*Review* **Serotonergic, Dopaminergic, and Noradrenergic Modulation of Erotic Stimulus Processing in the Male Human Brain**

**Heiko Graf 1,\*, Kathrin Malejko 1, Coraline Danielle Metzger 2,3,4, Martin Walter 5,6, Georg Grön <sup>1</sup> and Birgit Abler <sup>1</sup>**


Received: 8 February 2019; Accepted: 12 March 2019; Published: 14 March 2019

**Abstract:** Human sexual behavior is mediated by a complex interplay of cerebral and spinal centers, as well as hormonal, peripheral, and autonomic functions. Neuroimaging studies identified central neural signatures of human sexual responses comprising neural emotional, motivational, autonomic, and cognitive components. However, empirical evidence regarding the neuromodulation of these neural signatures of human sexual responses was scarce for decades. Pharmacological functional magnetic resonance imaging (fMRI) provides a valuable tool to examine the interaction between neuromodulator systems and functional network anatomy relevant for human sexual behavior. In addition, this approach enables the examination of potential neural mechanisms regarding treatment-related sexual dysfunction under psychopharmacological agents. In this article, we introduce common neurobiological concepts regarding cerebral sexual responses based on neuroimaging findings and we discuss challenges and findings regarding investigating the neuromodulation of neural sexual stimulus processing. In particular, we summarize findings from our research program investigating how neural correlates of sexual stimulus processing are modulated by serotonergic, dopaminergic, and noradrenergic antidepressant medication in healthy males.

**Keywords:** erotic stimulus processing; serotonin; noradrenaline; dopamine; fMRI; healthy; human

#### **1. Introduction**

Human sexual behavior is mediated by the integration of endocrine, vascular, peripheral, and central nervous mechanisms. The brain is considered as the "master organ" of sexual functioning [1] and is involved in all successive steps of human sexual behavior [2]. Electrophysiological and behavioral studies provided considerable insights into human sexual function, but underlying neural substrates were largely unknown until functional neuroimaging methods were widely introduced into neuroscientific research. Since then, the basic principles of neural processing of sexual stimulation were described in several studies [2–4]. However, empirical evidence regarding the effects of neuromodulators on these neural mediators was scarce.

Pharmacological functional magnetic resonance imaging (pharmaco-fMRI) provides a valuable tool to examine modulatory effects of different neurotransmitter systems on neural signatures of sexual function. Apart from clarifying these basic principles and the complex interaction between neuromodulators and functional network anatomy of sexual behavior, investigations with pharmaco-fMRI also have the potential to elucidate the neural correlates of treatment-related sexual dysfunction.

Various psychiatric disorders are commonly accompanied by sexual dysfunction, and have an immediate impact on subjective well-being and quality of life [5,6]. Of note, sexual dysfunction also occurs as a frequent side effect of psychopharmacological treatment and considerably compromises adherence to therapy. Clinical observational studies suggest that about 40% of patients with psychopharmacological antidepressant treatment discontinue their medication due to treatment-related sexual dysfunction [7]. Thus, the crucial implication of sexual dysfunction as a disease- and treatmentrelated symptom motivated the investigation of the underlying neural mechanisms.

In this review, we introduce common concepts of sexual behavior and evidence regarding neural substrates of sexual responses. We shortly discuss the challenges investigating the neuromodulation of neural sexual stimulus processing by pharmaco-fMRI. In particular, we summarize our research program that focused on how these neural correlates were modulated by serotonergic, dopaminergic, and noradrenergic antidepressant medication in healthy male subjects.

#### **2. Conceptualizing Sexual Behavior and Neural Responses**

Despite its debut already in the 1960s, the most commonly used model to conceptualize sexual activity is still the sexual response cycle by Masters and Johnson [8]. The term "sexual response" denotes the set of behaviors and functions related to sexual stimulation and the pursuit of a sexual goal. Based on their observations, Masters and Johnson [8] defined four different phases of sexual responses that refer to the sequence of physical and emotional changes during sexual arousal and activity. They distinguished a period of sexual desire and arousal, followed by a plateau, culminating in orgasm and ending in a refraction period. Kaplan proposed a slightly modified triphasic model comprising sexual desire, excitement, and orgasm [9]. However, these models were criticized for the linear sequence of the phases that may, for example, not be entirely transferable to female sexual responses.

Neuroimaging techniques such as positron emission tomography (PET) or functional magnetic resonance imaging (fMRI) made valuable contributions to identify underlying neural correlates of sexual responses. As outlined and summarized by Reference [10], the behavioral and neurofunctional principles underlying the sexual response cycle largely overlap with those related to other primary rewards such as food [10]. Analogous to concepts related to other rewards, Georgiadis and Kringelbach [10] suggested that sexual responses may be characterized by terms of motivation–consummation–satiety or wanting–liking–inhibition. Linking psychological with physiological and neurofunctional processes in more detail, a meta-analysis of functional imaging studies on sexual arousal conceptualized the neurophenomenological model of sexual arousal [2]. The model suggests a cognitive component, comprising the appraisal of and the attention to a subsequent sexual stimulus, which is represented by neural reactivity within the orbitofrontal cortex (OFC), the inferior temporal cortices, the inferior and superior parietal lobules, premotor and supplementary motor areas, and within the cerebellum. An emotional component representing sexual pleasure and hedonic qualities of sexual arousal as a primary reward is suggested to be mediated by neural activations of the amygdala, the insula, and primary and secondary somatosensory cortices. Neural processes comprising goal-directed behavior and the perceived urge to express overt sexual behavior are represented by activations within the anterior cingulate cortex (ACC), the claustrum, the posterior parietal cortex, the hypothalamus, the substantia nigra, and the ventral striatum. The autonomic/neuroendocrine component is thought to be mediated by activations within the ACC, the anterior insula, the putamen, and hypothalamus, and is supposed to lead subjects to a state of physiological readiness for sexual behavior [2,3,11].

A more recent meta-analysis [12] distinguished brain networks underlying psychosexual and physiosexual arousal. Hereby, the psychosexual network was suggested to include the lateral prefrontal cortex and the hippocampus (cognitive and memory-guided evaluations), the occipitotemporal cortex, superior parietal lobules (sensory processing), the amygdala and the thalamus (relevance detection and affective evaluation), the hypothalamus (autonomic responses), basal ganglia (sexual urge), and the anterior insula (awareness of sexual arousal). Physiosexual processes were conceptualized within a network comprising the subgenual anterior cingulate cortex (sgACC; autonomic and corresponding emotion regulation), the anterior midcingulate cortex (aMCC; initiation of copulatory behavior), the putamen and claustrum (sexual urge), the anterior insula (awareness of rising sexual desire and engendered bodily reactions), the insular cortex (somatosensory information), and the operculum (monitoring bodily changes during sexual arousal). Of note, the putamen and the claustrum were identified as brain regions that connect both psychosexual und physiosexual networks, with potentially dissociable functions. While the putamen is thought to orchestrate the integration of sensorimotor information in the context of sexual desire, the putamen might be responsible for cross-modal processing between and within the networks of sexual arousal.

Most of these studies summarized neural sexual responses that were investigated using visual stimuli. However, slightly divergent patterns of brain activations were reported due to different stimulus content (e.g., sexual intensity), presentation mode (visual static images versus dynamic video sequences), or design type (block versus event-related design) [13]. While sexual motivation and wanting is reliably induced by visual sexual stimulation, genital stimulation is usually required to enter the consummatory plateau [4] of the sexual response cycle. Indeed, the use of other types of stimulus material (e.g., haptic or acoustic) is limited by the circumstances of neuroimaging methods, like noise and motion sensitivity. A few studies simultaneously recorded fMRI blood oxygenation level dependent (BOLD) signals elicited by visual stimuli and the corresponding time course of penile tumescence to investigate neural substrates of orgasm and erection. Neural activations within the ACC, the insula, amygdala, hypothalamus, and secondary somatosensory cortices were considered to be associated with penile erection [14,15]. Neural activations in mid-anterior and medial subregions of the OFC were suggested to relate specifically to orgasm [16]. Only few neuroimaging studies investigated sexual inhibition/refraction; however, these were mainly in subjects with low sexual desire. These studies indicate that sexual inhibition is mediated by prefrontal hyperactivity [17,18]. Accordingly, volitional inhibition of sexual arousal in healthy subjects was indeed accompanied by increased activations within the superior parietal, the ventrolateral prefrontal [19], and the inferior frontal cortex [20]. Moreover, it was suggested that both intended and unintended sexual inhibition are related to an exaggerated activity within the neural network of sexual interest that may, however, prevent a shift to the neural sexual consummation network [4]. Investigating neural responses a few minutes after ejaculation, one fMRI study linked activation of the amygdala, the temporal lobes, and the septal area specifically to sexual satiety [21].

