PFI_12mmX178mm.

pdf + eps format

Journal of Child Psychology and Psychiatry 58:8 (2017), pp 913–921 doi:10.1111/jcpp.12724

Affective empathy, cognitive empathy and social

attention in children at high risk of criminal behaviour
Lisette van Zonneveld,1,2 Evelien Platje,1,2 Leo de Sonneville,1,2 Stephanie van Goozen,1,3
and Hanna Swaab1,2
1
Department of Clinical Child and Adolescent Studies, Leiden University, Leiden; 2Leiden Institute for Brain and
Cognition, Leiden, The Netherlands; 3School of Psychology, Cardiff University, Cardiff, UK

Background: Empathy deficits are hypothesized to underlie impairments in social interaction exhibited by those who
engage in antisocial behaviour. Social attention is an essential precursor to empathy; however, no studies have yet
examined social attention in relation to cognitive and affective empathy in those exhibiting antisocial behaviour.
Methods: Participants were 8- to 12-year-old children at high risk of developing criminal behaviour (N = 114, 80.7%
boys) and typically developing controls (N = 43, 72.1% boys). The high-risk children were recruited through an
ongoing early identification and intervention project of the city of Amsterdam, focusing on the underage siblings or
children of delinquents and those failing primary school. Video clips with neutral and emotional content (fear,
happiness and pain) were shown, while heart rate (HR), skin conductance level (SCL) and skin conductance
responses (SCRs) were recorded to measure affective empathy. Answers to questions about emotions in the clips were
coded to measure cognitive empathy. Eye-tracking was used to evaluate visual scanning patterns towards social
relevant cues (eyes and face) in the clips. Results: The high-risk group did not differ from the control group in social
attention and cognitive empathy, but showed reduced HR to pain and fear, and reduced SCL and SCRs to pain.
Conclusions: Children at high risk of developing criminal behaviour show impaired affective empathy but
unimpaired social attention and cognitive empathy. The implications for early identification and intervention
studies with antisocial children are discussed. Keywords: Criminality; antisocial behaviour; empathy; eye gaze;
psychophysiology.

Large cities such as Amsterdam in the Netherlands

Introduction
are confronted by serious criminal problems caused
A small group of children is at high risk of persistent
by groups of severe and persistent offenders, who
antisocial behaviour, including future involvement
come from families, which frequently operate off the
in the criminal justice system. More effective strate-
radar from health and social services. The Preventive
gies for targeting these children at an early sensitive
Intervention Trajectory (PIT) is a project of the
period for intervention may provide crucial oppor-
municipality of the city of Amsterdam that targets
tunities, not only to help these children attain a
children at risk of future criminal behaviour. These
more positive developmental trajectory but also to
children are the underage siblings of young offend-
diminish the enormous negative impact their beha-
ers, have delinquent parents or fail at school because
viour can have on society. Recent reviews of evi-
of severe absenteeism or extreme antisocial beha-
dence from neuroscience (Fairchild, van Goozen,
viour. While these children might have behavioural
Calder, & Goodyer, 2013), clinical science, forensic
problems, they often have no diagnosis, nor do their
psychology and criminology (Skeem, Scott, & Mul-
families actively seek help from clinicians, which
vey, 2014) indicate that high-risk children have
substantially increases the risk of an unfavourable
poorer parental supervision, come from more disad-
social developmental trajectory (Farrington, Piquero,
vantaged neighbourhoods, have greater problems
& Jennings, 2013; Loeber & Stouthamer-Loeber,
with emotional functioning, and exhibit alterations
1998). Building on a theoretical model of the devel-
in brain structure and function compared to other
opment of antisocial behaviour in children that
young people. However, research challenges the
focuses on the mediating role of neurocognitive and
notion that high-risk children inevitably mature into
emotional processes (Van Goozen, Fairchild, Snoek,
adult offenders (Odgers et al., 2007), which raises
& Harold, 2007), the aim of the PIT project is to
the possibility that well-targeted intervention could
actively seek and target these children, assess their
create a turning point in antisocial behaviour for
socioemotional functioning and provide them with
high-risk children. The period between childhood
an opportunity to redirect them onto a more adap-
and early adolescence is a time when children are
tive, prosocial pathway through directed interven-
particularly adept at specific kinds of social and
tions (Van Goozen & Fairchild, 2008).
emotional learning. This creates a window of oppor-
Recognition of others’ emotions and empathy are
tunity for intervention.
learned through experience and based on the grad-
ual refinement with age of children’s production and
Conflict of interest statement: No conflicts declared. recognition of emotional signals (Klinnert & Campos,

© 2017 Association for Child and Adolescent Mental Health.

Published by John Wiley & Sons Ltd, 9600 Garsington Road, Oxford OX4 2DQ, UK and 350 Main St, Malden, MA 02148, USA
914 Lisette van Zonneveld et al. J Child Psychol Psychiatr 2017; 58(8): 913–21

