The Psychopharmacology of Sexual Behavior

James G. Pfaus and Barry J. Everitt

The history of research into the effects of drugs on human sexual behavior has been concerned primarily with the search for an aphrodisiac. More recently, psychoactive drugs used in the treatment of psychiatric and neurological disorders have been seen to have (usually unwanted) side effects that include impaired libido in men and women or, occasionally, increases in sexual interest that have been much publicized in the press (e.g., the effects of L-dopa in Parkinsonian patients in the 1960s). Although modern medical and scientific opinion has tended to dismiss the idea of chemical stimulation of sexuality as improbable, at least as far as human sexuality is concerned, there has in recent years been a resurgence of interest in the development of such drugs within the pharmaceutical industry. This has in part resulted from the medicalization of male sexual dysfunction, which has emerged over the past 10 years with the widespread involvement of urology and vascular surgery in this clinical field, and with the discovery that injection of various smooth-muscle-relaxing drugs into the corpora cavernosa of the penis is effective in producing erection. The development of a drug that could be taken by mouth that would enhance sexual responses without other undesirable side effects would have major clinical as well as commercial consequences.

In men with sexual dysfunction, dopamine agonists and a2-adrenoceptor antagonists are currently being explored as therapeutic agents, a development based largely on the effects of such drugs on animal, in particular rodent, sexual behavior. This brings into focus the question concerning the relevance of such observations to human sexuality, since the behavioral parameters measured in rats, such as mounting and intromission latencies, have no obvious counterpart in men.

Recently, there have been new conceptual developments in the analysis of rodent sexual behavior and its neurochemical determinants (19, 52, 57), which, as a result of their relative sophistication, may more effectively address issues of importance in understanding the neural and neurochemical basis of human sexuality. In particular, these methods allow distinctions to be made between the psychological processes underlying appetitive and consummatory elements of sexual behavior and thereby the more effective study of their neural basis. To date, such methods have been developed mainly for male sexual behavior, although similar methods for females are emerging. They include (a) instrumental behavior of male rats maintained by sexual reinforcement (the opportunity to copulate with a female in heat) presented under a second-order schedule of reinforcement (19, 20); (b) conditioned level changing in a bilevel chamber (44); (c) conditioned place preference acquired by allowing sexual interaction in a distinctive environment (19, 52); (d) sexual partner preference, that is, the preference of a male for an estrous versus an anestrous female, or a female for an active or inactive male (see ref. 19). These procedures allow appetitive, or incentive motivational, components of sexual behavior to be studied independently of the ability to copulate and thereby represent a useful development in dissecting the neurochemical basis of sexual motivation and performance (19) (see also Behavioral Techniques in Preclinical Neuropsychopharmacology Research and Animal Models of Psychiatric Disorders).

The comparable attempts to produce models of human sexuality have been fundamentally different. In particular, they have focused not on patterns of behavior, such as mounting, intromission, or lordosis, but on patterns of physiological response, in particular genital responses and other manifestations of arousal. This in part reflects the assumption that, for the human subject, lying between a physiological state of arousal and a behavioral response is a largely imponderable complex of cognitive processes, with a wide variety of psychosocial influences. More concretely, it can be said that, whereas there are very specific behavioral or motor patterns involved in animal, especially rodent, sexual behavior (such as mounting, intromission, or lordosis, some of which are clearly reflexive and hormonally determined), there are no comparably predictable motor patterns in the human repertoire. There is certainly no human (or primate) counterpart of lordosis. The nearest measure of a predictable mounting pattern is the semivoluntary tendency to pelvic thrusting which is a feature of human sexual arousal, but which in fact has been largely ignored in the human literature.

In addition, the measurement of physiological responses in a sexual situation, at least within a laboratory setting, is technically much more feasible in human subjects. In fact, the animal, and in particular the rodent literature has not to date considered genital responses or other physiological parameters of arousal, with the exception of the specific reflexive erections of the male rat, which are frequently elicited in a nonsexual context and which have very uncertain relevance to their normal sexual behavior (see ref. 68 for discussion). Bancroft (6) has suggested that a complex set of processes be considered in the context of human sexual response, involving cognition, some form of central arousability of a specifically sexual kind, peripheral physiological manifestations of this state, and a variety of behaviors (see ref. 21). For example, in the case of a mans' sexual response, in a cognitive domain, self-ratings of frequency of sexual thoughts and associated excitement have been obtained. Behaviorally, assessment of the frequency and quality of various sexual acts, such as sexual intercourse or masturbation, has proved useful, whereas two principle psychophysiological methods have been employed: the measurement of spontaneous erections during sleep, or nocturnal penile tumescence (NPT), and the measurement of erectile responses to various types of erotic stimuli, recorded in a laboratory setting in a more or less standardized fashion.

The number of chemical transmitters, including neuropeptides, that have been implicated in the neurobiology of sexual behavior has grown exponentially since the first demonstration of cholinergic and monoaminergic effects on masculine and feminine sexual behavior in rats and humans. In this chapter, we consider some recent data on the monoaminergic neurotransmitters dopamine (DA), norepinephrine (NE), and serotonin (5-hydroxytryptamine or 5-HT), as well as some neuropeptides and other agents reported to affect sexual behavior, referring where appropriate to relevant studies in humans. The review is not intended to be exhaustive, but to illustrate advances based on the development of behavioral and neural methodologies.

Apomorphine, a mixed D1 and D2 dopamine receptor agonist, increases the likelihood of spontaneous erections in normal human volunteers and men with psychogenic erectile dysfunction (32, 67). The D2 side effects of such agonists, such as nausea or vestibular disturbance, make such positive drug effects difficult to elicit. Thus, whereas agonists, such as bromocriptine, are effective in reducing prolactin levels and increasing sexual desire in men with hyperprolactinemia, they are highly likely to produce overpowering side effects in men with normal prolactin levels (21). Recently, a new dopamine agonist, shown to be effective in enhancing sexual behavior in male rats, has been evaluated in the treatment of human male sexual dysfunction. However, D2 side effects were frequent and often severe (Bancroft, unpublished).

