International Journal of Psychophysiology 79 (2011) 364–370

Contents lists available at ScienceDirect

International Journal of Psychophysiology

j o u r n a l h o m e p a g e : w w w. e l s ev i e r. c o m / l o c a t e / i j p s yc h o

Anger and cardiovascular startle reactivity in normotensive young males

Steffen Richter a,⁎, Hans-Christian Deter b, Miriam Rudat b, Hartmut Schächinger a,
Frank Zimmermann-Viehoff b, Cora Weber b
a
University of Trier, Institute of Psychobiology, Department of Clinical Physiology, Trier, Germany
b
Department of Psychosomatic Medicine and Psychotherapy, Charité Universitätsmedizin, Campus Benjamin Franklin, Berlin, Germany

a r t i c l e i n f o a b s t r a c t

Article history: Anger has been implicated in the etiology of hypertensive disease. Trait anger has been linked to enhanced
Received 18 June 2010 cardiovascular responsiveness. However, whether this association reflects differences in context appraisal or a
Received in revised form 13 December 2010 general hyper-reactivity of the cardiovascular system remains unclear. We studied the cardiovascular
Accepted 15 December 2010
response to acoustic startle probes in 76 healthy Caucasian males in different affective contices (pleasant,
Available online 22 December 2010
neutral, and unpleasant). All participants completed the State-Trait-Anger-Expression-Inventory (STAXI) by
Keywords:
Spielberger and the results were analysed with stepwise regression analysis according to the anger scores and
Anger traditional risk factors for hypertension. Our study reveals differential modulation of the cardiovascular
Startle reflex response to startle stimuli by affective pictures in the dimensions “valence” for heart rate and “arousal” for
Cardiovascular reactivity blood pressure. Anger-in was identified as the most important determinant for blood pressure responses in
Affective modulation unpleasant context, while anger-out was associated with less cardiovascular activation in neutral context.
This is the first study that relates trait anger to cardiovascular reactivity and affective reflex modulation in
normotensive subjects. We could demonstrate an interaction of affective context and trait anger for
cardiovascular (hyper-)reactivity. Increased cardiovascular reactivity for higher scores of anger-in in
unpleasant context may indicate enhanced sympathetic reactivity and constitute a risk factor for the
development of essential hypertension.
© 2010 Elsevier B.V. All rights reserved.

1. Introduction the arteries (Schlaich et al., 2004). Furthermore, enhanced sympa-

The prevalence of hypertension has been rising continuously in
modern society. The contribution of psychosocial factors such as
suppressed anger has been suggested early in hypertension research
(Alexander, 1939). Anger has repeatedly been linked with the
etiology of essential hypertension (Deter et al., 2001; Dimsdale
et al., 1986). A recent review found a consistent association of trait
anger and ambulatory blood pressure (Schum et al., 2003). It has been
hypothesized that trait anger causes increased physiological arousal
by interacting with situational factors (Jorgensen et al., 1996). In line
with this, trait anger has been linked to enhanced cardiovascular
reactivity (Bongard et al., 1998; Vogele et al., 1997) and increased
sympathetic activity (McCraty et al., 1995). Indeed, a functional blood
pressure rise in stressful situations can precede a permanently
elevated blood pressure that is solidified by structural changes of
Abbreviations: ECG, electrocardiogram; EMG, electromyogram; BMI, body mass
index; SU, standardized units; mmHg, millimeter mercury; bpm, beats per minute;
IAPS, International Affective Picture System; STAXI, State-Trait-Anger-Expression-
Inventory; SAM, Self-Assessment Manikin questionnaire.
⁎ Corresponding author. University of Trier, Department of Psychobiology, Johanniterufer
15, 54290 Trier, Germany. Tel.: +49 6512013661; fax: +49 6512013737.
E-mail address: richters@uni-trier.de (S. Richter).
thetic activity can promote later hypertension without preceding
blood pressure elevations. Enhanced sympathetic activity unfolds
deleterious effects on the cardiovascular system (Mancia et al., 1999),
promoting media sclerosis and heightened vascular resistance
(Ferriss, 1978; Markovitz et al., 1993).
However, little is known about the affective modulation of blood
pressure responsiveness and the relationship between anger and
cardiovascular reactivity in a healthy population. Especially “anger-
in”, i.e. anger directed towards the own person, held or suppressed
anger feelings, and “anger out”, i.e. overtly and directly expressed
anger towards other persons or environmental objects (Forgays et al.,
1997) appear to be positively correlated with resting systolic blood
pressure (Bongard and al'Absi, 2005; Igna et al., 2009; Vogele and
Steptoe, 1992). Negative correlations with resting diastolic blood
pressure have been reported (Schum et al., 2003; Spicer and
Chamberlain, 1996), but contradictory findings exist (Holroyd and
Gorkin, 1983; Igna et al., 2009; Shapiro et al., 1997). “Anger-in” is also
positively correlated with systolic blood pressure reactivity during
stressful tasks (Burns, 1995; Holroyd and Gorkin, 1983; Mills and
Dimsdale, 1993; Vogele and Steptoe, 1992), while mixed results exist
for “anger-out”. Although consistently associated with higher systolic
blood pressure at rest, both positive (Burns, 1995; Faber and Burns,
1996; Vogele and Steptoe, 1992) and negative (Haeri et al., 1996;

