part0026

All of the theories of sexual offending against children discussed in the previous chapter include pedophilia as an important motivation and explanatory factor. Pedophilia still needs an explanation, however, preferably including those who do not commit sexual offenses against children. In this chapter, I discuss two main approaches to explaining pedophilia. The first explains pedophilia as a neurodevelopmental disorder, encompassing evidence of prenatal and perinatal risk factors and evidence of differences in brain structure and/or function. The second views pedophilia as a behavioral disorder related to perturbations in social and sexual development, beginning with early exposure to sex, childhood sexual abuse, and then later onset of peer relationships. Learning is central to the behavioral disorder approach: Early exposure to sex or childhood sexual abuse is presumed to be associated with greater sexual response to children because of conditioning, where child-related cues are associated with sexual arousal and reward through physical stimulation and later fantasy and masturbation.

Neither approach is deterministic, in the sense that no unique neurodevelopmental risk factor or neurostructural or neurofunctional difference is surely indicative of pedophilia, and in the sense that many children who are sexually abused or have unusual childhood sexual experiences do not go on to develop pedophilia. Instead, the presence of risk factors influences the probability that someone will develop pedophilia, and an interaction is likely wherein individuals are predisposed to develop pedophilia because of pre- or perinatal risk factors that change brain development, and this predisposition is activated by unusual early experiences. In this integrated perspective, at-risk individuals who do not have unusual childhood sexual experiences are less likely to develop pedophilia, and children who are sexually abused or who have precocious sexual experiences but do not have the neurodevelopmental differences do not go on to develop pedophilia either.

Before elaborating on explanations for pedophilia, I want to examine the question of whether pedophilia is a mental disorder. Some have argued that the designation of sexual attraction to prepubescent children as a mental disorder is arbitrary, reflecting contemporary social norms and biases (e.g., Sandfort, Brongersma, & van Naerssen, 1991). However, the designation of a phenomenon as a disorder can be made on the basis of biological pathology , which can be defined as a disturbance in a mechanism or process that interferes with the ability of that mechanism or process to perform as designed by natural selection (Spitzer & Wakefield, 2002; Wakefield, 1992). Under this definition, I believe pedophilia qualifies: Pedophilia is statistically rare in the general population; it interferes with reproductively viable sex, especially in its exclusive form; and it has been negatively sanctioned across time and cultures. The sine qua non of a Wakefieldian test is whether someone would prefer sex with a prepubescent child when sexually mature partners are potentially available, not whether an otherwise teleiophilic man might experience some sexual response to a prepubescent child. The fact that I have argued that pedophilia can be viewed as akin to a sexual orientation regarding age does not disqualify considering it also as a mental disorder (Seto, 2012, 2017b); the Wakefieldian definition of disorder is agnostic about whether the sexual attraction to children is experienced like an orientation or about its neurodevelopmental and/or social learning etiologies.

Both pedophilia and sexual offending against prepubescent children are puzzling from a Darwinian perspective. Being sexually attracted to nonfertile, prepubescent children, especially if it is a preference for these children over sexually mature persons, or the exclusive form of pedophilia, is evidence of biological maladaptation. In this section, I go further into the logic for this view. Readers who are unfamiliar with modern evolutionary theories are referred to books by Buss (2015), Ridley (1994), and G. C. Williams (1996). In the forensic realm, Wrangham and Peterson (1996) wrote a fascinating book offering a Darwinian perspective on male violence, and my colleagues have written excellent Darwinian analyses of rape and juvenile delinquency, respectively (Lalumière, Harris, Quinsey, & Rice, 2005; Quinsey, Skilling, Lalumière, & Craig, 2004).

Another important consideration is that inclusive fitness theory predicts incest avoidance because of negative impacts on Darwinian fitness (the genetic contribution of an individual to the next generation, relative to the rest of the population) due to inbreeding depression and psychological or physical harm to genetically related child victims (Hamilton, 1964a, 1964b). The key premise of inclusive fitness theory is that preferential treatment of kin over nonkin would be selected for; individuals who did not discriminate in sharing food, shelter, or other resources would be less likely to pass on genes than those who preferred kin. Concomitantly, a person’s fitness is affected not only by direct reproduction but also by indirect reproduction via genetic relatives (hence, inclusive). Incest avoidance and incest taboos serve to protect inclusive fitness by reducing the likelihood of reproducing with close genetic relatives (Thornhill, 1993). I explore these ideas further in the next chapter .

