Cognitive, Affective, & Behavioral Neuroscience (2022) 22:112–122 https://doi.org/10.3758/s13415-021-00936-3 RESEARCH ARTICLE Divergent effects of oxytocin on “mind-reading” in healthy males Ana Macchia 1,2 & Paul Theo Zebhauser 3,4 & Stephanie Salcedo 5 & Bethany Burum 5 & Edward Gold 6 & Miguel Alonso-Alonso 6 & Alvaro Pascual-Leone 7,8,9 & Daniel Gilbert 5 & Anna-Katharine Brem 1,6,10,11 Accepted: 15 July 2021 / Published online: 13 September 2021 # The Author(s) 2021 Abstract The neuropeptide oxytocin (OT) has been associated with a broad range of human behaviors, particularly in the domain of social cognition, and is being discussed to play a role in a range of psychiatric disorders. Studies using the Reading The Mind In The Eyes Test (RMET) to investigate the role of OT in mental state recognition reported inconsistent outcomes. The present study applied a randomized, double-blind, cross-over design, and included measures of serum OT. Twenty healthy males received intranasal placebo or OT (24 IU) before performing the RMET. Frequentist and Bayesian analyses showed that contrary to previous studies (Domes et al., 2007; Radke & de Bruijn, 2015), individuals performed worse in the OT condition compared to the placebo condition (p = 0.023, Cohen’s d = 0.55, 95% confidence interval [CI] [0.08, 1.02], BF10 = 6.93). OT effects did not depend on item characteristics (difficulty, valence, intensity, sex) of the RMET. Furthermore, OT serum levels did not change after intranasal OT administration. Given that similar study designs lead to heterogeneous outcomes, our results highlight the complexity of OT effects and support evidence that OT might even interfere with social cognitive abilities. However, the Bayesian analysis approach shows that there is only moderate evidence that OT influences mind-reading, highlighting the need for larger-scale studies considering the discussed aspects that might have led to divergent study results. Keywords Oxytocin . Theory of mind . RMET . Social cognition “Mind-reading” often is mentioned in the context of science component of social cognition are prominent in a range of fiction and referred to as a supernatural or telepathic ability to psychiatric disorders (Luyten & Fonagy, 2015). The ability quickly detect another's thoughts or feelings (Barrett et al., to simulate the internal state of others is commonly operation- 1882). In reality, humans lack such direct access to other alized using the "Reading the Mind in the Eyes Test" minds. Humans use hints to be able to "read" or "interpret" (RMET), a test that probes recognition of complex emotional the internal states of others, which provides the basis for ap- expressions (Baron-Cohen et al., 2001). Low-level perceptual propriate behavioral responses. Impairments in this core mechanisms, as well as emotion recognition and complex * Anna-Katharine Brem 6 Berenson-Allen Center for Noninvasive Brain Stimulation, Division katy.brem@gmail.com of Interventional Cognitive Neurology, Department of Neurology, Beth Israel Deaconess Medical Center, Harvard Medical School, 1 Max Planck Institute of Psychiatry, Munich, Germany Boston, MA, USA 7 2 Clinic for Psychiatry/Psychotherapy III, Ulm University, Guttmann Brain Health Institute, Institut Guttmann, Universitat Ulm, Germany Autonoma Barcelona, Barcelona, Spain 8 3 Department of Neurology, Klinikum Rechts der Isar der Technischen Department of Neurology, Harvard Medical School, Boston, MA, Universität München, Munich, Germany USA 9 4 Department of Psychiatry and Psychotherapy, Klinikum Rechts der Hinda and Arthur Marcus Institute for Aging Research and Center for Isar der Technischen Universität, Munich, Germany Memory Health, Hebrew SeniorLife, Boston, MA, USA 10 5 Department of Psychology, Harvard University, Cambridge, MA, University Hospital of Old Age Psychiatry, Murtenstrasse 21, USA 3008 Bern, Switzerland 11 Department of Neuropsychology, Lucerne Psychiatry, Lucerne, Switzerland Cogn Affect Behav Neurosci (2022) 22:112–122 113 mental state deduction, requiring theory of mind (ToM), have expressions. Nevertheless, the study indicated that individuals been suggested to be involved in the execution of the RMET with lower baseline empathy showed larger improvement in (Mitchell & Phillips, 2015). RMET-performance after OT administration. “Mind-reading” is only one in a range of human behaviors In sum, previous studies investigating the effect of OT on that have been associated with the neuropeptide oxytocin “mind-reading” in healthy males indicated that (1) effects of (OT), which further includes loving, trusting, caregiving, par- intranasally administered OT on “mind-reading” are inconsis- enting, sexual behavior, and aggression (de Jong & Neumann, tent, (2) if OT has any effect on “mind-reading” it might be 2017; Declerck et al., 2020; Hurlemann et al., 2010). Recent more pronounced for emotions that are difficult to interpret, findings indeed point toward a diverse portfolio of effects and and that (3) not all participants benefit from OT, but maybe a more general role of OT in social behavior (Harari-Dahan & especially those with lower baseline empathic abilities. Bernstein, 2014; Shamay-Tsoory & Abu-Akel, 2016). By op- The primary goal of the present study was to reexamine the erating as a hormone and neurotransmitter, oxytocin affects effect of OT on “mind-reading.” The present study employed the brain and has various effects on peripheral organs. In the a similar study design as previous studies by Domes et al. central nervous system, OT receptors can be mainly (2007) and Radke and de Bruijn (2015). To quantify evidence found in limbic areas and the basal forebrain of the for the effect of OT on “mind-reading,” we additionally pro- human brain (Baribeau & Anagnostou, 2015), emphasiz- vide Bayesian statistics for ours and previous studies (Domes ing the role of OT in emotional processing and regula- et al., 2007; Radke and de Bruijn, 2015). We investigated the tion (Tully et al., 2018). effect of OT versus placebo (PLC) performance on the RMET In recent years, a plethora of studies have investigated ef- in young healthy men, the association between empathic abil- fects of OT administration in healthy subjects and different ities and OT effects, and extended the design to include pe- psychiatric conditions on behavior, including ToM and emo- ripheral serum measures of OT. We included OT blood mea- tional processing. Mixed results have been reported in depres- sures to detect changes in OT blood levels after intranasal sion (MacDonald et al., 2013), schizophrenia (Brambilla et al., administration and examine possible associations with behav- 2016; Feifel et al., 2010), autism (Parker et al., 2017), eating ioral changes. disorders (Leppanen et al., 2017a), and addiction (Woolley et al., 2016). A meta-analysis that summarized the effect of intranasal OT on ToM in 14 studies concluded that neither healthy samples nor clinical populations benefited from OT Methods when interpreting complex emotions (Leppanen et al., 2017b). Problematically, there have been few replication attempts in Sample the field of intranasal OT research, and the few that have been implemented were largely unsuccessful (Declerck et al., 2020; A total of 24 healthy males (M = 22.85 years, SD = 3.5) were Mierop et al., 2020). There have been various considerations randomly assigned to receive 24 IU of intranasal OT to improve the precision of OT research. By including previ- (Syntocinon, Novartis, Basel, Switzerland) by taking three ous knowledge about a research field and providing evidence puffs per nostril (4 IU per puff) or PLC (saline solution that for or against a null hypothesis, Bayesian data analysis might contained all ingredients except the neuropeptide) on two be one of such tools to enhance the reproducibility of OT study visits that were scheduled one week apart. Further tasks research (Winterton et al., 2021). unrelated to our main research question were administered In a pioneering study, Domes et al. reported that OT im- after the end of the second visit and on a third visit (results proves the ability to decipher mental states using the RMET will be reported separately). Participants were recruited (Domes et al., 2007). The authors investigated healthy males through advertisements on university campuses and were in a double-blind, placebo-controlled, within-subject study, 20 compensated with $50 per session. Visits took place at the out of 30 individuals performed better on the RMET after Harvard Thorndike Clinical Center at Beth Israel Deaconess intranasal OT administration, with a mean increase of approx- Medical Center, Boston, MA. Exclusion criteria for participa- imately 3% of correct responses. These findings were corrob- tion were any current or past history of psychiatric illness, orated in a between-subject design, but in this study enhancing unstable medical condition, smoking, nasal pathologies, and effects of OT became only apparent in participants with lower current abuse or dependence in the past 6 months of drugs or baseline empathy and in more difficult items (Feeser et al., alcohol. Participants received instructions to abstain from food 2015). Subsequently, Radke and de Bruijn (2015) undertook for 9 hours, from drinks (other than water) for 2 hours, and a replication and extension of Domes’ et al. (2007) study and from alcohol and caffeine for 24 hours before study visits. All found no effect on “mind-reading,” even when considering study visits were scheduled to start at 9 am. Before each study item difficulty and carefully evaluating several other parame- visit participants were reminded of intake limitations via ters, such as valence, intensity, and sex of the presented facial phone calls. 114 Cogn Affect Behav Neurosci (2022) 22:112–122 The local Institutional Review Boards of Beth Israel are reported in the supplemental material. A research nurse Deaconess Medical Center and Harvard University approved instructed and supervised individuals on the administration the study. All participants gave written informed consent be- of OT. OT, as well as the PLC, were acquired through the fore the study onset according to the Declaration of Helsinki. same company (Victoria Pharmacy, Zurich, Switzerland) and presented in identical containers. Using appropriate pro- Instruments cedures, subjects are not able to detect differences between OT and PLC (MacDonald et al., 2011). After assessing vital The RMET measures how well people are able to “read” facial signs (oral temperature, blood pressure, heart rate), the nurse expressions by subtle social cues. It consists of photos of inserted the intravenous catheter (IV) and a distractor movie people’s eye regions depicting different mental states. showing nature scenes started. A total of 50 cc of blood was Participants have to choose the most appropriate mental state drawn before and after intranasal administration of either OT from a selection of four possible answers as quickly as possi- or placebo. Baseline blood draws began 15 minutes after ble. The computer-based revised English version (Baron- inserting the IV. In order to reduce baseline variability, the Cohen et al., 2001) consists of 36 items and one practice trial. baseline draw was divided into three draws of 10 cc spaced Stimuli were always presented in the same order and presen- 5 min apart providing us with a mean baseline OT level tation time was not time-limited. RMET items can be classi- (Lefevre et al., 2017). Immediately after the baseline blood fied in terms of the following variables: task difficulty (easy draw, participants received OT/PLC and after a 45-min break, vs. difficult), intensity (i.e., strength of the emotional state; the fourth blood draw (20 cc) took place before starting the low vs. high), valence (i.e., pleasantness of the emotional ex- tasks. Human interaction was kept to a minimum during the pression; positive vs. negative), and sex (female vs. male eye entire session. Blood draws were performed while subjects regions). To investigate mood state, the English version of the were reclining in a bed. Serum OT levels were measured fol- Multidimensional Mood Questionnaire (MDMQ) (Steyer lowing extraction using an enzyme immunoassay kit from et al., 1994) was administered after the RMET in both exper- Assay Designs (Ann Arbor, MI). The intra-assay coefficient imental conditions (OT, PLC). It contains 30 items with bipo- of variation (CV) was 8.7-12.4% and the inter-assay CV was lar dimensions assessing good and bad mood, alertness and 5.2-14.5%. The sensitivity was 7.0 pg/ml. tiredness, and calm and nervous states. Online questionnaires assessed self-reported trait empathy with the Interpersonal Reactivity Index (IRI) (Davis, 1980) and the Empathy Statistical analyses Quotient (EQ) (Baron-Cohen & Wheelwright, 2004). The IRI is a multidimensional tool to measure empathy consisting Data were analyzed using frequentist and Bayesian analysis of four subscales: a perspective taking scale, an empathic con- methods. Frequentist data analysis was performed with cern scale, a personal distress scale, and a fantasy scale (Davis, RStudio (R Version 3.5.1) and reanalyzed independently 1980). The EQ is a 60-item questionnaire with a maximum using IBM SPSS Statistics (Version 23.0) by two different score of 80 (Baron-Cohen & Wheelwright, 2004). Impulsivity authors. Bayesian data analysis was conducted with JASP was measured with the Barratt Impulsivity Scale (BIS) (Patton (JASP Team, 2020). Normality was examined with the et al., 1995). Thirty items are categorized into six first-order Shapiro-Wilk test (Tables S1-S3 in the supplemental material) factors, which build the second-order factors "attentional impul- and visual distributions. We used additional non-parametric siveness," "motor impulsiveness," and "nonplanning impulsive- methods for normality deviations. Results of tests dealing with ness." Participants, study procedures, and psychological tasks of non-normal data were comparable with parametric tests our study and previous studies are summarized in Table 1. (details are provided in the supplemental material). The sam- ple size decreased to 20 individuals due to drop out (n = 2) and technical problems (n = 2). One participant did not answer the Design online questionnaires but was included in the main analyses. Box and density plots are implemented to elucidate our data. Figure 1 shows the detailed study procedure. The study Compared with boxplots, density plots show smoother distri- followed