Cardiac reactivity and preserved performance under stress:
Two sides of the same coin?
Nathalie Pattyn a,b,c,d,
⁎, Olivier Mairesse a,e
, Aisha Cortoos a,f
, José Morais c
, Eric Soetens b
, Bart Roelands d
,
Annick van den Nest a
, Régine Kolinsky c
a
VIPER Research Unit, Royal Military Academy, Renaissancelaan, 30, 1000 Brussel, Belgium
b
Dept of Biological Psychology, Vrije Universiteit Brussel, Pleinlaan, 2, 1050 Brussel, Belgium
c
Unité de Recherches en Neurosciences Cognitives, Université Libre de Bruxelles, CP 191, avenue F.D. Roosevelt, 50, 1050 Bruxelles, Belgium
d
Department of Human Physiology, Vrije Universiteit Brussel, Pleinlaan 2, 1050 Brussel, Belgium
e
CHU Brugmann, Sleep Unit, Belgium
f
UZ Brussel, Sleep Unit, Belgium
a b s t r a c ta r t i c l e i n f o
Article history:
Received 10 July 2012
Received in revised form 11 March 2013
Accepted 12 March 2013
Available online xxxx
Keywords:
Stress
Performance
Cardiac reactivity
Stroop
In the present experiment, cognitive control under stress was investigated using a real-life paradigm, namely
an evaluation flight for military student pilots. The magnitude of cognitive interference on color–word,
numerical and emotional Stroop paradigms was studied during a baseline recording and right before the
test flight. Cardio-respiratory parameters were simultaneously assessed during rest and the performance of
the Stroop tasks. Cognitive data suggested a different speed/accuracy trade-off under stress, and no modulation
of the interference effect for color words or numerical stimuli. However, we observed a major increase
in error rates for specific emotional stimuli related to the evaluation situation in the stress condition. The
increase in cognitive interference from emotional stimuli, expressed as an increase in error rates, was correlated
to the decreased cardiac reactivity to challenge in the stress situation. This relationship is discussed in
the framework of Sanders' (1983) model of stress and performance. In terms of future research, this link
warrants a fruitful lead to be followed for investigating the causal mechanism of performance decrements
under the influence of stress.
© 2013 Published by Elsevier B.V.
1. Introduction
Ever since Yerkes and Dodson (1908), the effects of stress on cognitive
performance have been investigated in both experimental and applied
psychology. However, the concepts stress and performance have
too often been stocked with idiosyncratic or unspecific connotations,
hampering comparison of the available results. Operationalization of
stress or arousal in research paradigms indeed is quite complex. While
questionnaires rely on subjective evaluation, physiological measures
like heart rate variability quantify a systemic outcome. Operational definitions
based on response have been prominent in psychophysiological
literature (Sanders, 1983), ever since Selye (1956) introduced the
concept of stress as the response of the body to any demand made
upon it. Intervening variable-definitions (Cox, 1978, in Sanders, 1983)
on the other hand, have become very influential in clinical and coping
literature, whereas this conceptualization has been surprisingly absent
in psychophysiological research on stress and performance. The two
most influential contemporary stress and performance models did nonetheless
underscore an intervening variable concept, termed ‘resource recruitment’
in Hockey's (1997) compensatory control regulation and
‘effort’ in Sanders' (1983) cognitive-energetical model. In order to
allow inferences about mechanisms at play, psychophysiological investigations
of stress and performance should frame the design and dataanalysis
within such models, which happens surprisingly seldom in the
most recent research.
With regard to performance, as mentioned by Kofman et al.
(2006), very few studies examined the effects of stress on executive
functions. Indeed, considering the importance of stress in applied research,
for example on aviation or traffic safety (e.g. Matthews et al.,
1998), and considering the involvement of higher order cognitive
functions such as planning, monitoring and cognitive control for an
adequate performance in the aforementioned settings, there is a remarkable
lack of experimental results on the effects of stress on executive
functions. According to Matthews et al. (1997), stress mainly
affects performance in applied settings through cognitive interference.
In experimental paradigms investigating executive functions,
distinct variants of the Stroop task (Stroop, 1935) are known to trigger
this type of cognitive interference.
International Journal of Psychophysiology xxx (2013) xxx–xxx
⁎ Corresponding author at: VIPER Research Unit, Royal Military Academy, Renaissancelaan,
30, 1000 Brussel, Belgium. Tel.: +32 2 629 15 94, +32 476 66 12 03 (mobile); fax: +32 2
629 24 89.
E-mail address: nathalie.pattyn@rma.ac.be (N. Pattyn).
INTPSY-10614; No of Pages 8
0167-8760/$ – see front matter © 2013 Published by Elsevier B.V.
http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
Contents lists available at SciVerse ScienceDirect
International Journal of Psychophysiology
journal homepage: www.elsevier.com/locate/ijpsycho
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
In the classical color–word Stroop task, subjects are instructed to
name the color in which a word appears. Congruent stimuli display coherence
between semantics and appearance (e.g. the word red in red),
whereas incongruent stimuli do not (e.g. the word red in blue). The numerical
Stroop variant (Pavese and Umiltà, 1998) requires participants
to respond to the amount of characters presented onscreen; these
can be either congruent (three times 3) or incongruent (four times 3).
The response conflict elicited by incongruent stimuli materializes as
longer reaction times (RTs) and lower accuracy (for a review, see
MacLeod, 1991). The emotional Stroop variant includes (among the
color words) negatively valenced emotional words, related to a particular
individual's area of concern, causing interference as a result of attentional
bias toward threat related expressions (McKenna and Sharma,
1995).
