23(5) 459 –463
© The Author(s) 2012
Reprints and permission:
The notion that individual animals, including humans, pursue
dominant social positions to ensure access to resources and
reproductive advantage is of great scientific and societal inter-
est (Archer, 2006; Bos, Panksepp, Bluthé, & van Honk, 2011;
Eisenegger, Haushofer, & Fehr, 2011; Josephs, Newman,
Brown, & Beer, 2003; Josephs, Sellers, Newman, & Mehta,
2006; Mazur & Booth, 1998). All the way through vertebrate
phylogeny, from reptiles to mammals, the steroid hormone tes-
tosterone has been identified as a driving force for engaging
and prevailing in confrontations for social dominance (Archer,
2006), which underlie the formation of social hierarchies
(Mazur & Booth, 1998). For millions of years, testosterone
evidently has acted on evolutionarily primordial brain mecha-
nisms that motivate animals to increase and maintain social
status and power.
In humans, though, the expanded neocortex is thought to be
in control of primordial drives, and testosterone’s effects on
social behavior are said to have shifted to the promotion of feel-
ings of superiority, strength, anger, and low anxiety. In turn,
these consciously experienced motivational states are said to
direct voluntary control of behavior dealing with social chal-
lenges and threats (Eisenegger et al., 2011; Josephs et al., 2003;
Josephs et al., 2006; Mazur & Booth, 1998). This notion, how-
ever, is currently under debate because it is based on merely
correlational evidence; consequently, one cannot exclude the
possibility that testosterone regulates status-seeking behaviors
in humans unconsciously and automatically without affecting
conscious motivational states (Bos et al., 2011).
In earlier research, we showed that salivary testosterone lev-
els were associated with attentional vigilance to angry faces
(van Honk et al., 1999), and that testosterone administration
increased cardiac reactivity to angry faces (van Honk et al.,
2001). A third study demonstrated that testosterone administra-
tion increases amygdala reactivity to angry (relative to happy)
faces (Hermans, Ramsey, & van Honk, 2008). These findings
converge to suggest that testosterone enhances vigilance toward
social signals of dominance (i.e., angry faces). In these studies,
however, the facial expressions were perceived consciously,
whereas our hypothesis has been that testosterone increases
vigilance, or dominance, primarily through automatic, uncon-
scious mechanisms (van Honk & Schutter, 2007; van Honk,
Schutter, Hermans, & Putman, 2004). Although other research-
ers found correlational support for this hypothesis (Wirth &
Schultheiss, 2007), it has not yet been confirmed with causal
methodology. Here, we report a placebo-controlled study of the
effects of testosterone administration in which we not only used
infrared eye tracking to measure a social-dominance behavior
David Terburg, Utrecht University— Psychonomy, Heidelberglaan 2,
Utrecht, Utrecht 3584 CS, The Netherlands
Testosterone Affects Gaze Aversion From
Angry Faces Outside of
David Terburg1,2, Henk Aarts1, and Jack van Honk1,2
1Department of Psycholog y, Utrecht University, and 2Department of Psychiatry and Mental Health, University of Cape Town
Throughout vertebrate phylogeny, testosterone has motivated animals to obtain and maintain social dominance—a fact
suggesting that unconscious primordial brain mechanisms are involved in social dominance. In humans, however, the prevailing
view is that the neocortex is in control of primordial drives, and testosterone is thought to promote social dominance via
conscious feelings of superiority, indefatigability, strength, and anger. Here we show that testosterone administration in
humans prolongs dominant staring into the eyes of threatening faces that are viewed outside of awareness, without affecting
consciously experienced feelings. These findings reveal that testosterone motivates social dominance in humans in much the
same ways that it does in other vertebrates: involuntarily, automatically, and unconsciously.
social interaction, neuroendocrinology, facial expressions, eye movements, aggressive behavior
Received 7/20 /11; Revision accepted 11/17/11
460 Terburg et al.
that was much more ecologically valid than the measures in our
earlier studies, but also used a backward-masking technique to
ensure that the facial expressions did not reach consciousness.