a randomized, double-blind, cross-over design. bution of the data. It applies kernel density estimation to re- Participants answered questionnaires online through trieve the probability density function of the data. Moreover, Research Electronic Data Capture (REDCap) (Harris et al., density distributions determine the distribution shape and are 2009) after the second visit. Possible adverse events related useful to visualize differences between the two conditions. To to OT administration (irritation of the nasal mucosa, nausea, determine whether OT enhances RMET-performance, we per- headache, change in mood, allergic dermatitis, and other formed a two-sided paired t-test (OT vs. PLC). A sample of 20 events) were assessed using a questionnaire before and after participants in a within-participants design can only reliably each session. No significant adverse events occurred. Details detect (i.e., 80% power) an effect size of 0.66 or higher when Cogn Affect Behav Neurosci (2022) 22:112–122 115 Table 1 Characteristics of the current sample and study design compared to previous studies Current Radke & de Bruijn (2015) Domes et al. (2007) Sample (final analysis) 20 healthy males 24 healthy males 30 healthy males Mean age, yr (SD) 22.85 (±3.5) 21.5 (±1.9) 25.3 (±2.2) OT dose and visit interval 24IU, 1 week (administration 24IU, 2 weeks 24IU, 1 week order randomized) (administration order randomized) administration order balanced) RMET start (min after OT administration) 45 50 or 65 (counterbalanced) 45 Activities during waiting period (min after Watching movie with nature scenes Answering questionnaires; free to Reading news magazines administration) read/study/other occupations Country US Netherlands Germany Exclusion criteria Psychiatric disease; unstable medical Age <18 or >30 yr; neurological/ Medical/psychiatric disease; condition; smoking; nasal endocrine disease; medication use; medication use; smoking; pathologies; smoking >5 cigarettes/day; participation substance abuse drug/alcohol abuse (current abuse or in pharmacological study within 2 in the past 6 months) months before inclusion; sickness on test day (fever, cold, allergic rhinitis) Abstinence Food: 9 h Food: 2 h Food: 2 h Drink (except water): 2 h Drink (except water): 2 h Drink (except water): 2 h Alcohol/caffeine: 24 h Alcohol/caffeine: 24 h Alcohol/caffeine: 2 h Nicotine: exclusion criteria Nicotine: 24 h Nicotine: 2 h Recruitment Advertisement on university campus Advertisement on university campus Advertisement on university and online postings and online recruitment system campus Questionnaires EQ EQ MDMQ IRI IRI MDMQ VAS BIS STAXI SAS Incentive Financial compensation (50$) Financial compensation (50€) None BIS, Barratt Impulsiveness Scale; EQ, Empathy Quotient; IRI, Interpersonal Reactivity Index; MDMQ, Multidimensional Mood Questionnaire; RMET, Reading the Mind in the Eyes Test; STAXI, State-Trait Anxiety Inventory; SAS, Social Anxiety Scale; VAS, visual analogue scale for nonspecific effects of OT. performing a paired-samples t-test. In order to complement Williams (2018). We additionally analyzed the data using the frequentist results, we additionally provide Bayesian data “Oosterwijk prior” (t-distribution centered at 0.35, scaling fac- analysis. It allows making use of previous knowledge about tors of 0.102, 3 degrees of freedom). The “Oosterwijk prior” the research field and quantifies statistical evidence for and can be used if there is a lack of explicit prior knowledge and is against the alternative hypothesis (Etz & Vandekerckhove, considered to be a plausible prior in the field of psychology 2018). Bayesian paired samples t-test was applied to all three and biobehavioral science (Gronau et al., 2020; Quintana & studies (our study, Domes et al., 2007; Radke & de Bruijn, Williams, 2018; Stefan et al., 2019). 2015) using the Summary Stats module (Ly et al., 2018). As Separate (Bayesian) repeated-measure Analyses of prior distribution we used the average effect size of OT studies Variance (ANOVAs) for the potential moderating variables in healthy individuals (Cohen’s d = 0.28) as the central loca- task difficulty (easy vs. difficult), intensity (low vs. high), tion with a Cauchy scale of 0.1, as suggested by Quintana and valence (positive vs. negative), and sex (female vs. male eye Oxytocin acve Blood draw Online Oxytocin or Distractor Blood draw 4 RMET Consent 1+2+3 Quesonnaires Placebo movie (20 cc) MDMQ (10 cc each) Every 5 min 45 min Fig. 1 Detailed study procedure for each visit. After giving written, alone and were not allowed to use a mobile phone to minimize human informed consent, the IV was inserted and three baseline blood samples interaction. After the waiting period, a fourth blood sample (post) was (pre) were collected, spaced 5 min apart. Participants randomly received drawn. Then, the RMET and the Multidimensional Mood Questionnaire intranasal PLC or OT. While waiting 45 min for the OT effect to fully (MDMQ) (Steyer et al., 1994) were performed. Other questionnaires were emerge (Spengler et al., 2017), individuals watched a distractor movie administered via an online tool. Both the OT and the PLC visit followed showing nature scenes. During the waiting period, subjects were left the same procedure. 116 Cogn Affect Behav Neurosci (2022) 22:112–122 Table 2 Summary of statistics and descriptive for previous studies Current Radke et al. (2015) Domes et al. (2007) OT vs. PLC OT vs. PLC OT vs. PLC RMET in % t-test 72.92 (7.9) vs. 78.06 (7.1) 68.9 (12.6) vs.67.8 (12.6) 72.4 (8.6) vs. 69.4 (8.1) t(19) = 2.47, p = 0.023 (two-sided), t(23) = 0.465, p = 0.643 (two-sided), t(29) = −2.18, p = 0.019 (one-sided), d = 0.55, 95% CI [0.08, 1.02] d = 0.09, 95% CI [-0.31, 0.49] d = 0.40, 95% CI [−0.77, 0.02] BF10 (BF01) Informed prior: 6.93 (0.14); Informed prior: 0.60 (1.66); Informed prior: 3.29 (0.30); Mdn = 0.34; 95% CI [0.13, 0.75] Mdn = 0.24; 95% CI [−0.09, 0.44] Mdn = −0.30; 95% CI [−0.68, −0.02] Oosterwijk prior: 9.70 (0.10); Oosterwijk prior: 0.534 (1.87); Oosterwijk prior: 3.52 (0.28); Mdn = 0.39; 95% CI [0.18, 0.66] Mdn = 0.29; 95% CI [0.01, 0.48] Mdn = −0.22; 95% CI, [−0.60, −0.01] RMET item classification in % Easy vs. difficult 81.5 (±8.8) vs. 69.4 (±12.8) 73.6 (±2.5) vs. 63.1 (±2.6) N.A. Positive vs. negative 78.8 (±11.5) vs. 73.2 (±11.4) 72.8 (±3.0) vs. 63.4 (±2.4) N.A. High vs. low 80.4 (±8.7) vs. 70.6 (±11.4) 74.3 (±2.2) vs. 62.4 (±2.8) N.A. Female vs. male 78.4 (±10.0) vs. 72.9 (±10.0) 68.6 (±2.6) vs. 68.0 (±2.6) N.A. MDMQ Arousal 33.9 (±4.7) vs. 34.6 (±3.5) N.A. 34.3 (±4.7) vs. 34.0 (±4.8) Wakefulness 35.8 (±2.8) vs. 33.9 (±4.3) N.A. 28.8 (±6.8) vs. 28.2 (±7.1) Mood 30.8 (±3.3) vs. 31.4 (±3.1) N.A. 35.6 (±4.1) vs. 34.2 (±4.5) Questionnaires EQ 37.2 (±7.6) 33.8 (±10.8) N.A. IRI-PT 23.1 (±3.1) 22.7 (±5.2) N.A. IRI-EC 22.6 (±2.0) 21.9 (±2.8) N.A. IRI-F 22.2 (±4.6) 22.5 (±4.8) N.A. IRI-PD 18.4 (±3.2) 16.8 (±4.3) N.A. BIS-total 63.0 (±9.0) N.A. N.A. BIS-attentional 16.9 (±3.2) N.A. N.A. BIS-motor 23.1 (±3.7) N.A. N.A. BIS-nonplanning 23.0 (±4.7) N.A. N.A. Means with standard deviations except for t-test results. Effect size d according to Cohen (Cohen, 1988): small effect d = 0.2, medium effect d = 0.50, large effect d = 0.80. BF10: Bayes factor to quantify evidence for the alternative