See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/319701198

Abnormal emotional processing in maltreated

children diagnosed of Complex Posttraumatic
Stress Disorder

Article in Child Abuse & Neglect September 2017

DOI: 10.1016/j.chiabu.2017.09.020

CITATIONS READS

0 85

6 authors, including:

Maite Ferrn Lorenzo Livianos

University of Southampton Hospital Universitari i Politcnic la Fe
56 PUBLICATIONS 973 CITATIONS 59 PUBLICATIONS 626 CITATIONS

SEE PROFILE SEE PROFILE

Luis Rojo Moreno Ana C. Garca-Blanco

University of Valencia Instituto de Investigacin Sanitaria La Fe
132 PUBLICATIONS 756 CITATIONS 30 PUBLICATIONS 108 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Emotional processing in bipolar disorder View project

First episode psychosis and substance use View project

All content following this page was uploaded by Ana C. Garca-Blanco on 22 September 2017.

The user has requested enhancement of the downloaded file.

 Child Abuse & Neglect 73 (2017) 4250

Contents lists available at ScienceDirect

Child Abuse & Neglect

journal homepage: www.elsevier.com/locate/chiabuneg

Full Length Article

Abnormal emotional processing in maltreated children diagnosed MARK

of Complex Posttraumatic Stress Disorder
Clara Berta, Maite Ferrinb,c, Mara Barbera,d, Lorenzo Livianosa,d, Luis Rojoa,d,

Ana Garca-Blancoe,f,
a
Department of Psychiatry and Clinical Psychology, University and Polytechnic Hospital La Fe, Valencia, Spain
b
University of Southampton, UK
c
Institute of Psychiatry, London, UK
d
Department of Medicine, University of Valencia, Valencia, Spain
e
Neonatal Research Unit, Health Research Institute La Fe, Valencia, Spain
f
Department of Personality, Evaluation and Psychological Treatment, University of Valencia, Valencia, Spain

AR TI CLE I NF O AB S T R A CT

Keywords: Maltreated children usually show a specic pattern of emotional and behavioral symptoms that
Maltreatment exceed those relating to posttraumatic stress disorder (PTSD). These symptoms have been dened
Emotion regulation as Complex PTSD (CPTSD). The underlying attentional mechanisms of abnormal emotional
Attention bias processing and their relation to the clinical presentation of CPTSD are not well understood. A
Complex posttraumatic stress disorder
visual dot-probe paradigm involving pre-attentive (i.e., 500 ms) and attentive (i.e., 1500 ms)
Depression
presentation rates of neutral versus emotional (i.e., angry, happy or sad) facial expressions was
applied. Twenty-one maltreated CPTSD children were compared with twenty-six controls. The
results are as follows: an attention bias away from threatening faces and an attentional bias
towards sad faces were observed in maltreated CPTSD children during pre-attentive and attentive
processing. Whereas the attentional bias away from angry faces was associated with social
problems, the attentional bias towards sad faces was associated with depressive and withdrawn
symptoms. Therefore, CPTSD children develop maladaptive negative cognitive styles, which may
underlie not only social problems (by a cognitive avoidance of threatening stimuli) but also
depressive symptoms (by a cognitive approach to sad stimuli). Attention processing abnormal-
ities should be considered as therapeutic targets for new treatment approaches in this population.

1. Introduction

Children exposed to repeated, prolonged, or multiple ways of interpersonal trauma, often occurring under circumstances in which
escape is not possible, commonly develop more complex and serious symptoms that exceed normal Posttraumatic Stress Disorder
(PTSD) (Cloitre et al., 2012). Indeed, the eleventh version of the International Classication of Diseases (ICD-11) has proposed that
Complex PTSD (CPTSD) captures the severe psychological damage that occurs in children who experience chronic interpersonal
trauma. Thus, the ICD-11 has described PTSD and CPTSD as trauma-related disorders. Nevertheless, relevant dierential features
should be considered between both mental health diseases. First, whereas PTSD has been related to single-incident stressors, CPTSD
has been associated with sustained exposure to repeat or multiple types of trauma. Second, while PTSD comprises re-experiencing the
traumatic event, avoidance of traumatic reminders, and hypervigilance, CPTSD includes the PTSD symptoms in addition to the

Corresponding author at: Health Research Institute La Fe, Av. Fernando Abril Martorell, 106, 46010 Valencia, Spain.
E-mail address: ana.garcia-blanco@uv.es (A. Garca-Blanco).

http://dx.doi.org/10.1016/j.chiabu.2017.09.020
Received 14 February 2017; Received in revised form 10 September 2017; Accepted 13 September 2017
0145-2134/ 2017 Elsevier Ltd. All rights reserved.
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