The emotional Stroop version therefore allows inferences about
attentional biases, reflected as longer response latencies to name
the ink color of emotional words as compared to neutral words. As
Williams et al. (1997) summed up, emotional Stroop interference
could be attributed to variance in state or trait emotion, to variance
in the particular situation in which the task is performed, or to the
specific nature of the negatively valenced words used. Impact of negative
valence requires the presented material to be accustomed to
both the subject's current concerns, and the situation, as in Ray's
(1979) pioneering emotional Stroop challenge.
In addition to aforementioned Stroop interferences, effects from
negative priming and inverse negative priming have been investigated
as well. Negative priming refers to a slowed response time to a target
stimulus that has been previously ignored (correct response for
stimulus S is the inhibited response for item S-1, e.g. yellow presented
in blue after blue presented in red), and involves inhibition of a mechanism
of selective attention (Tipper, 1985). Inverse negative priming further
adds the reciprocal variation of relevant and irrelevant dimension
(yellow presented in blue after blue presented in yellow). Examining
negative priming effects thus provides more insight in the quality of
cognitive control for selecting relevant information.
Moreover, the Stroop tasks are among the most widely applied
paradigms to elicit stress in laboratory conditions, for the purpose of
investigating autonomic reactivity to mental stress (e.g. Akerstedt
et al., 1983; Kamarck et al., 1994; Heims et al., 2006; Wright et al.,
2007). In this line of research, reactivity is conceptualized as “an
acute and relatively rapid change in a cardiovascular parameter as a
function of the presentation of a stressor” (Hughdahl, 1995). This
concept of reactivity could be applied to quantify the “recruitment
of resources” from Hockey's model, or the “effort” from Sander's
model. Kofman et al. (2006) emphasized that, since both the regulation
of the autonomic stress response and the inhibition of prepotent
responses activate common prefrontal cortical regions, particularly
the anterior cingulate cortex (ACC), an interaction between these
processes could indeed be expected on this neural basis.
In order to further explore the interplay between real-life stress and
cognitive interference, within a psychophysiological frame linking reactivity
and the quality of performance, student pilots were subjected to
color–word, numerical and emotional Stroop tasks, once in a baseline
recording and once right before their Progress Test General Flying
(PTGF). PTGF is the most feared examination flight in the basic flight
training, as failure to pass might force the student pilot out of the
training. To assess stress induction efficiency, prior to experimental
measurements, subjects were administered the State Trait Anxiety
Inventory (STAI) (Spielberger, 1983). Cardio-respiratory parameters
were recorded both during rest and execution of the cognitive tests.
We expected faster responses on both the color–word and the numerical
Stroop task under stress, but increased interference effects as a result
of a decline in cognitive control. Furthermore, we expected these effects
to be related to the magnitude of stress reactivity, as quantified by the
cardio-respiratory parameters, fitting within the resource recruitment
or effort notion.
2. Method
2.1. Subjects
Student pilots (N = 12) from the Belgian Air Force in their basic
flight training, aged 19 to 25 years (mean = 22.5), all medically fit
to fly and free of significant medical antecedents, with normal vision,
participated.
2.2. Procedure
The total duration of the procedure approximated 40 min. Prior
to cognitive testing, participants were equipped with the LifeShirt system
(VivoMetrics, Inc.). After a rest recording period of 5 min, participants
completed the STAI questionnaire. The cognitive battery was computer
driven and lasted for approximately 20 min. Onscreen instructions were
followed by a series of 7 cognitive tasks in the following sequence: a Stroop
color–word task with neutral words (S1-N) among the color words on a
white background, a Stroop task including emotional words among
the color words (S1-E) on a white background, a similar Stroop
task with neutral words (S2-N) on a black background and a Stroop
task with emotional words (S2-E) on a black background. Subsequently,
two recognition tasks (Rec1 and Rec2) were presented,
each including neutral and emotional words from the lists presented
in the four previous tasks, as well as new words. The last Stroop paradigm
was a numerical Stroop task (Num). Task presentation (lists and
stimuli) was counterbalanced, to control for potential order effects.
The procedure was applied in a repeated measure design: the baseline
recording took place after approximately one third of the flight training,
the stress-condition recording was planned just before the PTGF, a
major stress-inducing flight evaluation. This evaluation flight would always
take place as the first flight of the day, therefore, all recordings
started around 09.00 AM (thus avoiding circadian interference),
ended around 09.40 AM, after which the student pilot started with
the briefing for his PTGF. Recording sessions were separated by minimum
2 and maximum 5 months. The memory tasks will not be
discussed in the present paper. In the description and discussion of results,
task will refer to the cognitive tasks, and test will be used to qualify
the pre-test condition (i.e. before the evaluation flight).
2.3. Task description
In the color–word Stroop task, words were presented in the middle
of the screen, in bold Courier New font, 14 points, under a vertical
visual angle of 2° 03′. The response–stimulus interval (RSI) was
32 ms; response times and error rates were recorded. After detailed
and standardized onscreen instructions, a 60 trial practice block was
inserted, providing subjects with performance feedback. Subjects
were instructed to respond to the color in which a word appeared
as quickly and accurately as possible, using color-labeled keys on
the keyboard. The Stroop task consisted of two lists with stimuli
presented on either a white or a black background, counter-balanced
between participants. Half of the participants got to see the list on a
white background first, whereas the other half received first the list
on a black background. Furthermore, the contents of each list were
again counterbalanced, meaning each list was presented on the white
background half of the time, and on the black background the other
half. Each list contained 14 general emotional, 7 pilot specific emotional,
7 student pilot specific emotional, and 28 neutral words, as well as 30
congruent, 30 incongruent, 5 negative priming and 15 inverse negative
priming trials. Stimuli appeared in red, blue, yellow or green, in a
pseudo-randomized order, limiting consecutive appearance of same
color to 2.
