Impaired executive control of emotional information in social anhedonia
We examined the executive control of emotional information and its relationship to social functioning in individuals at risk for schizophrenia, defined by high social anhedonia (SA). Using the same structure as the Attentional Network Test (ANT), we developed a measure of executive control of emotional information (ANT-Emotion) in which subjects identify the direction of an arrow flanked by irrelevant angry or neutral faces. Subjects completed the ANT, ANT-Emotion, and the Social Adjustment Scale, Self-Report (SAS-SR), a measure of social functioning. While there were no group differences in the alerting, orienting, and executive control networks assessed by the ANT, high SA individuals exhibited a specific impairment in the executive control of emotional information. High SA individuals also reported poorer social functioning. However, executive control of emotional information did not mediate the relationship between SA and social functioning. These findings indicate that, in high-risk populations, the impaired ability to inhibit emotional information allows negative affective stimuli to exert inappropriate influence on cognitive processes. These results are consistent with studies indicating similar findings in schizophrenia patients, suggesting that impaired inhibition of negative emotion may be part of the liability for the disorder.
Impaired executive control of emotional information in social anhedonia
Laura M. Tully
, Sarah Hope Lincoln, Christine I. Hooker
Psychology Department, Harvard University, Massachusetts, USA
Received 15 January 2011
Received in revised form 20 August 2011
Accepted 12 December 2011
Attention network test
We examined the executive control of emotional information and its relationship to social functioning in individuals
at risk for schizophrenia, deﬁned by high social anhedonia (SA). Using the same structure as the Attentional Network
Test (ANT), we developed a measure of executive control of emotional information (ANT-Emotion) in which sub-
jects identify the direction of an arrow ﬂanke d by irrelevant angry or neutral faces. Subjects completed the ANT,
ANT-Emotion, and the Social Adjustment Scale, Self-Report (SAS-SR), a measure of social functioning. While there
were no group differences in the alerting, orienting, and executive control networks assessed by the ANT, high SA
individuals exhibited a speciﬁc impairment in the executive control of emotional information. High SA individuals
also reported poorer social functioning. However, executive control of emotional information did not mediate the
relationship between SA and social functioning. These ﬁndings indicate that, in high-risk populations, the impaired
ability to inhibit emotional information allows negative affective stimuli to exert inappropriate inﬂuence on cogni-
tive processes. These results are consistent with studies indicating similar ﬁndings in schizophrenia patients, sug-
gesting that impaired inhibition of negative emotion may be part of the liability for the disorder.
© 2012 Elsevier Ireland Ltd. All rights reserved.
Social anhedonia (SA), a lack of pleasure from social interactions
that is distinguishable from social anxiety (Brown et al., 2007), is
thought to be a core feature of schizophrenia that contributes to the
liability of developing the disorder, and a key factor underlying the
social difﬁculties seen in the illness (Meehl, 1962). SA predicts social
functioning in individuals with schizophrenia (Blanchard et al., 1998)
and high-risk populations (Cohen et al., 2006), is associated with ele-
vated symptoms (Horan et al., 2007), and predicts conversion to
schizophrenia-spectrum disorders (Kwapil, 1998). Similar to schizo-
phrenia patients, socially anhedonic individuals also exhibit deﬁcits
in attention (Gooding et al., 2006b) and emotion processing
(Phillips and Seidman, 2008). In line with data demonstrating social
cognitive processes act as mediators between cognitive impairments
and functional outcome (Green et al., 2000) the present research ex-
amines attentional processes in relation to emotional information in
individuals high in social anhedonia. We hypothesize that attentional
deﬁcits affect social functioning via a speciﬁc domain of social cogni-
tion, the executive control of socially relevant emotional information,
and that this relationship is part of an underlying vulnerability ob-
servable in psychosis-prone populations.
1.1. Attention & the attention network test
Attentional deﬁcits in schizophrenia and psychosis-prone populations
are well documented (Heinrichs and Zakzanis, 1998; Nuechterlein et al.,
1998) but the mechanisms by which attentional processes inﬂuence so-
cial functioning are unknown. To understand attentional mechanisms
and how they contribute to social impairments in SA a measure that
can dissociate between speciﬁc domains of attention is needed. A promis-
ing candidate is the Attention Network Test (ANT) (Fan et al., 2002), an
experimental measure of three proposed attention networks: alerting,
orienting and executive control (Posner and Petersen, 1990). The alert-
ing network manages the ability to achieve and maintain an alert
state, the orienting network manages the a bility to select and focus
on the to-be-attended stimulus, and the executive control network
manages the ability to resolve conﬂict among responses and conse-
quently the regulation of cognitions and emotions (Posner et al.,
There is increasing evidence for a speciﬁcdeﬁcit in the executive
control network, as measured by the ANT, in schizophrenia (Wang
et al., 2005; Gooding et al., 2006a; Urbanek et al., 2009). However, it
is unclear whether this deﬁcit is a consequence of schizophrenia, or
if it is part of the liability for the disorder. Although executive control
has been examined using the ANT in positive schizotypy (Wan et al.,
2006), and individuals high in physical anhedonia (Dubal et al.,
2000), to our knowledge ANT performance has yet to be examined
in SA. Moreover, little is known about the contribution of deﬁcits in
the executive control network to social impairments in SA as execu-
tive control mechanisms have not yet been directly tied to social cog-
nition and functioning deﬁcits. We sought to examine 1) whether
Psychiatry Research 197 (2012) 29–35
⁎ Corresponding authorat: Psychology Department, Harvard University, 33 Kirkland St.,
William James Hall 806, Cambridge, MA 02138, USA. Tel.: +1 857 207 5509.