presence of symptoms on the three self-organization domains (i.e., emotional dysregulation, negative self-concept, and interpersonal
disturbances) (Maercker et al., 2013). Recent research suggests that these self-organization disturbances can be explained by the
dierential impact of trauma on children and adults (Pynoos et al., 2009) rather than by trauma frequency (i.e., single or sustained).
Noteworthy, childhood is a period in which individuals are still biologically, psychologically, and socially immature (Van der Kolk,
2003). Thus, this brain immaturity may explain why maltreated children are more sensitive to stressful and detrimental occurrences
than adults. This in turn increases their vulnerability to an impaired identity conguration, negative cognitive styles (i.e., altered
attributions about themselves and others), and emotional dysregulation (Cloitre et al., 2009). Unlike PTSD, CPTSD has been asso-
ciated with long-term psychosocial and functional impairment, which might be partially explained by these self-organization dis-
turbances (DAndrea, Ford, Stolbach, Spinazzola, & Van der Kolk, 2012).
Among the self-organization disturbances of CPTSD, emotional dysregulation has been proposed as an underlying mechanism of
negative self-concept and interpersonal disturbances (Gilbert et al., 2009; Scott, Smith, & Ellis, 2010; Widom, DuMont, & Czaja,
2007). At the theoretical level, information processing models (e.g., Beck, 1976) suggest that the manner in which information is
attended to, interpreted, and remembered is congruent with mood (i.e., individuals with a low mood have a preference for negative
information). Thus, emotional dysregulation conducts biases in emotional information processing (Dalgleish, Moradi, Taghavi,
Neshat-Doost, & Yule, 2001), leading to a negative self-concept (see Infurna et al., 2016, for a meta-analysis) and interpersonal
disturbances (Pollak, Vardi, Ptzer Bechner, & Curtin, 2005; Pollak, 2008; Shackman, Shackman, & Pollak, 2007). Although these
cognitive biases are a signicant factor in emotional vulnerability and elicit abnormal behavioral responses, how emotional dysre-
gulation occurs in CPTSD is still under discussion (Cloitre, Gavert, Brewin, Bryant, & Maercker, 2013; Elklit, Hyland, & Shevlin, 2014;
Knefel & Lueger-Schuster, 2013). As such, the assessment of how emotionally relevant stimuli from the social context are attended to
and processed, and their association with a specic pattern of clinical symptoms, may be essential to understanding emotional
dysregulation in maltreated CPTSD children.
Over last decade, emotional information processing has been exhaustively investigated in relation to acute trauma and PTSD,
which has been characterized by a bias toward threat-related stimuli (Dalgleish et al., 2001; Elsesser, Sartory, & Tackenberg, 2005).
Emotional processing in PTSD has been studied mainly through behavioral experimental paradigms. The dot-probe task has been
considered an excellent behavioral technique for examining how emotional stimuli capture attention (Garca-Blanco, Garca-Blanco,
Fernando, & Perea, 2016). During the dot-probe task, two cues of dierent valence (i.e., neutral versus emotional) are presented
simultaneously. The presentation rate of cues determines whether the attentional processing is automatic or controlled (i.e., under or
over 1 s, respectively) (Yiend, 2010). Thus, emotional cues can either be displayed by using a pre-attentive presentation rate, in
which stimuli presented for shorter durations (i.e., 500 ms) are used to evaluate automatic and earlier stages of processing, or by
using an attentive presentation rate, in which stimuli presented for extended durations (i.e., up to 1000 ms) are used to assess
controlled and higher-order regulatory inuences (Mc Crory et al., 2013). As soon as the stimuli disappear, a dot-probe (target)
replaces one of the stimuli. This trial can be either an emotional trial (e.g., when the target replaces the emotional stimulus) or a
neutral trial (e.g., when the target replaces the neutral stimulus). Participants are instructed to indicate the location at which the
target has appeared. Faster responses in emotional trials indicate an attentional bias toward emotional stimuli, whereas faster re-
sponses in neutral trials indicate an attentional bias away from emotional stimuli.
Similar to PTSD, recent data suggests that early exposure to maltreatment is related to a bias to threat-related stimuli (Gibb,
Schoeld, & Coles, 2009; Pine et al., 2005). Nevertheless, the direction of this abnormal threat-related processing (i.e., avoidance
versus approach) is not entirely consistent since it has provided contradictory evidence of the association between complex trauma
exposure and threatening attention bias (Gibb et al., 2009; Pine et al., 2005). Two plausible hypotheses have been proposed for these
discordant ndings: i) the stage of the assessed attentional processing, which is determined by the presentation rate of the emotional
cues (Yiend, 2010); ii) recent or distal trauma exposure, which is determined by whether the assessed individuals are children or
adults with a history of complex trauma (Elsesser et al., 2005).
The rst hypothesis proposes that the direction of this abnormal threat-related processing depends on the stimulus presentation
rate (Yiend, 2010). For instance, whereas Gibb et al. (2009) found an attentional approach toward threatening faces at an attentive
presentation rate, Pine et al. (2005) documented attention avoidance away from threatening faces at a pre-attentive presentation rate.
Gibb et al. (2009) applied a dot-probe task to adults with a history of childhood abuse. Attentional biases to threatening, happy, and
sad faces were assessed during an attentive (i.e., 1000 ms) presentation rate. The results showed attention bias toward threatening
cues in adults who had suered maltreatment relative to the control group. However, no signicant group dierences relative to sad
or happy faces were obtained. In contrast, Pine et al. (2005) administered a dot-probe task to maltreated children, most of whom
presented concomitant PTSD. Attentional biases to threatening and happy faces were assessed during a pre-attentive (i.e., 500 ms)
presentation rate. Unlike the control group, maltreated children performed attention bias away from threatening faces. Moreover,
those children who had been more intensely maltreated and met the PTSD criteria demonstrated a greater avoidance of threatening
faces. No dierences emerged in their attention responses to happy faces.
A second hypothesis suggests that the period of time since the traumatic event occurred could determine the direction of at-
tentional bias. In other words, recent versus distal trauma might explain if threatening stimuli are avoided or attended to (Fani,
Bradley-Davino, Ressler, & McClure-Tone, 2011). Gibb et al. (2009) assessed adults with a history of trauma, whereas Pine et al.
(2005) assessed maltreated children at the moment of testing. To check this hypothesis, Elsesser et al. (2005) administered a dot-
probe task to recent acute and single trauma survivors at short lapses since the trauma and at three months later. Attentional biases to
trauma-related pictures were assessed during a pre-attentive (i.e., 500 ms) presentation rate. Trauma victims attended away from
threat-related pictures shortly after the traumatic experience but toward threatening pictures three months after the trauma occurred.
Another important gap in studies concerning children exposed to complex trauma is that the attentional processes have been

