Skip to Content


MDS makes every effort to publish accurate information on the website. "Google Translate" is provided as a free tool for visitors to read content in one's native language. Translations are not guaranteed to be 100% accurate. Neither MDS nor its employees assume liability for erroneous translations of website content.

Main Content

Hypersexuality in Parkinson's Disease: Systematic Review and Report of 7 New Cases

View the entire article with references and supplemental information on the Wiley Online Library. 
Note:  Reference links embedded in the article below will also take you to the article on the Wiley Online Library.

Return to Table of Contents

Authors:  David Codling, Philip Shaw and Anthony S. David

Article first published online:   30 APR 2015 | DOI: 10.1002/mdc3.12155


Hypersexuality (HS) was one of the earliest examples of an impulse control disorder (ICD) or behavior to be associated with treatment for Parkinson's disease (PD), with an estimated prevalence of approximately 3.5%. Here, we report on a systematic review of the published literature of HS in PD with a view to uncovering evidence as to whether it is distinct from other ICDs. In addition, we report on 7 new cases that had broad neuropsychological testing, including a gambling test, which taps into reward and inhibitory mechanisms. The review uncovered a number of case series and cohorts that comment on the prevalence of HS, but very few made systematic comparisons with other ICDs, although younger onset and male sex are usually noted. A few studies have begun to map out a neuropsychological profile for HS, and our own cases show particular deficits in learning from negative outcomes, but, overall, there are insufficient data to draw firm conclusions. Functional imaging has shown patterns of increased content-specific activation in response to sexual material and this might relate to increased dopamine release. We conclude with a brief survey of the neurobiology of sexuality, which suggests possible avenues for further research and treatment of HS.

Hypersexuality (HS) is usually considered to constitute a marked increase in sexual interest, arousal, and behavior, which has adverse consequences for the patient and their partner or carers, and is out of keeping with premorbid personality. It is often characterized by a preoccupation with sexual thoughts, frequent demands, and desire for sexual practice that might be quite different from those previously engaged in, and currently, habitual use of sex lines and Internet pornography or contact with sex workers. HS is classified as an impulse control disorder (ICD), along with problem gambling (PG), compulsive shopping (CS), compulsive eating (CE), and compulsive hobbyism and punding (repetitive goal-less tasks), complicating the treatment of Parkinson's disease (PD), and has been operationally defined.[1] ICD in PD may be seen as an example of disinhibitory psychopathology[2]—a failure of executive function to inhibit reinforcing responses—although alternative theories regard it as an increase in incentive salience akin to that observed in substance misuse disorders.[3]

HS, as a complication of treated PD, is rare in absolute terms, with estimates of prevalence from postal questionnaires and clinical surveys of approximately 2% to 4% (see below[1, 4, 5]), with prevalence among users of dopamine agonists (DAs) reported as 7.2%[1] and 3.5%[4] in two of these articles (Cooper et al., not drawing the distinction). A recent review offered an overall rate of 3.5% based on 29 epidemiological studies.[6] Variations in rates may be explained by differential rates of reporting, which may underestimate prevalence in certain populations for whom these symptoms might be seen as shameful, decreasing disclosure. Hypersexuality in PD includes not only quantitative changes in behavior, but also encompasses a range of qualitative alterations in sexuality,[7, 8] such as transvestism[9] or paraphilias,[10, 11] and even gender identity disorders.[12] Case descriptions of HS in PD often note that patients deny having such impulses before treatment for PD.[13] Cases have been described with levodopa[13-15] and DBS,[16, 17] although the majority are observed with use of DAs.[18-20] This correlation with DA use is more established in pathological gambling.[21]

Hypersexuality in the General Population

Prevalence estimates of compulsive sexual behavior range from 3% to 6% in the general U.S. population[22, 23] and was first described in 1886,[24] although operational criteria are a relatively new phenomenon.[25] The most established view of HS in the general population is that of a behavioral addiction, with significant comorbidity with substance misuse disorders, and higher scores than population average on impulsivity, as well as the similarities (appetitive cravings and persistently engaging in the activity despite adverse social consequences) between the disorders.[26] Nevertheless, in proposing Diagnostic and Statistical Manual of Mental Disorders, Fifth Edition (DSM-5), criteria, Kafka[27] preferred the term hypersexual disorder, rather than “sexual addiction,” reflecting ongoing uncertainty. Alternative theories are that HS represents an impulsive/compulsive phenomenon[28] or that it represents an obsessive-compulsive pathology. This is born from the high levels of comorbidity with anxiety disorders and a theory that compulsive fantasizing may be a means of decreasing anxiety.[26]