E-mail address: email@example.com (L.M. Tully).
0165-1781/$ – see front matter © 2012 Elsevier Ireland Ltd. All rights reserved.
Contents lists available at SciVerse ScienceDirect
journal homepage: www.elsevier.com/locate/psychres
executive control deﬁcits are present in high SA individuals, and
2) the role of the executive control network in a speciﬁc social cognitive
process, the executive control of emotional information.
1.2. Executive control of emotional information
Executive control of emotional information – operationalized here
as the ability to control the extent that emotion inﬂuences cognition –
is critical for successful decision making and social interactions; if an
individual is unable to inhibit irrelevant emotional information in the
social array this could adversely affect social interactions. In schizophre-
nia, neuroimaging studies show dysfunctional activity in the lateral pre-
frontal corte x (Barch, 2005), a region involved in the regulation of
emotiononbehavior(Ochsner and Gross, 2005). In behavioral studies,
individuals with schizophrenia show increased latencies on the Stroop
task for threat-related or paranoid words compared to depressed indi-
viduals and healthy controls (Bentall and Kaney, 1989), and are more
inﬂuenced by negative affective primes when rating the trustworthiness
of neutral faces (Hooker et al., 2011). Collectively, these ﬁndings are in-
dicative of deﬁcits in the executive control of emotional information in
schizophrenia, and that these deﬁcits directly affect social behavior, po-
tentially contributing to social impairments. Thus, we were interested in
investigating whether psychosis-prone individuals also exhibit impair-
ments in the executive control of emotional information, and whether
these impairments relate to social functioning.
High SA is associated with executive functioning impairments (Tallent
and Gooding, 1999) and aberrant processing of affective information
(Kerns and Berenbaum, 2000), indicating that dysregulation of inhibitory
mechanisms, such as executive control, and emotion processing is present
premorbidly. However, the interplay between affective information and
executive control in SA has received limited attention in the literature. If
high SA is related to difﬁculties in controlling the inﬂuence of emotional
information on behavior, it may contribute to social deﬁci ts. Thus, we
were interested in whether executive control of emotional information
is impaired in socially anhedonic individuals, and whether this mediates
the relationship between SA and social functioning.
1.3. The present research
The present research examined attentional processes, speciﬁcally
the executive control of emotional information, and social functioning
in high SA individuals and healthy controls. We developed a direct mea-
sure based on the ANT: the ANT-Emotion, a ﬂanker task with orienting
cues in which participants identify the direction of an arrow ﬂanked by
irrelevant neutral or angry faces. SA was assessed with the Revised So-
cial Anhedonia Scale (Eckblad et al., 1982). We also administered the
ANT as a measure of pure attentional processes, and the Social Adjust-
ment Scale–Self Report (Weissman et al., 1978)asameasureofsocial
We directly tested the following predictions: 1) Compared to con-
trols, high SA individuals will exhibit attentional deﬁcits on both the
ANT, speciﬁcally in executive control, and the ANT-Emotion, speciﬁcally
in the executive control of emotional information. 2) High SA individuals
will report poorer social functioning compared to controls. 3) Executive
control of emotional information will mediate the relationship between
SA and social functioning.
34 high SA and 29 control partic ipants were recrui ted from the Greater Boston
Area. Participants were recruited as part of a larger stud y (N = 108) investigating
the relationship between psychosis-proneness, social cognition, and social func-
tioning in which they were screened for SA with the Revised Social Anhedonia
Scale (RSAS) (Eckblad et al., 1982), and also completed the Schizotypal Personality
Questionnaire (SPQ) (Raine, 1991). Candidate high SA participants were selected
from the larger study sample using the cut-off of 1.96 standard deviations above
the mean for their gender from no rmative data re ported by Kwapil (1998):scores
of 16 and above for females; 20 and above for males. In order to compare high SA
individuals to indiv iduals who are not at risk for sch izophren ia, candidate control
participants were individuals who scored below 1.96 standard devia tions from the
mean on the RSAS and within the 10% cutoffs of the SPQ from the normative data
reported by Raine (1991) (SPQ scores less than 8 or greater than 42). Thus, con-
trol participants were individuals with no elevated risk for schizophrenia, scoring
within the normal range of the SPQ and the RSAS. Excl usion criteria were: E nglis h
as a second language, IQ below 70, history of head trauma, neurological illness,
current or past axis I disorders, and active subst ance abuse wi thin the past six
Demographic information and group differences are presented in Table 1. The
study was approved by the ethical review board at Harvard University. After study pro-
cedures were explained, participants gave written informed consent.
2.2. Materials & assessments
2.2.1. Social anhedonia
The Revised Social Anhedonia Scale (RSAS) (Eckblad et al., 1982)isa40item
true/false self-report scale comprising of questions measuring individuals’ de-
creased pleasure from social interactions. Exampleitemsinclude:“Iattachvery
little importance to having close friends” (keyed tru e) and “Just being with friends
canmakemefeelreallygood” (keye d false).
