Nearly fifty years ago, the term “pheromone” was coined to define substances secreted by an organism that, even in minute amounts, release specific behaviors in another individual of the same species (Karlson and Lüscher, 1959). The animal kingdom is full of such chemical communication, from attracting silk moth mates (Vogt and Riddiford, 1981) to initiating suckling behavior in rabbit pups (Schaal et al., 2003). This abundance has naturally led to questions about the existence of human pheromones (Hays, 2003, Wysocki and Preti, 2004). However, it is unreasonable to believe that human behavior could be altered in such an automated fashion. Rather, human chemical communication is likely to be subtler, an idea that has led to the classification modulator pheromone. A modulator pheromone modifies “ongoing behavior or a psychological reaction to a particular context without triggering specific behaviors or thoughts” (McClintock, 2002). The natural human compound Δ4,16-androstadien-3-one (androstadienone) is a candidate to fit this definition, though the precise modulation of psychological states by the compound has yet to be definitively determined.
Androstadienone, a non-androgenic derivative of gonadal progesterone, has been identified in plasma (males: 98 ng/100 mL blood, females: 36 ng/100 mL; Brooksbank et al., 1972) and sweat (440 μM/mL of apocrine sweat; Gower et al., 1994, Labows, 1988), as well as male semen (Kwan et al., 1992) and axillary hair (Nixon et al., 1988). Even when presented in minute amounts, androstadienone modulates psychological, physiological and hormonal responses compared to control-treated sessions (Jacob and McClintock, 2000, Jacob et al., 2001a, Bensafi et al., 2003, Bensafi et al., 2004b, Lundström and Olsson, 2005, Villemure and Bushnell, 2007, Wyart et al., 2007). Notably, these effects can be obtained with vapors alone and are absent for compounds with similar chemical structure (androstenol) or odor qualities (muscone) (Jacob et al., 2002).
Biopsychological responses to androstadienone are dependent on individual characteristics and environmental context. During a close social interaction, the emotional states of men and women are differentially affected by the compound, with females generally reporting a more positive emotional state after androstadienone presentation and men reporting more negative or no change in emotion (e.g. Jacob and McClintock, 2000, Villemure and Bushnell, 2007). However, androstadienone-related mood differences are influenced by experimental factors, including the sex of the experimenter (Jacob et al., 2001a) or the emotional content of a film (Bensafi et al., 2004a). In a distinctly negative context, passive inhalation of an androstadienone mixture induces a more intense perception of pain by women (Villemure and Bushnell, 2007). Sexual preference also plays a role, altering the neural response to androstadienone in the hypothalamus in both men and women (Savic et al., 2001, Savic et al., 2005, Berglund et al., 2006).
Such a wide variety of results begs for a clarification of the precise psychological mechanisms through which androstadienone is operating. While previous research has often examined effects of androstadienone in light of its possible sexual properties (e.g. Savic et al., 2001, Bensafi et al., 2004b), here we aim to determine whether effects of androstadienone are present in a wider context. This goal is an important step in determining precisely how androstadienone modulates psychological processes. Rather than assuming the compound only exerts its effects in the sexual or social domain, we broadened the scope of the investigation in order to examine whether the effects of androstadienone can be simplified into a single, over-arching explanation. To this end, we propose three distinct, competing hypotheses regarding the psychological effects of androstadienone. We designed a complementary set of four studies aimed at uniquely accepting one of these three hypotheses.
The first hypothesis is that androstadienone exerts its effects by improving general attention to the external environment. This idea implies that androstadienone induces people to be more cognizant of and reactive to the world them. Lending credence to this concept is the observation that women score lower on a “negative-confused” factor (Jacob and McClintock, 2000, Jacob et al., 2002) and more “focused” (Lundström et al., 2003) with androstadienone.
Second, androstadienone may solely enhance attention to emotional information. With this hypothesis, mood effects are due to greater awareness and subsequent evaluation of both positive stimuli, such as interaction with an opposite sex experimenter (Jacob et al., 2002), and negative signals, including a painful heat stimulus (Villemure and Bushnell, 2007). This proposal is intriguing, considering that the initial psychological findings were interpreted as specific changes in emotional state, with opposite effects on men and women. Under this hypothesis, the effects on men and women could be the same, and previously reported sex differences would have resulted from sex differences in response to the environment, and not androstadienone itself.
Last, social attention may be specifically altered by androstadienone. Social interactions clearly plays a role in biopsychological response to androstadienone (Jacob et al., 2001a), as does sexual preference (Savic et al., 2005, Berglund et al., 2006), a prominent aspect of sociality. Enhanced attention to social information may be driving these effects (Gulyás et al., 2004).
