Pedophilia is Linked to Reduced Activation in
Hypothalamus and Lateral Prefrontal Cortex
During Visual Erotic Stimulation
Martin Walter, Joachim Witzel, Christine Wiebking, Udo Gubka, Michael Rotte, Kolja Schiltz,
Felix Bermpohl, Claus Tempelmann, Bernhard Bogerts, Hans Jochen Heinze, and Georg Northoff
Background: Although pedophilia is of high public concern, little is known about underlying neural mechanisms. Although pedophilic
patients are sexually attracted to prepubescent children, they show no sexual interest toward adults. This study aimed to investigate the
neural correlates of deficits of sexual and emotional arousal in pedophiles.
with emotional and erotic pictures with functional magnetic resonance imaging.
Results: Regions showing differential activations during the erotic condition comprised the hypothalamus, the periaqueductal gray, and
dorsolateral prefrontal cortex, the latter correlating with a clinical measure. Alterations of emotional processing concerned the amygdala–
hippocampus and dorsomedial prefrontal cortex.
Conclusions: Hypothesized regions relevant for processing of erotic stimuli in healthy individuals showed reduced activations during
sexual interest of these patients toward adults.
Key Words: Dorsolateral prefrontal cortex, fMRI, hypothalamus,
neural mechanism, pedophilia, periaqueductal gray
et al. 2005). These patients are sexually attracted to prepubescent
children and can be characterized by emotional immaturity and
attentional deficits (Cohen et al. 2002; Fagan et al. 2002),
although they show no sexual interest toward adults.
Case reports described changes in the prefrontal cortex and
the medial temporal cortex in single pedophilic patients
(Burns et al. 2003; Dressing et al. 2001; Mendez et al. 2000).
Together with the ventromedial hypothalamus, the periaque-
ductal gray (PAG), and the insula, these regions have been
shown to be implicated in sexual arousal in healthy subjects,
including its vegetative-autonomic and emotional and motiva-
tional components (Karama et al. 2002; Mouras et al. 2003;
Redoute et al. 2000; Stoleru et al. 1999). In contrast, studies
investigating deficits in neural correlates of sexual arousal in
pedophilia have not yet been reported.
Our study aimed to test the hypothesis of whether pedophilic
patients show altered neural activity in those brains regions
implicated in sexual arousal. We addressed this question by
studying 13 pedophilic patients and 14 closely matched healthy
control subjects. With functional magnetic resonance imaging
edophilia is a psychiatric disorder of high public concern,
because 1–2 in 10 children have been sexually ap-
proached and abused by an adult (Fagan et al. 2002; Freyd
(fMRI), neural activity was recorded during visual presentation of
erotic and non-erotic emotional pictures of adults.
Methods and Materials
The group of pedophilic patients consisted of 13 male pa-
tients meeting the diagnostic criteria for pedophilia according to
DSM-IV (American Psychiatric Association 2000). Unmedicated
patients, in part under supportive psychotherapy, were recruited
from the State Forensic Hospital Uchtspringe, Germany. Exclu-
sion criteria were other psychiatric or neurological diseases, a
history of alcoholism or drug abuse, and previous antiandrog-
All patients had committed sexual offenses involving child
victims of ? 10 years of age. The number of victims sexually
assaulted by the offenders ranged from 1 to about 10 (mean 4.3,
SD 2.9). Patients were scored in a routine procedure including a
structured professional judgment (de Vogel et al. 2004) accord-
ing to the “Sexual Violence Risk 20” checklist (Boer et al. 1997)
and the “Multiphasic Sex Inventory (MSI)” (Nichols and Molinder
The control group was matched to the patient group for
group size, age, years of education, and verbal and general
intelligence. Except for one pedophilic patient, all subjects were
The study was approved by the local ethics advisory board of
the Medical School, Otto-von-Guericke-University Magdeburg,
and written informed consent was obtained from all participants.
All patients participated voluntarily without any benefits.
Subjects were asked to view a total of 256 erotic, emotional,
and neutral photographs from the International Affective Picture
System (IAPS; Lang et al. 1997) that were projected on a mirror
mounted on a standard headcoil. Erotic pictures depicting only
adults were previously validated for sexual arousal with ratings
by 21 healthy subjects. Erotic and emotional picture sets were
From the Department of Psychiatry (MW, CW, KS, BB, GN); Department of
Magdeburg; State Hospital for Forensic Psychiatry of Saxonia Anhaltina
(JW, UG), Uchtspringe; Department of Psychiatry and Psychotherapy
Address reprint requests to Georg Northoff, M.D., Ph.D., Department of
Psychiatry, University of Magdeburg, Leipziger Strasse 44, 39120 Mag-
deburg Germany; E-mail: firstname.lastname@example.org.