2.2.2. Psychiatric illness
Axis I disorders were assessed using the Structured Clinical Interview for DSM-IV
Axis I Disorders (First et al., 2002); clinical interviews were conducted by trained mas-
ter's level doctoral students and supervised by a licensed clinical psychologist (CH).
Reliability assessments conducted by an independent clinician on a random sample
of ten clinical interviews revealed a kappa of 0.67, indicating substantial diagnostic
agreement (Landis and Koch, 1977).
Full scale IQ scores were estimated using the vocabulary and matrix reasoning subtests
of the Weschler Abbreviated Intelligence Scale (WASI) (Wechsler, 1999).
2.2.4. Social functioning
Social functioning was assessed with the Social Adjustment Scale – Self Report (SAS-SR)
(Weissman et al., 1978). The SAS-SR consists of 54 questions assessing six areas of function-
ing: work, social and leisure activities, relationships with extended family, role as marital
partner, parental role, and role within the family unit. Each area is rated across four catego-
ries of assessment: performance at expected tasks, level of conﬂict with people, interper-
sonal relations, and feelings and satisfactions. Area scores are averaged together to create
one composite score of social functioning.
2.3. Tasks and stimuli
2.3.1. The ANT
The ANT (Fan et al., 2002) assesses the efﬁciency of the orienting, alerting, and execu-
tive control attentional networks by measuring differences in reaction times to indicate the
direction of a central arrow across cue and ﬂanker conditions. Participants identify the di-
rection (left, right) of a target arrow that appears above or below a central ﬁxation cross.
The arrow is preceded by an asterisk cue that either alerts or orients participants to the up-
coming target. There are four cue types: no cue (neither alerting nor orienting), central cue
(on the ﬁxation cross; alerting), double cue (two asterisks presented simultaneously above
and below ﬁ
xation; alerting), or spatial cue (a single asterisk above or below ﬁxation in the
location of the upcoming target; orienting). Immediately after the asterisk cue, the target
arrow appears and is ﬂanked by one of three ﬂanker types: congruent (arrows in the
same direction as the target arrow), incongruent (arrows in the opposite direction as the
target arrow) and neutral (horizontal lines). The alerting and orienting networks are
assessed via the different cue conditions, the executive control network is assessed via
the different ﬂanker conditions. Measures of the efﬁciency of the three attentional networks
are obtained via simple subtractions of reaction times between conditions: alerting scores
are calculated by subtracting reaction times in the double cue condition from reaction
times in the no cue condition; orienting scores are calculated by subtracting reaction
times in the spatial cue condition from reaction times in the center cue condition; executive
control scores are calculated by subtracting reaction times in the congruent ﬂanker condi-
tion from reaction times in the incongruent ﬂanker condition.
The task consists of 288 trials; 72 in each of the four cue conditions. Each trial is 4000 ms
beginning with a ﬁxation cross, followed by one of the four cue types presented for 100 ms,
followed by an average interval of 400 ms after which one of the three target stimuli is pre-
sented for 1700 ms or until the subject responds (congruent, incongruent, neutral). The task
was presented on an IBM ThinkPad laptop using e-prime professional version 2.0.
2.3.2. The ANT-Emotion
Using the same structure as the ANT, the ANT-Emotion is designed to asse ss
the orienting an d executive control attentiona l networks in r elation to emotional
information. Orienting to emotional information is assessed by using faces (neutral, angry)
as spatial cues for the upcoming target. Executive control of emotional information is
assessed by using faces (neutral, angry) as ﬂankers of the central target arrow. Participants
30 L.M. Tully et al. / Psychiatry Research 197 (2012) 29–35
indicate the direction of a central arrow ﬂanked by one of three ﬂanker types: irrelevant
angry faces, irrelevant neutral faces, or neutral lines. The ﬂanker conditions are preceded
by one of three cue types (asterisk, neutral face, angry face) presented in one of two cue po-
sitions (central, spatial – asinglecueaboveorbelowﬁxation orienting participants to the
location of the upcoming target). Face cues never precede face ﬂankers: in face ﬂanker con-
ditions the cue presented is always an asterisk; in neutral line ﬂanker conditions the cue
presented is either an angry or neutral face (see Fig. 1a and b for task illustration).
Measures of the efﬁciency of each network are obtained through simple subtractions
of reaction times between conditions. Executive control of emotional information scores
are calculated by subtracting reaction times in the neutral face ﬂanker condition from re-
action times in the angry face ﬂanker condition. Orienting to emotional information scores
are calculated by subtracting reaction times in the neutral face spatial cue condition from
the angry face spatial cue condition. Orienting to emotional information was included in
the task in order to dissociate the attentional capture of emotional information from the
executive control of emotional information. If negative emotional faces are simply more sa-
lient for high SA participants we would expect faster orienting to emotional information
compared to controls; if, as predicted, high SA participants have difﬁculty inhibiting nega-
tive emotional faces we would expect impaired executive control of emotional information.
The face stimuli are colored photographs of unfamiliar male and female faces taken from
the NIMSTIM face set (Tottenham et al., 2009); 24 faces (12 female) were selected. In an effort
to reﬂect the race and ethnicity demographics of the Greater Boston Area 50% are Caucasian, 25%
are Asian/other, and 25% are African American (retrieved August 13, 2009 from http://www.
helloboston.com/Census.Cfm). We chose to use open-mouth angry face stimuli as they are
more reliably identiﬁed as angry than the closed-mouth angry stimuli (Tottenham et al.,
2009). Similarly, we chose to use angry faces because they are more reliably identiﬁed correctly
than fear faces (Tottenham et al., 2009). Each face is shown in both an angry expression and a
neutral expression, totaling 48 face stimuli. Face photographs were made uniform by centering
thefaceina1.68″ ×2.42″ boxwithablackbackground.