#### **3. Neuromodulation of Sexual Responses**

Despite these valuable insights arising from neuroimaging studies into potentially underlying neural correlates of sexual responses, modulatory effects of neurotransmitter systems or monoaminergic drugs like antidepressants on these neural substrates are largely unknown. Most of the evidence regarding the neuromodulation of sexual functions stems from animal studies (e.g., References [22,23]) or clinical observations in patients during the treatment with psychoactive drugs (e.g., References [7,24–26]). Understanding the underlying mechanisms is indeed of great relevance considering the high prevalence of psychopharmacologically related sexual dysfunction, quite likely arising from central nervous rather than peripheral mechanisms [27].

Apart from sexual hormones and neuropeptides, central monoamines and catecholamines that are commonly modulated by psychopharmacological agents exert a pivotal role in the neuromodulation of sexual behavior. Here, we concentrate on dopamine, serotonin, and noradrenaline as the most commonly altered neuromodulator systems in psychopharmacotherapy. While an elevated central dopaminergic neurotransmission was observed to be accompanied by increased sexual interest, serotonergic agents are associated with an opposite pattern of behavior [27–29]. Considering the overlap of behavioral and neurofunctional principles of sexual functioning with other primary rewards [10], the favorable effects of dopamine on sexual behavior seem plausible. Accordingly, the antidepressant and selective noradrenaline and dopamine reuptake inhibitor (SSNDRI) bupropion is associated with subjectively improved sexual functioning, such as the ability to achieve and maintain an erection and orgasm, along with increased sexual satisfaction [30]. Moreover, dopamine-agonist treatment in Parkinson's disease is frequently accompanied by the clinical observation of hypersexuality [31]. In contrast, up to 70% of patients with schizophrenia report sexual dysfunction under treatment with antidopaminergic antipsychotics like haloperidol [32]. Apart from hyperprolactinemia due to dopamine D2-receptor blockage in the tuberoinfundibular pathway [33], the inhibitory effects of dopamine antagonists on the mesolimbic/mesocortical reward system are considered as a crucial mechanism underlying antipsychotic related sexual dysfunction [34].

The considerable impact of the neuromodulator serotonin in mediating sexual activity was recognized by the rising prevalence of sexual dysfunction during antidepressant medication, in particular with selective serotonin reuptake inhibitors (SSRIs) [35]. Although the stimulation of some specific serotonin receptor subtypes, e.g., 5-HT2c- or 5-HT1A-receptors, may facilitate erection or ejaculation, primary central serotonergic effects are thought to be inhibitory. These effects are presumably mediated via decreased dopamine release in mesolimbic regions [28,36] and by suppressing spinal ejaculatory centers [37]. Accordingly, up to 80% of patients treated with the SSRI sertraline report sexual dysfunction and, in particular in young patients, antidepressant-related decrease in sexual function is one of the most relevant side effects [24,38]. Apart from the immediate negative impact on the quality of life [39,40], antidepressant-related sexual dysfunction is also one of the major reasons that lead to non-adherence to treatment [41], especially after remission of depressive symptoms. Since early discontinuation compared to the recommended maintenance therapy over several months is related to increased rates of relapse [42], the side effect compromises the overall success of antidepressant treatment.

Compared to serotonin, the contribution of the neuromodulator noradrenaline in mediating sexual responses is less well understood. Clinical observations assume a favorable effect of selective noradrenaline reuptake inhibitors (SNRIs) on sexual functions compared to SSRIs based on lower rates of sexual dysfunction under SNRIs [24,43,44]. In line with this, actual sexual activity is related with an increase in plasma noradrenaline levels during orgasm with a subsequent rapid decline [45]. However, the limited available data regarding the effects of SNRIs on sexual functioning compromise definite conclusions [24].

#### **4. Challenges**

The conclusions regarding the effects of monoaminergic psychopharmaceuticals on sexual functions are mainly based on clinical observations and may be confounded by the disease itself. Most studies did not assess baseline sexual function before the initiation of medication. However, up to 75% of patients with major depression report sexual dysfunction prior to antidepressant treatment, in particular decreased sexual interest [46,47]. Thus, the mechanisms related to sexual dysfunction under monoaminergic agents have also to be investigated in healthy subjects to exclude confounds by the disease itself. Moreover, to meet clinical conditions as much as possible, but also to reach steady-state conditions, multi-dose trials over several days rather than single-dose applications are required to investigate neural correlates of sexual responses under antidepressants. Another limitation often arises from the study design, especially when two agents are compared with each other or relative to placebo in two different study groups. These study designs limit the capability to differentiate effects of group from those of medication, even when randomization was applied to minimize between-group effects. Also, between-group designs usually require larger sample sizes to reduce putative and systematic effects of group. Thus, apart from placebo-controlled investigations in

healthy subjects under subchronic administration of study medication, repeated measures within one group (within-subject and cross-over) may represent the most desirable study design to investigate psychopharmacological effects on neural responses of sexual behavior.

#### **5. Serotonergic, Dopaminergic, and Noradrenergic Neuromodulation of Sexual Responses**

One of the first studies investigating sexual dysfunction under monoaminergic agents and underlying neural correlates was conducted in 2009 [48] in male patients with major depression. Neural activations under visual erotic stimulation in nine patients taking SSRIs (six took paroxetine and three fluoxetine) and in 10 patients taking mirtazapine, which blocks central adrenergic and serotonin receptors, were compared to 10 healthy controls. This study demonstrated decreased neural activation within the ACC, the OFC, the insula, and the caudate nucleus in the SSRI-group compared to controls. These brain regions with attenuated responses were related to attentional and motivational components of the sexual response cycle. Neural activations in the group treated with mirtazapine were relatively lower than in controls but still elevated compared to those treated with SSRIs. Sexual dysfunction as assessed by questionnaires was significantly more frequent in depressed patients compared to controls, but did not differ between the two treatment groups. This study provided first evidence for the potential underlying neural correlates of sexual dysfunction in depression while under antidepressant treatment. However, the study design was not in the position to distinguish effects of disease from treatment-related effects on sexual functions.

We, therefore, investigated a sample of 18 healthy heterosexual males using fMRI and a randomized placebo-controlled within-subject cross-over study design. Participants were investigated after subchronic administration of the SSRI paroxetine, the SSNDRI bupropion, and placebo. Each treatment was applied for seven days separated by a wash-out time of at least 14 days [49]. During fMRI, we used a dynamic visual erotic stimulus paradigm consisting of erotic and non-erotic video clips. Erotic video clips depicted sexual interactions between one man or two women (petting, oral sex, and vaginal intercourse) extracted from commercial adult films. Non-erotic video clips showed men and women in emotionally neutral interactions. Subjective behavioral changes in sexual interest, sexual arousal, the ability to achieve orgasm, the ability to achieve and maintain an erection, and overall sexual satisfaction during drug administration were assessed by the Massachusetts General Hospital Sexual Functioning Questionnaire (MGH-SFQ) [50]. We demonstrated significantly attenuated neural activations within the sgACC, the pgACC, the aMCC, the pMCC, the nucleus accumbens, the midbrain, and the amygdala under the SSRI during visual erotic stimulation. In line with these neural alterations under the SSRI, we found a decrease in subjective sexual functions under paroxetine compared to placebo. In particular, we observed a significant decrease in subjective sexual arousal and the ability to achieve an orgasm under the SSRI compared to placebo.