We administered testosterone and placebo to 20 healthy
volunteers and tested effects on performance on a social-
dominance task, as well as self-reports on a widely used inven-
tory that assesses conscious motivational states of anger, vigor,
fatigue, anxiety, and depression. In the social-dominance task,
faces were presented outside of conscious awareness, and eye
movements were tracked to assess participants’ inclination to
either gaze away from (submission) or endure (dominance)
face-to-face status threats in the form of angry stares (Terburg,
Hooiveld, Aarts, Kenemans, & van Honk, 2011). Thus, we
measured genuine gaze aversion from masked angry faces and
tested the causal role of testosterone in promoting social domi-
Participants and design
Twenty healthy volunteers (age range: 20–25 years) received
sublingual testosterone and placebo in counterbalanced order,
with the two tests separated by 1 week. We exclusively
recruited women using single-phase contraceptives for several
reasons. First, this minimized menstrual-cycle effects on basal
hormone levels. Second, the magnitude and time course of the
neurophysiological effects of testosterone have been estab-
lished only in women (Tuiten et al., 2000). Third, basal testos-
terone levels in females have been shown to correlate with
both aggressive behavior and implicit measures of dominance
(Cashdan, 1995; Dabbs & Hargrove, 1997; Josephs et al.,
2003; Josephs et al., 2006).
Sublingual drug samples consisted of 0.5 mg of testosterone,
5 mg of cyclodextrin (carrier), 5 mg of ethanol, and 0.5 ml of
water. The placebo samples were the same except that testos-
terone was omitted. Sublingual administration of testosterone
induces behavioral and physiological effects, as indexed by
subjective and vaginal arousal to erotic stimuli, that peak after
4 hr (Tuiten et al., 2000). Accordingly, experimental testing
was started 4 hr after drug (and placebo) administration. Note
that this method has been used successfully in more than a
dozen studies on social and emotional aspects of human
behavior (Bos et al., 2011).
Conscious assessment of mood state
Before performing the social-dominance task, participants
completed the Profile of Mood States (Shacham, 1983), a vali-
dated 30-item questionnaire that indexes consciously experi-
enced anger, anxiety, depression, fatigue, and vigor, using
visual analogue scales.
The stimuli for the social-dominance task included angry,
happy, and neutral faces of five men and five women. On each
trial, a gray mask with a central fixation point was followed by
a face that was presented in blue, green, or red for 33 ms before
a mask stimulus of the same color; the masks and face had
similar luminance properties. At the bottom of each face and
mask display were three circles; participants were instructed
that when the central stimulus turned from gray to a color, they
should avert their gaze from the central fixation point to the
circle with the corresponding color (see Fig. 1a). The differ-
ence in gaze-aversion latency between angry and happy
expressions in this task is a reliable index of dominance
motives (see Terburg et al., 2011). Facial expressions were
presented in a fixed sequence that was repeated five times
(NxxyNyyxNNyyxNxxyN; N = neutral; x and y = angry and
happy, counterbalanced across the two sessions). This order
ensured that all combinations of successive trial types occurred
equally often, allowing us to analyze trials following a neutral
baseline separately and eliminating trial-by-trial interference
of emotionally conflicting information (Etkin, Egner, Peraza,
Kandel, & Hirsch, 2006; Kunde & Mauer, 2008).
Gaze movements were recorded with a Tobii-1750 eye
tracker (Tobii Technology, Danderyd, Sweden), and gaze-
aversion latency was defined as the time between face onset
and first gaze on the target circle. Latencies more than 3 stan-
dard deviations from an individual’s mean were excluded
Emotion awareness check
At the end of the final session, participants were asked whether
they had seen the emotional expressions during the task. Sub-
sequently, all 30 face stimuli (10 faces × 3 emotions) were
presented again, masked, and participants were instructed to
identify each facial expression as happy, angry, or neutral, in a
Mean latencies on angry-face and happy-face trials were
baseline-corrected by subtracting the mean latency on neutral-
face trials and then entered in a 2 (emotion: angry vs. happy) ×
2 (drug condition: testosterone vs. placebo) repeated measures
analysis of variance. The Emotion × Drug Condition interac-
tion was significant, F(1, 19) = 8.84, p = .008, ηp
2 = .32. Post
hoc paired t tests confirmed that after testosterone administra-
tion, t(19) = 3.06, p = .006, but not after placebo, t(19) = −1.33,
p = .201, gaze aversion from angry faces was slower than gaze
aversion from happy faces (see Fig. 1b).
Next, we assessed angry- and happy-face trials that fol-
lowed neutral-face trials separately. This analysis revealed
a main effect of emotion, F(1, 19) = 5.06, p = .037, ηp
.21, which was explained by the Emotion × Drug Condition
Testosterone Predicts Gaze Aversion 461
interaction, F(1, 19) = 5.74, p = .027, ηp
2 = .23 (see Fig. S1 in
the Supplemental Material available online). Post hoc paired t
tests confirmed that testosterone administration slowed down
gaze aversion from angry faces, t(19) = 2.13, p = .046, and not
from happy faces, t(19) = 0.10, p = .992. In sum, although
slower gaze aversion from angry faces compared with happy
faces can be interpreted as reflecting either dominance or
reduced reward sensitivity (Terburg et al., 2011), this anger-
specific effect confirms that testosterone promotes dominance-
related gaze behavior.
