Startle modulation in childhood anxiety disorders1
Baseline and Affective Startle Modulation by Angry and Neutral Faces
in 4-8 Year Old Anxious and Non-Anxious Children
Allison M. Waters1, David L. Neumann1, Julie Henry1,
Michelle G. Craske2, & Edward M. Ornitz3
1School of Psychology, Griffith University, Gold Coast, Australia
2Department of Psychology, University of California, Los Angeles, USA
3Department of Psychiatry and Biobehavioral Sciences, University of California,
Los Angeles, USA
Address for Correspondence:
Allison M. Waters Ph.D., School of Psychology, Griffith University, Gold Coast Qld 4222,
Australia. Phone: +61-7-5552-8132, Email: email@example.com
Startle modulation in childhood anxiety disorders2
The present study examined the magnitudes of startle blink reflexes and electrodermal responses
in 4 to 8 year old high anxious children (N = 14) and non-anxious controls (N = 11). Responses
were elicited by 16 auditory startle trials during a baseline phase prior to an affective modulation
phase involving 12 startle trials presented during angry and neutral faces. Results showed
significant response habituation across baseline trials and equivalent response magnitudes
between groups during the baseline phase. The modulation of response magnitudes during angry
and neutral faces did not differ significantly in either group. However, high anxious children
showed larger responses overall compared with non-anxious control children during the
affective modulation phase. Moreover, greater anxiety severity and larger startle reflexes were
associated with poorer accuracy in rating neutral faces as neutral in high anxious children.
Results may reflect elevated reactivity to threat contexts in 4 to 8 year old high anxious versus
Startle modulation in childhood anxiety disorders3
Baseline and Affective Startle Modulation by Angry and Neutral Faces
in 4-8 Year Old Anxious and Non-Anxious Children
Anxiety disorders are the most commonly diagnosed psychiatric disorders and one of the
most significant health problems in terms of global burden of disease, exceeding the vast
majority of physical health diseases (Murray & Lopez, 1996). Childhood-onset anxiety is a
debilitating condition affecting up to 15-20% of youths and is a significant risk factor for other
emotional and behavioural disorders including adolescent and adult anxiety (McGee, Feehan,
Williams, & Anderson, 1992), depression (Hayward, Killen, Kraemer, & Taylor, 2000; Pine,
Cohen, Gurley, Brook, & Ma 1998), eating disorders (Patton, 1998) and substance disorders
(Merikangas, Avenevoli, Dierker, & Grillon, 1999). Childhood anxiety disorders are also
associated with debilitating academic and vocational functioning (Kessler, Foster, Saunders, &
Stang, 1995), impaired social competence (Spence, Donovan, & Brechmann-Touissant, 1999),
and if left to persist into adulthood, a range of socio-economic costs including unemployment,
days lost from work, hospitalisation and medication (Waghorn, Chant, White, & Whiteford,
Although research on childhood anxiety disorders lags behind research on adult anxiety
in general, a combination of both genetic and environmental influences are thought to play
contributory roles to the development and maintenance of childhood anxiety disorders (see
Craske & Waters, 2005, for a review). The understanding of neurophysiological processes that
underlie childhood anxiety disorders is particularly limited and based primarily on studies of
children in mid- to late-childhood and adolescence (see Ornitz, 1999, for review). The current
study extends this literature by examining baseline and affective modulation of the startle reflex
in high anxious and low anxious control children between 4 and 8 years of age.
Startle modulation in childhood anxiety disorders4
The magnitude of the human startle reflex indexes defensive responding to aversive
stimuli and associated contexts (see Davis, 1998, and Grillon & Baas, 2003, for reviews).
Rodent studies have shown that modulation of the startle reflex by aversive contexts (such as
long duration bright lights) is mediated by the bed nucleus of the stria terminalis, whereas fear-
induced modulation of the startle reflex by explicit threat cues (such as a cue previously paired
with an electric shock) is mediated by the central nucleus of the amygdala (Davis, 1998; Walker,
Toufexis, & Davis, 2003). As tendencies to react with fear and avoidance in a range of situations
associated with anxiety-provoking stimuli are hallmarks of emotional and behavioural
characteristics of anxious children (American Psychiatric Association, 1994), the study of startle
reflex modulation by explicitly aversive stimuli and associated contexts may elucidate the
neurophysiological processes that underlie anxiety disorders in children. Moreover, as an
involuntary response present from birth, the startle reflex is a neurophysiological measure
especially suited to the study of defensive responding in young children (see Ornitz, 1999, for a
Numerous studies have demonstrated increased startle reflex magnitude in adults with
anxiety disorders in contexts associated with threat but not in response to explicit threat cues.
For example, Grillon and colleagues demonstrated that adults with panic disorder and post-
traumatic stress disorder (PTSD) showed sustained elevations in “baseline” startle reflexes
elicited at the commencement of the experimental session that later involved the delivery of
unpleasant electric shocks signalled by a cue. In contrast, participants with anxiety disorders did
not differ from controls in startle reflex magnitude when elicited during explicit cues of threat
(e.g., Grillon, Ameli, Goddard, Woods, & Davis, 1994; Grillon, Morgan, Davis & Southwick,
1998; Grillon & Ameli 1998; Grillon & Morgan, 1999). Baseline startle reflexes were not
elevated when participants with PTSD were explicitly informed that no aversive stimuli would
be presented during the entire experimental procedure (e.g., Grillon et al., 1998). Another study
of police officers (Pole, Neylan, Best, Orr, & Marmar, 2003) demonstrated that severity of
Startle modulation in childhood anxiety disorders5
PTSD symptoms correlated with startle reflex magnitude during low and medium threat
conditions, when informed that no shock would be delivered, but did not correlate with startle
reflex magnitude during high threat conditions that explicitly signalled potential shock. Thus, in
all of these studies, adults with anxiety disorders did not show larger startle reflexes compared
with low anxious adults in response to cues signalling explicit threat of electric shock. Instead,
they showed elevated responding in contexts associated with threat (see Grillon, 2002).
Elevated baseline startle reflexes in individuals with anxiety disorders is thought to
represent increased anxiety about the laboratory context associated with threat of an aversive
stimulus (Grillon et al., 1998b). Indeed, Grillon, Baas, Lissek, Smith, and Milstein (2004)
concluded that “sustained contextual anxiety, but not phasic explicit cue fear, differentiates
anxiety-disordered from non-anxiety-disordered individuals” (p. 916). This conclusion reflects
what has been called the “strong situation” effect (Lissek, Pine, & Grillon, 2006), in which
anxious and non-anxious individuals respond equally to intensely aversive stimuli, whereas only
anxious individuals also respond strongly to low intensity aversive stimuli relative to non-
A small literature exists on fear-potentiated startle modulation in children with most
studies based on children at risk for anxiety disorders by virtue of parental anxiety. For example,
child and adolescent offspring of parents with anxiety disorders (age range 7 to 18 years) have
exhibited elevated startle reflexes during the preceding baseline phase as well as during
darkness-induced fear-potentiation protocols and during threat of an air blast to the larynx
compared with low-risk offspring (Grillon, Dierker, & Merikangas, 1997; 1998; Merikangas et
al., 1999). Similarly, whereas there were no group differences in startle reflex magnitudes
during pictures of fear-relevant visual (i.e., snake picture) and auditory (i.e., 1000 Hz, 100 dB
tone) stimuli, offspring of anxious parents (age range 7 to 12 years) displayed significantly
higher electrodermal activity during the resting baseline and during the inter-trial intervals
(Turner, Beidel, & Roberson-Nay, 2005).