Neural activations within the anterior but also rather rostral subdivisions of the ACC were previously found to be modulated by SSRIs during emotional aversive stimuli [51]. Within the context of sexual behaviour, neural activations within the ACC are associated with autonomic components of sexual responses [2,3]. Moreover, neural activity within the pgACC is related to the interaction of subjective sexual intensity and its hedonic and emotional value [52]. The results, therefore, suggested an altered neural reactivity within brain regions linked to autonomic and emotional components of sexual responses under SSRIs. In particular, attenuated neural activations within the pMCC under the SSRI were correlated with paroxetine blood-serum levels and with detrimental overall subjective functions under this antidepressant. In addition, by demonstrating attenuated neural activations within the nucleus accumbens under the SSRI, we found evidence for a diminished neural motivational component of sexual responses. This attenuation may relate to the close interaction and opposing effects between dopaminergic and serotonergic systems [53–55]. Increasing levels of serotonin as seen under SSRIs seem to dampen the functioning of the dopaminergic reward system [54,56]. To further examine whether the SSRI-related attenuation of the dopaminergic reward system and, in particular, within the nucleus accumbens might be mediated by other brain regions as observed in secondary rewards [57], we applied a psychophysiological interaction approach [58]. Indeed, we observed a significantly elevated negative reciprocal interaction between the anteroventral prefrontal cortex (avPFC) and the nucleus accumbens under the SSRI that was also associated with impulsivity as a personality trait. Thus, an increase in PFC activation may mediate the dampening effects of SSRIs on the human reward system and associated functions, e.g., sexual satisfaction.

In line with the opposing effects of serotonin and dopamine on reward-related functions and neural activity, we observed slightly enhanced and prolonged neural activations within the pMCC and within subcortical regions such as the midbrain, the amygdala, and the thalamus under the SSNDRI bupropion compared to placebo. Subjective sexual functions were indeed unimpaired under this agent in accordance to clinical studies, suggesting bupropion as a treatment alternative in patients with SSRI-related sexual dysfunction [59]. The dopaminergic agents also reveal favourable effects on sexual functions as compared to SSRIs [59–62]. The elevated neural activation pattern as found in our study and, in particular, within the ventral striatum and the midbrain as dopaminergic reward-related brain regions may represent a neural correlate of increased responsiveness to sexual stimuli arising from the dopaminergic properties of bupropion. Moreover, with concomitant activations within the amygdala that were previously related to perceived sexual arousal and to orgasmic pleasure [63], and neural activations within the thalamus and cortical regions such as the MCC, we observed activations within a neural network referred to as the salience network, which integrates homeostatic autonomic functions, emotion, and reward processing [64] (see Figure 1).

**Figure 1.** Implications of subchronic steady state serotonergic, noradrenergic, and dopaminergic stimulation on subjective sexual functions and neural responses to erotic stimulation in healthy subjects. For further information on the neurophenomenological model of sexual arousal, see Stoléru et al. [2]. SgACC = subgenual anterior cingulate cortex, pgACC = pregenual anterior cingulate cortex, amy = amygdala.

Apart from these diverging effects of the neuromodulators serotonin and dopamine, a unidirectional neural activation was found under both the SSRI and the SSNDRI within the aMCC and, thus, in a brain region associated with attentional top-down control [65]. However, the video-clip task limited the specific investigation of attentional components of sexual responses. We, therefore, investigated the same sample of 18 healthy male subjects with fMRI under the two antidepressants paroxetine and bupropion compared to placebo. During fMRI, we now used an established visual

erotic picture task [52,66] consisting of erotic and non-erotic pictures of positive emotional content taken from the International Affective Picture System (IAPS) [67]. Of note, half of the stimuli of each condition (erotic, non-erotic) were announced. The implementation of these anticipatory periods allowed the reliable investigation of attentional processes [68]. In general, anticipation is regarded as preceding attention to an upcoming predicted stimulus [69,70] and numerous studies showed neural parallels between anticipatory and attentional processes [68,71]. Under both serotonergic and dopaminergic antidepressants, we revealed attenuated neural activations within the fronto-parietal and cingulo-opercular neural network, essential for task initiation and adjustment, as well as for the maintenance of attention [65]. Accordingly, these network alterations were accompanied by unidirectional detrimental effects on the behavioral level under both agents in terms of prolonged reactions in a divided attention task.

Beneficial effects of increasing dopaminergic neurotransmission on attention and prefrontal cortical functions were conceptualized as an inverted u-shaped curve [72], whereby either too low or too high levels of dopamine [73] led to a worsening of prefrontal cortex functioning. Thus, one may argue that an increase in dopaminergic neurotransmission in healthy subjects as induced by the SSNDRI bupropion may have increased the responsivity of the neural attention network beyond the optimum and led to detrimental attentional functioning on a behavioral level. A similar response pattern was shown for increases in noradrenergic neurotransmission and other cognitive functions such as error monitoring [74]. In addition, it is of note that an increase in dopaminergic neurotransmission in prefrontal regions is not only described for bupropion but also for paroxetine via indirect pathways [75,76], supporting our observation regarding similar attention network alterations. In line with this, detrimental sustained attention was also found in other studies under SSRI administration in healthy subjects [77,78].

Apart from the serotonergic and dopaminergic antidepressants, we further investigated neural effects of noradrenergic antidepressants. Within a randomized placebo-controlled within-subject crossover design, 19 healthy heterosexual male subjects were investigated after subchronic administration of the selective noradrenaline reuptake inhibitor (SNRI) reboxetine and the second-generation antipsychotic amisulpride. During fMRI, we again used the dynamic erotic video-clip task. Noradrenergic agents and, in particular, reboxetine were thought to exert less detrimental effects on sexual functioning compared to serotonergic agents [24,79]. However, this assumption was mainly derived from investigations in depressive patients that demonstrated greater improvement in sexual satisfaction, in the ability to become sexually excited [80], and in achieving orgasm [43] under reboxetine. In contrast to these beneficial effects on sexual functions, we observed a significant decrease in overall subjective sexual function under the noradrenergic agent reboxetine compared to placebo and amisulpride in healthy subjects. In particular sexual arousal, the ability to achieve orgasm and penile erection [81] decreased. These results were, however, in line with other previous clinical reports of prolonged orgasm [82], erectile dysfunction [83], and anorgasmia [43] under this drug. On the neural level, we revealed diminished neural activations within the caudate nucleus under reboxetine compared to placebo that were significantly associated with the decreased sexual interest under this agent. With regard to erotic stimulation, caudate nucleus activation was linked to goal-directed behavior and reward [84]. Whereas ventral parts of the striatum are commonly associated with the expectation and the receipt of incentives, dorsal striatal/caudate nucleus activation was associated with motivational rather than reward processing [85]. Thus, our findings may support the notion that an increase in noradrenergic neurotransmission might have detrimental effects on motivational components of sexual responses along with diminished subjective sexual functioning.

It is of note that we did not find significant neural alterations during visual erotic stimulation and in subjective sexual functions under the antipsychotic drug amisulpride compared to placebo. The antipsychotic drug amisulpride has high and selective affinity to postsynaptic D2- und D3-receptors [86–88] and it is known for its capacity to induce sexual dysfunction mainly due to the blockage of dopamine D2-receptors [33] with secondary increases of prolactin levels [34]. The lack of significant alterations in neural visual erotic stimulus processing along with unchanged subjective

sexual functions in our study was most likely due to the low dosage of 200 mg/day amisulpride for seven days. Antipsychotic effects of amisulpride were reported for high dosages from about 400 to 600 mg/day due to reliable D2-receptor occupancy [86]. In contrast, lower dosages (50 to 200 mg/day) as used in our study are thought to primarily block presynaptic dopamine autoreceptors with the consequence of mild pro-dopaminergic effects [86,89,90] that may have left sexual functions and corresponding neural correlates unimpaired in our sample of healthy male subjects.