These hypotheses were evaluated with three computerized studies and an evaluation of psychological state. These tests clarified how participants react to and encode information while exposed to minute amounts of androstadienone, undetectable as an odor, and during a separate carrier control condition. Study 1 measured the automatic attention deployed towards subliminal positive and negative facial expressions by the time it takes to react to the location of a dot probe (Dot Probe Study; MacLeod et al., 1986, Mogg and Bradley, 1999). Study 2 quantified the strength of cognitive and emotional information as distractors in a simple reaction time task, identification of the ink color in which a word is printed (Stroop Study; Stroop, 1935, Gotlib and McCann, 1984). Relative attention to social and nonsocial stimuli that are emotionally neutral was assessed in Study 3, by requiring participants to maintain attention to images while viewing subsequent images (2-back Study; Gevins et al., 1990). Finally, Study 4 assessed participant mood and attentiveness in the absence of extensive social interaction. Performance during control and androstadienone sessions was contrasted to determine the viability of each proposed hypothesis.
The advantage of this suite of studies is that each hypothesis predicts a distinct, cohesive set of behavioral results. The hypothesis that androstadienone enhances general attention predicts that reaction speed will improve with androstadienone for each study, regardless of task or stimulus type. Alternatively, if solely attention to emotional information is improved, reaction times (RTs) will be altered only when emotional stimuli are involved (Study 1 and the Emotional Stroop task in Study 2). Finally, if androstadienone exerts its effects by altering how social information is attended, then RT differences will be solely observed in Study 3, between social and nonsocial stimuli.
Participant mood and perceived attention were assessed to investigate the role of androstadienone in direct modulation of emotional state in a situation with limited social interaction with an experimenter. This evaluation of mood was designed to determine whether androstadienone acts as a direct modulator of the emotional state, particularly for women (Jacob and McClintock, 2000, Bensafi et al., 2004b, Villemure and Bushnell, 2007), or if salient social and emotional cues must be present in order for androstadienone to modulate mood. Furthermore, these data will inform why androstadienone-driven sex differences in mood have been reported. The compound may have a direct, sex-specific effect on mood, or androstadienone may have the same psychological effect on men and women that differentially alters mood due to unrelated sex differences in how the experimental environment is perceived.
Fifty participants (20 men and 30 women) completed the study, ranging in age from 18 to 39 (mean ± SEM: 20.9 ± 0.6). Participants were students and employees recruited through advertisements from the University of Chicago area. They provided informed consent approved by the University of Chicago Biological Sciences Division Internal Review Board and were paid $30 upon completion of the experiment.
Participants did not take hormones (including oral contraceptives), were not smokers and had no history
Study 1. Dot Probe: specific effect of androstadienone on automatic attention to social emotional information
Social feedback from facial expressions is an important source of emotional information. Darwin maintained that these expressions provide a rapid source of nonverbal emotional communication (Darwin, 1872), and indeed, such communication is powerful and automatic: visual attention is preferentially directed to processing emotional faces (Hansen and Hansen, 1988, Stenberg et al., 1998, Eastwood et al., 2001, Vuilleumier, 2002). Emotional faces, therefore, provide a powerful source for testing how
Study 2. Cognitive and Emotional Stroop: effect of androstadienone on attention to positive and negative lexical information
The Stroop effect (Stroop, 1935) occurs when a separate psychological process interferes with the stated goal of the task, inducing competition for attentional resources. The classic cognitive Stroop experiment requires that the participant identify the ink color of presented words (MacLeod, 1991, for review). Participants are slower to identify ink color if the word spells out a different color (e.g. “blue” in red ink) than if the word reads the correct color or is a nonword. Attention to
Study 3. Neutral Face and Shape 2-back: effect of androstadienone on attention to social and nonsocial images
The previous two studies offer convergent evidence that androstadienone modulates processing of emotional stimuli, whether social (faces) or lexical (written word). However, no evidence was provided to reject the hypothesis that androstadienone enhances social cognition as well. To establish that social cognition is not affected, we measured the effect of the steroid on attention to emotionally neutral social and nonsocial images. A common working memory task requires participants to
Study 4. Psychological states: attention and mood and odor perception
Androstadienone typically raises positive mood in women over the course of an experiment (Jacob and McClintock, 2000, Jacob et al., 2002, Lundström and Olsson, 2005), with either no change or a more negative mood for men (Jacob and McClintock, 2000, Bensafi et al., 2004b). The importance of social context (e.g. experimenter gender) in mediating these mood effects (Jacob et al., 2001a), taken with our current conclusion that androstadienone enhances attention to emotional signals, points to
This set of studies indicates that passive inhalation of the natural human compound androstadienone augments attention specifically to emotional information. First, passively inhaled androstadienone drew attention towards subliminal emotional facial expressions presented on the left or the right, shifting attention spatially. The shifted spatial attention decreased reaction time for identifying a dot probe subsequently presented on the same side as the emotional face, and slowed reaction time
We thank the Brain Research Foundation and the Institute for Mind and Biology for funding support. We are also grateful to Joscelyn Hoffman, Elissa Patterson, Hannah You and Emily Hudson for their assistance with this project.