hypothesis relative to the null hypothesis; BF01: Bayes factor to quantify evidence for the null hypothesis relative to the alternative hypothesis; Mdn = median effect size and 95% credible intervals of the posterior distribution of the effect size; BIS, Barratt Impulsiveness scale; EQ, empathy quotient; IRI, Interpersonal Reactivity Index; IRI-PT, IRI perspective taking; IRI-EC, IRI empathic concern; IRI-F, IRI fantasy; IRI-PD, IRI personal distress; MDMQ, Multidimensional Mood Questionnaire; N.A., not available; OT, oxytocin; PLC, placebo; RMET, Reading the Mind in the Eyes Test. regions), as well as the experimental condition (OT vs. PLC) (Bayesian) Pearson correlations were applied to investi- were performed in order to obtain model comparisons. Item gate associations between empathy (EQ; IRI), impulsive- classification was obtained from Radke and de Bruijn who ness (BIS) and the RMET in both experimental conditions had followed earlier studies (Radke and de Bruijn, 2015). (PLC, OT). Additional partial correlations (controlling for We compared all possible models with main and interaction RMET performance in the PLC condition) were used to effects of item classification (difficulty, valence, intensity, investigate the relationship between empathy and RMET sex) and condition (OT, PLC) against the model providing performance. Pooled pre- (first, second, and third blood mea- the strongest evidence for the null hypothesis (i.e., null mod- sure) and post- (fourth blood measure drawn 45 min later) OT el). Each of the candidate models was considered equally serum measures were compared with a (Bayesian) paired- probable and therefore assigned to the same prior probability samples two-sided t-test. Additionally, (Bayesian) paired t- (i.e., 1/5 = 0.2). We preserved default Cauchy priors in JASP tests were used to determine whether the pre- and post-OT with a width factor of 0.5 for fixed effects, 1.0 for random serum levels differed between the conditions. No effects, and 0.354 for interaction effects. predetermined hypotheses were used to guide the analysis of Differences of the MDMQ in the OT and PLC condi- OT serum levels. Results of the blood level analysis are pro- tion were tested with (Bayesian) two-sided paired t-tests. vided in the supplemental material. Cogn Affect Behav Neurosci (2022) 22:112–122 117 Results over another. Using the Oosterwijik prior yields similar Bayes Factors for each study (our study, Domes et al., RMET 2007; Radke and de Bruijn, 2015) (see Table 2 for details). Plots including the posterior- and prior density Table 2 depicts our data in comparison to earlier studies. distribution of the effect sizes are presented in the supple- Unlike in previous studies, total RMET scores were signifi- mental material (Figures S1-S6). cantly reduced after OT administration compared with PLC administration, indicated by a medium effect , t(19) = 2.47, p RMET item classification = 0.023, d = 0.55, 95% confidence interval [CI] [0.08, 1.02] (Table 2; Figure 2). Considering an average effect size of d = Repeated measure ANOVAs revealed main effects of condi- 0.28 of OT studies in healthy individuals (Walum et al., 2016), tion (OT vs. PLC) and most item classifications (task difficul- our data show that the alternative hypothesis (i.e., OT influ- ty, task intensity, sex), except for valence (Table S4, ences mind-reading) is 6.93 times more likely than the null supplemental material). In both conditions, participants model (i.e., there is no effect of OT on mind-reading). This is achieved higher RMET scores in easier compared with more in line with the frequentist data analysis and can be interpreted difficult items (F(1, 137.5) = 19.83, p ≤ 0.001, ηp2 = 0.51), in as moderate evidence for a negative effect of OT on “mind- items with higher intensity compared to items with lower in- reading” according to the classification of Lee and tensity (F(1, 137.5) = 24.32, p ≤ 0.001, ηp2 = 0.56), and in Wagenmakers (2013; adjusted from Jeffreys, 1961). Domes’ items that depicted a female rather than a male eye region et al. (2007) study shows that it is 3.29 times more plausible (F(1, 135.8) = 8.14, p = 0.010, ηp2 = 0.30). None of the that there is a positive effect of OT on “mind-reading” than a condition-classification-interactions reached significance (all null-effect, which also coincides with the frequentist analysis. ps > 0.30), suggesting that OT does not affect the recognition In contrast, Radke’s study provides anecdotal evidence for a of mental states depending on difficulty, valence, intensity, null-effect (BF10 = 1.67) and does not favor one hypothesis and sex of the presented facial expression. Every Bayesian ANOVA compared five models: (1) the null model, (2) the model with condition (PLC vs. OT), (3) the model with task classification (difficulty, valence, intensity, sex), (4) the model with both main effects (condition, task classification), and (5) the model with both main effects (condition, task classifica- tion), and their interaction (Table S5, supplemental material). For the model comparisons, including task difficulty, and in- tensity, the models, including the two main effects—condition and task classification—provide the strongest support against the null model. Regarding the task valence and the item clas- sification of sex, the models with the main effect of condition provided the strongest support, although the evidence favor- ing these models was small. Mood state The subdomain Wakefulness was significantly higher in the OT condition, t(19) = −2.17, p = 0.043, d = 0.49, 95% CI [0.02, 0.94], BF10 = 0.11, Mdn = −0.44; 95% CI [−0.88, −0.01]. However, according to the Bayes factor it is 8.8 (BF01) times more likely that there was no difference in Wakefulness between the conditions. Notably, participants in the current study scored higher in Wakefulness in both the OT and the PLC study visit compared with Domes’ et al. study (OT: t(48) = −4.35, p < 0.001, d = 1.27, 95% CI [−0.64, 1.87], BF10 = 2.76, Mdn = −0.99; 95% CI [−1.67, −0.18]), for Fig. 2 Total RMET scores in the OT and PLC conditions. Boxplots (top) PLC: t(48) = −3.21, p = 0.002, d = 0.93, 95% CI [0.33, 1.52], and Density distributions (bottom) indicate that participants receiving OT performed worse in the RMET. Density plots show that RMET scores in BF 10 = 0.27, Mdn = −0.45; 95% CI [−1.20, 0.30]). the OT condition are concentrated on a lower interval compared with the Correlational analysis of the suspected association between PLC condition Wakefulness and the RMET score in the PLC and OT 118 Cogn Affect Behav Neurosci (2022) 22:112–122 condition did not reveal a significant association in the current Discussion study (PLC: r = −0.06, p = 0.786, 95% CI [−0.39, 0.49], BF10 = 0.29; OT: r = −0.30, p = 0.195, 95% CI [−0.65, 0.16], BF10 We investigated the influence of OT on “mind-reading,” as = 0.61). The MDMQ subdomains Mood and Arousal did not assessed with the RMET, explored its relationship with em- differ significantly between PLC and OT study visits (all ps > pathic abilities, and extended our results by including periph- 0.28, all BFs10 > 1.42, see Table 2 for descriptive statistics). eral serum OT measures. Opposed to prior results that report- ed an improvement of mentalizing abilities (Domes et al., 2007) or null-effects (Radke & de Bruijn, 2015), our findings Trait empathy point toward an impairing effect of OT. Individuals performed significantly worse after administration of intranasal OT com- None of the online questionnaires measuring trait empathy pared with PLC. Specifically, we found a Bayes factor of 6.93, (EQ, IRI) or impulsiveness (BIS) were related to the RMET meaning that there is 6.93 times more evidence for the alter- score in the PLC condition (all rs < |0.33|, all ps > 0.17, all native hypothesis that OT influences mind-reading. Domes` BFs10 < 0.68 ). In the OT condition, the IRI subdomain et al. (2007) study shows that the alternative hypothesis is 3.29 Perspective Taking (IRI-PT) correlated negatively with the times more likely than the null model providing evidence for a RMET score (r = −0.63, p = 0.004, 95% CI [−0.84, −0.24], positive effect of OT on “mind-reading,” whereas Radke and BF10 = 13.41), indicating that subjects with higher self- de Bruijn’s (2015) study does not favor one hypothesis over reported perspective taking abilities achieved a lower RMET another. Furthermore, OT did not affect mentalizing abilities score in the OT condition (Figure 3). depending on item classification, which is in accordance with This was accompanied by a Bayes factor of 13.41, indicat- Radke and de Bruijn’s (2015) findings and is opposed to other ing strong evidence that there exists an association between studies (Domes et al., 2007; Feeser et al., 2015). However, perspective taking as measured by the IRI and the RMET RMET item classification influenced participant’s accuracy score in the OT condition. Except for very small prior widths, independently of OT effects as in Radke and Bruijn’s study the robustness check shows that the Bayes factor is relatively (2015), showing that participants found it more difficult to stable across different prior widths (supplemental material, label difficult emotions, less intense emotions, and emotions Figure S7). When controlling for PLC RMET performance expressed by a male’s eye region. Furthermore, we found no in a partial correlation, the IRI-PT scale remained significantly significant change from pre- to post-OT serum levels after correlated with the RMET (r = −0.63, p = 0.004, 95% CI intranasal OT administration. [−0.85, −0.23]). Other scales of the IRI (IRI-EC, IRI-F, IRI- By reporting negative effects of OT on “mind-reading,” PD), the EQ, and the BIS were not significantly related to this study further reinforces uncertainties regarding OT effects RMET performance in the OT condition (all rs < |0.40|, all in the field of empathy research. In the following, we discuss ps > 0.09, all BFs10 < 1.10). several factors that might explain diverging and converging results in the three studies that investigated this question (our study, Domes et al., 2007; Radke and de Bruijn, 2015) and subsequently present more general considerations. (1) Study populations and sample size: our sample com- prised a mixed population compared with student-only popu- lations in the other two studies and our sample size was sim- ilarly small. Small sample sizes have been generally criticized in studies that investigate intranasal OT effects (Walum et al., 2016). However, despite the small sample size, we report a medium-effect size (d = 0.55), which can be considered quite large compared to the average effect sizes (d = 0.28) reported in the OT literature (Walum et al., 2016). Additionally, the finding of this study fits the conclusion of a recent systematic review, which showed that statistical significance is generally unrelated to sample size in intranasal OT research (Mierop Fig. 3 Correlation of the RMET score and perspective taking (IRI-PT). et al., 2020). Scatterplot with regression lines for the RMET score in the PLC (red dots, (2) Nonspecific OT effects: in our study, but not Domes solid line) and the OT (blue dots, dotted line) condition and their relation et al. (2007) study, participants reported higher wakefulness in to the perspective taking scale of the IRI (IRI-PT). In the OT but not the PLC condition, the RMET score was negatively correlated with the OT compared to the PLC condition. Wakefulness is asso- perspective taking. Note that one participant did not answer the online ciated with the activity of an evolutionarily conserved arousal questionnaires. system that regulates different neural populations and Cogn Affect Behav Neurosci (2022) 22:112–122 119 neurotransmitter secretion (de Lecea et al., 2012) and could empathic abilities, personality traits, degree of psychopathol- possibly interact with OT effects. Differences in wakefulness ogy, or the present psychological state (Shamay-Tsoory & could therefore have led to differential results in the two stud- Abu-Akel, 2016). Our sample showed high baseline empathic ies. Nevertheless, we could not confirm a direct link between abilities and the BIS impulsivity score was consistent with RMET performance and wakefulness. results found in healthy males (Stanford et al., 2009). In so- (3) Trait empathic abilities: in line with Radke and de cially proficient individuals, OT application possibly causes Bruijn’s (2015) study, our findings (EQ; IRI) were not related an impairing effect on “mind-reading” by inducing oversensi- to the RMET score in the PLC condition. However, we found tivity to emotions and thereby impairing their accurate classi- evidence for a negative relationship between Perspective fication. Lastly, we think that further nonspecific effects could Taking (IRI-PT) and the RMET score after OT administration, have influenced task performance. For instance, participants suggesting that OT impaired mind-reading in individuals with in our study were not allowed to eat before study onset for a higher perspective taking skills. A similar negative relation longer time window compared with the other studies. was found in Radke and de Bruijn’s study for the IRI subscale Moreover, they received financial compensation and were Empathic Concern (IRI-EC). Differences in empathy scores more wakeful. Because OT is associated with a broad range between studies might partially explain contradictory effects of human behaviors, such additional confounding variables of OT on RMET scores in the different studies. It has been could potentially explain these divergent study results. reported that individuals with low baseline empathic abilities Beyond that, more general considerations have to be are more likely to profit from exogenous OT (Feeser et al., taken into account. Psychological tasks, such as the 2015). Notably, our participants showed higher self-measured RMET, might not be an ideal measure to assess the effect empathy and RMET scores than subjects participating in of OT on social cognition. Considering that there is Radke and de Bruijn’s (2015) study. OT modulates its own context-dependent variability of OT effects (Bartz et al., release resulting in a non-linear relationship or a saturation 2011), mental state recognition of faces presented out of curve that could be related to impairing effects of OT in indi- context and without movement or body language might viduals with high empathic abilities. Such impairing effects differ entirely from mental state recognition in a real so- might reflect a disturbance of homeostatic processes. cial context. Moreover, the RMET depends on language Furthermore, the finding that the EQ is not related to the abilities (Olderbak et al., 2015; Peterson & Miller, 2012) RMET performance is consistent with earlier findings and higher-order executive brain functions (Launay et al., (Launay et al., 2015; Radke & de Bruijn, 2015). 