For the numerical Stroop task, as described by Pavese and Umiltà
(1998), participants had to respond to the amount of stimuli present
on the screen. Stimuli were either numbers (2, 3, 4 or 5) or crosses
2 N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
(X), grouped by two, three, four or five. Each stimulus string appeared at
every possible location in the horizontal visual range. The experimental
list comprised 250 trials, with 20% congruent, 20% neutral, and 60% incongruent
trials, presented in a randomized order. The responsestimulus
interval (RSI) was 32 ms as well. Recorded variables were response
times and error rates. The experimental block was preceded by a
60 trial practice block, during which feedback was provided to
participants. Subjects responded through the numerically labeled keys
on the keyboard.
2.4. Apparatus
All experimental sessions were run at the same location and the
same time of day, starting between 08.30 and 10.30. During the testing,
subjects wore a headset to minimize possible noise disturbance.
Stimulus presentation, timing and data recording were controlled
using E-Prime software (Schneider et al., 2003) on a Sony VAIO
laptop computer, with a 15.4′ monitor, at 40 cm viewing distance.
Cardiorespiratory parameters were recorded non-invasively through
the LifeShirt system (VivoMetrics, Inc.). A standard single lead ECG
was recorded with a sampling frequency of 200 Hz and re-sampled at
1000 Hz for R-wave detection. Respiratory movements were measured
by respiratory inductive plethysmography; abdominal and ribcage
excursions were recorded at 50 Hz. All data were visually inspected
for artifacts; ectopic beats or erroneous R-wave detections were manually
corrected (removal of erroneous detection/artifact followed by a
cubic spline interpolation; corrections b 1%).
Through a derivative based algorithm R-waves were detected, and RR
intervals calculated. Other computed outcome variables were respiratory
frequency (F_resp), tidal volume (TV), and respiratory sinus arrhythmia
(RSA), which was calculated with the peak-valley method. As variation
in TV across experimental conditions was more relevant than absolute
volume values per se, the Qualitative Diagnostic Calibration (a proprietary
relative calibration of volume in arbitrary units implemented in the
software Vivologic 2.9.3) was applied to individual data-files.
2.5. Stimuli
Color words used in the classical Stroop task (yellow, green, red
and blue), were supplemented with neutral and emotional words,
matched for word length and frequency across all conditions and
lists with the use of the Dutch CELEX database (Baayen et al., 1995).
Wherever possible, stimuli were also matched for familiarity according
to the Hermans and De Houwer rating (1994). As for emotional content,
there were 4 conditions of words: neutral, general emotional, pilot
specific emotional, and student pilot specific emotional. As for congruence,
there were four conditions of color words: congruent, incongruent,
negative priming and inverse negative priming. All words are
listed in Appendix 1. General emotional words were selected from the
database of Hermans and De Houwer (1994), pilot specific and student
pilot specific emotional words were generated by three independent
pilots who did not participate in the experiment. Neutral words were
matched to these selected stimuli, on word length and chosen from
average frequency.
2.6. Cognitive data-analysis
Color Stroop and emotional Stroop effects were analyzed separately.
We aimed at investigating differences between distinct levels of interference,
and therefore performed ANOVA on the color Stroop data with conflict
as a four level factor (levels: congruent, incongruent, negative
priming and inverse negative priming). Congruence and priming effects
were explored separately so as to permit investigation of incremental
conflict levels. ANOVA on emotional Stroop data was performed with
emotion as a four level factor (levels: neutral, general emotional, pilot
specific emotional and student pilot specific emotional words). Except
when explicitly mentioned, all testing was two-sided testing. Post-hoc investigation
of significant differences was done with the ANOVA contrasts,
and all statistical testing applied the significance level of 0.05.
2.7. Physiological data-analysis
Considering the known trade-off in psychophysiology between the
need for longer recordings from a statistical point of view versus the
need to analyze a period with a signal as close as possible to stationarity,
2-minute sequences are recommended as an acceptable compromise
(Berntson et al., 1997). Hence, the subtests were programmed to approximate
this duration. These 2-minute sequences were thus extracted from
the recordings of the different subtests, starting after the practice trials
when applicable, or at the beginning of the task. A repeated measures 8
(Test) ∗ 2 (Session) MANOVA was first performed on all physiological
variables to map interdependency patterns before performing separate
ANOVA's, if necessary complemented with ANCOVA's. Post-hoc
investigation of significant differences was done with the ANOVA
contrasts, and all statistical testing applied the significance level of
0.05.
3. Results
3.1. Anxiety scores
The baseline session yielded mean scores for trait and state anxiety
of 42.64 (SEM = 1.42) and 43.92 (SEM = 1.23) respectively, whereas
pre-exam recording returned mean scores of 40.64 (SEM = 1.23)
and 50.15 (SEM = 2.79) for trait and state. Under the a priori assumption
of elevated anxiety scores in the pre-exam recording,
a one-tailed t-test confirmed the state anxiety increase to be significant
[t(11) = 2.18; p = 0.021; η2
= 0.30].
3.2. Color–word Stroop task
A 4 (condition) ∗ 2 (session) ANOVA on RTs yielded a significant
effect of condition [F(3,33) = 3.13; p = 0.039; η2
= 0.221], but no
significant effect of session [F(1,11) = 2.17; p > 0.1], nor any interaction
between session and condition [F(3,33) b 1]. As depicted in
Fig. 1, the pre-exam condition does not modulate the interference
effect. With respect to accuracy, the stress-level does not seem to
modulate the interference effect in congruent and incongruent word
conditions, as depicted in Fig. 2. A 4 ∗ 2 ANOVA on error rates after
arcsine transformation returned none of the observed variations
significant: condition [F(3,33) = 1.99; p > 0.1], session [F(1,11) b 1]
and their interaction [F(3,33) b 1] all failed to reach significance.