0167-8760/$ – see front matter © 2010 Elsevier B.V. All rights reserved.
doi:10.1016/j.ijpsycho.2010.12.004
 S. Richter et al. / International Journal of Psychophysiology 79 (2011) 364–370
Holroyd and Gorkin, 1983; Vogele and Steptoe, 1992) correlations
with systolic blood pressure reactivity to stress have been reported.
Assessing cardiovascular stress reactivity in subjects differing in
trait anger may potentially be biased by differences in task perception.
The startle reflex represents a valuable tool for the investigation of
cardiovascular responses to external stimulation and the modulation
of such responses by affective context. In contrast to other stimuli that
are commonly used to elicit a cardiovascular reaction (e.g. math task,
cold pressor, interview stress, etc.), the startle reflex is brainstem
relayed and therefore does not require higher cognitive processing of
the eliciting stimulus itself.
The startle reflex is part of the defensive reflex system and
accompanied by a cardiovascular response consisting of a rise in heart
rate and blood pressure (Girard et al., 2001; Holand et al., 1999a;
Holand et al., 1999b). The amplitude of the motor response (as
measured by electromyogram of the musculus orbicularis oculi) is
enhanced in negative affective context (match condition: when the
context matches the aversive character of the reflex) and diminished
in positive affective context (mismatch condition). The same pattern
of modulation has been found for heart rate responses to startle
stimuli (Bradley et al., 1990; Cook et al., 1992; Gautier and Cook,
1997). Although some studies reported inconsistent results for
reactions in pleasant context (Bradley and Lang, 2000; Sanchez
et al., 2002), these studies demonstrate an affective modulation of the
heart rate reaction that is elicited by the brainstem mediated startle
reflex. The affective modulation of the blood pressure response to
startle stimuli has, to our knowledge, not been investigated yet, but
there is evidence that enhanced affective startle modulation is
associated with generally enhanced blood pressure reactivity. Gautier
and Cook found enhanced affective startle modulation in individuals
with enhanced blood pressure response to mental stress (Gautier and
Cook, 1997). Also, enhanced affective startle modulation has been
linked to physiological characteristics such as salt sensitivity and
psychological states or traits such as anxiety, irritability, and anger
(Buchholz et al., 2001) that in turn are considered risk factors for
hypertension.
In this study, we investigated the affective modulation of heart rate
and blood pressure responses to startle stimuli in normotensive
subjects. We furthermore investigated how anger affects the
cardiovascular response to startle stimuli. Trait "anger expression"
(with the dimensions “anger-out" and “anger-in”) as well as trait
“anger control” were assessed with the German version of the State-
Trait-Anger-Expression-Inventory (STAXI). Pictures of different emo-
tional content (valence) and acoustic startle stimuli were presented to
the participants while startle magnitude, heart rate, and blood
pressure were recorded. Self-ratings of the visual stimuli were
obtained to control for differences in perceived affective stimulation
and correlated with anger scores.
We hypothesized that the startle reflex would be accompanied by
a significant cardiovascular response (i.e. an increase of heart rate and
blood pressure) modulated by affective context (i.e. an increase in
unpleasant context, and a decrease in pleasant context). We
furthermore assumed that trait anger would have a significant effect
on the cardiovascular response to startle stimuli. We hypothesized
that this effect would be independent of affective context in case of a
general hyper-reactivity of the cardiovascular system, or depending
on affective context in case of an underlying difference in context
appraisal.
2. Methods
2.1. Participants
We recruited healthy male volunteers for the study via advertise-
ment on university black boards. The relationship between trait anger
and blood pressure may be affected by gender (Durel et al., 1989;
365
Ewart and Kolodner, 1994; Shapiro et al., 1995), hereditary factors
(Lawler et al., 1998; Shapiro et al., 1995; Vogele and Steptoe, 1992),
and ethnic background (Durel et al., 1989; Ewart and Kolodner, 1994;
Shapiro et al., 1995). To control for these factors, only healthy
Caucasian males were included into the study sample and family
history of hypertension was assessed during the medical screening.
We included 80 volunteers. Four of these had to be excluded from
statistical analysis. Two participants demonstrated repeatedly elevat-
ed blood pressure in the experimental setting (N160 mmHg) and two
others had to be excluded due to technical difficulties during the
recordings. Applicants with known cardiovascular disease, other
acute or chronic illnesses, smokers, and shift workers were excluded.
All participants underwent routine physical examination. Participants
had a mean age of 25.5 ± 2.6 years and a BMI of 22.8 ± 2.3 kg/m².
Mean systolic blood pressure on physical examination prior to the test
was 110 ± 7.8 mmHg, mean diastolic blood pressure 55.8 ± 5 mmHg,
and mean heart rate 57.7 ± 8.6 beats per minute. Mean values for
physical activity (subjective rating scale, see below) were 1.9 ± .78
arbitrary units. The study was approved by the local ethics committee
of the hospital (University Hospital Benjamin Franklin, Berlin,
Germany). All participants gave written consent before being
included and were free to leave the study at any point.
2.2. Stimulus material
The affective pictures were selected from the International
Affective Picture System (IAPS) (Lang et al., 1988). The IAPS is a
well established method for affective stimulation. The system
contains a large number of slides that have been rated for valence
(i.e. the emotion elicited by the slide ranging from “unpleasant = 0” to
“pleasant = 9”) and arousal (i.e. how much arousal is elicited by the
slide ranging from “low = 0” to “high = 9”). Pleasant (e.g. nature
scenes, erotic pictures), neutral (e.g. household items, colored
textures), and unpleasant pictures (e.g. fierce animals, mutilation
pictures) were presented.1 Slides were selected to have equal ratings
for valence within each category (pleasant: 7.6 (sd = 0.54); unpleas-
ant: 2.8 (sd = 0.65); neutral: 5.09 (sd = 0.28)) based on the original
IAPS ratings by Lang et al. (1988). Positive and negative slides were
selected to have comparable ratings for arousal (5.84 (sd = 1.05) and
5.93 (sd = 0.92), respectively). Neutral slides are generally rated as
less arousing (3.22 (sd = 1.0)). The startle stimuli consisted of 95 dB
white noise of 50 milliseconds duration and instantaneous rise time
(noise generator provided by Dr I. Curio, Medizinelektronik, Bonn,
Germany).
2.3. Questionnaires
All participants completed a questionnaire assessing traditional
risk factors for hypertensive disease, i.e. family history of hyperten-
sion, age, BMI, and physical activity. Participants were categorized as
positive for “family history of hypertension” if at least one first grade
relative had a history of clinically relevant hypertensive disease.
Physical activity was assessed on a subjective rating scale ranging
from 0 (no physical activity) to 3 (high level of physical activity, e.g.
regular endurance training).
1
Complete list of IAPS-pictures used in the experiment: Original ratings for valence and
arousal are presented in brackets (valence/arousal). Pictures in brackets were not included
into statistical analysis, because they were followed or preceded by another startle in too
close temporal relationship. Pleasant pictures: 1710 (8.02/5.53), 8200 (7.15/6.33), 7580
(7.4/5.08), 4250 (8.39/7.02), 1750 (7.89/4.21), 7230 (7.42/5.81), 2160 (6.87/5.31), 4180
(8.21/7.43), 1600 (6.95/4.08). Unpleasant pictures: 6360 (2.63/5.8), 6230 (2.73/7.1), 2800
(2.31/4.94), 6540 (2.53/6.51), 1120 (4.73/6.6), 9140 (2.56/4.9), 1300 (4.06/6.9), 2120
(3.65/4.93), 3030 (2.31/6.39). Neutral pictures: 7550 (5.39/4.48), 7010 (4.95/1.55), 7130
