ArticlePDF AvailableLiterature Review

Abstract and Figures

Although sexual behavior is controlled by hormonal and neurochemical actions in the brain, sexual experience induces a degree of plasticity that allows animals to form instrumental and Pavlovian associations that predict sexual outcomes, thereby directing the strength of sexual responding. This review describes how experience with sexual reward strengthens the development of sexual behavior and induces sexually-conditioned place and partner preferences in rats. In both male and female rats, early sexual experience with partners scented with a neutral or even noxious odor induces a preference for scented partners in subsequent choice tests. Those preferences can also be induced by injections of morphine or oxytocin paired with a male rat's first exposure to scented females, indicating that pharmacological activation of opioid or oxytocin receptors can "stand in" for the sexual reward-related neurochemical processes normally activated by sexual stimulation. Conversely, conditioned place or partner preferences can be blocked by the opioid receptor antagonist naloxone. A somatosensory cue (a rodent jacket) paired with sexual reward comes to elicit sexual arousal in male rats, such that paired rats with the jacket off show dramatic copulatory deficits. We propose that endogenous opioid activation forms the basis of sexual reward, which also sensitizes hypothalamic and mesolimbic dopamine systems in the presence of cues that predict sexual reward. Those systems act to focus attention on, and activate goal-directed behavior toward, reward-related stimuli. Thus, a critical period exists during an individual's early sexual experience that creates a "love map" or Gestalt of features, movements, feelings, and interpersonal interactions associated with sexual reward.
Content may be subject to copyright.
Who, What, Where, When (and Maybe Even Why)? How
the Experience of Sexual Reward Connects Sexual Desire,
Preference, and Performance
James G. Pfaus Tod E. Kippin Genaro A. Coria-Avila
`ne Gelez Veronica M. Afonso Nafissa Ismail
Mayte Parada
Published online: 9 March 2012
Springer Science+Business Media, LLC 2012
Abstract Although sexual behavior is controlled by hormonal
and neurochemical actions in the brain, sexual experience induces a
degree of plasticity that allows animals to form instrumental and
Pavlovian associations that predict sexual outcomes, thereby
directing the strength of sexual responding. This review describes
how experience with sexual reward strengthens the development of
sexual behavior and induces sexually-conditioned place and part-
ner preferences in rats. In both male and female rats, early
sexual experience with partners scented with a neutral or even
noxious odor induces a preference for scented partners in sub-
sequent choicetests. Those preferences canalso be induced by
injectionsof morphine or oxytocin paired with a malerat’s first
exposure to scented females, indicating that pharmacological
activationof opioid or oxytocinreceptors can‘‘standin’’for the
sexual reward-related neurochemical processes normally acti-
vated by sexual stimulation. Conversely, conditioned place or
partner preferences can be blocked by the opioid receptor antag-
onist naloxone. A somatosensory cue (a rodent jacket) paired
with sexual reward comes to elicit sexual arousal in male rats,
such that paired rats with the jacket off show dramatic copula-
tory deficits. We propose that endogenous opioid activation
forms the basis of sexual reward, which also sensitizes hypo-
thalamic and mesolimbic dopamine systems in the presence of
cues that predict sexual reward. Those systems act to focus atten-
tion on, and activate goal-directed behavior toward, reward-related
stimuli. Thus, a critical period exists during an individual’s early
sexual experience that creates a‘love map’or Gestalt of features,
movements, feelings, and interpersonal interactions associated
with sexual reward.
Keywords Sexual reward Sexual preference
Conditioning Critical periods Opioids Dopamine
We have evolved a nervous system that acts in the interest
of our gonads, and one attuned to the demands of repro-
ductive competition.
—M. T. Ghiselin (1973)
Erotic fetichism makes an idol of physical or mental qual-
ities of a person or even merely of objectsbecause they
awaken mighty associations withsexual pleasure.
—R. v. Krafft-Ebing (1886/1929)
The first cut is the deepest
—Cat Stevens (1967)
Imagine yourself walking along a busy street when suddenly
you smell something that reminds you of your very first love.
J. G. Pfaus (&)M. Parada
Center for Studies in Behavioral Neurobiology,
Department of Psychology, Concordia University,
7141 Sherbrooke W., Montreal, QC H4B 1R6, Canada
T. E. Kippin
Department of Psychology, University of California
at Santa Barbara, Santa Barbara, CA, USA
G. A. Coria-Avila
Programa de Neurobiologı
´a, Universidad Veracruzana, Xalapa,
H. Gelez
Pelvipharm, Domaine CNRS, Gif-sur-Yvette, France
V. M. Afonso
Department of Psychology, University of Toronto Mississauga,
Mississauga, ON, Canada
N. Ismail
Department of Psychology, Center for Neuroendocrine Studies,
University of Massachusetts, Amherst, MA, USA
Arch Sex Behav (2012) 41:31–62
DOI 10.1007/s10508-012-9935-5
The odor is familiar, comforting, and despite the terrible ado-
lescent angst that accompanied the breakup, it elicits a complete
and happy memory of exciting experiences you shared. To another
person, that smell may well rekindle horrible memories of abuse
and pain. And to a third person, the smell has no meaning what-
soever. Now consider another scenario. You are at a bar with a
good friend watching others and selecting those who are the
sexiest. Although at first you both agree on some rudimentary
symmetries and body types, the discussion quickly degenerates
around particular features: you like brunettes while your friend
likes blondes, you like almond-shaped eyes, while your friend
likes round eyes, you like long faces while your friend likes round
faces, you like a particular set of dimples while your friend cannot
stand dimples, you like a particular size of breasts or pectorals,
while your friend likes skinny. You like the one dressed in leather,
while your friend likes the one in the business suit. It is even worse
when the talk turns to personality characteristics. You like them
loud, unbridled, and maybe a bit dangerous, while your friend
likes them demure, soft-spoken, reliable, and boring. Eventually,
you agree to disagree on who is hot and who is not.
If the ultimate drive for reproductive success causes us to select
potential sex partners with epigamic characteristics that denote
good genetic material, health, wealth, higher social status, and
good potential parental investment (e.g., Buss, 1994), why would
we ever have a difference of opinion? Why would we have ‘a
billion wicked thoughts’’(Ogas & Gaddam, 2011) instead of just a
few that denote reproductive success? Why would we ever need
to learn anything about, or from, sex? We should simply be drawn
to men that are big and strong enough to tear the flesh off bears,
and/or younger women with ample breasts and optimal waist-to-
hip ratios, and then copulate as much as possible, preferably in the
position(s) that provide the most effective stimulation for impreg-
nation, and, of course, only when his sperm count is high and she
is ovulating. But who are we to pick? Do we all possess the same
degree of desired physical characteristics as those we want to
mate with? Of course, the answer to that rhetorical question is no.
Moreover, mating is clearly not the only reason we have sex and,
in fact, may be second to a more proximate master: pleasure (with
Tantric liberation coming a distant third). Indeed, sexual behavior
in some species facilitates sociosexual goals; for example, brief
female–female mounting in bonobos that promotes peaceful
interactions among males (Furuichi, 2011; Hohmann & Fruth,
A nervous system that is prepared to copulate also needs to be
flexible and learn strategies that maximize copulatory success in
an ever-changing world. But what defines success? In order to
have sex, we must first be able to respond to hormonal and neu-
rochemical changes that signal our own sexual desire and arousal
and distinguish it from other sympathetic activation, such as
anxiety.This ability underlies our moment-to-momentlevel of
sexual arousability (as conceived by Whalen, 1966) and
defines a large part of the internal state that is commonly
referred to as‘sex drive.’ Second, we must be able to make
sense of external cues that signal sexual attraction and recep-
tivity in others (e.g., Salu, 2011). This ability requires a complex
mix of instinct, learning, and feedback: a neural organization
that allows for incentive-based motivation and expectancy
(Bindra, 1968,1974; Bolles, 1972;Toates,1986). We must be
able to identify external stimuli that predict where potential sex
partners can be found, to seek out, solicit, court, or otherwise
work to obtain sex partners, distinguish external cues and behav-
ioral patterns of potential sex partners from those that are not
sexually receptive, and to pursue sex partners once sexual
contact has been made (Pfaus, Kippin, & Coria-Avila, 2003). In
this way, we move in time from distal to proximal to interactive,
with an ever-increasing load on sensory processing and motor
sophistication as we interact more and more with a moving and
sentient sensory stimulus.
Neural mechanisms also exist that allow sexual responding
to become habitual or automated with practice and such processes
may underlie the ability of desired features in others to be iden-
tified and responded to quickly (e.g., Ortigue & Bianchi-Dem-
icheli, 2008) and of sexually experienced individuals to be less
affected by treatments that disrupt sexual responding in sexually
¨ve individuals (e.g.,Pfaus & Wilkins, 1995). Similarly, neural
mechanisms exist that allow the stimulation received during sexual
contact to be perceived as rewarding. Such reward alters sub-
sequent behavior, for example, by contributing to the formation
of preferences for salient stimuli associated with positive sexual
reinforcement (Pfaus, Kippin, & Centeno, 2001), just as indi-
viduals develop preferences for stimuli associated with drug or
food reward (Berridge, 2009; Robinson & Berridge, 2001).
Many of these aspects of sexual responding go well beyond the
traditional focus on copulation or genital reflexes/blood flow.
Although some appetitive responses that individuals make prior
to copulation are not specific to sexual behavior, they can be con-
sidered‘‘sexual’’ if theyare conditioned using sexual reward as the
positive reinforcer (Pfaus, 1999). This is as true for appetitive bar
pressing in male rats (e.g., Everitt, Fray, Kostarczyk, Taylor, &
Stacey, 1987) as it is for giving flowers and remembering birth-
days in men. Thus, we learn about who makes us feel desire, about
what we like, and where welikeit,andwelearnwhattodotoget
It is becoming increasingly clear that there is a critical period of
sexual behavior development that forms around an individual’s
first experiences with sexual arousal and desire, masturbation,
orgasm, and sexual intercourse itself. During this period, the sen-
sory and motor mechanics of the behavior become integrated and
crystallized along with the development of preferences for ideal
activities and physical features of a partner. Such preferences
often violate societal‘‘norms’’ (e.g., as in the development of
fetishes or paraphilias) and so-called‘‘evolutionary laws’’ regard-
ing features that represent genetic and reproductive strength,
appearing more to be based on an egocentric evaluation of salient
reward- or pleasure-related characteristics that differ from one
individual to another. The formation of such experience-based
32 Arch Sex Behav (2012) 41:31–62
preferences can be found in historical texts, such as Stendhal’s
‘‘ P r i n c i p l e ’’ i n h i s w o r k De L’Amour (1822), described explicitly
in case histories, such as those in Krafft-Ebing’s Psychopathia
Sexualis (1886/1929), and more theoretically in the‘‘love maps’
proposed by Money (1986). This critical period may well form on
the foundation laid by previous critical periods, especially those
associated with attraction to other- or same-sex individuals, which
itself may form on the foundation laid by a critical period for gender
typical versus atypical behavior, and the sense of self as‘‘female’’or
‘‘ m a l e . ’’
In this article, we propose a neurodevelopmental trajectory of
sexuality that considers how innate predispositions are forged into
successful sexual behaviors, expectations, and desires, based on
an individual’s experience with reward and capacity to link reward
to cues that predict it. The integration of genetic and experiential
factors in determining behavior is consistent with older ideas of the
complementary nature of ultimate and proximate causality (Tin-
bergen, 1963) and new ideas concerning the role of epigenetic fac-
tors (e.g., Bale, 2011). A popular reading of genetic causality seems
to infer that behaviors are fixed or hardwired in the brain, whereas
learning infers that such brain mechanisms and the behaviors they
underlie are flexible. We want to emphasize at the outset that even
so-called‘fixed action patterns’’ are subject to experience and that
subsequent epigenetic changes in gene transcription and protein
synthesis in neurons can alter the expression of behaviors differ-
ently in different individuals. Evolutionary pressures alter the costs
and benefits of any behavior, and experience with reward (and
possibly punishment) maintains the cost-benefit ratio. We note
that this ratio may change in different environmental conditions,
sometimes quickly and radically. Those who can learn to respond
in the wake of sudden changes to a niche (e.g., a decrease in pop-
ulation density) will likely out-reproduce those who do not learn.
Thus, ultimate and proximate causes of behavior are comple-
mentary. We posit that reward mechanisms sit in the middle of
these two domains.
Genetic and Hormonal Conditions and Predispositions
Sexual activity or partner preferences that stray from so-called
‘normal’ are often attributed to genetic or hormonal differ-
ences. Such differences are believed to create different brains
that process the world differently, or to generate physiological
factors that predispose individuals to certain sexual dysfunctions,
such as lifelong premature ejaculation (Waldinger, 2008)orerec-
tile dysfunction (Bancroft & Janssen, 2000).