If pedophilia is indeed maladaptive, how does it persist over time? Different ideas have been proposed. The leading candidate is that male pedophilia is the maladaptive extreme of a male tendency to be attracted to youthfulness cues. Robust evidence indicates that, on average, heterosexual men prefer female partners who are younger and heterosexual women prefer male partners who are older (Kenrick & Keefe, 1992). Female fertility is correlated with youthfulness, and thus men who preferred youthful partners who were nonetheless sexually mature were more likely to produce offspring and pass on the preference than those who were indifferent to cues of youthfulness (see Symons, 1979; Trivers, 1972). Not surprisingly, there is ample historical and cross-cultural evidence that men prefer cues of youthfulness—smooth skin, lustrous hair, neotenous (childlike) facial features, playfulness, vivacity—that are accompanied with cues of sexual maturity—full breasts, waist-to-hip ratio close to .67, and adult size. Pedophilia may represent a malfunction of this youthfulness-seeking propensity, such that men with pedophilia respond strongly to cues of youthfulness even in the absence of cues of sexual maturity (Quinsey & Lalumière, 1995; Seto, 2008). 1

Men who seek youthful sexual partners with neotenous faces but sexually mature bodies are more likely to successfully reproduce than men who are indifferent to these age-related cues. Malfunctions can sometimes occur, however, resulting in someone who is highly attuned to neotenous faces but not responsive to sexually mature bodies, resulting in pedophilia. Other kinds of malfunctions—for example, an indifference to neoteneous faces—could reflect chronophilias, in which a man prefers postmenopausal or older women who are past their reproductive years over young sexually mature women (Seto, 2017b). (Note that I am talking about preferences, not attraction; many men are rightly attracted to middle-aged or older women, but would they sexually prefer the older or younger version of that attractive person?)

Research on the biological substrates of youthfulness detection, including research on the brain regions selectively activated by human faces and body shape, could reveal if this explanation is correct. For example, Ponseti et al. (2015) showed that pedophilic men differ from nonpedophilic men in their brain activation to child versus adult faces, focusing on areas associated with reward and sexual response generally. An interesting question—if there were sufficient resolution—is whether pedophilic and nonpedophilic men differ specifically in their processing of child versus adult faces, because of a module specifically designed for face recognition (see Kanwisher, 2000). It would also be fascinating to further distinguish child and adult faces by maturity category; for example, are brain activations different for young sexually mature adults compared with elderly adults?

A by-product explanation suggests there is a trade-off between the costs of pedophilia (in terms of the individual’s reproductive fitness) and some indirect benefit to him or to his kin (recall the definition of inclusive fitness). Thus, pedophilia could be maintained in the general population, even though it is expected to reduce the reproductive fitness of the pedophilic individual, as long as it is offset by benefits to kin who share some of the genes underlying pedophilia but who do not have the condition. In other words, the direct fitness costs of preferring prepubescent children or having no sexual attraction to any other type of person could be offset if the affected person’s brothers were more successful in reproducing, for example, by having a more finely tuned sensitivity to youthfulness. This is a testable hypothesis and may tie into other research on the brain structure or functional differences between pedophilic and nonpedophilic men: Maybe the brain areas that have been implicated may be involved in the processing of potential partner cues, including youthfulness.

Researchers may eventually discover that pedophilia is not associated with by-product benefits but instead is a result of early maternal or child exposure to certain pathogens that have coevolved with people and thus engage us in an ongoing arms race. Ewald (1996, 2002) and others have argued that many chronic disorders are associated with pathogen exposure, thereby explaining a range of different facts, including the persistence of serious disorders that have substantial fitness costs, high prevalences of certain acute infections in the history of those with the chronic condition, familial aggregation among nongenetically related members, as with spouses who develop the same disease; and disease prevalence in some small communities in close contact with each other. More recently, Washington (2015) reviewed the evidence for mental disorders. The best-known example involves schizophrenia, which occurs in approximately 1% of the male population and is associated with dramatically lower reproduction because of reduced functioning because of delusions, hallucinations, or actively disordered thinking; not forming sexual or stable romantic relationships; and being less likely to have children of one’s own (e.g., Haukka, Suvisaari, & Lönnqvist, 2003). The onset of schizophrenia is associated with exposure of the person with schizophrenia’s mother to Toxoplasma gondii, a parasite carried by domesticated cats and passed on through their feces. Pathogens could perturb fetal development by impairing the mother’s health or could even have direct effects on fetal neurodevelopment. If the pathogen exposure hypothesis for pedophilia is valid, future research may reveal that the mothers of persons with pedophilia were more likely to have been exposed to certain pathogens during their pregnancies. This would not be the only potential explanation, however, as pedophilia may involve multiple causal pathways. Washington (2015) suggested that pathogen exposure might explain 10% to 20% of cases of mental illness, such as schizophrenia, autism, and a rapid onset form of tic or obsessive compulsive disorder because of strep throat in some children (a form of pediatric autoimmune neuropsychiatric disorders associated with streptococcal infections, a concept with substantial controversy in the medical field). In other words, the large majority of cases with these disorders involve other causes.