43
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

mainly focused on threat-related processing, whereas attentional biases to sad information have been poorly considered. This concept
is important due to the high risk of depression, the negative self-concept, and the interpersonal disturbances observed in children who
have suered from complex trauma (Duque & Vzquez, 2015). Romens and Pollak (2012) applied a dot-probe task to maltreated
children after the induction of a sad emotional state. Attentional biases to threatening, sad, and happy faces were assessed during an
attentive (i.e., 1500 ms) presentation rate. Children with higher levels of maltreatment and higher rumination traits showed an
attention bias toward sad faces at an 8-min delayed period after the sad induction. The authors concluded that detecting a sad
attention bias in maltreated children could be used as a predictor to identify subjects who might present a higher risk of depression in
the future. Maltreated children may show not only a threatening bias, which is associated with PTSD symptoms (Gibb et al., 2009;
Pine et al., 2005), but also a sad bias, which might be associated with the development of depressive symptoms (Romens & Pollak,
2012).
Taken together, research on attentional biases associated with complex trauma remains controversial due to the heterogeneous
methodology in relation to the type of emotional stimuli (e.g., threatening versus sad), the current or past trauma, and the type of
stimulus presentation rate used (e.g., pre-attentive versus attentive). Therefore, this is the rst study to: (i) examine attention bias
when using a dot-probe task that involves neutral versus angry, sad, and happy faces; (ii) assess an homogeneous group of CPTSD
children with recent maltreatment experiences; and (iii) use both pre-attentive (i.e., 500 ms) and attentive (i.e., 1500 ms) pre-
sentation rates.
Firstly, we hypothesize that maltreated CPTSD children would show an abnormal processing of negative information, consisting
of avoidance of angry faces (Pine et al., 2005) and an approach to sad faces (Romens & Pollak, 2012). Our second hypothesis is that
according to the stimulus presentation rate, the avoidance of angry faces would show at pre-attentional (Pine et al., 2005) but not
during attentional presentation stages (Gibb et al., 2009) as abnormal attentional biases in CPTSD children are automatic reactions
that are not amenable to controlled and higher-order regulatory inuences. Our third hypothesis explores the association between
attentional bias according to the type of stimuli used (e.g., angry, threatening and sad faces) and specic clinical symptoms in the
children. Maltreated children may show not only a threatening bias, which is associated with PTSD symptoms (Gibb et al., 2009; Pine
et al., 2005), but also a sad bias, which might be associated with the development of depressive symptoms (Romens & Pollak, 2012).
According to information processing models (Beck, 1976), attentional biases constitute a factor of emotional vulnerability and elicit
abnormal behavioral responses (Pollak et al., 2005; Pollak, 2008; Shackman et al., 2007), which are conditioned by attitudes about
oneself, the world, and the future (Dalgleish et al., 2001; Infurna et al., 2016). Previous studies using the dot-probe task have also
found a link between attentional biases toward sad faces in adolescents with depressive symptoms (Klein, de Voogd,
Wiers, & Salemink, 2017) and toward angry faces in young people with disruptive mood dysregulation disorder (Stoddard et al.,
2016), suggesting that attention biases might play an important role in these clinical conditions. We thus aim to explore whether
these attentional biases toward sad, threatening, and happy faces in children with CPTSD are in any way correlated with any clinical
symptoms.

2. Method

2.1. Participants

Forty-seven children took part in the experiment. Twenty-one children with documented exposure to maltreatment were admitted
to a specic program for maltreated children in an outpatient child and adolescent mental health service. The study procedures were
approved by the hospitals Bioethics Committee. All children had been removed from their homes because of conditions of extreme
negligence. They were all exposed to long-term physical abuse, and three of them also suered sexual abuse. They were safeguarded
in a young oenders institution after all parental responsibility was removed. The state had court-awarded custody of all of them. In
order to avoid any impact of drugs on cognitive function, the children were assessed before any psychotropic medication was
prescribed. Information regarding maltreatment and a psychiatric mental health assessment were collected from two sources. First,
the social services case les were used to obtain a precise characterization of the childrens maltreatment history. Second, this
characterization was conrmed through consecutive accurate psychiatric, psychological, and social non-structured interviews.
An additional control group of non-maltreated children was composed by 26 subjects, 17 of whom had been recruited from a
neighboring local public primary school. The remaining participants in the control group were self-referred by means of adver-
tisements in the community and word-of-mouth. The control group was comparable in terms of age, gender, handedness, and eth-
nicity. Informed consent was obtained from all the participants as well as from their legal representatives. Fig. 1 shows the selection
process of the nal sample.

2.2. Inclusion criteria

All participants in the clinical group had received the diagnosis of CPTSD in compliance with the ICD-11 proposal by the referring
clinicians prior to the study. In addition to their PTSD symptoms (re-experiencing the traumatic event, avoidance of traumatic
reminders, and excessive hypervigilance), the CPTSD children also showed at least one symptom from each of the following domains:
i) emotional dysregulation (violent outbursts, excessive crying, anhedonia, self-destructive behavior, dissociation, emotional numbing);
ii) negative self-concept (perception of a diminished or defeated sense of self, persistent negative beliefs about oneself along with
feelings of guilt and shame); and iii) interpersonal problems (inability to build or maintain close and intimate personal bonds).
The diagnosis of CPTSD was veried by expert clinical assessment and by the standardized procedure applied by Cloitre et al.,

44
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

Fig. 1. Selection process of the nal sample.