BIOL PSYCHIATRY 2007;62:698–701
© 2007 Society of Biological Psychiatry
matched for emotional arousal and valence, and all three picture
types were presented in a randomized order across all eight runs,
each picture lasting 5 sec (see also Supplement 1). A button press
at appearance of each picture was required to control subjects’
vigilance in the scanner.
To control for potential effects of preceding attention on
emotional picture viewing (Bermpohl 2006), one half of the
pictures were preceded by an expectancy period of 4–6 sec.
Following Kastner et al. (1999), the type of the subsequently
presented picture was indicated by presentation of specific white
arrows on a dark background. In contrast, to control for general
effects of the expectancy period, an ambiguous expectancy
period was followed by random presentation of stimuli.
After the picture presentation, a fixation cross was presented
for a variable duration of 8–10 sec (jitter steps of .5 sec) to make
use of this interval variation in the data analysis (Sakai and
Reaction times of button presses were recorded during the
fMRI scan. After the scan, subjective ratings of sexual and
emotional intensity and valence were obtained for each partici-
pant with a visual analogue scale. Mean averages of reaction
times and the three ratings were calculated for each stimulus type
and compared with two sample t tests, setting the level of
significance at p ? .05.
Data acquisition was conducted on a 1.5 Tesla General
Electric Signa scanner (Signa, General Electric Medical System,
Milwaukee, Wisconsin). Imaging procedures included collection
of inversion recovery T1 weighted echo planar images coplanar
with the functional images and echo planar functional images
(402 volumes, 23 slices with 3.125-mm in-plane resolution, 5-mm
thickness, 1-mm gap; T2* weighted gradient echo sequence:
repetition time 2 sec, echo time 40 msec). Image processing and
statistical analyses were carried out with SPM2 (Wellcome De-
partment of Imaging Neuroscience, London; Friston et al. 1995,
A total of 3160 (8 ? 395) volume images were realigned to the
first image to correct for head movement, mean-adjusted by
proportional scaling, resliced, and normalized into standard
stereotactic MNI space (Montreal Neurological Institute, isotro-
phic 3-mm resolution). Transformed data sets were smoothed
with a Gaussian kernel of 8 mm (full-width half-maximum).
Subject-specific low-frequency drifts in signals were removed by
a high pass filter of 128 sec, and effects of separate picture
viewing conditions were calculated with the General Linear
Modelling implemented in SPM2. Group comparisons were
tested with a two sampled t test in a random effects model. All
whole brain analyses were performed at p ? .005 uncorrected
and a cluster threshold (k) of 10 contiguous voxels. Resulting
clusters were entered into a region of interest (ROI) analysis. For
correlation analysis, a simple regression of MSI sub scores of
sexual contact with children with subjects’ contrast images was
performed on a whole brain level. For resulting regions, ex-
tracted percent signal changes were tested with Spearman cor-
relation with a two-sided p ? .05.
Previously matched for emotional valence and arousal values
provided by the IAPS catalogue (Lang et al. 1997), the behavioral
data revealed significantly higher sexual arousal and emotional
intensity ratings for erotic pictures than for non-erotic emotional
pictures in both groups of healthy and pedophilic subjects,
although these differences did not differ in extent from healthy
subjects to control subjects. No significant group differences in
ratings of both sexual arousal and emotional intensity of erotic
and non-erotic emotional pictures or in reaction times were
obtained between groups.
the contrast sexual arousal ? emotional arousal between healthy
and pedophilic subjects. Pedophilic patients showed significantly
lower signal intensities (peak voxel x, y, z, Z scores) in the
hypothalamus (?3, ?3, ?15; Z ? 2.74), the dorsal midbrain with
peak activation in the PAG (3, ?30, ?18; Z ? 3.34), the dorsolateral
prefrontal cortex (DLPFC) (?27, 42, 36; Z ? 3.18), the right lateral
parietal (60, ?51, 6; Z ? 3.31), the right ventrolateral (57, 27, 0; Z ?
3.34), and the right occipital cortex (24,?78, 6; Z ? 3.43) as well as
the left insula (?33, 12, ?3; Z ? 2.90) for the sexual arousal
condition than healthy subjects (p ? .005 uncorrected, k ? 10
voxels, Figure 1, Supplement 2).
Signal differences in the hypothalamus and left DLPFC were
observed in both the expected and the unexpected (i.e., ambig-
uous) mode when analyzed separately. The PAG, insula, and
right ventrolateral and right parietal cortex showed differential
activations only in the unexpected (i.e., ambiguous) mode,
whereas a cluster in the left orbitofrontal cortex showed differ-
ential activation (?36, 30, ?18; Z ? 2.83) only found in the
Correlation between clinical-behavioral measures and re-
gional signal changes during sexual (compared with emotional)
arousal revealed a significant negative relationship also at p ?