ICDs in PD

Behavioral symptoms in PD have attracted much attention since the landmark publication in 2000 by Giovannoni et al.,[29] and coinciding with the increasing use of potent[30] DA medications. Although ICDs are present in the population, patients treated with DAs have been shown to be at increased risk of ICDs above that of the general population, as well as of untreated PD.[31] Neurobiological mechanisms for ICDs have been postulated with localized brain abnormalities centering on the STN,[32] ventral striatum,[33] and frontal cortex.[34, 35] Cognitive testing and functional imaging data have implicated such domains as general impulsivity[35-37] and frontal lobe dysfunction,[38] poor negative reinforcement sensitivity,[39, 40] increased positive reinforcement sensitivity,[33, 34, 41, 42] poor recall, lack of ability to estimate risk,[43] and increased temporal discounting,[44] although Siri et al.[45] found preserved executive function despite increased aggressiveness and impulsivity.

HS can be distinguished from other ICDs having been identified in an earlier era,[7, 8, 13, 14, 46-48] but as with other ICDs, reports indicate that these impulses generally recede on withdrawal of the dopamine replacement or agonist drug,[49, 50] although HS may persist.[51] Early literature on dopamine replacement in PD also describes hypomanic symptoms.[47]

Understanding HS

Potential mechanisms by which HS might be mediated can be drawn from comparisons with fields such as addictions,[3] anxiety disorders, and ICDs,[52] as well as through research into psychological aspects of sexuality.[53, 54] Dopamine in the mesolimbic area is recognized as a mediator of reinforcement and is implicated in drug addiction[55] and is also recognized to play an important role in the regulation of sexuality.[54] However, hypersexual behavior has also been described as part of a general loss of impulse control associated with lesions in the prefrontal cortex.[55, 56]

There is, however, a growing evidence base for a regulatory mechanism specific for sexuality. Bancroft et al.[53] provide a compelling argument that sexual behavior is mediated by separate systems of inhibitory and excitatory controls. For example, individuals with erectile dysfunction score highly on measures of inhibition proneness, whereas males with hypersexual behavior show higher sexual excitation scores.[57] Given data from animal studies in which environmentally induced sexual inhibition could be overcome using amphetamine in rats,[58] this offers an intriguing possible alternative explanation for the link between dopamine and hypersexual behavior. Studies in humans have shown some effect of amphetamine[59] and yohimbine[60]—both of which are known to be dopaminergic, in the treatment of erectile dysfunction. Finally, it has been noted that HS can occur independently of any improvement in sexual function that dopamine replacement might incur.[15, 61, 62]

Aims of the Review

Given the differing characteristics of HS, its distinctive neurobiology, and the recent explanatory models put forward by Bancroft et al., we decided to search the published literature for evidence that HS fits into the current ICD paradigm, as well as whether a more complex explanation is required for this variety of behaviors. A systematic literature review was conducted using terms related to PD and HS, looking for studies that considered HS in relation to other ICDs. We also report on 7 new cases of our own from a clinical case register that had additional neuropsychological testing, including a gambling test that taps into neuropsychological mechanisms governing response to risk and reward.


An initial search was run using MEDLINE, EMBASE, and PsycINFO 1984 to the present (May 2013) in order to identify previous meta-analyses or systematic reviews on HS and PD. Next, in order to identify original contributions to this topic, the same three databases were searched by OVID using Medical Subject Headings (MeSH) terms Parkinson's disease AND hypersexuality OR paraphilia OR impulse control disorder OR ICD OR QUIP OR dopamine dysregulation syndrome OR impulsive-compulsive OR impulsivity. The title and abstract of the articles were then analyzed to look for relevance. All articles that made any reference to ICDs in general or HS (or another term denoting this, e.g., compulsive sexual behavior, or a subset thererof, e.g., paraphilias) were considered. Those found to be relevant were vetted against the inclusion criteria. Bibliographies from relevant articles found in the primary search were reviewed and additional references obtained. Inclusion criteria were:

  1. HS assessed using standardized criteria. Given that there is not currently any single recognized set of criteria, those using any validated criteria set (e.g., the MIDI or QUIP criteria) were included.
  2. Cases should consist entirely of idiopathic PD (iPD) and relate to pharmacological treatment.
  3. Article distinguishes HS from other ICDs.
  4. An attempt was made to distinguish some etiologically relevant criterion.
  5. Study was not a case report or case series of n < 5.