Startle modulation in childhood anxiety disorders6
A significant limitation in applying fear-potentiation protocols with young children is
that threat of electric shock, air blasts to the larynx, and periods of time in complete darkness
may be too aversive for young children to tolerate. One avenue for overcoming this limitation
has been to employ a picture viewing paradigm that assesses affective modulation of the startle
reflex (e.g., Bradley, Cuthbert, & Lang, 1993; Cuthbert, Bradley, & Lang, 1996). An extensive
literature with adults has shown repeatedly that blink reflexes are larger while viewing
unpleasant pictures in comparison to blinks elicited during neutral and pleasant pictures
(Cuthbert et al., 1996). Affective modulation of the blink reflex is thought to reflect the
activation of defensive motivational processes by the match in affective valence between
unpleasant picture stimuli and the aversive auditory startle-eliciting stimulus (Bradley, Cuthbert,
& Lang, 1999). Moreover, adult studies have demonstrated a robust relationship between
enhanced affective startle modulation and high fearfulness (e.g., Cook, Hawk, Davis, &
Stevenson, 1991; Hawk, Stevenson, & Cook, 1992).
Studies of affective modulation in normally developing children have produced results
divergent from those with adults. McManis and colleagues were unable to demonstrate startle
facilitation during unpleasant compared with neutral pictures in 7 to 10 year old children (e.g.,
McManis, Bradley, Cuthbert, and Lang, 1995) and found that girls but not boys showed the
expected increase in startle reflex magnitude during unpleasant than pleasant pictures
(McManis, Bradley, Berg, Cuthbert, and Lang (2001). Cook, Hawk, Hawk, and Hummer (1995)
also did not find the adult pattern using affectively valent script-induced imagery in school-age
children; almost identical startle magnitude was found during imagery of pleasure, joy, sadness,
fear, and anger. Finally, Waters, Lipp, and Spence (2005) found that startle reflex magnitude did
not differ significantly during unpleasant compared with neutral or pleasant pictures in 8 to 12
year old children.
Studies examining anxiety-related differences in affective modulation of startle in
children have also produced conflicting results. Cook et al. (1995) found that startle responses of
Startle modulation in childhood anxiety disorders7
children who scored higher on a fear survey schedule were smaller during unpleasant than
pleasant imagery and were smaller in high fear compared with low fear children. Waters et al.
(2005) found that anxious children showed larger blink reflexes overall during unpleasant,
neutral and pleasant pictures compared with low anxious children. As affective startle
modulation is strongest for startle reflexes elicited during highly arousing affective stimuli
(Cuthbert et al., 1996), inconsistent results between children and adults have been attributed to
the less arousing picture stimuli used with children compared with adults and the unreliability of
imagery procedures with children (see McManis et al., 2001; Waters et al., 2005).
Emotional face stimuli in a picture viewing paradigm may be an approach that
overcomes these limitations with very young children. Other advantages include the greater
ecological validity of face stimuli than other emotion-evoking stimuli (Mogg & Bradley, 1998)
and that emotional faces, such as angry ones, receive preferential processing over other salient
stimuli due to their evolutionary association with threat to the safety of humans across the
lifespan (see Öhman & Mineka, 2001, for a review). Recent functional neuroimaging and
cognitive science studies with children and adolescents between 8 and 18 years of age show
perturbed amygdala activation and attention allocation to angry and fearful faces compared with
neutral ones in anxious versus non-anxious youths (e.g., Hadwin et al., 2003; Stirling, Eley, &
Clark, 2006; Monk et al., 2006; Thomas et al., 2001; Waters & Lipp, in press; Waters, Mogg,
Bradley, & Pine, in press). Moreover, recent studies of startle modulation by emotional faces
with non-selected adults demonstrated larger startle reflex magnitudes when participants viewed
angry faces versus other types of expressions (i.e., fearful, neutral, and happy) (Springer, Rosa,
McGetrick, & Bowers, 2007). These findings suggest that emotional faces may possess greater
sensitivity than broad-based affective picture stimuli for assessing neurophysiological processes
that underlie anxiety in children, such as enhanced defensive responding to threat and associated
The present study
Startle modulation in childhood anxiety disorders8
This study examined baseline and affective startle modulation in young anxious and non-
anxious children between 4 and 8 years of age. This age range was selected to expand on
previous studies that have not assessed children below 7-8 years of age. The study assessed
whether high anxious children would show larger startle reflexes compared with non-anxious
control children during a baseline phase prior to an affective modulation phase in which angry
faces were presented. The study also examined whether high anxious children would show
enhanced affective startle modulation relative to non-anxious control children, as indexed by
larger startle reflexes during angry compared with neutral faces. Angry faces were selected
because they are stronger signals of the presence of threat than are fearful faces (Whalen et al.,
2001) and have been shown to elicit larger startle reflex magnitudes than fearful, happy or
neutral expressions in adults (Springer et al., 2007). Moreover, in the absence of affective
modulation studies in children using face stimuli, we compared responses during angry faces
with those during neutral faces based on the cognitive science literature which has shown
reliable anxiety-related attentional bias effects for angry faces compared with neutral ones in
children (e.g., Hadwin et al., 2003; Stirling et al., 2006; Monk et al., 2006; Waters et al., in
press). Skin conductance responses were also recorded as an adjunct or complementary measure
to the startle blink reflex because it will reflect a defensive response to the startle eliciting
stimulus. Similar results across the two measures were expected.
Participants were 25 children aged 4 years, 4 months to 8 years, 6 months (12 girls; 13
boys). Of these children, 14 had a clinically-significant diagnosis of an anxiety disorder (M age
= 6.08, SD = 1.44; 7 girls; 7 boys) and 11 were non-anxious controls (M age = 6.00; SD = 1.41;
5 girls; 6 boys). Written informed consent was obtained from parents prior to children’s
participation in an experimental protocol that was approved by the Institutional Human
Research Ethics Committee. High anxious children were referred by paediatricians, local
Startle modulation in childhood anxiety disorders9
community mental health agencies, school guidance counsellors and parents to the Griffith
University Child and Adolescent Anxiety Disorders Research Program for assessment and
treatment. Non-anxious control children were recruited from first year undergraduate
psychology students who had children 4 to 8 years old. These students received course credit in
exchange for their child’s participation.
All children were born in Australia and spoke English as their first language. Twenty-
one children (85%) lived with parents who were married and four children (16%) had parents
who were separated/divorced. Children tended to come from average income Australian families
according to the Daniel Prestige Scale (1983), a measure of Australian occupational prestige.
There were no significant differences between groups on demographic variables.