2015; Mitchell & Phillips, 2015), which again might be (4) Waiting period: Keeping human interaction to a mini- related to previously mentioned sample characteristics mum possibly influenced the OT effect in our study, as posi- (see discussion point 1). Studies investigating the effects tive and negative social stimuli likely affect the OT system of OT on other ToM measures also report mixed findings. (Neumann & Landgraf, 2012). This was handled differently in Woolley et al. (2014) included “The awareness of Social other studies; for example, participants in Radke and de Inference Test,” which is a dynamic emotion recognition Bruijn’s study (2015) completed questionnaires or could task consisting of several short videoclips. The authors study, read, or engage in other activities, while participants found no improvement of emotion recognition after intra- in Domes’ et al. study (2007) could read news magazines. nasal OT administration. Another study reported Recent overarching models of the effects of OT on human impairing effects of OT on emotion recognition using behavior might help us to understand the impairing effects of t h e “Pe rce ivi ng Emo tio ns ” an d “ U nd er s t an di n g OT on “mind-reading.” First, the strictly nonsocial (i.e., keep- Emotions” tasks of the Mayer-Salovey-Caruso Emotional ing social interaction to a minimum) and even threatening Intelligence Test. Participants receiving intranasal OT rat- environment (intravenous blood draw) could have negatively ed facial emotions as more intense than those receiving influenced the effect of OT. According to the allostatic theory PLC. This effect was associated with a decrease in accu- (Quintana & Guastella, 2020), OT facilitates stability in racy (Cardoso et al., 2014). By contrast, Bartz et al. changing environments to promote survival. Our study setting (2019) found that intranasal OT improves empathic accu- possibly induced an adaptive response to the threat, hereby racy of participants who rated the valence of feelings of promoting nonsocial behavior in terms of withdrawal and a persons discussing emotional autobiographic events. This reduced ability of mental state recognition. Similarly, this effect was limited to men scoring higher on autistic traits. study setting could also explain the negative findings within In accordance with our study, the authors report slightly the general approach-avoidance model, suggesting that OT worse task performance for socially proficient men after motivates approach to rewards and reinforces avoidance to OT administration. threats (Harari-Dahan & Bernstein, 2014). Finally, the social Studies investigating OT are usually limited to male sam- salience hypothesis states that OT modulates the salience of ples to keep hormonal variance low. However, males show social stimuli depending on interindividual properties, such as differential brain activation patterns compared to females after 120 Cogn Affect Behav Neurosci (2022) 22:112–122 OT administration (Domes et al., 2007; Tully et al., 2018) and Acknowledgements The authors thank Gregor Domes and Sina Radke for providing detailed information about their studies, including the higher levels of testosterone can have an opposing action to RMET item classification, additional details on participant recruitment, OT (Crespi, 2016). In healthy women, the effect of intranasal and activities during the waiting period. OT might be entirely different. Moreover, it needs to be considered that the effect of OT on Author contributions SS, BB, MA, APL, DG, and AKB designed the RMET performance could simply reflect a regression to the study. AKB and EG collected the data. AM and PZ analyzed the data. AM and PZ prepared the initial draft report. AKB edited the report, which mean. Statistical variation could be mistaken as real change was critically reviewed by all authors. AM, PZ, and AKB had full access and lead to the incorrect conclusion that OT influences “mind- to all the data in the study and take responsibility for the integrity of the reading.” data and the accuracy of the data analysis. Even though intranasal OT can produce effects in brain physiology, there is limited evidence that it can reliably ma- Funding Open Access funding enabled and organized by Projekt DEAL. This study was supported by an award of the Mind/Brain/Behavior nipulate behavioral effects (Winterton et al., 2021). Interfaculty Initiative of Harvard University. AKB was moreover sup- Furthermore, it remains unclear whether peripheral OT can ported by the Young Academics Support and the Stiefel-Zangger be used as a valid and reliable marker of OT circulation. Foundation of the University of Zurich, Switzerland, and the Swiss Regarding OT plasma levels, it has been suggested that they National Foundation (PBZHP1_147196). are not representatively mirroring levels in the brain even though increased plasma OT levels have been reported after Declaration of interest None. intranasal OT application (Gossen et al., 2012; Striepens et al., 2013). Open Access This article is licensed under a Creative Commons We found no significant change of OT serum levels after Attribution 4.0 International License, which permits use, sharing, adap- intranasal application, which might be related to methodolog- tation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, pro- ical issues. Recent publications report that different OT assays vide a link to the Creative Commons licence, and indicate if changes were lack reliability and yield widely varying OT values (Lefevre made. The images or other third party material in this article are included et al., 2017; Poljak & Sachdev, 2021). In this study, baseline in the article's Creative Commons licence, unless indicated otherwise in a variability in OT levels was high and mostly exceeded the credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by typical range found in healthy males when applying radioim- statutory regulation or exceeds the permitted use, you will need to obtain munoassay methods. Therefore, we postulate a cautious inter- permission directly from the copyright holder. To view a copy of this pretation on the association of OT serum levels with behav- licence, visit http://creativecommons.org/licenses/by/4.0/. ioral measures. References Conclusions Baribeau, D. A., & Anagnostou, E. (2015). Oxytocin and vasopressin: Linking pituitary neuropeptides and their receptors to social neurocircuits. Frontiers in Neuroscience, 9, 335. https://doi.org/10. As opposed to previous results, we found a moderate 3389/fnins.2015.00335 impairing effect on “mind-reading”-performance after OT ad- Baron-Cohen, S., & Wheelwright, S. (2004). The empathy quotient: An ministration. Our findings add to the heterogeneity of earlier investigation of adults with asperger syndrome or high functioning autism, and normal sex differences. Journal of