The initial search revealed 1,219 articles. After being scanned for relevance, several[7-12, 49, 50, 61-78] were rejected for being case reports, five[46, 79-82] were review articles, and 1,181 were irrelevant to the study. Twenty-three articles were studied further, with the majority being rejected for not examining etiological or pathological factors, followed by not using recognized or any specific criteria for HS, with the rest failing to draw conclusions specifically related to HS. These articles that explored the area in question, but did not fulfill criteria for inclusion, are tabulated in Supporting Appendix 1.

The references of the accepted articles were scanned and a further two relevant articles were found. Table 1 summarizes the characteristics of eight studies that looked at prevalence and etiological associations of HS comparative to other ICDs that met our criteria. Two categories of articles were recognized: cross-sectional studies, in which etiological factors relating to the different ICDs were explored, and case-control studies, which attempted to explore the pathology underlying the process.

Table 1. Studies of hypersexuality in PD meeting criteria for systematic review
Reference Study Type Case Selection Location Evaluation Participants Comparisons
  1. MMSE: Mini-mental status exam; VIQ: verbal IQ; PIQ: performance IQ; NART: National Adult Reading Test; FH: family history; UPDRS: Unified Parkinson's disease rating scale; OP: out-patient.

Fan et al.[83] Case control iPD on medication using patient database China Initial screening and Voon criteria 400 Demographics, LEDD, smoking, and alcohol status
Solla et al.[84] Cohort Fulfilling Brain Bank criteria for PD, presenting to PD clinic Italy Voon criteria 349 Demographics, medication, neuropsychiatric conditions, MMSE
Voon et al.[1] Cohort iPD, presenting to movement disorder clinic Canada Voon criteria 297 Psychiatric disorders, demographics, and medications
Cooper et al.[5] Cohort study All iPD registered at an outpatient clinic USA Voon criteria 141 Demographics, health history, NPI, and BDI
Lee et al.[87] Cross-sectional iPD, presenting to movement disorder clinic Korea MIDI 1,167 Length of PD, age of onset, motor symptoms, and LEDD
Weintraub et al.[4] Cross-sectional Patients with PD receiving follow-up at 46 centers USA/Canada MIDI 3,090 Demographics including education, FH of mental health problems, alcohol, and smoking
Vitale et al.[89] Case–control iPD attending PD OP clinic Italy MIDI 63 UPDRS, LEDD, LEDD-DA, frontal lobe: WCST, TMT, Stroop
Politis et al.[93] Case–control iPD with and without hypersexuality UK Voon criteria 24 fMRI of patients viewing neutral, reward, and erotic stimuli

Fan et al.[83] studied the demographics and treatment regimens of patients with ICDs, as well as smoking and alcohol status, in a Chinese population. A screening questionnaire for ICDs was sent to 400 patients with iPD, with those positive for ICDs compared to those without. Criteria for HS proposed by Voon et al.[1] were used. The study found that HS had the highest prevalence of any ICD (1.92%) and showed an increased prevalence, compared to controls, in contrast to other ICDs, which did not.

Solla et al.[84] took 349 subjects presenting consecutively to a PD clinic who fulfilled criteria for iPD. They then screened these patients for ICDs, using the Voon et al.[1] criteria for HS. They were concurrently assessed for motor complications, neuropsychiatric disorders, dementia, and PD severity. Of the ICDs, HS was the only one significantly associated with motor complications, with all ICDs associated with longer disease duration and higher l-dopa equivalent daily dose (LEDD), calculated according to previous criteria.[85]

Voon et al.[1] used a tripartite method of assessment of HS based around DSM criteria[86] and enquiry surrounding degree of functional impairment caused by the symptoms with concurrent mania/hypomania as an exclusion criterion. They studied 297 patients attending a PD clinic, who filled in a survey based around ICD criteria. They found a lifetime prevalence of 2.4%, but 7.2% in DA-treated PD, which was found to be associated with increased LEDD, independent of whether it was a consequence of increased DA use. Six of seven patients were also found to have comorbid depression. HS had a statistically significant association with male sex and earlier-onset PD.

In a cross-sectional study of 1,167 patients presenting to a PD clinic in Korea, which used the Minnesota Impulsive Disorders Interview (MIDI) as a screening tool, Lee et al.[87] found 33 HS patients (2.8%). HS and PG were both associated with male gender (75.8% for HS). Total LEDD was calculated, as well as LEDD accounted for by agonists. All ICDs had nonsignificantly higher agonist usage than controls.