Diagnostic status of high anxious children was determined using the parent interview
schedule of the Anxiety Disorders Interview Schedule for DSM-IV (ADIS-C; Silverman &
Albano, 1996) in which a clinician severity rating (CSR) of 4 or greater (scale 0 to 8) for at least
their principal diagnosis was used to determine clinical significance (Silverman & Albano,
1996). Thus, inclusion criteria for anxious children was a principal diagnosis of an anxiety
disorder with an ADIS-C CSR of 4 or greater, in the absence of externalising disorders,
developmental disorders, psychosis, organic brain damage or vision impairments. Major
depressive disorder (MDD) was not a reason for exclusion as long as it was not the principal
diagnosis. No child met criteria for MDD. Thus, of the 14 high anxious children, six had a
principal diagnosis of specific phobia, three had generalised anxiety disorder, three had
separation anxiety disorder, and two had social phobia. The mean ADIS-C CSR of children’s
principal anxiety diagnosis was 6.43 (SD = 1.91) and comorbidity between anxiety disorders
was high with children having an average of 3.64 anxiety diagnoses (SD = 2.13), predominantly
specific phobias. The six children with a principal diagnosis of specific phobia all had comorbid
diagnoses of either generalised anxiety disorder, social phobia or separation anxiety disorder,
suggesting that emotional faces would be as pertinent to these children as those with principal
Startle modulation in childhood anxiety disorders 10
anxiety diagnoses that are interpersonally-focused. Parents of high anxious children also
completed the parent version of the Spence Children’s Anxiety Scale (SCAS-P; Nauta et al.,
2004; Spence, 1998) which is a questionnaire containing a total score and subscale scores
corresponding to each of the DSM-IV anxiety disorder diagnoses.
The non-anxious control group was determined by total scores within the non-clinical
range on the SCAS-P (Nauta et al., 2004). Consistent with this, the mean SCAS-P total score of
31.53 (SD = 13.52) for high anxious children and 16.22 (SD = 9.71) for non-anxious control
children are comparable to the normative means for anxiety-disordered children (M = 31.8, SD =
14.1) and non-clinical controls respectively (M = 14.2; SD = 9.7; Nauta et al., 2004).1
Stimuli and Apparatus
Electrophysiological materials, equipment and data acquisition. Auditory startle-
eliciting stimuli (105 dB, 0 ms rise time, 50 ms white noise bursts) were presented binaurally
through stereophonic headphones (Sony, Model MDRV700). A Dell 19” CRT colour monitor
was used to present the silent animal documentary used during the baseline phase and angry and
neutral faces used during the modulation phase. To record electromyogram (EMG) activity of
the orbicularis oculi associated with the startle reflex blink, two miniature Ag/AgCl electrodes
(Med Assoc., TDE-023-48) were placed beneath the right eye approximately 10 mm apart edge
to edge, and 9 to 11 mm below the lower lid margin (e.g., Ornitz, Guthrie, Kaplan, Lane, &
Norman, 1986). The lateral electrode was placed 5 mm medial to the outer canthus. Vertical (lid
movement) and horizontal (left and right eye gaze direction) electro-oculogram (EOG) activity
associated with blink activity and eye movement was recorded for data quality control purposes
(see Response Definitions, Data Screening and Statistical Analysis section), using two Ag/AgCl
electrodes 1 cm in diameter. The vertical EOG electrodes were placed above the left eyebrow
and on the cheek, equidistant from the pupil. The horizontal EOG electrodes were placed at the
outer canthi. Skin conductance responses (SCRs) were recorded from two Ag/AgCl electrodes 1
cm in diameter (Grass, F-E9M-60-5), placed on the distal phalanx of the index and middle
Startle modulation in childhood anxiety disorders 11
fingers of the non-dominant hand. A ground electrode was placed on the centre of the forehead.
The impedance level of electrodes was 15 KOhm or less.
Data from EMG, EOG, and SCR recordings were acquired using a Grass Instruments
Amplifier System (Model 15RXI) and were digitized and sampled on-line using National
Instruments LabVIEW Programming Software (v7) installed on a Dell Precision Workstation
computer. EMG was full-wave rectified. EMG and EOG measures were AC amplified at a gain
of 10,000 for EMG and horizontal EOG and 5000 for vertical EOG, using a Grass Instruments
Model 15A54 Quad AC amplifier installed in the Grass Instruments Model 15 Amplifier
System. The low and high frequency cut-off values were 30 Hz and 1000 Hz for EMG, .01 Hz
and 30 Hz for vertical EOG, and 10 Hz and 30 Hz for horizontal EOG. SCR was DC amplified
at a gain of 2000.
Face Stimuli. Still pictures of 3 male and 3 female young adults of Caucasian
appearance taken from the NimStim Face Stimulus Set (www.macbrain.org) were used for each
category of angry and neutral facial expressions (12 faces in total). Validity ratings of these
faces indicate high agreement by both children and adults about the expressed facial emotions
(Tottenham, Borscheid, Ellertson, Marcus, & Nelson, 2002). All facial images were digitized
files sized to 506 pixels wide x 650 pixels high presented in full colour surrounded by a dark
grey background in the centre of the 19” Dell computer monitor.
Subjective Ratings of Face Stimuli. Children provided ratings of the emotion shown on
each of the 12 faces. Children were shown a colour picture of each face in random order and
were asked to say what emotion they thought each face was showing, if any. The research
assistant recorded if children correctly identified the relevant faces as angry or neutral and
recorded verbatim any other emotions children reported. Children were told that they could say
they were unsure if they did not know what emotion the face showed. All responses were
subsequently coded into categories of neutral, angry, scared, sad, happy, surprised, and unsure.
Startle modulation in childhood anxiety disorders 12
Upon initial contact with parents of high anxious children, a brief telephone screening
was conducted based on the ADIS-C to ensure that anxiety was the child’s principal problem
and eligible families were invited to attend an assessment session at the Griffith University
Psychology Clinic. High anxious children completed the experiment in a research laboratory
opposite the psychology clinic in which parents completed questionnaires and the ADIS-C
interview. The parent interview schedule of the ADIS-C was administered with the attending
parent/s by a doctorate-level student in clinical psychology trained to reliability in the ADIS-C
administration by matching diagnoses and clinical severity ratings for each diagnosis from
interviews conducted by a trained diagnostician. All ADIS-C interviews were reviewed during
clinical supervision for reliability and validity.
When parents of non-anxious control children contacted the research team in response to
study advertisements, the same brief telephone screening was conducted to ensure that anxiety
was not a principal concern for their child, and an appointment time was made with eligible
families. Non-anxious control children completed the experiment in the same research
laboratory as anxious children while their parents completed the SCAS-P.
After orientation to the laboratory setting, children and parent/s were informed that
children’s physical responses would be recorded while they heard loud, crackly sounds through
headphones and watched a silent animal movie followed by seeing some faces of people
showing different feelings, including calm and angry ones, presented on the computer monitor.
Children were seated in an air-conditioned, experimental room adjacent to the control room and
were in constant contact with a research assistant via a closed circuit camera and intercom
system. After electrode connection, children completed a 5 min rest period during which time
they sat quietly watching a silent DVD animal documentary on the computer monitor. The
computer monitor was set approximately 1 m in front of the child at eye level. Children were
then fitted with headphones and were informed they would hear a number of loud crackly
sounds through the headphones while they continued to watch the silent animal documentary.
Startle modulation in childhood anxiety disorders 13
Children received a single startle stimulus to reduce initial reactivity prior to completing the
experiment; it was excluded from analyses.