1987). Young children who are good in recognizing for antisocial boys, who are known to have low verbal
other people’s emotions are more socially skilled and IQ and problems with self-reflection, which could
popular (Manstead & Edwards, 1992), but the result in unreliable self-reported affective empathy
reverse process also exists. Children who are (Bowen, Morgan, Moore, & van Goozen, 2014; Tyson,
adversely treated or exposed to aberrant emotional 2005). Similarly, studies that used physiological
signals exhibit a range of emotional difficulties measures to assess affective empathy often did not
(Pollak, Cicchetti, Hornung, & Reed, 2000). There is include measures of cognitive empathy. These affec-
substantial evidence that individuals who engage in tive empathy studies reported that children with
inappropriate interpersonal behaviour, such as disruptive behaviour disorders (De Wied, Boxtel,
aggression or antisocial behaviour, have problems Posthumus, Goudena, & Matthys, 2009; De Wied,
in emotion recognition and empathy (Marsh & Blair, van Boxtel, Matthys, & Meeus, 2012; De Wied, van
2008). The reasoning is that if one cannot correctly Boxtel, Zaalberg, Goudena, & Matthys, 2006), and
identify distress caused to another person, one is children with conduct disorder with and without CU-
more likely to continue with the harmful or distress- traits (Marsh, Beauchaine, & Williams, 2008) dis-
ing behaviour. As it is assumed that empathy deficits played decreased physiological responses and thus
underlie the impairments in social interaction less affective empathy in response to negative emo-
related to antisocial behaviour (Blair, 2005), the tions. With the present study, we extend the existing
aim of the present study was to examine the role of literature by using objective physiological measures
empathy in children at high risk of developing future for affective empathy, combined with both cognitive
criminal behaviour. and affective empathy.
Empathy is distinguished into affective and cogni- In order to understand someone’s emotions and
tive empathy (Singer, 2006). Affective empathy is the respond empathetically, initial attention to socially
capacity of an individual to experience what it feels relevant cues is crucial. From early age on, humans
like for another person to experience a certain have a preference towards social information (Chita-
emotion (e.g. Blair, 2005; De Waal, 2008; Smith, Tegmark, 2016), which can be referred to as social
2009), while cognitive empathy is the capacity of an attention. Faces, in particular the eyes, play a key
individual to understand what others’ emotions and role in providing information about the mental and
thoughts might be, without being emotionally emotional state of another person during social
involved (e.g. Bartoli & Wendt, 2014; Blair, 2005; interaction (Emery, 2000; Klein, Shepherd, & Platt,
Bons et al., 2013; Dadds, El Masry, Wimalaweera, & 2009), and attention to the eyes is considered
Guastella, 2008; De Vignemont & Singer, 2006; necessary for the recognition of facially expressed
Lovett & Sheffield, 2007; Singer, 2006). Several emotions (Bons et al., 2013). Social attention can
studies have examined both affective and cognitive therefore be seen as an essential precursor of an
empathy in children with antisocial behaviour and empathic response. In a previous study, community
found impaired affective empathy but unimpaired children with high CU-traits showed deficits in
cognitive empathy in children with conduct disorder attention to the eyes compared to children low on
and high levels of callous-unemotional (CU) traits CU-traits, particularly for fearful faces (Dadds et al.,
(Anastassiou-Hadjicharalambous & Warden, 2008; 2008). When these children were instructed to
Schwenck et al., 2012), in children with conduct direct their attention to the eyes, their fear recogni-
disorder with or without ADHD (Van Goozen et al., tion was as accurate as that of controls. The authors
2016), in those with conduct problems and high therefore concluded that the fear recognition prob-
levels of CU-traits (Pasalich, Dadds, & Hawes, 2014) lems in children with CU-traits are partly due to a
and in children with psychopathic tendencies or failure in attention towards the eyes (Dadds et al.,
conduct problems recruited from the community 2006).
(Jones, Happ
e, Gilbert, Burnett, & Viding, 2010), in Although previous studies on social attention
line with the empathy imbalance theory of Smith often used static (facial) stimuli, we used stimuli
(2009, 2010). In addition, studies that examined that represent dynamic social situations to evoke an
only cognitive empathy showed normal cognitive empathetic response, making the current design
empathy (Sutton, Reeves, & Keogh, 2000; Wood- more sensitive to examine the role of social attention
worth & Waschbusch, 2008). However, these studies in cognitive and affective empathy (Chevallier et al.,
investigated empathy by using questionnaires and 2015). The current study examined the role of social
failed to measure affective empathy with physiolog- attention and empathy in response to different
ical measures. Although physiological arousal is not emotionally meaningful events in children at high
synonymous with affective empathy, it certainly risk of future criminal behaviour, and predicted in
represents a reliable, objective and direct measure line with Herpers, Scheepers, Bons, Buitelaar, and
of affective empathy (Bons et al., 2013), and has Rommelse (2014) and Dadds et al. (2006) that the
often been related to antisocial behaviour (e.g. Gao, high-risk children would have impaired social atten-
Raine, Venables, Dawson, & Mednick, 2010; Van tion and affective empathy, but unimpaired cogni-
Goozen, 2015). Furthermore, verbal reports of one’s tive empathy compared to typically developing
own experienced emotion(-s) are difficult, especially controls.

© 2017 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12724 Attention and empathy in at-risk children 915