A number of genetic and neural correlates of male homosexuality
have been reported as‘‘female’’phenotypes (assumed to be due to
either perinatal feminization and/or demasculinization). For exam-
ple, relative to heterosexual men, gay men possess a smaller
female-like interstitial nucleus III of the anterior hypothalamus
(LeVay, 1991), a larger female-like suprachiasmatic nucleus
(Swaab, Gooren, & Hoffman, 1992), and in male-to-female trans-
sexuals a smaller female-like central region of the bed nucleus of
the stria terminalis (Zhou, Hofman, Gooren, & Swaab, 1995). Of
course, it is not known whether those neuroanatomical differences
actually contribute to the sexual orientation of the individual or
whether they simply reflect an unrelated endpoint of some dif-
ferentiated developmental process. Biomarkers have also been
suggested for a distinct (and not putatively feminized or dem-
asculinized) ‘gay’’ genetic makeup in the expression of the
X-linked marker Xq28 (Hamer, Hu, Magnuson, Hu, & Pat-
tatucci, 1993), in the 2D:4D ratio of gay men relative to heter-
osexual men (e.g., Robinson & Manning, 2000), in a leftward
directional handedness asymmetry (Martin, Puts, & Breedlove,
2008), hair-whorl direction (Rahman, Clarke, & Morera, 2009),
visuospatial performance tasks (Collaer, Reimers, & Manning,
2007), and in the fraternal birth order effect (Blanchard, 2007;
Blanchard & Lippa, 2007;Bogaert&Skorska,2011;Rahman,
2005; Vasey & VanderLaan, 2007). Handedness and general IQ
also appear as correlates in incarcerated male pedophiles (Cantor,
Blanchard, Bobichaud, & Christensen, 2005; Cantor, Klassen
et al., 2005), and cerebral white matter deficiencies in the supe-
rior fronto-occipital and right arcuate fasciculi have been found
in incarcerated male pedophiles relative to an incarcerated non-
pedophilic male population (Cantor et al., 2008), suggesting that
frontal output (responsible for executive function and behav-
ioral inhibition) of pedophilic individuals is compromised.
Sex Differences and Steroid Hormones
Another source of innate differences in sexual function involves
the pattern of circulating steroid hormone levels that differ
between the sexes. This occurs in two phases in mammals: (1) an
‘organizational’phase in which the male phenotype differen-
tiates from the female‘‘default’through the combined action of
androgens and estrogens during a critical perinatal period that
‘sculpt’’ a male brain; and (2) during a subsequent‘‘activational’
phase in which gonadal steroid secretion acts on those differ-
entiated brain regions to activate the neuroendocrine systems nec-
essary for reproduction and the behaviors that will bring it about
(Bonthuis et al., 2010; Lenz & McCarthy, 2010; Nelson, 2005).
Thus, the brains of male mammals are differentiated from the
female default by early hormone actions, which lead to different
patterns of sexual responding in adulthood driven by two distinct
steroid hormone secretion patterns, one continuous (male) and
one cyclic (female).
The cyclic nature of estradiol, testosterone, and progesterone
action in females leads to changes in sexual responding and increases
in sexual arousal and desire around the time of ovulation in all ver-
tebrate species, including humans (Slob, Ernste, & van der Werff
ten Bosch, 1991; Stanislaw & Rice, 1988), although a smaller
increase has been reported around the time of menstruation
Arch Sex Behav (2012) 41:31–62 33
(Singer & Singer, 1972). The hormonal milieu during the pe-
riovulatory follicular phase alters the way in which visual sexual
stimuli are processed in women (Gizewski et al., 2006;Krug,
Pihal, Fehm, & Born, 2000;Mass,Ho
¨lldorfer, Moll, Bauer, &
Wolf, 2009;Ruppetal.,2009), which presumably leads to a shift
in the incentive value of the stimuli. Analogous findings have
been reported in our primate cousins, for example, in approa-
ches and solicitations made around the time of the mid-cycle
estradiol peak in rhesus macaques (Wallen, Winston, Gaventa,
Davis-DaSilva, & Collins, 1984) and in the appetitive and con-
summatory sexual behaviors that characterize the periovulatory
period of female rats (e.g., Beach, 1976;Erskine,1989; McClin-
tock, 1984;Pfaff,1980,1999;Pfausetal.,2003). This contrasts
with the relatively stable and continuous testicular androgen secre-
tion in mammalian males (and its reduction to neural andro-
gens or aromatization to neural estradiol in different regions of the
brain) that maintains sexual arousability and responsiveness in a
relatively continuous manner (Hull & Dominguez, 2007).
Biological Differences in Mating Strategy
Although a number of innate differences in mating strategies
between species have been described, perhaps the most studied
is the partner preference displayed by monogamous prairie voles
relative to promiscuous meadow or montane voles (Lim, Ham-
mock, & Young, 2004; Lim & Young, 2004;Winslow,Has-
tings, Carter, Harbaugh, & Insel, 1993; Witt, Carter, & Walton,
1990; Young & Wang, 2004). Differences in functional vaso-
pressin and oxytocin receptor densities have been described
between these two species of vole (Insel, Winslow, Wang,
Young, & Hulihan, 1995) and, in particular, greater vasopressin
1A receptor density in the ventral pallidum (a motor output r egion
that contributes to goal-directed behavior) of monogamous prai-
rie voles compared to promiscuous voles (Lim, Hammock et al.,
2004). Indeed, overexpression of this receptor in meadow voles
by injection of an adenoviral vector that contains the whole recep-
tor gene resulted in male meadow voles developing a significant
partner preference for the first female they mate with (Lim, Wang
et al., 2004). These data suggest strongly that a genetic difference
in receptor density for those two neuropeptides results in a pre-
disposition toward social recognition, affiliation, and pair-bond-
ing, which, in turn, underlies the display of particular reproductive
strategies in a species.
Temperament and Response Style
Finally, what is called‘‘temperament’’ is believed to reflect innate
differences in brain function that start in childhood and lead to
individual differences in sensory processing and motor response
styles. This results in stable personality traits such as extroversion,
introversion, impulsivity, and inhibition (Kagan, 1994). One of
the first classification schemes was that of the ancient ‘‘humoral
doctrine’’of Hippocrates (460–370 BCE). This doctrine described
four human personality traits (sanguine, choleric, melancholic,
and phlegmatic) that allegedly explained both individual and group
differences in responding. Sexually speaking, someone who is san-
guine and prone to humor and passion would likely process sexual
stimulation more readily than the calm phlegmatic person who is
prone to sluggishness. That latter individual may require more sen-
sory stimulation to achieve a requisite level of arousal for sexual
responding and may gravitate to highly arousing, naughty, even
violent sex play. Likewise, an introverted individual may have less
sexual experience than an extroverted individual based simply on
exposure to social situations and concomitant availability of potential
sex partners. The two may possess different sexual attitudes and
different levels of sexual arousal and desire (Meston & Buss,
2007). Indeed, it is possible that whatever the genetic bases of
temperament are also produce differences in the activation pat-
terns of neurochemical systems that excite or inhibit sexual
behavior (Pfaus, 2009). However, introversion in women is cor-
related with a history of childhoodsexualabuse(Talbotetal.,
2000), making it difficult to know what comes first, response
style or particular sexual experiences that may reinforce response
style. Differences in temperament can also change how drugs
affect sexual behavior (P faus, Wilkins, et al., 2010), with one
person’s‘‘aphrodisiacproducing a loss of sexual function in
In summary, innate genetic predispositions are assumed to
underlie different sexual responses and partner preferences.
Those differences may form generally around different hor-
mone actions that sexually differentiate the brain and induce a
differential activation of excitatory or inhibitory neurochemical
systems, which leads to specific types of sensory processing
and/or motor function. Together, they create individual brains
with different sexual phenotypes and proneness to sexual exci-
tation and inhibition, either as a spontaneous ‘‘drive’’ or in the
presence of incentive sexual stimuli. Such differences are believed
to underlie sexual orientation in humans and different species-
specific mating strategies, such as monogamy or polygamy.
Associative Conditions
Another innate characteristic of brains is that they learn and
remember. Brain function (and, therefore, behavior) is modified
to a large extent by experience. The ability to form operant asso-
ciations (Skinner, 1938) between motor actions (responses) and
their consequences (reinforcers) and Pavlovian associations
(Pavlov, 1927) between a neutral external cue that s erves as a con-
ditioned stimulus (CS) and different emotional states like plea-
sure, reward, or aversion, induced by an unconditioned stimulus
(UCS), require a brain capable of experience-based synaptic plas-
ticity, as in the formation of Hebbian circuits and neural networks
that sensitize neural responses to stimuli (Hebb, 1949; Johnson,
34 Arch Sex Behav (2012) 41:31–62
LeDoux, & Doye
`re, 2009). There are a number of molecular
events that shape such plasticity. These include: (1) the activation
of growth factors that direct the formation of new synapses; (2)
augmentation of intracellular events, such as transmitter receptor
proliferation and the availability of second messengers linked to
receptors; and (3) epigenetic factors, including DNA methylation
or histone acetylation of genes to suppress or enhance the gen-
eration of proteins they code for that ultimately facilitate or inhibit
individual responses that comprise the behavior. All three pro-
cesses are stimulated and sensitized during phases of neuronal acti-
vation following the first encounter with a new sensory stimulus
(see, e.g., McEwen, 2010). The emotional state induced by the
consequences of that encounter feed back to strengthen or inhibit
the likelihood of future encounters. Because the molecular pro-
tein synthetic events that subserve those phenomena take time,
along with a degree of repeated experience with the conditions,
the final form of the response accrues across a ‘‘learning curve’
that can be defined individually or for a group of individuals
exposed to the same stimuli and consequences. Regardless of
the species, all neurons store the memory of salient events in this
way (Kandel, 2001). Thus, activation of the network itself by a
formerly ‘‘neutral’’ CS associated with it generates, as Pavlov
said, a‘‘representation of the UCS’’which serves to predict, acti-
vate, and prime the emotional state. An incentive motivational
account of Pavlovian conditioning (e.g., Bindra, 1968)hasthe
CS activating a ‘‘central motivational state,’’ which moves the
body in the direction of the goal (e.g., activates appetitive
responses) and prepares it viscerally to act on the goal (e.g.,
salivation in response to a CS that predicts food; erection in
response to a CS that predicts sexual gratification).
As with Pavlovian conditioning, a motor response that results
in the arrival of a needed or preferred stimulus becomes strength-
ened, such that doing it on its own after it is learned can be rein-
forcing to a certain extent. Motor patterns also become automated
with repeated activation, as in the development of ‘‘habits’’ or
‘routines’’ that require very little cognitive loading to be engaged
once they are formed (Hull, 1943; Tolman, 1932). This also makes
them extremely difficult to break without an extensive program of
extinction and/or alternate reinforcement and, even then, spon-
taneous recovery can occur, showing the primacy of early relative
to subsequent experiences. Thus, associative conditioning creates
a brain in which the ends‘‘justify’’the means and in which the cues
that predict the ends become conditioned incentives. To a brain
that generates some degree of self-awareness (as in humans),con-
ditioning results in circular reasoning and tautologies. When do
we ever thank our bodies and ourselves for our pleasurable responses
to salient erotic stimulation? Instead, we ascribe their valence and
intensity to our sex partner(s), as in the phrase‘‘you make me feel
so good.’ Of course, when such stimulation loses its arousing
potential, as often happens when it is repeated over and over again
the same way (and at intervals that produce tolerance, rather than
sensitization), we tend to blame that loss on our sex partners no
longer arousing us as they once did.
Setting the Stage
If we were creating a sexually reproductive universe, would we
make sexual responses as‘‘innate’’ as possible to assure that every-
one would do the same thing at the right time with presum-
ably the most optimal mate(s)? After all, like a male black widow
spider or praying mantis, we may get only one crack at it. Yet,
anyone who has had the opportunity to watch utterly inept sex-
ually naı
¨ve male rats try to copulate no doubt wondered why there
are so many rats. Like language, sexual desire, preference, and
performance must be learned and crystallized before they appear
as a relatively‘‘automatic’’set of species-specific responses. Essen-
tially, genetic factors and early hormonal differentiation help to
set the stage upon which experience creates the play. But the
actors have to rehearse before the play can be performed properly.
The sexual brain must have feedback systems that coordinate,
facilitate, and crystallize the learning of appropriate and compe-
tent sexual responses. Those systems must link together the acti-
vation of both sympathetic and parasympathetic divisions of the
autonomic nervous system by desired sexual stimuli with appe-
titive sexual behaviors directed at those stimuli and drive them
into a coordinated set of consummatory (e.g., copulatory) responses.
What kind of feedback is required and how can this be studied
Obviously, human sexual behavior is best studied in humans,
but it is impossible to control the conditions under which humans
have their first sexual experiences. Far too much occurs randomly
and chaotically despite the best-laid plans. Therefore, we depend
on animal models in which different aspects of sexual experience
can be controlled methodically (for comprehensive reviews, see
˚gmo & Ellingsen, 2003;Pfausetal.,2003). Sexual behavior in
animals, like that of humans, can be divided along appetitive and
consummatory lines (Pfaus, 1999; Pfaus, Damsma, et al., 1990)
and can be studied with respect to the stimuli animals are given to
respond to and the motor patterns that become strengthened and
automated with repeated sexual experience. Although the behav-
ioral structure of sexual responding may be the same for all
vertebrates, it is true that laboratory rats are not humans. Drawing
conclusions from one species and applying it to another must
be done with caution and an eye toward predictions made
from one species about the functions of another, e.g., mecha-
nisms underlying erection in males or sexual solicitations in
females (Beach, 1950; Pfaus et al., 2003). We do not know if rats
experience ‘‘orgasm’’ as we do, although at another level of
analysis they clearly experience a sexual reward state that aug-
ments their responsiveness to cues associated with it.