Weintraub et al.[4] performed a cross-sectional study of 3,090 patients in 46 different movement disorder centers in North America. HS was identified using the MIDI as well as assessment by a specialist consultant using the patient notes. Demographic data were taken and LEDD was calculated; 3.5% had “compulsive sexual behavior.” HS was associated with male sex. HS was the only one of the ICDs measured not to be related to gambling problems in a first-degree relative. A selection of 282 ICD cases identified in this study and 282 PD controls went on to have further tests.[88] As a group, ICD patients scored more highly on measures of impulsivity and general psychopathology, although the subgroup with HS (n = 47) differed from PD controls only on state anxiety, except that all but 1 was male.

Vitale et al.[89] took 49 patients with ICDs—PG, HS, CE, and multiple ICDs—all diagnosed using the MIDI, as well as 14 parkinsonian controls without ICDs for neuropsychological profiling. This consisted of memory testing and tests of frontal/executive function: Wisconsin Card Sorting Test (WCST)[90]; Trail Making Test (TMT)[91]; Rey-Osterrieth Complex Figure Test; matrices; and Stroop Color Word Test. Whereas there was no significant difference in demographics or disease burden associated with PD between groups, there were differences in neuropsychological profile. All ICD cases scored worse on the Rey figure test and TMT, suggesting difficulties with spatial planning and set shifting. Patients with HS, however, scored significantly worse on memory function and on the Stroop test. The researchers suggest that this indicates that HS, more than other ICDs, is primarily a disorder of impulse control (see a previous work[92]).

A study specifically into HS was performed by Cooper et al.,[5] who mailed 400 patients with questions regarding demographics, the Neuropsychiatric Inventory (NPI), Beck's Depression Inventory (BDI), as well as a screening questionnaire for HS, which included a list of recognized behaviors. Those who were positive and accepted were then invited for interview against the Voon criteria. Overall, 15 of the 141 returners were screened positive for HS, although only 6 of these were positive on interview. Of these, correlation was found with earlier onset of disease, but there was no correlation found with marital status or psychiatric illness, as measured on the BDI and NPI. Significantly, there was no correlation found with gender or use of DAs.

Politis et al.[93] used functional MRI (fMRI) in 12 PD patients with and 12 without HS to examine functional activation when they watched neutral and then erotic stimuli. Those patients with HS were found to have greater activation of limbic, paralimbic, temporal, occipital, somatosensory, and prefrontal cortices, which the researchers suggest correspond to emotional, cognitive, autonomic, visual, and motivational processes. Enhanced activation in cingulate and orbitofrontal cortices strongly correlated with increased sexual desire. The study had relatively few patients and found diffuse changes in activation, making the drawing of straightforward conclusions difficult. It is also difficult to determine whether these patients are merely more “reward sensitive” in general or whether this is specific to erotic cues, owing to the fact that only erotic cues were given.

Mamikonyan et al.[94] followed up patients diagnosed with ICDs, contacting them by telephone on average 29.3 months after diagnosis with an ICD. Fifteen subjects were successfully contacted and were administered a MIDI as well as being asked about subjective severity of ICD symptoms at follow-up. Seven patients were identified with HS, 6 of whom had their DA decreased or discontinued at follow-up with a lower contribution of DA to total LEDD. Of these, 6 had complete remission of symptoms and no longer met MIDI criteria for compulsive HS, and 1 described a subjective partial remission. Similar results were found in CG and CS. For ICDs as a whole, LEDD was not statistically different, although DA LEDD was decreased and LEDD increased with statistical significance. The study suggests that DA LEDD is implicated in the genesis and continuation of HS symptoms as well as those of other ICDs (see also a previous work[51]).

New Cases

A database of 592 patients with iPD attending the regional Neurosciences Clinic at King's College Hospital in London was screened to identify patients whose notes explicitly mentioned symptoms suggestive of HS or associated behaviors. The patients identified were then interviewed by a psychiatrist using the Structured Clinical Interview for DSM-IV supplemented by a detailed psychosexual history. Seven patients met the criteria for HS and their demographic, clinical, and neuropsychological details are given in Table 2.