Unconscious Gaze-Aversion Bias (ms)
Anger Anxiety Depression Fatigue Vigor
Conscious Mood Score
Fig. 1. Illustration of the experimental method and results. In each trial of the social-dominance t ask (a),
participants watched a meaningless gray picture turn blue, green, or red, at which point they were to shift their
gaze downward, as fast as possible, to the circle with the corresponding color. Crucially, during the color transition,
a facial expression was presented too quickly to be consciously perceived; thus, the downward gaze shif t was an
implicit act of gaze aversion from a social signal of reassurance (happy expression), a neutral signal (neutral
expression), or a face-to-face status threat (angry expression, which rendered the gaze shift an unconscious act
of submission; f igure adapted from Terburg, Hooiveld, Aar ts, Kenemans, & van Honk, 2011) . The graphs present
(b) the mean difference in baseline-corrected gaze-aversion latency bet ween angry and happy faces (angry – happy) and
(c) the mean self-reported mood states in the two drug conditions (testosterone vs. placebo). Error bars represent
standard errors of the mean.
462 Terburg et al.
None of the participants reported awareness of the facial
expressions, but 5 scored significantly above chance level on
the awareness check (i.e., > 14 correct; chance level = 10 cor-
rect; binomial test with n = 30, one-tailed α = .05). Crucially,
the effect of testosterone on gaze aversion remained signifi-
cant (tested with one-tailed Wilcoxon signed-ranks tests
because of the small sample size and directed hypotheses) for
both these participants (Z = −2.02, p = .022, n = 5) and those
who were not aware of the facial expressions (Z = −1.70, p =
.044, n = 15).
Finally, there were no effects of drug condition on self-
reported mood states (all ps > .5; see Fig.1c).
Our results show that after testosterone administration, partici-
pants reflexively maintain eye contact when unconsciously
confronted with angry faces. Crucially, this unconscious dis-
play of dominance in face-to-face confrontations (Terburg
et al., 2011) was accompanied by neither increased anger and
vigor nor decreased anxiety, fatigue, or depression. This find-
ing indicates that these consciously experienced motivational
states do not underlie testosterone-induced social-dominance
Slower gaze aversion from angry than from happy faces has
been shown to be independently related to dominance motives
and reduced reward sensitivity (Terburg et al., 2011). On the
basis of these findings taken by themselves, we cannot exclude
the possibility that our results are due to testosterone speeding
up gaze aversion from happy faces. However, testosterone
administration has previously resulted in increased reward sen-
sitivity and appetitive motivation (Hermans et al., 2010; van
Honk, Schutter, Hermans, Putman, Tuiten, & Koppeschaar,
2004), which makes the latter explanation unlikely. Most impor-
tant, the effect of testosterone in the baseline-corrected analysis
was anger-specific, which confirms our hypothesis that testos-
terone specifically induces dominance-related gaze behavior.
Although our drug-administration method generally yields
effects similar to those of endogenous testosterone in females as
well as males (Bos et al., 2011; van Honk & Schutter, 2007),
future research should confirm that the results we obtained are
also observed in males.
These results extend our previous findings on vigilance to
consciously processed angry faces after testosterone adminis-
tration (Hermans et al., 2008; van Honk et al., 2001; van Honk
et al., 1999), by showing that testosterone promotes domi-
nance behavior toward unconsciously perceived angry faces
as well. Our results add to the ongoing debate on whether tes-
tosterone promotes dominance through complex psychologi-
cal mechanisms (Mazur & Booth, 1998) or reflexive biological
mechanisms (van Honk, Schutter, Hermans, & Putman, 2004).
Moreover, we have shown not only that testosterone affects
vigilance to anger, but also that the hormone genuinely pro-
motes social-dominance behavior by restraining gaze aversion
when individuals are confronted with angry eye contact
(Terburg et al., 2011). Although conscious psychological
mechanisms unmistakably play a role in the urge for social
status (Eisenegger et al., 2011; Mazur & Booth, 1998), testos-
terone’s promotion of human social-dominance behavior evi-
dently precedes these higher-order mechanisms. The present
study thus provides compelling evidence that testosterone acts
directly—involuntarily, automatically, and unconsciously—
on social dominance in humans through phylogenetically
ancient pathways shared with other vertebrate species (Archer,
2006; Bos et al., 2011).
Declaration of Conflicting Interests
The authors declared that they had no conflicts of interest with
respect to their authorship or the publication of this article.
This research was supported by grants from the Netherlands
Organization for Scientific Research (VICI-453-06-002; Brain &
Cognition 056-24-010) and the Hope-for-Depression Research
Additional supporting information may be found at http://pss.sagepub
Archer, J. (2006). Testosterone and human aggression: An evalua-
tion of the challenge hypothesis. Neuroscience & Biobehavioral
Reviews, 30, 319–345.