Learning What to Do
In his seminal analysis of conditioning and copulatory behavior in
male rats, Larsson (1956) demonstrated learning curves for base-
line rates of mounts, intromissions, and ejaculations in male rats.
Baseline rates were typically achieved in 3–5 trials to ejaculation
Arch Sex Behav (2012) 41:31–62 35
and were faster if males achieved more ejaculations in their initial
trials. Notably, sexually naı
¨ve males typically took a long time to
sniff and lick the female’s anogenital region before mounting
whereas some males never copulated at all despite repeated expo-
sure to females and their progressively intense attempts to solicit
copulation. Some sexually sluggish or non-copulating males
(referred to colloquially as‘‘duds’relative to the‘‘studs’’ that cop-
ulate) could be stimulated to mount if unavoidable, low-level
footshock, short-term pain (e.g., tail-pinch), or neutral stimuli
paired with them were applied (Barfield & Sachs, 1968;Cag-
giula, 1972; Crowley, Poplaw, & Ward, 1973). Such stimuli also
reduced the number of intromissions required for ejaculation in
sexually active males (Beach & Fowler, 1959; Sachs, Macaione,
& Fegy, 1974).
In other experiments, repeated exposure of the males to the
copulation testing chambers for severaldays prior to their first
exposure to the female increased the proportion that mounted
to 100%, suggesting thatthe stress of a novel environment may
preclude copulation in sensitive males (Pfaus & Wilkins, 1995).
Endogenous opioids are activated in response to novelty
stress (Izquierdo & McGaugh, 1987) and may disrupt copulation
in sexually naı
¨ve males by lreceptor activation in an important
hypothalamic region known as the medial preoptic area (mPOA)
(Hughes, Everitt, & Herbert, 1990; Matuszewich & Dornan,
1992). Administration of the opioid receptor antagonist naloxone
to sexually naı
¨ve males increased the proportion of males that cop-
ulated on their first test in a manner identical to that of preexposure
to the testing chamber (Pfaus & Wilkins, 1995). In fact, naloxone
had no effect in that study if males had been preexposed. In addition
to genetic reactivity to novelty stress, sexually sluggish males were
found to have decreased nuclear estrogen receptor in the mPOA
relative to sexually active males (Clark, Davis, & Roy, 1985), sug-
gesting a genetic factor that would either fail to activate excitatory
pathways, or fail to disinhibit them in the presence of stressful
inhibitory external cues.
Learning the appropriate set of appetitive responses is crucial to
gaining access to sex partners and learning how to copulate prop-
erly is crucial to successful reproduction. Both types of responses
become ‘‘crystallized’’ in an animal’s behavioral repertoire.
Many examples of this exist in the literature. In rats, those behav-
iors include performance in obstruction boxes (Jenkins, 1928;
Moss, 1924; Stone, Barker, & Tomlin, 1935; Warner, 1927),
straight-alley running (Beach & Jordan, 1956; Sheffield,
Wulff, & Backer, 1951;Ware,1968), maze learning (Drewett,
1973; Eliasson & Meyerson, 1975; Hetta & Meyerson, 1978;
Kagan, 1955; Meyerson & Lindstrom, 1973; Warner et al.,
1991; Whalen, 1961), crossing of electrified grids (Moss,
1924), nose-pokes and other attempts to‘get to’a potential sex
partner behind a wire-mesh screen (Damsma, Pfaus, Wenkstern,
Phillips,& Fibiger,1992; Pfaus, Damsma, Wenkstern,& Fibiger,
1995; Pfaus, Mendelson, & Phillips, 1990), digging through sand
(Anderson, 1938), bar-pressing for a sex partner (Beck, 1971,
1974,1978; Beck & Chmielewska, 1976; French, Fitzpatrick, &
Law, 1972; Jowaisas, Taylor, Dewsbury, & Malagodi, 1971;
Larsson, 1956; Sachs et al., 1974; Schwartz, 1956)orforacue
light associated with the arrival of a sex partner (Everitt, 1990;
Everitt & Stacey, 1987; Everitt et al., 1987), and psychomotor
stimulation in anticipation of the arrival of a sex partner (Men-
delson & Pfaus, 1989). To gain access to receptive females, male
guinea pigs lean to run an alley (Seward & Seward, 1940), male
pigeons learn to peck keys (Gilbertson, 1975), and male stickle-
back fish learn to swim through rings (Sevenster, 1973). Male
songbirds learn to sing particular songs that attract females
(MacDougall-Shackleton, 2009;Mooney,2009).
Larsson (1956) also foundthat, once established,individual
patterns of copulatory behavior appear stable and relatively resis-
tant to environmental change. Similarly, Pattij et al. (2005) demon-
strated stable individual differences in ejaculatory latency in male
rats that resembled those found in humans using the stopwatch
method to determine intravaginal ejaculatory latency (e.g., Wal-
dinger, McIntosh, & Schweitzer, 2009), and could be assigned
statistically to groups of rapid, normal, and delayed ejaculators.
In fact, such assignment predicted certain drug effects, such as
the facilitation of ejaculation by the presynaptic 5-HT1a agonist
drug 8-OH-DPAT (which occurred in normal and delayed ejacu-
lators, but not in rapid ejaculators). However, the early conditions
under which copulatory patterns are established can alter the way
that they appear in subsequent tests. For example, Larsson dis-
covered an ‘‘enforced interval effect’ (EIE) in which males that
had the receptive female pulled out of the test arena for durations
of up to 1 min would ejaculate with progressively fewer intromis-
sions. Once established, this effect would last despite the fact that
males had free access to females. However, although males given
early experience with free copulation showed an EIE within a cop-
ulatory session, they would resume their ‘‘normal’’ pattern under
conditions of free copulation.
In an elegant and well-controlled extension of the EIE,
Silberberg and Adler (1974) trained sexually naı
¨ve male rats to cop-
ulate in three different conditions. Control males were allowed
to copulate freely with sexually receptive females for 30 min on
each of 20 tests of sexual behavior. Experimental males were
allowedtoachievesevenintromissions before the receptive female
was removed during each test. A third group was yoked in time to
the experimental group, but not in terms of a particular number of
intromissions prior to the female being removed. A large and sig-
nificant proportion of males in the experimental group consistently
achieved ejaculation prior to the 7th intromission relative to males
in the other groups. Intromissions thus could be interpreted as ‘‘oper-
ants’’that brought about ejaculation and the intromission pattern
(along with the ejaculation latency) could come under external
To examine the development of particular patterns of sexual
responding, we have used unilevel pacing chambers with two
36 Arch Sex Behav (2012) 41:31–62
types of Plexiglas divider, one with four holes and another with
one hole, that separate the chamber into two equal sides. The
holes are small so that only the female can cross from one side to
the other. This sequesters the male to one side and allows the
female to control or‘pace’’ the initiation and rate of copulation.
However, this creates two distinctive pacing conditions, one in
which the female has free access to the male (4-hole) and one in
which she has relatively restricted access (1-hole) given that the
male typically puts his head in the a hole (presumably trying to
get to the other side). Of course, from the male’s perspective, the
female is always available in the 4-hole condition whereas he
has to wait longer in the 1-hole condition. Male rats given their
first nine trials of sexual activity in one of the two pacing con-
ditions developed significantly different patterns of copulatory
activity (Ismail, Zhao, & Pfaus, 2008). Males trained in the
4-hole condition developed faster ejaculation latencies and
fewer intromissions before ejaculation relative to males
trained in the 1-hole condition(Fig. 1). Notably, when males in
each group were switched to the other condition on the 10th
trial, they maintained their previously-established ejaculatory
pattern, despite the difference in condition.
Beach (1956) proposed that two separate but interactive
sexual mechanisms existed in male rats, an arousal mechanism
(AM) and a copulatory mechanism (CM). The AM integrated
distal olfactory, auditory, and visual cues from receptive females.
When the cue strength became sufficiently intense, the AM
activated the CM to initiate copulatory responding (mounts and
intromissions). The CM then integrated tactile stimulation from
the penis with each vaginal intromission, leading eventually to
ejaculation and the highly stereotyped post-ejaculatory refractory
period. Given that sexually naı
¨ve males can show preferences for
receptive over non-receptive females and that they are responsive
to sex odors from receptive females, the first unconditioned rein-
forcer in the cascade of sexual reinforcers would be sex odors.
The goal would be getting to them. In Beach’s terms, this would
be the primary function of the AM.
Fig. 1 Development of male rat sexual behavior in 1-hole vs. 4-hole
unilevel pacing conditions. aAverage frequencies of mounts, intromis-
sions, and ejaculations during the nine conditioning trials in the two
pacing conditions. bAverage mount, intromission, and ejaculation
latencies during the nine conditioning trials in the two pacing conditions.
cAverage time (s) the females spent away from males during the nine
conditioning trials in the two pacing conditions. dEjaculation latencies
in the two pacing conditions across the nine conditioning trials and
during the 10th test when the males were switched into the other pacing
Arch Sex Behav (2012) 41:31–62 37
Whalen (1961) asked what the necessary stimulation must be
for the development of copulatory behavior. The answer was penile
stimulation. Whalen varied whether males achieved mounts with-
out intromission, intromissions without ejaculation, or intromis-
sions with ejaculation with sexually receptive females. On a final
test, males were allowed to copulate to ejaculation with receptive
females. Many rats that achieved only mounts during their sexual
experience trials did not copulate whereas rats that achieved in-
tromissions with or without ejaculation were able to copulate to
ejaculation normally. Thus, exposure to sex odors alone was not
sufficient to crystallize patterns of copulation; sensory feedback
from penile stimulation was necessary. This makes penile stim-
ulation a second goal or reinforcer in the cascade (and the first of
the CM).
Whalen also asked whether mounts with or without intro-
mission could augment the ability of male rats to learn to run a T
maze and find a sexually receptive female. Males with 1 or 4
mounts alone, or 1 or 4 intromissions, both scored higher than
chance (choosing the sidewith the receptivefemale more often
than the non-receptive female). However, experience with intro-
missions produced a significantly higher percentage of ‘‘correct’
responses, indicating that tactile stimulation of the penis summed
with mounts (and other close interaction with the female during
training) to augment the learning. Lo
´pez, Olster, and Ettenberg
(1999) asked whether sexually naı
¨ve rats would run faster in a
straight-arm runway if their prior copulatory experience was
intromissions with or without ejaculation. Only rats that achieved
ejaculation showed faster running times. Thus, it would seem that a
cascade of reinforcing events, from perception of sex odors to
chasing receptive females to penile stimulation during mounting to
ejaculation, is necessary for the normal display of appetitive and
consummatory sexual responses.
Resistance to Disruption
Sexual experience has profound effects on physiology and makes
animals relatively resistant to treatments that disrupt sexual behav-
ior. For example, compared with sexually inexperienced males,
sexually experienced males have larger testes (Drori & Folman,
1964), heavier penises (Herz, Folman, & Drori, 1969), lighter
body weights (Siegel, Nunez, & Wade, 1981), and increased secre-
tions from accessory sex glands (Drori & Folman, 1964). Sex-
ual experience has also been shown to ameliorate the disruptive
effects of anosmia (Thor & Flannelly, 1977), castration (Lisk &
Heiman, 1980;Fig.2), penile deafferentation (Lodder, 1975),
penile anesthesia with lidocaine (Fig. 3), and age (Gray, Smith,
Dorsa, & Davidson, 1981). Gonadally intact, sexually experi-
enced males prefer the odors of receptive females over those of
non-receptive females whereas sexually naive or castrated males
do not display a significant preference (Carr, Loeb, & Dissinger,
1965; Carr, Loeb, & Wyllie, 1966).
We have shown that sexually naive males are susceptible to
the disruptive effects of novelty stress on copulation whereas
males with 10 prior multi-ejaculatory sexual experiences are not
(Pfaus & Wilkins, 1995). Although both sexually naive and expe-
rienced males that were placed into a novel testing apparatus
displayed behaviors consistent with a fear state (e.g., freezing
followed by exploration near the walls of the chamber), the sex-
ually experienced males responded immediately to the intro-
duction of a sexually receptive female with the initiation of copu-
lation whereas most of the sexually naive males ignored the
receptive female for a long period of time, and, as noted above,
Fig. 2 Effects of castration with or without dihydrotestosterone treatment
on consummatory sexual behaviors of male rats. Top Proportion of males dis-
playing mounts, intromissions, and ejaculations a month after castration.
Bottom Frequencies of mounts, intromissions, and ejaculations a month
after castration. Maleswere given either 1 or 10 trialsof sexual behaviorto
one ejaculation each prior to castration. *p\.05 compared to oil-treated
38 Arch Sex Behav (2012) 41:31–62
some sexually naı
¨ve males did not initiate copulation, although
novelty-induced opioid transmission in certain regions of the
brain leads to an endogenous state in which copulatory behavior
is suppressed, the induction of this state in males with sexual
experience was not sufficient to suppress copulation once the
female was placed into the chamber.
