Eye tracking and child sexual offenders: a systematic review

Figures & data

Table 1. Summary of the studies included in the literature review with samples, stimuli and results.

References	Sample (effectives)	Stimuli	Results
Fromberger et al. [36]	Three groups: paedophiles (n = 22); non-paedophilic forensic controls (n = 8); healthy controls (n = 52).	Pairs of pictures of children and adults at the same time (girl vs. woman, boy vs. man).	Paedophiles had a significantly longer viewing time for child stimuli (P = 0.037), a significantly shorter fixation latency for child stimuli (P = 0.002), and a significantly longer relative fixation time for child stimuli (P = 0.039). Receiver operating characteristic (ROC) analyses showed that fixation latency measurement discriminated between paedophiles and non-paedophiles with high accuracy (area under the curve, AUC = 0.902). Relative fixation time and viewing time also discriminated between paedophiles and non-paedophiles, (respectively AUC = 0.828 and AUC = 0.759)
Fromberger et al. [43]	Three groups: paedophiles (n = 22); forensic inpatients (n = 8); healthy controls (n = 52).	Pairs of pictures of children and adults at the same time (girl vs. woman, boy vs. man). Areas of interest were defined on each pictures (head, breasts, waist, and pubic region).	Paedophiles had a significantly shorter average entry time for the child pictures (P = 0.005); forensic and healthy controls had a significantly shorter average entry time for the adult pictures (respectively P = 0.021 and P < 0.001). Paedophiles showed the shortest entry time to the child’s head (P < 0.001), followed by the child’s pubic region and waist. Paedophiles showed a significantly longer relative fixation time for the child’s pubic region than other areas of interest (P < 0.001) but it is the same with adult’s pubic region (P < 0.001).
Jordan et al. [44]	Three groups: paedophiles (n = 22); forensic inpatients (n = 7); healthy controls (n = 50).	Pairs of three-dimensional cube figures and pictures of children (boys and girls) and adults (men and women). Rotation stimuli and one sexual distractor were showed in same time.	Paedophiles viewed sexual distractor (adult and child) significantly earlier than the forensic control group (P = 0.042) and paedophiles exhibited significant longer fixation latencies for mental rotation figures compared to both control groups (P = 0.009 with forensic control and P < 0.001 with non-forensic controls). Then, eye movements were computed in an attentional control index (ACI). Results indicate a significantly weaker control of eye movements in paedophiles compared to non-paedophiles. The ACI fixation latency with adult sexual distractor discriminated between paedophiles and non-paedophiles, sensitivity is hight (90.9%) and specificity is moderate (77.4%). With child sexual distractor, the ACI discriminated with hight accuracy between paedophiles and non-paedophiles, sensitivity and specificity are high (respectively 90.9% and 84.9%). ROC analysis with fixation time showed moderate accuracy with ACI for adult sexual distractor to discriminate paedophiles and non-paedophiles (sensitivity 71.9%, specificity 63.9%). With child sexual distractor, the ROC analysis discriminated with moderate accuracy between paedophiles and non-paedophiles (sensitivity 84.2% and specificity 63.6%).
Jordan et al. [45]	Three groups: paedophiles (n = 22); forensic inpatients (n = 7); healthy controls (n = 50).	Pairs of three-dimensional cube figures and pictures of children (boys and girls) and adults (men and women). Rotation stimuli and one sexual distractor were showed in same time.	Paedophiles exhibited shortest fixation latencies towards sexual distractor compared to both control groups, but it was only significant compared with the forensic control group (P = 0.042). Paedophiles differed from both control groups with longer relative fixation times to sexual distractor, but it was only significant compared to forensic controls (P = 0.014). Moreover, on age preference index (API) was computed, and ROC-analysis revealed that the API for the relative fixation time discriminated between paedophiles and non-paedophiles with moderate accuracy (AUC = 0.697, P = 0.007). The API differed between paedophiles and non-paedophiles with moderate sensitivity and specificity (respectively 71.9% and 63.6%).
Jordan et al. [46]	Three groups: outpatients with a sexual interest in children (n = 11); paedophilic forensic inpatients (n = 22); control group (n = 60) divided into eight forensic inpatients and 52 healthy subjects.	First step: pairs of pictures of children and adults at the same time (girl vs. woman, boy vs. man). Second step: sexual distractor task, one picture of character was showed with one pair of three-dimensional figures.	Outpatients with a sexual interest in children and paedophiles inpatients both viewed child stimuli earlier (only significant for paedophiles P < 0.05). Paedophiles inpatients viewed the child sexual distractor earlier (P < 0.05) and cognitive stimuli later than both other groups (P < 0.05).
Hall et al. [47]	Two groups: 13 child sexual offenders; 13 non offenders.	Pictures from Internet fashion catalogues, classified in pre-pubescent children, adults in their early 20s and adults in their late 30s or early 40s.	Child sexual offenders and non-offenders viewed adult stimulus without significant difference. With child pictures, child sexual offenders viewed more often toward the upper body of the female child figure than male child figure (P = 0.04). There is no significant difference in other regions, even with post hoc analyses.

Fromberger P, Jordan K, Steinkrauss H, et al. Diagnostic accuracy of eye movements in assessing pedophilia. J Sex Med. 2012;9:1868–1882.

 PubMed Web of Science ®Google Scholar

Fromberger P, Jordan K, Steinkrauss H, et al. Eye movements in pedophiles: automatic and controlled attentional processes while viewing prepubescent stimuli. J Abnorm Psychol. 2013;122:587–599.

 Web of Science ®Google Scholar

Jordan K, Fromberger P, von Herder J, et al. Impaired attentional control in pedophiles in a sexual distractor task. Front Psychiatry. 2016;7:193.

 Web of Science ®Google Scholar

Jordan K, Fromberger P, von Herder J, et al. Can we measure sexual interest in pedophiles using a sexual distractor task? J Foren Psy. 2016;01:1–2.

 Google Scholar

Jordan K, Fromberger P, Müller I, et al. Sexual interest and sexual self-control in men with self-reported sexual interest in children — a first eye tracking study. J Psychiatr Res. 2018;96:138–144.

 PubMed Web of Science ®Google Scholar

Hall CJ, Todd E, Guo K. Gaze patterns to child figures reflect deviant sexual preference in child sex offenders—a first glance. J Sex Med. 2015;21:303–307.

 Google Scholar