3.3. Emotional Stroop task
Session did not impact on the emotional content interference effects,
as visualized in Fig. 3. RT variation did not reach significance:
the 4 (emotion) ∗ 2 (session) ANOVA did not yield any significant
effect. The accuracy data, on the other hand, did show the expected
interference pattern, as depicted in Fig. 4. Pre-exam recordings
showed an increase in error rates as compared to the baseline condition,
with highest increase in errors in the ‘emo-stud’ condition,
containing words related to student pilot evaluation. A 4 ∗ 2
ANOVA after arcsine transformation returned a significant effect of
session [F(1,11) = 7.18; p = 0.021; η2
= 0.395] and of emotion
[F(3,33) = 55.61; p b 0.001; η2
= 0,835], as well as a significant interaction
between both [F(3,33) = 3.78; p = 0.02; η2
= 0.256].
Only the contrast analysis on student pilot specific emotional stimuli
[F(1,11) = 13.82; p = 0.003] showed these to distinguish significantly
between sessions.
3N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
3.4. Numerical Stroop test
As shown in Fig. 5, session did not seem to modulate the interference
effect. A 2 (condition) ∗ 2 (session) ANOVA on RTs confirmed the significant
effect of condition [F(1,11) = 60.2; p b 0.001; η2
= 0.858], but
the absence of a significant session effect [F(1,11) = 1.57; p > 0.1],
and a lack of interaction between both [F(1,11) = 1.32; p > 0.1].
As for accuracy, as shown on Fig. 6, session did not impact on error
rate, but interference in the incongruent condition seemed substantially
higher in the pre-exam test session. A 2 (condition) ∗ 2 (session)
ANOVA after arcsine transformation confirmed the significant
main effect of condition [F(1,11) = 38.62; p b 0.001; η2
= 0.794],
the lack of significant main effect of session [F b 1], and the absence
of a significant interaction between both [F b 1].
3.5. Physiological data
Mean values and respective standard deviations for the recorded
and computed physiological parameters for both baseline and preexam
recordings are reported in Table 1, according to the different
subtest rest, S1-N, S1-E, S2-N, S2-E, Rec1, Rec2 and Num. These physiological
data have been reported in details elsewhere (Pattyn et al.,
2010). To summarize, two different effects showed when comparing
baseline and pre-exam recordings. Firstly, the mean rest values differ,
most markedly for RRI (772 ms for the pre-exam vs 851 ms for
the baseline recording) and for RSA (133 ms for the pre-exam
vs 189 ms for the baseline recording), indicating the expected higher
activation during the pre-exam session. Secondly, the size of the initial
reactivity – the difference between rest values and the first test
presentation – due to cognitive test presentation decreases in the
pre-exam condition, most markedly for RRI.
A repeated measures 8 (Test) ∗ 2 (Session) MANOVA on the
variables described in Table 1 showed a significant effect of
both Test [F(7,143) = 28.73; p b 0.001; η2
= 0.584], and Session
[F(1,143) = 3.8; p = 0.003; η2
= 0.12], but no interaction
[F(7,143) b 1]. The univariate ANOVAs for Test showed a significant
effect for F_resp [F(7,143) = 17.59; p b 0.001; η2
= 0.463]
and for RSA [F(7,143) = 9.01; p b 0.001; η2
= 0.306]. The univariate
ANOVAs for Session showed a significant effect for RRI
[F(1,143) = 11.17; p = 0.001; η2
= 0.072] and for RSA [F(1,143) =
10.35; p = 0.002; η2
= 0.067]. Separate MANOVAs for the baseline
and pre-exam recordings showed similar effects of Test, respectively
Fig. 1. RTs (ms) for the different conditions of the Stroop color–word test for student
pilots on baseline and pre-test recordings.
Fig. 2. Error rates (%) for the different conditions of the Stroop color–word task for
student pilots on baseline and pre-test recordings.
Fig. 3. RTs (ms) for the different conditions of the emotional Stroop task for student
pilots on baseline and pre-test recordings.
Fig. 4. Error rates (%) for the different conditions of the emotional Stroop task for
student pilots on baseline and pre-test recordings.
4 N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
[F(7,71) = 16.97; p b 0.001; η2
= 0.626] and [F(7,71) = 14.91;
p b 0.001; η2
= 0.592], with univariate ANOVAs for both showing significant
variations for F_resp and RSA. As we previously discussed in details
(Pattyn et al., 2010), the analysis strategy of using MANOVAs first
aimed at mapping the internal dependencies of the different outcome
variables. Indeed, as we have shown, the RSA reactivity is mainly dependent
on respiration, and these two can thus not be treated as independent
variables.
Considering the fact that the data showed no effect of presentation
of emotional words, no effect of task difficulty nor of task switching,
neither during the baseline nor during the pre-exam session, the
main effect in the physiological data is thus the initial reactivity
between rest recordings and recordings during presentation of the
first task, with a tendency to return to rest values along time-on-task.
Despite the overall MANOVA showing no significant interaction
between session and test, the initial reactivity in RRI differs between
baseline and pre-exam recordings, and RRI is the parameter for which
this difference appears most clearly. The difference in reactivity
is most noticeable when the RRI values described in Table 1 are
expressed in proportional changes when compared to rest values.