Brain abnormalities have been suspected as a cause of pedophilia since the condition was first described by von Krafft-Ebing (1906/1999). For example, Burns and Swerdlow (2003) described a man in his 40s who showed no evidence of pedophilic interest or behavior but then showed an onset of sexual interest in children that disappeared upon the removal of a right orbitofrontal tumor. At the same time, the orbitofrontal region is associated with inhibition of behavior, so it is possible that this man had a preexisting sexual interest in children that was never expressed, or was successfully concealed, but became apparent with the appearance of the tumor. Other, similar case examples exist (e.g., Mendez & Shapira, 2011).

The brain must be involved in pedophilia because all psychological traits and behavior involve the brain. The question is not whether the brain is involved; the key questions are, What structural or functional features are associated with pedophilia, and are they specific to pedophilia or involved more generally in atypical sexuality or all sexuality? And, given almost all neuroscientific research has been conducted on clinical or correctional samples of pedophilic offenders, can brain structure or function differences associated with pedophilia be further distinguished from those associated with criminality?

Consistent with the idea of prenatal factors influencing the onset of pedophilia (or, at least, the expression of pedophilic behavior; see below), the Centre for Addiction and Mental Health (CAMH) sexology research group, led by Ray Blanchard and then James Cantor, has produced results showing that pedophilic sex offenders differ in their rates of nonright-handedness, height and weight, and minor physical anomalies (Dyshniku, Murray, Fazio, Lykins, & Cantor, 2015; Fazio, Dyshniku, Lykins, & Cantor, 2017). 2 Handedness emerges prenatally and non–right-handedness is an indication of neurodevelopmental perturbations. Greater nonright-handedness is associated with mental disorders such as schizophrenia (Sommer, Aleman, Ramsey, Bouma, & Kahn, 2001) and is associated with sexual orientation for gender (Lalumière, Blanchard, & Zucker, 2000). Minor physical anomalies are small deviations in structure that arise because of developmental perturbations during the first or early second trimester (e.g., wide-set eyes, low-set ears, curved pinky fingers; Waldrop & Halverson, 1971). Most people have a few minor physical anomalies, but individuals with neurodevelopmental conditions such as schizophrenia, autism, and attention deficit/hyperactivity disorder have more minor physical anomalies, on average, than unaffected controls (e.g., Ismail, Cantor-Graae, & McNeil, 1998; Waldrop & Halverson, 1971; Weinstein, Diforio, Schiffman, Walker, & Bonsall, 1999). I had speculated in the 2008 edition that researchers would expect to find differences in measures of neurodevelopmental perturbation, such as fluctuating asymmetry and minor physical anomalies, and this is indeed what Dyshniku et al. (2015) found.

This evidence argues again for prenatal influences on pedophilia, which could involve maternal illness, pregnancy complications, or other early influences. However, we were not able to detect differential associations between prenatal and perinatal risk factors in a population-representative study of sexually and nonsexually violent offenders from Swedish national registries (Babchishin et al., 2017). One explanation is that we were not able to restrict our analyses to pedophilic sex offenders, even when we conducted analyses only on offenders with child victims (only some of whom would be expected to have pedophilia, as I discussed in previous chapters).

In their meta-analysis of 236 samples, Cantor, Blanchard, Robichaud, and Christensen (2005) found that adult sex offenders scored significantly lower than other offenders, who in turn scored lower than nonoffending controls. Similarly, adolescent sex offenders scored lower than other adolescent offenders. Supplementary analyses suggested the effect might be driven by the presence of sex offenders with child victims, because they scored lower than other sex offenders on measures of intelligence, and there was a positive association between intelligence score and the age cutoff used by original study authors to define child victims; in other words, samples comprising offenders against younger children—who would therefore be more likely to be pedophilic—had lower intelligence scores than those who offended against older children or adolescents. 3

Other research by the CAMH group has extended these results in studies using large samples of pedophilic offenders. Offenders who are diagnosed with pedophilia on the basis of sexual history and phallometrically assessed sexual arousal have lower intelligence scores than nonpedophilic offenders (Cantor et al., 2004). Moreover, in the same study, number of child victims under the age of 12 was negatively related to intelligence score, whereas number of older victims was not; also, the number of adult sexual partners was positively related to intelligence scores, indicating the relationship was not simply an effect of sexual history. Cantor et al. (2004) also found a similar pattern of results for measures of verbal and visuospatial learning and memory. Intelligence is both moderately heritable and influenced prenatally (Nisbett et al., 2012).