The startle baseline phase involved the recording of physiological responses to 16
startle stimuli while children continued to view the silent animal documentary to maintain their
focus (e.g., Ornitz, Russell, Yuan, & Liu, 1996). The mean interstimulus interval (ISI) between
startle trials was 22 sec (range 20-24 sec). Children were then instructed they would continue to
hear the crackly sounds through the headphones while faces appeared on the computer screen
showing different feelings. Elapsed time between the two phases was no longer than 2 min on
average. During the affective modulation phase, physiological responses were recorded to a
further 12 startle stimuli also with a mean ISI of 22 sec (range 20-24 sec) delivered during 6
angry faces (3 male; 3 female) and 6 neutral faces (3 male; 3 female) shown individually on the
computer monitor for 6 s each. Angry and neutral faces were presented across 3 blocks with one
male and one female face per block. Angry and neutral face blocks were presented in alternating
order and were counterbalanced across participants. A startle stimulus was delivered at random
between 3 and 5 sec after the onset of each face. A white fixation cross set against a dark grey
background was displayed in-between face presentations and no startle stimuli were delivered
during these periods. After the affect modulation phase, the headphones and electrodes were
removed. The children were presented with each of the 12 faces again in the same order as
before and asked to say out loud what feeling they thought was shown on each face. Their
responses were recorded verbatim. Children were then debriefed and returned to their parents.
All anxious children commenced group-based, cognitive-behavioural treatment for anxiety
disorders within two weeks after participating in this experiment.
Response definitions, data screening and statistical analysis
The startle reflex was measured by electromyogram (EMG) activity of the orbicularis
oculi and by skin conductance responses (SCR). Eye blink measures the muscle response
whereas SCR measures the sympathetically mediated arousal response to a startle stimulus (e.g.,
Startle modulation in childhood anxiety disorders 14
Samuels, Hou, Langley, Szabadi, & Bradshaw, 2007). Electro-oculogram (EOG) activity
associated with vertical lid movement and left and right eye gaze direction was measured for
quality control purposes when quantifying EMG (Gehricke, Ornitz, & Siddarth, 2002).
Startle blink reflex responses. After application of a 2 ms moving average, EMG onsets
were computed as departures that were two SDs above the mean pre-startle EMG (computed for
200 ms before the startle stimulus) and did not drop below that level for more than 10 ms within
a 20-80 ms response onset window following startle stimulus onset (e.g., Ornitz et al., 1986).
Amplitudes of EMG were defined as the difference between the mean pre-startle EMG and the
peak of the response, expressed in microvolts (V). Data were log transformed to normalise the
distribution (Ornitz et al., 1986).
Trials were rejected if (1) the tonic EMG exceeded 5 V or the vertical EOG revealed
spontaneous blinks or saccades during the 200 ms pre-startle period, (2) if the onset of the startle
blink response seen on either the EMG or vertical EOG channels was earlier than 20 ms
following startle stimulus onset, or (3) if the behavioural observations recorded during the
experimental session via closed-circuit video camera indicated gross body movement or
drowsiness prior to the startle stimulus. Behavioural observations leading to trial rejection were
augmented by additional channels of recording (electrocardiogram and horizontal EOG) that
were useful in revealing excessive muscle tension, trunk or head movement, and gaze deviations
before or during the startle trials. Drowsiness was also documented by slow rolling eye
movement in the EOG channels. Six percent of trials were rejected using these criteria. Trials
were scored as zero (9% of all trials) if no observable blink activity was evident in the EMG
channel during the 20-80 ms response onset window and there was no reason to reject the trial.
Participants with more than one-third of trials rejected or scored as zero (N = 2 high anxious
children; N = 1 non-anxious control children) were excluded from analyses. Thus, analyses of
startle reflex magnitude data were based on 12 high anxious children and 10 non-anxious
Startle modulation in childhood anxiety disorders 15
SCR. The magnitude of SCRs elicited by the startle stimulus was defined as the
difference between the trough and apex of the curve, expressed in microsiemens (S), and
commencing within 1-4 s following startle stimulus onset (Christie & Venables, 1980; Neumann
& Waters, 2006). Trials were rejected if behavioural observations recorded during the
experiment and activity in other channels indicated excessive drowsiness, movement, or
behaviour such as deep sighing, coughing and sneezing. Trials were scored as zero if there was
no observable SCR activity within the 1-4 s onset window. For consistency across measures,
analyses of SCR data were based on data from the same 12 anxious children and 10 control
children who had available startle magnitude data. SCR magnitude data were square root
transformed to normalise the distribution (Christie & Venables, 1980; Neumann & Waters,
Labelling of face stimuli. The sum of correct labelling of angry and neutral faces was
computed for each participant and expressed as a percentage of the total number of labels for
each type of face. Thus, higher percentage values reflected greater accuracy in labelling the
faces as either angry or neutral. For consistency across measures, analyses of the labelling data
were based on responses from the same 12 anxious children and 10 control children who had
available startle reflex and SCR data.
Analyses. Startle reflex and SCR magnitude data acquired during the startle baseline
phase were divided into four blocks averaged over four trials per block and were analysed with
separate 4 [Block; first, second, third, fourth] x 2 [Group: high anxious; low anxious control]
repeated-measures analyses of variance (ANOVA). Independent samples t-tests were performed
on startle reflex and SCR magnitude data during the first block of baseline trials to test for group
differences in initial reactivity. Startle reflex and SCR magnitude data acquired during the affect
modulation phase were averaged separately across the six trials for angry and neutral faces as
preliminary analyses revealed that two consecutive trials of the same emotional face over three
blocks did not produce a block effect. Similarly, as preliminary analyses indicated that whether
Startle modulation in childhood anxiety disorders 16
the face was male or female had no effect on results, face gender was not included as a factor in
the final analyses. Thus, data were analysed with separate 2 [Face Emotion: angry; neutral] x 2
[Group: high anxious; low anxious control] repeated measures ANOVAs. Percent correct
labelling of angry and neutral faces were analysed using a 2 [Face Emotion: angry; neutral] x 2
[Group: high anxious; low anxious control] repeated measures ANOVA. Correlation analyses,
corrected for multiple comparisons with Bonferroni adjustments, were also performed between
startle reflex and SCR data and children’s percent correct labelling of the face stimuli and
SCAS-P total scores.
Startle baseline phase
Startle reflexes. As reported in previous studies of children (e.g., Ornitz et al., 1996),
there was wide variation in startle reflex amplitude, with mean responses varying from 5.3 to
359.1 V. Log transformations reduced this variability substantially. Figure 1, left panel,
displays the mean startle reflex magnitudes expressed in log transformed units across four
blocks of four trials. The repeated measures ANOVA revealed a significant Block main effect,
F(3, 18) = 3.67, p = .03, ηp2= .38, reflecting that startle reflex magnitude decreased in all
children across consecutive blocks of trials. However, there were no main effects or interactions
involving the Group factor, all F’s < 1.92, ns. The independent samples t-test comparing groups
over the first block of trials was not significant, t = .30, ns.
Insert Figure 1 about here
SCR. Figure 1, right panel, displays the mean square root transformed SCR magnitude
data across four blocks of four trials. The repeated measures ANOVA revealed a significant
Block main effect, F(3, 18) = 11.90, p < .001, ηp2= .67. As with EMG magnitude, there were no
Startle modulation in childhood anxiety disorders17
significant main effects or interactions with Group, all F’s < 1.09, ns. The independent samples
t-test comparing groups over the first block of trials was not significant, t = .35, ns.
