Intranasal Oxytocin Administration in 83 Predominantly White Middle Aged Men

My unwieldy subheading includes only a small subset of the 161 variables in the study of Van Cappellen et al. (2016). Granted, some of these variables (e.g, the answers to individual items on questionnaires) were never examined in isolation — they were part of a composite score. Nonetheless, I think we can tick the “Ridiculously large numbers of variables” bullet point. We also have “Intranasal oxytocin administration” and “Small n candidate gene studies” (with n's below 20 in some cells). Bonus bullet point of “Between subjects design” is a personal pet peeve. I'd really like to see some within-subjects studies.

And there's a mysterious element to some of the data not included in this paper:

The data presented here are part of a larger study testing additional hypotheses not related to the present ones. For the larger study and to test a larger model, based on power calculation, a sample of 240 participants was targeted with a breakdown female-male of 125-115... Data collection ... stopped at 239 but despite recruitment effort, the sample is skewed toward females. This report focuses only on the 83 males who took part in the study... Growing evidence suggests that the effects of oxytocin are different for males and females (Feng et al., 2015) and most of the current evidence on intranasal oxytocin’s psychological effects, which support the current hypotheses, come from studies with exclusively male samples. A separate analysis of female participants, controlling for a series of additional variables related to natural variations in oxytocin is ongoing.

The larger study also included a task using Chinese pictographs, since the ability to read Chinese pictographs was an exclusionary criterion “applied to another task unrelated to the current investigation.” I'm generally not a study pre-registration evangelist, but one can really see the point here.

In Oxytocin We Doubt

I'll conclude on a pessimistic note (what else is new?). Some highly critical reviews of the oxytocin literature have appeared recently.

Evans SL, Dal Monte O, Noble P, Averbeck BB. Intranasal oxytocin effects on social cognition: a critique. Brain Res. 2014 Sep 11;1580:69-77.

Leng G, Ludwig M. Intranasal Oxytocin: Myths and Delusions. Biol Psychiatry. 2016 Feb 1;79(3):243-50.

Despite widespread reports that intranasal application of oxytocin has a variety of behavioral effects, very little of the huge amounts applied intranasally appears to reach the cerebrospinal fluid. However, peripheral concentrations are increased to supraphysiologic levels, with likely effects on diverse targets including the gastrointestinal tract, heart, and reproductive tract. The wish to believe in the effectiveness of intranasal oxytocin appears to be widespread and needs to be guarded against with scepticism and rigor. Preregistering trials, declaring primary and secondary outcomes in advance, specifying the statistical methods to be applied, and making all data openly available should minimize problems of publication bias and questionable post hoc analyses. Effects of intranasal oxytocin also need proper dose-response studies, and such studies need to include control subjects for peripheral effects, by administering oxytocin peripherally and by blocking peripheral actions with antagonists. Reports in the literature of oxytocin measurements include many that have been made with discredited methodology. Claims that peripheral measurements of oxytocin reflect central release are questionable at best.

--- there is a retort to Leng and Ludwig by Quintana and Woolley:  Intranasal Oxytocin Mechanisms Can Be Better Understood, but Its Effects on Social Cognition and Behavior Are Not to Be Sniffed At.

Walum H, Waldman ID, Young LJ. Statistical and Methodological Considerations for the Interpretation of Intranasal Oxytocin Studies. Biol Psychiatry. 2016 Feb 1;79(3):251-7.

...Our conclusion is that intranasal OT studies are generally underpowered and that there is a high probability that most of the published intranasal OT findings do not represent true effects. Thus, the remarkable reports that intranasal OT influences a large number of human social behaviors should be viewed with healthy skepticism...

McCullough ME, Churchland PS, Mendez AJ. Problems with measuring peripheral oxytocin: can the data on oxytocin and human behavior be trusted? Neurosci Biobehav Rev. 2013 Sep;37(8):1485-92.


It might be time to order Liquid Trust...

Footnotes

¹ Actually, the number of articles is closer to ten, but I'll just list these four for now.

1. Van Cappellen P, Way BM, Isgett SF, Fredrickson BL. Effects of Oxytocin Administration on Spirituality and Emotional Responses to Meditation. Soc Cogn Affect Neurosci. 2016 Jun 17. PMID: 27317929.
 
2. Haas BW, Filkowski MM, Cochran RN, Denison L, Ishak A, Nishitani S, Smith AK. Epigenetic modification of OXT and human sociability. Proc Natl Acad Sci. 2016 Jun 20. PMID: 27325757.
 
3. Simons RL, Lei MK, Beach SR, Cutrona CE, Philibert RA. Methylation of the oxytocin receptor gene mediates the effect of adversity on negative schemas and depression. Dev Psychopathol. 2016 Jun 20:1-12. PMID: 27323309.
 
4. Gao S, Becker B, Luo L, Geng Y, Zhao W, Yin Y, Hu J, Gao Z, Gong Q, Hurlemann R, Yao D, Kendrick KM. Oxytocin, the peptide that bonds the sexes also divides them. Proc Natl Acad Sci. 2016 Jun 20. PMID: 27325780.

² Hyperbolic and inaccurate. They called methyl groups "proteins".


References

Haas BW, Filkowski MM, Cochran RN, Denison L, Ishak A, Nishitani S, Smith AK. Epigenetic modification of OXT and human sociability. Proc Natl Acad Sci. 2016 Jun 20. PMID: 27325757.

Van Cappellen P, Way BM, Isgett SF, Fredrickson BL. Effects of Oxytocin Administration on Spirituality and Emotional Responses to Meditation. Soc Cogn Affect Neurosci. 2016 Jun 17. PMID: 27317929.


Further Reading

The Ed Yong Collection:

Oxytocin: the hype hormone

One Molecule for Love, Morality, and Prosperity?

Oxytocin: Still Not a Moral Molecule

The Weak Science Behind the Wrongly Named Moral Molecule

The Neurocritic Back Catalogue:

Oxytocin and Mind Reading...

LEARN TO TRUST YOUR OWN EYES

ABC News Says: 'Trust Drug' Oxytocin Unbelievable For Now

Paul Zak, Oxytocin Skeptic?

Your Brain on Coupons?

Neuroskeptic Posts:

More Doubts Over The Oxytocin And Trust Theory
(about: Failed Replication of Oxytocin Effects on Trust)

Psychologists Throw Open The “File Drawer”
(about: Is there a publication bias in behavioral intranasal oxytocin research on humans?)

Sofia Deleniv: The Dark Side of Oxytocin

Nature News: Neuroscience: The hard science of oxytocin

Old ad for Liquid Trust