There is evidence for the impact of early childhood (postnatal) brain development as well. The CAMH group found that pedophilic sex offenders were more likely than nonpedophilic offenders to report head injuries resulting in at least momentary unconsciousness—a potential source of postnatal brain damage—before the age of 13 but did not differ in their reports of head injuries after age 13 (Blanchard et al., 2002, 2003). The fact that there was no difference in head injuries after age 13 challenges an alternative explanation that pedophilic sex offenders are more likely to report head injuries, because there is no a priori reason, such as socially desirable responding, to distinguish injuries before age 13 from those that occur after. This finding also counters the idea that the association may reflect accident proneness as a cause, because accident proneness could still be present after age 13. Instead, this finding suggests that head injury during a critical, sensitive period of brain development, before the onset or completion of puberty, is important. Perhaps relatedly, given evidence that traumatic experiences can affect brain development, childhood sexual abuse seems to have a stronger association with later sexual offending when it occurs between the ages of 3 and 7 (Grabell & Knight, 2009).

The CAMH group also pursued this research using more direct means of assessing brain structure (Cantor et al., 2008). In their literature review, the group identified three leading explanations about the development of pedophilia. The first explanation implicated frontal brain regions, viewing pedophilic sexual interest and behavior as an inability to inhibit (e.g., Flor-Henry, Lang, Koles, & Frenzel, 1991; Graber, Hartmann, Coffman, Huey, & Golden, 1982; Stone & Thompson, 2001). Many men show some sexual response to prepubescent and pubescent children, and one possible explanation for the difference between pedophilic and nonpedophilic men is that the former are less able to inhibit their sexual response to nonadults. Some evidence supports this idea, for example, MRI differences in the orbitofrontal cortex (decision making, sensory integration) when comparing 18 pedophilic individuals with 24 nonoffending controls (Schiffer et al., 2007).

The second explanation implicated temporal-limbic structures that are important in regulating sexual behavior (e.g., Hucker et al., 1986; Kolárský, Freund, Machek, & Polák, 1967; Langevin, Wortzman, Dickey, Wright, & Handy, 1988; Langevin, Wortzman, Wright, & Handy, 1989). Temporal-limbic dysfunction might explain the association between mood problems and paraphilic sexual offending (Kafka & Hennen, 2002). This idea is consistent with a small study showing differences in limbic structures between 15 pedophilic individuals and 15 nonoffending controls (Schiltz et al., 2007). However, a potential criticism of both Schiffer et al. (2007) and Schiltz et al. (2007) is that differences might be due to criminality rather than pedophilia per se, because people in the pedophilic group were identified because of sexual offending and were compared to nonoffending men.

The third explanation implicated sex-dimorphic structures affected by brain masculinization. The CAMH group backed this hypothesis, predicting that pedophilia would be more strongly correlated with the volumes of sexually dimorphic brain structures, with any differences in frontal or temporal areas limited to sexually dimorphic structures. Unexpectedly, the results did not support any of these hypotheses; instead, the scans revealed that 65 pedophilic sex offenders had significantly less white matter (the neuron axons connecting brain structures) than 58 nonpedophilic, non–sex offender controls in two major tracts, the superior occipito-frontal fasciculus and the right arcuate fasciculus (Cantor et al., 2008). Both of these white matter tracts connect cortical regions that are involved in the evaluation of stimuli as being sexually relevant, suggesting that pedophilic men may differ from other men by being less coordinated in their processing of sexual cues (see also Poeppl et al., 2015, which replicated results suggesting dysconnectivity in a “sex network”; not replicated by Gerwinn et al., 2015). These white matter differences might be relevant to how facial, body, and other cues of potential attractiveness are perceived and integrated.

White matter differences were not found in two prior studies with smaller samples, using more conservative statistical thresholds because no differences were expected in white matter (Schiffer et al., 2007; Schiltz et al., 2007). No differences were found in gray matter regions either, unlike other studies with smaller samples that compared pedophilic offenders with nonoffending controls. The results of Cantor et al. (2008) were replicated by the same team using a more sensitive method for examining white matter, diffusion tensor imaging, in a comparison of 24 pedophilic sex offenders and 32 nonoffending controls (Cantor et al., 2015). Only one third of the persons with pedophilia admitted to their sexual attraction to children, and almost two thirds, 62%, had committed child pornography offenses. Some of the differences between groups might be explained by criminality rather than pedophilia per se.

Results are summarized in Table 5.1 . Interpretation of this evidence regarding brain activation differences is limited by many neuroimaging studies having small samples and using different measures. A recent meta-analysis of functional imaging studies found no significant differences in activation foci between 58 pedophilic and 65 nonpedophilic participants drawn from seven studies (Polisois-Keating & Joyal, 2013). However, this was an analysis based on activation in response to preferred sexual stimuli compared with nonsexual stimuli; that is, persons with pedophilia did not differ in their activation for child versus nonsexual stimuli from those of nonpedophilic persons for adult versus nonsexual stimuli. A different question is whether persons with pedophilia differ from others in their activations to child versus adult stimuli. Here, researchers have found significant group differences, although studies vary in which areas are implicated (Poeppl et al., 2011; Schiffer et al., 2007; Walter et al., 2007).