2013, that is, a combination of the Modied PTSD Symptom Scale-Self-Report Severity (MPSS-SR; Falsetti, Resnick,
Resick, & Kilpatrick, 1993) and the Brief Symptom Inventory (BSI; Derogatis & Melisaratos, 1983) after the performance of a latent
prole analysis. Both the MPSS-SR and BSI are self-report instruments in which responses are recorded on a Likert scale. The MPSS-SR
is a brief scale in which participants are asked about the severity of each of the 17 PTSD symptoms outlined in the DSM-IV using a
Likert scale ranging from 0 = not at all to 4 = extremely. The internal consistency was supported by alpha coecients ranged from
0.96 to 0.97 and the inter-rater reliability was supported by intraclass correlation coecients from 0.21 to 0.62 (Falsetti et al., 1993;
Ruglass, Papini, Trub, & Hien, 2014). The BSI consists of a 53-item self-report psychological symptom inventory with nine primary
symptom dimensions (Somatization, Obsessive-Compulsive, Interpersonal Sensitivity, Depression, Anxiety, Hostility, Paranoid
Ideation, and Psychoticism) in which participants are asked about the extent to which a problem bothered or distressed them using a
5-point frequency rating scale from 0 = not at all to 4 = extremely. Internal consistency coecients ranged from 0.71 to 0.82 and
test-retest reliability ranged from 0.68 to 0.91 (Derogatis & Melisaratos, 1983).

2.3. Exclusion criteria

None of the participants exhibited head trauma or a neurological history, major medical disorders, use of medication that could
inuence cognition (e.g., psychotropic medication, treatment with corticosteroids), or diculty distinguishing colors (e.g., color-
blindness). Apart from CPTSD, other comorbid psychiatric diagnoses in the CPTSD group were considered as exclusion criteria. All
subjects from the control group lived with their families of origin. The exclusion criteria for the control group were a previous history
of maltreatment or neglect and/or a previous psychiatric history.
Subclinical symptoms were assessed using the Child Behavior Checklist (CBCL; Achenbach and Edelbroch, 1991), which was
completed by every legal child representative (a social worker was assigned to every child at the time of being safeguarded by an
institution). The CBCL obtains information from adults on problematic behavior in children at ages between 6 and 18 through eight
syndrome scales (Anxious/Depressed, Depressed/Withdrawal, Somatic Complaints, Social Problems, Thought Problems, Attention
Problems, Rule-Breaking Behavior, and Aggressive Behavior). The internal consistency coecient ranged from 0.72 to 0.96 and the
inter-rater and test-retest reliabilities ranged from 0.93 to 1.00 (Achenbach & Edelbroch, 1991). The demographic and clinical data
for the nal sample are presented in Table 1.

2.4. Materials

In our experiment, the emotional stimuli serving as cues consisted of 84 color photographs of male and female facial expressions
taken from the FACES database (Ebner, Riediger, & Lindenberger, 2010). Two faces appeared as cues in each trial, namely, an

45
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

Table 1
Sociodemographic data and clinical symptoms for each group.

Control CPTSD
(n = 26) (n = 21) p

Female (%) 69.2% 57.1% 0.391

Age (years) [M(SD)] 9.7 (1.4) 10.8 (3.2) 0.118
CBCL scores [M(SD)]
Anxious/Depressed 55.0 (3.8) 64.0 (8.4) 0.000
Withdrawn/Depressed 51.5 (2.2) 75.38 (12.2) 0.000
Somatic Complains 55.2 (3.4) 61.2 (8.4) 0.000
Social Problem 53.0 (2.2) 71.8 (11.0) 0.000
Thought Problems 54.8 (5.6) 64.7 (8.5) 0.000
Attention Problems 52.8 (2.2) 69.1 (9.9) 0.000
Rule-Breaking Behavior 52.9 (4.3) 66.8 (11.3) 0.000
Aggressive Behavior 53.4 (2.1) 60.9 (6.9) 0.000
PTSD symptoms [M(SD)] 5.3 (3.2) 14.9 (5.0) 0.001
Re-experiencing 1.3 (1.8) 4.0 (2. 9) 0.001
Avoidance 2.0 (2.0) 5.3 (2.5) 0.001
Hypervigilance 2.9 (1.3) 5.6 (2.4) 0.001
Self-organization symptoms [M(SD)] 5.7 (3.2) 19.1 (2.9) 0.001
Emotional dysregulation 1.6 (1.7) 6.4 (1.5) 0.001
Negative self-concept 2.0 (1.8) 6.0 (1.8) 0.001
Interpersonal problems 2.2 (1.6) 6.7 (1.6) 0.001

CPTSD: Complex Posttraumatic Stress Disorder; CBCL: Child Behavior Checklist; M: mean; SD: Standard Deviation.

emotional face (happy, angry, or sad) and a neutral face. We selected a total of 12 happy, 12 angry, and 12 sad faces and 48 neutral
images (36 for the control and 12 for the practice trials). Each emotional face was matched with the neutral control face of the same
actor. Each participant was presented with three types of experimental trials: 12 happyneutral, 12 angryneutral, and 12 sad-
neutral cues. Each pair of cued faces was presented three times during the experiment (i.e., there were 36 trials per condition). In
addition, six pairs of neutral faces were presented before the experimental trials as a practice block.