Systemic treatment with a range of dopaminergic drugs has long been known to affect profoundly the display of sexual behavior in rodents. Recently, it has become apparent that these effects of dopamine receptor blockade are not simple; with more careful attention to differential effects on D1 and D2 receptors as well as dosage, discrete effects on precopulatory behaviors have been demonstrated. For example, in male rats all neuroleptic drugs decreased rates of conditioned level changing, and atypical neuroleptics delayed dose-dependently the initiation of copulation, but had little effect on copulatory behavior once it was initiated (58). Metoclopramide, on the other hand, reduced dose-dependently the number of intromissions to ejaculation but had no effect on initiation latencies, whereas the typical neuroleptics haloperidol and pimozide both prolonged mount and intromission latencies and also altered the number of intromissions preceding ejaculation (58). The different behavioral effects of these drugs may be interpreted in the light of recent in vivo neurochemical data indicating the differential release of dopamine in the preoptic area and the dorsal and ventral striatum that is correlated with relatively discrete epochs of sexual interaction (see below). Thus, those dopaminergic antagonists that more or less selectively block dorsal–striatal (i.e., caudate–putamen) dopamine receptors affect only copulatory responses; in fact, they may actually decrease the ejaculation threshold in terms of the number of intromissions required to reach it. By contrast, those drugs that predominantly affect ventral–striatal dopamine receptors, such as the atypical neuroleptics (e.g., clozapine), delay the initiation of copulation with no other behavioral effects. In other words, appetitive rather than copulatory elements of sexual behavior are primarily affected.

The special sensitivity of such precopulatory responses to dopaminergic blockade was further demonstrated in experiments utilizing a second-order schedule of sexual reinforcement. Thus, the mixed D1/D2 dopaminergic receptor antagonist, a-flupenthixol, dose-dependently decreased in response to a receptive female. The drug also prolonged mount and intromission latencies, although at doses slightly greater than those that decreased instrumental responses. More importantly, these effects on appetitive or precopulatory behavior were achieved at doses of the drug that had no significant effect on mounts or intromissions. Only at the highest doses tested were all measures of sexual behavior affected, when the majority of males did not intromit or ejaculate (19).

It is important to consider the neural site of action of these behavioral effects of dopaminergic drugs. However, a number of investigators have also explored actions within the medial preoptic area (mPOA). Infusion of dopamine receptor agonists into this area selectively enhanced measures of copulatory behavior, such as intromission rates and efficiency, but did not affect latencies to mount and intromit (8). Dopaminergic receptor antagonists, or presynaptic doses of the agonists, tended to have opposite effects. All these effects were suggested to depend upon alterations in the autonomic control of penile reflexes, for example, lengthening the latency to erection. This demonstrates the often neglected importance (at least in animal studies) of genital changes in mediating the apparent motivational effects of drugs when measured only as alterations in initiation latencies.

It is evident that the marked changes in appetitive aspects of masculine sexual behavior that follow systemic (8, 58, 19) or intracerebroventricular (icv) infusion (8) of dopaminergic drugs do not obviously follow direct intrahypothalamic treatment, although infusions of haloperidol into the mPOA reduced rates of conditioned level changing (58), an appetitive response that does not appear to depend upon alterations in penile responsiveness. Clearly, sites of action outside the hypothalamus must mediate these effects of dopaminergic drugs on noncopulatory measures of sexual activity, the striatum being an obvious candidate.

Manipulating dopamine in the ventral striatum affects appetitive sexual responses, but not copulation itself. Thus, infusing the dopamine releaser d-amphetamine into the nucleus accumbens dose-dependently increased instrumental responding for access to an estrous female, reduced latencies to mount and intromit, but did not alter copulatory responses (such as mounts, intromissions and ejaculation), hit rate, or ejaculation latency (Fig. 1) (19). Conversely, lesioning the dopaminergic innervation of the ventral striatum by infusing the neurotoxin 6-hydroxydopamine (6-OHDA), significantly lengthened mount and intromission latencies, also without altering copulatory performance (19). Similarly, infusing the predominantly D2-receptor antagonist, haloperidol, into the ventral striatum reduced rates of conditioned level changing but did not affect measures of copulation (58).

The impact of these manipulations on incentive motivational processes in a sexual context was further demonstrated by a procedure that effectively devalued as sexual incentives the receptive females with which neuroleptically treated males were interacting. This was achieved by injecting hormone-primed females systemically with the neuroleptic, a-flupenthixol. This treatment selectively abolishes the female's proceptive soliciting responses but actually enhances her display of immobile lordosis postures (18). In this condition, mounts and intromissions only occur if the male initiates them and, even in normal males, this results in prolonged mount and intromission latencies (19). Males infused into the ventral striatum with the D2-dopamine receptor antagonist, raclopride, were markedly affected by this coincident treatment of females with a-flupenthixol, such that latencies to mount and intromit were more than doubled (Fig. 1). Some males were so affected by this relative immobilization of females, that mounting and intromitting were actually prevented, even though males treated in this way showed only modest increases in their latencies to mount and intromit when with untreated females who were actively soliciting (19). These data demonstrate the importance of the dopaminergic innervation of the ventral striatum in the display of appetitive responses to sexual incentive stimuli.

The ability of estrogen to stimulate DA release and to augment DA release and behavior in response to amphetamine has been well established (reviewed in ref. 7). Thus it would appear that brain DA systems are influenced by estrogens, and this may be one mechanism underlying the control of female sexual behavior, although this has been difficult to establish. Systemic administration of DA agonists can facilitate or inhibit lordosis behavior in ovariectomized rats primed with estrogen and progesterone or estrogen alone (18). Paradoxically, systemic administration of a range of doses of DA antagonists also facilitates lordosis, although the behavioral signature of the effect is different. Whereas DA agonists can increase lordosis quotients and proceptivity counts, DA antagonists increase lordosis quotients and the duration that female rats hold the lordosis posture, but abolish proceptive behavior (18). These latter effects have led to the suggestion that DA facilitates the active behavioral components of female sexual behavior, that is, proceptivity, but inhibits the passive components, that is, lordosis (10). Alternatively, the effects of DA antagonists could be viewed as decreasing the ability of the female to disengage from lordosis once it is initiated or to switch between lordosis and proceptive pacing, two mutually exclusive behavioral sequences.

Both the facilitatory and inhibitory effects of DA agonists and antagonists appear to occur through an action on D2 receptors, as selective D1 agonists and antagonists do not affect measures of sexual proceptivity or receptivity in female rats (23, 24). However, there is little agreement concerning the involvement of pre- and postsynaptic DA receptors. Few studies have examined the central sites of action of DA in female sexual behavior. Neurochemical lesions of the mesolimbic DA pathways with 6-OHDA facilitate lordosis behavior in ovariectomized, estrogen-primed rats (66), suggesting that DA release in the nucleus accumbens or other projection regions of this pathway are inhibitory to lordosis. However, others have found no effect of mesolimbic 6-OHDA lesions on lordosis or proceptive behaviors following a similar level of depletion of DA and its acid metabolites (25). No studies have yet examined the effect of DA agonists or antagonists in other regions of the brain or on other forms of appetitive sexual responding in females.