(4.79/3.54), 9070 (4.88/3.51), 7830 (5.31/4.40), 7090 (4.95/2.30), 7100 (5.29/3.08), 2190
(4.73/2.27), 7190 (5.5/3.89)
366 S. Richter et al. / International Journal of Psychophysiology 79 (2011) 364–370
Furthermore, all participants completed the German version of the
State-Trait-Anger-Expression-Inventory (STAXI (Spielberger, 1988);
German translation by Schwenkmezger and Hodapp (1991). The
STAXI measures trait anger expression including the dimensions
“anger-out”, that is anger directed towards the outside world (objects
or persons) and “anger-in”, that is anger directed towards the own
person or suppressed anger, as well as the trait “anger control”, a
measure for a successful and healthy control of anger. Subjective
ratings of arousal and valence induced by each picture were assessed
by the Self-Assessment Manikin questionnaire (SAM) (Bradley and
Lang, 1994) immediately after the experiment.
2.4. Physiological data
During the startle trial heart rate was continuously recorded by
ECG with standard Ag/AgCl electrodes in lead II configuration. Blood
pressure was continuously recorded by a Finapres device (Ohmeda,
USA). Electromyographic recording of muscular activity from the
musculus orbicularis oculi was obtained as index of reflex strength. All
raw signals were recorded at 1000 Hz by DOS software (Turbolab,
Stemmer PC-Systeme, Puchheim, Germany). Signals were amplified
(TG424, Schwarzer, Munich, Germany) and in addition to the
electronically recorded data recorded on paper by a polygraph
(UniEDAipt UD210, Schwarzer, Munich, Germany). The electromyo-
graphic data were amplified and rectified using a root-mean square
rectifier and integrator (time constant, 10 ms).
2.5. Study protocol
Participants presented at the laboratory at a separate day to
familiarize with the location, for medical screening and resting blood
pressure measurement. To obtain reliable blood pressure measure-
ments, blood pressure was measured in a supine position every minute
over 1 h with a Dinamap device (Dinamap, Criticon, Tampa, USA). The
mean of these measurements was taken as resting blood pressure.
On the experimental day, participants were welcomed and seated in
a quiet room with dim light. Electrodes for the ECG and EMG as well as
the blood pressure monitor were attached. A series of 42 pictures from
the International Affective Picture System (IAPS) (Lang et al., 1988)
was presented on a 19-inch standard PC screen at a distance of
approximately 30 in. The only instruction given to the participants was
to view the pictures during the whole presentation period. The picture
sequence was divided into three blocks of 14 pictures each. In each block
one picture of each valence was missing, to enhance the unpredictability
of the valence category for the last pictures. Thus, 5 pictures of two
categories and 4 pictures of the remaining category were presented
during each block in semi-randomized counterbalanced order. The
blocks were presented in randomized order. Each picture was presented
for 6 s, followed by 14 s of dark screen. Startle probes were delivered
binaurally over monophonic headphones. Startle probes were pre-
sented during three pictures of each valence category in each block
(65% of the pictures) and three times in the dark period (21.4% of all
dark periods). The probe onset varied from 2 to 5 s after picture onset
(or offset for startle probes in the dark period).
2.6. Data reduction and analysis
Questionnaire data were manually transferred into SPSS for
further analysis. Heart rate was derived from ECG data by calculating
the R to R distances by Turbo Pascal based software. The heart rate and
blood pressure responses were calculated as change from the values
immediately before the respective startle stimulus. Blink magnitude
was calculated as the maximum of ms 20 to 150 following the startle
stimulus, expressed as change from the median of 50 ms before the
upstroke of the blink response. EMG data were excluded if they
fluctuated by more than 30% of the amplitude within 50 ms before
response onset. EMG data were z-transformed for further analysis.
Because z-transformation matches all participants in response
magnitude and may therefore preclude the detection of differences
due to trait anger, an additional correlation analysis of absolute (non-
transformed) EMG data with trait anger scores was performed. ECG
and blood pressure were offline controlled for artifacts by visual
artifact control. The beat-to-beat values were automatically adjusted
to values per second by Turbo Pascal based software, and a period of
10 s after the startle stimulus was included in the analysis. All further
analysis was based on second-by-second data.
Statistical analysis was performed by SPSS+ version 15 (SPSS Inc.,
Chicago, Illinois, USA). Statistical analysis of startle EMG responses and
valence and arousal ratings of visual stimuli were done by an ANOVA of
repeated measures design with the within subjects factor “affective
context” (“neutral”, “pleasant”, and “unpleasant”). Subjective valence
and arousal ratings of visual stimuli were then correlated with anger
scores by bivariate correlation analysis. Statistical significance of the
cardiovascular response and its affective modulation were tested using a
3 × 10 ANOVA of repeated measures design with the within-subjects
factors “affective context” and “response” (cardiovascular response over
10 s after the startle stimulus). Analysis of contrasts (Bonferroni
adjusted pairwise comparisons) was calculated for all ANOVAs. The
impact of anger scores (anger-in, anger-out, and anger-control) and
classical risk factors for hypertension (age, BMI, physical activity and
family history of hypertension) on the maximal cardiovascular response
after startle stimuli in different affective context was tested with a
stepwise regression analysis (linear regression model). All predictor
variables were processed in a separate model for each valence category
and entered (or excluded) according to their statistical contribution in
explaining the variance in the cardiovascular response. Collinearity
diagnostics were calculated for all regression models. Results were
Greenhouse-Geisser corrected, whenever indicated by Mauchly's test of
sphericity. Values of p b 0.05 were considered statistically significant.
3. Results
3.1. Questionnaires
Subjective ratings of the visual stimuli were significantly different for
valence (F(2,150) = 815.18, p b .001, η² = .916) and arousal (F(2,150) =
247.24, p b .001, η² = .767). Analysis of contrast demonstrated that
valence ratings for pleasant (7.05 ± .71 AU) and unpleasant (2.54 ±
.76 AU) pictures were significantly different from ratings for neutral
(5.43 ± .63 AU) pictures (pb .001 and p b .001) and from one another
(pb .001). Unpleasant (5.23± 1.17 AU) pictures were rated as most
arousing, with significant differences to pleasant (3.99 ± 1.22 AU) and
neutral (2.42 ± 1.05 AU) pictures (p b .001 and p b .001). Pleasant
pictures were rated as significantly more arousing than neutral pictures
(pb .001).
Anger-control was positively correlated with anger-in (r = .30,
p = .008) and negatively correlated with anger-out (r = −.47,
p b .001). Anger scores were not significantly correlated with classical
risk factors for hypertension, i.e. age, BMI, physical activity or family
history for hypertension. Anger-out scores were positively correlated
with valence ratings of pleasant pictures (r = .23, p = .043), while anger-
in scores were positively correlated with arousal ratings of unpleasant
pictures (r= .24, p = .034), but not with valence ratings. BMI correlated
positively with positive valence ratings (r = .25, p = .029). Family
history of hypertension correlated positively with neutral (r=.32,
p = .005), but negatively with positive ratings (r =−.29, p = .010).
3.2. Startle EMG responses
Startle eye blink magnitude was enhanced in unpleasant context
and diminished in pleasant context (F(2, 150) = 13.55, p b .001,
η² = .153), although analysis of contrasts revealed significant
 S. Richter et al. / International Journal of Psychophysiology 79 (2011) 364–370 367