Methods duration for each clip separately. These percentages of total

fixation were used as indices of social attention.
Participants
Data were gathered from children recruited through the PIT. Affective empathy. Affective arousal was operationalized
This is a large ongoing project of the municipality of the city of as the contrast between baseline and the emotional clips in
Amsterdam, the Netherlands. Participants were the underage cardiovascular and electrodermal activity. Electrodes to mea-
siblings of young offenders, children of delinquent parents or sure electrodermal activity were placed on the middle phalanges
children who fail at school due to severe unauthorized absen- of the ring finger and index finger of the nondominant hand.
teeism (e.g. truancy) or because of extreme antisocial beha- Before the electrodes were attached, the participants were asked
viour. The total sample consisted of 157 children (123 boys to wash their hands carefully. Two electrodes measuring cardio-
and 34 girls) with a mean age of 10.37 years (SD = 1.35). The vascular activity were placed on the chest and left ribs under-
high-risk group consisted of 114 children (92 boys and 22 neath the clothes. Participants were instructed not to move or
girls) with a mean age of 10.40 years (SD = 1.38). The control talk during the measurement. Physiological data were recorded
group [N = 43; 31 boys and 12 girls; mean age of 10.27 years continuously with AcqKnowledge software version 4.3.1. (BIO-
(SD = 1.29)] was recruited through the same schools that were PAC Systems Inc., Goleta, CA, USA). Recordings were acquired
attended by the children in the high-risk group. The Dutch through a galvanic skin response amplifier (GSR100C), electro-
version of the Teacher Report Form (TRF; Achenbach & cardiogram amplifier (ECG100C) and a BIOPAC data acquisition
Rescorla, 2001; Verhulst, Van der Ende, & Koot, 1997) was system (MP150 Windows) with a sampling rate of 200 Hz. The
used to confirm risk status of the participants; all high-risk physiological monitoring equipment was synchronized with
children scored in the borderline or clinical range on the the Tobii software by manually assigned event markers repre-
aggression and/or rule-breaking behaviour scales (t-score senting the start and end of each clip. In AcqKnowledge, a 0.5 Hz
≥65); and their average internalizing problem behaviour score highpass filter and a 50 Hz notch filter were applied to stabilize
was in the normal range. All control children scored within the the cardiovascular signal. The recorded physiological data were
normal range on all problem scales (t-score <65). The Dutch further processed with a script in MATLAB Release 2012b (The
version of the Child Behaviour Checklist (CBCL; Achenbach & MathWorks, Inc., Natick, MA). A forward and reverse first order
Rescorla, 2001; Verhulst, Van der Ende, & Koot, 1996) was lowpass digital Butterworth filter with a cut-off frequency of
used to identify the problem behaviour reported by the parents 0.33 Hz was used on the raw electrodermal signal, which
of the high-risk group. removed all high frequency noise while ensuring precisely zero
Children were eligible to participate if they were between 8 phase distortion. To detect skin conductance responses (SCRs),
and 13 years old and spoke and understood the Dutch a phasic channel was created by filtering the tonic channel
language. Other exclusion criteria were the use of stimulants using a zero-phase forward and reverse digital infinite impulse
and a known DSM classification; based on these criteria, three response filter with a cut-off frequency of one. Each section on
participants were excluded. Written informed consent was the phasic channel that lies above the threshold was considered
obtained from the parents and from the children if they were 12 a SCR. The SCR peak was located at the maximum value of
years or older. Ethics approval for this study was obtained this interval, as measured on the tonic channel (Boucsein, 1992;
from Leiden University’s Education and Child Studies Ethics Society for Psychophysiological Research Ad Hoc Committee on
Committee. Electrodermal Measures, 2012). Motion artefacts were visually
identified and excluded from the data. We used heart rate (HR)
as cardiovascular response variable and skin conductance
Procedure level (SCL) and skin conductance responses (SCRs) as electro-
dermal response variables.
Following informed consent, an appointment was made at
school, where the tests were administered following a standard
protocol. All participants were individually assessed in a quiet Cognitive empathy. After each clip, participants were
room. The assessors were two trained graduate students under asked questions concerning the type and intensity of the
supervision of a clinical investigator (LvZ). emotions of the main character in the clip and the reasons for
these emotions. The answers were scored with a coding system
taking into account four elements of empathy: (a) the correct
Instruments target emotion, (b) correct similar and relevant other emotion(-
s), (c) the intensity of the emotion(-s) and (d) the explanation for
Stimuli. We showed four video clips: a neutral baseline clip the causes of the emotion (for details see, Van Goozen et al.,
(180 s) displayed an aquarium with fishes (Coral Sea Dream- 2016; Van Rijn et al., 2014). Answers were independently
ing, Small World Music Inc.) to obtain baseline cardiovascular coded by two coders; the interrater agreement was 73.1%.
and electrodermal activity; three emotional clips with different
emotional contents were presented in random order. The target
emotions in these clips were fear (103 s), happiness (87 s) and
Intellectual functioning. Intellectual functioning was
assessed with the Dutch version of the Wechsler Intelligence
pain (101 s). The three clips were derived from existing movies
Scale for Children (WISC-III; Kort et al., 2005). Two subtests,
(for details, see Van Rijn, Barendse, van Goozen, & Swaab,
Block Design (perceptual organization skills) and Vocabulary
2014).
(verbal skills), were used to estimate full scale IQ (FSIQ;
Campbell, 1998).
Social attention. Social attention was assessed by means
of visual scanning patterns towards social relevant cues, that
is eyes and face. The visual scanning patterns were measured Statistical analyses
with a Tobii T120 eye-tracker (Tobii Technology, Sweden) using
the I-VT fixation filter. After a 5-point calibration procedure, There were no outliers or violations of statistical assumptions.
the clips were shown on a LCD screen, placed at approximately Due to technical difficulties, HR data were not available for one
65 cm distance from the participant. In Tobii Studio (version participant, SCL and SCR data were not available for 54
3.0.2.), dynamic areas of interest (AOIs) were drawn by hand in participants (44 high risk; no differences in aggression
all clips for the eyes and face (without the eyes) (Hessels, (p = .638) or rule-breaking behaviour (p = .472) within the
Kemner, Boomen, & Hooge, 2015). To control for slight high-risk group between those with and without data), and
differences in AOI durations between clips, total fixation eye-tracking data were unavailable for five participants.
duration was computed as the percentage of the total AOI A priori, the high-risk and control groups were compared on

© 2017 Association for Child and Adolescent Mental Health.