Yohimbine, an a2 receptor antagonist, has long been regarded as an aphrodisiac. Recently, a number of placebo-controlled studies of its effects on erectile dysfunction have been reported (65, 74). Although each study has certain methodological problems, the consistency of the results suggests a positive effect, at least in a subgroup of men with erectile dysfunction. Equally important has been the striking lack of side effects in contrast to comparable attempts to evaluate dopamine agonists. The main problem in evaluating this evidence is that there was no underlying rationale for using the drug. In each case, men with erectile dysfunction were involved resulting in heterogeneous groups. It was assumed that the drug would enhance erectile response, but no consideration of the type of erectile response or circumstances relevant to the response were taken into account (see ref. 21). Recent studies, however, have confirmed the potential of a2 receptor antagonists, especially in a younger group of men with psychogenic impotence (Bancroft, unpublished).

In rodents, there are marked effects of a2 adrenoceptor agonists and antagonists on measures of masculine sexual behavior. For example, yohimbine increased the proportion of sexually naive male rats that mounted, intromitted, and ejaculated and decreased latencies to initiate copulation (14, 15). Idazoxan, enhanced copulatory rate and decreased ejaculation latency (15). Agonists at the a2 receptor had essentially opposite effects (15) that were interpreted as being mediated by changes in sexual arousal.

However, the sexual specificity of the effect remains unclear. It seems unlikely that a drug interacting with diffuse, noradrenergic projection systems that innervate the entire neuraxis are having an effect exclusively on sexual behavior. Furthermore, changes in sexual responsiveness may be mediated by spinal and/or peripheral autonomic effects. Nevertheless, as indicated above, positive effects on sexual responsiveness in men have been obtained with minimal side effects.

There is considerable evidence that the hormonal changes that underlie lordosis behavior and certain neuroendocrine reflexes, such as the preovulatory luteinizing hormone (LH) surge and pseudopregnancy, are associated with altered norepinephrine transmission. Systemic treatment with a and b receptor agonists and antagonists clearly modulate feminine sexual behavior, but no clear picture emerges (22, 43). Central effects of adrenergic treatments on female sexual behavior have not been studied in detail. Infusion of the a1 antagonist prazocin into the ventromedial hypothalamus (VMH), but not the mPOA, inhibited lordosis (22, 17), whereas infusions of the a2 antagonist idazoxan or the b antagonist metoprolol into the VMH had only a small inhibitory effect in some animals (17). In contrast, infusions of metoprolol into the mPOA inhibited lordosis in most rats (17). Together, these results have been suggested to indicate that stimulation of a1 receptors in the VMH plays some part in the hormonal facilitation of lordosis behavior, whereas stimulation of b receptors in the mPOA, and to a lesser extent in the VMH, appears to inhibit lordosis. Nothing is known about the role of norepinephrine in other forms of appetitive sexual responding in female rats.

There is minimal methodologically-sound evidence on the effects on human sexuality of manipulating the serotoninergic system. Ware et al. (76) compared the effects of trazadone, which inhibits serotonin reuptake and decreases 5-HT2 receptor binding, with trimipramine and placebo in six normal volunteers. Trazadone had an enhancing effect on NPT that was independent of any direct effect on rapid-eye-movement (REM) sleep, but no effect on waking erections was reported. Otherwise, there are only anecdotal reports of patients taking serotoninergic drugs such as fluoexetine, trazadone, and fenfluramine, which suggest both negative and positive effects on sexual arousal and performance measures (see ref. 67).

The long history of facilitative effects on masculine sexual behavior of impairing serotonin transmission have been complemented recently by the observation that the selective 5-HT1A receptor agonist, 8-OH-DPAT, markedly facilitated sexual behavior when given systemically to male rats (1, 27). The effect is a dramatic one; treated subjects were seen to ejaculate after only one or two intromissions, rather than the more usual ten or twelve. The apparently paradoxical effect of this drug, namely that a 5-HT receptor agonist has a facilitatory effect on copulatory behavior similar to, or even greater than, global 5-HT depletion while being opposite to the effects of other direct, or indirect, nonspecific 5-HT agonists has yet fully to be explored. However, it has been convincingly demonstrated that the facilitatory effects of 8-OH-DPAT depend upon actions at the 5-HT1A autoreceptor on midbrain raphé serotonin neurons resulting in the inhibition of their activity (27) (see also Molecular Biology of the Dopamine Receptor Subtypes, Dopamine Receptors: Clinical Correlates, Serotonin Receptor Subtypes and Ligands, and Serotonin Receptors: Signal Transduction Pathways).

An extensive literature suggests both inhibitory and facilitatory effects of serotoninergic transmission on feminine sexual behavior, depending on the receptor subtype involved, for example, activity at 5-HT1A and possibly 5-HT3 receptors inhibits, whereas activity at either 5-HT1B, 5-HT1C, or 5-HT2 receptors facilitates, lordosis (45). However, Mendelson (45) has suggested that there is little reason to suspect that the powerful effects of serotonergic drugs on lordosis reflect a physiological role of brain serotonin systems in this behavior.

Experimentally, morphine generally inhibits mounting, intromissions, and ejaculation in male rats, whereas naloxone given systemically (often in high doses which are not selective for opiate receptors) tends either to have no effect, or to facilitate aspects of sexual behavior, for example in sexually sluggish rats (see ref. 53). These effects of predominantly m-receptor agonist and antagonist treatments have also been studied using some of the methods described above (instrumental behavior, conditioned place preference, and partner preference), initially by assessing the effects of naloxone given both systemically and into the medial preoptic/anterior hypothalamic area (mPOA) as well as the effects of the endogenous opioid peptide, b-endorphin, infused within the mPOA. The mPOA was studied both because of its well established, central importance in the neural system underlying masculine sexual behavior and also because of its relatively rich innervation by proopiomelanocortin (POMC) containing neurons of the ventral hypothalamus (19, 20).