To further investigate neural responses to visual erotic stimulus processing including preceding attention and their modulation by noradrenergic agents, we also applied the abovementioned erotic picture paradigm with anticipatory periods [91]. Notably, upon static rather than previously applied dynamic visual erotic stimulation, we observed additional treatment effects of the noradrenergic agent reboxetine compared to placebo during visual erotic stimulation by diminished neural activations not only within the caudate nucleus, but also within the ventral striatum/nucleus accumbens, the pgACC, the aMCC, and the OFC. In addition, decreases in subjective sexual arousal correlated with attenuated neural activations within the posterior insula, a region that is repeatedly associated with sexual arousal and penile response [14,84]. Thus, our results support the notion of detrimental effects of noradrenergic agents on emotional, motivational, and autonomic neural components of sexual responses, along with decreased subjective sexual function (see Figure 1). In addition, they also underpin the implication regarding stimulus presentation mode in investigating neural substrates of erotic stimulus processing considering that treatment effects of noradrenergic agents were found within a broader neural network during static rather than dynamic visual erotic stimulation.

Similar to the investigation by erotic video stimulation, we also found no significant neural alteration under amisulpride compared to placebo. Moreover, in contrast to serotonergic and predominantly dopaminergic antidepressants, neither the noradrenergic agent reboxetine nor the antipsychotic amisulpride led to neural alterations during the anticipation of erotic stimuli, in line with unimpaired attentional functions on a behavioral level in this study. However, it is of note that major nodes of the neural network altered by the noradrenergic agent reboxetine compared to placebo such as the ventral striatum, the pgACC, aMCC, and the OFC highly resemble those brain regions that were also modulated by serotonergic agents upon erotic video stimulation in our previous investigation.

While it remains speculative, one may argue that either monoaminergic modulation ends up via similar neural pathways and presumably also on a molecular level. Interactions of both the serotonergic and noradrenergic system with dopaminergic projections were extensively studied [54,92], and a modulation of one system will invariably influence the transmission of the other. Here, the human reward system may represent a major or common final pathway. The specific increase in serotonergic and noradrenergic turnover under paroxetine and reboxetine, respectively, dampened the neural activity within the dopaminergic human reward system and, in particular, within the nucleus accumbens. However, this attenuation is potentially restricted to the processing of specific rewards or reinforcers such as sexual stimuli or primary rewards, considering that the serotonergic and noradrenergic attenuation of neural activity within the nucleus accumbens was not evident when processing monetary rewards as secondary reinforcers [93,94].

#### **6. Perspectives**

Our project using pharmacological and task-based fMRI identified neuromodulatory effects of monoamines and catecholamines on neural sexual responses and potential neural proxies for the development of sexual dysfunction under antidepressants. Insights from this methodological approach mainly concern basic research; however, some aspects might be transferred to clinical practices. Considering that task-based fMRI may not be easily implemented in clinical routines due to its complexity and dependency on a subject's motivation and performance, resting-state fMRI may provide a valuable alternative. Accordingly, we investigated healthy subjects using resting-state fMRI [95] and demonstrated that more impaired subjective sexual function under serotonergic agents was predicted by low baseline functional connectivities under placebo. In particular, functional connectivities of the sublenticular extended amygdala with midbrain, pgACC, and the insula revealed a predictive potential for the development of SSRI-related decreases in sexual functioning. Although these results await empirical replication in larger samples, they may support the idea of a potentially valuable contribution of imaging techniques in the prediction of pharmaco-related sexual dysfunction within the context of personalized medicine.

It is of note that the investigations presented were exclusively conducted in healthy male subjects and the conclusions drawn may not be transferable to females. Within the past years, gender and sex aspects were widely recognized in scientific research and, with regard to sexual responses, sex differences are proposed to not only occur on behavioral and downstream peripheral, but also on the neural level. Relative to men, meta-analyses suggest a less consistent and decreased neurofunctional activation in subcortical regions in women during sexual arousal [96,97]. In addition, female sex hormones appear to play a crucial role in mediating particularly cortical activations in response to sexual stimulation [97–99]. Moreover, there is evidence for an interaction between sex hormones and the dominant neurotransmitters such as serotonin and dopamine [100]. These observations suggest a divergent monoaminergic neuromodulation of erotic stimulus processing in females. Consequently, the investigation of females under different levels of monoaminergic or catecholaminergic neurotransmitter levels in combination with different hormonal states is highly encouraged as a future research topic.

#### **7. Conclusions**

Within a broader research program, we investigated healthy male subjects under visual erotic stimulation by fMRI and different antidepressant medication to disentangle effects of monoaminergic and catecholaminergic neuromodulatory substances on neural substrates of sexual responses. After increasing serotonergic neurotransmission, we observed attenuated neural activations within cerebral networks previously related to motivational, emotional, and autonomic components of sexual behavior along with diminished subjective sexual functions. Psychophysiological interaction analyses revealed that the dampening of the motivational component and, in particular, human reward system activation was presumably mediated by an increase in prefrontal cortex activation as a potential correlate of increased cognitive control under serotonergic agents. Of note, neural motivational and emotional components, as well as subjective sexual functions, were either unaffected or even increased under dopaminergic stimulation. Apart from these divergent effects on erotic stimulus processing, both serotonergic and dopaminergic stimulation diminished neural attention network activation during the anticipation of visual sexual stimuli, along with a decrease in behavioral measures of attention. Investigating the noradrenergic neuromodulation of neural substrates of erotic stimulus processing revealed similar neural alterations as serotonergic agents, and showed again attenuation of neural emotional and motivation components along with a decrease in subjective sexual functions. However, neural activations during the anticipation of sexual stimuli and behavioral attentional functioning were not altered by a noradrenergic agent. Thus, our results provided evidence for the neuromodulatory effects of serotonergic, noradrenergic, and dopaminergic agents on neural substrates of erotic stimulus processing. Considering the overlay of neuromodulatory effects of serotonergic and noradrenergic neurotransmission, this may suggest that both monoaminergic modulations end up via similar neural pathways and presumably affect dopaminergic projections within the human reward system. Notably, the dampening of the human reward system by both serotonergic and noradrenergic agents was, however, restricted to the processing of visual sexual stimuli as primary reinforcers and was not evident during processing of monetary rewards as secondary reinforcers.

From a basic research perspective, we demonstrate that modulations in sexual functioning on the subjective behavioral level are indeed closely linked to cerebral networks that mediate motivational, emotional, autonomic, and attentional components of the sexual response. Our data emphasize the hypothesis that altered cerebral reactivity rather than peripheral effects might be the key to explain side effects of monoaminergic substances on sexual functioning.

**Author Contributions:** Conceptualization, H.G., M.W., G.G. and B.A.; writing—original draft preparation, H.G., G.G. and B.A.; writing—review and editing, H.G., K.M., C.D.M, G.G. and B.A.; visualization, H.G.; supervision, H.G., G.G. and B.A.; project administration, H.G.

**Acknowledgments:** We thank C. Hiemke and his stuff at the University of Mainz, Germany, Department of Psychiatry and Psychotherapy, for measuring drug serum levels in our research project.

**Conflicts of Interest:** The authors declare no potential (including financial) conflicts of interests related to this work.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

### *Review* **Online Porn Addiction: What We Know and What We Don't—A Systematic Review**

#### **Rubén de Alarcón 1, Javier I. de la Iglesia 1, Nerea M. Casado <sup>1</sup> and Angel L. Montejo 1,2,\***


Received: 27 November 2018; Accepted: 10 January 2019; Published: 15 January 2019

**Abstract:** In the last few years, there has been a wave of articles related to behavioral addictions; some of them have a focus on online pornography addiction. However, despite all efforts, we are still unable to profile when engaging in this behavior becomes pathological. Common problems include: sample bias, the search for diagnostic instrumentals, opposing approximations to the matter, and the fact that this entity may be encompassed inside a greater pathology (i.e., sex addiction) that may present itself with very diverse symptomatology. Behavioral addictions form a largely unexplored field of study, and usually exhibit a problematic consumption model: loss of control, impairment, and risky use. Hypersexual disorder fits this model and may be composed of several sexual behaviors, like problematic use of online pornography (POPU). Online pornography use is on the rise, with a potential for addiction considering the "triple A" influence (accessibility, affordability, anonymity). This problematic use might have adverse effects in sexual development and sexual functioning, especially among the young population. We aim to gather existing knowledge on problematic online pornography use as a pathological entity. Here we try to summarize what we know about this entity and outline some areas worthy of further research.