differences for pleasant versus neutral context only (p b .001). Mean

z-standardized EMG responses were − 0.04 ± .4 standardized units
(SU) in neutral, − 0.23 ± .41 SU in pleasant and 0.01 ± .49 SU in
unpleasant context. Importantly, EMG responses to startle stimuli
(both z-transformed and absolute values) were not significantly
correlated to any trait anger dimension in any affective context,
suggesting similar reflex magnitudes.

3.3. Cardiovascular responses to startle stimuli

We observed a significant increase of heart rate (F(9,675) = 6.52,

p b .001, η² = .080), systolic (F(9,675) = 22.97, p b .001, η² = .234) and
diastolic (F(9,675) = 14.15, p b .001, η² = .159) blood pressure after
the startle stimuli (main effect “response”). Heart rate responses were
increased by unpleasant context (maximum 3.1 ± 1.9 bpm) and
decreased by pleasant context (maximum 2.3 ± 2.1 bpm) as com-
pared to neutral context (maximum 2.7 ± 2.0 bpm) (F(2,150) = 7.02,
p = .001, η² = .086; main effect “affective context”). This pattern of
modulation paralleled the EMG responses. Similar to EMG responses,
analysis of contrasts revealed significant differences for pleasant
versus neutral context only (p = 0.035).
Systolic blood pressure responses were increased by unpleasant
(maximum 3.3 ± 3.4 mmHg) as well as pleasant (maximum 3.6 ±
2.9 mmHg) context, compared to neutral context (maximum 2.2 ±
2.4 mmHg) (F(2,150) = 7.06, p = .001, η² = .086; main effect “affective Fig. 1. Change of heart rate during 10 s after the startle stimulus. Lines represent change
of heart rate (mean of all participants, n = 76). Different lines represent different
context”). Analysis of contrasts revealed significant differences between
affective contices, error bars represent 95% confidence interval of mean.
the responses in unpleasant versus neutral (p= 0.037) and pleasant
versus neutral context (p= 0.001). The same pattern of modulation was
observed for diastolic blood pressure, with an increase in unpleasant a positive correlation with family history of hypertension (ß = .218,
(maximum 1.9 ± 1.6 mmHg) and pleasant context (maximum 2.0 ± t = 2.080, p = .041) and age (ß = .214, t = 2.049, p = .044) in
1.7 mmHg), compared to neutral context (maximum 1.5 ± 1.4 mmHg) unpleasant context. Although diastolic blood pressure responses
(F(2,150) = 3.75, p = .026, η² = .048; main effect “affective context”). were best predicted by anger-out in neutral and anger-in in
Analysis of contrasts revealed a significant difference of responses in unpleasant context, we also observed a negative correlation with
unpleasant versus neutral context (p= 0.037). Heart rate and blood physical activity (ß = −.220, t = −2.042, p = .045) in neutral and
pressure responses in different affective contices are summarized in with BMI (ß = −.237, t = −2.202, p = .031) in unpleasant context.
Figs. 1, 2 and 3.
3.6. Impact of affective pictures without startle stimulus presentation
3.4. Impact of trait anger on the cardiovascular response to startle
stimuli Since the presentation of affective pictures may itself impact on
heart rate and blood pressure, we analyzed the response to pictures
Stepwise regression analysis revealed that in neutral context heart
rate responsiveness was negatively correlated with anger-out (ß=
−.235, t = −2.127, p = .037), while systolic blood pressure respon-
siveness was positively correlated with anger-in (ß= .314, t = 3.006,
p = .004). For diastolic blood pressure a negative correlation with
anger-out (ß= −.312, t = −2.897, p = .005) was found in neutral
context. Trait anger did not correlate with heart rate responsiveness in
pleasant or unpleasant context. However, systolic blood pressure
responsiveness was positively correlated with anger-in in unpleasant
(ß= .306, t = 2.760, p = .007) but not in pleasant context. Also diastolic
blood pressure responses were positively correlated with anger-in
(ß= .348, t = 3.233, p = .002) in unpleasant but not in pleasant context
(Figs. 4 and 5).