Infused bilaterally into the mPOA, b-endorphin dose-dependently inhibited sexual behavior (Fig. 1). Mounts and intromissions eventually ceased to occur and the latencies to mount, intromit, and ejaculate were prolonged, eventually to the duration of the 15-min test session or even longer (29, 30). Subsequent experiments explored the nature of this effect, especially assessing whether males lose interest in females as a result of the treatment. Careful observation of precopulatory, investigative responses revealed that males infused with b-endorphin actively investigated and pursued females and, to some extent, made abortive mounting attempts (29); changes in behavior that were seen in a more emphatic way following lesions of the mPOA (20). Analysis of the behavioral sequence revealed that when a treated male and female first made contact, the pattern of interaction was not significantly different from that seen with control males, but that the sequence broke down at the point when investigative responses usually switched to the copulatory responses of mounting and intromitting; thus, b-endorphin infused into the preoptic area appears not to affect sexual interest or arousal, but instead the transition between investigative and copulatory responses of mounting and intromitting (29, 72). These males also showed evidence of thwarting a motivated response tendency in that irrelevant behaviors, such as scratching and grooming, emerged at very high rates in b-endorphin-treated males who could not copulate (29). In addition, these sexually inhibited males now showed a release of the drinking of a preferred sweet solution, which is normally suppressed in the presence of a receptive female with which the male would usually copulate (29).

Infusion of b-endorphin into the mPOA had no effect on instrumental responding for a receptive female presented under a second-order schedule of sexual reinforcement, nor on the expression of a place preference conditioned by sexual interaction with such a female. However, the same treatment rapidly abolished a male's preference for a receptive, over an unreceptive, female tethered in either side of the place preference apparatus (Fig. 2) (30, 19).

A further dimension to the analysis of the effects of b-endorphin followed its infusion into the mPOA after an intromission, rather than before interaction with a female had begun. The peptide no longer had an inhibitory effect on mounting, intromission, or ejaculation. Furthermore, the inhibitory effects of b-endorphin were lost or reduced even if a delay of 2 hr or so was interposed between a single intromission and the infusion, provided the male was retested with the same female with which he had intromitted previously. If a different female was placed with the male following intromission, then the inhibitory effect on copulatory behavior of b-endorphin reappeared (72). These results clearly suggest that b-endorphin does not simply act to prevent copulation, because, if the male is allowed to begin interacting sexually with an individual female, the otherwise inhibited behavior can be emitted normally and ejaculation will occur.

Taken together, the results of these experiments reveal the remarkable behavioral specificity of the effects of b-endorphin infused into the mPOA. The peptide does not, apparently, influence appetitive aspects of sexual behavior nor reward-related processes, so far as these are assessed in the place preference procedure. Instead, intrahypothalamic b-endorphin appears to prevent the display of the copulatory reflexes, which together form the consummatory elements of the sexual response sequence. However, even here the effects of the peptide are not absolute, because if the male is allowed to engage sexually with a female, then the inhibitory effects of the peptide are themselves inhibited. The impaired preference for a receptive female that follows b-endorphin infusion appears to be related to the failure to initiate the copulatory sequence—presumably only if the peptide is infused prior to an intromission, although this has not been tested explicitly by studying its effects on preference when infused after an intromission.

Therefore, a neural mechanism may exist in the mPOA that allows the appropriate behavioral responses of mounting and intromitting to be matched to a relevant sexual incentive stimulus, and this is what is impaired in the b-endorphin-treated male rat. This may suggest a particular relationship between the intrinsic sensory properties of a receptive female and the species-specific motor output of copulatory reflexes, since acquired motor responses (e.g., bar-pressing or conditioned approach) for conditioned incentives, such as a light CS or the properties of the preferred place, are completely unaffected by the same b-endorphin manipulation of the mPOA. The behavior of a rat in the operant and place preference procedures may, therefore, be controlled by the same mechanisms that underlie unconditioned appetitive or preparatory responses, and these appear not to reside within the mPOA, because lesions of the structure are also without effect on these parameters (19, 20). Reduced preference for an estrous female after b-endorphin infusions, however, may be related more to the switch from sexual to ingestive responses, which is seen to follow manipulations of testosterone or hypothalamic b-endorphin in male rats (29).

The effects of naloxone on sexual behavior have also been studied using a number of these behavioral procedures, and the results also indicate the complexities underlying the superficially simple effects of the systemically administered opioid antagonist. Systemic naloxone may, in some circumstances, facilitate sexual behavior in male rats (see above), but if given to intact, sexually active males these effects are vanishingly small (29). However, the same treatment reduced instrumental responses for a female presented under a second-order schedule of reinforcement, promptly abolished a previously acquired conditioned place preference, but had no effect on partner preference (19, 31). Infusing naloxone bilaterally into the mPOA resulted in a powerful facilitation of copulatory behavior; males required fewer intromissions to ejaculate with a much reduced latency, yet had no effect on instrumental behavior or a conditioned place preference. Partner preference was seen to be reduced following this treatment, but analysis of the data revealed this to be an epiphenomenon of the increased sexual activity, since males spent more of the test period in a state of refractoriness and therefore away from both females in the neutral compartment of the choice apparatus (19, 20, 31).

It is clear from these results that opioid mechanisms within the POA do not seem primarily to be involved with incentive motivational responses to sexual stimuli but are more involved with consummatory sexual responses. However, more interesting, perhaps, is the indication that such incentive, or reward-related, responses are sensitive to systemic naloxone. In addition, it has been shown that this sensitivity is much greater in animals that were recently castrated (41, 46). Because castration also profoundly affected a sexually conditioned place preference (19, 31) and instrumental sexual responses (19), whereas mPOA lesions, as well as intra-mPOA opioid manipulations, were without effect on these measures, opioid involvement in sexual reward-related processes may both be sex-hormone-dependent and involve extrahypothalamic substrates (20). The ventral–tegmental area dopaminergic system innervating ventral–striatal and limbic structures is an obvious focus for experiments investigating this problem.

Indeed, in a particularly interesting series of experiments, Mitchell and Stewart (48, 49) have demonstrated that infusions into the ventral–tegmental area of morphine and dynorphin1–13 increased the number of males that mounted and showed female-directed behavior (47). Morphine, but not dynorphin1–13, increased DA metabolism in the nucleus accumbens, indicating that the effects of these opioid peptides infused into the A10 region may have DA-dependent and DA-independent actions on sexual behavior. In addition, it was demonstrated that masculine sexual behavior was facilitated when males were placed in an environment that previously had been associated with systemic injections of morphine (49). This important observation demonstrates that the conditioned reinforcers established through the pairing of a previously arbitrary constellation of cues with the positive incentive effects of morphine can significantly affect, in this case facilitate, sexual behavior that is under the control of the conditioned and unconditioned incentive properties of an estrous female.

Infusion of another member of the POMC peptide family, melanocyte-stimulating hormone (MSH), into the mPOA results in a markedly different pattern of effects than those seen following infusions of b-endorphin. This peptide facilitates, rather than inhibits, sexual behavior: ejaculation and intromission latencies, as well as the postejaculatory interval are shortened, ejaculation occurs after fewer intromissions, and the number of ejaculations occurring within a 15-min test is increased (29). Although this same treatment has no direct effects on place or partner preference, it results in a small, but significant, increase in responding under the second-order schedule of sexual reinforcement.