**Keywords:** online pornography; addiction; cybersex; internet; compulsive sexual behavior; hypersexuality

#### **1. Introduction**

With the inclusion of "Gambling Disorder" in the "Substance Use and Addictive Disorders" chapter of the DSM-5 [1], the APA publicly acknowledged the phenomenon of behavioral addiction. Furthermore, "Internet Gaming Disorder" was placed in Section 3—conditions for further study.

This represents the ongoing paradigm shift in the field of addictions that relates to addictive behavior, and paves the way for new research in the light of cultural changes caused by the new technologies.

There is apparently an existing common neurobiological [2] and environmental [3] ground between the varying addictive disorders, including both substance abuse and addictive behavior; this can manifest as an overlapping of both entities [4].

Phenomenologically, behaviorally addicted individuals frequently exhibit a problematic consumption model: impaired control (e.g., craving, unsuccessful attempts to reduce the behavior), impairment (e.g., narrowing of interests, neglect of other areas of life), and risky use (persisting intake despite awareness of damaging psychological effects). Whether these behaviors also meet physiological criteria relating to addiction (tolerance, withdrawal) is more debatable [4–6].

Hypersexual disorder is sometimes considered one of those behavioral addictions. It is used as an umbrella construct that encompasses various problematic behaviors (excessive masturbation, cybersex, pornography use, telephone sex, sexual behavior with consenting adults, strip club visitations, etc.) [7]. Its prevalence rates range from 3% to 6%, though it is difficult to determine since there is not a formal definition of the disorder [8,9].

The lack of robust scientific data makes its research, conceptualization, and assessment difficult, leading to a variety of proposals to explain it, but is usually associated with significant distress, feelings of shame and psychosocial dysfunction [8], as well as other addictive behaviors [10] and it warrants direct examination.

Concurrently, the rise of the new technologies has also opened up a pool of problematic addictive behavior, mainly Internet Addiction. This addiction may focus on a specific application on the internet (gaming, shopping, betting, cybersex . . . ) [11] with potential for risk-addictive behavior; in this case, it would act as a channel for concrete manifestations of said behavior [4,12]. This means inevitable escalation, providing new outlets for established addicts as well as tempting people (due to increased privacy, or opportunity) who would not have previously engaged in these behaviors.

Online pornography use, also known as Internet pornography use or cybersex, may be one of those Internet-specific behaviors with a risk for addiction. It corresponds to the use of Internet to engage in various gratifying sexual activities [13], among which stands the use of pornography [13,14] which is the most popular activity [15–17] with an infinite number of sexual scenarios accessible [13,18–20]. Continued use in this fashion sometimes derives in financial, legal, occupational, and relationship trouble [6,21] or personal problems, with diverse negative consequences. Feelings of loss of control and persistent use despite these adverse results constitute "online sexual compulsivity" [22] or Problematic Online Pornography Use (POPU). This problematic consumption model benefits from the "Triple A" factors [23].

Due to this model, pornography-related masturbation may be more frequent nowadays, but this is not necessarily a sign of pathology [21]. We know that a considerable proportion of young male population access Internet for pornography consumption [24,25]; in fact, it is one of their key sources for sexual health [26]. Some have expressed concern about this, addressing the time gap between when porn material is consumed for the first time ever, and an actual first sexual experience; specifically, how the former can have an impact on sexual development [27] like abnormally low sexual desire when consuming online pornography [28] and erectile dysfunction, which has spiked dramatically among young men in the past few years when compared to a couple decades ago [29–33].

We systematically reviewed the existing literature on the subject of POPU to try and summarize the various recent advances made in terms of epidemiology, clinical manifestations, neurobiological evidence that supports this model of problematic use, its diagnostic conceptualization in relation to hypersexual disorder, its proposed assessment instruments and treatment strategies.

#### **2. Methods**

We performed the systematic review following PRISMA guidelines (Figure 1). Given the relatively new body of evidence regarding this subject, we conducted our review with no specific time-delimitation. Priority was placed upon literature reviews and articles published via a newest to oldest methodology, preferentially for already published reviews on the subject. PubMed and Cochrane were the main databases used, though a number of articles were compiled through cross-referencing.

**Figure 1.** PRISMA flow diagram.

Since our focus was mainly online pornography and addictive sexual behavior, we excluded those articles that had only a peripheral association with it in our search: those with a focus on generalized Internet addiction, those centered on the pornographic equivalent of varying paraphilias, and those that approached the subject from a social perspective.

The following search terms and their derivatives were used in multiple combinations: cybersex, porn\* (to allow for both "pornography" and "pornographic"), addict\* (to allow for both "addiction" and "addictive"), online, internet, sex, compulsive sex, hypersexuality. The reference management tool Zotero was used to build a database of all articles considered.

#### **3. Results**

#### *3.1. Epidemiology*

Pornography consumption in the general population proves difficult to be adequately measured, especially since the rise of the Internet and the "triple A" factors which have allowed for both privacy and ease of access. Wright's study about the use of pornography in U.S. male population using the General Social Survey (GSS) [34], and Price's study (which expands upon Wright's by distinguishing among age, cohort, and period effects) [35] constitute some of the few, if not the only ones, existing sources that track pornography use in the general population. They show the overall increasing consumption of pornography over the years, especially among male population in contrast to females. This is particularly prevalent among young adults, and it steadily decreases with age.

Some interesting facts about pornography consumption tendencies stand out. One of them is that the 1963 and 1972 male cohort showed only a very small decline on their usage from the year 1999 onwards, suggesting that porn consumption among these groups has remained relatively constant since [35]. The other one is that 1999 is also the year the tendency for women aged 18 to 26 to consume pornography became three times as likely than the ones aged 45 to 53, instead of just two times as likely as it used to be up until that point [35]. These two facts could be related to changing tendencies in pornography consumption motivated by technology (switching from the offline to the online model of consumption), but it is impossible to know for sure since the original data does not account for differences in both offline and online variants when tracking pornography usage.

As for POPU, there is no clear and reliable data in the literature reviewed that can offer a solid estimation of its prevalence. Adding up to the already mentioned motives for lack of data on general pornography consumption, part of it might stem from the perceived taboo nature of the topic at hand by possible participants, the wide range of assessment tools used by researchers, and the lack of consensus on what actually constitutes a pathological usage of pornography, which are all issues also reviewed further into this paper.

The vast majority of studies pertaining POPU or hypersexual behavior prevalence use convenience samples to measure it, usually finding, despite population differences, that very few users consider this habit an addiction, and even when they do, even fewer consider that this could have a negative effect on them. Some examples:


The only study with a representative sample to date is an Australian one, with a sample of 20,094 participants; 1.2% of the women surveyed considered themselves addicted, whereas for the men it was 4.4% [38]. Similar findings also apply to hypersexual behavior outside of pornography [39].

Predictors for problematic sexual behavior and pornography use are, across populations: being a man, young age, religiousness, frequent Internet use, negative mood states, and being prone to sexual boredom, and novelty seeking [17,37,40,41]. Some of this risk factors are also shared by hypersexual behavior patients [39,42].

#### *3.2. Ethiopathogenical and Diagnostic Conceptualization*

Conceptualizing pathological behaviors continues to be a challenge today. While several attempts have been made regarding hypersexual behavior, the lack of robust data as of now explains the fact that there's no consensus on this matter [9]. POPU comprises a very specific set of sexual behaviors that involve technology. Due to problematic technology use (especially online technology) being relatively recent, we need first to talk about hypersexual behavior not related to technology in order to understand the place of online pornography in it.

Sexuality as a behavior is vastly heterogeneous, and its potential pathological side has been studied for centuries [43]. Therefore, it represents a challenge to models trying to adequately define it, since it can incorporate practices ranging from solitary fantasizing to sexual violence [21]. It is also difficult to define what constitutes an actual dysfunction and manage to avoid the possible misuse of that definition to stigmatize and pathologize individuals [44]. For example, some set the limit between normal and pathological sexual behavior at more than seven orgasms in a week [43] (p. 381), but this approach focusing on quantity can be dangerous, since what constitutes normal and pathological behavior can vastly vary between individuals. This lack of uniformity and consistency in its classification may hinder future research on investigating hypersexual behavior [45] and ignore the quality aspects that focus on the negative emotions associated with it [46,47]. There have been proposals to redeem this issue using certain tools, already developed as part of the hypersexual disorder proposal used in the DSM-5 field trial [43,47].