Learned Inhibition
Sometimes animal have to learn what not to do. Sexually expe-
rienced male rats learn not to attempt copulation with sexually
non-receptive females. This was shown when males that had
10 prior multiejaculatory trials at 4-day intervals with females
were presented with a sexually non-receptive female (Pfaus &
Pinel, 1989). All males attempted to mount those females despite
their attempts being thwarted aggressively by the females’ defen-
sive behaviors. When males were presented sequentially with sex-
ually non-receptive and receptive females in alternating trials, the
males learned not to mount the non-receptive females over the
course of approximately five trials while maintaining high rates of
sexual responding on trials with sexually receptive females. Inter-
estingly, a low dose of alcohol increased the proportion of males
that attempted to mount the non-receptive females in a subsequent
test, showing that alcohol possesses disinhibitory properties only if
animals were inhibited in the first place.
When presented with a sexually inactive male rat, sexually
receptive and experienced females will solicit the males with
increasing vigor before mounting them repeatedly (Afonso &
Pfaus, 2006;Beach,1968). This ‘‘female–male mounting’’
(FMM) behavior was displayed only when the gonadally-intact
females were sexually receptive, when ovariectomized (OVX)
females were primed with estrogen, and was terminated abruptly
if the males finally‘‘got it’’and mounted back. Thus, it is regarded
as a‘‘super-solicitation’’behavior. Importantly, FMM is displayed
by sexually naı
¨vefemales presented with a castrated, sexually inac-
tive male on their first exposure to males. However, if the males
never mounted back (as the long-term castrates do not), the FMM
behavior diminished over five successive trials (Afonso, Bablekis,
&Pfaus,2006). It is as if the females learned that this behavior does
not successfully stimulate the males to mount them, although it
could be restored if the males mounted intermittently. Fully recep-
tive female rats also learned to suppress both appetitive and con-
summatory aspects of sexual behavior if they are paired with an
explicitly nonrewarding state. We have created such a state with
injections of the opioid receptor antagonist drug naloxone. Females
were given their first six experiences of copulation with sexually
vigorous males at 4-day intervals in unilevel pacing chambers
either under the influence of naloxone or saline. On the 7th test,
all rats were injected with saline. Females that had experienced
sex without opioid reward showed dramatic reductions in solic-
itations, hops and darts, lordosis, and increased defensive responses
compared to control rats and received fewer intromissions and
ejaculations from the males (Fig. 4). As the half-life of naloxone is
approximately 1–2 h in plasma (Fishman, Roffwarg, & Hellman,
1973), it is highly improbable that the reduction was due to a long-
term or sustained action of naloxone itself. It is more likely that it
stemmed from an expectancy of nonreward during sex.
Neural Correlates
Whereas critical periods of perinatal and pubertal steroid hor-
mone action alter brain structure and function within hypotha-
lamic, limbic, and cortical regions associated with sexual arousal
and desire, experience with salient sexual stimulation (and reward)
also changes or refines brain structure and neurochemical function
(Pfaus, 2009). For example, whole-brain endorphin content
increases significantly in male rats following repeated ejacula-
tions (Szechtman, Hershkowitz, & Simantov, 1981). Copulation of
sexually naı
¨verats to one ejaculation results in sensitized dopamine
release and greater induction of the immediate-early gene product
Fos (a marker of neuronal activation) in the nucleus accumbens
(NAc) in response to a female behind a screen, relative to males
allowed to intromit only during their first experience or that remain
sexually naı
¨ve (Biały & Kaczmarek, 1996; Bradley & Meisel,
´pez & Ettenberg, 2002). Males allowed 10 multiejacu-
latoryexperiences with sexuallyreceptive females haveincreased
numbers of dendritic spines (and therefore synapses) in the NAc
relative to males allowed one multiejaculatory experience. Males
Fig. 3 Effect of the topical anesthetic lidocaine (5% topical cream) or saline
applied to the penis of male rats on the proportion of males that mount, intro-
mit, and ejaculate as a function of prior sexual experience. Males received
either 1, 2, 3, 4, 5, or 10 prior experiences to one ejaculation before the appli-
cation of lidocaine or saline. For percentages of 60 or lower, p\.05
Arch Sex Behav (2012) 41:31–62 39
allowed 16 multiejaculatory experiences have increased glucose
metabolism in limbic structures relative to sexually naı
¨ve males
or those allowed three multiejaculatory experiences (Sakata,
Gonzalez-Lima, Gupta, & Crews, 2002). Those data suggest that
experience with ejaculation in males activates endogenous opioid
reward systems and sensitizes mesolimbic systems associated with
incentive motivation in male rats (Robinson & Berridge, 1993).
Experience with paced copulation in females appears to activate the
samereward and incentive systems. Females that mount sexually
inactive males show increased Fos induction in the mPOA, ven-
tromedial hypothalamus (VMH), and medial amygdala, but
females that either do not mount or that have learned not to mount
over five trials with castrated, sexually inactive males do not show
Fos induction in those regions (Afonso, Lehmann, Tse, Woehr-
ling, & Pfaus, 2009).
Sexual experience also activates the cerebellum. Sexually
¨ve male rats allowed to copulate to one ejaculation at 4-day
intervals showed activation of the cerebellar vermis (notably lobule
7) whereas males given non-copulatory exposure to females had
relatively little activation (Manzo et al., 2008). Using an identical
paradigm, dynamic increases were found in GAD-65 (assessed by
RT-PCR), the enzyme that converts glutamate into the inhibitory
neurotransmitter GABA, during the first three ejaculations expe-
rienced by males that were sexually naı
¨ve at the beginning of the
experiment (Bolivar-Duarte, Silva, Manzo, & Pfaus, 2012); how-
ever, this increase had dropped back to baseline levels by the fourth
ejaculation. Thus, the acquisition of baseline rates of sexual behav-
ior in male rats appears to correspond to dynamic reorganization of
the cerebellar vermis, consistent with the critical role of the cerebellum
in motor habit learning (Salmon & Butters,1995). It is tantalizing
to consider that the particular pattern of sexual responding that
crystallizes in animals is mediated by dynamic changes in cere-
bellar function during their initial sexual experiences.
Learning Where to Do It
Whereas the crystallization of responses to unconditioned sexual
incentives and copulatory behavior itself appears to be condi-
tioned by genital stimulation, learning where to find sex partners,
or optimal places to experience copulation, are conditioned by sex-
ual pleasure or reward. Contextual factors, such as settings, are
important components of positive sexual experiences for both
men and women (e.g., Basson, 2001;Hoon,1984;Kinsey,
Pomeroy, & Martin, 1948; Kinsey, Pomeroy, Martin, & Geb-
hard, 1953; McCarthy, 1977). Salient cues in the environment
may be associated with sexual reward in such a way that they
increase arousal or desire directly in their presence. Such is the
case with places that are associated with sexual gratification.
Animals often display a preference to remain in a context that
has been paired consistently with access to a reward (e.g., drugs
of abuse, highly palatable foods, a mate) over a context that has
not. This conditioned place preference (CPP) is typically demon-
stratedinanapparatuswithtwodistinctive c ompartments that are
connected to either side of a third neutral compartment (Fig. 5).
During training, the compartments are paired differentially with
unconditional stimuli (e.g., one side is paired with a sex partner,
food, or a rewarding drug, and the other side is paired with either
Solicitations LordosisMagnitude 3
Hops and Darts Intromissions
Defensive Responses Ejaculations
Fig. 4 Effects of acquiring
sexual experience under the
influence of saline or naloxone
(5 mg/kg, ip) on appetitive and
consummatory sexual behaviors
in OVX female rats primed fully
with estradiol and progesterone.
Females received six
multiejaculatory experiences at
4-day intervals prior to the final
test in which all rats received an
injection of saline
40 Arch Sex Behav (2012) 41:31–62
nothing or a control manipulation). On the final test, the animal is
placed into the neutral compartment with the two doors on either
side opened to allow free access to either compartment. CPP is
said to have developed if the subject spends significantly more
time in the reward-paired compartment than the other compart-
ment. Stimuli or events that are capable of supporting CPP are
referred to as‘‘rewards’’rather than ‘‘reinforcers,’’because the sub-
ject has never been required to move into the paired compartment
to experience them. Thus, CPP is not reinforced, per se, because it
is displayed spontaneously on the final test. However, the increased
time spent in the side paired with reward is clearly conditional upon
the Pavlovian association of those contextual cues with the reward
In male rats, sexual CPPs have been established using two dif-
ferent conditioning procedures. In one, copulation to ejaculation
is allowed to occur within one of the distinctive environments and
this environment is subsequently preferred over one in which no
copulation occurred (Everitt, 1990). CPPs developed by this
procedure are referred to as‘‘copulatoryCPPs.’’Copulatory CPPs
can be maintained by intromissions alone whereas prevention of
intromission disrupts a previously established CPP (Hughes et al.,
1990). In a second procedure, male rats are allowed to copulate to
ejaculation in a separate testing chamber and are then transferred
immediately to one distinctive compartment of the CPP apparatus.
As with the first testing procedure, the other distinctive compart-
ment is paired on intermediate days with a control condition (usu-
allyno copulation). Followingsuch training, the compartment
paired with copulation is preferred over the other compartment
˚gmo&Berenfeld,1990). A CPP induced bythisprocedure
isreferred to asa‘‘post-ejaculatoryCPP.’’Demonstrations of post-
ejaculatoryCPPs might appearpuzzling at first glancebecause the
CS (i.e., the distinctive environment) is presented after the UCS
(copulation to ejaculation), in what learning theorists call‘‘back-
ward conditioning’ (that would not be expected to yield condi-
tional responding to the environment). However, if the neural
reward state induced by ejaculation is the unconditional stimulus,
then the pairing of environmental cues with it is simultaneous.
Thus, post-ejaculatory CPP can be accounted for by the rules of
Pavlovian conditioning.
CPPs have also been demonstrated in female rats and ham-
sters. Oldenburger, Everitt, and de Jonge (1992) found that when
copulation occurred within one of the distinctive compartments
of a CPP apparatus, female rats showed a weak CPP. Subse-
quently, Paredes and Alonso (1997) and Paredes and Vazquez
(1999) demonstrated a robust CPP in female rats that depended on
whether the females were able to pace the rate of copulation with-
Black Wall
Bilevel pacing chamber
Unilevel pacing chamber
Place preference apparatus
Fig. 5 Apparatus used to test CPP. Rats are allowed different types of copu-
latory experience in either bilevel or unilevel pacing chambers after which
they are transferred to one side of a CPP box. On alternating trials, rats receive
15 min of no copulation in the pacing chamber, after which they are
transferred to the other side of the CPP box. On the final test, rats are placed
into the central start compartment, the guillotine doors raised, and rats are
allowed to roam freely from one side to the other over a specified amount of
time. The amount of time spent in each compartment is marked by photo-
beam breaks between the compartments, and tabulated automatically
Arch Sex Behav (2012) 41:31–62 41
out having to employ defensive behaviors. This was accomplished
using unilevel pacing chambers bisected by a Plexiglas divider with
one or more small holes that only the female can pass through
(Erskine, 1985,1989; Paredes & Alonso 1997; Paredes &
Vazquez, 1999). The male was sequestered on one side of the cham-
ber and the female was then free to pace the copulatory contact by
running from side to side whenever she wanted. Like males,
females acquired a strong preference for a distinctive environ-
ment only if they were placed into the CPP box immediately after
paced copulation. No preference was found if the copulation was
unpaced prior to placement in the CPP box (meaning that it had
occurred in the same pacing chamber but without the divider).
Thus, for a female rat, CPP develops only if she has been able to
control the initiation and rate of copulation freely without hav-
ing to use defensive behaviors.
Although a sexually vigorous male rat is a clear UCS for
approach and solicitation in female rats (e.g., A
˚gmo, 1999), contex-
tual cues associated with pacing elicit a conditioned sexual reward
state in those females. However, these results may also indicate the
presence of an unconditional aversive state during unpaced copu-
lation. To examine this possibility, Afonso, Woehrling, and Pfaus
(2006) allowed female rats to copulate in two unilevel pacing
conditions using Plexiglas dividers that had either 4 holes or 1 hole.
This was doneto eliminate the possibility of an‘‘aversive’’state
resulting from unpaced copulation. Trials were conducted sequen-
tially at 4-day intervals and each pacing condition was paired with
one of the distinctive sides of a CPP apparatus, in a counterbalanced
fashion. Control groups contrasted the 4-hole or 1-hole condition
with a no-divider condition (as was done by Paredes & Alonso,
Control females developed significant CPP for either the
1-hole or 4-hole condition, relative to unpaced copulation with no
divider. Those control data replicate the findings of Paredes and
Alonso (1997) and indicate that both the 4-hole and 1-hole con-
dition were rewarding relative to the unpaced (no divider) con-
dition. However, they do not rule out the possibility that the real
distinction being made was between an aversive condition (unp-
aced copulation) and a rewarding condition (paced copulation).
This was addressed in the group allowed to contrast the 4-hole vs.