Systemic administration of the k-receptor agonist U-50,488H, decreased sexual behavior in male rats, and this effect was prevented by systemic naloxone or intracranial infusions of a k-receptor antagonist, nor-binaltorphimine (NBNI). The latter compound markedly facilitated female-directed behavior and also prevented the effects of systemically administered U-50,488H when infused either into the ventral–tegmental area or the mPOA (33).

From these results, it is apparent that there are quite complex opioid influences on sexual behavior and that the sites of action of m- and k-receptor agonists and antagonists determine markedly different effects on sexual behavior, some of which may require interactions with DA neurons in the ventral–tegmental area. What remains unclear are the circumstances under which these opioid systems are activated so as to modulate the neural systems underlying sexual behavior. One possibility is that they come into play, especially the POMC-containing neurons of the hypothalamus, under conditions of stress to mediate the inhibition of reproductive function (26).

Systemic administration of opiate agonists, such as morphine or heroin, inhibits the sexual behavior of female monkeys, dogs, and rodents (reviewed in ref. 53). In female rats, acute systemic morphine decreases lordosis quotients and reflex scores, and abolishes proceptive pacing and solicitation. These effects are reversible with naloxone or naltrexone. However, there have been no consistent reports of an effect of systemically administered opioid antagonists on female sexual behavior, suggesting that endogenous opioids do not exert a tonic inhibitory influence, at least not in females that display full receptivity and proceptivity.

In contrast to an exclusive inhibitory effect of systemic morphine, centrally administered opioid agonists may inhibit or facilitate the sexual behavior of female rats depending upon the receptor type and brain area stimulated. For example, b-endorphin can inhibit or facilitate lordosis behavior following infusion of comparable doses into the third ventricle (79) or lateral ventricles (54), respectively. However, low doses of b-endorphin and the m-selective agonist morphiceptin inhibit lordosis, whereas high doses of these agonists facilitate lordosis following infusion into the lateral ventricles (54). Lateral ventricular infusions of the m-selective agonist D-Ala2-MePhe4-Gly-ol5-enkephalin (DAMGO) inhibit lordosis in rats primed with estrogen and progesterone, but not in rats primed with estrogen alone (54, 59), suggesting that progesterone enhances the ability of m agonists to exert their inhibitory effects. Infusions of morphine to the ventromedial hypothalamus or mesencephalic central grey inhibit lordosis behavior (75), as do infusions of b-endorphin into the medial preoptic area, VMH, mesencephalic central gray (MCG), or mesencephalic reticular formation (71, 79). The inhibitory effect of b-endorphin and morphiceptin may occur through an interaction with high-affinity m1 receptors, as the selective m1 antagonist naloxazone can reverse the inhibitory effect of these peptides (54, 79). However, unlike m agonists, central infusions of the a -selective agonists D-Ser2-Leu5-Thr6-enkephalin (DSTLT) and D-Pen2-D-Pen5enkephalin (DPDPE) into the lateral ventricles facilitate lordosis behavior (54, 59), as do infusions of the k-selective agonists U50-488 and leumorphin (59). Infusions of DPDPE into the lateral ventricles also facilitate proceptive behaviors (59).

Central administration of opioid antagonists also facilitates or inhibits female sexual behavior depending upon the brain area and hormonal state of the animal. For example, infusions of naloxone intrathecally or into the mesencephalic central grey can facilitate lordosis in female rats primed with estrogen alone (71, 77), whereas infusions of naloxone into the lateral ventricles inhibit lordosis in females primed with estrogen and progesterone (34). Although naloxone binds with highest affinity to m receptors, an effect on other receptors cannot be ruled out. The effects of selective opioid receptor antagonists have not been examined in females.

It is difficult at present to incorporate these effects of opioid drugs on feminine sexual behavior within a clear conceptual framework, in marked contrast, therefore, to the situation in males.

Oxytocin is a nonapeptide secreted by the posterior pituitary, but it is also released into the brain from the terminals of neurons located in the supraoptic and paraventricular hypothalamus, the latter of which projects axons inter alia to the ventrolateral regions of the VMH, lateral septum, brainstem and spinal cord autonomic regions (see Rinaman et al.). Both estrogen and testosterone stimulate the synthesis of oxytocin binding sites, and recent evidence links oxytocin with the facilitation of sexual behavior in both male and female rats, and perhaps with the stimulation of sexual desire and arousal in humans (see also Vasopressin and Oxytocin in the Central Nervous System).

Systemic injections of oxytocin, either intravenous (iv) or intraperitoneal (ip), reduce the number of intromissions required for ejaculation, decrease the ejaculation latency, and increase the number of ejaculations in a timed test (5, 74). Central infusions of oxytocin into the ventricles also produce a syndrome of yawning and penile erections in male rats that is displayed even in the absence of a receptive female (2). The effects of icv infusions of oxytocin on male sexual behavior seem to be of a dual nature depending upon the dose and brain area. Very low doses of oxytocin infused into the lateral ventricles stimulate male sexual behavior by decreasing both the ejaculation latency and postejaculatory interval (2, 4). In contrast, higher doses infused into the third ventricle increased the mount and intromission latencies and the postejaculatory interval (73). The infusion of lower doses into the third ventricle did not affect sexual behavior. Infusion of very low doses of the oxytocin antagonist d(CH2)5Tyr(Me)-[Orn8]-vasotocin dramatically and dose-dependently inhibited penile erections and copulatory behavior in sexually experienced male rats (2, 3). This antagonist also prevented the induction of the yawning and penile erection syndrome by icv infusions of oxytocin. These data suggest that endogenous oxytocin serves to facilitate certain components of sexual arousal and copulatory behavior. Hughes et al. (28) demonstrated that oxytocin levels in cerebrospinal fluid (CSF) more than doubled after 5 min of copulatory behavior and more than tripled 20 min after an ejaculation. However, although lesions of the lateral and posterior paraventricular hypothalamus prevented the rise in CSF oxytocin levels during copulation, and produced a small but significant increase in the mount and intromission latencies, those lesions had little effect on copulatory behavior once it was initiated. Moreover, those lesions decreased the postejaculatory interval. Thus, it remains to be established whether endogenous oxytocin release is necessary or sufficient for male sexual behavior. The effect of oxytocin on measures of conditioned sexual arousal in male rats is not yet known.