Bos, P. A., Panksepp, J., Bluthé, R. M., & van Honk, J. (2011). Acute
effects of steroid hormones and neuropeptides on human social-
emotional behavior: A review of single administration studies.
Frontiers in Neuroendocrinology, 33, 17–35.
Cashdan, E. (1995). Hormones, sex, and status in women. Hormones
and Behavior, 29, 354–366.
Dabbs, J. M., Jr., & Hargrove, M. F. (1997). Age, testosterone, and
behavior among female prison inmates. Psychosomatic Medi-
cine, 59, 477–480.
Eisenegger, C., Haushofer, J., & Fehr, E. (2011). The role of testos-
terone in social interaction. Trends in Cognitive Sciences, 15,
Etkin, A., Egner, T., Peraza, D. M., Kandel, E. R., & Hirsch, J. (2006).
Resolving emotional conflict: A role for the rostral anterior cin-
gulate cortex in modulating activity in the amygdala. Neuron, 51,
Hermans, E. J., Bos, P. A., Ossewaarde, L., Ramsey, N. F., Fernandez,
G., & van Honk, J. (2010). Effects of exogenous testosterone on
the ventral striatal BOLD response during reward anticipation in
healthy women. NeuroImage, 52, 277–283.
Hermans, E. J., Ramsey, N. F., & van Honk, J. (2008). Exogenous
testosterone enhances responsiveness to social threat in the neural
circuitry of social aggression in humans. Biological Psychiatry,
Testosterone Predicts Gaze Aversion 463
Josephs, R. A., Newman, M. L., Brown, R. P., & Beer, J. M. (2003).
Status, testosterone, and human intellectual performance: Stereo-
type threat as status concern. Psychological Science, 14, 158–163.
Josephs, R. A., Sellers, J. G., Newman, M. L., & Mehta, P. H. (2006).
The mismatch effect: When testosterone and status are at odds.
Journal of Personality and Social Psychology, 90, 999–1013.
Kunde, W., & Mauer, N. (2008). Sequential modulations of valence
processing in the emotional Stroop task. Experimental Psychol-
ogy, 55, 151–156.
Mazur, A., & Booth, A. (1998). Testosterone and dominance in men
[Target article and commentaries]. Behavioral and Brain Sci-
ences, 21, 353–397.
Shacham, S. (1983). A shortened version of the Profile of Mood
States. Journal of Personality Assessment, 47, 305–306.
Terburg, D., Hooiveld, N., Aarts, H., Kenemans, J. L., & van Honk,
J. (2011). Eye tracking unconscious face-to-face confrontations:
Dominance motives prolong gaze to masked angry faces. Psy-
chological Science, 22, 314–319.
Tuiten, A., Van Honk, J., Koppeschaar, H., Bernaards, C., Thijssen,
J., & Verbaten, R. (2000). Time course of effects of testosterone
administration on sexual arousal in women. Archives of General
Psychiatry, 57, 149–153.
van Honk, J., & Schutter, D. J. L. G. (2007). Vigilant and avoidant
responses to angry facial perceptions: Dominance and submission
motives. In E. Harmon-Jones & P. Winkielman (Eds.), Social
neuroscience: Integrating biological and psychological expla-
nations of social behavior (pp. 197–223). New York, NY: Guil-
van Honk, J., Schutter, D. J. L. G., Hermans, E. J., & Putman, P. (2004).
Testosterone, cortisol, dominance, and submission: Biologically
prepared motivation, no psychological mechanisms involved.
Behavioral and Brain Sciences, 27, 160–162.
van Honk, J., Schutter, D. J. L. G., Hermans, E. J., Putman, P., Tuiten,
A., & Koppeschaar, H. (2004). Testosterone shifts the balance
between sensitivity for punishment and reward in healthy young
women. Psychoneuroendocrinology, 29, 937–943.
van Honk, J., Tuiten, A., Hermans, E., Putman, P., Koppeschaar, H.,
Thijssen, J., . . . van Doornen, L. (2001). A single administration
of testosterone induces cardiac accelerative responses to angry
faces in healthy young women. Behavioral Neuroscience, 115 ,
van Honk, J., Tuiten, A., Verbaten, R., van den Hout, M., Koppe-
schaar, H., Thijssen, J., & de Haan, E. (1999). Correlations
among salivary testosterone, mood, and selective attention to
threat in humans. Hormones and Behavior, 36, 17–24.
Wirth, M. M., & Schultheiss, O. C. (2007). Basal testosterone moder-
ates responses to anger faces in humans. Physiology & Behavior,