.001, uncorrected, between the MSI subscale for the sexual abuse
of children and signal intensities in the DLPFC and the left
occipital cortex. The higher the score for sexual abuse of
children, the lower the signals obtained only during sexual
arousal (Figure 1 and Supplement 2).
To investigate neural correlates of emotional processing, signal
changes during emotional pictures were compared with viewing of
intensities (peak voxel x, y, z, Z-scores) in the dorsomedial prefron-
tal cortex (DMPFC) (?6, 66, 24; Z ? 3.63), the retrosplenial cortex
(12, ?45, 3; Z ? 2.96), and the left amygdala–hippocampal
complex (?21, ?21, ?15 Z ? 3.52) in the non-erotic emotional
condition (p ? .005 uncorrected, k ? 10 voxels).
When compared with neutral pictures, erotic pictures re-
vealed similar patterns of differential activation, like the erotic ?
emotional picture contrast, but showed additional effects for the
left amygdala and left parahippocampal gyrus corresponding to
regions of the emotional ? neutral contrast and further in the
right parietal cortex (24, ?60, 60; Z ? 3.41; p ? .005 uncorrected,
k ? 10 voxels).
Our results demonstrate, for the first time, abnormal neural
activity in subcortical and cortical regions in pedophilia during
sexual arousal. Subcortical regions like the hypothalamus and
the dorsal midbrain are involved in the vegetative-autonomic
component of sexual arousal in healthy subjects (Bancroft et al.
2005; Ferretti et al. 2005; Karama et al. 2002). Although some
caution regarding the level of significance or potential group
differences in baseline perfusion rates is warranted, our findings
indicate that pedophilic patients remain unable to recruit vege-
tative-autonomic regions during stimulation with sexually arous-
ing stimuli of adults, which might account for their lack of sexual
interest toward adults. This is further supported by observation
M. Walter et al.
BIOL PSYCHIATRY 2007;62:698–701 699
of vegetative-autonomic (i.e., phallometric) abnormalities in
pedophilia (Cohen et al. 2002).
Subcortical regions involved in vegetative-autonomic pro-
cessing are controlled by cortical regions like the DLPFC that
modulate sexual arousal (Beauregard et al. 2001). Reduced
activation in left DLPFC and correlation with a clinical measure of
abuse thus suggest abnormal cortical control of sexual arousal in
pedophilia. However, the relationship between cortical and
subcortical regions (i.e., their functional and effective connectiv-
ity) (Morgane et al. 2005) during sexual arousal needs to be
investigated to further support this hypothesis. The specificity of
these findings and their relation to abnormal sexual urges toward
children remain to be investigated with corresponding pictorial
stimuli of children, which is ethically problematic. Differences in
patterns for the expected and the ambiguous mode signify the
influence of attentional processes and underline the importance
of further investigations.
We also observed reduced activation in the DMPFC, the hip-
pedophilic patients compared with healthy control subjects during
non-erotic emotional stimulation. These regions have been associ-
ated with the emotional component of sexual arousal (Ferretti et al.
2005; Karama et al. 2002; Redoute et al. 2000) as well as with
emotional processing in general (Northoff et al. 2000; Phan et al.
2002). Altered neural activity in these regions might reflect some of
the non-erotic emotional abnormalities observed in pedophilia, like
lack of assertiveness and emotional immaturity. The relevance of an
interaction of emotional and erotic processing, especially in the
amygdala and parahippocampal gyrus, was further supported by
the erotic ? neutral picture contrast.
Neural differences contrast with our behavioral findings in
pedophilic patients, which did not differ in their ratings of
sexual arousal and emotional intensity from healthy subjects.
This limits relation of neural findings to suppression or denial
of adult stimuli, but it contradicts with characteristic clinical
criteria and therefore complicates interpretation. Together
with the clinical measures, such neural–behavioral dissocia-
tion suggests that the validity of the pedophilic patients’
ratings might at least be questioned. Assessment of altered
neural activation could therefore be considered a complemen-
tary tool to investigate the pedophilic patient’s “true” feelings
of sexual arousal.
In summary, we observed abnormal activation in subcortical
(i.e., hypothalamus) and cortical (i.e., left DLPFC, DMPFC)
regions in pedophilia during visual-erotic stimulation with adults.