916 Lisette van Zonneveld et al. J Child Psychol Psychiatr 2017; 58(8): 913–21

age, gender and intellectual functioning. We next examined (.30 < p < .82). In Figure 1, the means and standard
total fixation duration on the total screen to control for errors of measurement (SEM) are shown for the
potential differences in attention. To analyse group differences,
percentages of total fixation as a function of Group,
we performed a two-way repeated measures analysis of vari-
ance (RM-ANOVA) with AOI (eyes, face) and Emotion (fear, AOI and Emotion. The RM-ANOVA results revealed no
happiness, pain) as within-subject factors and Group as main effect of Group on social attention (p = .527);
between-subjects factor. Subsequently, we performed three however, there was a significant effect of Emotion
RM- ANOVAs to investigate differences between groups in HR, (F(2,300) = 358.43, p < .001, ƞp2 = .823), AOI (F(1,150) =
SCL and SCR, respectively, in response to the emotion clips,
290.92, p < .001, ƞp2 = .66) and a significant Emotion
with Emotion (fear, happiness, pain) as within-subject factor
and Group as between-subjects factor. A simple contrast was by AOI interaction (F(2,300) = 137.36, p < .001, ƞp2 =
used with baseline as reference for each of the three emotions. .662), indicating that differences in fixation duration
Post hoc group differences in baseline were examined. Lastly, between eyes and face were largest for the negative
we compared mean cognitive empathy scores for the two emotions (Figure 1). No significant Group by Emotion
groups with a MANOVA. Significance level was set at a < .05. A
(p = .135), or Group by AOI (p = .152) interactions were
false discovery rate (FDR) control as described by Glickman,
Rao, and Schultz (2014) to correct for multiple testing was found.
used. Effect sizes were calculated using partial eta squared
(ƞp2) with ƞp2 ~ .03 representing a small effect, ƞp2 ~ .06
representing a moderate effect, and ƞp2 ≥ .14 a large effect Affective empathy
(Cohen, 1992).
There were no group differences in HR (p = .431),
SCL (p = .135), or SCR (p = .087) at baseline. With
regard to HR, there was no effect of Group (p = .971),
Results but there was a significant main effect of Emotion
Descriptive statistics (F(3,462) = 8.37, p < .001, ƞp2 = .052) and a signif-
Descriptive data for gender, age, FSIQ, externalizing icant Emotion by Group interaction (F(3,462) = 5.08,
problem behaviour and internalizing problem beha- p = .003, ƞp2 = .032). Subsequent simple contrasts
viour are shown in Table 1. The high-risk and showed a significant Emotion by Group interaction
control groups did not differ in age or gender, but effect for fear (F(1,154) = 7.70, p = .006, ƞp2 = .048)
the high-risk group had a significantly lower esti- and pain (F(1,154) = 9.62, p = .002, ƞp2 = .059), but
mated FSIQ; the high-risk group also scored signif- not for happiness (p = .023). Figure 2 illustrates that
icantly higher on TRF aggression, rule-breaking HR increased during emotion exposure in the control
behaviour, total externalizing behaviour and total group, whereas it decreased in the high-risk group.
internalizing behaviour (Table 1). As expected, With regard to SCL, the results showed no main
parents of the high-risk children reported less effect of Group (p = .655), but a significant effect of
problem behaviour (Maggression = 58.81, SD = 8.80; Emotion (F(3,291) = 134,10, p < .001, ƞp2 = .58), and
Mrule-breaking = 58.30, SD = 6.85) compared to a significant Emotion by Group interaction
teachers for aggression (t(1,112) = 14.21, p < .001, (F(3,291) = 7.86, p = .001, ƞp2 = .075). Subsequent
d = 1.7) and rule-breaking behaviour (t(1,112) = simple contrasts showed a significant Emotion by
13.13, p < .001, d = 1.6). Because IQ was not corre- Group interaction effect for pain (F(1,97) = 17.37,
lated with any of the social attention or empathy p < .001, ƞp2 = .152), reflecting a smaller increase in
variables, IQ was not included as a covariate in SCL during the pain clip in the high-risk group than in
subsequent analyses. the control group, but no such pattern was observed
for fear (p = .086) or happiness (p = .105).
With regard to SCR, the results showed no main effect
Social attention
of Group (p = .492), but a significant effect of Emotion
First, we examined the total fixation duration to the (F(3,291) = 130.90, p < .001, ƞp2 = .574), and a signif-
total screen, controlled for the duration of the clips. icant Emotion by Group interaction (F(3,291) = 3.69,
The groups did not differ in attention to the total screen p = .012, ƞp2 = .037). Subsequent simple contrasts

Table 1 Descriptive statistics for gender, age, FSIQ, aggression, rule-breaking behaviour, total externalizing behaviour and total
internalizing behaviour for the two groups

Risk group Control group

M SD M SD t/v2 test p

Gender (% boys) 80.7% 72.1% v2 (1,157) = 1.36 .243

Age (years) 10.40 1.38 10.27 1.29 t (155) = 0.53 .598
FSIQ 83.01 12.54 104.45 15.43 t (155) = 8.95 <.001
Aggression TRF (t-score) 75.64 (61.4%) 10.77 52.51 (0.0%) 4.35 t (1,155) = 19.17 <.001
Rule-breaking TRF (t-score) 69.98 (50.8%) 7.54 52.33 (2.33%) 3.90 t (1,155) = 19.12 <.001
Total Internalizing (t-score) 61.09 (43.9%) 7.76 49.84 (4.65%) 8.14 t (1,155) = 7.99 <.001
Total externalizing (t-score) 73.51 (93.0%) 7.58 48.49 (2.33%) 7.09 t (1,155) = 18.76 <.001

Between the brackets are the percentages of the children within the clinical range displayed.