Systemic or intracranial infusions of oxytocin in male prairie voles (Microtus ochrogaster) inhibit copulation (80). However, this may be the result of a concomitant increase in social affiliative behaviors (e.g., side-by-side contact). Although oxytocin agonists and antagonists have not been tested on human subjects, exposure to erotic stimuli and masturbation to orgasm in men leads to an increase in plasma oxytocin levels (13).

Oxytocin and its receptor are highly regulated in different brain regions during the estrous cycle, an effect that is likely due to changes in estrogen and progesterone. Numbers of oxytocin immunoreactive cells, and levels of oxytocin receptor mRNA, are stimulated by estrogen and enhanced by progesterone. Estrogen stimulates the transcription of oxytocin binding sites in the VMH, and progesterone quickly modifies the effect of estrogen by spreading the active oxytocin receptor sites laterally to innervate oxytocin terminal regions (69). This suggests that estrogen and progesterone may act to synchronize the availability of endogenous oxytocin with its receptor.

Lordosis behavior is facilitated dramatically by oxytocin. Systemic injections of the nonapeptide increase the frequency of lordosis responding in estrogen-primed females (5, 11). Similarly, icv infusions increase lordosis behavior in rats treated chronically with estrogen or treated with estrogen and progesterone, suggesting that progesterone may exert a permissive effect on the actions of exogenous oxytocin (11). Caldwell (12) identified the mPOA as an important site for the regulation of lordosis behavior by oxytocin in female rats. Infusions of the nonapeptide into the mPOA increased lordosis quotients in rats treated with estrogen and given multiple tests of copulation. Most importantly, this facilitation occurred at doses that were ineffective in other sites, such as the third ventricle, VMH, ventral–tegmental area, or the mesencephalic central grey. However, a recent study provided evidence that oxytocin in the mPOA and VMH contribute to different aspects of lordosis behavior in rats primed with estrogen and progesterone. Infusions of very low doses into the VMH facilitated the duration of each lordosis, without affecting the lordosis frequency, whereas infusions of higher doses into the mPOA facilitated the lordosis frequency but not the duration of each lordosis (70). Additionally, oxytocin infusions into the VMH of estrogen-primed female prairie voles reduced rates of aggression and increased the amount of physical contact that females made with males, although it led to a faster termination of estrus (81). Nothing is known about the possible effects of oxytocin on appetitive aspects of sexual behavior in female rats.

Sexual stimulation during copulation increases plasma oxytocin levels in female rats, as does vaginal distention or vaginocervical stimulation. In human females, masturbation to orgasm also increases plasma oxytocin levels (13). Thus, taken together, the data in males and females suggest that oxytocin may form part of a neurochemical axis that participates in the desire to affiliate with a sexual partner, to engage in sexual contact, and to achieve sexual satiety after extended matings.

The effects of alcohol on sexual behavior are of interest for at least two reasons. First, alcohol consumption has been repeatedly implicated in the etiology of various sexual disorders and dysfunctions in humans, such as inhibited sexual desire, erectile failure, delayed ejaculation, and inhibited orgasm. Second, alcohol has long been associated with sexual disinhibition and has been reported to enhance sexual arousal and behavior in some individuals with preexisting sexual dysfunctions, such as premature ejaculation, inhibited sexual desire, or inhibited orgasm.

Only two published experiments exist in which the effects of alcohol on human sexual behavior have been examined directly, one in men (38) and one in women (39). In both studies, the effects of a range of doses of alcohol were examined on the ability of subjects to masturbate to orgasm while viewing a sexually explicit film. In men, alcohol dose-dependently delayed ejaculation, reduced the intensity of orgasm, and decreased both physiological and subjective measures of sexual arousal. Nearly identical effects were observed in women, although alcohol increased, rather than decreased, their subjective levels of sexual arousal. These reports generally support clinical observations that alcohol can disrupt sexual activity; however, their relevance to other forms of sexual behavior is unclear. Furthermore, it has been difficult to find unambiguous evidence of the disinhibition of sexual arousal commonly associated with low-to-moderate doses of alcohol (see also Caffeine-A Drug of Abuse? and Phencyclidine).

The effects of a range of doses of alcohol, administered ip, on the copulatory behavior of sexually active male rats were studied by Pfaus and Pinel (55). Low doses of alcohol increased the mount, intromission, and ejaculation latencies, and reduced the percentage of rats that achieve ejaculation in a 30-min test (55). A moderate dose intensified these effects further and increased the postejaculatory interval, whereas a high dose abolished all copulatory activity. No disinhibitory effect was observed at any dose. However, sexually active rats do not show evidence of sexual inhibition; therefore, a second study was conducted in rats that had learned to suppress their copulatory advances toward sexually nonreceptive females. Sexually active males were given sequential access to sexually receptive and nonreceptive females at 48-hr intervals. Although all males attempted to mount the nonreceptive females in the first training session, none attempted to do so by the sixth and subsequent sessions, despite being fully sexually active during interspersed sessions with receptive females.

The effects of low and moderate doses of alcohol in these males were examined in two tests following this training period, one during exposure to receptive females, and one during exposure to nonreceptive females. As in the first experiment, both doses of alcohol disrupted sexual behavior when males were exposed to receptive females. However, the low dose dramatically released mounting behavior and subsequent ejaculations from inhibition during the test with nonreceptive females, despite vigorous defensive and rejection responses made by the females and despite the fact that none of the males gained vaginal intromission. The moderate dose did not produce any disinhibition of sexual behavior during this test. These data indicate that alcohol can produce a disinhibition of sexual behavior in the male rat but only under conditions in which the rat's behavior is already inhibited.

Another critical feature of alcohol's effects on male sexual behavior is that tolerance may accrue to its disruptive effects. Although empirical reports are sparse in the clinical literature, tolerance to a variety of effects of alcohol have been noted in the animal literature, including its anticonvulsive effects and its disruption of balance and motor coordination. Such tolerance is said to be contingent, that is, its development depends largely upon the occurrence of some experience or behavior during periods of drug exposure.

The development of tolerance to alcohol's disruptive effect on sexual behavior has been examined in male rats (63). Rats were administered a moderate dose of alcohol either before or after they engaged in a 30-min test of sexual behavior with receptive females. The control group received an equal volume of saline. Tests were conducted every 4 days, to approximate the normal estrous cycle of the female. This dose of alcohol increased the mount, intromission, and ejaculation latencies and increased the postejaculatory interval in the alcohol-before group on the first tolerance-development trial, although these disruptive effects became progressively less pronounced in subsequent trials. By the fifth trial, these measures were indistinguishable from control values and from the animals' own no-drug baseline values. No lingering effect of this dose of alcohol was observed on the sexual behavior of the alcohol-after group. On the final test, this dose of alcohol was administered to rats in the three groups before the test of sexual behavior. Significant increases in the mount, intromission, and ejaculation latencies, and in the postejaculatory intervals, were observed in rats of the saline control and alcohol-after groups, but not in rats of the alcohol-before group, thus demonstrating that contingent tolerance to alcohol's disruptive effects had occurred. These results suggest that some of the variability and inconsistency in the magnitude and nature of alcohol's disruptive effect on human sexual behavior may be attributed to differences in the frequency with which individuals have previously engaged in sexual activity while intoxicated.