Hypersexuality generally acts as an umbrella construct [7]. Its nomenclature is still a matter of debate to this day, and it is frequent to encounter several terms that refer to the same concept: compulsive sexual behavior, sex addiction, sexual impulsivity, hypersexual behavior or hypersexual disorder. Some authors, while recognizing the value of the terms "addiction" and "compulsivity", prefer to draw attention to the issue of control and its possible loss or compromise as the primary concern about this behavior, thus referring to it as "out of control sexual behavior" [45,48,49].

Although definitions are not uniform, they usually focus on the frequency or intensity of symptoms [46] of otherwise normal urges and fantasies, that would result in dysfunction. This differentiates it from paraphilic sexual behavior, though the need for a better clarification of possible differences, similarities, and overlap between the two types still persists [45].

Usually included in hypersexual behavior are excessive masturbation and various sexual related behaviors, like dependence on anonymous sexual encounters, repetitive promiscuity, internet pornography, telephone sex, and visiting strip clubs [43,44,49–51]. Bancroft particularly thought that, in using Internet, both masturbation and these sexual activities could blend themselves, stating that men "use it as an almost limitless extension of their out of control masturbatory behavior".

While the possibility to diagnose hypersexual behavior was always available with "sexual disorder not otherwise specified" in the DSM [1], Kafka [43] tried to propose it as a diagnostic entity for the DSM-5. He presented a set of criteria for it, as part of the sexual disorders chapter. These proposed models included hypersexual behavior as: (1) sexually motivated, (2) a behavioral addiction, (3) part of the obsessive-compulsive spectrum disorder, (4) part of the impulsivity-spectrum disorders, and (5) an "out of control" excessive sexual behavior. This proposal was ultimately rejected due to several reasons; the main was said to be absence of consolidated epidemiological and neuroimaging data regarding this behavior [52,53], but also its potential for forensic abuse, a not specific enough set of diagnostic criteria, and potential politic and social ramifications of pathologizing an integral area of behavior to human life [54]. It is interesting to compare it to the other two previous set of criteria present in the reviewed literature, those of Patrick Carnes and Aviel Goodman [9]. All three share the concepts of loss of control, excessive time spent on sexual behavior and negative consequences to self/others, but diverge on the other elements. This reflects in broad strokes the lack of consensus in conceptualizing hypersexual behavior across the years. Currently, the main options propose hypersexual behavior either as an impulse control disorder or a behavioral addiction [55].

From an impulse control disorder perspective, hypersexual behavior is generally referred to as Compulsive Sexual Behavior (CSB). Coleman [56] is a proponent of this theory. While he includes paraphilic behavior under this term [57], and they may coexist in some cases, he distinctly differentiates it from nonparaphilic CSB, which is what we want to focus on in this review. Interestingly, nonparaphilic hypersexual behavior is usually as frequent, if not more, than some paraphilias [43,58].

However, more recent definitions of CSB usually refer to multiple sexual behaviors that can be compulsive: the most commonly reported being masturbation, being followed by compulsive use of pornography, and promiscuity, compulsive cruising, and multiple relationships (22–76%) [9,59,60].

While there are definite overlaps between hypersexuality and conditions such as obsessive-compulsive disorder (OCD) and other impulse control disorders [61], there are also some notable differences pointed out: for example, OCD behaviors do not involve reward, unlike sexual behavior. Moreover, while engaging in compulsions might result in temporary relief for OCD patients [62], hypersexual behavior is usually associated by guilt and regret after committing the act [63]. Also, the impulsivity that can sometimes dominate the patient's behavior is incompatible with the careful planning that is sometimes required in CSB (for example, in regards to a sexual encounter) [64]. Goodman thinks that addiction disorders lie at the intersection of compulsive disorders (which involve anxiety reduction) and impulsive disorders (which involve gratification), with the symptoms being underpinned by neurobiological mechanisms (serotoninergic, dopaminergic, noradrenergic, and opioid systems) [65]. Stein agrees with a model combining several ethiopathogenical mechanisms and proposes an A-B-C model (affective dysregulation, behavioral addiction, and cognitive dyscontrol) to study this entity [61].

From an addictive behavior standpoint, hypersexual behavior relies on sharing core aspects of addiction. These aspects, according to the DSM-5 [1], refer to the mentioned problematic consumption model applied to hypersexual behavior, both offline and online [6,66,67]. Evidence of tolerance and withdrawal in these patients might probably be key in characterizing this entity as an addictive disorder [45]. Problematic use of cybersex is also often conceptualized as a behavioral addiction [13,68].

The term "addiction" applying to this entity is still subject to great debate. Zitzman considers that the resistance to use the term addiction is "more a reflection of cultural sexual liberality and permissiveness than any lack of symptomatic and diagnostic correspondence with other forms of addiction" [69]. However, the term needs to be used with caution, since it can be interpreted as a justification for an irresponsible search for gratification and hedonist pleasure, and blame the disruptive consequences on it.

There has long been a debate between Patrick Carnes and Eli Coleman over the diagnostics of hypersexual behavior. Coleman has considered hypersexuality to be driven by the need to reduce some type of anxiety, not by sexual desire [56] having classified it in seven subtypes (one of them being use of online pornography) [57], while Carnes (who defined addiction as "a pathological relationship with a mood altering experience") finds similitudes to other behavioral addictions like gambling, focusing on the loss of control and continued behavior despite negative consequences [70].

A thorough review of the literature by Kraus [71], concluded that despite these similitudes, significant gaps in the concept's understanding complicate its classification as an addiction. The main concerns are aimed towards quantity of large-scale prevalence, longitudinal and clinical data (defining main symptoms and its diagnostic limits), supported by neuropsychological, neurobiological, and genetic data, as well as some information regarding possible treatment screening and prevention, and points to digital technology in hypersexual behavior as a key point for future research.

The rise of the Internet increases the possibilities for sexual interactions, and not just online pornography (webcamming, casual sex websites). Even whether Internet use represents a conduit for other types of repetitive behavior (e.g., sexual behavior or gambling) or constitutes a different entity in its own right is still debated [72]. Nevertheless, if the case is the former, the previous evidence and considerations could very well apply to its online counterpart.

There is currently a need for empirically derived criteria that takes into account unique factors characterizing online (versus offline) sexual behaviors, since most of them do not have an offline version that can be compared to [73]. So far, there have been mentions of new phenomena when dealing with online sexual behavior, like the presence of online dissociation [74], which causes to "be mentally and emotionally detached when engaged, with compromised time and depersonalization". This dissociation has already been described in relation to other online activities [75], which supports the notion that cybersex problematic use could be related to both internet and sex addiction [76].

Finally, we have to mention that a diagnostic entity called "compulsive sexual behavior disorder" is being included in the upcoming definitive edition of ICD-11, in the "impulse control disorders" chapter [77]. The definition can be consulted at https://icd.who.int/dev11/l-m/en#/http%3a%2f% 2fid.who.int%2ficd%2fentity%2f1630268048.

The inclusion of this category in the ICD-11 may be a response to the relevance of this issue and attest to its clinical utility, whereas the growing but yet inconclusive data prevents us from properly categorizing it as a mental health disorder [72]. It is believed to provide a better tool (yet in refinement process) for addressing the needs of treatment seeking patients and the possible guilt associated [78], and also may reflect the ongoing debates regarding the most appropriate classification of CSB and its limited amount of data in some areas [55,71] (Table 1). This inclusion could be the first step towards recognizing this issue and expanding on it, one key point being undoubtedly its online pornography subtype.



#### *3.3. Clinical Manifestations*

Clinical manifestations of POPU can be summed up in three key points:


Having some very specific pornographic content interests has been associated with an increase in reported problems [17]. It has been debated if these clinical features are the consequence of direct cybersex abuse or due to the subjects actually perceiving themselves as addicts [91].