Because these regions are implicated in the vegetative-auto-
nomic and emotional components of sexual arousal, our findings
indicate possible neural correlates of lack of sexual interest
toward adults in pedophilic patients. This might not only con-
tribute to a better understanding of the pathophysiology of
pedophilia but might also serve to establish future therapy and
thus to lower public concern about this thus far rather neurosci-
entifically neglected psychiatric disorder.
The study was financially supported by a grant from the DFG
to GN (No 304/4-3).
viewing) when compared with pedophilic patients (p ? .005, k ? 10 voxels). The time course for this contrast is plotted over the next 16 sec (mean ? SEM,
children in the multiphasic sexual inventory (MSI, r ? .57, p ? .05). The brain section shows voxels revealed by the simple regression of MSI scores and the
contrast (erotic ? emotional picture viewing) (p ? .001, k ? 10 voxels).
700 BIOL PSYCHIATRY 2007;62:698–701
M. Walter et al.
We thank the staff from the Forensic Clinic Uchtspringe and Download full-text
the Department of Neurology II for their skillful assistance.
Supplementary material cited in this article is available
Mental Disorders, Text Revision. Washington, DC: American Psychiatric
Beauregard M, Levesque J, Bourgouin P (2001): Neural correlates of con-
scious self-regulation of emotion. J Neurosci 21:RC165.
Bermpohl F. Pascual-Leone A, Amedi A, Merabet LB, Fregni F, Gaab N, et al.
(2006): Attentional modulation of emotional stimulus processing: an
fMRI study using emotional expectancy. Hum Brain Mapp 27:662–677.
Boer DP, Hart SD, Kropp PR, Webster CD (1997): Manual for the Sexual Vio-
Vancouver, British Columbia, Canada: Institute Against Family Violence.
Burns JM, Swerdlow RH (2003): Right orbitofrontal tumor with pedophilia
symptom and constructional apraxia sign. Arch Neurol 60:437–440.
Cohen LJ, Nikiforov K, Gans S, Poznansky O, McGeoch P, Weaver C, et al.
(2002): Heterosexual male perpetrators of childhood sexual abuse: A
preliminary neuropsychiatric model. Psychiatr Q 73:313–336.
SVR-20 and Static-99 in a Dutch sample of treated sex offenders. Law
Dressing H, Obergriesser T, Tost H, Kaumeier S, Ruf M, Braus DF (2001):
Homosexual pedophilia and functional networks—An fMRI case report
and literature review. Fortschr Neurol Psychiatr 69:539–544.
Fagan PJ, Wise TN, Schmidt CW Jr., Berlin FS (2002): Pedophilia JAMA 288:
Ferretti A, Caulo M, Del Gratta C, Di Matteo R, Merla A, Montorsi F, et al.
(2005): Dynamics of male sexual arousal: Distinct components of brain
activation revealed by fMRI. Neuroimage 26:1086–1096.
Friston KJ, Frith CD, Turner R, Frackowiak RS (1995): Characterizing evoked
hemodynamics with fMRI. Neuroimage 2:157–165.
Event-related fMRI: Characterizing differential responses. Neuroimage
Karama S, Lecours AR, Leroux JM, Bourgouin P, Beaudoin G, Joubert S,
Beauregard M (2002): Areas of brain activation in males and females
during viewing of erotic film excerpts. Hum Brain Mapp 16:1–13.
Kastner S, Pinsk MA, De Weerd P, Desimone R, Ungerleider LG (1999): In-
creased activity in human visual cortex during directed attention in the
absence of visual stimulation. Neuron 22:751–761.
Lang PJ, Bradley MM, Cuthbert BN (1997): International Affective Picture
Mendez MF, Chow T, Ringman J, Twitchell G, Hinkin CH (2000): Pedophilia
the limbic forebrain/limbic midbrain. Prog Neurobiol 75:143–160.
magnetic resonance imaging study. Neuroimage 20:855–869.
& Molinder Assessments.
Northoff G, Richter A, Gessner M, Schlagenhauf F, Fell J, Baumgart F, et al.
(2000): Functional dissociation between medial and lateral prefrontal
cortical spatiotemporal activation in negative and positive emotions: A
combined fMRI/MEG study. Cereb Cortex 10:93–107.
Phan KL, Wager T, Taylor SF, Liberzon I (2002): Functional neuroanatomy of
Brain processing of visual sexual stimuli in human males. Hum Brain
Sakai K, Passingham RE (2003): Prefrontal interactions reflect future task
operations. Nat Neurosci 6:75–81.
cessed July 2004.
Stoleru S, Gregoire MC, Gerard D, Decety J, Lafarge E, Cinotti L, et al. (1999):
Neuroanatomical correlates of visually evoked sexual arousal in human
M. Walter et al.
BIOL PSYCHIATRY 2007;62:698–701 701