Very little work has been done to examine the effects of alcohol on female sexual behavior. In both female hamsters and rats, moderate-to-high doses of alcohol inhibited receptive and proceptive behaviors (62). However, in the latter study, a low dose of alcohol facilitated proceptive behavior in ovariectomized female rats who normally display a small amount of proceptive behavior following treatment with estrogen and a moderate dose of progesterone. Thus, alcohol may have synergistic effects with progesterone in the disinhibition of proceptive behaviors.

The analysis of drug effects on behavior gives rise to hypotheses concerning transmitter activity during ongoing behavior. Although ex vivo studies, which rely on postmortem tissue analysis of transmitter levels using high-performance liquid chromatography (HPLC) with electrochemical detection (ED), have provided suggestive data in this regard, they do not permit a clear resolution of which behaviors contribute to the increases or decreases in transmitter levels. The more recent development of techniques to monitor extracellular concentrations of neurotransmitters in vivo, most notably microdialysis and voltametry, has moved us closer to resolving these issues. Often, however, these techniques provide information that in some ways underscores the limits of the pharmacological approach (see Modification of Central Catecholaminergic Systems by Stress and Injury: Functional Significance and Clinical Implications).

Microdialysis has been used to examine extracellular concentrations of DA, its acid metabolites DOPAC and homovanillic acid (HVA), and the serotonin metabolite 5-hydroxyindoleacetic acid (5-HIAA) during preparatory and consummatory aspects of sexual activity in male rats. Recall that DA antagonists are capable of reducing rates of conditioned level changing in male rats at doses that do not affect the initiation and subsequent performance of copulatory behavior (58). This led us to suggest that, as with feeding, preparatory sexual behavior is highly dependent upon the functional integrity of brain DA systems, whereas consummatory sexual behavior is less dependent (50). Accordingly, we predicted that DA release would be more pronounced during preparatory sexual behavior compared to copulation. Exactly the opposite occurred. Dialysates from the nucleus accumbens of sexually active male rats revealed a small increase in DA and its metabolites during the preparatory phase, but a dramatic and sustained increase during active copulation (Fig. 3) (56). In contrast, dialysates from the dorsal striatum revealed a small and progressive increase in DA and its metabolites throughout the test session that was not correlated with any particular aspect of copulation (Fig. 3). Subsequent work established that the increases were not due to general locomotor activity, nor were they due to the novelty of the testing chamber (16), and that the increased DA transmission in the nucleus accumbens during these phases of sexual activity did not require sexual experience (40, 78). These results suggest another interpretation of the role of DA in sexual behavior. Given that DA receptor antagonists bind competitively with DA for occupancy at DA receptors, the greater sensitivity of preparatory behaviors to the disruptive effects of these drugs may reflect the lower concentration of extracellular DA that is available to compete with the drug. In contrast, consummatory behaviors may be less susceptible to disruption by DA antagonists because DA release is high during such behavior. However, to date consummatory elements of masculine sexual behavior have not been seen to be disrupted following DA receptor antagonist infusions into, or 6-OHDA-induced lesions of the dopaminergic innervation of the ventral striatum.

The use of in vivo voltametry has further enhanced our understanding of the role of DA in male sexual behavior. This technique has shown a small phasic rise in DA oxidation current in the nucleus accumbens that is of short duration during the preparatory phase, but a larger and sustained increase during copulation (Fig. 3) (61). However, the DA oxidation current declines precipitously after ejaculation, during the absolute refractory period, but rises again before the male reinitiates another copulatory series (Fig. 3). The dynamic nature of DA transmission in the nucleus accumbens contrasts with that of the dorsal striatum, which, as observed with microdialysis, showed a small but progressive increase throughout the test that was not correlated with any specific phase of sexual behavior. The DA oxidation current in the nucleus accumbens of male rats also appears to be sensitive to olfactory or pheromonal cues provided by estrous females (47), although the increases observed during the presentation of estrous vaginal secretions on a glass slide are less than those obtained during the presentation of an estrous female (61). In contrast, DA oxidation currents did not increase in the dorsal striatum during the presentation of sexually relevant stimuli.

In females, DA antagonists facilitate lordosis but inhibit proceptivity, suggesting that DA is inhibitory with respect to lordosis, but may facilitate proceptive behaviors. In contrast, NE may facilitate lordosis without affecting proceptive behaviors. Dialysis samples have been taken from the nucleus accumbens and dorsal striatum during ongoing sexual behavior in ovariectomized rats fully primed with estrogen and progesterone (Pfaus, Wenkstern, and Fibiger, unpublished data). Dopamine transmission increased comparably in both regions during sexual behavior, although the effect in the nucleus accumbens seemed to be related to the incentive quality of the stimulus male, whereas the effect in the striatum seemed to be related to the ability of the female to pace the copulatory contact (50). Increases in DA transmission have also been reported during sexual behavior in female Syrian hamsters, who assume the lordosis posture for very long durations (42). Both DA and norepinephrine transmission increase in the VMH of female rats during sexual behavior (75). These latter results are easier to interpret, as both DA and noradrenergic agonists are reported to facilitate lordosis following infusion into the VMH.

In summary, in vivo techniques are fast becoming a major experimental tool, one which can provide direct assessments of neurotransmitter activity during behavior and which can refine our understanding of pharmacological manipulations.

