Oxytocin modulates human chemosensory decoding of sex in a dose-dependent manner
Figures
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig1-v1.tif/full/617,/0/default.jpg)
Experimental procedure and design.
(A) Schematic illustration of a trial in the gender identification task. Each trial began with a 500 ms fixation cross, followed by a dynamic point-light walker (PLW) presented for 500 ms (0.5 walking cycle). Participants then pressed one of two buttons to indicate whether it was a male or a female walker. The physical gender of the PLW was denoted by a Z score and ranged in seven equal steps from feminine (−0.45 SD) to masculine (0.45 SD) with the center (0) individually adjusted to approximately perceived gender neutrality in the absence of drug treatment and olfactory stimulus. (B) Participants’ gender judgments for PLWs were fitted with a psychometric function that contained two parameters: point of subjective equality (PSE) and difference limen. The PSE is the physical gender of a PLW (Z value) that corresponds to a probability of 50% on the fitted psychometric function (gray circle on the x-axis), where the participant perceived the PLW as equally masculine and feminine. The difference limen is half the interquartile range of the fitted function. (C) Participants’ PSEs when smelling either chemosignal were compared with those when smelling the carrier control alone (gray circle). A negative PSE shift relative to the carrier control condition (dark gray arrow) corresponds to an overall leftward shift of the psychometric curve and reflects an increased tendency to judge the PLWs as male, hence a masculine bias in gender perception. Conversely, a positive PSE shift (light gray arrow) indicates a feminine bias in gender perception. All curves here are hypothetical. (D) Experiments 1 and 2 tested heterosexual and homosexual men, respectively. Drug treatment (60 µg atosiban, 24 IU oxytocin, 24 IU vasopressin) served as a between-subjects factor whereas olfactory condition (androstadienone, estratetraenol, carrier control) served as a within-subjects factor. Participants performed the experimental blocks of the gender identification task 35 min after nasal drug administration, under the continuous exposure of an olfactory stimulus. (E) Experiments 3 and 5 tested high AQ (AQ scores ≥ 25) and low AQ (AQ scores < 25) heterosexual men, respectively. Both drug treatment (no drug, 12 IU oxytocin, 24 IU oxytocin, 12 IU vasopressin, 24 IU vasopressin) and olfactory condition (estratetraenol, carrier control) were manipulated in a within-subjects fashion.
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig2-v1.tif/full/617,/0/default.jpg)
Oxytocin, vasopressin, and subconscious chemosensory decoding of sex in heterosexual and homosexual men.
(A–D, F–I) Androstadienone- and estratetraenol- induced visual gender judgment biases in the absence of drug treatment (A and F, adapted from Zhou et al., 2014 for comparison) and after the nasal administrations of 60 µg atosiban (B, G), 24 IU oxytocin (C, H), and 24 IU vasopressin (D, I) in heterosexual (Experiment 1, A–D) and homosexual (Experiment 2, F–I) men. Gender identification performances under the exposures of androstadienone, estratetraenol, and the carrier control are respectively fitted with sigmoidal curves (blue solid curves, red solid curves, and gray dashed curves, respectively). Insets show the androstadienone- and estratetraenol- induced proportional ‘male’ biases at the gender-neutral point of the point-light walkers (PLWs), that is, androstadienone- and estratetraenol- induced differences in the proportion of ‘male’ responses at Z = 0 relative to the carrier control condition. (E, J) Androstadienone- and estratetraenol- induced overall point of subjective equality (PSE) shifts with respect to the carrier control after the nasal administrations of 60 µg atosiban, 24 IU oxytocin (OT), and 24 IU vasopressin (VP) in heterosexual (E) and homosexual (J) men. A positive PSE shift indicates a feminine bias, that is, a bias toward perceiving the PLWs as more feminine, whereas a negative PSE shift indicates a masculine bias, that is, a bias toward perceiving the PLWs as more masculine. Dashed box: data from our earlier study (Zhou et al., 2014); error bars: SEMs adjusted for individual differences; *: p<0.05; **: p≤0.01; ***: p≤0.005.
-
Figure 2—source data 1
Experiments 1 and 2.