1-hole condition. In this group, females developed significant
CPP for the 4-hole condition relative to the 1-hole condition, sug-
gesting strongly that copulatory CPP reflects a true sexual reward
state in females. Similarly, Jenkins and Becker (2003) found that
female rats developed significant CPP for paced relative to unpaced
mating, butalso for unpaced matingin which the experimenter
removed the male for a period that approximated the female’s
imposed interintromission interval,relative to unpaced mating
in which male removal did not occur. Thus, female rats develop
CPP for sex at their own preferred intervals. Taken together with
the results of Matthews et al. (1997), these data suggest that reward
comes from the sexual stimulation that females receive, namely
mounts with intromission, so long as that stimulation occurs at the
desired time intervals.
What about paced copulation leads to CPP in females? Meerts
and Clark (2009) reported that artificial vaginocervical stimulation
(VCS) applied with a 1 ml syringe plunger at 200 g of pressure for
2 s at 30-s intervals, for a total of 15 stimulations, induced a reli-
one. Given that VCS stimulates both internal clitoris and cervix,
we asked whether clitoral stimulation (CLS) alone could induce
CPP (Cibrian-Llanderal et al., 2010; Parada, Chamas, Censi, Coria-
Avila, & Pfaus, 2010). In these studies, CLS was administered either
with a lubricated paintbrush or a small cotton-tipped vibrator at
preferred intervals for 10–15 min over 5–6 reinforced sessions.
Both types of stimulation induced robust CPP. Importantly, clito-
ral reward can be induced in OVX females with or without hor-
mone priming (Parada, Abdul-Ahad, Censi, Sparks, & Pfaus, 2011),
indicating that sexual reward is independent of steroid priming,
although such priming would normally be required for females
to experience CLS from mounts with pelvic thrusting (Pfaff,
Montgomery, & Lewis, 1977).
Domjan and Hall (1986)demonstratedthatmaleJapanese
quail will stay in the vicinity of a window in their home cage
through which they could see a sexually receptive female during a
pre-copulatory period. However, this behavior developed only if
the males had the opportunity to copulate with the female after the
pre-copulatory period. A variant of this procedure, similar to that
used to study anticipatory motor responding in rats, was used by
Balthazart, Reid, Absil, Foidart, & Ball (1995) to study the role of
hormones and brain dopamine systems in conditioned sexual b ehav-
ior in quail. Male quail were placed into a chamber that contained
a window and sliding door at one end through which the male could
see a sexually receptive female. After a 10-min period, the sliding
door opened and the animals could interact freely. As in Domjan
and Hall (1986) and Mendelson and Pfaus (1989), only males that
copulated with the females during this period developed the behav-
ior, in this case, a preference to stay close to the window in thepre-
copulatory period of subsequent tests. Castrated males did not
develop this conditioned proximity behavior nor did males that
did not copulate. Castration also reduced the time spent near the
window males trained prior to castration, and subsequent replace-
ment with testosterone or estradiol restored the behavior. Sub-
sequently, Castagna, Ball, and Balthazart (1997) reported that nom-
ifensine, a dopamine re-uptake inhibitor, decreased the appetitive
social proximity response, but increased the frequency of mount
attempts. In contrast, amfonelic acid, a compound that enhances
dopaminergic tone, increased aspects of both appetitive and consum-
matory sexual behaviors. Thus, brain dopamine systems in birds and
mammals seem to have analogous functions in the control of
appetitive or conditioned sexual approach behaviors.
Although both copulatory and post-ejaculatory CPP pro-
cedures produce effects of similar magnitude in male rats, there
are differences in the underlying neurobiology. The opioid recep-
tor antagonist naloxone disrupts both copulatory and post-ejac-
ulatory CPPs, but in different ways. A
˚gmo and Berenfeld (1990)
found that the development of post-ejaculatory CPP was blocked
42 Arch Sex Behav (2012) 41:31–62
by injections of naloxone prior to each training session. Conversely,
the development of a copulatory CPP was unaffected by naloxone
prior to each training session (Meharra & Baum, 1990). However,
once a copulatory CPP had developed, its expression was blocked
by naloxone injections prior to the final test (Hughes et al., 1990;
Meharra & Baum, 1990).
There is also evidence that the site of action of naloxone is
different forthese effects. A
˚gmo and Gomez (1993) found that
naloxone’s disruption of the development of post-ejaculatory
CPP occurred following infusions into the mPOA whereas infu-
sions of naloxone into this brain region did not disrupt the expres-
sion of a copulatory CPP (Hughes et al., 1990). In females, pac-
ing-related CPP was blocked by systemic injections of naloxone
(Paredes & Martinez, 2001), or following infusions of naloxone
to the mPOA, VMH, or medial amygdala, but not the NAc (Garcia-
Horsman, A
˚gmo, & Paredes, 2008). These data suggest that com-
mon opioid systems in the brains of male and female rats are acti-
tute a primary reward signal. Indeed, as mentioned above, repeated
ejaculations increase whole-brain endorphin content in male rats
(Szechtman et al., 1981). As with approach behaviors toward sex-
related odors, castration disrupts the expression of a copulatory
CPP on the first postoperative test in male rats (Hughes et al., 1990;
Miller & Baum, 1987), and acquisition of a copulatory CPP was
blocked by naloxone in castrated, but not gonadally-intact, male
rats (Meharra & Baum, 1990). Endocrine responses have also been
examined in male rats following exposure to contextual stimuli
associated with copulation. Kamel, Mock, Wright, & Frankel
(1975) reported that serum testosterone, luteinizing hormone,
and prolactin levels were elevated after 45 min of exposure to an
arena in which prior copulation occurred. Although dopamine
antagonists have not been reported to alter the development or
expression of copulatory CPPs in either male or female rats, or
postejaculatory CPPs in male rats (A
˚gmo & Berenfeld, 1990;
Garcia-Horsman & Paredes, 2004), Meisel, Joppa, and Rowe
(1996) found that the development of a copulatory CPP in female
hamsters was blocked by injections of the D2-receptor antago-
nists sulpiride or raclopride prior to each training session.
In summary, sexual reward in the form of ejaculation in male
rats, or paced copulation in female rats, involves the activation of
brain opioid systems. A similar effect may underlie the copulatory
reward in male and female Japanese quail. This is reminiscent of
the generalization made by some opiate addicts that the drug
effect reminds them of the pleasure experienced at orgasm (Pfaus
& Gorzalka, 1987), an effect referred to explicitly as a‘pharma-
cogenic orgasm’ by Chessick (1960). As with the reward state
induced by heroin or cocaine, both male and female rats spend
more time in contexts associated with state induced by sexual
reward. Moreover, conditioned contextual stimuli associated with
those reward states activate pituitary and gonadal hormone release
in male rats, suggesting a priming or ‘‘occasion-setting’’ response
that gets male rats ready to copulate.
Learning When to Do It
When to have sex seems straightforward in terms of arousal,
desire, and opportunity. If you have all three, you are all set. But
that is a big‘‘if’’! Sexual arousalis almost always conditional upon
the provoking stimuli being adequate to activate the necessary
sympathetic (e.g., increased heart rate) andparasympathetic (e.g.,
genital blood flow) divisions of the autonomic nervous system.
Desire often depends on whether the stimuli are competent as incen-
tives and what else might be going on at the time. And opportunity
is often fleeting.
Sexual arousal and desire are constrained further by experience.
In addition to habituation that occurs when one does the sa me th ing
over and over, or the inhibition that occurs during a refractory
period, the reward-related occasion setting mentioned above
may well underlie the ability of externalcues (and in their most
obvious manifestation,fetish objects) to controlsexual arousal
and thereby control when to have sex. Inanimate objects, such
as articles of clothing, particular forms and textures, and, in fact,
any antecedent tactile, olfactory, auditory or visual stimulation,
can during an animal’s initial experiences with sexual reward,
come to control both arousal and desire. Thus, the development of
paraphilias may well follow the same pattern of Pavlovian and
operant sexual learning discussed above.
Some opportunity is constrained by hormonal activation.
Obviously, having a sexual‘heat’ makes things simple, con-
straining sexual arousal, desire, and behavior to the periovu-
latory period when estrogens and progestins can stimulate it (Pfaff,
1980,1999). However, despite the fact that women’s desire peaks
around the time of ovulation (Stanislaw & Rice, 1988), we and
some of our primate cousins (e.g., rhesus macaques, Goy, 1978)
can have sex anytime throughout the cycle. And males can learn
when it is appropriate and not appropriate to attempt copulation
depending on the receptive versus defensive behaviors of the
female (Pfaus & Pinel, 1989). Male rats also display ‘‘psycho-
genic’’or noncontact erections in response to unconditioned sex
odors of sexually receptive females (Sachs, Akasofu, Citron,
Daniels, & Natoli, 1994), an effect that is facilitated mildly by
brief olfactory, visual, and auditory contact with inaccessible
receptive females, moderately by experience with mounting alone,
and dramatically by full copulatory experience to ejaculation
(Sachs & Liu, 1997). Thus, although erections can be induced
by presumably‘prepotent’’ unconditional olfactory stimuli, the
stimuli become far more effective—perhaps even sensitized—
by pairing them with the state induced by sexual reward.
What about other sensory cues? In his classic work Psychop-
athia Sexualis,Krafft-Ebing(1886/1929) presented numerous case
histories (almost all male) of fetish development in which partic-
ular inanimate objects, actions or states of arousal induced by
painful stimulation (e.g., spanking, piercing the skin, etc.) or
other tactile sensations (e.g., the feeling of feathers) became
associated with an individual’s first sexualerection or first set
Arch Sex Behav (2012) 41:31–62 43
of sexual experiences to orgasm. In some of those cases, the
initial experience occurred prior to the normative age of puberty,
suggesting that the individual may well have been in precocious
puberty or that at least his genital responses were in a state of
relative sexual awakening. Often, the individual would mastur-
bate to ejaculation while reminiscing about the early event and/or
fantasizing about experiencing it again with others. Of course, the
fetish became problematic when it was attempted with an unwill-
ing partner or when the individual was caught engaging the fetish
in public (e.g., hiding in bushes and spilling water from a pitcher
on an unsuspecting woman’s bustier). The case histories were
often accompanied by a psychological profile of the individual,
noting a history of odd behavior and any potential eugenic abnor-
malities (e.g., low level cognitive functioning) that ran in the
family. Given that such behavior was not‘‘normal’’in its day, and
that the individual appeared to suffer from some genetic basis for
abnormal behavior in general, it was easy to conclude that the
sexual behavior was a‘‘psychopathic’’ symptom of a moregeneral
genetic-based syndrome. And, after all, how could it serve repro-
duction if an individual can’t get an erection unless he or his part-
ner wears leather or lace? Surely, such objects did not play a central
role in our species’ evolutionary sexual history.
An alternate hypothesis from an incentive-based analysis is
that anything that predicts sexual reward becomes an occasion
setter for sexual arousal (and maybe desire too). But here is
where the situation gets dicey. Whereas sexual reward may be
studied objectively in animals, in humans the notion is muddled in
peculiar definitions, along with particular phases of the human sex-
ual response (Pfaus, 1999). It is also the case that human subjects in
laboratory settings show relatively weaker effects than are typically
observed in animals. This may be due to the fact that the typical
human subject has already acquired sexual experience prior to the
study, an effect that reduces the possibility of robust conditioning.
Thus, the sexual UCSs used may not create a‘‘strong enough’
sexual reward state to condition the particular CSs, especially
as they are not occurring during an early critical period of sexual
behavior development.
There have been several notable attempts to condition sexual
arousal in humans (see Hoffmann, 2012). Rachman (1966)and
Rachman and Hodgson (1968) found that following pairing with
erotic visual slides, a pair of women’s boots was able to elicit erec-
tions in men. Similarly, McConaghy (1970,1974)demonstrated
conditioned erection elicited by colored circles or squares paired
previously with erotic videos or still pictures in heterosexual and
homosexual men. A particularly informative study by Kantoro-
witz (1978) examined the nature of association between the UCS
and conditioned arousal induced by still pictures. For each sub-
ject, three different slides were paired with the plateau, abso-
lute refractory, and later resolution stages during and after mas-
turbation to orgasm. During subsequent testing, stimuli paired with
the plateau phase produced an increase in penile erection whereas
stimuli paired with the refractory phase produced a decrease in
erection. Stimuli paired with the later resolution phase had no
effect. Remarkably, those responses were still present after 3
months. Lalumie
`re and Quinsey (1998) reported significant con-
ditioned genital arousal in heterosexual men to a picture of a mod-
erately attractive, partially nude woman that was paired with a
video depicting highly arousing sexual interaction. A control
group that received access to the picture alone (without the
video) showed habituation.
Regarding women, Letourneau and O’Donohue (1997) failed
to find significant effects of conditioning on sexual arousal.
However, it was noted that the UCSs (erotic films) produced only
moderate levels of arousal whereas in studies with male subjects
such stimuli produced high levels of arousal. Thus, this failure to
demonstrate conditioned arousal in women may have been due to
an ineffective UCS. Indeed, Both, Laan et al. (2008) found that
neutral pictures of male headshots paired with 2 s of intensely
pleasurable vibrotactile CLS produced greater vaginal pulse ampli-
tude (VPA) during extinction in the paired versus unpaired groups.