In baseline recordings, the initial reactivity in RRI is expressed as a
mean decrease of 10%, whereas in pre-exam recordings, this initial
reactivity represents only a decrease of 2.3%. This interesting effect
might be blurred by interindividual difference, and would thus require
a within-subject standardization to show significance, which
is a repeatedly acknowledged issue in psychophysiological research
(e.g., Bush et al., 1993). The range-correction procedure (Lykken
and Venables, 1971, in Grossman and Kollai, 1993) was thus applied
to the RRI for initial reactivity data. The result of this rangecorrection
procedure is depicted in Fig. 7.
This effect was examined through a 2 (Session) ∗ 2 (Test) repeated
measures ANOVA, which showed a significant effect of test
[F(1,10) = 22.38; p = 0.001; η2
= 0.72], no significant effect of
session [F(1,10) = 2.29; p > 0.1] and a significant interaction between
session and test [F(1,10) = 10.62; p = 0.01; η2
= 0.54].
3.6. Linking the cognitive and physiological data
Results in the stress condition are characterized by higher error
rates, an increased resting activation (RRI & RSA) and a decreased
Fig. 5. RTs (ms) for the numerical Stroop task for student pilots on baseline and
pre-test recordings.
Fig. 6. Error rates (%) for the numerical Stroop task for student pilots on baseline and
pre-test recordings.
Table 1
Summary of means and standard deviations for cardio-respiratory parameters throughout the sequence of cognitive tests, for baseline and pre-exam recordings. All data were
computed based on 2 minute segments.
Rest S1-N S1-E S2-N S2-E Rec1 Rec2 Num
RRI (ms)
baseline
851 765 803 795 807 841 877 828
111 127 110 109 96 110 115 120
RRI (ms)
exam
772 753 747 754 758 779 774 770
111 107 98 95 98 120 103 99
F_resp (min−1
)
baseline
9.5 19.1 17.6 17.9 17.9 15.3 16.0 17.6
2.5 3.1 3.4 4.1 3.1 2.5 2.9 2.9
F_resp (min−1
)
exam
10.4 19.8 18.1 17.1 17.4 16.0 15.4 17.5
2.5 2.5 2.2 4.1 3.3 2.1 2.8 4.3
TV (a.u.)
baseline
19242 19628 19622 19594 19552 19884 20258 19670
4759 4949 5005 5009 5023 5018 5140 4989
TV (a.u.)
exam
17864 18066 18040 18089 18063 18177 18230 18256
5 734 5 868 5 879 5 856 5 879 5 947 5977 5973
Ti/Ttot
baseline
0.36 0.40 0.38 0.39 0.38 0.37 0.37 0.39
0.05 0.04 0.05 0.06 0.04 0.04 0.04 0.04
Ti/Ttot
exam
0.36 0.41 0.39 0.38 0.39 0.37 0.38 0.39
0.05 0.03 0.02 0.05 0.03 0.04 0.07 0.05
RSA (ms)
baseline
189 63 77 67 69 82 95 74
110 35 42 35 33 49 59 46
RSA (ms)
exam
133 41 44 57 55 56 62 55
100 21 21 43 43 28 35 40
5N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
reactivity. The question remains whether these changes are merely
co-occurring or whether there is a relationship between them. To investigate
this, we performed a correlation analysis on the significant
changes from baseline to stress, being the increase in error rate for
specific emotional stimuli, the decrease in resting RRI and resting
RSA, and the reactivity decrease in RRI. This yielded a significant correlation
between the increase in error rate and the reactivity decrease
(r = 0.42; p = 0.03), meaning that the more reactivity decreases in
the stress condition, the more the error rate on specific stimuli increases.
The increase in error rate did not correlate significantly
with neither the decrease in resting RRI nor the decrease in resting
RSA.
4. Discussion
The aim of the present study was to investigate the effect of a
naturalistic stress paradigm on Stroop interferences, and whether
this effect was related to the autonomic response to stress. We hypothesized
that the effect of stress would fasten the responses, but
show an increased interference effect due to a decline in cognitive
control. Furthermore, we expected this decline in cognitive control
to be related to stress reactivity, as measured by the cardio-respiratory
parameters.
The STAI results showed that the chosen paradigm indeed elicited situational
stress, and was thus suited to the purpose of our investigation.
The apparent speed–accuracy trade-off in the stress condition
was not significant. Exam-related stress did not significantly alter
performance on the classical color–word Stroop, nor was the size of
the interference modulated. Insertion of emotional words did not significantly
alter RTs either. However, and contrary to earlier studies
where mainly RT alterations were found (Ray, 1979; Rutherford et
al., 2004), we observed a remarkable increase in error rates on the
student-pilot specific emotional stimuli, which lead us to conclude
that a specific subject-related emotional interference effect might be
sensitive to particular stress contexts. In summary, we did not replicate
Kofman et al.'s (2006) facilitatory effect of stress on executive
function, as shorter RTs did not reach significance, and co-occurred
with increased error rates. In fact, the more specific, self-referent,
and situationally relevant the emotional stimuli were, the larger
the interference, expressed as error rates, under stressful testing
conditions.
Several limitations in the present study should be addressed. The
fact that there was no control for repetition of testing does not allow
to rule out practice and carry-over effects between baseline and preexam
testing. However, the minimal period of two months between
two test sessions seems, in our opinion, sufficient to minimize such
effects. Furthermore, these effects alone could not explain our findings,
as responses to emotional words were slower and less accurate during
retest. The idea that the use of similar word material might have
boosted the emotional effect is not in line with previous research
(for a review, see Williams et al., 1997), which tends to show that
repetition might attenuate emotional responses through habituation.
Another limitation of the present study is the small number of participants.