Affective modulation phase
Startle reflexes. Figure 1, left panel displays the mean log transformed startle reflex
magnitude values during angry and neutral faces for each group. As can be seen, high anxious
children showed larger startle reflexes overall compared with non-anxious control children,
significant Group main effect, F(1, 20) = 4.39, p = .04, ηp2= .18. There were no significant main
effects or interactions involving the Face Emotion factor, all F’s < 2.20, ns.
SCR. Figure 1, right panel also shows the mean square root transformed SCR magnitude
values during angry and neutral faces for both groups. Similar to EMG magnitude data, the
analysis revealed a significant Group main effect, reflecting larger SCRs in high anxious
compared with non-anxious control children, F(1, 20) = 4.71, p = .04, ηp2= .19, whereas no
other effects were significant, all F’s < 1.81, ns.
Response habituation across the entire experimental procedure
Additional analyses were performed to examine group differences in startle reflex and
SCR habituation across the entire experimental procedure. As is evident in Figure 1, analyses
for each group separately revealed that significant Block main effects for startle reflexes and
SCR were associated with significant linear components in non-anxious control children (EMG:
F(1, 9) = 10.84, p = .009, ηp2= .45; SCR: F(1, 9) = 25.80, p < .001, ηp2= .74) but significant
linear and quadratic components in high anxious children (EMG: F(1, 11) = 2.90, p = .05, ηp2=
.30; F(1, 11) = 3.21, p = .04, ηp2= .31; SCR: F(1, 11) = 18.64, p < .001, ηp2= .63; F(1, 11) =
11.46, p = .006, ηp2= .51). These results suggest that whereas non-anxious control children
habituated across the entire experiment, habituation was interrupted by the presentation of
emotional faces in high anxious children
Labelling of emotional faces
Startle modulation in childhood anxiety disorders18
Table 1 displays the mean percent correct labelling of angry and neutral faces for both
groups. The analysis revealed a significant main effect of Face Emotion, F(1, 20) = 16.58, p <
.001, ηp2= .45, reflecting significantly more accurate labelling of angry faces (97%) compared
with neutral faces (64.4%) by all children combined. No other effects were significant, all F’s <
1, ns. Of the incorrect labelling of neutral faces (see Table 1), 13.6% were labelled as sad,
14.4% as happy, and 3.8% as angry. Independent samples t-tests revealed no significant group
differences in the emotion children incorrectly reported for neutral faces (all p > .29), and there
were no significant differences in the percentages of different emotions chosen within each
group, all F’s < 1.92, ns.
Correlations between startle reflex and SCR magnitude data and percent correct labelling
of angry and neutral faces for each group separately revealed a significant negative correlation
for high anxious children between startle reflex magnitude during neutral faces and neutral
labelling (r = -.60, p = .01). There were no significant correlations for non-anxious control
children (all r-values < .29, ns) or for other measures in anxious children (all r-values < .41, ns).
The significant correlation suggested that high anxious children who showed larger startle
reflexes to neutral faces were also less accurate in labelling the faces as neutral. Additional
correlations assessed the relationship between startle reflex and SCR magnitude and percentages
of incorrect labelling of neutral faces as happy, angry and sad. Whereas there were no
significant correlations for non-anxious control children (all r-values < .47, ns), larger startle
reflex magnitudes were significantly associated with increased labelling of neutral faces as sad
in high anxious children (r = .68, p = .02). There were no other significant correlations for
anxious children (all r-values < .46, ns).
Correlations between children’s anxiety severity and startle reflexes, SCR and correct
and incorrect labelling of angry and neutral faces were performed using the SCAS-P total score
given that parents of children in both groups completed this measure and because the ADIS-C
Startle modulation in childhood anxiety disorders 19
CSR ratings for high anxious children only has five values in the clinical range (i.e., 4 – 8), with
the majority of anxious children (78%) obtaining the median score (6) or above. Correlations
between SCAS-P total scores and anxious children’s labelling of neutral faces revealed a
significant negative correlation (r = -.62, p = .01), indicating that more severely anxious children
were less likely to label neutral faces as neutral. There were no other significant correlations (all
r-values < .40).
Results revealed that 4 to 8 year old high anxious and non-anxious control children
showed significant habituation of startle blink reflexes and electrodermal responses over 16
trials during the baseline phase. During the affective modulation phase, responses were
undifferentiated during angry faces compared with neutral ones in both groups of children.
However, responses were larger overall in high anxious children compared with non-anxious
controls during this phase; a group difference that was not observed during the baseline phase.
Finally, the labelling data revealed that both groups were highly accurate (97%) in judging
angry faces, whereas they were marginally better than chance at categorising neutral faces
(65%). For 35% of neutral faces, children labelled them as emotional (happy, sad, or angry).
Moreover, in high anxious but not non-anxious control children, those with greater anxiety
severity and larger startle blink reflexes were less accurate in labelling neutral faces as neutral
and were more likely to label them as sad.
Results of the present study indicate that after receiving explicit instruction about the
impending presentation of emotional faces, startle blink reflexes and skin conductance responses
were larger in high anxious than non-anxious control children. Moreover, this appeared to
reflect that whereas non-anxious control children habituated across the entire experiment, i.e.,
they showed no effect of the introduction of the emotional faces, habituation was interrupted by
the presentation of emotional faces in high anxious children (see Figure 1). These results suggest
that in an experimental procedure in which an affective modulation phase follows a startle
Startle modulation in childhood anxiety disorders 20
baseline phase, elevated anxious responding to emotionally threatening and neutral faces can
interfere with response habituation in high anxious children.
These findings present important similarities with previous studies of fear-potentiated
startle modulation in older at-risk youths and anxious adults (e.g., Grillon et al., 1994; 1997,
1998; 2005; Grillon & Ameli, 1998; Merikangas et al., 1999). In all of those studies, anxious
and at-risk participants have shown elevations of the startle reflex during contexts associated
with threat. One could argue that larger response magnitudes during angry and neutral faces in
high anxious compared with non-anxious control children in the present study reflects a similar
context effect since elevated responding was present also during neutral faces. A cue-specific
effect would suggest that startle reflexes would have been larger during angry than neutral faces.
However, prior studies have found that startle reflex magnitudes are elevated during the baseline
phase prior to, as well as throughout, fear-potentiation protocols involving threat of electric
shock, blasts of air to the larynx, or periods in complete darkness (e.g., Grillon et al., 1994;
1997; Merikangas, 1999). In the present study, there were no group differences in response
magnitudes during the baseline phase. One explanation for this difference is that instruction
about forthcoming presentations of emotional faces is less intensely aversive compared with
upcoming electric shocks, air blasts and darkness. All of these “traditional” aversive stimuli
have direct associations with potential physical harm and may evoke stronger defensive
responding earlier in the experimental session in anxious individuals. Notably, in a study of
aversive learning in anxious and non-anxious children (age range 7-14 years) (Liberman, Lipp,
Spence, & March, 2006), significant group differences in startle reflex or SCR magnitudes were
not found during a 12 trial startle baseline phase prior to a Pavlovian conditioning task involving
a loud tone unconditional stimulus; a stimulus that also is less aversive than electric shock, air
blasts and darkness (see Neumann & Waters, 2006). Another obvious difference between the
present and previous studies in which elevated baseline responding in anxious and at-risk
children was found is children’s age. Instruction about an event that will occur later in the
Startle modulation in childhood anxiety disorders 21
experimental session may be too abstract for very young children to comprehend. Further
studies with adults and older youths that compare startle modulation by emotional faces versus
the more traditional stimuli (e.g., electric shock) would help clarify these possibilities.