#### *3.4. Neurobiological Evidence Supporting Addiction Model*

Collecting evidence about POPU is an arduous process; main data on this subject is still limited by small sample sizes, solely male heterosexual samples and cross-sectional designs [71], with not enough neuroimaging and neuropsychological studies [4], probably due to conceptual, financial and logistic obstacles. In addition, while substance addiction can be observed and modeled in experimental animals, we cannot do this with a candidate behavioral addiction; this may limit our study of its neurobiological underpinnings [72]. Current knowledge gaps regarding the research of hypersexual behavior, as well as possible approaches for addressing them, are expertly covered and summarized in Kraus' article [71]. Most of the studies found in our research pertain hypersexual behavior, with pornography being only one of its accounted accessories.

This evidence is based on an evolving understanding of the neural process among addiction-related neuroplasticity changes. Dopamine levels play an important part in this sexual reward stimuli, as observed already in frontotemporal dementia and pro-dopaminergic medication in Parkinson's disease being linked with sexual behavior [92,93].

The addictive process with online pornography may be amplified by the accelerated novelty and the "supranormal stimulus" (term coined by Nobel prize winner Nikolaas Tinbergen) that constitutes Internet pornography [94]. This phenomenon would supposedly make artificial stimuli (in this case, pornography in the way it is mostly consumed today, its online form) override an evolutionarily developed genetic response. The theory is that they potentially activate our natural reward system at higher levels than what ancestors typically encountered as our brains evolved, making it liable to switch into an addictive mode [2]. If we consider online porn from this perspective, we can start seeing similarities to regular substance addicts.

Major brain changes observed across substance addicts lay the groundwork for the future research of addictive behaviors [95], including:


These brain changes observed in addicts have been linked with patients with hypersexual behavior or pornography users through approximately 40 studies of different types: magnetic resonance imaging, electroencephalography (EEG), neuroendocrine, and neuropsychological.

For example, there are clear differences in brain activity between patients who have compulsive sexual behavior and controls, which mirror those of drug addicts. When exposed to sexual images, hypersexual subjects have shown differences between liking (in line with controls) and wanting (sexual desire), which was greater [8,100]. In other words, in these subjects there is more desire only for the specific sexual cue, but not generalized sexual desire. This points us to the sexual cue itself being then perceived as a reward [46].

Evidence of this neural activity signalizing desire is particularly prominent in the prefrontal cortex [101] and the amygdala [102,103], being evidence of sensitization. Activation in these brain regions is reminiscent of financial reward [104] and it may carry a similar impact. Moreover, there are higher EEG readings in these users, as well as the diminished desire for sex with a partner, but not for masturbation to pornography [105], something that reflects also on the difference in erection quality [8]. This can be considered a sign of desensitization. However, Steele's study contains several methodological flaws to consider (subject heterogeneity, a lack of screening for mental disorders or addictions, the absence of a control group, and the use of questionnaires not validated for porn use) [106]. A study by Prause [107], this time with a control group, replicated these very findings. The role of cue reactivity and craving in the development of cybersex addiction have been corroborated in heterosexual female [108] and homosexual male samples [109].

This attentional bias to sexual cues is predominant in early hypersexual individuals [110], but a repeated exposure to them shows in turn desensitization [111,112]. This means a downregulation of reward systems, possibly mediated by the greater dorsal cingulate [107,113,114]. Since the dorsal cingulate is involved in anticipating rewards and responding to new events, a decrease in its activity after repeated exposure points us to the development of habituation to previous stimuli. This results in a dysfunctional enhanced preference for sexual novelty [115], which may manifest as attempts to overcome said habituation and desensitization through the search for more (new) pornography as a means of sexual satisfaction, choosing this behavior instead of actual sex [20].

These attempts at novelty seeking may be mediated through ventral striatal reactivity [116] and the amygdala [117]. It is known that the viewing of pornography in frequent users has also been associated with greater neural activity [99], especially in the ventral striatum [116,118] which plays a major role in anticipating rewards [119].

However, connectivity between ventral striatum and prefrontal cortex is decreased [103,113]; a decrease in connectivity between prefrontal cortex and the amygdala has also been observed [117]. In addition, hypersexual subjects have shown reduced functional connectivity between caudate and temporal cortex lobes, as well as gray matter deficit in these areas [120]. All of these alterations could explain the inability to control sexual behavior impulses.

Moreover, hypersexual subjects showed an increased volume of the amygdala [117], in contrast to those with a chronic exposure to a substance, which show a decreased amygdala volume [121]; this difference could be explained by the possible neurotoxic effect of the substance. In hypersexual subjects, increased activity and volume may reflect overlapping with addiction processes (particularly supporting incentive motivation theories) or be the consequence to chronic social stress mechanisms, such as the behavioral addiction itself [122].

These users also have shown a dysfunctional stress response, mainly mediated through the hypothalamus–pituitary–adrenal axis [122] in a way that mirror those alterations seen in substance addicts. These alterations may be the result of epigenetic changes on classic inflammatory mediators driving addictions, like corticotropin-releasing-factor (CRF) [123]. This epigenetic regulation hypothesis considers both hedonic and anhedonic behavioral outcomes are at least partially affected by dopaminergic genes, and possibly other candidate neurotransmitter-related gene polymorphisms [124]. There is also evidence of higher tumor necrosis factor (TNF) in sex addicts, with a strong correlation between TNF levels and high scores in hypersexuality rating scales [125].

#### *3.5. Neuropsychological Evidence*

In regard to the manifestations of these alterations in sexual behavior, most neuropsychological studies show some kind of indirect or direct consequence in executive function [126,127], possibly as a consequence of prefrontal cortex alterations [128]. When applied to online pornography, it contributes to its development and maintenance [129,130].

The specifics of this poorer executive functioning include: impulsivity [131,132], cognitive rigidity that impedes learning processes or the ability to shift attention [120,133,134], poor judgment and decision making [130,135], interference of working memory capacity [130], deficits in emotion regulation, and excessive preoccupation with sex [136]. These findings are reminiscent of other behavioral addictions (such as pathological gambling) and the behavior in substance dependencies [137]. Some studies directly contradict these findings [58], but there may be some limitations in methodology (for example, small sample size).

Approaching the factors that play a role in the development of hypersexual behavior and cybersex, there are a number of them. We can think of cue-reactivity, positive reinforcement and associative learning [104,109,136,138,139] as the core mechanisms of porn addiction development. However, there may be factors of underlying vulnerability [140], like: (1) the role of sexual gratification and dysfunctional coping in some predisposed individuals [40,141–143] whether it is a consequence of trait impulsivity [144,145] or state impulsivity [146], and (2) approach/avoidance tendencies [147–149].

#### *3.6. Prognosis*

Most of the studies referenced use subjects with a long-term exposure to online pornography [34,81,113,114], so its clinical manifestations appear to be a direct and proportional consequence of engaging in this maladaptive behavior. We mentioned difficulty in obtaining controls to establish causation, but some case reports suggest that reducing or abandoning this behavior may cause improvement in pornography-induced sexual dysfunction and psychosexual dissatisfaction [79,80] and even full recovery; this would imply that the previously mentioned brain alterations are somewhat reversible.

#### *3.7. Assessment Tools*

Several screening instruments exist for addressing CSB and POPU. They all rely on the responder's honesty and integrity; perhaps even more so than regular psychiatry screening tests, since sexual practices are the most humbling due to their private nature.

For hypersexuality, there are over 20 screening questionnaires and clinical interviews. Some of the most notable include the Sexual Addiction Screening Test (SAST) proposed by Carnes [150], and its later revised version SAST-R [151], the Compulsive Sexual Behavior Inventory (CSBI) [152,153] and the Hypersexual Disorder Screening Inventory (HDSI) [154]. The HDSI was originally used for the clinical screening of the DSM-5 field proposal of hypersexual disorder. While further explorations of the empirical implications regarding criteria and the refinements of cutoff scores are needed, it currently holds the strongest psychometric support and is the best valid instrument in measuring hypersexual disorder [151].

As for online pornography, the most used screening tool is the Internet Sex-screening test (ISST) [155]. It assesses five distinct dimensions (online sexual compulsivity, online sexual behavior-social, online sexual behavior-isolated, online sexual spending and interest in online sexual behavior) through 25 dichotomic (yes/no) questions. However, its psychometric properties haven only been mildly analyzed, with a more robust validation in Spanish [156] that has served as a blueprint for posterior studies [157].