One of the most exciting techniques to come from the revolution in molecular biology is the use of antisense oligonucleotides to halt the transcription of particular genes (see Basic Concepts and Techniques of Molecular Genetics and ref. 37). Antisense to mRNA for glutamic acid decarboxylase (GAD), the enzyme that converts glutamate to GABA, was shown to inhibit lordosis behavior in ovariectomized female rats made continuously receptive with chronic estrogen implants (35). The effect was maximal 24 hr after infusions into either the hypothalamus or the MCG, and took between 4 and 6 days to recover. Control animals that received the same nucleotides, but in a scrambled order, did not show a reduction in lordosis behavior following infusion. These results are similar to those obtained following infusion of GABA receptor antagonists to those regions. In contrast, infusion of GAD antisense to the mPOA did not affect lordosis behavior, a finding consistent with the fact that GABA in this region does not affect lordosis. Thus, the inhibition observed following infusion of antisense was predictable, site-specific, and reversible. Successful modulation of female sexual behavior was also accomplished by infusion of antisense to progesterone receptor mRNA to the VMH of ovariectomized female rats primed with estrogen and progesterone (51, 64). Antisense infusions reduced lordosis behavior and blocked the induction of proceptive behaviors following progesterone treatment. Infusion of the scrambled control sequence had little effect. Finally, McCarthy et al. (36) demonstrated that infusions of estrogen receptor (ER) mRNA antisense into the hypothalamus protected neonatal females from the masculinizing effects of neonatal androgen treatment, including the masculinization of their sexual behavior, locomotion, and volume of sexually dimorphic brain nuclei. In contrast, infusions of ER mRNA antisense to the hypothalamus of neonatal males demasculinized their physical appearance (i.e., they had shorter anogenital distances) and disrupted their ability to show male sexual behavior 60 days after antisense treatment.

Antisense technology, although still in its formative stage, shows much promise in providing us with the ability to limit or halt gene expression selectively within discrete brain regions in vivo. As more and more genes that encode specific proteins in the brain are cloned and sequenced, the causal relationship between their expression and their effects on sexual behavior can be explored, rather than inferred from the administration of relatively nonselective drugs.

There has been substantial progress in defining the consequences for the sexual behavior of males and females of manipulating monoamine, amino acid, and peptide transmitters in the CNS. However, the conceptual framework within which these data exist is unclear in many instances. This is in part from limitations in behavioral methods and because only some aspects of reproductive behavior have been studied. For example, in females, the lordosis reflex has been investigated to the virtual exclusion of any other aspect of feminine sexual behavior. Such limitations make between-sex and, especially, between-species comparisons very difficult indeed, and this is one reason for the slow progress in developing clinically viable drugs with which to treat disorders of sexual arousal and performance.

At a neural level of analysis, there is also a problem of interpretation of such psychopharmacological data which concerns the nature of the neural system that underlies the expression of sexual behavior. A common interpretation of drug effects on sexual behavior is as follows: receptor-binding drug x enhanced behavior y (e.g., mounting/lordosis), thus indicating that transmitter z has an inhibitory role in the control of this parameter of sexual behavior. This argument, that the action of the drug defines the (opposite, in this case) functional significance of the affected transmitter as an element of a neural system is, of course, both circular and flawed. Furthermore, experiments involving lesions and hormonal implantation in the CNS that have defined critical elements of the neural system underlying sexual behavior in males and females have not generally mapped onto the neurochemically defined systems with any degree of consistency. Nor is it clear whether the functions subserved by monoaminergic, peptidergic, or amino acid-containing neurons are related to hormone-sensitive mechanisms, to the sensory systems defined by lesion studies as critical for sexual behavior, or to more general processes such as arousal, reward (often called incentive motivation), or attention (see refs. 5, 19).

A relatively novel method that has successfully revealed the parts of the brain that are engaged during sexual interaction is the immunocytochemical visualization of the protein product of the immediate-early gene, c-fos (6; see also Proto-Oncogenes: Beyond Second Messengers). In these experiments, it was demonstrated that neurons in the mPOA are activated during sexual interaction in male rats and that this activation, revealed as increases in Fos-immunoreactivity in many neurons, depends upon the relay of olfactory information via the medial amygdala (in which c-fos was also induced by appropriate olfactory sexual stimuli, similarly the bed nucleus of the stria terminalis) and genital somatosensory information relayed via the midbrain central tegmental field (in which c-fos was induced by intromissions). Lesions of each of these structures comprising nodes in the system severely impair copulation, but do not affect appetitive responses to sexual stimuli (e.g., ref. 19). By contrast, exposure to conditioned cues (i.e., previously neutral stimuli reliably paired with copulation) strongly increased the expression of c-fos in neurons in the nucleus accumbens and basolateral amygdala, but not in those diencephalic and midbrain structures activated following copulation itself (Everitt and Baum, unpublished). Manipulations of these latter structures have little or no effect on copulation, but profoundly affect appetitive sexual responses. It seems a similar system is activated by coitus and/or cervical stimulation in female rats, although the analysis has not extended beyond this to other aspects of sexual behavior (60). How do the neurochemical systems addressed by drug treatments that have been reviewed here interface with these critical neural structures revealed by other experimental approaches to be essential for the display of discrete patterns of sexual response?

In large part, this is the challenge of the next generation of progress. In the case of DA and the ventral striatum, real progress has been made in integrating the effects of dopaminergic drugs on sexual behavior in male rats with theories of incentive motivation and reward (19, 61). This is much less the case for noradrenergic and serotoninergic systems. The effects of neuropeptides on sexual behavior are difficult in most cases to relate to hypothalamic or related limbic mechanisms underlying sexual behavior. Herbert (26) has argued that the high information content of peptide molecules and their preferential distribution in limbic structures indicate a chemical code of behavior and that given peptides have coordinated central and peripheral roles that have been conserved through evolution. But in terms of sexual behavior, this claim is far from secure. For example, LH releasing hormone (LHRH) clearly controls the gonadal axis, but the function of central LHRH neurons in sexual behavior in males is virtually unknown, and it is still unclear whether, despite early interest, LHRH neurons in females are necessary or sufficient for lordosis to occur, even though the peptide influences the occurrence of this response.

The list of drugs and peptides that affect the display of lordosis is enormous and growing. Is it the case that all of the chemically identified neurons, monoaminergic, peptidergic, and more, with which such drugs and peptides interact are really to be considered as part of a single neural system that is normally engaged during the display of a simple, hormone-dependent reflex such as lordosis? Can the same conclusion be drawn concerning the neural basis of sexual behavior in males, which is also profoundly affected by a similarly wide range of compounds? If so, then it becomes crucial to identify the contexts in which each chemically defined unit is brought into play. Not only are we a long way from such an endpoint, but it seems unlikely to be a generally fruitful task. In the case of peptide neurons, there is no doubt that part of the problem in establishing their functional role within such a neural network is greatly limited by the lack of selective tools with which to manipulate them, especially receptor antagonists. Antisense technology is likely to be especially important in this regard. In the case of monoamines, attempts to integrate drug effects on sexual behavior within prevailing theories of the functions of such systems are more advanced, but there are still wide gaps in our understanding.

The neural tools and behavioral approaches with which to make such advances are increasingly available. The next phase must concern itself more with hypothesis testing and less with the phenomenology of drug effects on sexual behavior, dramatic though they are.

published 2000