Startle reflex magnitude and electrodermal responses were undifferentiated during angry
and neutral faces in both high anxious and non-anxious control children, findings that differ
from a recent study with adults showing larger startle reflexes during angry faces than other
emotional expressions (e.g., Springer et al., 2007). Instead, the present results appear consistent
with mixed results on affective startle modulation in children (e.g., Cook et al., 1995; McManus
et al., 1995; Waters et al., 2005). The lack of differential modulation to angry and neutral faces
in children, but not adults, may reflect that children perceive neutral faces as more negative than
do adults. Supporting this interpretation was the analysis of the children’s labelling data, which
indicated that they interpreted neutral faces with only 65% accuracy, instead describing about
one-third of faces (35%) as emotional (predominantly happy and sad). Inspection of children’s
incorrect labelling data confirmed this was not associated with particular neutral faces over
others. That there were no significant group differences is consistent with recent evidence of
unperturbed facial expression labelling in anxious children relative to controls (Guyer et al.,
2007). Thus, whereas children were highly accurate in discriminating neutral and angry
expressions, neutral faces were emotionally ambiguous to children.
Startle reflexes can be potentiated by emotional processes, such as the affective valence
of foreground stimuli (e.g., Bradley et al., 1999) and by attentional properties, such as stimulus
interest value (e.g., Lipp, Siddle, & Dall, 1998). One possibility is that startle reflexes were
enhanced during neutral faces due to increased vigilance associated with attempts to interpret
ambiguous expressions. Thomas et al. (2001) found greater amygdala activation during neutral
than fearful faces and poor categorisation of these expressions in children compared with
reversed effects in adults. They concluded that because neutral faces were highly ambiguous to
children, they produced amygdala activation due to continued efforts to interpret them.
Startle modulation in childhood anxiety disorders 22
However, larger startle reflex magnitudes and increased anxiety severity in high anxious
children were associated with less accurate judgments of neutral faces and increased
interpretations of them as sad. Therefore, another possibility is that emotionally negative
interpretations of neutral faces may have modulated startle reflexes, particularly in high anxious
children, which would suggest an emotional modulation account (e.g., Bradley et al., 1999). As
the present study cannot elucidate the underlying processes that mediate these results, further
research with anxious and non-anxious children in this age range will be required before firm
conclusions can be drawn.
Although the present study extends previous investigations of defensive responding in
threat contexts in anxious adults and older at-risk youths to very young children, there are
notable limitations. The present study does not address the aetiology of group differences
between anxious and control children. Although studies of older at-risk offspring of anxious
parents demonstrate similar findings (e.g., Grillon et al., 1997; 1998; 2005; Merikangas et al.,
1999), studies of younger children at risk are required to examine whether sensitivity to threat
contexts and emotionally ambiguous faces reflects on neurophysiological vulnerabilities to
childhood anxiety disorders. Future investigations will also need to include other negative
emotional faces (e.g., sad faces) to examine differential reactivity to varying negative human
emotions (e.g., Springer et al., 2007). Given that neutral faces appear to be highly ambiguous to
children, future study designs should compare responses elicited during negative emotional
faces with those elicited during happy expressions (Balaban, 1995) or during intertrial intervals,
which may be a better “neutral” condition than potentially ambiguous neutral faces. It is also
possible that making judgments of faces as “neutral” may be difficult for children in part
because of demand characteristics and because the majority of faces were emotional which tend
to be more salient in general. Future studies should also include a dimensional rating of emotion
elicited by the faces (e.g., from unpleasant to pleasant). Such a rating may be more sensitive to
anxiety-related differences in young children than categorical labelling of the faces. Another
Startle modulation in childhood anxiety disorders 23
consideration for future studies is whether adult versus child angry faces differentially influence
responding in anxious relative to control children. Adult angry faces may be more likely than
child angry faces to provoke anxiety and contextual threat in anxious children. Studies with
larger sample sizes and equivalent age and sex distributions will also be important for
examining developmental effects (Balaban, 1995; Springer et al., 2007) and to extend on
previous studies showing sex differences in face processing (e.g., Thomas et al., 2001; McClure,
2000; McClure, et al., 2004) and fear-potentiated startle (e.g., Grillon et al., 1998; 2005).
Moreover, as adult studies have shown differences in the physiology of various anxiety
disorders (e.g., Cuthbert et al., 2003), future studies should include a larger sample that would
permit these comparisons in anxious children. Future research may also benefit from utilising
sound stimuli instead of, or in conjunction with, face stimuli. For example, Neumann, Waters
and Westbury (in press) have shown that an unpleasant sound of metal scraping on slate was
both tolerable and highly effective in eliciting unconditional responses in adolescents. Lau et al.
(in press) have shown that a scream sound coupled with a fearful face produced anxiety-related
differences in a fear conditioning paradigm in adolescents. Future designs should also include
more than three blocks of two trials for each emotional face and attempt to assess children’s
judgments of emotional faces “on line” rather than after the experimental procedure. Finally,
examination of the time course of processing emotional faces is an important direction for future
studies given previous evidence that anxious children and adults show elevated startle
magnitudes during unpleasant compared with neutral pictures and words very early during
stimulus processing (i.e., 60 ms lead intervals) (e.g., Aitken, Siddle, & Lipp, 1999; Waters,
Lipp, & Cobham, 2000; Waters et al., 2005).
Startle modulation in childhood anxiety disorders 24
Aitken, C. J., Siddle, D. A. T., & Lipp. O. V. (1999). The effects of threat and non-threat word
lead stimuli on blink modification. Psychophysiology, 36(6), 699-705.
American Psychiatric Association. (1994). Diagnostic and statistical manual of mental
disorders (4thed.). Washington, DC: Author.
Balaban, M. T. (1995). Affective influences on startle in five-month-old infants: Reactions to
facial expressions of emotions. Child Development, 66, 28-36.
Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1993). Pictures as prepulse: Attention and
emotion in startle modification. Psychophysiology, 30, 541-545.
Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1999). Affect and the startle reflex. In M. E.
Dawson, A. M. Schell, & A. H. Bohmelt. (Eds.), Startle modification: Implications for
neuroscience, cognitive science, and clinical science. (pp. 157-183). New York:
Cambridge University Press.
Cook, E. W., III, Hawk, L. W., Davis, T. L., & Stevenson, V. E. (1991). Affective individual
differences and startle reflex modulation. Journal of Abnormal Psychology, 100, 5-13.
Cook, E. W., III, Hawk, L. W., Hawk, T. M., & Hummer, K. (1995). Affective modulation of
startle in children [Abstract]. Psychophysiology, 32, S25.
Craske, M. G., & Waters, A. M. (2005). Panic disorder, phobias, and generalized anxiety
disorder. In S. Nolen-Hoeksema, T. Cannon, T. Widiger, T. Baker, S. Luthar, S. Mineka,
R. Munoz, & D. Salmon (Eds.). Annual Review of Clinical Psychology, 1, 197-225.
Cuthbert, B. N., Bradley, M. M., & Lang, P. J. (1996). Probing picture perception: Activation
and emotion. Psychophysiology, 33, 103-111.