2.5. Procedure

All the children were individually assessed by the same clinical psychologist in a quiet noise-isolated room. In order to test for
patterns of attention processes during the emotional stimuli, we used the visual dot-probe task (McLeod, Mathew, & Tata, 1986).
Stimulus presentations and the recording of responses were controlled by the DMDX software (Forster & Forster, 2003). In each trial,
the participants were instructed to look at a xation point (+) in the center of the screen, which was presented for 500 ms.
Thereafter, two faces were presented simultaneously at dierent screen locations (up and down), which were two cued stimuli with
dierent emotional valences (i.e., one neutral and one emotional) randomly displayed for 500 ms or 1500 ms. The images then
disappeared, and a green or red square replaced one of the two stimulieither emotional (i.e., emotion trial) or neutral (i.e., neutral
trial). The participants were told to press a button to indicate the color of the square as quickly and as accurately as possible. This trial
can be: an emotional trial (i.e., the target replaces the emotional stimulus) or a neutral trial (i.e., the target replaces the neutral
stimulus). The subjects were instructed to press a button to indicate the location in which the target appears. Faster responses in
emotional trials signal an attentional bias toward emotional stimuli, whereas faster responses in neutral trials signal an attentional
bias away from emotional stimuli. The sequence of stimulus presentation is shown in Fig. 2.
The task comprised of one practice block followed by nine test blocks composed of 12 experimental trials (4 happyneutral, 4
angryneutral, and 4 sadneutral), which were randomly displayed within each block. Thus, a total of 114 trials (108 study + 6
ller) were displayed. The vertical location and type of face (emotional or neutral) replaced by the square were balanced across trials,
with the constraint that each type of face appeared in each of two positions on half of the trials, and the square replaced the emotional
cues on the other half. The presentation order of the blocks was randomized across participants. The variation in the image locations
and the randomization of the trials guaranteed that the participants were not able to use any predetermined scanning strategy. The
whole session lasted approximately 2530 min.

2.6. Data analysis

Probe response times (RTs) were calculated for correct responses (i.e., errors responses were excluded from further latency
analyses). Preliminary analyses showed that both groups showed very low error rates (less than 5%) and that there were no dif-
ferences between groups and conditions (all Fs < 1). Before examining the bias scores, the dot-probe data were cleaned following the
procedures of previous studies on children (Marotta et al., 2013). Very short RTs (less than 200 ms) or those exceeding 2.5 standard
deviations above the participants means were excluded to ensure that the latencies were based on actual responses to the probe
locations. For each participant, the mean RT in each condition (for happy, angry, and sad faces at 500 and 1500 ms) was calculated.
As suggested by Behrmann et al. (2006), the dierence in proportion between the emotional (i.e., where the probe replaced an

46
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

Fig. 2. The stimulus presentation sequence in a neutral trial.

emotional face) and neutral trials (i.e., where the probe replaced a neutral face) was calculated to estimate the bias scores [(mean RT
neutral trials/mean RT emotional trials*100)-100)] to control for the RT dierences between the CPTSD children and the healthy
children (699 ms and 674 ms, respectively). Positive bias scores indicate an attentional bias toward a particular emotional face,
whereas negative bias scores indicate an attentional bias away from an emotional face.

3. Results

The RT mean for each condition is shown in Table 2. The mean (presented in percentages) RT dierences are shown in Fig. 3.
First, the dierences (in percentages) were analyzed in a 2 (Group: CPTSD, Control) x 3 (Valence: happy, angry, sad) x 2 (Cue
presentation rate: 500 ms, 1500 ms) omnibus analysis of variance (ANOVA) in which Group was the between-subjects factor, and the
Valence and Cue presentation rates were the within-subjects factors. Simple eects tests were conducted in the case of signicant
interactions. Second, one-sample t-tests were used to determine whether the bias score was statistically dierent from zero for each
group. Third, bivariate correlations were conducted to check the relation between signicant dot-probe bias scores and the CBCL
symptoms in the CPTSD group.
We failed to nd a Cue presentation rate eect, [F(1.45) = 2.20, p = 0.145, 2 = 0.05], a Valence eect (F < 1), a Group eect
[F(1.45) = 1.37, p = 0.248, 2 = 0.03], a Cue presentation rate x Group interaction (F < 1), a Valence x Cue presentation rate
interaction (F < 1), or a three-way Cue presentation rate x Valence x Group interaction (F < 1). However, a statistically signicant
Valence x Group interaction was found [F(2.90) = 7.13, p = 0.001, 2 = 0.14].
In order to further examine this interaction, we conducted separate ANOVAs using Valence x Group as factors. Interestingly, for
the angry faces, the CPTSD group showed a more negative bias score than the control group, [F(1.46) = 13.72, p = 0.001]. For the
sad faces, the CPTSD group showed a more positive bias score than the control group, [F(1,46) = 4.94, p = 0.031]. For the happy
faces, no dierences were found between the CPTSD and control groups, [F(1,46) = 1.45, p = 0.235]. Thus, an attentional bias away
from angry faces but an attentional bias toward sad faces was found in CPTSD children relative to healthy children.
In addition, one-sample t-tests were used to determine whether the bias score was statistically dierent from zero for each group.

Table 2
Response time (ms) to dot-probe task for each group.

Cue Control CPTSD

(n = 26) (n = 21)

Valence Presentation rate Emotional Trial [M(SD)] Neutral Trial [M(SD)] Emotional Trial [M(SD)] Neutral Trial [M(SD)]

Happy Short 637 (138) 651 (145) 657 (176) 648 (171)
Long 708 (141) 695 (143) 720 (183) 701 (183)
Sad Short 657 (144) 641 (133) 670 (235) 691 (245)
Long 705 (146) 687 (137) 722 (241) 728 (235)
Threat Short 645 (144) 664 (144) 685 (179) 660 (168)
Long 699 (158) 710 (160) 734 (179) 715 (170)

CPTSD: Complex Posttraumatic Stress Disorder; M: mean; SD: Standard Deviation.

47
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

Fig. 3. Bias scores (with Standard Error) for the Valence x Group interaction.