The task consists of 288 trials; 96 in each of the three cue types. Each trial begins
with a ﬁxation cross for 400 ms, followed by one of the four cue types presented for
100 ms, followed by an interval of 400 ms after which one of the three target stimuli
(neutral line ﬂanker, neutral face ﬂanker, angry face ﬂanker) is presented for
1700 ms or until the subject responds. The task was presented on an IBM ThinkPad
laptop computer using e-prime professional version 2.0.
2.4. Data analysis
Data analysis was conducted with SPSS 18.0. All variables were screened for normalcy
and outliers, deﬁned as 2.5 or more standard deviations away from the mean of each group.
Outliers were replaced with the group mean accordingly. Eight scores were replaced: ﬁve
ANT scores (four high SA), and three ANT-Emotion scores (one high SA). Independent t-
tests and χ
analyses were used to assess group differences on demographic variables. Uni-
variate Analysis of Variance (ANOVAs) and Multivariate Analysis of Covariance (MANCO-
VAs) were used to assess group differences on the ANT and ANT-Emotion. Pearson
correlations and linear regressions were calculated to assess the proposed mediation
model between SA, executive control of emotional information, and social functioning.
3.1. Hypothesis 1: compared to controls, high SA participants will exhibit
attentional deﬁcits on both the ANT, speciﬁcally in the executive control
network, and the ANT-Emotion, speciﬁcally in the executive control of
Attention network scores and test statistics for group differences on
the ANT are presented in Table 2. Due to group differences in IQ, IQ was
entered as a covariate of no interest in all analyses. Results showed that
there was no difference between high SA and control participants on
any of the ANT scores. Speciﬁcally, ANOVAs revealed no group differ-
ences in overall reaction times [F(1,54)= 0.397, P =0.531] or accuracy
[F(1,54)=0.036, P =0.85]. MANCOVA with attention network scores
of the ANT (alerting, orienting, executive control) as dependent
variables revealed no main effect of group [F(3,52)=0.767, P=0.518]
or IQ [F(3,52)=0.563, P=0.642]. Similarly, follow-up ANOVAs
revealed no group differences on alerting [F(1,54)=2.373, P=0.129],
orienting [F(1,54)=0.42, P=0.839], or executive control network
scores [F(1,54)=0.127, P= 0.723]. Given that low IQ is a characteristic
feature of schizophrenia liability (Woodberry et al., 2008), removing
variance associated with IQ might be removing meaningful variance as-
sociated with the illness-related deﬁcit under investigation (Miller and
Chapman, 2009). Therefore, we also conducted these analyses without
IQ as a covariate; results remained the same. Thus, our hypothesis
that high SA individuals would exhibit deﬁcits in the executive control
network for non-emotional information was not supported (see Fig. 2).
Group means and test statistics for the ANT-Emotion are pre-
sented in Table 2. IQ was entered in all analyses as a covariate of no
interest. Results showed a single group difference in executive control
of emotional information, and no group differences on any other vari-
ables. Speciﬁcally, ANOVAs revealed no group differences in overall
reaction times [F(1,59)= 0.343, P =0.56], accuracy [F(1,59)= 0.001,
P= 0.976], or orienting to emotional information [F(1,59)= 0.139,
P= 0.71]. There were, however, group differences in executive con-
trol of emotional information: compared to controls, high SA partici-
pants were signiﬁcantly slower to respond in the angry face ﬂanker
condition versus the neutral face ﬂanker condition [F(1,59) =5.478,
P= 0.023] (see Fig. 3). To examine whether this effect was driven
by differences in IQ we conducted these analyses without IQ as a co-
variate. Results remained the same: compared to controls, high SA
participants showed impaired executive control of emotional infor-
mation [F(1,60)= 3.971, P= 0.05].
scores (see Table 1), it is possible that other aspects of psychosis-
proneness contribute to this group difference in executive control of
emotional information. However, although SPQ scores did relate to SA
(r=0.76, Pb 0.0001) and social impairments (r=0.52, Pb 0.0001),
there was no signiﬁcant relationship with executive control of emotional
information (r=− 0.02, P= 0.89). This indicates that the group differ-
ence in executive control of emotional information is not driven by
other schizotypal traits that are measured in the SPQ.
Collectively these ﬁndings demonstrate that while high SA partici-
pants do not exhibit attentional deﬁcits in relation to non-emotional in-
formation, they do exhibit a speciﬁcdeﬁcit in the executive control of
emotional information. Moreover, this deﬁcit appears to be driven by
SA, rather than psychosis-proneness more generally.
3.2. Hypothesis 2: high SA individuals will report poorer social functioning
compared to controls
Mean social functioning and group differences are displayed in
Table 1. Consistent with our hypothesis, independent t-tests revealed
high SA participants reported signiﬁcantly poorer social functioning than
controls [t(59)= 4.245, Pb 0.0001]. This replicates previous ﬁndings that
Demographics and sample characteristics.