Cuthbert, B. N., Lang, P. J., Strauss, C., Drobes, D., Patrick, C. J., & Bradley, M. M. (2003).
The psychophysiology of anxiety disorders: Fear memory imagery. Psychophysiology,
Daniel, A. E. (1983). Power, Privilege, and Prestige. Melbourne: Longman Cheshire.
Startle modulation in childhood anxiety disorders 25
Davis, M. (1998). Are different parts of the extended amygdala involved in fear versus anxiety?
Biological Psychiatry, 44, 1239-1247.
Gehricke, J-G., Ornitz, E. M., & Siddarth, P. (2002). Differentiating between reflex and
spontaneous blinks using simultaneous recording of the orbicularis oculi
electromyogram and the electro-oculogram in startle research. International Journal of
Psychophysiology, 44, 261-268.
Grillon, C. (2002). Startle reactivity and anxiety disorders: Aversive conditioning, context, and
neurobiology. Biological Psychiatry, 52(10), 958-975.
Grillon, C., & Ameli, R. (1998). Effects of threat and safety signals on startle during anticipation
of aversive shocks, sounds, or airblasts. Journal of Psychophysiology, 12, 329-337.
Grillon, C., Ameli, R., Goddard, A., Woods, S., & Davis, M. (1994). Baseline and fear-
potentiated startled in panic disorder patients. Biological Psychiatry, 35, 431-439.
Grillon, C., & Baas, J. M. (2003). A review of the modulation of startle by affective states and
its application to psychiatry. Clinical Neurophysiology, 114, 1557-1579.
Grillon, C., Baas, J. M., Lissek, S., Smith, K., & Milstein, J. (2004). Anxious responses to
predictable and unpredictable aversive events. Behavioral Neuroscience, 118, 916-924.
Grillon, C., Dierker, L., & Merikangas, K. R. (1997). Startle modulation in children at risk for
anxiety disorders and/or alcoholism, Journal of the American Academy of Child and
Adolescent Psychiatry, 36, 925-932.
Grillon, C., Dierker, L., & Merikangas, K. R. (1998). Fear-potentiated startle in adolescent
offspring of parents with anxiety disorders. Biological Psychiatry, 44, 990-997.
Grillon, C., & Morgan, C. A. (1999). Fear-potentiated startle conditioning to explicit and
contextual cues in Gulf war veterans with posttraumatic stress disorder. Journal of
Abnormal Psychology, 108, 134-142.
Startle modulation in childhood anxiety disorders 26
Grillon, C., Morgan C. A., Davis M., & Southwick, S. M. (1998). Effects of experimental
context and explicit threat cues on acoustic startle in Vietnam veterans with
posttraumatic stress disorder. Biological Psychiatry, 44, 1027-1036.
Grillon, C., Warner, V., Hill, J., Merikangas, K. R., Bruder, G. E., & Tenke, C. E. et al. (2005).
Families at high and low risk for depression: Three-generation startle study. Biological
Psychiatry, 57, 953-960.
Guyer, A. E., McClure, E. B., Adler, A. D., Brotman, M. A., Rich, B. A., Kimes, A. S., Pine, D.
S, Ernst, M., & Liebenluft, E. (2007). Specificity of facial expression labelling deficits in
childhood psychopathology, Journal of Child Psychology and Psychiatry, 48, 863-871.
Hadwin, J. A., Donnelly, N., French, C. C., Richards, A., Watts, A., & Daley, D. (2003). The
influence of children's self-report trait anxiety and depression on visual search for
emotional faces. Journal of Child Psychology and Psychiatry, 44(3), 432-444.
Hawk, L. W., Stevenson, V. E., & Cook, E. W., III (1992). The effects of eyelid closure on
affective imagery and eye blink startle. Journal of Psychophysiology, 6, 299-310.
Hayward, C., Killen, J. D., Kraemer, H. C., & Taylor, C. B. (2000). Predictors of panic attacks
in adolescents. Journal of American Academy of Child Adolescent Psychiatry, 39(2),
Kessler, R. C., Foster, C. L., Saunders, W. B., & Stang, P. E. (1995). Social consequences of
psychiatric disorders I: Educational attainment. American Journal of Psychiatry, 46, 1567-
Lau, J. Y. F., Lissek, S., Nelson, E., Lee, Y., Roberson-Nay, R., Poeth, K., Jenness, J., Ernst, M.,
Grillon, C., & Pine, D. S. (in press). Fear conditioning in adolescents with anxiety disorders:
Results from a novel experimental paradigm. Journal of the American Academy of Child and
Startle modulation in childhood anxiety disorders 27
Liberman, L. C., Lipp, O. V., Spence, S. H., & March, S. (2006). Evidence for retarded
extinction of aversive learning in anxious children. Behaviour Research & Therapy, 44,
Lipp, O. V., Siddle, D. A. T., & Dall, P. J. (1998). The effect of foreground stimulus modality
on blink magnitude startle modulation. Psychophysiology, 34, 340-347.
Lissek, S., Pine, D. S., & Grillon, C. (2006). The strong situation: A potential impediment to
studying the psychobiology and pharmacology of anxiety disorders. Biological
Psychology, 72, 265-270.
McClure, E. B. (2000). A meta-analytic review of sex differences in facial processing, and their
development in infants, children, and adolescents. Psychological Bulletin, 3, 424-453.
McClure, E. B., Monk, C. S., Nelson, E. E., Zarahn, E., Leibenluft, E., Bilder, R. M., Charney,
D. S., Ernst, M., & Pine, D. S. (2004). A developmental examination of gender
differences in brain engagement during evaluation of threat. Biological Psychiatry,
McGee, R., Feehan, M., Williams, S. & Anderson, J. (1992). DSM-III disorders from age 11 to
age 15 years. Journal of the American Academy of Child & Adolescent Psychiatry, 31,
McManis, M. H., Bradley, M. M., Berg, W. K., Cuthbert, B. N., & Lang, P. J. (2001).
Emotional reactions in children: Verbal, physiological, and behavioral responses to
affective pictures. Psychophysiology, (38), 222-231.
McManis, M. H., Bradley, M. M., Cuthbert, B. N., & Lang, P. J. (1995). Kids have feelings too:
Children’s physiological responses to affective pictures [Abstract]. Psychophysiology,
Merikangas, K. R., Avenevoli, S., Dierker, L. & Grillon, C. (1999). Vulnerability factors among
children at risk for anxiety disorders. Society of Biological Psychiatry, 99, 172-179.
Startle modulation in childhood anxiety disorders 28
Mogg, K., & Bradley, B. P. (1998). A cognitive-motivational analysis of anxiety. Behaviour
Research and Therapy, 36, 809-848.
Monk, C. S., Nelson, E. E., McClure, E. B., Mogg, K., Bradley, B. P., Leibenluft, E., Blair, R. J.
R., Chen, G., Charney, D. S., Ernst, M., & Pine, D. S. (2006). Ventrolateral prefrontal
cortex activation and attentional bias in response to angry faces in adolescents with
generalized anxiety disorder. American Journal of Psychiatry, 163, 1091-1097.