In another study, stimuli were presented briefly (30 ms), followed
by a masking stimulus (Both, Spiering, et al., 2008). Although only
the paired group showed increased VPA to the CS during the
first extinction trial, there was no increase in the conscious
affective value of the stimulus. Finally, Hoffmann, Janssen,
and Turner (2004) varied the duration and relevance of a CS
(abdominal area vs. a gun) that was paired with short erotic film
clips in both men and women. Interestingly, when the stimuli
were presented subliminally for brief durations, the relevant
abdominal stimulus increased arousal in both men and women.
However, when the stimuli were presented for longer periods
prior to the erotic film clips, a sex difference emerged in which
the relevant CS alone (abdominal area) induced genital arousal
in men, but the presumably irrelevant stimulus alone (gun)
induced genital arousal in wo men. This latter effect may indicate
that women require CSs that increase autonomic arousal to a
higher extent than men, a potential corollary of the‘discordance’
experienced by women, but not men, between genital and
subjective sexual arousal (Chivers & Rosen, 2010).
Several studies have attemptedto demonstrate instrumental
control of sexual arousal in men and women. Rosen, Shapiro,
and Schwartz(1975) found that given feedback andcontingent
monetary reinforcement, men learned to become sexually aroused
in the absence of erotic stimuli. Other studies have found that men
can suppress (Rosen, 1973; Rosen & Kopel, 1977)orincrease
(Reynolds, 1980) penile erection with instructional feedback;
however, those studies failed to demonstrate learning effects
across trials. Given similar instructions, women can increase
their vaginal pulse in the absence of erotic stimulation (Zing-
heim&Sandman,1978) or decrease vaginal pulse in the presence
of erotic stimulation (Cerny, 1978), but, again, no learning effects
were reported. In summary, the evidence regarding instrumental
control of sexual arousal is limited to the one report in which
monetary reinforcement and feedback were provided.
Can animals show sexual arousal in response to cues asso-
ciated with sexual arousal or reward? Graham and Desjardins
44 Arch Sex Behav (2012) 41:31–62
(1980) placed sexually naı
¨ve male rats in a chamber scented
with a methyl salicylate (wintergreen oil) CS for 7 min before
transferring them to an adjacent chamber with an unscented
receptive female for 15 min of copulation (UCS). Control males
received CS alone, UCS alone or handling. Training occurred
once a day for 14 days. On the 15th day, males were placed in
either the CS or UCS chamber and trunk blood was taken within
15 min to assay for luteinizing hormone and testosterone. Expo-
sure to females increased concentrations of both hormones in
plasma whereas the CS was effective only in males in the paired
group. Sachs and Garinello (1978) reported that placing male
rats into a chamber in which copulation had previously occurred
dramatically reduced the latency for males to display penile erec-
tions. Male rats trained to associate an almond odor with copu-
lation to ejaculation (see below) initiated copulation faster with an
almond-scented receptive female than with an unscented recep-
tive female (Kippin & Pfaus, 2001b). Similarly, male marmosets
trained to copulate to ejaculation with lemon-scented receptive
females display erections and increased place preference in response
to the odor alone relative to unpaired males (Snowdon, Tannenbaum,
Schultz-Darken, Ziegler, & Ferris, 2011).
Placement of male rats into a holding cage has been used as a
conditioned contextual stimulus to signal non-copulatory expo-
sure to a receptive female behind a wire-mesh screen (Zamble,
Hadad, Mitchell, & Cutmore, 1985; Zamble, Mitchell, & Find-
lay, 1986). Placing males into the holding cage prior to a final test
of unrestricted copulation resulted in significantly shorter laten-
cies to intromit and ejaculate relative to males placed directly from
their home cages into the copulatory condition. Subsequent studies
found that second-order conditioned stimuli (e.g., a plastic toy fish
found in the holding cage) were effective in enhancing the same
measures of arousal (Zamble et al., 1985) and such conditioning in
sexually naive or sluggish males could increase the proportion of
males that copulated on a subsequent test (Cutmore & Zamble,
1988). Hollis, Cadieux, and Colbert (1989) demonstrated that
repeatedly pairing a light with non-contact exposure to a receptive
female resulted in conditioning of sexual behavior in male gou-
ramies, a type of Labyrinth fish. They found that males receiving
the conditioning treatment displayed significantly lower latencies
to initiate copulation and lower levels of aggression towards
females when the CS was presented before access to a female.
Similar results have been demonstrated in Japanese quail. Male
quail that received repeated exposure to females following the
presentation of a CS light displayed significantly shorter laten-
cies to initiate copulation when the stimulus was present com-
pared to when it was absent (e.g., Domjan, O’Vary, & Greene,
Finally, if male rats are given their first copulatory trials
wearing a Velcro tethering jacket, the jacket will become a CS?
for sexual arousal. Sexually naı
¨ve males in the paired group were
given nine multi-ejaculatory trials at 4-day intervals with thejacket
on, whereas males in the unpaired group received the same number
of trials without the jacket. On the 10th trial, males in both groups
were assigned randomly to two subgroups, one with the jacket on
and one with the jacket off, and given a final 30-min test of copu-
lation. Males that had been trained without the jacket displayed
normal baseline rates of copulatory behavior with or without the
jacket, as did males trained with the jacket that had the jacket on
during the final test (Fig. 6). In contrast, males trained with the
jacket that had the jacket off during the final test displayed a severe
copulatory deficit, including significantly increased mount, intro-
mission, and ejaculation latencies, a decreased number of condi-
tioned level changes, and a reduction in the proportion of males
that achieved ejaculation during the test. The pattern suggests sig-
nificantly diminished sexual arousal and desire and a failure to
engage sympathetic tone for ejaculation (Pfaus, 2009). Thus, the
jacket had become a CS?for sexual arousal only in the paired
group that had their early formative sexual experiences with it.
The ability of neutral olfactory and tactile stimuli associated
with sexual reward to direct sexual arousal in males of a variety of
species indicates a high degree of plasticity in the neural systems
that control sexual behavior in general, rather than a specific genetic
difference that makes some males susceptible to conditioning.
Could this plasticity extend to the development of preferences
for individual partner-related cues?
Learning Who to Do It With
Although it is clear that animals learn what to do sexually and
where and whento do it, the questionof who they prefer to do it
with seems like it should be less affected by experience and more
by innate, genetic, and species-specific causes geared toward repro-
ductive success. And in the case of homosexuality, it would seem
that innate causes play a large role in determining sexual orientation
long before sexual activity and sexual reward is initiated. Indeed,
sexual partner preferences that are not geared toward reproduction
(as in homosexuality) or that are highly constrained by the presence
or absence of inanimate objects (as in many paraphilias) should
either not exist or should confer some other reproductive benefit.
One such benefit could be related to kin selection (e.g., Wilson,
1975), as in the‘gay uncle hypothesis’that has received much
attention but little empirical support in Western culture (Bobrow
& Bailey, 2001), although it appears to be supported among the
Samoan fa’afafine (Vasey & VanderLaan, 2010). Yet, from a
reward-based incentive view, thereisabsolutelynoreasonwhy
partner-related cues that can be detected at a distance (such as sex,
gender, race, hair color or length, facial features, body type) or more
proximally (such as a multitude of personality characteristics)
cannot also become conditioned by early experience with sexual
reward, although there are important developmental constraints
that make the preference for gender far less amenable to condi-
tioning (see below).
Evidence of this in animals would provide strong support for
the alternative hypothesis put forth here that the‘‘genetic’ basis
of sexual preference rests in reward-based learning, as do
Arch Sex Behav (2012) 41:31–62 45
virtually all other motivational and regulatory functions per-
formed by organisms on this planet. This does not make it a
‘choice’’ except at an epiphenomenological level of interpreta-
tion. Once conditioned, sexual preferences should feel as natu-
ral and‘‘hardwired’ as one’s food preferences. And, in humans,
the conditioning does not occur overnight, but rather through a pro-
gressive and iterative process that begins well before the peri-
adolescent awakening of sexual arousal, and becomes relatively
renedandfixedduringformativeexperiences with masturbation
and orgasm prior to actual sexual activity with other people, when
further refinement and fixing occur. Interpersonal sexual reward
would thus reinforce the already-established template.
Features Conditioned by Sexual Reward
To study the effects of Pavlovian conditioning on sexual prefer-
ences, we took as our lead the findings of Graham and Desjardins
(1980) above. Could a male rat’s primary sexual experienceswith
scented females come to elicit a preference for scented partners in
Kippin, Talianakis, Schattmann, Bartholomew, and Pfaus
(1998) trained male rats to associate an almond odor smeared
on the back of the neck and anogenital area of females that were
either sexually receptive or non-receptive. Males in the Paired
group received sequential access to scented receptive and unscented
non-receptive females at 4-day intervals for nine trials each. Males in
the Unpaired group received the opposite order of association,
such that the scent was on the non-receptive females and no scent
was on the receptive females. Another group of males received
Random pairing of almond odor with the receptive and non-
receptive females. On the final test, males were placed into a
large openfield for a 5-min acclimation period,after which two
receptive females were placed equidistant from the male, one
scented and the other unscented. The choice of female for first
mount, intromission, and ejaculation and the distribution of
mounts, intromissions and ejaculations throughout the 30-min
open field test were recorded.
Males in all groups copulated with the females closest to
them, displaying no significant preference of female to receive
their mounts and intromissions. However, males in the Paired
group ejaculated preferentially with the scented females, choos-
ing them significantly more than the unscented females to receive
their first ejaculations, and more often to receive their ejaculations
throughout the test. Males in the Unpaired had the opposite prefer-
ence, choosing to ejaculate first, and more times overall, with the
unscented females. Males in the Random group did not snow any
ejaculatory preference. Interestingly, males in the Unpaired group
displayed a preference to mount and intromit with the unscented
females on the final test, suggesting that the almond odor had
acquired conditioned inhibitory properties that dampened their
interest in now-receptive females bearing the odor. In fact, those
males showed robust avoidance behavior toward the scented
Conditioning Group Conditioning Group
Control Jacket Control Jacket
Control Jacket Control Jacket
Conditioning Group Conditioning Group
Level changes Intromission latency
Number of intromissions Number of ejaculations
Jacket off
Jacket on
Fig. 6 Appetitive and consummatory sexual behaviors by male rats given their first nine trials of sexual behavior with or without a rodent jacket. On
the final test, rats in each group either had the jacket on or off. *p\.05 between jacket on and jacket off conditions
46 Arch Sex Behav (2012) 41:31–62
females during the final test (an effect that was responded to with
increased solicitations on the part of those females).
Subsequent studies established that the conditioned ejacula-
tory preference (CEP) developed only if males were in the vicin-
ity of the females during their post-ejaculatory refractory period
(Kippin & Pfaus, 2001a) and was more robust if males had more
ejaculations (Kippin, Samaha, Sotiropoulos, & Pfaus, 2001). It
did not develop if males were allowedonlytointromitorifthe
female was removed immediately after ejaculation. An analysis
of the conditioned response showed that males became‘‘choosy’
during the last three mounts prior to ejaculation and directed those
mounts at the female bearing the familiar cue (Kippin & Pfaus,
2001b). This is a period in which males are sensitive to their impend-
ing ejaculation and emit a 50 kHz call that attracts females imme-
diately prior to ejaculation (McGinnis & Vakulenko, 2003). Using
unilevel pacing chambers bisected by a Plexiglas partition that con-
tained either 1 or 4 holes, males were shown to develop significant
CEP only when trained in the 1-hole condition, in which they had to
wait longer between intromissions, had more intromissions before
ejaculation, and longer ejaculation latencies compared to males in
the 4-hole condition (Ismail, Gelez, Lachapelle, & Pfaus, 2009),
suggesting that the ejaculatory reward state may rest on increased
parasympathetic arousal during copulation, something that occurs
naturally in bilevel pacing chambers as the males must chase the
females in order to copulate with them (Pfaus, Mendelson et al.,
1990). Finally, Erskine (2005) raised the issue of whether such
partner preferences might be displayed in more natural contexts
(e.g., between different strains of partner). So, instead of neutral
olfactory cues on the same strain of partner, we varied the strain
of partner (pigmented Long-Evans rats vs. albino Wistar rats) asso-
ciated with the ejaculatory reward state (Ismail, Jones, Graham,
Sylvester, & Pfaus, 2011). Pigmented Long-Evans males received
nine multi-ejaculatory copulatory training trials with either pig-
mented or albino females in a 1-hole pacing chamber. In a final
open field test with two females, one pigmented and one albino,
the males chose the strain of female associated with the 1-hole con-
dition significantly for their first ejaculation. However, the males
chose pigmented females significantly for their second ejacula-
tion only if they were associated with the 1-hole condition. This
suggests that a degree of ‘‘assortative’’ choice (in which animals
have a natural preference for less related members of their own strain,
or at least that look a little like them) may strengthen the develop-
ment of CEP (for a discussion of optimal in- and out-breeding as
it relates to assortative mating, see Bateson, 1983).