Control group High SA
N 29 34
Gender (F/M) 20 F/9 M 19 F/15 M χ
(1)=1.136, P = 0.287
Age: mean (S.D.), [range] 28.76 (13.81),[18–65] 30.47 (14.04),[19–64] t(61)=0.486, P = 0.629
Education: mean (S.D.), [range] 14.86 (2.01), [12 –20] 14.53 (2.16), [10–20] t(60)=0.612, P = 0.543
WASI IQ: mean (S.D.), [range] 119.52 (10.77), [92–137] 111.76 (11.89), [82–132] t(61)=2.692, P = 0.009, d =0.69
Social anhedonia: mean (S.D.), [range] 6.21 (3.82), [0–14] 26.85 (6.41), [18–40] t(61)=15.184, P b 0.0001, d = 3.89
SPQ: mean (S.D.), [range] 19.97 (9.56), [9–40] 41.06 (15.47), [11 – 70] t(61) =6.372, P b 0.0001, d=1.63
Social impairment: mean (S.D.), [range] 56.30 (11.67), [36–84] 70.68 (14.19), [39–109] t(59)=4.245, P b 0.0001, d = 1.11
Cohen's d effect size.
Two subjects did not complete the Social Adjustment Scale.
31L.M. Tully et al. / Psychiatry Research 197 (2012) 29–35
high SA is associated with poorer social functioning (Blanchard et al.,
1998; Cohen et al., 2006).
3.3. Hypotheses 3: executive control of emotional information will mediate
the relationship between social anhedonia and social functioning
Following the steps required for determining mediation (see
Baron and Kenny, 1986), we conducted a series of correlation and
regression analyses across all 63 participants to determine if there
were signiﬁcant associations between 1) SA and social functioning,
2) SA and executive control of emotional information, 3) executive
control of emotional information and social functioning, and 4) SA
and social functioning whilst controlling for executive control of emo-
tional information. All four steps must be met for a mediation model
to be supported. A variable is considered a statistical mediator if the
relationship between the independent (i.e. SA) and the dependent
Fig. 1. a: Illustration of an ANT-Emotion trial assessing orienting to emotional information. Participants see a neutral or angry face cue followed by a target arrow ﬂanked with neutral
lines. Scores for orienting to emotional information are obtained through simple subtractions of RTs: RT angry face cue – RT neutral face cue. Lower scores indicate an orienting bias to
negative emotional faces. b: Illustration of an ANT-Emotion trial assessing executive control of emotional information. Participants see an asterisk cue followed by a target arrow ﬂanked
by neutral or angry faces. Scores for executive control of emotional information are obtained through simple subtractions of RTs: RT angry face ﬂankers – RT neutral face ﬂankers. Higher
scores indicate reduced ability to inhibit negative emotional faces.
32 L.M. Tully et al. / Psychiatry Research 197 (2012) 29–35
(i.e. social functioning) variable is signiﬁcantly reduced when the me-
diator (i.e. executive control of emotional information) is controlled
for (Baron and Kenny, 1986). Pearson correlations revealed social anhe-
donia correlated signiﬁcantly with social impairment (r =0.59;
Pb 0.0001), but executive control of emotional information did not corre-
late signiﬁcantly with either social anhedonia (r= 0.21, P=0.1) or social
impairment (r=− 0.17, P= 0.19). Linear regression revealed social an-
hedonia signiﬁcantly predicted social functioning [F(1,59)=31.215,
Pb 0.0001; R
=0.35] such that higher social anhedonia predicted great-
er social impairment, satisfying step 1 of the mediation model. However,
social anhedonia did not signiﬁcantly predict executive control of emo-
tional information [F(1,60)=2.741, P=0.103; R
executive control of emotional information did not signiﬁcantly predict
social functioning [F(1,58)=1.803, P = 0.185; R
=0.03]. Thus, the me-
diation model was not supported. These non-signiﬁcant ﬁndings could
be due to a small sample size; power analysis indicates that our sample
size has sufﬁcient power to detect large and medium effects (0.99 and
0.70 respectively), but low power to detect small effects (0.12).
This study examined social functioning and executive control of
emotional information in individuals at risk for developing schizo-
phrenia. Attentional processes were assessed using the ANT and our
new measure of executive control of emotional information, the
ANT-Emotion. Two main ﬁndings emerged: ﬁrst, while there were
no group differences in alerting, orienting, or executive control net-
works assessed by the ANT, high SA individuals did exhibit a speciﬁc
impairment in the executive control of emotional information rela-
tive to controls. Second, high SA individuals reported signiﬁcantly
greater social functioning impairments relative to controls.
These ﬁndings suggest that there are speciﬁcdeﬁcits in the execu-
tive control of emotional information observable in individuals with
vulnerability factors for schizophrenia. This deﬁcit was only apparent
in the presence of irrelevant angry faces, and not in the executive con-
trol network assessed by the ANT. This suggests that high SA individuals'
executive control may be intact but vulnerable, such that deﬁcits are not
detected by pure cognitive tasks, rather they only emerge when affec-
tive and cognitive information interacts. Moreover, high SA participants
did not exhibit faster orienting to emotional information indicating that
their slower reaction times in the presence of angry face ﬂankers is not
simply due to an orienting bias towards angry faces. We interpret these
ﬁndings as indicating that the impaired ability to inhibit emotional
information allows negative affective stimuli to exert inappropriate in-
ﬂuence on cognitive processes. This is consistent with prior studies indi-
cating that schizophrenia patients have an impaired ability to regulate
the inﬂuence of irrelevant negative affective information on cognitive
Attention network RT scores and accuracy of high SA and control participants.