One-sample t-tests showed that bias scores for angry faces were signicantly smaller than zero in the CPTSD group [t(20) = 2.38,
p = 0.027] (i.e., we found a bias away from angry faces) and signicantly higher than zero for angry faces in the control group [t(25)
= 2.88, p = 0.008] (i.e., we found a bias toward angry faces). No bias scores reached signicant dierences from zero for happy and
sad faces, both in the CPTSD group (all ps > 0.12) and the control group (all ps > 0.15).
Finally, bivariate Pearson correlations between the CBCL subscales and the angry bias scores in the CPTSD group indicated that
social problems were signicantly correlated with the bias scores for angry faces (r = 0.646, p = 0.002). That is, higher scores for
social problems were associated with more negative bias scores for threatening faces (i.e., higher avoidance). No other CBCL sub-
scales correlated with an angry bias score (all ps > 0.10). Additionally, we computed Pearsons coecient to examine the re-
lationship between the CBCL subscales and the sad bias score in the CPTSD group. Depressed symptomatology was signicantly
correlated with sad bias scores (r = 0.570, p = 0.007). Higher scores on the depressive/withdrawn subscale were associated with
more positive bias scores for sad faces (i.e., higher approach), however, no other CBCL subscales correlated with sad bias scores (all
ps > 0.10).

4. Discussion

The present study examined attentional bias toward both positive and negative stimuli (e.g., angry and sad faces) in a sample of
children with CPTSD and used both automatic and controlled stages of processing. The main ndings can be summarized as follows.
We had initially hypothesized that maltreated CPTSD children would show an abnormal processing of negative information, con-
sisting of avoidance of angry faces while approaching sad faces. We had also predicted that the avoidance of angry faces would show
at pre-attentional more than during attentional presentation rates. Finally, we had suggested that attention biases could be directly
related to concomitant clinical presentation.
As hypothesized, in our study, CPTSD children did demonstrate abnormal emotional face processing relative to the control group,
which was conditioned by the valence (e.g., happy, angry, and sad) of the face. First, the dierences between the groups were
particularly clear in relation to angry and sad faces. Second, according to the direction of attentional biases and in comparison with
children in the control group, maltreated CPTSD children showed an attentional bias away from angry faces but an attentional bias
toward sad faces. However, the dierences were found during both pre-attentional and attentional presentation rates. Finally, an
association between the avoidance of threatening faces and social problems and approaching sad faces and depressive symptoms was
observed in the present study.
With regard to abnormal threat-related processing, the CPTSD children in the current study performed attention avoidance of
angry faces regardless of the stimulus presentation rate. This abnormal threat-related processing is consistent with previous data in
children with current maltreatment experiences (Pine et al., 2005). However, in contrast to Gibb et al. (2009), we also found that the
attentional bias away from threatening faces was maintained during the attentive presentation rates. Thus, contrary to our initial
hypothesis, cognitive processing bias in the present study was not determined by the stimulus presentation rate. These results might
be more in line with the second hypothesis that the direction of threat-related attentional bias may have a higher association with the

48
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

period of time since the traumatic event occurred than to the processing stimulation rate itself (Elsesser et al., 2005; Fani et al., 2011).
Several explanations could account for the attentional biases away from angry faces in CPTSD children. The lack of attention to
threatening cues in recent maltreated children may represent a maladaptive mechanism to environmental adversity, which further
impacts on the development of adaptive strategies to identify and cope with threatening situations (hman, 2009; Pollak & Sinha,
2002). A reduced ability to detect anger aective cues would therefore result in a biased and negative perception of future events
(Pollak & Sinha, 2002). In our sample, the greater the avoidance of angry faces, the more important the social problems in children
with CPTSD. A plausible interpretation is that maltreated CPTSD children may present a higher threshold of detecting anger and
hostility during social interactions, which would in turn lead them into dangerous situations (Pine et al., 2005). Conversely, over
time, individuals might develop some adaptive strategies, which will be represented by preferential attention to threatening faces.
Thus, the development of strategies would be a key factor in understanding the dierences between studies using recent and distal
trauma (hman, 2009).
CPTSD children also performed an attention approach to sad faces. Importantly, this attentional bias toward sad faces was
associated with more withdrawn and depressive symptoms. Similar to the present results, Romens and Pollak (2012) found that
maltreated children with depressive traits such as rumination show a dierent attention approach to sad information. An inter-
pretation of the current data can also be placed in the context of previous research with children at risk of depression, who pre-
ferentially attended to sad faces (Hankin, Gibb, Abela, & Flory, 2010). Early negligent experiences such as complex trauma may
predispose subjects to depression in the future (Widom et al., 2007). Noteworthy, all children comprising our sample met the criteria
for CPTSD according to the new ICD-11 proposal, though not for other psychiatric disorders such as depression. The ICD-11 criteria
emphasize that aective instability and emotional liability are clinical manifestations of CPTSD, so it is probable that those children
who showed stronger attention to sad stimuli may suer depressive symptomatology.
The present study does have some limitations. First, the relatively small sample size might have reduced the power to detect other
real dierences between the groups. However, our sample, which was composed entirely of children exposed to an intensity of
maltreatment severe enough to be removed from their parents tutelage, as well as meeting the criteria for CPTSD according to the
new ICD-11 proposal, was fairly homogeneous to detect dierences between the groups. Second, we utilized a cross-sectional study
design, and a follow-up study in the current sample has not been performed. As such, the possibility of issuing causality between
maltreatment and attention bias is restricted. For this reason, an important goal will be to perform future longitudinal or naturalistic
studies in order to illustrate causal inferences and to provide valuable data about CPTSD and its clinical presentation. Furthermore, a
follow-up of the current sample, as performed by Rutter and Sonuga-Barke (2010) with the English and Romanian adoptee cohort,
could give us additional clarity into the causal pathways underlying the development of CPTSD and future psychiatric conditions.
Third, while the control group was selected for comparability in terms of age, gender, handedness, and ethnicity, the partial re-
cruitment of the control group from a public school and the possible dierences in socioeconomic status between the control and the
CPTSD groups could constitute a cohort-eect limitation. Moreover, the possibility of comparing children with clinical depression/
anxiety who have not experienced previous traumatic events could shed light on the cognitive aspects underlying the etiology of
CPTSD (Maercker et al., 2013).
In conclusion, these ndings suggest that CPTSD children who have been exposed to maltreatment develop an abnormal pro-
cessing of negative information, regardless of the stage of attentional processing. Therefore, the identication of attention bias may
be useful in the characterization of CPTSD children. In fact, the attention bias away from angry faces and the attention bias toward sad
faces may underlie concomitant clinical presentation such as social problems and depressive symptoms, respectively (see Garca-
Blanco, Perea, & Livianos, 2013, for evidence regarding the inuence of mood on attention). Our ndings might represent a potential
for future research and clinical intervention. Concerning intervention implications, attention bias abnormalities should be seen as
targets for novel treatment in CPTSD children (DAndrea et al., 2012; Pine et al., 2005). In this context, recent studies have found that
attentional bias can modify and improve aective symptoms (Wells & Beevers, 2010). In relation to research implications, further
research should examine whether attentional bias modication can alleviate aective symptomatology and social problems in CPTSD
children. Additionally, a further follow-up of this sample would conrm whether attention bias away from angry faces and attention
bias toward sad faces increase vulnerability regarding future psychiatric disorders (Garca-Blanco et al., 2013; Young & Widom, 2014;
Pine et al., 2005).