© 2017 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12724 Attention and empathy in at-risk children 917

showed a significant Emotion by Group interaction High-risk group Control group

Percentage differences between

2
effect for pain (F(1,97) = 5.79, p = .018, ƞp2 = .056),
but not for fear (p = .234) or happiness (p = .916),

baseline and emotion

* *
1
reflecting fewer SCRs during the pain clip in the high-
risk group than in the control group.
0

Cognitive empathy –1

Mean cognitive empathy scores for the two groups

–2
and the three emotions are shown in Table 2. No Fear Happiness Pain
group differences were found (p = .887).
Figure 2 The differences in percentage of heart rate responses
for the three clips versus baselines for the two groups. *p < .01.
For visualization purposes the differences in percentage between
Discussion the emotions and baselines are shown
A small group of children is at high risk of persistent
antisocial behaviour and a criminal career. Early in contrast to some studies, for example Dadds et al.
intervention may provide crucial opportunities to (2008, 2006) who reported that fear recognition
prevent the detrimental effects on the children problems in children with CU-traits are partly due to
themselves as well as on society (Van Goozen, a failure in attention towards the eyes. Correcting
2015). The current study is embedded within the attention towards the eyes subsequently seemed to
PIT, a project that targets children at risk of future correct some of the recognition impairments (Dadds
criminal behaviour. The current study specifically et al., 2006). It is possible that the attentional mech-
focused on empathy deficits underlying the impair- anisms underlying CU-traits differ from those under-
ments in social interaction related to antisocial lying aggressive, antisocial or criminal behaviour, and
behaviour (Blair, 2005). The results show that the future research would need to examine this. However,
high-risk children had impaired affective empathy we found no deviance in social attention and this
when viewing emotional clips. Specifically, they suggests that the type of training for children with CU-
showed reduced HR to pain and fear, and reduced traits, as recommended by Dadds et al. (2006), would
SCL and SCRs to pain compared to controls. How- not necessarily benefit all children with conduct
ever, importantly, they did not differ from controls on problems and/or who are at high risk of future
measures of social attention and cognitive empathy, criminal behaviour. These results are in need of
nor in affective empathy to happiness. replication. Future studies should also include chil-
To the authors’ knowledge, this is the first study dren at high risk of developing criminal behaviour and
that examined the role of social attention using eye- assess their levels of CU-traits to cross-validate these
tracking methodology and empathy in response to findings.
different emotionally meaningful events in a group of In line with previous studies on cognitive and
children at high risk of developing criminal behaviour. affective empathy (see review; Herpers et al., 2014),
Although social attention to emotionally charged and Smith’s empathy imbalance theory (Smith,
events is required to ensure an empathic response, 2009, 2010), our results revealed significant differ-
we did not find evidence of impaired social attention in ences in affective empathy but not in cognitive
the high-risk group, suggesting that social attention empathy. These results indicate that high-risk chil-
does not account for the deficits in affective empathy dren showed adequate recognition and good under-
that we observed in these children. These results are standing of the emotions presented in the video clips

High-risk group Control group

Percentages of total fixation (%)

60

50

40

30

20

10

0
Eyes fear Face fear Eyes Face Eyes pain Face pain
happiness happiness
AOIs by Emotion

Figure 1 Percentages of total fixation (M, SEM) for areas of interest and emotion for the high-risk group and control group

© 2017 Association for Child and Adolescent Mental Health.

918 Lisette van Zonneveld et al. J Child Psychol Psychiatr 2017; 58(8): 913–21

Table 2 Means, standard deviations and ANOVAs for the three emotions

High-risk group (N = 114) Control group (N = 42)

M (SD) Range M (SD) Range F p

Fear 4.70 (1.52) 0–7 4.50 (1.76) 0–7 (1,155) = .46 .483
Happiness 4.75 (1.13) 0–7 4.64 (1.19) 0–6 (1,155) = .29 .589
Pain 4.82 (1.53) 0–8 4.79 (1.42) 2–8 (1,155) = .02 .886

Maximum range was 0–9. Scores were normally distributed and there were no ceiling effects, skewness ranged between 0.4 and
1.5.