3.5. Impact of other risk factors for hypertensive disease

Stepwise regression analysis furthermore revealed that heart rate

responsiveness in all affective contices was influenced by physical
activity (ß = −.231, t = −2.088, p = .040 for neutral, ß = −.307, t =
−2.779, p = .007 for pleasant and ß = −.303, t = −2.734, p = .008 for
unpleasant context), with smaller heart rate responses for higher
physical activity. Although anger-in had the highest predictive value
for systolic blood pressure reactivity in all affective contices, we also
Fig. 2. Change of systolic blood pressure during 10 s after the startle stimulus. Lines
found a negative correlation with physical activity (ß = −.253, t = represent change of systolic blood pressure (mean of all participants, n = 76). Different
−2.292, p = .025) in neutral context. Furthermore, there was a lines represent different affective contices, error bars represent 95% confidence interval
negative correlation with BMI (ß = −.264, t = −2.508, p = .014) and of mean.
368 S. Richter et al. / International Journal of Psychophysiology 79 (2011) 364–370

Fig. 4. Correlation of maximum systolic blood pressure responses to startle stimuli

during unpleasant context and anger-in (mean of all participants, n = 76).

Fig. 3. Change of diastolic blood pressure during 10 s after the startle stimulus. Lines
represent change of diastolic blood pressure (mean of all participants, n = 76). Different
lines represent different affective contices, error bars represent 95% confidence interval
of mean.
without startle stimulus presentation. A 3 × 4 ANOVA of repeated
measures design was conducted with the within-subjects factors
“affective context” and “response” (cardiovascular response over the
4 s during which startle stimuli were presented, i.e. second 2 to 5 after
picture onset). No significant impact of affective context was found for
heart rate or diastolic blood pressure. However, significant differences
were found for systolic blood pressure (F(2,150) = 5.47, p = .011,
η² = .068; main effect “affective context”). Analysis of contrasts
indicated a significant increase of systolic blood pressure in pleasant
compared to neutral or unpleasant context (p = 0.021). Regression
analyses were performed in a similar way as for the response to startle
stimuli. Anger out was negatively correlated with the systolic (ß =
−.277, t = − 2.481, p = .015) and diastolic (ß = −.234, t = − 2.074,
p = .042) blood pressure response to unpleasant pictures. Further-
more, age was negatively correlated with diastolic blood pressure
responses to neutral pictures (ß = −.230, t = − 2.034, p = .045), and
physical activity was positively correlated with systolic blood
pressure responses to pleasant pictures (ß = .250, t = 2.220, p = .029).
Collinearity diagnostics were performed on all regression models.
Only the systolic blood pressure in unpleasant context showed a
Condition Index above 30, indicating a moderate degree of collinearity
(CI: 34.98). However, variance–decomposition proportions associated
with the high Condition Index were less than 0.50 for all included
variables, thus none of the variables was excluded from the model.
dimension of the foreground stimuli) that was found for EMG and
heart rate responses (Bradley and Lang, 2000; Lang et al., 1990; Miller
et al., 2002; Vanman et al., 1996), we found the blood pressure
responses enhanced in both pleasant and unpleasant context. Both
pleasant and unpleasant pictures were rated as more arousing than
neutral pictures, as previously described (Lang et al., 1988). We
may therefore assume that the observed pattern reflects a modulation
by the arousal inherent in the foreground pictures. An influence of
arousal on the startle reflex has been described before (Sanchez-
Navarro et al., 2008), but it has never been demonstrated for the
cardiovascular response to startle stimuli.
It should be noted that the presentation of affective pictures itself
elicits a cardiovascular response and thus may have influenced the
cardiovascular response to startle stimuli. In accordance, we observed
a significant increase of systolic (but not diastolic) blood pressure in
response to the presentation of pleasant (compared to neutral)
pictures without startle stimulus presentation. This increase of
systolic blood pressure may reflect an activation of the sympathetic
nervous system that would be in line with the increased response to
the subsequent startle stimulus presentation. Increased baseline

