Am J Physiol Regul Integr Comp Physiol 314: R761–R769, 2018.

First published February 14, 2018; doi:10.1152/ajpregu.00430.2017.

RESEARCH ARTICLE Cardiovascular and Renal Integration

Role of respiration in the cardiovascular response to orthostatic and

mental stress
Michal Javorka,1,2 Fatima El-Hamad,3 Barbora Czippelova,1,2 Zuzana Turianikova,1,2 Jana Krohova,1,2
Zuzana Lazarova,1,2 and X Mathias Baumert3
1
Biomedical Center Martin, Comenius University in Bratislava, Jessenius Faculty of Medicine, Martin, Slovakia; 2Department
of Physiology, Comenius University in Bratislava, Jessenius Faculty of Medicine, Martin, Slovakia; and 3School of Electrical
and Electronic Engineering, The University of Adelaide, South Australia, Australia
Submitted 6 December 2017; accepted in final form 11 February 2018

Javorka M, El-Hamad F, Czippelova B, Turianikova Z, Kro-
hova J, Lazarova Z, Baumert M. Role of respiration in the cardio-
vascular response to orthostatic and mental stress. Am J Physiol Regul
Integr Comp Physiol 314: R761–R769, 2018. First published Febru-
ary 14, 2018; doi:10.1152/ajpregu.00430.2017.—The objective of this
study was to determine the response of heart rate and blood pressure
variability (respiratory sinus arrhythmia, baroreflex sensitivity) to
orthostatic and mental stress, focusing on causality and the mediating
effect of respiration. Seventy-seven healthy young volunteers (46
women, 31 men) aged 18.4 ⫾ 2.7 yr underwent an experimental
protocol comprising supine rest, 45° head-up tilt, recovery, and a
mental arithmetic task. Heart rate variability and blood pressure
variability were analyzed in the time and frequency domain and
modeled as a multivariate autoregressive process where the respira-
tory volume signal acted as an external driver. During head-up tilt,
tidal volume increased while respiratory rate decreased. During men-
tal stress, breathing rate increased and tidal volume was elevated
slightly. Respiratory sinus arrhythmia decreased during both interven-
tions. Baroreflex function was preserved during orthostasis but was
decreased during mental stress. While sex differences were not ob-
served during baseline conditions, cardiovascular response to ortho-
static stress and respiratory response to mental stress was more
prominent in men compared with women. The respiratory response to
the mental arithmetic tasks was more prominent in men despite a
significantly higher subjectively perceived stress level in women. In
conclusion, respiration shows a distinct response to orthostatic versus
mental stress, mediating cardiovascular variability; it needs to be
considered for correct interpretation of heart rate and blood pressure
phenomena.
baroreflex; blood pressure variability; causality; head-up tilt; heart
rate variability; mental stress; respiratory pattern
INTRODUCTION
Many previous studies have assessed oscillation in cardio-
vascular measures [heart rate variability (HRV) and blood
pressure variability (BPV)] with the aim of estimating the
autonomic nervous system response to various stressors. The
two most important outcomes of the analysis, cardiac barore-
flex sensitivity and respiratory sinus arrhythmia magnitude,
both expressing predominantly the parasympathetic influences
on the heart, are typically analyzed by time or frequency
domain methods in response to passive or active orthostasis
and mental stress (e.g., Refs. 2, 4, 13, 18, 25). While these
analyses are frequently performed in laboratory settings, prac-
tical applications are limited, due to high inter- and intraindi-
vidual differences that result in poor sensitivity and specificity
for diagnostic purposes.
Methodological advances in the analysis of cardiovascular
oscillations have resulted in a better understanding of their
origin, making it necessary to reconsider previous results
obtained with various stressors by taking into consideration the
following important observations.
First, during the last decade, it has become evident that
closed-loop interaction of heart rate and blood pressure needs
to be considered in any analysis of spontaneous baroreflex
sensitivity. While blood pressure influences heart rate via
baroreflex, which has been traditionally considered to be the
dominant mechanisms of coupling, an increasing number of
studies show that feedforward mechanisms elicited by Frank-
Starling and Windkessel effects constitute significant contribu-
tions to the heart rate-blood pressure relationship (11, 30).
Therefore, the analysis of interaction between heart rate and
blood pressure oscillations with the aim of probing baroreflex
sensitivity demands closed-loop assessment of the causal in-
fluences in both directions (19).
Second, previous studies have demonstrated the important
influence of respiration and respiratory pattern changes on
cardiovascular oscillations (6, 29). This is especially important
when respiratory pattern changes in response to a stimulus are
to be expected, e.g., during mental stress (14, 22). Furthermore,
the sex-specific influence on autonomic nervous system control
at rest and during various challenges is often neglected (23).
The aim of our study was to demonstrate the effect of
respiration and causal relationships in heart rate and blood
pressure control by using two established test paradigms during
physical (orthostatic) and cognitive (mental arithmetic) stress.
We hypothesized that analysis of tidal volume and respiratory
rate will give important insights into heart rate and blood
pressure stress response. Additionally, sex differences in car-
diovascular stress response were analyzed.
METHODS