but had specific problems with empathizing Our sample consisted of children who are at high
and experiencing others’ negative emotions. These risk of developing criminal behaviour. The high risk
results in combination with our findings on unim- consists of the severity of their externalizing beha-
paired social attention suggests that impaired affec- viour as reported by their teachers in combination
tive empathy is the key empathy component that is with the parental ignorance of these problems as they
related to antisocial behaviour (Blair, 2005). reported their child’s behaviour to be in the normal
Elaborating on these results, when a child does range, which might explain why they did not actively
not empathize with the distress caused by their seek help. Teachers are often considered more reli-
aggressive behaviour, they are more likely to con- able informants; their report of children’s behaviour
tinue the display of harmful behaviour (Marsh & is more objective and they can compare the beha-
Blair, 2008). Our results show that the affective viour of each child against that of many others. The
response was significantly smaller for high-risk severe behavioural problems of these children as
children, in particular in response to seeing some- reported by the teachers, in combination with their
one else in pain or fear. Problems in affective parents’ unawareness of these problems, could neg-
empathy in children with CD (Van Goozen et al., atively impact their future social development (Van
2016) or psychopathic traits in response to negative Goozen et al., 2007). Early identification of these
emotions have been observed before (Lockwood, children is crucial in order to provide tailored inter-
Bird, Bridge, & Viding, 2013). However, we ventions to prevent them from drifting towards a
observed similar deficits in children who are not criminal career (Van Goozen & Fairchild, 2008).
psychopathic and also do not have a diagnosis of The findings of the current study indicate that the
CD, showing that these affective processes could empathy impairment that presumably plays a signif-
play a role in a much larger range of problem icant role in antisocial development is primarily a
behaviours. Future research would benefit from deficit in affective empathy rather than in social
incorporating not only social attention and physio- attention or cognitive empathy. This has implications
logical assessments but also fMRI, to investigate the for the development of interventions, which focus
possible relation between functional brain networks specifically on enhancing emotional awareness and
and affective empathy. This would eventually affective empathy. For example, emotion awareness
increase our insight in the underlying brain mech- programs in clinical samples of aggressive children
anisms of empathy. (Van Baardewijk, Stegge, Bushman, & Vermeiren,
The study also had several limitations. First, we 2009) and young offenders (Hubble, Bowen, Moore, &
did not assess motor empathy, which precluded van Goozen, 2015) have been found to be successful
obtaining information about the ability to express in attenuating aggression or severity of crimes com-
empathic facial reactions. As this is an important mitted. Moreover, there are preliminary indications
component of empathy (Van der Graaff et al., 2015), that empathy and compassion training result in
information on motor empathy could confirm the increased affective response and functional activity
finding that affective empathy is the key empathy in brain areas involved in emotion processing (Kli-
component related to antisocial behaviour. Future mecki, Leiberg, Ricard, & Singer, 2014). Programs
research on high-risk samples should therefore aim that target an increase in emotion awareness could
to incorporate this measure of empathy. Another thus be an important component in future interven-
limitation is the loss of data on electrodermal activity tion and prevention research.
due to technical difficulties. However, it was verified
that the data loss was random, and that participants
without electrodermal data did not differ from those Conclusion
for whom data were available on key outcome This study found impaired affective empathy but
measures. Third, we were not able to include a unimpaired social attention and cognitive empathy
questionnaire measure of CU-traits, nor did we in a sample of children considered to be at high risk for
obtain information about affective empathy using a future criminal behaviour because they are the under-
self-report measure. Future studies should aim to age siblings of young offenders, they have delinquent
include these self-report measures. parents or fail at school because of severe absenteeism

© 2017 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12724 Attention and empathy in at-risk children 919

or extreme antisocial behaviour. The findings high- Preventief Interventie Team (PIT), and particularly Ros-
light not only the important role of emotion function, aly Brandon, in facilitating and funding our research.
specifically reduced affective response, in the devel- They had no role in the study design, the collection,
opment of antisocial behaviour, but also suggest that analysis or interpretation of the data, or the writing of
interventions should directly target these affective the manuscript.
processes in order to influence the development of
behaviour in a more prosocial direction.
Correspondence
Lisette van Zonneveld, Faculty of Social and Beha-
Acknowledgements vioural Sciences, PO Box 9555, 2300 RB Leiden,
The authors gratefully acknowledge the municipality of The Netherlands; Email: e.m.van.zonneveld@fsw.leide
the city of Amsterdam, RVE Onderwijs, Jeugd en Zorg/ nuniv.nl

Key points
• Deficits in empathy are hypothesized to play a key role in the impairments in social interaction shown by those
who engage in antisocial behaviour. Although evidence highlights the role of affective and/or cognitive
empathy in antisocial development, its precursor – social attention – has not yet been investigated.
• The role of social attention, and affective and cognitive empathy was studied in a group of children at high
risk of developing future criminal behaviour.
• Findings indicate a specific deficit in affective empathy for negative emotions, but no impairment in social
attention, cognitive empathy or affective empathy for happiness.
• Interventions aimed at preventing future problem behaviour should focus on enhancing emotional awareness
and/or affective empathy.

spectrum disorder using eye-tracking: Stimulus type mat-

References ters. Autism Research, 8, 620–628.
Achenbach, T.M., & Rescorla, L.A. (2001). Manual for the Chita-Tegmark, M. (2016). Social attention in ASD: A review
ASEBA school-age forms & profiles. Burlington, VT: Univer- and meta-analysis of eye-tracking studies. Research in
sity of Vermont, Research Center for Children, Youth, and Developmental Disabilities, 48, 79–93.
Families. Cohen, J. (1992). A power primer. Psychological Bulletin, 112,
Anastassiou-Hadjicharalambous, X., & Warden, D. (2008). 155.
Cognitive and affective perspective-taking in conduct-dis- Dadds, M.R., El Masry, Y., Wimalaweera, S., & Guastella, A.J.
ordered children high and low on callous-unemotional traits. (2008). Reduced eye gaze explains “fear blindness” in
Child and Adolescent Psychiatry and Mental Health, 2, 1–11. childhood psychopathic traits. Journal of the American
Bartoli, A.P.J., & Wendt, G.W. (2014). Neuropsychology of Academy of Child and Adolescent Psychiatry, 47, 455–463.
psychopathic traits in children. The Routledge International Dadds, M.R., Perry, Y., Hawes, D.J., Merz, S., Riddell, A.C.,
Handbook of Biosocial Criminology, 218, 218–235. Haines, D.J., . . . & Abeygunawardane, A.I. (2006). Attention
Blair, R.J.R. (2005). Responding to the emotions of others: to the eyes and fear-recognition deficits in child psychopa-
Dissociating forms of empathy through the study of typical thy. The British Journal of Psychiatry, 189, 280–281.
and psychiatric populations. Consciousness and Cognition, De Vignemont, F., & Singer, T. (2006). The empathic brain: How,
14, 698–718. when and why? Trends in Cognitive Sciences, 10, 435–441.
Bons, D., van den Broek, E., Scheepers, F., Herpers, P., De Waal, F.B. (2008). Putting the altruism back into altruism:
Rommelse, N., & Buitelaaar, J. (2013). Motor, emotional, The evolution of empathy. Annual Review of Psychology, 59,
and cognitive empathy in children and adolescents with 279–300.
autism spectrum disorder and conduct disorder. Journal of De Wied, M., Boxtel, A.V., Posthumus, J.A., Goudena, P.P., &
Abnormal Child Psychology, 41, 425–443. Matthys, W. (2009). Facial EMG and heart rate responses to
Boucsein, W. (1992). Electrochemical activity. New York: emotion-inducing film clips in boys with disruptive behavior
Plenum Press. disorders. Psychophysiology, 46, 996–1004.
Bowen, K.L., Morgan, J.E., Moore, S.C., & van Goozen, S.H. De Wied, M., van Boxtel, A., Matthys, W., & Meeus, W. (2012).
(2014). Young offenders’ emotion recognition dysfunction Verbal, facial and autonomic responses to empathy-eliciting
across emotion intensities: Explaining variation using psy- film clips by disruptive male adolescents with high versus
chopathic traits, conduct disorder and offense severity. low callous-unemotional traits. Journal of Abnormal Child
Journal of Psychopathology and Behavioral Assessment, Psychology, 40, 211–223.
36, 60–73. De Wied, M., van Boxtel, A., Zaalberg, R., Goudena, P.P., &
Campbell, J.M. (1998). Internal and external validity of seven Matthys, W. (2006). Facial EMG responses to dynamic
Wechsler Intelligence Scale for Children - Third Edition short emotional facial expressions in boys with disruptive
forms in a sample of inpatients. Psychological Assessment, behavior disorders. Journal of Psychiatric research, 40,
10, 431–434. 112–121.
Chevallier, C., Parish-Morris, J., McVey, A., Rump, K.M., Emery, N.J. (2000). The eyes have it: The neuroethology,
Sasson, N.J., Herrington, J.D., & Schultz, R.T. (2015). function and evolution of social gaze. Neuroscience and
Measuring social attention and motivation in autism Biobehavioral Reviews, 24, 581–604.