Whereas this is a common drawback of applied research
targeting very specific populations, the levels of significance obtained
for the emotional effect, as well as the important effect sizes, warrant
the robustness of the findings. This small sample size may however account
for the lack of significance of the speed/accuracy trade-off in the
non-emotionally related stimuli. With regard to the range-correction
procedure this is in our opinion necessary and justified because, as
Fig. 8. “A cognitive-energetical linear stage model of information processing and stress. The cognitive level consists of computational processing stages derived by means of the
additive factor method (Sternberg, 1969). There are three energetical supply mechanisms, two of which are basal (arousal and activation) and coupled to respectively input and
output processing stages. The basal mechanisms are coordinated and supervised by effort, which is also directly linked to the central stage of response choice. Apart from direct
energetical supply to this stage, effort serves the function of keeping the basal mechanisms at optimal value. Information about the state of the basal mechanisms is mediated
by an evaluation mechanism.”
From Sanders (1983).
Fig. 7. Range-corrected RRI variation between rest and test recordings, for baseline and
pre-exam sessions.
6 N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
has been pointed out by several other authors in psychophysiology (e.g.
Ohman et al., 2000), the range of variation we target with subtle
changes in mental activity or overall activation are only representative
of a minor portion of the full scale of potential cardiac reactivity.
Hence, these subtle variations might be blurred by the known
interindividual variability, or by larger effects, such as these of a
real-life stress situation. Interindividual standardization procedures,
like the range-correction for the RRI, allow for a more precise investigation
of these effects.
The high error rates for the SP-specific stimuli, which are already
higher in the baseline condition, are probably related to the specificity
of the material. One could argue that the fact these words are in English
makes them stand out when compared to the neutral and emotional
stimuli, however, the pilot-specific stimuli are also presented in English
(indeed, for these categories, there is no available translation, since
during their briefings, flight, debriefings, classes and general instruction
student pilots work in English) and do not show the same pattern of results.
Previous preliminary results on university students (Pattyn, 2007)
allow us to conclude that it is the specificity of the presented material
that is the discriminant feature in eliciting the increase in error rates.
Overall, these results suggest a modulation of cognitive control related
to emotional material under the influence of stress: our subjects
seem less able to suppress irrelevant stimulus dimensions, only when
these are related to their current concern.
We had expected that the effect of stress on performance would be
related to the physiological reactivity. Indeed, physiological reactivity
range appeared to be related to the quality of cognitive performance.
This might seem surprising at first, however, it is in line with previous
psychophysiological findings. In Kennedy and Scholey's (2000) serial
subtraction task, higher increases in heart rate correlated with a better
performance (higher accuracy and throughput), and lower heart rate
baselines coupled to higher reactivity in response to cognitive processing,
manifested as important performance mediators. These authors
thus concluded that an individual's physiological efficiency, both in
terms of increase in heart rate and glucose utilization during task, is predictive
of cognitive performance. Hansen et al. (2003) investigated
vagal influences on working memory and attention, and found high
HRV individuals to perform better than low HRV subjects on both a
working memory and a sustained attention task. Carter and Pasqualini
(2004) observed a positive correlation between success on a gambling
task and anticipatory autonomic response. These results not only confirm
the association between performance quality and elicited physiological
response, but also suggest a causal link between both.
Despite the fact that correlation may not be mistaken for causality,
the observed relationship between the performance decrement
(increased error rates) and the decrease in heart rate reactivity
does fit within Sanders' model of performance. Indeed, one of our research
questions was whether the performance decrement observed
in the stress-condition, would be related to the physiological response
to the task. Sanders (1983) applied his model, depicted in Fig. 8, to explain
raised error rates under the influence of “stress”, that is, a rise in
arousal will trigger a signal to the activation system, enhancing response
readiness and response execution, which causes an increase in error
rates in the case of stronger demands on response choice. In order to
avoid unacceptable error rates, immediate arousal in conditions of already
increased activation should be suppressed and this may be accomplished
by uncoupling both systems through effort (Sanders, 1983).
Although our results only allow to suggest this as a hypothesis for
further research, the decreased arousal (decreased RRI-reactivity) in
the pre-exam session might be a suppression of the arousal response
on the background of a higher activation (higher resting heart rate
and lower resting parasympathetic tone). These results challenge the
traditional view on cardiac reactivity, stemming from stress research,
as a dispositional marker of pathology. In the present results, when
the system functions in a “nominal” state, reactivity, i.e. arousal, is
high. However, in the “stress” condition, arousal due to mental
challenge is decreased. As has been stated by various authors (e.g.
Pool, 1989; Thayer and Lane, 2000; Van Diest et al., 2006) healthy systems
are characterized by variability and flexibility, which allow to
adapt to varying environmental demands.
The interaction between cognitive control, emotion regulation and
autonomic reactivity suggested in the present findings warrants this
lead to be fruitful for future research to further unveil the relationship
between stress and performance.
Acknowledgments
This work was supported by a Prodex grant 90030 (European
Space Agency/Belgian Federal Government) and by grant ERM-HF10
(Belgian Department of Defense). Dr Pattyn's work is supported by
Table 2
Dutch neutral and general emotional words, along with affectivity and familiarity
ratings (Hermans and De Houwer, 1994) and the word frequency, as identified through
the CELEX database.