Association with reward can also reduce the impact of aver-
sive stimuli. For example, although rats initially reject a bitter
morphine solution that replaces their drinking water, they learn to
like the bitter taste as the morphine reward state sensitizes over
time (Zellner, Berridge, Grill, & Ternes, 1985). The sexual reward
state in males is clearly powerful enough to condition a preference
for a neutral odor, so we asked whether it could also blunt the
impact of an aversive odor like cadaverine, a polyamine that is
produced in decaying flesh. Rats are well known to bury aversive
smelling objects such as noxious food, dead conspecifics, and anes-
thetized conspecifics speared with cadaverine (Wilkie, MacLen-
nan, & Pinel, 1979), so cadaverine is considered unconditionally
We gave male rats their first nine sexual experiences with
either cadaverine-scented females (unanesthetized, of course),
or unscented control females. A third group of males was given
habituation to the cadaverine odor in their home cages and copu-
lation trials with unscented females (Pfaus et al., 2001). On the
10th trial, males from the three groups were placed into an open
field and allowed to copulate freely with two receptive females,
one scented with cadaverine and one unscented. Males in the control
group pursued the unscented females selectively and ejaculated
exclusively with them. Males thathadbeenhabituatedintheir
home cages did the same. In contrast, males that had copulated pre-
viously with cadaverine-scented females pursued, copulated, and
ejaculated with both females, indicating that the aversive properties
of cadaverine had been diminished after pairing with sexual reward.
Four days later, males received a wooden dowel that had been sat-
urated in cadaverine into their home cages. Males in the control
and habituated groups avoided the dowel, and some attempted to
bury it. In contrast, all males in the cadaverine-paired group made
contact with the dowel and more than half picked it up and gnawed
on it, as they would if the dowel had been laced with something
hedonically positive like estrous vaginal secretions or chocolate.
First and foremost, these findings demonstrated that a simple
Pavlovian conditioning paradigm can modify male rat sexual strat-
egies away from an assumed innate polygamous preference for
novel females toward a preference for familiar females. Second,
they showed that male rats can distinguish copulation from mat-
ing, as the conditioned preference was not displayed in the Paired
group for choice of female for mounts and intromissions but rather
for ejaculations. Third, the association of the odor with sexually
non-receptive females induced a conditioned inhibitory state
that compelled males to avoid sexually receptive females bearing
the odor on the final test. Fourth, the state induced by ejaculation
was the critical UCS, especially if it occurred on a background of
high arousal. Finally, an unconditionally aversive odor (cadaver-
ine) was made less aversive and possibly conditionally appetitive
by pairing with sexual reward.
Female rats also show olfactory conditioned partner pref-
erence formales associated witha pacing-induced rewardstate
(Coria-Avila, Ouimet, Pacheco, Manzo, & Pfaus, 2005). This
was accomplis hed easily in unileve l pacing chambers in whi ch
the paced condition involved the placement of either a 1-hole
or 4-hole Plexiglas divider through which the female could reg-
ulate the initiation and rate of copulation. The non-paced condition
involved copulation in the same chamber but without the divider.
Females in the Paired group were given paced copulation with
males that have had the almond odor smeared on their necks and
anogenital area versus non-paced copulation with males that
have had distilled water smeared on the same areas. After four
paced versus non-paced trials, females were placed into a large
Arch Sex Behav (2012) 41:31–62 47
open field with two tethered males, one scented and the other
unscented,and choice ofmale to solicit,copulate with, andreceive
ejaculations from, was recorded. Females for which the odor was
paired with the paced condition selectively solicited, copulated
with, and received ejaculations from the scented male. Females
that had the odor explicitly unpaired or paired randomly with
pacing did not display a preference.
As with males, females showed a similar preference for strain
cues associated with paced copulation (Coria-Avila et al., 2006),
although it was stronger if the strain associated with paced cop-
ulation was their own. Interestingly, in that study, pigmented or
albino females solicited whichever strain of male was associated
with paced copulation, but received ejaculations preferentially
from males of their own strain and only if that male had been
associated with paced copulation. This also revealed a degree of
assortative choice, especially for mating, and showed that females,
like males, can differentiate copulation from mating. Finally, female
rats that experienced manual CLS in the presence of a cotton gauze
pad soaked in almond extract chose to copulate selectively with
almond-scented males over unscented males during their first
sexual experience in a large open field with both males (Parada
et al., 2011). Interestingly, they did not show a preference to
receive the scented male’s ejaculations, suggesting that the VCS
received from males during paced copulation induced a further
reproductive or mate choice. It is not yet known whether this stems
from specific stimulation of the cervix (and pelvic nerve) or from full
stimulation of internal and external aspects of the clitoris. Experience
with paced, relative to non-paced, copulation in unilevel cham-
bers induces significant neurogenesis in the granular layer of the
accessory olfactory bulbs (Corona, Larriva-Sahd, & Paredes, 2011),
a region known to contain intrinsic memory systems related to
pheromonal stimulation and recognition of conspecifics (Brennan,
Kaba, & Keverne, 1990; Brennan & Keverne, 1997). Thus, as
Darwin (1859) suggested, females are indeed more‘‘choosy’’than
males in terms of choosing a male to copulate and mate with. How-
ever, like males, they base this choice on the external (epigamic)
characteristics of the male associated with their own experience of
a copulatory reward state.
Femalesalso learn inhibitoryassociations. In addition to the
effect of nonreward induced by training under the influence of
the opioid receptor antagonist naloxone, sexual frustration in
the presence of an inaccessible male can also induce an inhibitory
state. Parada et al. (2011) gave sexually naı
¨ve female rats five trials
of CLS in the presence of a sexually active male scented with
almond behind a screen. On alternating days, the females received
sham CLS in the presence of an unscented male behind the screen.
During the final open field test with two males, one scented and the
other unscented, females solicited selectively the unscented male
and showed a trend to receive that male’s ejaculations preferen-
tially. At first glance, these data seem at odds with the fact that CLS
induces a reward state. However, it was noted that females attempted
to solicit the males behind the screen following CLS during the
training trials, which, of course, were not successful because the
male was behind a screen. Thus, it is likely that the female was in
a state of frustrated non-reward that she associated with the odor
and generalized to the choice of male for her first sexual experience.
Neural Correlates
One of the strengths of the olfactory conditioned partner preference
paradigm is that the odor can be presented alone in the animal’s bed-
ding or on a cotton gauze pad and neuroanatomical and neuro-
chemical activation in response to the odor examined. This takes
the mess of copulation (and interpretation based on different types
or amounts of sensory stimulation) out of the analysis and makes
the study far more cognitive in terms of examining responses made
to a second-order conditioned cue. In fact, it is now possible using
this paradigm to relate areas of the rat brain activated by the con-
ditioned odorwith areas of the human brain activated by visual
erotic cues.
We have found significant activation of the immediate-early
gene product Fos following presentation of the odor in paired
versus unpaired groups in a number of brain regions that con-
stitute a sexual circuit similar to that found activated in fMRI
studies of sexual arousal and desire in men and women (e.g.,
Karama et al., 2002; Ortigue & Bianchi-Demicheli, 2008; Paul
et al., 2008). These regions included cortical structures such as
the medial prefrontal cortex, piriform cortex, insula, and anterior
cingulate gyrus; limbic/motor regions such as the olfactory tuber-
cle, nucleus accumbens, dorsal striatum, lateral septum, basolateral
and medial amygdala; hypothalamic regions such as the mPOA,
bed nuclei of the stria terminalis, anterior and lateral hypothalamus,
and the midbrain anterior ventral tegmentum (a region that gives
rise to the mesolimbic dopamine (DA) system) of males (Kippin,
Cain, & Pfaus, 2003)andfemales(Coria-Avila&Pfaus,2007)
after presentation of the paired odor alone relative to the odor
unpaired or random. Previous studies showed these regions to
be activated by unconditioned olfactory or genitosensory stimulation
during copulation (reviewed in Pfaus & Heeb, 1997); thus, their
conditional activation by the odor suggests that they are activating
a representation of the UCS, as Pavlov (1927) originally suggested
for all CS?s. Notably, a nearly identical set of regions is activated
by stimuli that predict cocaine or heroin reward (Childress et al.,
2008; Garavan et al., 2000), suggesting that these regions func-
tion together as a general system for appetitive arousal and desire
related to reward.
We suggest that the conditioned cue acts as a‘‘priming stimulus’
to activate cortical, limbic, and hypothalamic circuits involved in
the facilitation of sexual arousal and desire and/or in the suppression
of inhibitory systems (Fig. 7). Some of those circuits involve selec-
tive processing of the sexual stimulation that generates the reward
state (UCS), the olfactory stimulus (CS), and a system that inte-
grates the CS and UCS so thatanimals can focus their attention
and engage appropriate forward-directed locomotion toward
the CS whenit is present (Robinson& Berridge, 1993).What is
immediately obvious fromthe figure is thatsystems for sexual
48 Arch Sex Behav (2012) 41:31–62
reward and incentive responding overlap with systems pro-
posed forsexual (and maternal)bonding (e.g., Young& Wang,
2004). These involve the interaction of at least three neuro-
chemical systems, including DA, hypothalamic oxytocin (OT),
and opioids that inhibit hypothalamic structures like the mPOA,
but sensitize mesolimbic DA systems through a process of dis-
inhibition (Kalivas & Stewart, 1991). This is strikingly similar
Fisher’s (2000) proposal of three primary emotional systems for
mating, reproduction, and parenting.
As reviewed above, inhibition of opioid receptors with nal-
oxone inhibits the development of sexual CPPs in both male and
female rats and produces an expectancy of non-reward that inhib-
its sexual desire and behavior in female rats that have their first
sexual experiences with the drug. Naloxone administration dur-
ing training also inhibits the development of sexually-conditioned
ejaculatory and partner preferences in male and female rats (Coria-
Avila et al., 2008;Ismail, Girard-Be
´riault, Nakanishi, & Pfaus,
2009). In these studies, naloxone or saline was administered imme-
diately before a male rat’s first 10 multiejaculatory sexual experiences
with a scented partner or during a female’s first five experiences
of paced copulation with a scented partner (relative to five trials
of non-paced copulation with unscented males that were inter-
spersed). Rats were then given a final test in a large open field with
two sexually receptive and vigorous partners, one scented and
the other unscented, following saline injections. Rats trained with
saline injections showed the partner preference reviewed above
whereas rats trained with naloxone did not display the pref-
erence for the scented partner. In fact, maleschose the unscented
female significantly more frequently for their first ejaculation
and females solicited the unscented male significantly more
frequently and received more mounts, intromissions, and ejacu-
lations from that male.
We have also found conditioned DA release in the NAc of
Paired-trained male rats in response to the almond odor, but not in
Unpaired-trained males (Fig. 8) and conditioned activation of Fos
within parvocellular OT neurons of the paraventricular nucleus
(PVN) that likely project to other forebrain regions. Fos was also
found activated by the conditioned odor in magnocellular
CS detector
Olfactory cue
UCS detector
Sexual reward
CS-UCS Integrator
Pir Ctx
DA +++
MSH ++
Appetitive responses
Sexual rewar
(and arousal)
DA +++
DA +++
Fig. 7 Neural systems which are critical for the display of conditioned
olfactory preferences in the rat. Appetitive behaviors made toward
conditioned stimuli lead to sexual reward that is processed by three
interactive systems. Two systems process olfactory stimuli and sexual
reward relatively independently, whereas a third, mesolimbic DA
system, acts to integrate both the conditioned olfactory cue and its
rewarding sexual outcome. Three common regions, the piriform cortex,
mPOA, and VTA, are activated in male and female rats by conditioned
olfactory stimuli. Opioid actions in the VTA potentiate mesolimbic DA
activation, whereas opioid actions in the mPOA inhibit sexual arousal
and desire. Neurotransmitter systems or their receptors in red/lighter are
excitatory for sexual motivation whereas those in blue/darker are inhi-
bitory. Note that opioids can be excitatory in the VTA, inhibitory in the
mPOA, or either in the VMH (depending on the receptor type). Dopa-
mine, gonodotropin releasing hormone, melanocyte stimulating hor-
mone, noradrenaline, and oxytocin are excitatory whereas serotonin,
opioids, and the endocannabinoidsare inhibitory. ACC anterior cingulate
cortex; ArcN arcuate nucleus of the hypothalamus; CB1 cannabinoid
type 1 receptor; CPu caudate-putamen (striatum); DA dopamine; ddelta
opioid receptors; GnRH gonadotropin releasing hormone; LS lateral
septum; MeApd posterior-dorsal nucleus of the medial amygdala; mPOA
medial preoptic area; MSH melanocyte stimulating hormone; lmu opi-
oid receptors;NAcc nucleus accumbens; NE noradrenaline;OT oxytocin;
PirCtx piriform cortex; PVN paraventricular nucleus of the hypothala-
mus; Tu olfactory tubercle; VMH ventromedial nucleus of the hypothal-
amus; VP ventral pallidum; VTA ventral tegmental area; 5-HT seroto-
nin. Adapted from Pfaus, Ismail, and Coria-Avila (2010) (Color figure
Arch Sex Behav (2012) 41:31–62 49
vasopressin (VP) neurons of the supraoptic nucleus that likely
project into the posterior pituitary. Systemic injections of OT
(5 lg/kg), but not vasopressin, to sexually naı
¨ve male rats during
their first exposure to almond-scented receptive females resul-
These data suggest that conditioning in the putatively polyg-
amous rat sensitizes a circuit similar to that activated in monog-
amous Prairie voles during their formative sexual experiences
(Aragona et al., 2006; Young, Murphy Young, & Hammock,
2005; Young & Wang, 2004) and following parturition (Wang,
Liu, Young, & Insel, 2000). The results of Aragona et al. (2006)
are particularly instructive, as the sexual bond formation was
inhibited by activation of D1 receptors, but facilitated by the
activation of D2 receptors. This suggests a neural reorganization
in mesolimbic terminals after formative sexual experiences that
‘seals the bond,’’ making it less likely for other stimuli to acquire
associative strength.