Control group High SA group Differences between groups
ANT: mean (S.D.)
Alerting 33.98 (19.25) 42.88 (23.49) F(1,54) = 2.373, P=0.129
Orienting 46.30 (25.88) 49.55 (26.84) F(1,54) = 0.42, P = 0.839
Executive control 114.18 (40.31) 112.75 (41.76) F(1,54) = 0.127, P=0.723
Overall RT 549.42 (70.23) 566.72 (72.12) F(1,54) = 0.397, P=0.531
Accuracy (%) 97.53 (2.24) 97.53 (4.49) F(1,54)=0.036, P = 0.850
ANT-Emotion: mean (S.D.)
Executive control of emotional information − 0.76 (16.98) 7.85 (16.86) F(1,59) = 5.478, P =0.023, η
Orienting to emotional information 0.03 (15.07) − 0.04 (17.40) F(1,59) =0.139, P =0.71
Overall RT 520.87 (73.1) 541.59 (88.14) F(1,59) =0.343, P = 0.56
Accuracy (%) 95.92 (2.86) 95.99 (2.53) F (1,59) =0.001, P = 0.976
Six control participants did not complete the ANT.
=partial eta squared effect size.
IQ was entered as a covariate of no interest. There was no main effect of IQ. Results remained the same in all analyses when IQ was not entered as a covariate.
One subject did not complete the ANT-Emotion.
Fig. 2. There were no differences between control and high SA participants on any of
the three attentional networks of the ANT, indicating that high SA is not characterized
by attentional deﬁcits in relation to nonemotional stimuli.
Fig. 3. High SA participants demonstrated impaired executive control of emotional
information: compared to control participants, high SA participants were signiﬁcantly
slower to respond to the target in the angry face ﬂanker condition versus the neutral
face ﬂanker condition.
33L.M. Tully et al. / Psychiatry Research 197 (2012) 29–35
processes: patients show increased latencies on the emotional Stroop
(Bentall and Kaney, 1989), and rate faces as less trustworthy following
negative affective primes (Hooker et al., 2011). Moreover, this impair-
ment may be speciﬁcally associated with SA: Suslow et al. (2003) exam-
ined the inﬂuence of subliminal affective face primes on valence
judgments of neutral Chinese characters in schizophrenia patients with
and without SA. Compared to patients without SA, high SA patients
showed persistent negative evaluative shifts, indicating that impaired
regulation of irrelevant negative affect may be characteristic of high SA.
Our results are consistent with this, and indicate that it may also be char-
acteristic of individuals at risk for developing schizophrenia.
Our ﬁndings also suggest that the deﬁcit in executive control may
be speciﬁc to the inhibition of socially relevant emotional informa-
tion. Using the emotion Stroop with non-socially speciﬁc negative
words, Mohanty et al. (2008) did not ﬁnd inhibitory deﬁcits in SA.
However, our ﬁndings, and those of Suslow et al. (2003), suggest
that there is a deﬁcit in the presence of socially relevant stimuli (i.e.
negative emotional faces). This could contribute to the social
withdrawal and dysfunction characteristic of high SA: if individuals
struggle to inhibit socially relevant negative affective information, it
follows that they would withdraw from unnecessarily unpleasant
experiences (Suslow et al., 2003). Future studies could examine this
hypothesis further by adapting the ANT-Emotion to contain both so-
cial and non-social ﬂankers, or by using an emotional Stroop with
categories of social and non-social negative words.
Consistent with our second hypothesis, high SA related to poorer
social functioning. This replicates previous ﬁndings that high SA is as-
sociated with social impairments (Blanchard et al., 1998; Cohen et al.,
2006), providing further evidence for the presence of social deﬁcits in
people at increased risk for schizophrenia. However, executive control
of emotional information did not mediate the relationship between SA
and social functioning. There are two possible explanations for this.
First, it could be that there is a mismatch between the speciﬁcity of
the domain measured by the ANT-Emotion and the broad assessment
of social functioning obtained by the SAS-SR. We chose the SAS-SR be-
cause it is a standard measure used in the ﬁeld. However, like many so-
cial functioning assessments, it is broad in scope. Our ﬁnding suggests
that developing measures of social functioning that more accurately re-
ﬂect the underlying social cognitive processes necessary for successful
social interactions may result in more success in revealing any underly-
ing relationship between SA, social cognitive deﬁcits, and social impair-
ments. Second, although our sample size is appropriate to detect
mediation of large and medium effects, it is not big enough to detect
small effects. Future research should examine the relationship between
executive control of emotional information and social functioning using
alternate measures of social impairments in sample sizes large enough
to detect small effects.
Several limitations of the current study should be acknowledged.
First, the ANT-Emotion is a newly created task; replications in inde-
pendent samples are needed and improvements may be warranted.
Of particular interest is the effect of stimulus onset asynchrony
(SOA). Unlike the ANT, the ANT-Emotion uses a constant SOA
(400 ms). This might allow participants to predict the onset of target
stimuli, possibly masking any orienting biases that might exist in the
sample, or even reducing group differences on executive control of
emotional information. However, despite these potential improve-
ments, the task provides valid and compelling data that high SA indi-
viduals have impaired executive control of emotional information.
