A multiverse meta-analysis of oxytocin administration studies.

Kang H, Deilhaug E, Walle KM, Sartorius AI, Quintana DS

Biological psychology21 Aug 2025Cited by 1 (Europe PMC)

PMID40848964DOI10.1016/j.biopsycho.2025.109112PepBankPUB-40848964

Publisher page PubMedAsk about this study

Summary

Oxytocin administration is a valuable approach for experimentally increasing central oxytocin levels to investigate its impact on brain function and behavior. Despite considerable research interest, results have shown considerable variability across studies. Several meta-analyses have been conducted to help address this inconsistency. However, conducting a meta-analysis requires researchers to make numerous decisions, such as defining inclusion criteria and analytical techniques. In response to this challenge, a 'multiverse' approach can be used to simultaneously explore a range of different potential and realistic scenarios by systematically varying these analytical decisions. In this study, we performed a multiverse meta-analysis to investigate how researcher decisions can influence outcomes in meta-analyses of oxytocin administration studies. We systematically varied inclusion criteria, data synthesis models, and methods for correcting publication bias across 530 effect sizes derived from 185 studies, conducting 256 distinct meta-analyses. Our results demonstrate that summary effect estimates can vary considerably depending on meta-analytic decisions (from d = -0.16 to d = 1.45). Notably, meta-analyses with neurotypical populations tended to yield larger effect sizes than non-neurotypical populations, and multiple administrations studies tended to yield larger effect sizes than single administration studies. Furthermore, summary effect estimates can substantially vary according to the applied publication bias correction method. Our findings also demonstrated that the current evidence mostly supports the presence of the effect of oxytocin across domains, regardless of meta-analytic choices, as over 90% of the observed meta-analyses exceeded the range of bootstrapped meta-analyses assuming a null effect. These results warrant future research on how different contexts may impact the robustness of the effects of oxytocin, while underscoring the value of meta-analysis pre-registration to facilitate the transparent evaluation of meta-analyst decision-making.

CC BYText from Europe PMC. Highlighted sentences are the source of results.

Results from this study4

4 results from studies that gave the peptide. 4 results from people.

01HumanSystematic reviewEfficacy outcome
In a multiverse meta-analysis of oxytocin administration studies, summary effect estimates varied considerably with analytic decisions.
In plain wordsIn a review of oxytocin studies, the overall effect size changed a lot depending on how the analysis was set up.
Peptide
Oxytocin
Outcome
MixedSummary effect size range across meta-analytic decisions
Study
n: not reported in source
Outcome group
Social behavior and cognition · Nervous system and cognition
Studied dose
Dose not stated
Measurements
Summary effect size, lowest
−0.16 d
across meta-analytic choices
Summary effect size, highest
1.45 d
across meta-analytic choices
Source sentences
Summary
summary effect estimates can vary considerably depending on meta-analytic decisions (from d = -0.16 to d = 1.45).
1 context sentence(design, population, supporting numbers)
Summary · context
We systematically varied inclusion criteria, data synthesis models, and methods for correcting publication bias across 530 effect sizes derived from 185 studies, conducting 256 distinct meta-analyses.
Show in the study text Report a problemFND-4084896401
02HumanSystematic reviewEfficacy outcome
Meta-analyses of oxytocin administration in neurotypical populations tended to give larger effect sizes than in non-neurotypical populations.
In plain wordsOxytocin studies in neurotypical people tended to show larger effects than studies in non-neurotypical people.
Peptide
Oxytocin
Outcome
Not applicableEffect size in neurotypical versus non-neurotypical populations
Study
n: not reported in source
Comparator
Non-neurotypical populations
Outcome group
Social behavior and cognition · Nervous system and cognition
Studied dose
Dose not stated
Source sentences
Summary
meta-analyses with neurotypical populations tended to yield larger effect sizes than non-neurotypical populations
1 context sentence(design, population, supporting numbers)
Summary · context
We systematically varied inclusion criteria, data synthesis models, and methods for correcting publication bias across 530 effect sizes derived from 185 studies, conducting 256 distinct meta-analyses.
Show in the study text Report a problemFND-4084896402
03HumanSystematic reviewEfficacy outcome
Studies with multiple oxytocin administrations tended to give larger effect sizes than studies with a single administration.
In plain wordsStudies that gave oxytocin several times tended to show larger effects than studies that gave it once.
Peptide
Oxytocin
Outcome
Not applicableEffect size in multiple versus single administration studies
Study
n: not reported in source
Comparator
Single administration studies
Outcome group
Social behavior and cognition · Nervous system and cognition
Studied dose
Dose not stated
Source sentences
Summary
multiple administrations studies tended to yield larger effect sizes than single administration studies
1 context sentence(design, population, supporting numbers)
Summary · context
We systematically varied inclusion criteria, data synthesis models, and methods for correcting publication bias across 530 effect sizes derived from 185 studies, conducting 256 distinct meta-analyses.
Show in the study text Report a problemFND-4084896403
04HumanSystematic reviewEfficacy outcome
Over 90% of the observed meta-analyses exceeded the range of bootstrapped null-effect meta-analyses, so the evidence mostly supports an effect of oxytocin across domains.
In plain wordsMost analyses of oxytocin studies showed more effect than chance alone would give, whatever analysis choices were made.
Peptide
Oxytocin
Outcome
Not applicableMeta-analyses exceeding the null-effect range
Study
n: not reported in source
Outcome group
Social behavior and cognition · Nervous system and cognition
Studied dose
Dose not stated
Measurements
Meta-analyses exceeding null range
> 90%
bootstrapped null
Source sentences
Summary
the current evidence mostly supports the presence of the effect of oxytocin across domains, regardless of meta-analytic choices, as over 90% of the observed meta-analyses exceeded the range of bootstrapped meta-analyses assuming a null effect
1 context sentence(design, population, supporting numbers)
Summary · context
We systematically varied inclusion criteria, data synthesis models, and methods for correcting publication bias across 530 effect sizes derived from 185 studies, conducting 256 distinct meta-analyses.
Show in the study text Report a problemFND-4084896404

AI wrote and checked these results. AI can make mistakes, so check important results against the source.

Version history

PepBank keeps every version of the data on this study. Status changes from Retraction Watch, Europe PMC and Crossref are listed with their source.

Version 2Current
11 Oct 2026, 00:50 UTC
Study results updated
Version 1
9 Oct 2026, 02:47 UTC
First added · source update from Europe PMC

Subjects

Terms as the source lists them.

Publication types

MeSH terms

Keywords

PepBank is a research reference. Talk to a doctor before you use any peptide.