Autism and findings and provide evidence that OT might even interfere Developmental Disorders, 34(2), 163–175. https://doi.org/10. with social cognitive abilities. This emphasizes the impor- 1023/B:JADD.0000022607.19833.00 tance of identifying intra- and interindividual factors that de- Baron-Cohen, S., Wheelwright, S., Hill, J., Raste, Y., & Plumb, I. (2001). termine OT effects. Specifically, different sample characteris- The “Reading the Mind in the Eyes” test revised version: A study with normal adults, and adults with asperger syndrome or high- tics, including baseline empathic abilities as well as differ- functioning autism. Journal of Child Psychology and Psychiatry, ences in study designs, might explain divergent results regard- 42(2), 241–251. https://doi.org/10.1111/1469-7610.00715 ing OT-effects on RMET-performance. Future studies should Barrett, W. F., Gurney, E., & Myers, F. W. (1882). First report on investigate individual hormonal, situational, and cognitive thought-reading. Proceedings of the Society for Psychical factors that determine OT effects on behavioral outcomes. Research, 1, 13–34. Bartz, J. A., Zaki, J., Bolger, N., & Ochsner, K. N. (2011). Social effects Furthermore, a better understanding of pharmacokinetic dy- of oxytocin in humans: Context and person matter. Trends in namics of different pathways into the central nervous system Cognitive Sciences, 15(7), 301–309. https://doi.org/10.1016/j.tics. and the relationship between peripheral and central OT levels 2011.05.002 is key to understand behavioral effects. Bartz, J. A., Nitschke, J. P., Krol, S. A., & Tellier, P.-P. (2019). Oxytocin Selectively Improves Empathic Accuracy: A Replication in Men and Novel Insights in Women. Biological Psychiatry: Cognitive Supplementary Information The online version contains supplementary Neuroscience and Neuroimaging, S2451902219300473. https:// material available at https://doi.org/10.3758/s13415-021-00936-3. doi.org/10.1016/j.bpsc.2019.01.014 Cogn Affect Behav Neurosci (2022) 22:112–122 121 Brambilla, M., Manenti, R., de Girolamo, G., Adenzato, M., Bocchio- socially reinforced learning and emotional empathy in humans. Chiavetto, L., & Cotelli, M. (2016). Effects of intranasal oxytocin on Journal of Neuroscience, 30(14), 4999–5007. https://doi.org/10. long-term memory in healthy humans: A systematic review: 1523/JNEUROSCI.5538-09.2010 Oxytocin and memory. Drug Development Research, 77(8), 479– JASP Team. (2020). JASP (Version 0.14) [Computer software]. https:// 488. https://doi.org/10.1002/ddr.21343 jasp-stats.org/ Cardoso, C., Ellenbogen, M. A., & Linnen, A.-M. (2014). The effect of Launay, J., Pearce, E., Wlodarski, R., van Duijn, M., Carney, J., & intranasal oxytocin on perceiving and understanding emotion on the Dunbar, R. I. M. (2015). Higher-order mentalising and executive Mayer-Salovey-Caruso Emotional Intelligence Test (MSCEIT). functioning. Personality and Individual Differences, 86, 6–14. Emotion, 14(1), 43–50. https://doi.org/10.1037/a0034314 https://doi.org/10.1016/j.paid.2015.05.021 Cohen, J. (1988). Statistical power analysis for the social sciences. Lee, M. D., & Wagenmakers, E.-J. (2013). Bayesian cognitive modeling: Lawrence Erlbaum A practical course. Cambridge university press. Crespi, B. J. (2016). Oxytocin, testosterone, and human social cognition: Lefevre, A., Mottolese, R., Dirheimer, M., Mottolese, C., Duhamel, J.-R., Oxytocin and social behavior. Biological Reviews, 91(2), 390–408. & Sirigu, A. (2017). A comparison of methods to measure central https://doi.org/10.1111/brv.12175 and peripheral oxytocin concentrations in human and non-human Davis, M. H. (1980). A multidimensional approach to individual differ- primates. Scientific Reports, 7(1), 1–10. https://doi.org/10.1038/ ences in empathy. JSAS Catalog of Selected Documents in s41598-017-17674-7 Psychology, 10, 85. Leppanen, J., Cardi, V., Ng, K. W., Paloyelis, Y., Stein, D., Tchanturia, de Jong, T. R., & Neumann, I. D. (2017). Oxytocin and aggression. In R. K., & Treasure, J. (2017a). Effects of intranasal oxytocin on the Hurlemann & V. Grinevich (Eds.), Behavioral Pharmacology of Interpretation and Expression of Emotions in Anorexia Nervosa. Neuropeptides: Oxytocin (Vol. 35, pp. 175–192). Springer Journal of Neuroendocrinology, 29(3). https://doi.org/10.1111/jne. International Publishing. https://doi.org/10.1007/7854_2017_13 12458 de Lecea, L., Carter, M. E., & Adamantidis, A. (2012). Shining Light on Leppanen, J., Ng, K. W., Tchanturia, K., & Treasure, J. (2017b). Meta- Wakefulness and Arousal. Biological Psychiatry, 71(12), 1046– analysis of the effects of intranasal oxytocin on interpretation and 1052. https://doi.org/10.1016/j.biopsych.2012.01.032 expression of emotions. Neuroscience & Biobehavioral Reviews, Declerck, C., Boone, C., Pauwels, L., Vogt, B., & Fehr, E. (2020). A 78, 125–144. https://doi.org/10.1016/j.neubiorev.2017.04.010 registered replication study on oxytocin and trust. Nature Human Luyten, P., & Fonagy, P. (2015). The neurobiology of mentalizing. Behaviour, 4(6), 646–655. https://doi.org/10.6084/M9.FIGSHARE. Personality Disorders: Theory, Research, and Treatment, 6(4), 11980368 366–379. https://doi.org/10.1037/per0000117 Domes, G., Heinrichs, M., Michel, A., Berger, C., & Herpertz, S. C. Ly, A., Raj, A., Etz, A., Marsman, M., Gronau, Q. F., & Wagenmakers, (2007). Oxytocin Improves “Mind-Reading” in Humans. E.-J. (2018). Bayesian reanalyses from summary statistics: A guide Biological Psychiatry, 61(6), 731–733. https://doi.org/10.1016/j. for academic consumers. Advances in Methods and Practices in biopsych.2006.07.015 Psychological Science, 1(3), 367–374. https://doi.org/10.1177/ Etz, A., & Vandekerckhove, J. (2018). Introduction to bayesian inference 2515245918779348 for psychology. Psychonomic Bulletin & Review, 25(1), 5–34. MacDonald, E., Dadds, M. R., Brennan, J. L., Williams, K., Levy, F., & https://doi.org/10.3758/s13423-017-1262-3 Cauchi, A. J. (2011). A review of safety, side-effects and subjective Feeser, M., Fan, Y., Weigand, A., Hahn, A., Gärtner, M., Böker, H., re a c t i o n s t o i n t ra n a s a l o x y t o c i n i n h u m an r es e a r c h . Grimm, S., & Bajbouj, M. (2015). Oxytocin improves mentalizing Psychoneuroendocrinology, 36(8), 1114–1126. https://doi.org/10. – Pronounced effects for individuals with attenuated ability to em- 1016/j.psyneuen.2011.02.015 pathize. Psychoneuroendocrinology, 53, 223–232. https://doi.org/ MacDonald, K., MacDonald, T. M., Brüne, M., Lamb, K., Wilson, M. P., 10.1016/j.psyneuen.2014.12.015 Golshan, S., & Feifel, D. (2013). Oxytocin and psychotherapy: A Feifel, D., Macdonald, K., Nguyen, A., Cobb, P., Warlan, H., Galangue, pilot study of its physiological, behavioral and subjective effects in B., Minassian, A., Becker, O., Cooper, J., Perry, W., Lefebvre, M., males with depression. Psychoneuroendocrinology, 38(12), 2831– Gonzales, J., & Hadley, A. (2010). Adjunctive intranasal oxytocin 2843. reduces symptoms in schizophrenia patients. Biological Psychiatry, Mierop, A., Mikolajczak, M., Stahl, C., Béna, J., Luminet, O., Lane, A., 68(7), 678–680. https://doi.org/10.1016/j.biopsych.2010.04.039 & Corneille, O. (2020). How can intranasal oxytocin research be Gossen, A., Hahn, A., Westphal, L., Prinz, S., Schultz, R. T., Gründer, G., trusted? A systematic review of the interactive effects of intranasal & Spreckelmeyer, K. N. (2012). Oxytocin plasma concentrations oxytocin on psychosocial outcomes. Perspectives on Psychological after single intranasal oxytocin administration – A study in healthy S c i e n ce , 1 5 ( 5 ) , 1 2 2 8 – 1 2 4 2 . h t t p s : / / d o i . o r g / 1 0 . 