The ANT-Emotion is based on the ANT – a well-validated and reliable
measure of attention networks, of which the executive control piece
is based on another well-validated task – the Eriksen ﬂanker
(Eriksen and Eriksen, 1974). Furthermore, the interference effect of
irrelevant affective information has been established in other tasks
such as the emotional Stroop, therefore it is unlikely that our ﬁndings
are a consequence of an invalid task. A second limitation is that the
ANT-Emotion currently only uses angry faces, hence it is unclear if
our ﬁndings would generalize to other facial expressions, both nega-
tive and positive. In light of the “anhedonia paradox” (Horan et al.,
2008), which suggests that the anticipation of reward, not consum-
mation, is the primary deﬁcit in schizophrenia, the relationship be-
tween SA and the regulation of positive emotional information is of
particular interest. Our ﬁndings indicate that high SA is associated
with difﬁculties regulating negative affective information and could
suggest a complementary deﬁcit in the facilitation of positive affec-
tive information, thought to underlie the observed deﬁcit in anticipa-
tory pleasure in anhedonia (Pizzagalli, 2010). Future studies could
examine both inhibitory and facilitatory mechanisms in the context of
emotional information and SA to delineate this further. Finally, there
were group differences on the SPQ. Although these differences are
expected given the close relationship between positive and negative as-
pects of schizotypy (Wuthrich and Bates, 2006), and they were shown to
have no impact on the group differences in executive control of emo-
tional information, future studies could investigate the speciﬁcinﬂuence
of different aspects of psychosis-proneness on executive control of emo-
In summary, the current study demonstrates a deﬁcit in the execu-
tive control of socially relevant emotional information in socially anhe-
donic individuals. Additionally, SA predicted poorer social functioning,
consistent with the idea that social impairments are a core characteristic
of psychosis-prone populations (Meehl, 1962). These ﬁndings have im-
plications for how high-risk populations process negative affective infor-
mation in social contexts. Further research is needed to identify the
precise nature of these deﬁcits, and how they might contribute to the so-
cial dysfunction characteristic of schizophrenia.
Barch, D.M., 2005. The cognitive neuroscience of schizophrenia. Annual Review of Clinical
Psychology 1 (1), 321–353.
Baron, R.M., Kenny, D.A., 1986. The moderator-med iator variable distinction in social
psychological research: conceptual, strategic, and statistical considerations. Journal
of Personality and Social Psychology 51 (6), 1173–1182.
Bentall, R.P., Kaney, S., 1989. Content speciﬁc information processing and persecutory
delusions: an investigation using the emotional Stroop test. The British Journal of
Medical Psychology 62, 355–364.
Blanchard, J.J., Mueser, K.T., Bellack, A.S., 1998. Anhedonia, positive and negative affect,
and social functioning in schizophrenia. Schizophrenia Bulletin 24 (3), 413–424.
Brown, L.H., Silvia, P.J., Myin-Germeys, I., Kwapil, T.R., 2007. When the need to belong goes
wrong: the expression of social anhedonia and social anxiety in daily life. [Article].
Psychological Science, 18(9). Wiley-Blackwell, pp. 778–782.
Cohen, A.S., Leung, W.W., Saperstein, A.M., Blanchard, J.J., 2006. Neuropsychological
functioning and social anhedonia: results from a community high-risk study.
Schizophrenia Research 85 (1–3), 132–141.
Psychophysiology 37 (05), 711.
Eckblad, M., Chapman, L. J., Chapman, J. P., Mishlove, M. 1982. The Revised Social Anhedonia
Scale. Unpublished Test.
Eriksen, B., Eriksen, C., 1974. Effects of noise letters upon the identiﬁcation of a target
letter in a nonsearch task. Perception & Psychophysics 16 (1), 143–149.
Fan, J., McCandliss, B.D., Sommer, T., Raz, A., Posner, M.I., 2002. Testing the efﬁciency
and independence of attentional networks. Journal of Cognitive Neuroscience 14
First, M.B., Spitzer, R.L., Gibbon, M., Williams, J.B.W., 2002. Structured Clinical Interview
for DSM-IV-TR Axis I Disorders, Research Version, (SCID-I/P). Biometrics Research,
Patient Edition. New York State Psychiatric Institute, New York.
Gooding, D.C., Braun, J.G., Studer, J.A., 2006a. Attentional network task performance in
patients with schizophrenia-spectrum disorders: evidence of a speciﬁcdeﬁcit.
Schizophrenia Research 88 (1–3), 169–178.
Gooding, D.C., Matts, C.W., Rollmann, E.A., 2006b. Sustained attention deﬁcits in relation
to psychometrically identiﬁed schizotypy: evaluating a potential endophenotypic
marker. Schizophrenia Research 82 (1), 27–37.
Green, M.F., Kern, R.S., Braff, D.L., Mintz, J., 2000. Neurocognitive deﬁcits and functional
outcome in schizophrenia: are we measuring the “right stuff”? Schizophrenia Bulletin
26 (1), 119–136.
Heinrichs, R.W., Zakzanis, K.K., 1998. Neurocognitive deﬁcit in schizophrenia: a quantitative
review of the evidence. Neuropsychology 12 (3), 426–445.
Hooker, C.I., Tully, L.M., Verosky, S.C., Fisher, M., Holland, C., Vinogradov, S., 2011. Can I
trust you? Negative affective priming inﬂuences social judgments in schizophrenia.
Journal of Abnormal Psychology 120 (1), 98–107.