Conict of interest

The authors declare that they have no conict of interest.

Acknowledgements

This study was funded by 43-CONTROL-DES-PEREA-GARCIA-2015-A from VLC-BIOMED (University of Valencia and
University & Polytechnic Hospital La Fe, Spain); JR17/00003 Juan Rods from the Instituto Carlos III (Spanish Ministry of Economy
and Competitiveness).

References

hman, A. (2009). Of snakes and faces: An evolutionary perspective on the psychology of fear. Scandinavian Journal of Psychology, 50, 543552.
Achenbach, T. M., & Edelbroch, C. S. (1991). Manual for the child behavior checklist and revised child prole. Burlington: University of Vermouth.

49
C. Bert et al. Child Abuse & Neglect 73 (2017) 4250

Beck, A. T. (1976). Cognitive Therapy and the Emotional Disorders. New York: New American Library.
Behrmann, M., Avidan, G., Leonard, G. L., Kimchi, R., Luna, B., Humphreys, K., et al. (2006). Congural processing in autism and its relationship to face processing.
Neuropsychologia, 44, 110129.
Cloitre, M., Stolbach, B. C., Herman, J. L., van der Kolk, B., Pynoos, R., Wang, J., et al. (2009). A developmental approach to complex PTSD: Childhood and adult
cumulative trauma as predictors of symptom complexity. Journal of Traumatic Stress, 0, 110.
Cloitre, M., Courtois, C. A., Ford, J. D., Green, B. L., Alexander, P., Briere, J., et al. (2012). The ISTSS expert consensus treatment guidelines for complex PTSD in adults.
[Retrieved November 5].
Cloitre, M., Gavert, D. W., Brewin, C. R., Bryant, R. A., & Maercker, A. (2013). Evidence for proposed ICD-11 PTSD and complex PTSD: A latent prole analysis.
European Journal of Psychotraumatology, 4, 20706.
DAndrea, W., Ford, J., Stolbach, B., Spinazzola, J., & van der Kolk, A. (2012). Understanding interpersonal trauma in children: why we need a developmentally
appropriate trauma diagnosis. American Journal of Orthopsychiatry, 82, 187200.
Dalgleish, T., Moradi, A. R., Taghavi, M. R., Neshat-Doost, H. T., & Yule, W. (2001). An experimental investigation of hypervigiance for threat children and adolescents
with post-traumatic stress disorder. Psychological Medicine, 31, 541547.
Derogatis, L. R., & Melisaratos, N. (1983). The brief symptom inventory: An introductory report. Psychological Medicine, 13, 595605.
Duque, A., & Vzquez, C. (2015). Double attention bias for positive and negative emotional state in clinical depression: Evidence from an eye-tracking study. Journal of
Behavioral Therapy & Experimental Psychiatry, 46, 107114.
Ebner, N. C., Riediger, M., & Lindenberger, U. (2010). FACES-a database of facial expressions on young, middle-aged: and older women and men: Development and
validation. Behavior Research Methods, 42, 351362.
Elklit, A., Hyland, P., & Shevlin, M. (2014). Evidence of symptom proles consistent with posttraumatic stress disorder and complex posttraumatic stress disorder in
dierent trauma samples. European Journal of Psychotraumatology, 5, 24221.
Elsesser, K., Sartory, G., & Tackenberg, A. (2005). Initial symptoms and reactions to trauma-related stimuli and the development of posttraumatic stress disorder.
Depression and Anxiety, 21, 6170.
Falsetti, S. A., Resnick, H. S., Resick, P. A., & Kilpatrick, D. G. (1993). The modied PTSD symptom scale: A brief self-report measure of posttraumatic stress disorder.
The Behavioral Therapist, 16, 161162.
Fani, N., Bradley-Davino, B., Ressler, K. J., & McClure-Tone, E. B. (2011). Attention bias in adult survivors of childhood maltreatment with and without posttraumatic
stress disorder. Cognitive Therapy and Research, 35, 5767.
Forster, K. I., & Forster, J. C. (2003). DMDX: A windows display program with millisecond accuracy. Behavior Research Methods, Instruments & Computers, 35, 116124.
Garca-Blanco, A., Perea, M., & Livianos, L. (2013). Mood-congruent bias and attention shifts in the dierent episodes of bipolar disorder. Cognition & Emotion, 27,
11141121.
Garca-Blanco, A., Garca-Blanco, M. C., Fernando, B., & Perea, M. (2016). Inuence of computer feedback on attentional biases to emotional faces in children.
Computers in Human Behavior, 64, 881887.
Gibb, B. E., Schoeld, C. A., & Coles, M. E. (2009). Reported history of childhood abuse and young adults information-processing biases for facial displays of emotion.
Child Maltreatment, 14, 148156.
Gilbert, R., Widom, C. S., Browne, K., Fergusson, D., Webb, E., & Janson, S. (2009). Burden and consequences of child maltreatment in high-income countries. Lancet,
373, 6881.
Hankin, B. L., Gibb, B. E., Abela, J. R. Z., & Flory, K. (2010). Selective attention to aective stimuli and clinical depression among youths: Role of anxiety and
specicity of emotion. Journal of Abnormal Psychology, 119, 491501.