Table 2. Clinical and neuropsychological features of PD cases with hypersexuality
Variable Case 1 Case 2 Case 3 Case 4 Case 5 Case 6 Case 7


Gender M M M M M M M
Age (yr)
Onset of PD 50 61 42 54 32 43 20
Onset of HS 53 65 60 58 38 55 30
At assessment 54 70 63 60 41 58 51
H & Y stage 2 3 3 2 2.5 3 4
Medication at onset of symptoms

















Associated symptoms
Sexual behavior and other impulse control problems


Pathological gambling

Phone lines

Change in orientation



Possible pedophilic tendencies

Phone lines

Incessant demands for sex


Benzhexol misuse



Incessant demands

May have assaulted daughter


Phone lines

Including gambling; misuse of DA meds

Excessive masturbation

Frequent demands


Excessive masturbation; frequent demands

Misuse of DAergic meds

Mood disturbance Depression Hypomania Hypomania
Other Pathological jealousy toward wife Hypochondrical delusions and second person AHs
FSIQ (VIQ/PIQ) 98 (98/98) 114 (127/93) 103 (105/101) 98 (106/95) 84 (88/80) 58 (62/55) 80 (84/77)
Premorbid (NART) 102 113 109 120 92 78 94
Logical (immediate/delayed) 9/9 7/6 8/4 6/4 10/9 3/2 4/4
Executive function
Hayling Low average Low average Low average Low average Poor Poor
Brixton Impaired Low average Poor Poor Abnormal Impaired
Stroop (centile) 25th 25th–50th 10th–25th 10th–25th <2nd <2nd
Gambling task Impaired Impaired Impaired Intact Impaired Impaired N/A

All patients reported an intense preoccupation with sex, with mounting levels of tension until they found a sexual outlet, which resulted in a transient sense of relief. The most frequent sexual activities were masturbation (most patients trying to masturbate at least twice a day) and use of telephone and Internet sex lines. Five patients described sex with prostitutes (which only 1 admitted to premorbidly) and 4 were noted to proposition female friends and clinical staff. Several patients developed paraphilias; 1 described a completely new interest in masochistic sex; another admitted to indecently exposing himself on several occasions; and 1 patient drilled holes in the bathroom walls of his house so he could watch his unsuspecting partner undress. Another was accused of sexually assaulting his daughter by inappropriate physical contact. The behaviors led to relationships with partners ending in 4 cases.

In 5 of these patients, there was no clear temporal relationship between adjustments in medication and the onset of hypersexual symptoms, nor in fluctuations of the levels of medication and symptom intensity. All patients had their dopaminergic medication reduced, but in 4 this merely changed the ability to act upon sexual desire, not the intense sexual preoccupations themselves. Misuse of medication was present in 3 cases and 1 of these patients linked his escalating libido to intentional overuse of anticholinergic, but not dopaminergic, medication.

Neuropsychological Testing

We tested 6 of our hypersexual patients' performance on a short battery of neuropsychological tests, including one designed to examine decision making under ambiguous circumstances. We used the Iowa Gambling Task[95] (IGT) in order to examine patient tendency to balance large rewards/losses over smaller rewards/losses. Other neuropsychological tests, including an estimate of IQ using the shortened version of the Wechsler Adult Intelligence Scale—Second Edition (WAIS-II), the Wechsler Memory Scale (logical memory immediate and delayed), and a short executive battery that included the Hayling Sentence Completion test (which requires subjects to inhibit an obvious missing word at the end of a sentence), the Brixton Spatial Anticipation test[96] (which examines the ability to detect rules in sequences non-verbal symbols—somewhat analogous to the WCST)—and the Stroop colour-word interference test (an index of selective attention[92]). The standard clinical scoring of the Stroop was used, namely, number of correct responses in 45s converted into a centile score.

These patients were compared to a control group of 11 subjects with PD with no sexual dysfunction (non-HS PD) who were matched as a group for demographic (age, 57 years, and gender), clinical stage of PD (median H & Y stage = 3; range, 2–4), and medication. Seven of the eleven controls were on agonist medication in addition to optimized doses of l-dopa; 3 were on entacapone and 1 on benzhexol and 1 on amantadine. Regarding neuropsychological variables, WAIS IQ showed an HS mean of 90.7 (range, 58–114) versus a non-HS PD mean of 94.0 (range, 68–112; t test, > 0.1). The groups had a similar spread of low average to definitely impaired scores on the Hayling and Brixton Tests (although this was not subjected to statistical comparison).