Androstadienone- and estratetraenol- induced visual gender judgment biases after the nasal administrations of 60 µg atosiban, 24 IU oxytocin, and 24 IU vasopressin in heterosexual and homosexual male participants.
- https://cdn.elifesciences.org/articles/59376/elife-59376-fig2-data1-v1.data.xlsx
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig3-v1.tif/full/617,/0/default.jpg)
Central tendencies of androstadienone- and estratetraenol- induced point of subjective equality (PSE) shifts in heterosexual and homosexual men across drug conditions.
Each subfigure shows the bivariate distributions of bootstrapped sample means for heterosexual men (1000 cyan dots) and homosexual men (1000 lime dots) plotted against the horizontal and vertical axes representing androstadienone- and estratetraenol-induced PSE shifts, respectively, following the nasal administration of 60 µg atosiban, 24 IU oxytocin, or 24 IU vasopressin. A positive value on either axis indicates a feminine bias, that is, a bias toward perceiving the point-light walkers (PLWs) as more feminine, whereas a negative value indicates a masculine bias, that is, a bias toward perceiving the PLWs as more masculine.
-
Figure 3—source data 1
Experiments 1 and 2.
Degrees of gender perception biases induced by chemosensory sexual cues in high AQ and low AQ individuals.
- https://cdn.elifesciences.org/articles/59376/elife-59376-fig3-data1-v1.data.xlsx
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig3-figsupp1-v1.tif/full/617,/0/default.jpg)
Histogram distributions (with normal curves) of 1000 bootstrapped sample means for the points of subjective equality (PSEs) under each combination of olfactory and drug conditions in heterosexual and homosexual men.
Note that the PSEs under the carrier control condition were comparable across drug treatments and between heterosexual and homosexual men (drug treatment: F2, 138 = 0.092, p=0.91; sexual orientation: F1, 138 = 0.011, p=0.92; interaction: F2, 138 = 0.22, p=0.80).
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig3-figsupp2-v1.tif/full/617,/0/default.jpg)
Central tendencies of the degrees of gender perception biases induced by chemosensory sexual cues in high AQ and low AQ individuals in Experiments 1 and 2.
Each plot shows the distributions of bootstrapped sample means for the vasopressin-treated high AQ and low AQ individuals. The x-axis represents the proportional gender perception bias at the gender-neutral point (Z = 0) of the point-light walkers (PLWs) (A) or the overall point of subjective equality (PSE) shift (B) induced by estratetraenol (for heterosexual men) or androstadienone (for homosexual men) with respect to the carrier control. Zero on the x-axis marks an absence of chemosensory influence. A positive effect of estratetraenol on heterosexual men (i.e. biasing them toward perceiving the PLWs as more feminine) or of androstadienone on homosexual men (i.e. biasing them toward perceiving the PLWs as more masculine) has a positive x value in (A) and a negative x value in (B).
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig4-v1.tif/full/617,/0/default.jpg)
Oxytocin, but not vasopressin, modulates chemosensory decoding of femininity in heterosexual men in a dose-dependent manner.
(A–E, G–K) Estratetraenol-induced visual gender judgment biases in the absence of drug treatment (A, G) and after the nasal administrations of 12 IU oxytocin (B, H), 24 IU oxytocin (C, I), 12 IU vasopressin (D, J), and 24 IU vasopressin (E, K) in high AQ (Experiment 3, A–E) and low AQ (Experiment 5, G–K) heterosexual men. Gender identification performances under the exposures of estratetraenol and the carrier control are fitted with sigmoidal curves (red solid curves and gray dashed curves, respectively). Insets show the estratetraenol-induced proportional ‘male’ biases at the gender-neutral point of the point-light walkers (PLWs), that is, estratetraenol-induced differences in the proportion of ‘male’ responses at Z = 0 relative to the carrier control condition. (F, L) Estratetraenol-induced overall point of subjective equality (PSE) shifts with respect to the carrier control in the absence of drug treatment and after the nasal administrations of 12 IU oxytocin, 24 IU oxytocin, 12 IU vasopressin, and 24 IU vasopressin in high AQ (F) and low AQ (L) heterosexual men. A positive PSE shift indicates a feminine bias, that is, a bias toward perceiving the PLWs as more feminine. Error bars: SEMs adjusted for individual differences; *: p<0.05; **: p≤0.01; ***: p≤0.005.