4. Discussion

Startle stimuli elicited significant heart rate and blood pressure

responses that were modulated by affective context. Trait-anger
differentially influenced the cardiovascular reaction to startle stimuli
depending on affective context. We could replicate a significant
cardiovascular response to startle stimuli. To our knowledge, there
have only been three previous publications reporting a blood pressure
response to startle stimuli (Girard et al., 2001; Holand et al., 1999a;
Holand et al., 1999b). However, this is the first study that investigated
the influence of affective context on the blood pressure response. In Fig. 5. Correlation of maximum diastolic blood pressure responses to startle stimuli
contrast to the expected pattern of modulation (along the valence during unpleasant context and anger-in (mean of all participants, n = 76).
 S. Richter et al. / International Journal of Psychophysiology 79 (2011) 364–370
blood pressure in turn activates the baroreflex and may therefore alter
the heart rate response to startle stimuli in pleasant context. This
finding may help to explain the inconsistent results for heart rate
responses in pleasant context reported in previous studies (Bradley
and Lang, 2000; Sanchez et al., 2002). The increased systolic blood
pressure in response to pleasant pictures may also have obscured
effects of anger-in on the blood pressure response in pleasant context.
In contrast, the blood pressure response did not differ between
unpleasant and neutral pictures, although unpleasant pictures were
rated as even more arousing than pleasant pictures. A possible
explanation may be that the cardiovascular response to affective
stimuli differs significantly for different subcategories of unpleasant
valence, such as anger, fear or disgust etc. (Christie and Friedman,
2004; Lerner et al., 2007). Importantly, EMG responses did not
correlate with trait anger, suggesting similar reflex magnitudes.
Arousal ratings for unpleasant pictures correlated positively with
anger-in score of the STAXI questionnaire that was also the most
important determinant of blood pressure responsiveness in unpleas-
ant context. These results underline the importance of a psychological
trait (anger-in) for sympathetic reactivity as reflected by blood
pressure reactivity, especially in situations that induce negative
emotions. In contrast heart rate responses, governed by vagal activity
(de Marneffe et al., 1993) were not affected, suggesting that
parasympathetic control of the heart is not influenced by trait
anger. These results support an association of trait anger and
increased sympathetic activity in the etiology of hypertensive disease.
They furthermore suggest that (subjective) arousal, rather than
emotion, constitutes the link between the psychological trait and
autonomic nervous system reactivity. However, it may be that
participants with higher scores of “anger-in” suppress their negative
emotions to a higher degree, resulting in similar valence ratings in
spite of stronger emotional reactions to foreground stimulation.
Heart rate responsiveness and blood pressure responsiveness in
neutral context were negatively correlated with physical activity,
probably reflecting an increased parasympathetic tone (Sharma and
Deepak, 2004; Sztajzel et al., 2006) or increased baroreflex control of
the heart rate (Raczak et al., 2005; Ueno and Moritani, 2003) that are
believed to mediate the beneficial effects of physical activity on the
cardiovascular system (Sharma and Deepak, 2004). Indeed, a recent
study found enhanced baroreflex sensitivity that was mainly
associated with vagal activation in physically active students
(Martin-Vazquez and Reyes Del Paso, 2010).
Our results match previous studies that consistently observed an
association of anger-in with higher blood pressure reactivity (Burns,
1995; Holroyd and Gorkin, 1983; Mills and Dimsdale, 1993; Vogele
and Steptoe, 1992). We could extend these findings by demonstrating
the importance of affective context for this relationship. The observed
association of anger-out with lower cardiovascular responses to
startle stimuli (as well as blood pressure responses to unpleasant
pictures) is in line with previous reports investigating cardiovascular
reactivity during stressful tasks (Holroyd and Gorkin, 1983; Vogele
and Steptoe, 1992). However, opposite results have also been
reported (Burns, 1995; Faber and Burns, 1996; Haeri et al., 1996;
Vogele and Steptoe, 1992). This inconsistency may be due to
differences in task design (Garcia-Leon et al., 2003; Vogele and
Steptoe, 1992). Positive correlations of trait anger and blood pressure
reactivity were most consistently reported for active coping tasks
(Burns, 1995; Faber and Burns, 1996; Haeri et al., 1996). In line with
this, Vogele and Steptoe found a positive correlation of anger-out and
systolic blood pressure reactivity during a mirror drawing task, but a
negative correlation during a non-verbal arithmetic task within the
same study population.
The startle reflex is brainstem relayed and responses are elicited
before a conscious evaluation of the stimulus is possible. However, the
continuing cardiovascular response may be modulated by higher
cognitive processes after the response onset, and the general setting
369
(i.e. passive listening to loud noises as employed in the present task)
may be more related to a passive coping task. Therefore, task design
may have contributed to the observed association of anger-out with
lower (instead of higher) blood pressure responses, as well as the fact
that anger-in (a trait more related to passive coping) emerged as a
better predictor for blood pressure reactivity in the present study.
The association of family history of hypertension with blood
pressure reactivity and subjective ratings of pleasant stimulus
material suggest a more general, hereditary (or acquired) difference
in autonomic nervous system responsiveness and affective stimulus
processing. Furthermore, trait anger was associated with higher
(anger-in) and lower (anger-out) blood pressure responses in neutral
context, and anger-out was associated with a more positive
evaluation of pleasant pictures. These findings support the assump-