Condition Stimulus Affective load Familiarity Frequency
Emo Aids 127 45 –
Emo Alcoholisme 142 413 52
Emo Bedreiging 173 378 787
Emo Begrafenis 175 413 906
Emo Bommen 134 303 433
Emo Braaksel 17 343 100
Emo Coma 158 314 103
Emo Drugs 169 431 483
Emo Executie 133 291 240
Emo Gangster 17 306 60
Emo Gezwel 144 331 162
Emo Gijzelaar 171 298 85
Emo Haat 136 45 1558
Emo Incest 12 341 77
Emo Kanker 136 445 750
Emo lijk 175 364 993
Emo Marteling 128 288 99
Emo Misdaad 152 369 674
Emo Moord 116 378 1609
Emo Ongeluk 153 503 1783
Emo Oorlog 123 43 7810
Emo Pedofiel 148 284 7
Emo Pijn 18 55 6313
Emo Sadist 172 363 42
Emo Slachting 148 313 84
Emo Stank 167 422 664
Emo Tandpijn 172 417 12
Emo Tiran 17 278 133
Emo Tumor 131 334 200
Neutral Absorptie 375 261 40
Neutral Accent 4 453 1145
Neutral Adelaar 452 294 142
Neutral Advocaat 367 388 1170
Neutral Agentschap 389 339 112
Neutral Ansjovis 338 313 81
Neutral Antilope 452 258 15
Neutral Autobus 386 469 94
Neutral Basketbal 433 445 12
Neutral Bladzijde 422 592 490
Neutral Boog 4 353 1071
Neutral Bord 408 588 1622
Neutral Circus 477 389 274
Neutral Cirkel 419 445 831
Neutral Condoom 456 473 63
Neutral Disco 409 436 45
Neutral Eicel 447 394 110
Neutral Gebeente 33 363 84
Neutral Geur 452 48 2452
Neutral Gist 397 352 97
Neutral Haring 345 317 180
(continued on next page)
Appendix 1
7N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008
a Euro Space Society grant through the Fonds voor Wetenschappelijk
Onderzoek. Dr Bart Roelands is a postdoctoral research fellow from
the FWO (Fonds voor Wetenschappelijk Onderzoek). Dr Kolinsky is
a Research Director from the FNRS (Fonds National de la Recherche
Scientifique). The authors wish to thank the 5th Squadron from the
Belgian Air Component Flying School (1 Wing), as well as the student
pilots who agreed to participate, for their support and collaboration.
References
Akerstedt, T., Gillberg, M., Hjemdahl, P., Sigurdson, K., Gustavsson, I., Daleskog, M., et al.,
1983. Comparison of urinary and plasma-catecholamine responses to mental stress.
Acta Physiologica Scandinavica 117, 19–26.
Baayen, R.H., Piepenbrock, R., Gulikers, L., 1995. The CELEX Lexical Database [CD-ROM].
Linguistic Data Consortium, University of Pennsylvania, Philadelphia.
Berntson, G.G., Bigger, J.T., Eckberg, D.L., Grossman, P., Kaufmann, P.G., Malik, M., et al.,
1997. Heart rate variability: Origins, methods, and interpretive caveats. Psychophysiology
34, 623–648.
Bush, L.K., Hess, U., Wolford, G., 1993. Transformations for within-subject designs — a
Monte-Carlo investigation. Psychological Bulletin 113, 566–579.
Carter, S., Pasqualini, M.C.S., 2004. Stronger autonomic response accompanies better
learning: a test of Damasio's somatic marker hypothesis. Cognition & Emotion
18, 901–911.
Grossman, P., Kollai, M., 1993. Respiratory sinus arrhythmia, cardiac vagal tone, and
respiration — within-individual and between-individual relations. Psychophysiology
30, 486–495.
Hansen, A.L., Johnsen, B.H., Thayer, J.E., 2003. Vagal influence on working memory and
attention. International Journal of Psychophysiology 48 (3), 263–274.
Heims, H.C., Critchley, H.D., Martin, N.H., Jager, H.R., Mathias, C.J., Cipolotti, L., 2006.
Cognitive functioning in orthostatic hypotension due to pure autonomic failure.
Clinical Autonomic Research 16, 113–120.
Hermans, D., De Houwer, J., 1994. Affective and Subjective familiarity ratings of 740
Dutch words. Psychologica Belgica 34, 115–139.
Hockey, G.R.J., 1997. Compensatory control in the regulation of human performance
under stress and high workload: a cognitive energetical framework. Biological
Psychology 45, 73–93.
Hughdahl, K., 1995. Psychophysiology: the mind–body perspective, 1st ed. Harvard
University Press, Cambridge, MA.
Kamarck, T.W., Jennings, J.R., Pogue-Geile, M., Manuck, S.B., 1994. A multidimensional
model for cardiovascular reactivity. Stability and cross-validation in 2 adult
samples. Health Psychology 13 (6), 471–478.
Kennedy, D.O., Scholey, A.B., 2000. Glucose administration, heart rate and cognitive
performance: effects of increasing mental effort. Psychopharmacology 149, 63–71.
Kofman, O., Meiran, N., Greenberg, E., Balas, M., Cohen, H., 2006. Enhanced performance
on executive function tasks associated with examination stress: evidence
from task switching and Stroop paradigms. Cognition & Emotion 20, 577–595.
MacLeod, C.M., 1991. Half a century of research on the Stroop effect — an integrative
review. Psychological Bulletin 109, 163–203.
Matthews, G., Desmond, P.A., Joyner, L.A., Carcary, W., Gilliland, K., 1997. A comprehensive
questionnaire measure of driver stress and affect. In: Rothengatter, T., Vaya,
E.C. (Eds.), Traffic and Transport Psychology. Elsevier, Amsterdam.
Matthews, G., Dorn, L., Hoyes, T.W., Davies, D.R., Glendon, A.I., Taylor, R.G., 1998. Driver
stress and performance on a driving simulator. Human Factors 40 (1), 136–149.