© 2017 Association for Child and Adolescent Mental Health.

920 Lisette van Zonneveld et al. J Child Psychol Psychiatr 2017; 58(8): 913–21

Fairchild, G., van Goozen, S.H.M., Calder, A.J., & Goodyer, Odgers, C.L., Caspi, A., Broadbent, J.M., Dickson, N., Han-
I.M. (2013). Research review: Evaluating and reformulating cox, R.J., Harrington, H., . . . & Moffitt, T.E. (2007). Predic-
the developmental taxonomic theory of antisocial beha- tion of differential adult health burden by conduct problem
viour. Journal of Child Psychology and Psychiatry, 54, subtypes in males. Archives of General Psychiatry, 64,
924–940. 476–484.
Farrington, D., Piquero, A.R., & Jennings, W.G. (2013). Pasalich, D.S., Dadds, M.R., & Hawes, D.J. (2014). Cognitive
Offending from childhood to late middle age: Recent results and affective empathy in children with conduct problems:
from the Cambridge study in delinquent development. New Additive and interactive effects of callous–unemotional traits
York: Springer Science & Business Media. and autism spectrum disorders symptoms. Psychiatry
Gao, Y., Raine, A., Venables, P.H., Dawson, M.E., & Mednick, Research, 219, 625–630.
S.A. (2010). Reduced electrodermal fear conditioning from Pollak, S.D., Cicchetti, D., Hornung, K., & Reed, A. (2000).
ages 3 to 8 years is associated with aggressive behavior at Recognizing emotion in faces: Developmental effects of child
age 8 years. Journal of Child Psychology and Psychiatry, 51, abuse and neglect. Developmental Psychology, 36, 679.
550–558. Schwenck, C., Mergenthaler, J., Keller, K., Zech, J., Salehi, S.,
Glickman, M.E., Rao, S.R., & Schultz, M.R. (2014). False Taurines, R., . . . & Warnke, A. (2012). Empathy in children
discovery rate control is a recommended alternative to with autism and conduct disorder: Group-specific profiles
Bonferroni-type adjustments in health studies. Journal of and developmental aspects. Journal of Child Psychology and
Clinical Epidemiology, 67, 850–857. Psychiatry, 53, 651–659.
Herpers, P.C.M., Scheepers, F.E., Bons, D.M.A., Buitelaar, Singer, T. (2006). The neuronal basis and ontogeny of empathy
J.K., & Rommelse, N.N.J. (2014). The cognitive and neural and mind reading: Review of literature and implications for
correlates of psychopathy and especially callous–unemo- future research. Neuroscience and Biobehavioral Reviews,
tional traits in youths: A systematic review of the evidence. 30, 855–863.
Development and Psychopathology, 26, 245–273. Skeem, J.L., Scott, E.S., & Mulvey, E.P. (2014). Justice policy
Hessels, R.S., Kemner, C., Boomen, C., & Hooge, I.T.C. (2015). reform for high-risk juveniles: Using science to achieve
The area-of-interest problem in eyetracking research: A large-scale crime reduction. Annual Review of Clinical Psy-
noise-robust solution for face and sparse stimuli. Behavior chology, 10, 709–739.
Research Methods, 48, 1–19. Smith, A. (2009). The empathy imbalance hypothesis of
Hubble, K., Bowen, K.L., Moore, S.C., & van Goozen, S.H.M. autism: A theoretical approach to cognitive and emotional
(2015). Improving negative emotion recognition in young empathy in autistic development. The Psychological Record,
offenders reduces subsequent crime. PLoS ONE, 10, 59, 489–510.
e0132035. Smith, A. (2010). Cognitive empathy and emotional empathy in
Jones, A.P., Happ
e, F.G.E., Gilbert, F., Burnett, S., & Viding, human behavior and evolution. The Psychological Record,
E. (2010). Feeling, caring, knowing: Different types of 56, 1.
empathy deficit in boys with psychopathic tendencies and Society for Psychophysiological Research Ad Hoc Committee
autism spectrum disorder. Journal of Child Psychology and on Electrodermal Measures (2012). Publication recommen-
Psychiatry, 51, 1188–1197. dations for electrodermal measurements. Psychophysiology,
Klein, J.T., Shepherd, S.V., & Platt, M.L. (2009). Social 49, 1017–1034.
attention and the brain. Current Biology, 19, R958–R962. Sutton, J., Reeves, M., & Keogh, E. (2000). Disruptive
Klimecki, O.M., Leiberg, S., Ricard, M., & Singer, T. (2014). behaviour, avoidance of responsibility and theory of mind.
Differential pattern of functional brain plasticity after com- British Journal of Developmental Psychology, 18, 1–11.
passion and empathy training. Social Cognitive and Affective Tyson, P. (2005). Affects, agency, and self-regulation: Com-
Neuroscience, 9, 873–879. plexity theory in the treatment of children with anxiety and