Other notable instruments are the problematic pornography use scale (PPUS) [158] which measures four facets of POPU (including: distress and functional problems, excessive use, control difficulties and use for escape/avoidance of negative emotions), the short internet addiction test adapted to online sexual activities (s-IAT-sex) [159], a 12-item questionnaire measuring two dimensions of POPU, and the cyber-pornography use inventory (CPUI-9) [160].

The CPUI-9 evaluates three dimensions: (1) access efforts, (2) perceived compulsivity, and (3) emotional distress. At first considered to have convincing psychometric properties [9], this inventory has more recently proved to be unreliable: the inclusion of the "emotional distress" dimension address levels of shame and guilt, which do not belong in an addiction assessment and thus skews the scores upward [161]. Applying the inventory without this dimension appears to accurately reflect to some extent compulsive pornography use.

One of the most recent is the pornographic problematic consumption scale (PPCS) [162], based on Griffith six-component addiction model [163], though it does not measure addiction, only problematic use of pornography with strong psychometric properties.

Other measures of POPU that are not designed to measure online pornography use but have been validated using online pornography users [9], include the Pornography Consumption Inventory (PCI) [164,165], the Compulsive Pornography Consumption Scale (CPCS) [166] and the Pornography Craving Questionnaire (PCQ) [167] which can assess contextual triggers among different types of pornography user.

There are also tools for assessing pornography users' readiness to abandon the behavior through self-initiated strategies [168] and an assessment of treatment outcome in doing so [169], identifying in particular three potential relapse motivations: (a) sexual arousal/boredom/opportunity, (b) intoxication/locations/easy access, and (c) negative emotions.

#### *3.8. Treatment*

Given that still many questions remain regarding the conceptualization, assessment, and causes of hypersexual behavior and POPU, there have been relatively few attempts to research possible treatment options. In published studies, sample sizes are usually small and too homogeneous, clinical controls are lacking, and the research methods are scattered, unverifiable, and not replicable [170].

Usually, combining psychosocial, cognitive–behavioral, psychodynamic, and pharmacologic methods is considered most efficient in treatment of sexual addiction, but this non-specific approach reflects the lack of knowledge about the subject [9].

#### 3.8.1. Pharmacological Approaches

The studies have centered on paroxetine and naltrexone thus far. One case series involving paroxetine on POPU helped to decrease the anxiety levels, but eventually failed to reduce the behavior by itself [171]. Additionally, using SSRIs to create sexual dysfunction through their side effects is apparently not effective, and according to clinical experience are useful only in patients with comorbid psychiatric disorders [172].

Four case reports involving naltrexone to treat POPU have been described. Previous findings have suggested that naltrexone could be a potential treatment for behavioral addictions and hypersexual disorder [173,174], theoretically reducing cravings and urges by blocking the euphoria associated with the behavior. While there is not yet a randomized controlled trial with naltrexone in these subjects, there are four case reports. Results obtained in reducing pornography use varied from good [175–177] to moderate [178]; at least in one of them the patient also received sertraline, so it is unclear how much can be attributed to naltrexone [176].

#### 3.8.2. Psychotherapeutic Approaches

Undoubtedly, psychotherapy can be an important tool in fully comprehending and changing a behavior. While cognitive-behavioral therapy (CBT) is considered by many clinicians to be useful in treating hypersexual disorder [179], a study that involved problematic online pornography users failed to achieve a reduction of the behavior [180], even if the severity of comorbid depressive symptoms and general quality of life was improved. This brings up the interest notion that merely reducing pornography use may not represent the most important treatment goal [170]. Other approaches using CBT to treat POPU have been made, but reoccurring methodological problems in this area prevent us from extracting reliable conclusions [181,182].

Psychodynamic psychotherapy and others like family therapy, couples' therapy, and psychosocial treatments modeled after 12 step programs may prove vital when addressing themes of shame and guilt and restoring trust among the users' closest relationships [170,172]. The only randomized controlled trial that exists with problematic online pornography users focuses on Acceptance and Commitment Therapy (ACT) [183], an improvement from their 2010 case series [184], which was the first experimental study to specifically address POPU. The study showed effective results but it is hard to extrapolate since the sample was again too small and focused on a very specific population.

The reported success with CBT, conjoint therapy and ACT might rely on the fact that are based on mindfulness and acceptance frameworks; depending on the context, increasing pornography use acceptance may be equally or more important than reducing its use [170].

#### **4. Discussion**

It seems that POPU is not only one subtype of hypersexual disorder, but currently the most prevalent since it also frequently involves masturbation. Although this is difficult to accurately determine given the anonymity and accessibility factors that make pornography use today so pervasive, we can at least confirm that the patron of consumption for pornography has changed for roughly the last decade. It would not be absurd to assume its online variant has had a significant impact on its consumers, and that the triple A factors enhance the potential risk for POPU and other sexual behaviors.

As we mentioned, anonymity is a key risk factor for this sexual behavior to develop into a problem. We need to keep in mind that statistics regarding this problem are obviously limited to people of legal age to engage in sexual activity, online or otherwise; but it does not escape us that sexual activity rarely starts after this threshold, and there is a likely chance that minors still in the process of sexual neurodevelopment are a particularly vulnerable population. The truth is that a stronger consensus

on what pathological sexual behavior constitutes, both offline and online, is necessary to adequately measure it in a representative manner and confirm how much of a problem it is in today's society.

As far as we know, a number of recent studies support this entity as an addiction with important clinical manifestations such as sexual dysfunction and psychosexual dissatisfaction. Most of the existing work is based off on similar research done on substance addicts, based on the hypothesis of online pornography as a 'supranormal stimulus' akin to an actual substance that, through continued consumption, can spark an addictive disorder. However, concepts like tolerance and abstinence are not yet clearly established enough to merit the labeling of addiction, and thus constitute a crucial part of future research. For the moment, a diagnostic entity encompassing out of control sexual behavior has been included in the ICD-11 due to its current clinical relevance, and it will surely be of use to address patients with these symptoms that ask clinicians for help.

A variety of assessment tools exist to help the average clinician with diagnostic approaches, but delimiting what is truly pathological and not in accurate manner is still an ongoing problem. So far, a crucial part of the three sets of criteria proposed by Carnes, Goodman, and Kafka include core concepts of loss of control, excessive time spent on sexual behavior and negative consequences to self and others. In some manner or other, they are also present in the majority of screening tools reviewed.

They may be an adequate structure in which to build upon. Other elements, that are considered with varying degrees of importance, probably signal us to take individual factors into account. Devising an assessment tool that retains some level of flexibility while also being significant for determining what is problematic is surely another of the current challenges that we face, and will probably go in hand with further neurobiological research that help us better understand when a specific dimension of common human life shifts from normal behavior to a disorder.

As for treatment strategies, the main goal currently focuses on reducing pornography consumption or abandoning it altogether, since clinical manifestations appear to be reversible. The way to achieve this varies accordingly to the patient and might also require some individual flexibility in the strategies utilized, with a mindfulness and acceptance-based psychotherapy being equally or more important than a pharmacological approach in some cases.

**Funding:** This research received no external funding.

**Conflicts of Interest:** Rubén de Alarcón, Javier I. de la Iglesia, and Nerea M. Casado declare no conflict of interest. A.L. Montejo has received consultancy fees or honoraria/research grants in the last five years from Boehringer Ingelheim, Forum Pharmaceuticals, Rovi, Servier, Lundbeck, Otsuka, Janssen Cilag, Pfizer, Roche, Instituto de Salud Carlos III, and Junta de Castilla y León.

#### **References**


© 2019 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).

MDPI St. Alban-Anlage 66 4052 Basel Switzerland Tel. +41 61 683 77 34 Fax +41 61 302 89 18 www.mdpi.com

*Journal of Clinical Medicine* Editorial Office E-mail: jcm@mdpi.com www.mdpi.com/journal/jcm

MDPI St. Alban-Anlage 66 4052 Basel Switzerland

Tel: +41 61 683 77 34 Fax: +41 61 302 89 18