Such an effect is consistent with modern theories of learning
(e.g., Harris, 2011; Rescorla & Wagner, 1972)andhasbeen
implicated in the susceptibility to drug addiction, especially in
terms of responding to cues that predict drug reward (Flagel et al.,
2010), and more generally in response to food-related cues
(Berridge, 2009; Blackburn, Pfaus, & Phillips, 1992). The inter-
action of OT and DA in the PVN, mPOA, VTA, and NAc of male
rats induces penile erections and links them to appropriate appe-
titive sexual behaviors (Succu et al., 2007). Thus, opioid reward
states may form the rudimentary mechanism of bonding because
they sensitize DA release in the presence of reward-related
cues compelling animals to focus their attention and goal-directed
behavior toward those cues. Activation of brain OT systems (by
DA or other means) adds a reduced social distance and bonding to
this neurochemical reward state. Given that pharmacological acti-
vation of opioid receptors induces a direct suppression of both hypo-
thalamic and pituitary OT secretion (Vuong, Van Uum, O’Dell,
Fig. 8 Dopamine release in the nucleus accumbens of male rats in response
to an almond odor paired or unpaired with sexual reward, estrous vaginal secre-
tions, or a lemon odor unpaired in both groups. Unilateral dialysates were taken
every 15 min and analyzed via high performance liquid chromatography.
Almond and lemon odors were presented on gauze. Estrous vaginal secretions
were presented on Q-tips after vaginal lavage
50 Arch Sex Behav (2012) 41:31–62
Lufty, & Friedman, 2010), sensitized and potentially reorganized
mesolimbic and hypothalamic DA systems must be a necessary
intermediary. This is consistent with a multifaceted role of mes-
olimbic DA in incentive salience and response initiation (e.g.,
Blackburn et al., 1992; Robinson & Berridge, 1993).
In social systems, assortativeness refers to a bias in favor of sim-
ilar characteristics. This bias could be toward any outwardly
definable (or distal) trait, including gender, race, size, age, dress,
and more proximal traits, including socioeconomic status, polit-
ical belief or religious orientation. In mating systems, assortative-
ness refers to choosing characteristics that are similar to oneself,
as in species or strain (race/size)whereas disassortativeness refers
to choosing characteristics that are dissimilar, such as the sex of
the partner in heterosexual circumstances. Such choice is believed
to reflect sexual imprinting, in which early postnatal (PN) condi-
tions require the newborn to recognize and stick close to its mother
or caregiver. Bateson (1983) defined heterosexual assortative mate
choice for different species along an optimality curve between
inbreeding and outbreeding, with a number of external and genetic
weighting coefficients that define the costs and benefits of both.
The choices are then made on the basis of dominant distal sensory
cue(s), such as coloration in the Japanese quail studied by Bateson.
But all is not quite as it should be. Many different external
cues can be imprinted duringthe early PN period depending on
who (or what) happens to be in the right place at the right time
(as in the famous picture of Konrad Lorenz with Greylag gos-
lings following him). Several important studies have demon-
strated that adult males mate preferentially with females that
have attributes similar to those of the female(s) that nursed them
early in life. Yamazaki et al. (1988) found that male mice nursed by
foster mothers chose to mate with females that resembled their
foster mother rather than females that resembled their biological
mother. Similarly, adult male Lesser Snow Geese prefer the col-
oration of the female that nursed them over a different coloration
(Cooke & McNally, 1975) and field studies (Cooke, Finney, &
Rockwell, 1976; Cooke, Mirsky, & Seiger, 1972). One of the most
provocative reports of sexual imprinting was that of Kendrick,
Hinton, Atkins, Haupt, and Skinner (1998), who demonstrated that
sexual partner preferences can be achieved between goats and
sheep using cross-fostering to manipulate the imprinting pro-
cess. In males and females of both species, sexual partner pref-
erences were toward members of the opposite sex of the species
of the foster, rather than biological, mother.
Is assortative mate choice dependent on mom per se or on a
state of reward that her features are associated with? We have
found that an olfactory cue associated with early PN reward in rats
directs mate choice in adulthood (Me
´nard et al., in preparation).
Newborn rats separated from their mothers display distress vocal-
izations that are reduced or eliminated by stroking their heads
and backs with a soft fine paintbrush. Thus, stroking is both
negatively reinforcing (in the reduction of acute distress) and
positively reinforcing as a reward (Sullivan & Hall, 1988). In the
´nard et al. study, we separated rat pups from their mothers for
15 min a day on PN days 1–10 and placed them on bedding
scentedwithlemon (Paired) orleft unscented(Unpaired). Pups
were then stroked gently for the 15min of maternal separation,
after which they were returned to their mothers. On PN 60, after
adolescence had been reached, the males were given their first
sexual experience in an open field with two sexually receptive
females, one scented with lemon and one unscented. All males in
the Paired group ejaculated preferentially with thelemon-scented
females whereas males in the Unpaired group did not display a
preference.Imprinting alters neurogenesis and synapse formation
in a variety of limbic and cortical brain structures associated
with se nsory coding, reward, emotion, and memory, especially the
hippocampus and related cortical structures (Bolhuis, 1999;Mori-
ceau & Sullivan, 2004). It is likely, therefore, that those changes
essentially sculpt the bias in brain.
Experience and Expectancy
Experience leads to expectancy. Reward to desire. In a Pavlovian
scheme, the CS comes to elicit a representation of the UCS and
conditioned responses that lead to reinforcement strengthen the
association. It is circular, but that is a fundamental property of the
experience-driven brain plasticity that underlies Hebbian syn-
apses. To the extent that‘first cuts are the deepest,’’ and that first
experiences with arousal and reward form a critical period of
behavioral development and crystallization, the onset of sexuality
then becomes a fluid critical period during which distal cues,
proximal cues, and interpersonal cues that signal sexual reward
are assembled into‘‘wholes’’ or sexualGestalts made up of desired
features in potential sex partners (assembled perhaps hand-in-hand
with desired activities with those partners). Individual differences
in arousability, attention, temperament, reward intensity, and inhibi-
tion would then determine the success and timing of early sexual
experiences. And if the process is iterative and additive, which it
must be by definition, then new and unexpected experiences of
greater sexual arousal and reward could refine the features further
(e.g., as suggested by Morin, 1995) and lead to a conscious aware-
ness that one has bonded sexually, emotionally, and romantically
with another individual. Conversely, experiences with sexual non-
reward could loosen or extinguish these associations, leading the
features that were once in the‘foreground’’into the‘background’
(to paraphrase Ko
¨hler, 1929). However, if the process of reward
learning is iterative and additive, it may never be completely pos-
sible to extinguish salient features entirely. Former drug addicts
always want or crave some aspect of the early drug effect, even if
the rewarding aspects of the drug had diminished prior to absti-
nence (Robinson & Berridge, 1993). This can be stimulated by
cues that predict drug reward, and also by stress or other unexpected
factors, and can lead to spontaneous recovery of drug taking
Arch Sex Behav (2012) 41:31–62 51
(Shaham, Erb, & Stewart, 2000), recovery of fears and phobias
(Jacobs& Nadel, 1985),and perhaps alsoto the sexualgestalts and
activities that we initially preferred.
Experience plays a vital role in some theories of sexual ori-
entation. In his‘exotic becomes erotic’’theory, Bem (1996,2000)
provides support for the argument that gender-atypical behavior in
childhood roots individuals into a biological framework in which
same-sex activities become stronger predictors of sexual reward.
This occurs in a social context in which gender role is fixed into
‘male’’ and‘‘female’’modes and imposed, and always confounded
with sexual behavior. A number of studies have shown a strong cor-
relation between gender nonconformity in childhood and sub-
sequent homosexual orientation (e.g., Bailey & Zucker, 1995)
and Bem’s own path analysis from a large sample of twins (Bem,
2000) reinforces this notion. So a first set of pieces in the puzzle of
sexual orientation (and preference) comes from early interactions
of genetic and sociocultural factors.
A second set of pieces comes from sexual experience itself.
Storms (1981) argued that sex drive development occurs in concert
with erotic development during a peri-adolescent period when
individuals are moving from predominantly homosocial to
heterosocial peer groupings. The content of an individual’s early
sexual fantasies and desires is then consciously assessed in the
context of actual sexual behavior and the dominant sociocultural
definition of the appropriate gender and sexual orientation for that
behavior, so that a convenient (and sometimes devastating) cogni-
tive‘‘label’’ends up incorporated in one’s self-definition. And given
the magnitude of sexual awakening in most people’s adolescent
experiences, this addition to the self-definition is usually large and
profound. One prediction in Storms’ theory is that sexual experi-
ences that occur in the early phase of adolescence may be more
likely to involve salient individuals in one’s same-sex social group.
Interestingly, the modal retrospective age of puberty onset reported
by the sample of gay men in Hamer et al. (1993) was between 11
and 12 years, somewhat earlier than the ‘‘normative’’12–13 years
typically reported for boys in Western countries (Slyper, 2006).
Of course, puberty does not occur suddenly but rather progresses
in a cascade of events related to brain maturation, pituitary pep-
tide hormone release, gonadal steroid hormone secretion, growth,
and metabolism. Nevertheless, this leads us to entertain a tanta-
lizing possibility that the X-linked genetic locus discovered by
Hamer et al. could be related to a‘‘female’’-like timing of puberty
onset. The Hamer et al. study can be mined for other gems of sup-
port for some of Bem’s and Storms’ ideas. The retrospective age
of onset of first same-sex attraction had three phases, one from
ages 5 to 8, a second from ages 9 to 13, and a third from 13 years
on. The retrospective age of self-acknowledgement of homo-
sexuality began just before the modal peak onset of puberty (approx-
imately between 10 and 11years), and rosesteadily to age 17,
then leveled off. Trailing this was the retrospective age of
acknowledgement of homosexuality to others, which had a
dynamic rise from ages 16 to 22,then leveled off in twophases,
one from ages 22 to 32, and then less steeply after age 32. One
might imagine that exotic was becoming erotic through iter-
ations of sexual reward achieved from formative same-sex sex-
ual activities (starting, for example, with arousal, then rehearsal
during masturbation, then actual sex play to genitally-based inter-
course to orgasm) roughly between the ages of 11 and 16. It is not
known from the data reported whether the men in that sample had
experienced gender atypical behavior in that first period of
attraction to other boys, but it is likely given the preponderance of
supportive data.
Critical Periods and Bias: Disentangling Orientation
from Preference
The data of Hamer et al. (1993) suggest the existence of at least
three critical periods during the trajectory toward full sexual
responsiveness: attraction, then arousal, and finally the emergence
of a sexual reward system linked predominantly to genital pleasure
and orgasm during and after puberty. The sexual nature of this last
system is explicit: by the time people have incorporated sexual
activity into their daily routines, they have already imbued it
with names and given it personal and social definitions and values
that are specific to it. Although the animal work reviewed above
indicates a high degree of plasticity regarding the development of
preferences for certain types of stimuli associated with sexual arousal
and reward during an animal’s first experiences with genital stimu-
lation and copulation, it does not reveal how sexual orientation
toward members of the opposite or same sex develops. The devel-
opment of sexual orientation then must form a ‘constraint’’ on
learning (Gallistel, 1980;Lashley,1951), biasing the subsequent
development of sexual preferences toward the preferred gender.
In fact, we argue that each critical period biases the formation of
the next, although the next critical period follows the rules of an
emergent property as far as the brain is concerned and cannot be
reduced solely to the organization and activation that occurred in
the previous period. That these critical periods emerge on a self-
defining continuum is consistent with the trajectory outlined in
Bancroft’s (2009) seminal review of sexual development.
To begin to understand how an orientation bias might develop,
we much consider the role of two earlier critical periods, one of
perinatal hormone-induced brain differentiation (McCarthy &
Arnold, 2011) followed by another within the first 2–3 years of
life linked to gender differentiation and the display of gender
typical versus atypical behaviors. A cascade of developmental
epochs or critical periods is depicted in Fig. 9(and is contrasted
with a neo-Piagetian cascade of stages of cognitive develop-
ment that occurs during the same time frame, e.g., Demetriou,
1998;Piaget,1967). These first two critical periods form accord-
ingto a relativebinary code linkedto‘‘male’’and‘‘female’’(Money
&Ehrhardt,1972). Thus, we refer to sexual differentiation ofanat-
omy and brain as the first critical period out of which should emerge
an anatomical‘boy’’or ‘‘girl’that we assume has a brain differen-
tiated accordingly. The second critical period differentiates gender-
52 Arch Sex Behav (2012) 41:31–62