Horan, W.P., Brown, S.A., Blanchard, J.J., 2007. Social anhedonia and schizotypy: the
contribution of individual differences in affective traits, stress, and coping. Psychiatry
Research 149 (1–3), 147–
34 L.M. Tully et al. / Psychiatry Research 197 (2012) 29–35
Horan, W.P., Blanchard, J.J., Clark, L.A., Green, M.F., 2008. Affective traits in schizophrenia
and schizotypy. Schizophrenia Bulletin 34 (5), 856–874.
Kerns, J.G., Berenb aum, H., 2000. Aberrant semantic and affective processing in people
at risk for psychosis. Journal of Abnormal Psychology 109 (4), 728–732.
Kwapil, T.R., 1998. Social anhedonia as a predictor of the development of schizophrenia-
spectrum disorders. Journal of Abnormal Psychology 107 (4), 558–565.
Landis, J.R., Koch, G.G., 1977. The measurement of observer agreement for categorical
data. Biometrics 33 (1), 159–174.
Meehl, P.E., 1962. Schizotaxia, schizotypy, schizophrenia. American Psychologist 17
Miller, G.A., Chapman, J.P., 2009. Misunderstanding analysis of covariance. Journal of
Abnormal Psychology 110 (1), 40–48.
of emotion-related effects on attentional processing in schizotypy. Schizophrenia Re-
search 103 (1–3), 129–137.
Nuechterlein, K.H., Asarnow, R.F., Subotnik, K.L., Fogelson, D.L., Ventura, J., Torquato, R.D.,
Dawson, M.E., 1998. Neurocognitive vulnerability factors for schizophrenia: conver-
gence across genetic risk studies and longitudinal trait-state studies. In: Lenzenweger,
M.F., Dworkin, R.H. (Eds.), Origins and Development of Schizophrenia. Advances in
Experimental Psychopathology. American Psychological Association, Washington,
DC, pp. 299–327.
Ochsner, K.N., Gross, J.J., 2005. The cognitive control of emotion. Trends in Cognitive
Sciences 9 (5), 242–249.
Phillips, L.K., Seidman, L.J., 2008. Emotion processing in persons at risk for schizophrenia.
Schizophrenia Bulletin 34 (5), 888–903.
Pizzagalli, D., 2010. The “anhedonia paradox” in schizophrenia: insights from affective
neuroscience. Biological Psychiatry 67, 899–901.
Posner, M.I., Petersen, S.E., 1990. The attention system of the human brain. Annual Review
of Neuroscience 13 (1), 25–42.
Posner, M.I., Rothbart, M., Vizueta, N., Levy, K., Evans, D., Thomas, K., Clarkin, J., 2002.
Attentional mechanisms of borderline personality disorder. Proceedings of the National
Academy of Sciences of the United States of America 99 (25), 16366–16370.
Raine, A., 1991. The SPQ: A Scale for the Assessment of Schizotypal Personality Based
on DSM-III-R Criteria. Schizophrenia Bulletin 17 (4), 555–564.
Suslow, T., Roestel, C., Arolt, V., 2003. Affective priming in schizophrenia with and
without affective negative symptoms. European Archives of Psychiatry and Clinical
Neuroscience 253 (6), 292–300.
Tallent, K.A., Gooding, D.C., 1999. Working memory and Wisconsin Card Sorting Test
performance in schizotypic individuals: a replication and extension. Psychiatry Research
89 (3), 161–170.
Tottenham, N., Tanaka, J., Leon, A.C., McCarry, T., Nurse, M., Hare, T.A., 2009. The NimStim
set of facial expressions: judgments from untrained research participants. Psychiatry
Research 168 (3), 242–249.
Urbanek, C., Neuhaus, A.H.M., Opgen-Rhein, C., Strathmann, S., Wieseke, N., Schaub, R., Hahn,
E., Dettling, M., 2009. Attention network test (ANT) reveals gender-speciﬁcalterations
of executive function in schizophrenia. Psychiatry Research 168 (2), 102–109.
Wan, L., Crawford, H.J., Boutros, N., 2006. P50 sensory gating: impact of high vs. low
schizotypal personality and smoking status. International Journal of Psychophysiology
60 (1), 1–9.
Wang, K., Fan, J., Dong, Y., Wang, C.-q., Lee, T.M.C., Posner, M.I., 2005. Selective impairment of
attentional networks of orienting and executive control in schizophrenia. Schizophrenia
Research 78 (2–3), 235–241.
Wechsler, D., 1999. Wechsler Abbrevi ated Scale of Intelligence. The Psychological
Corporation, San Antonio, TX.
Weissman, M.M., Prusoff, B.A., Thompson, W.D., 1978. Social adjustment by self-report
in a community sample and in psychiatric outpatients. The Journal of Nervous and
Mental Disease 166 (5), 317–326.
Woodberry, K.A., Giuliano, A.J., Seidman, L.J., 2008. Premorbid IQ in schizophrenia: a
meta-analytic review. The American Journal of Psychiatry 165 (5), 579–587.
Wuthrich, V.M., Bates, T.C., 2006. Conﬁrmatory factor analysis of the three-factor structure
of the schizotypal personality questionnaire and Chapman schizotypy scales. Journal of
Personality Assessment 87 (3), 292–304.
35L.M. Tully et al. / Psychiatry Research 197 (2012) 29–35