-
Figure 4—source data 1
Experiments 3 and 5.
Estratetraenol-induced visual gender judgment biases in the absence of drug treatment and after the nasal administrations of 12 IU oxytocin, 24 IU oxytocin, 12 IU vasopressin, and 24 IU vasopressin in high AQ and low AQ heterosexual male participants.
- https://cdn.elifesciences.org/articles/59376/elife-59376-fig4-data1-v1.xlsx
-
Figure 4—source data 2
Experiment 4.
Estratetraenol-induced visual gender judgment biases after the nasal administrations of saline, 12 IU oxytocin, and 24 IU oxytocin in high AQ heterosexual male participants.
- https://cdn.elifesciences.org/articles/59376/elife-59376-fig4-data2-v1.xlsx
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig4-figsupp1-v1.tif/full/617,/0/default.jpg)
Experiment 4 replicated the main findings of Experiment 3 that oxytocin modulates chemosensory decoding of femininity in high AQ heterosexual men in a dose-dependent manner.
(A–C) Estratetraenol-induced visual gender judgment biases after the nasal administrations of saline (A), 12 IU oxytocin (B), and 24 IU oxytocin (C). Sigmoidal curves and insets are as in Figure 4. (D) Estratetraenol-induced overall point of subjective equality (PSE) shifts with respect to the carrier control after the nasal administrations of saline (t23 = 0.65, p=0.52), 12 IU oxytocin (t23 = 3.03, p=0.006, Cohen’s d = 0.62), and 24 IU oxytocin (t23 = 0.24, p=0.82). Error bars: SEMs adjusted for individual differences; †: marginally significant; **: p≤0.01.
![](https://iiif.elifesciences.org/lax/59376%2Felife-59376-fig5-v1.tif/full/617,/0/default.jpg)
Comparison of dose-response relationships of oxytocin and vasopressin between high AQ and low AQ heterosexual men.
(A) Bivariate distributions of bootstrapped sample means for high AQ (Experiments 3–4, 1000 dark brown dots) and low AQ (Experiment 5, 1000 light brown dots) heterosexual men plotted against the horizontal and vertical axes, representing estratetraenol-induced point of subjective equality (PSE) shifts at baseline (no drug treatment/saline) and following 12 IU intranasal oxytocin, respectively. (B) AQ score was negatively correlated with estratetraenol-induced PSE shift at baseline (left panel) and positively correlated with the increase of estratetraenol-induced PSE shift post 12 IU oxytocin treatment (right panel). (C) Overall, exogenous oxytocin modulated estratetraenol-induced PSE shift in manners that were dose-dependent and contingent upon the recipient’s social proficiency (left panel), whereas exogenous vasopressin consistently showed no significant impact (right panel). A positive PSE shift indicates a feminine bias, that is, a bias toward perceiving the point-light walkers (PLWs) as more feminine. Error bars: SEMs adjusted for individual differences; *: p<0.05; **: p≤0.01; ***: p≤0.005.
Tables
Summary of the effects of androstadienone and estratetraenol on the recipients’ gender judgment criteria across various drug treatments.
Each cell represents results from 24 participants, respectively from Experiments 1 (heterosexual men), 2 (homosexual men), 3 (high AQ heterosexual men), and 5 (low AQ heterosexual men). −: no significant effect relative to the carrier control, +: a positive effect, that is, recipients biased toward perceiving the point-light walkers (PLWs) as more masculine (by androstadienone, A) or more feminine (by estratetraenol, E).
Baseline * | Atosiban | 24 IU OT | 24 IU VP | |
---|---|---|---|---|
Heterosexual men | A− E+ | A− E− | A− E− | A− E+ |
Homosexual men | A+ E− | A− E− | A− E− | A+ E− |
Baseline | 12 IU OT | 24 IU OT | 12 IU VP | 24 IU VP | |
---|---|---|---|---|---|
High AQ heterosexual men | E−† | E+ † | E−† | E− | E− |
Low AQ heterosexual men | E+ | E− | E− | E+ | E+ |
-
*Results from our earlier study (Zhou et al., 2014) for comparison.
†Results replicated in Experiment 4.