Klinnert, M.D., & Campos, J.J. (1987). Social referencing: The disruptive behavior disorders. Journal of the American
infant’s use of emotional signals from a friendly. Annual Psychoanalytic Association, 53, 159–187.
Progress in Child Psychiatry and Child Development, 1987, 26. Van Baardewijk, Y., Stegge, H., Bushman, B.J., & Vermeiren,
Kort, W., Schittekatte, M., Dekker, P.H., Verhaeghe, P., Com- R. (2009). Psychopathic traits, victim distress and aggres-
paan, E.L., Bosmans, M., & Vermeir, G. (2005). Handleiding sion in children. Journal of Child Psychology and Psychiatry,
en verantwoording WISC-IIINL. London: Pearson Assess- 50, 718–725.
ment. Van der Graaff, J., Meeus, W., de Wied, M., van Boxtel, A., van
Lockwood, P.L., Bird, G., Bridge, M., & Viding, E. (2013). Lier, P.A., Koot, H.M., & Branje, S. (2015). Motor, affective
Dissecting empathy: High levels of psychopathic and autistic and cognitive empathy in adolescence: Interrelations
traits are characterized by difficulties in different social between facial electromyography and self-reported trait
information processing domains. Frontiers in Human Neuro- and state measures. Cognition and Emotion, 1–17.
science, 7, 760. Van Goozen, S.H.M. (2015). The role of early emotion impair-
Loeber, R., & Stouthamer-Loeber, M. (1998). Development of ments in the development of persistent antisocial behavior.
juvenile aggression and violence -Some common miscon- Child Development Perspectives, 9, 206–210.
ceptions and controversies. American Psychologist, 53, 242– Van Goozen, S.H.M., & Fairchild, G. (2008). How can the study
259. of biological processes help design new interventions for
Lovett, B.J., & Sheffield, R.A. (2007). Affective empathy deficits children with severe antisocial behavior? Development and
in aggressive children and adolescents: A critical review. Psychopathology, 20, 941–973.
Clinical Psychology Review, 27, 1–13. Van Goozen, S.H.M., Fairchild, G., Snoek, H., & Harold, G.T.
Manstead, A.S., & Edwards, R. (1992). Communicative aspects (2007). The evidence for a neurobiological model of child-
of children’s emotional competence. International Review of hood antisocial behavior. Psychological Bulletin, 133, 149–
Studies on Emotion, 2, 167–196. 182.
Marsh, P., Beauchaine, T.P., & Williams, B. (2008). Dissocia- Van Goozen, S.H., Langley, K., Northover, C., Hubble, K.,
tion of sad facial expressions and autonomic nervous system Rubia, K., Schepman, K., . . . & Thapar, A. (2016). Identifying
responding in boys with disruptive behavior disorders. mechanisms that underlie links between COMT genotype
Psychophysiology, 45, 100–110. and aggression in male adolescents with ADHD. Journal of
Marsh, A.A., & Blair, R.J.R. (2008). Deficits in facial affect Child Psychology and Psychiatry, 57, 472–480.
recognition among antisocial populations: A meta-analysis. Van Rijn, S., Barendse, M., van Goozen, S.H.M., & Swaab, H.
Neuroscience and Biobehavioral Reviews, 32, 454–465. (2014). Social attention, affective arousal and empathy in

© 2017 Association for Child and Adolescent Mental Health.

doi:10.1111/jcpp.12724 Attention and empathy in at-risk children 921

men with klinefelter syndrome (47, XXY): Evidence from Afdeling Kinder- en Jeugdpsychiatrie, Sophia Kinderzieken-
eyetracking and skin conductance. PLoS ONE, 9, e84721. huis, Erasmus MC.
Verhulst, F.C., Van der Ende, J., & Koot, H.M. (1996). Woodworth, M., & Waschbusch, D. (2008). Emotional pro-
Handleiding voor de CBCL/4–18 [Manual for the Child cessing in children with conduct problems and callous/
Behavior Checklist]. Rotterdam, the Netherlands: Afdeling unemotional traits. Child: Care, Health and Development,
Kinder- en Jeugdpsychiatrie, Sophia Kinderziekenhuis, 34, 234–244.
Erasmus MC.
Verhulst, F.C., Van der Ende, J., & Koot, H.M. (1997). Accepted for publication: 16 February 2017
Handleiding voor de Teachers’ Report Form [Manual for First published online: 11 April 2017
the Teachers’ Report Form]. Rotterdam, the Netherlands:

© 2017 Association for Child and Adolescent Mental Health.

Copyright of Journal of Child Psychology & Psychiatry is the property of Wiley-Blackwell
and its content may not be copied or emailed to multiple sites or posted to a listserv without
the copyright holder's express written permission. However, users may print, download, or
email articles for individual use.