1 1 7 7 / men. Neuropeptides, 46(5), 211–215. https://doi.org/10.1016/j. 1745691620921525 npep.2012.07.001 Mitchell, R. L. C., & Phillips, L. H. (2015). The overlapping relationship Gronau, Q. F., Ly, A., & Wagenmakers, E.-J. (2020). Informed bayesian between emotion perception and theory of mind. Neuropsychologia, t-tests. The American Statistician, 74(2), 137–143. https://doi.org/ 70, 1–10. https://doi.org/10.1016/j.neuropsychologia.2015.02.018 10.1080/00031305.2018.1562983 Neumann, I., & Landgraf, R. (2012). Balance of brain oxytocin and Harari-Dahan, O., & Bernstein, A. (2014). A general approach-avoidance vasopressin: Implications for anxiety, depression, and social behav- hypothesis of oxytocin: Accounting for social and non-social effects iors. Trends in Neurosciences, 35(11), 649–659. https://doi.org/10. of oxytocin. Neuroscience & Biobehavioral Reviews, 47, 506–519. 1016/j.tins.2012.08.004 https://doi.org/10.1016/j.neubiorev.2014.10.007 Olderbak, S., Wilhelm, O., Olaru, G., Geiger, M., Brenneman, M. W., & Harris, P. A., Taylor, R., Thielke, R., Payne, J., Gonzalez, N., & Conde, J. Roberts, R. D. (2015). A psychometric analysis of the reading the G. (2009). Research electronic data capture (REDCap)—A mind in the eyes test: Toward a brief form for research and applied metadata-driven methodology and workflow process for providing settings. Frontiers in Psychology, 6. https://doi.org/10.3389/fpsyg. translational research informatics support. Journal of Biomedical 2015.01503 Informatics, 42(2), 377–381. https://doi.org/10.1016/j.jbi.2008.08. Parker, K. J., Oztan, O., Libove, R. A., Sumiyoshi, R. D., Jackson, L. P., 010 Karhson, D. S., Summers, J. E., Hinman, K. E., Motonaga, K. S., Hurlemann, R., Patin, A., Onur, O. A., Cohen, M. X., Baumgartner, T., Phillips, J. M., Carson, D. S., Garner, J. P., & Hardan, A. Y. (2017). Metzler, S., Dziobek, I., Gallinat, J., Wagner, M., Maier, W., & Intranasal oxytocin treatment for social deficits and biomarkers of Kendrick, K. M. (2010). Oxytocin enhances amygdala-dependent, response in children with autism. Proceedings of the National 122 Cogn Affect Behav Neurosci (2022) 22:112–122 Academy of Sciences, 114(30), 8119–8124. https://doi.org/10.1073/ administration in humans. Scientific Reports, 3(1). https://doi.org/ pnas.1705521114 10.1038/srep03440 Patton, J. H., Stanford, M. S., & Barratt, E. S. (1995). Factor structure of Tully, J., Gabay, A. S., Brown, D., Murphy, D. G. M., & Blackwood, N. the Barratt impulsiveness scale. Journal of Clinical Psychology, (2018). The effect of intranasal oxytocin on neural response to facial 51(6), 768–774. emotions in healthy adults as measured by functional MRI: A sys- Peterson, E., & Miller, S. F. (2012). The Eyes test as a measure of indi- tematic review. Psychiatry Research: Neuroimaging, 272, 17–29. vidual differences: How much of the variance reflects verbal IQ? https://doi.org/10.1016/j.pscychresns.2017.11.017 Frontiers in Psychology, 3, 220. https://doi.org/10.3389/fpsyg. Walum, H., Waldman, I. D., & Young, L. J. (2016). Statistical and meth- 2012.00220 odological considerations for the interpretation of intranasal oxyto- Poljak, A., & Sachdev, P. (2021). The need for a reliable oxytocin assay. cin studies. Biological Psychiatry, 79(3), 251–257. https://doi.org/ Molecular Psychiatry. https://doi.org/10.1038/s41380-021-01114-0 10.1016/j.biopsych.2015.06.016 Quintana, D. S., & Williams, D. R. (2018). Bayesian alternatives for Winterton, A., Westlye, L. T., Steen, N. E., Andreassen, O. A., & common null-hypothesis significance tests in psychiatry: A non- Quintana, D. S. (2021). Improving the precision of intranasal oxy- technical guide using JASP. BMC Psychiatry, 18(1), 178. https:// tocin research. Nature Human Behaviour, 5(1), 9–18. https://doi. doi.org/10.1186/s12888-018-1761-4 org/10.1038/s41562-020-00996-4 Quintana, D. S., & Guastella, A. J. (2020). An Allostatic Theory of Woolley, J. D., Arcuni, P. A., Stauffer, C. S., Fulford, D., Carson, D. S., Oxytocin. Trends in Cognitive Sciences, 24(7), 515-528. https:// Batki, S., & Vinogradov, S. (2016). The effects of intranasal oxyto- doi.org/10.1016/j.tics.2020.03.008 cin in opioid-dependent individuals and healthy control subjects: A Radke, S., & de Bruijn, E. R. A. (2015). Does oxytocin affect mind- pilot study. Psychopharmacology, 233(13), 2571–2580. https://doi. reading? A replication study. Psychoneuroendocrinology, 60, 75– org/10.1007/s00213-016-4308-8 81. https://doi.org/10.1016/j.psyneuen.2015.06.006 Woolley, J. D., Chuang, B., Lam, O., Lai, W., O’Donovan, A., Rankin, Shamay-Tsoory, S. G., & Abu-Akel, A. (2016). The social salience hy- K. P., Mathalon, D. H., & Vinogradov, S. (2014). Oxytocin admin- pothesis of oxytocin. Biological Psychiatry, 79(3), 194–202. https:// istration enhances controlled social cognition in patients with doi.org/10.1016/j.biopsych.2015.07.020 schizophrenia. Psychoneuroendocrinology, 47, 116–125. https:// Spengler, F. B., Schultz, J., Scheele, D., Essel, M., Maier, W., Heinrichs, doi.org/10.1016/j.psyneuen.2014.04.024 M., & Hurlemann, R. (2017). Kinetics and dose dependency of intranasal oxytocin effects on amygdala reactivity. Biological Psychiatry, 82(12), 885–894. https://doi.org/10.1016/j.biopsych. Open practices 2017.04.015 The data or materials for the experiments reported here are available Stanford, M. S., Mathias, C. W., Dougherty, D. M., Lake, S. L., on reasonable request from the corresponding author, and none of the Anderson, N. E., & Patton, J. H. (2009). Fifty years of the Barratt experiments was preregistered. Impulsiveness Scale: An update and review. Personality and Individual Differences, 47(5), 385–395. https://doi.org/10.1016/j. Author Note paid.2009.04.008 Data collection and preliminary analyses were supported by an award Stefan, A. M., Gronau, Q. F., Schönbrodt, F. D., & Wagenmakers, E.-J. of the Mind/Brain/Behavior Interfaculty Initiative of Harvard University. (2019). A tutorial on Bayes Factor design analysis using an in- AKB was moreover supported by the Young Academics Support and the formed prior. Behavior Research Methods, 51(3), 1042–1058. Stiefel-Zangger Foundation of the University of Zurich, Switzerland, and https://doi.org/10.3758/s13428-018-01189-8 the Swiss National Foundation (PBZHP1_147196). We have no conflicts Steyer, R., Schwenkmezger, P., Notz, P., & Eid, M. (1994). of interest to disclose. Testtehoretische Analysen des Mehrdimensionalen Befindlichkeitsfragebogen (MDBF) [Theoretical analysis of a mul- tidimensional mood questionnaire (MDBF)]. Diagnostica(40 (4)), 320–328. Striepens, N., Kendrick, K. M., Hanking, V., Landgraf, R., Wüllner, U., Maier, W., & Hurlemann, R. (2013). Elevated cerebrospinal fluid Publisher’s note Springer Nature remains neutral with regard to jurisdic- and blood concentrations of oxytocin following its intranasal tional claims in published maps and institutional affiliations.