In over 100 trials of the IGT, the hypersexual patients tended to pick more disadvantageous cards than the PD control group (t = 2.02; df = 15; = 0.06). The group difference reached significance in the latter stages of the task, with the PD control group behaving in a normative manner and choosing progressively more from the advantageous cards, in contrast to the hypersexual subjects who persisted in making disadvantageous choices (over the last 60 trials, hypersexual mean number of disadvantageous choices = 34; standard deviation [SD] = 3; PD controls mean = 26; SD = 8; t = 2.3; df = 15; = 0.03; see Fig. 1; scoring system explained in legend). This behavior cannot be attributed simply to the impairments in response inhibition or perseveration among the hypersexual patients given that the comparison group had similar levels of executive dysfunction.


Figure 1. Performance on the IGT. “Hypersexual” PD patients with hypersexual behavior (n = 6; solid line) and PD controls (n = 11; broken line). Y-axis shows mean net score from advantageous decks (C and D) minus disadvantageous decks (A and B) over the course of five blocks (each block represents 20 consecutive card selections). HS patients show significantly poorer learning.



Discussion: Gambling Task

The study adds HS to the list of impulse control and substances misuse disorders in which deficits in the IGT have been reported, implicating ventromedial prefrontal cortex (VMPFC) dysfunction in their pathogenesis.[95, 97] The study is small, so the results must be interpreted with caution, although, as our review shows, studies exploring cognitive underpinnings of specific ICDs are rare. As well as studies in PD per se,[98] there are two previous studies that have examined decision making under varying conditions of risk and ambiguity in PD patients with and without ICDs. Rossi et al.[99] compared 7 PD patients with PG (2 of whom were also said to have HS) and 13 without ICDs. In general, the differences between the groups were minor, although the PG group showed a significant tendency to select disadvantageous decks, compared to the PD controls. Performance on other decision tasks, including those where risk and advantages were explicit, did not differ between groups. Bentivoglio et al.[100] recruited 17 PD patients with ICDs (8 had “compulsive sexual behavior”) and contrasted them with 17 matched PD controls. They found no statistically significant neuropsychological differences, although there were trends for the ICD group to do somewhat worse on frontal-executive tests and to show some increases in risky decision making. Voon et al.[33] examined decision making using a gambling task in 14 PD patients with and 14 without ICD (though none had HS), of whom 11 from each group underwent a series of fMRI studies. Performance was studied ON and OFF DA. The results were complex, with the ICD patients prone to make more risky choices when starting at 0 dollars relative to when starting at a loss and showing reduced ventral striatal activity. ICD patients showed increased sensitivity to risk with agonists, unlike PD controls. The same PD patients with ICD who took part in a reinforcement learning paradigm[34] showed faster rates of learning from gain outcomes with DAs, compared to PD controls, and when OFF agonists. Agonists appeared to increase striatal prediction error activity, perhaps signifying a “better than expected” outcome and, by implication, encouraging further “bets.”

Neuropathologically, it might imply that the hypersexual PD patients, assuming they have most in common with other ICD patients, may have either a relatively heavy local loss of dopaminergic neurons in the VMPFC or in the striatal projections to this area, both of which could increase sensitivity to DAs. Therapeutically, it suggests that reduction of dopaminergic medication is only one component of treatment, and future therapies based on an appreciation of the cognitive substrate of the problem may be more effective.

General Discussion

We carried out a systematic literature review in order to see whether there was emerging evidence that HS is a distinct entity from other ICDs with more subtle deficits than the category nomenclature implies. However, this review has shown that there is as yet insufficient evidence to draw this conclusion. Review of the literature showed that there were few articles that actively sought to distinguish HS from its umbrella term. Over 30 studies were identified initially, which did this with most looking solely at prevalence and obtaining very different figures (see Supporting Appendix 1) and most (except for eight; Table 1) failing to meet our entry criteria for the review. There is a lack of a defined standard for the diagnosis of HS, with the QUIP, MIDI, and Voon et al.[1] criteria all being used as part of the definition.

Demographically, there seems to be a consensus that HS is associated with younger onset of disease,[1] with most suggesting a preponderance in male patients[1, 4, 19, 87, 100] and increased LEDD,[1, 84] whereas Solla et al.[84] also noted that it was related to increased motor complications, suggesting a possible direct effect of dopaminergic stimulation. Interestingly, it was the only ICD that displayed this in what was a numerically small study. The gender bias in the study seems less easy to explain, but would seem to indicate that there are potential physiological predisposing factors to HS. Hence, the main demographic risk factors previously uncovered for ICDs as a whole (earlier age of onset and male gender) seem to be especially relevant for HS.