tion of a general bias in affective stimulus processing associated with
trait anger. The trait anger-out may be associated with a bias towards
positive stimulus perception, while anger-in may be associated with a
bias towards negative stimulus perception. As previously hypothe-
sized, such differences in perception may cause greater arousal in
subjects with high anger-in (Jorgensen et al., 1996), resulting in
enhanced cardiovascular reactivity (Bongard et al., 1998; Vogele et al.,
1997), especially in unpleasant context, promoting the development
of later hypertensive disease (Schlaich et al., 2004). Thus, our results
back Alexander who earlier postulated such a role of anger in the
etiology of hypertension (Alexander, 1939).
Some limitations need to be considered. This study was not
designed to investigate the underlying mechanisms of autonomic
activation. We could demonstrate an impact of trait anger on
cardiovascular responses to startle stimuli in different affective
context, but the mechanisms remain to be clarified in future research.
The modulation of the blood pressure responses to startle stimuli by
arousal was not expected because similar patterns of modulation have
previously not been described in the literature. To further elucidate
the role of arousal in the modulation of the blood pressure responses
to startle stimuli, a study including systematic variation of the arousal
inherent in the presented stimuli would be needed. We examined
young healthy Caucasian males. Therefore, results cannot be gener-
alized to older people, females, participants with different ethnic
background or patients with cardiovascular diseases. Even though it
was our explicit goal to investigate the effect of anger in healthy
participants, a further investigation of these potential influencing
factors might be useful.
In conclusion, we could demonstrate that blood pressure
responses, in contrast to heart rate responses, were modulated by
the arousal inherent in the foreground stimulation. These findings
suggest sympathetic activation with higher arousal (independent of
valence) and parasympathetic (co-)activation in pleasant, but not in
unpleasant, context. Higher scores for anger-in were associated with
enhanced subjective arousal and enhanced blood pressure respon-
siveness, especially in unpleasant context, suggesting higher sympa-
thetic reactivity as a consequence of affective context. A positive
correlation of family history of hypertension and blood pressure
responsiveness in unpleasant context may suggest a hereditary
component. Our results support the hypothesized association be-
tween the psychological trait anger-in (that is associated with the
suppressed anger construct) in unpleasant context and enhanced
cardiovascular reactivity. Long-term follow up studies have to show if
this pattern of enhanced sympathetic reactivity in unpleasant context
is a risk factor for the development of essential hypertension.
Acknowledgement
The study was part of an ongoing project on salt sensitivity and
stress reactivity of Hans-Christian Deter supported by the Deutsche
Forschungsgemeinschaft (DE224/6-1).
370 S. Richter et al. / International Journal of Psychophysiology 79 (2011) 364–370
References
Alexander, F.G., 1939. Emotional factors in essential hypertension. Psychosom Med 1,
175–179.
Bongard, S., al'Absi, M., 2005. Domain-specific anger expression and blood pressure in
an occupational setting. J Psychosom Res 58, 43–49.
Bongard, S., al'Absi, M., Lovallo, W.R., 1998. Interactive effects of trait hostility and anger
expression on cardiovascular reactivity in young men. Int J Psychophysiol 28,
181–191.
Bradley, M.M., Lang, P.J., 1994. Measuring emotion: the Self-Assessment Manikin and
the Semantic Differential. J Behav Ther Exp Psychiatry 25, 49–59.
Bradley, M.M., Lang, P.J., 2000. Affective reactions to acoustic stimuli. Psychophysiology
37, 204–215.
Bradley, M.M., Cuthbert, B.N., Lang, P.J., 1990. Startle reflex modification: emotion or
attention? Psychophysiology 27, 513–522.
Buchholz, K., Schachinger, H., Wagner, M., Schorr, U., Sharma, A.M., Deter, H.C., 2001.
Enhanced affective startle modulation in salt-sensitive subjects. Hypertension 38,
1325–1329.
Burns, J.W., 1995. Interactive effects of traits, states, and gender on cardiovascular
reactivity during different situations. J Behav Med 18, 279–303.
Christie, I.C., Friedman, B.H., 2004. Autonomic specificity of discrete emotion and
dimensions of affective space: a multivariate approach. Int J Psychophysiol 51,
143–153.
Cook III, E.W., Davis, T.L., Hawk, L.W., Spence, E.L., Gautier, C.H., 1992. Fearfulness and
startle potentiation during aversive visual stimuli. Psychophysiology 29, 633–645.
de Marneffe, M., Gregoire, J.M., Waterschoot, P., Kestemont, M.P., 1993. The sinus node
function: normal and pathological. Eur Heart J 14, 649–654.
Deter, H.C., Buchholz, K., Schorr, U., Mathiak, K., Sharma, A.M., 2001. Salt-sensitivity and
other predictors of stress-related cardiovascular reactivity in healthy young males.
Clin Exp Hypertens 23, 213–225.
Dimsdale, J.E., Pierce, C., Schoenfeld, D., Brown, A., Zusman, R., Graham, R., 1986.
Suppressed anger and blood pressure: the effects of race, sex, social class, obesity,
and age. Psychosom Med 48, 430–436.
Durel, L.A., Carver, C.S., Spitzer, S.B., Llabre, M.M., Weintraub, J.K., Saab, P.G.,
Schneiderman, N., 1989. Associations of blood pressure with self-report measures
of anger and hostility among black and white men and women. Health Psychol 8,
557–575.
Ewart, C.K., Kolodner, K.B., 1994. Negative affect, gender, and expressive style predict
elevated ambulatory blood pressure in adolescents. J Pers Soc Psychol 66, 596–605.
Faber, S.D., Burns, J.W., 1996. Anger management style, degree of expressed anger, and
gender influence cardiovascular recovery from interpersonal harassment. J Behav
Med 19, 31–53.
Ferriss, J.B., 1978. The pathogenesis of essential hypertension. Invest Cell Pathol 1,
171–197.
Forgays, D.G., Forgays, D.K., Spielberger, C.D., 1997. Factor structure of the State-Trait
Anger Expression Inventory. J Pers Assess 69, 497–507.
Garcia-Leon, A., Reyes del Paso, G.A., Robles, H., Vila, J., 2003. Relative effects of