Murray, C. & Lopez, A. (1996). The global burden of disease: A comprehensive assessment of
mortality and disability from disease, injuries and risk factors from 1990 and projected to
2020. Cambridge, MA: Harvard University Press.
Nauta, M. H., Scholing, A., Rapee, R., Abbott, M., Spence, S. H., & Waters, A. M. (2004). A
parent-report measure of children's anxiety: Psychometric properties and comparison with
child-report in a clinic and normal sample. Behaviour Research and Therapy, 42, 813-839.
Neumann, D. L., & Waters, A. M. (2006). The use of an unpleasant sound as an unconditional
stimulus in a human Pavlovian conditioning procedure. Biological Psychology, 73, 175-185.
Neumann, D. L., Waters, A. M., & Westbury, H. R. (in press). The use of an unpleasant sound
as the unconditional stimulus in aversive Pavlovian conditioning experiments that involve
children and adolescent participants. Behaviour Research Methods.
Patton, G. C. (1998). American Journal of Public Health, 88(10), 1518-1522.
Öhman, A. & Mineka, S. (2001). Fears, phobias, and preparedness: Toward an evolved module
of fear and fear learning. Psychological Review, 108, 483-522.
Ornitz, E. M. (1999). Startle modification in children and developmental effects. In M. E.
Dawson, A. M. Schell, & A. H. Bohmelt. (Eds.), Startle modification: Implications for
neuroscience, cognitive science, and clinical science. (pp. 245-266). New York:
Cambridge University Press.
Ornitz, E. M., Guthrie, D., Kaplan, A. R., Lane, S. J., & Norman, L. (1986). Maturation of
startle modulation. Psychophysiology, 23(6), 624-634.
Startle modulation in childhood anxiety disorders 29
Ornitz, E. M., Russell, A. T., Yuan, H. & Liu, M. (1996). Autonomic, electro-
encephalographic, and myogenic activity accompanying startle and its habituation during
mid-childhood. Psychophysiology, 33, 507-513.
Pine, D., Cohen, P., Gurley, D., Brook, J., & Ma, Y. (1998). The risk for early-adulthood
anxiety and depressive disorders in adolescents with anxiety and depressive disorders.
Archives of General Psychiatry, 55, 56-64.
Pole, N., Neylan, T. C., Best, S. R., Orr, S. P. & Marmar, C. R. (2003). Fear-potentiated startle
and posttraumatic stress symptoms in urban police officers. Journal of Traumatic Stress,
Samuels, E.R., Hou, R.H., Langley, R.W., Szabadi, E., & Bradshaw, C.M. (2007). Modulation
of the acoustic startle response by the level of arousal: Comparison of clonidine and
modafinil in healthy volunteers. Neuropsychopharmacology, 1-17.
Silverman, W., & Albano, A. M. (1996). The Anxiety Disorders Interview Schedule for Children
(DSM-IV). San Antonio: Psychological Corporation.
Spence, S. H. (1998). A measure of anxiety symptoms among children. Behaviour Research and
Therapy, 36, 545-566.
Spence, S. H., Donovan, C., & Brechman-Toussaint, M. (1999). Social skills, social outcomes,
and cognitive features of childhood social phobia. Journal of Abnormal Psychology, 108,
Springer, U. S., Rosas, A., McGetrick, J., & Bowers, D. (2007). Differences in startle reactivity
during the perception of angry and fearful faces. Emotion, 7(3), 516-525
Stirling, L. J., Eley, T. C., & Clark, D. M. (2006). Preliminary evidence for an association
between social anxiety symptoms and avoidance of negative faces in school-age
children. Journal of Child and Adolescent Clinical Psychology, 35(3), 440-445.
Startle modulation in childhood anxiety disorders 30
Thomas, K. M., Drevets, W. C., Whalen, P. J., Eccard, C. H., Dahl, R. E., Ryan, N. D., et al.
(2001). Amygdala response to facial expressions in children and adults, Biological
Psychiatry, 49, 309-316.
Tottenham, N., Borscheild, A., Ellertsten, K., Marcus, D.J., Nelson, C.A. (2002). Catagorisation
of facial expressions in children and adults: establishing a larger stimulus set.
Unpublished work presented at Cognitive Neuroscience Annual Meeting, San Francisco.
Turner, S. M., Beidel, D. C., & Roberson-Nay, R. (2005). Offspring of anxious parents: Reactivity,
habituation, and anxiety-proneness. Behaviour Research & Therapy, 43, 1263-1279.
Venables, P. H., & Christie, M. J. (1980). Electrodermal activity. In I. Martin & P. H. Venables
(Eds.), Techniques in Psychophysiology, pp. 3-67. Chichester, UK: Wiley.
Waghorn, G., Chant, D., White, P., & Whiteford, H. (2004). Disability, employment and work
performance among people with ICD-10 anxiety disorders. Australian and New Zealand
Journal of Psychiatry, 39, 55-66.
Walker, D. L., Toufexis, D. J., & Davis, M. (2003). Role of the bed nucleus of the stria
terminalis versus the amygdala in fear, stress, and anxiety. European Journal of
Pharmacology, 463, 199 –216.
Waters, A. M., & Lipp, O. V. (in press). Visual search for emotional faces in children. Cognition
Waters, A. M., Lipp, O. V., & Cobham, V. E. (2000). Investigation of threat-related attentional
bias in anxious children using the startle eyeblink modification paradigm. Journal of
Psychophysiology, 14, 142-150.
Waters, A. M., Lipp, O. V., & Spence, S. H. (2005). The effects of affective picture stimuli on
blink modulation in adults and children. Biological Psychology, 68, 257-281.
Waters, A. M., Mogg, K., Bradley, B. P., & Pine, D. S. (in press). Attentional bias for emotional
faces in children with generalised anxiety disorder. Journal of the American Academy of
Child and Adolescent Psychiatry.
Startle modulation in childhood anxiety disorders 31
Although the psychometric properties of the SCAS-P have been derived from large
samples of anxious and non-anxious children between 6 and 18 years of age (e.g., Nauta et
al., 2004; Spence, 1998), the SCAS-P was completed by all parents in the present study to
avoid having incomparable anxiety measures for children below and above 6 years of age.
Parents of children younger than 6 years of age had no difficulty completing the SCAS-P in
relation to their child and their scores were similar to those relating to older children in their
Startle modulation in childhood anxiety disorders 32
Percent correct ratings of angry and neutral faces and percent of neutral faces judged
incorrectly as happy, sad or angry as a function of group.
(n = 10)
100 (00.00)66.66 (38.49) 14.66 (21.08) 15.00 (16.57)3.33 (2.22)
(n = 12)
94.44 (14.97) 62.5 (37.01) 20.83 (36.32)12.5 (20.26) 4.16 (2.99)
Startle modulation in childhood anxiety disorders 33
Figure 1 Mean EMG (left panel) and SCR (right panel) magnitudes in response to auditory
startle-eliciting sounds for high anxious (n = 12) and non-anxious control
children (n = 10) during the four blocks of trials in the baseline phase and during
angry and neutral faces in the affective modulation phase.
Startle modulation in childhood anxiety disorders34 Download full-text
1st 2nd 3rd 4th Angry Neutral
Mean SCR magnitude (uS[sqrt])
1st2nd 3rd 4th AngryNeutral
Mean EMG magnitude (uV[ln])