Cognitive Mechanisms

The current data on the pathological mechanisms of HS are quite sparse in the general population. However, from our systematic review of the literature, there were eight studies in PD (Table 1). Those that have looked at ICDs in general have concluded that they show increased impulsivity, decreased ability to learn from negative outcomes, and decreased ventral striatal activation, suggesting loss of inhibitory control. Vitale et al.[89] suggest primarily a deficit in impulse control with poor performance on the Stroop test and also poorer performance in memory testing, although evidence that resistance to attentional interference is a good measure of behavioral impulsivity is weak.[92] The generalizability of this study is limited by the small number of patients. The researchers' conclusion that development of ICDs is likely owing to pre-existing cognitive deficits, based on the lack of difference in LEDD between groups with and without ICDs, would be stronger if individuals underwent neuropsychological testing before developing ICDs. Voon et al.[88] found a consistent pattern in a larger group of patients with ICD, although neuropsychological variables did not point to profiles for each ICD subtype. There were some differences in terms of other psychiatric measures.

Recent efforts to understand the cognitive basis of behavioral addictions—such as HS—have emphasized the tendency to engage in immediately gratifying behavior, despite full awareness of adverse long-term consequences. Bechara et al.[95] have developed a paradigm to assess this tendency in which participants choose between packs of cards that bring high immediate monetary reward, but even higher intermittent associated losses (the “disadvantageous” packs) and packs that bring smaller immediate monetary gain, but even smaller losses (the “advantageous” packs). Whereas findings on the IGT and related gambling tasks have been mixed, both in detailed studies of PD patients per se[98] and those contrasting patients with and without ICDs,[33, 34, 45, 99, 100] our own data suggest that there are particular difficulties in this domain in HS patients that cannot be explained by general cognitive or executive dysfunction. However, in order to show that such deficits are specific to HS, carefully matched patients with other ICDs would need to be examined on this task as well.

Neuroimaging data from a smaller pool of patients, albeit one consisting entirely of HS patients, suggest that there may be some increased element of sexual desire, with patients showing increased activation of reward pathways in response to erotic cues.[93] Finally, in a PET study of 18 PD patients with a range of ICDs (5 of whom had HS), O'Sullivan et al.[102] used[11] C-raclopride PET to infer D2 receptor availability. ICD patients evinced greater reduction of raclopride-binding potential in the ventral striatum after reward-related cue exposure (which included some erotic images), relative to neutral cues, after l-dopa challenge. This was interpreted as a heightened response of striatal reward circuitry to reward-related visual cues consistent with a global sensitization to appetitive behaviors with dopaminergic therapy in ICDs.

The Neurobiology of HS

As with ICDs in general, there is currently a debate in the literature as to whether hypersexual disorder per se can be categorized as a behavioral addiction, an impulse control problem, or is more akin to an obsessive-compulsive disorder (OCD).[51, 103] Current understanding of the neurobiology of sexuality may help move this debate forward (for review, see a previous work[54]).

An established model of human sexual behavior uses the cycle of excitement, plateau, orgasm, and refraction and is largely based around reward. Animal studies have shown large increases in DA in the nucleus accumbens (NA) of rats around the time of orgasm, which rapidly decreases during the refractory phase.[104] This can be compared to the rush of euphoria produced by administration of opiates followed by a long period of relaxation. This is often referred to as the “pharmacological orgasm,” and endogenous opioids have been found to be released during ejaculation in male rats. This rather unique signature may be of value in differentiating HS from other ICDs. Alongside human data showing that individuals with high synaptic DA and strongest NA response to reward showed a lower age of first sexual intercourse and increased number of sexual partners,[105, 106] this gives a good theoretical underpinning for an addictive model of hypersexual behavior, which suggests that increased reinforcement may lead to an increase in sexual “wanting.”[107] It also hints at a possible role for opiate agonists in the treatment of HS—although this would require an extremely careful risk-benefit analysis.

It is also the case that hypersexual behavior shares many of the characteristics of an addiction with: continued engagement despite adverse consequences; appetitive urge or craving; and diminished self-control over engagement and compulsive engagement.[103] Brain-imaging studies indicate that sexual arousal and orgasm affect the mesolimbic reward systems, including the striatum and VMPFC,[54] which have been implicated in both substance misuse disorders and problem gambling, and are of interest in ICDs owing to the proposed increase ratio of D3 to other receptor subtypes in mesolimbic regions.[102, 108] Indeed, one of the distinguishing features of DA drugs versus l-dopa is affinity to D3, and hence the increased risk of ICD conferred by agonist treatment[4] has led to the inference that this receptor subtype is especially relevant to ICDs.[109] Nevertheless, whereas PG has been shown to correspond to high sensation seeking,[41, 88] this may be less true of hypersexual disorder.