harassment, frustration, and task characteristics on cardiovascular reactivity. Int J
Psychophysiol 47, 159–173.
Gautier, C.H., Cook III, E.W., 1997. Relationships between startle and cardiovascular
reactivity. Psychophysiology 34, 87–96.
Girard, A., Holand, S., Laude, D., Elghozi, J.L., 2001. Antihypertensive monotherapy and
cardiovascular responses to an acoustic startle stimulus. J Cardiovasc Pharmacol 37,
101–107.
Haeri, S.L., Mills, P.J., Nelesen, R.A., Berry, C.C., Ziegler, M.G., Dillon, E., Dimsdale, J.E.,
1996. Acute psychologic stress reactivity in blacks versus whites:relationship to
psychologic characteristics. Blood Press Monit 1, 27–32.
Holand, S., Girard, A., Laude, D., Meyer-Bisch, C., Elghozi, J.L., 1999a. Effects of an
auditory startle stimulus on blood pressure and heart rate in humans. J Hypertens
17, 1893–1897.
Holand, S., Girard, A., Meyer-Bisch, C., Elghozi, J.L., 1999b. Cardiovascular responses to a
acoustic startle stimulus in man. Arch Mal Coeur Vaiss 92, 1127–1131.
Holroyd, K.A., Gorkin, L., 1983. Young adults at risk for hypertension: effects of family
history and anger management in determining responses to interpersonal conflict.
J Psychosom Res 27, 131–138.
Igna, C.V., Julkunen, J., Vanhanen, H., 2009. Anger expression styles and blood pressure:
evidence for different pathways. J Hypertens 27, 1972–1979.
Jorgensen, R.S., Johnson, B.T., Kolodziej, M.E., Schreer, G.E., 1996. Elevated blood
pressure and personality: a meta-analytic review. Psychol Bull 120, 293–320.
Lang, P.J., Öhman, A., Vaitl, D., 1988. The International Affective Picture System
(photographic slides). University of Florida, Gainsville, Florida, USA.
Lang, P.J., Bradley, M.M., Cuthbert, B.N., 1990. Emotion, attention, and the startle reflex.
Psychol Rev 97, 377–395.
Lawler, K.A., Kline, K., Seabrook, E., Krishnamoorthy, J., Anderson, S.F., Wilcox, Z.C.,
Craig, F., Adlin, R., Thomas, S., 1998. Family history of hypertension: a
psychophysiological analysis. Int J Psychophysiol 28, 207–222.
Lerner, J.S., Dahl, R.E., Hariri, A.R., Taylor, S.E., 2007. Facial expressions of emotion reveal
neuroendocrine and cardiovascular stress responses. Biol Psychiatry 61, 253–260.
Mancia, G., Grassi, G., Giannattasio, C., Seravalle, G., 1999. Sympathetic activation in the
pathogenesis of hypertension and progression of organ damage. Hypertension 34,
724–728.
Markovitz, J.H., Matthews, K.A., Kannel, W.B., Cobb, J.L., D'Agostino, R.B., 1993.
Psychological predictors of hypertension in the Framingham Study. Is there
tension in hypertension? Jama 270, 2439–2443.
Martin-Vazquez, M., Reyes Del Paso, G.A., 2010. Physical training and the dynamics of
the cardiac baroreflex: a comparison when blood pressure rises and falls. Int J
Psychophysiol 76, 142–147.
McCraty, R., Atkinson, M., Tiller, W.A., Rein, G., Watkins, A.D., 1995. The effects of
emotions on short-term power spectrum analysis of heart rate variability. Am J
Cardiol 76, 1089–1093.
Miller, M.W., Patrick, C.J., Levenston, G.K., 2002. Affective imagery and the startle
response: probing mechanisms of modulation during pleasant scenes, personal
experiences, and discrete negative emotions. Psychophysiology 39, 519–529.
Mills, P.J., Dimsdale, J.E., 1993. Anger suppression: its relationship to beta-adrenergic
receptor sensitivity and stress-induced changes in blood pressure. Psychol Med 23,
673–678.
Raczak, G., Pinna, G.D., La Rovere, M.T., Maestri, R., Danilowicz-Szymanowicz, L.,
Ratkowski, W., Figura-Chmielewska, M., Szwoch, M., Ambroch-Dorniak, K., 2005.
Cardiovagal response to acute mild exercise in young healthy subjects. Circ J 69,
976–980.
Sanchez, M., Ruiz-Padial, E., Perez, N., Fernandez, P., Cobos, P., Vila, J., 2002. Modulacion
emocional de los reflejos defensivos mediante visualizacion de imagenes afectivas.
Psicotema 14, 702–707.
Sanchez-Navarro, J.P., Martinez-Selva, J.M., Torrente, G., Roman, F., 2008. Psychophysio-
logical, behavioral, and cognitive indices of the emotional response: a factor-analytic
study. Span J Psychol 11, 16–25.
Schlaich, M.P., Lambert, E., Kaye, D.M., Krozowski, Z., Campbell, D.J., Lambert, G.,
Hastings, J., Aggarwal, A., Esler, M.D., 2004. Sympathetic augmentation in
hypertension: role of nerve firing, norepinephrine reuptake, and angiotensin
neuromodulation. Hypertension 43, 169–175.
Schum, J.L., Jorgensen, R.S., Verhaeghen, P., Sauro, M., Thibodeau, R., 2003. Trait anger,
anger expression, and ambulatory blood pressure: a meta-analytic review. J Behav
Med 26, 395–415.
Schwenkmezger, P., Hodapp, V., 1991. A questionnaire for assessing anger and
expression of anger. Z Klin Psychol Psychopathol Psychother 39, 63–68.
Shapiro, D., Goldstein, I.B., Jamner, L.D., 1995. Effects of anger/hostility, defensiveness,
gender, and family history of hypertension on cardiovascular reactivity. Psycho-
physiology 32, 425–435.
Shapiro, D., Jamner, L.D., Goldstein, I.B., 1997. Daily mood states and ambulatory blood
pressure. Psychophysiology 34, 399–405.
Sharma, R.K., Deepak, K.K., 2004. A short duration of physical training benefits
cardiovascular performance. Indian J Physiol Pharmacol 48, 481–485.
Spicer, J., Chamberlain, K., 1996. Cynical hostility, anger, and resting blood pressure. J
Psychosom Res 40, 359–368.
Spielberger, C.D., 1988. State-Trait-Anger-Expression-Inventory (STAXI). Research
Edition ed, Psychological Assessment Ressources, Odessa, Florida, USA.
Sztajzel, J., Atchou, G., Adamec, R., Bayes de Luna, A., 2006. Effects of extreme endurance
running on cardiac autonomic nervous modulation in healthy trained subjects. Am
J Cardiol 97, 276–278.
Ueno, L.M., Moritani, T., 2003. Effects of long-term exercise training on cardiac
autonomic nervous activities and baroreflex sensitivity. Eur J Appl Physiol 89,
109–114.
Vanman, E.J., Boehmelt, A.H., Dawson, M.E., Schell, A.M., 1996. The varying time courses of
attentional and affective modulation of the startle eyeblink reflex. Psychophysiology
33, 691–697.
Vogele, C., Steptoe, A., 1992. Emotional coping and tonic blood pressure as
determinants of cardiovascular responses to mental stress. J Hypertens 10,
1079–1087.
Vogele, C., Jarvis, A., Cheeseman, K., 1997. Anger suppression, reactivity, and
hypertension risk: gender makes a difference. Ann Behav Med 19, 61–69.