McKenna, F.P., Sharma, D., 1995. Intrusive cognitions: an investigation of the emotional
Stroop task. Journal of Experimental Psychology: Learning and Cognition 21,
1595–1607.
Ohman, A., Hamm, A., Hugdahl, K., 2000. Cognition and the autonomic nervous system.
In: Cacioppo, J.T., Tassinary, L.G., Berntson, G.G. (Eds.), Handbook of Psychophysiology.
Cambridge University Press, Cambridge, MA, pp. 533–575.
Pattyn, N., 2007. Psychophysiological measures of cognitive performance in operational
conditions: applications to aviation and space environments. Doctoral Dissertation.
Vrije Universiteit Brussel and Royal Military Academy, Belgium.
Pattyn, N., Migeotte, P.F., Neyt, X., Van den Nest, A., Cluydts, R., 2010. Comparing reallife
and laboratory induced stress reactivity on cardio-respiratory parameters:
differentiation of a tonic and a phasic component. Physiology & Behavior 101,
218–223.
Pavese, A., Umiltà, C., 1998. Symbolic distance between numerosity and identity modulates
Stroop interference. Journal of Experimental Psychology. Human Perception
and Performance 24, 1535–1545.
Pool, R., 1989. Cardiac chaos. Science 243 (4897), 1419.
Ray, C., 1979. Examination stress and performance on a color–word interference test.
Perceptual and Motor Skills 49, 400–402.
Rutherford, E.M., MacLeod, C., Campbell, L.W., 2004. Negative selectivity effects and
emotional selectivity effects in anxiety: differential attentional correlates of state
and trait variables. Cognition & Emotion 18, 711–720.
Sanders, A.F., 1983. Towards a model of stress and human-performance. Acta
Psychologica 53, 61–97.
Schneider, W., Eschman, A., Zuccolotto, A., 2003. E-Prime Version 1.1 [Computer software].
Psychology Software Tools, Inc., Pittsburgh.
Selye, H., 1956. The Stress of Life. McGraw-Hill, Toronto.
Spielberger, C.D., 1983. State-Trait Anxiety Inventory. Consulting Psychologists Press
Inc.
Sternberg, S., 1969. The discovery of processing stages: Extensions of Donders' method.
In: Koster, W.G. (Ed.), Attention and performance II: Acta Psychologica, 30,
276–315.
Stroop, J.R., 1935. Studies of interference in serial verbal reactions. Journal of Experimental
Psychology 18, 643–662.
Thayer, J.F., Lane, R.D., 2000. A model of neurovisceral integration in emotion regulation
and dysregulation. Journal of Affective Disorders 61, 201–216.
Tipper, S.P., 1985. The negative priming effect: Inhibitory priming by ignored objects.
Quarterly Journal of Experimental Psychology 37A, 571–590.
Van Diest, I., Thayer, J.F., Vandeputte, B., Van De Woestijne, K.P., Van den Bergh, O.,
2006. Anxiety and respiratory variability. Physiology & Behavior 89, 189–195.
Williams, J.M.G., Watts, F.N., MacLeod, C., Mathews, A., 1997. Cognitive Psychology and
Emotional Disorders, 2nd ed. Wiley, Chichester, UK.
Wright, C.E., O'Donnell, K., Brydon, L., Wardle, J., Steptoe, A., 2007. Family history of cardiovascular
disease is associated with cardiovascular responses to stress in healthy
young men and women. International Journal of Psychophysiology 63, 275–282.
Yerkes, R.M., Dodson, J.D., 1908. The relation of strength of stimulus to rapidity of
habit-formation. Journal of Comparative Neurology and Psychology 18, 459–482.
Table 2 (continued)
Condition Stimulus Affective load Familiarity Frequency
Neutral Hazelnoot 458 419 20
Neutral Herfst 453 544 944
Neutral Hoed 433 442 1314
Neutral Honing 469 461 524
Neutral Hoofdsteun 453 395 6
Neutral Inkt 427 459 333
Neutral Inspanning 444 58 1186
Neutral Kaas 469 584 1837
Neutral Kapper 445 494 379
Neutral Klei 417 348 339
Neutral Krant 434 577 3041
Neutral Kruid 444 406 175
Neutral Magazine 442 511 49
Neutral Microscoop 4 336 136
Neutral Muren 342 527 2169
Neutral Naaimachine 4 355 77
Neutral Piano 514 464 643
Neutral Rok 425 497 903
Neutral Saffier 481 242 10
Neutral Sap 459 503 366
Neutral Schaar 389 527 218
Neutral Sigaar 323 359 751
Neutral Slaapzaal 361 331 148
Neutral Slager 37 478 303
Neutral SPIEREN 467 506 1035
Neutral Spinazie 448 428 130
Neutral Staal 339 33 327
Neutral Stoep 402 461 658
Neutral Straling 377 366 428
Neutral Streep 4 503 556
Neutral Tapijt 445 481 485
Neutral Tekstverwerker 413 413 49
Neutral Televisie 466 603 1954
Table 3
Specific emotional words presented to the student pilots
(SPs) in the experiment.
Pilot specific SP specific
Alert Test
Defect Rev
Emergency Evaluatie
Failure Commissie
Fire Buis
Mayday Schrapping
Alarm Onvoldoende
Birdstrike Instructeur
Abort Mislukken
Crash Retest
Stall Les
Checks Syllabus
Drill Falen
Incident Fout
8 N. Pattyn et al. / International Journal of Psychophysiology xxx (2013) xxx–xxx
Please cite this article as: Pattyn, N., et al., Cardiac reactivity and preserved performance under stress: Two sides of the same coin? International
Journal of Psychophysiology (2013), http://dx.doi.org/10.1016/j.ijpsycho.2013.03.008