Infurna, M. R., Reichl, C., Parzer, P., Schimmenti, A., Bifulco, A., & Kaess, M. (2016). Associations between depression and specic childhood experiences of abuse and
neglect: A meta-analysis. Journal of Aective Disorders, 190, 4755.
Klein, A. M., de Voogd, L., Wiers, R. W., & Salemink, E. (2017). Biases in attention and interpretation in adolescents with varying levels of anxiety and depression.
Cognition & Emotion, 3, 19.
Knefel, M., & Lueger-Schuster, B. (2013). An evaluation of ICD-11 PTSD and complex PTSD criteria in a sample of adult survivors of childhood institutional abuse.
European Journal of Psychotraumatology, 4, 22608.
Maercker, A., Brewin, C. R., Bryant, R. A., Cloitre, M., Reed, G. M., Van Ommeren, M., et al. (2013). Proposals for mental disorders specically associated with stress in
the ICD-11. Lancet, 381, 16831685.
Marotta, A., Pasini, A., Ruggiero, S., Maccari, L., Rosa, C., Lupiez, J., et al. (2013). Inhibition of return in response to eye gaze and peripheral cues in young people
with Aspergers syndrome. Journal of Autism and Developmental Disorders, 43, 917923.
Mc Crory, E. J., De Brito, S. A., Kelly, P. A., Bird, G., Sebastian, C. L., Mecheli, A. M., et al. (2013). Amygdala activation in maltreated children during pre-attentive
emotional processing. The British Journal of Psychiatry, 202, 269276.
McLeod, C., Mathew, A., & Tata, P. (1986). Attentional bias in emotional disorders. Journal of Abnormal Psychology, 95, 1520.
Pine, D. S., Mogg, K., Bradley, B. P., Montgomery, L., Monk, C. S., McClure, E., et al. (2005). Attention bias to threat in maltreated children: Implications for
vulnerability to stress-Related psychopathology. American Journal of Psychiatry, 162, 291296.
Pollak, S. D., & Sinha, P. (2002). Eects of early experience on childres recognition of facial displays of emotion. Developmental Psychology, 38, 784791.
Pollak, S. D., Vardi, S., Ptzer Bechner, A. M., & Curtin, J. J. (2005). Physically abused childrens regulation of attention in response to hostility. Child Development, 76,
968977.
Pollak, S. D. (2008). Mechanisms linking early experience and the emergence of emotions. Current Directions in Psychological Science, 17, 370375.
Pynoos, R. S., Steinberg, A. M., Layne, C. M., Briggs, E. C., Ostrowski, S. A., & Fairbank, J. A. (2009). DSM-V PTSD diagnostic criteria for children and adolescents: A
developmental perspective and recommendations. Journal of Traumatic Stress, 22, 391392.
Romens, S. E., & Pollak, S. D. (2012). Emotion regulation predicts attention bias in maltreated children at risk for depression. Journal of Child Psychology and Psychiatry,
53, 120127.
Ruglass, L. M., Papini, S., Trub, L., & Hien, D. A. (2014). Psychometric properties of the modied posttraumatic stress disorder symptom scale among women with
posttraumatic stress disorder and substance use disorders receiving outpatient group treatments. Journal of Traumatic Stress Disorders & Treatment, 4. http://dx.doi.
org/10.4172/2324-8947.1000139.
Rutter, M., & Sonuga-Barke, E. J. (2010). X. Conclusions: Overview of ndings from the era study, inferences, and research implications. Monographs of the Society for
Research in Child Development, 75, 212229.
Scott, K. M., Smith, D. R., & Ellis, P. M. (2010). Emotion regulation predicts attention bias in maltreated children at risk for depression. Archives of General Psychiatry,
67, 712719.
Shackman, J. E., Shackman, A. J., & Pollak, S. D. (2007). Physical abuse amplies attention to threat and increases anxiety in children. Emotion, 7, 838852.
Stoddard, J., Sharif-Askary, B., Harkins, E. A., Frank, H. R., Brotman, M. A., Penton-Voak, I. S., et al. (2016). An open pilot study of training hostile interpretation bias
to treat Disruptive Mood Dysregulation Disorder. Journal of Child and Adolescent Psychopharmacology, 26, 4957.
Van der Kolk, B. A. (2003). The neurobiology of childhood trauma and abuse. Child and Adolescent Clinics of North America, 12, 293317.
Wells, T. T., & Beevers, C. G. (2010). Biased attention and dysphoria: Manipulating selective attention reduces subsequent depressive symptoms. Cognition & Emotion,
24, 719728.
Widom, C. S., DuMont, K., & Czaja, S. J. (2007). A prospective investigation of major depressive disorder and comorbidity in abused and neglected children grow up.
Archives of General Psychiatry, 64, 4956.
Yiend, J. (2010). The eects of emotion on attention: A review of attentional processing of emotional information. Cognition & Emotion, 24, 347.
Young, J. C., & Widom, C. S. (2014). Long-term eects of child abuse and neglect on emotion processing in adulthood. Child Abuse & Neglect, 38, 13691381.

50

View publication stats