As noted earlier, an alternative model of hypersexual behavior involves loss of inhibition. Bancroft et al.[53] propose a theory of sexuality involving dual control that consists of both inhibition and excitation, which are dissociable from each other. It is argued that inhibition may occur at times in which alternative challenges or threats are requiring of attention. Fiorino et al.[58] also review animal models that show similar inhibition of sexual response when rats are placed in a novel environment, which is overcome when dopaminergic systems are sensitized using amphetamine.[58] As yet, though, small-scale human experiments have failed to replicate a decrease in inhibition proneness in a group of self-identifying sex addicts. Whether HS in PD affects inhibition and/or excitation could be simply tested using the Bancroft approach and may lead to behavior modification pathways—both physical and psychological (see a previous work[110]).

Clinical and population surveys have found a correlation between impulsiveness and hypersexual behavior.[52] This could suggest that hypersexual behavior is related to difficulties in regulating impulses per se, rather than particular regulation of sexual impulses. Theories around impulsivity have generally focused on the noradrenergic connections to the right inferior frontal cortex, and their role in prepotent motor disinhibition and delay discounting[111, 101]—which is linked with orbitofrontal circuitry and its serotonergic and dopaminergic control.[108] There is limited systematic evidence for this outside attention deficit hyperactivity disorder, although patients with prefrontal lesions have been found to exhibit hypersexual behavior alongside an array of executive functioning in set-shifting, emotional regulation, and planning.[55, 56] The orbitofrontal cortex has been linked to abnormalities on the IGT,[95] which were found in the new cases of HS reported here. As noted above, non-neurological samples of “sexually addicted” patients also rate themselves as having general problems with impulsivity and self-control,[52] but, curiously, do not show objective executive function impairments.[112]

Finally, the high rate of comorbidity between anxiety disorders and HS, also observed in PD,[88] may suggest an obsessive-compulsive pathology, which is postulated to be a form of compulsive fantasizing used to relieve anxiety, which is gradually positively reinforced by the relief from anxiety. However, phenomenologically, there does seem to be discrepancy between the ego-syntonic nature of sexual thoughts in hypersexual disorder, and the more ego-dystonic obsessions of OCD. Schwartz and Abramowitz[113] noted high levels of sexual arousal and pleasure from performing sexual acts among patients classified as having a nonparaphilic sexual addiction. This was in contrast to that reported by the patients with OCD.

The foregoing review leads us to the following conclusions and recommendations:

  1. There appears to be relatively mixed data as to the prevalence of HS in the PD population, with a mean of approximately 3.5%,[6] with variation possibly owing to the different criteria used, or cultural variation (e.g., previous works[114, 115]).
  2. There are quite convincing demographic data on patients with HS, who seem to be often males on DAs (but see a previous work[5]) with higher LEDDs with earlier-onset diseases; there is less information as to what processes underlie this.
  3. The development of a standard tool for the assessment of HS, as exists in CG and CS, would undoubtedly advance the field.
  4. Our new data, based on a small case series of 7 men, suggest that there are particular difficulties in decision making associated with negative consequences in HS PD patients, over and above other “executive” difficulties, and this requires replication. In fact, the only study to compare patients across ICDs found that those with HS were, if anything, more impaired on response inhibition and working memory.[89]
  5. Greater sensitivity to the motor side effects of LEDDs was raised in one study[84] and, if replicated, could be used to alert clinicians to vulnerability to HS in patients conforming to the relevant demographic profile.
  6. At the neurobiological level, HS seems to overlap with other reward mechanisms and incentive salience (“wanting” rather than “liking”)[93] and probably involves a wide interconnected brain network extending well beyond the basal ganglia, although the role of endogenous opioids may lead to novel therapeutic approaches.


We use cookies to give you the best possible experience with our website. These cookies are also used to ensure we show you content that is relevant to you. If you continue without changing your settings, you are agreeing to our use of cookies to improve your user experience. You can click the cookie settings link on our website to change your cookie settings at any time. Note: The MDS site uses related multiple domains, including and This cookie policy only covers the primary and domain. Please refer to the MDS Privacy Policy for information on how to configure cookies for all other domains on the MDS site.
Cookie PolicyPrivacy Notice