Subjects
Address for reprint requests and other correspondence: M. Javorka, Faculty
of Medicine, Biomedical Center Martin Comenius Univ., Mala Hora 4C, Seventy-seven healthy young volunteers (46 women, 31 men),
03601 Martin, Slovakia (e-mail: mjavorka@jfmed.uniba.sk). recruited from local elementary schools and high schools and the

http://www.ajpregu.org 0363-6119/18 Copyright © 2018 the American Physiological Society R761

Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
R762 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS
university, with a mean age of 18.4 yr (SD ⫽ 2.7 yr) participated in
this study. All subjects were normotensive (normal blood pressure
range for the given age on regular medical examinations and during
three sequential measurements separated by 5 min at the start of
examination protocol in the laboratory) and within the normal range
of body mass index for their age (7). Subjects were instructed not to
use substances influencing the autonomic nervous system or cardio-
vascular system activity. Women were examined in the proliferative
phase (6th to 13th day) of the menstrual cycle. All procedures were
approved by Ethical Committee of the Jessenius Faculty of Medicine,
Comenius University, and all participants or their legal guardian (for
minors) signed a written informed consent.
Data Acquisition
We recorded continuous finger arterial blood pressure [beat-to-beat
systolic and diastolic blood pressure (SBP and DBP)] noninvasively
by the photoplethysmographic volume-clamp method (Finometer Pro;
FMS) and RR interval by the ECG (horizontal bipolar thoracic lead;
CardioFax ECG-9620; NihonKohden). Respiratory volume (RV)
changes were recorded by respiratory inductive plethysmography
(RespiTrace 200; NIMS), using two belts (thoracic and abdominal).
Qualitative diagnostic calibration of the respiratory signal (9) was
followed by quantitative calibration, where participants breathed into
a 1,000-ml plastic bag. All signals were digitized and transferred to a
personal computer at a sampling rate of 1,000 Hz using a PowerLab
8/35 (ADInstruments) device.
For each breath, respiratory rate and tidal volume were measured
from the raw RV signal. Tidal volume was taken as the difference
between local minimum (onset of inspiration) and subsequent maxi-
mum (offset of expiration) of the RV signal. Respiratory rate in cycles
per minute was calculated as the reciprocal value of breathing cycle
duration (time distance between 2 consecutive onsets of inspiration).
Minute ventilation was calculated as a product of respiratory rate and
tidal volume for each breath. For further analysis, median values of all
respiratory measures as representative values for the given phase of
the study protocol were used. For further beat-to-beat analysis, the RV
signal (RV signal) was sampled at the R-wave occurrence in ECG; the
RR interval corresponds to the time interval to the subsequent R wave;
and the SBP value corresponds to the systolic blood pressure en-
trained by the RR interval.
Study Protocol
Participants were positioned on a tilt table with their feet
touching the footboard at the end of the table. A restraining strap
secured at the thigh level was used to provide additional support
and safety. The participants were asked to avoid moving and
speaking during the measurement. The study protocol consisted of
five consecutive phases: 1) supine rest (REST, 15 min); 2) head-up
tilt (HUT, the subject was tilted to 45° on a motor-driven tilt table
for 8 min to evoke mild orthostatic stress); 3) recovery from HUT
in the supine position (REC1, 10 min); 4) mental arithmetic task in
the supine position (MA, 6 min); and 5) recovery (REC2, 10 min).
During the MA task, participants were instructed to sum up
three-digit numbers until a one-digit number was reached and to
decide if the final one-digit number was odd or even. The numbers
were projected on the ceiling of the examination room. After decision-
making, participants had to click on the corresponding push button
Fig. 1. Study protocol timeline with 5 phases: Gray
boxes indicate the analyzed 300-beat-long intervals.
REST, supine rest; HUT, head-up tilt to 45°; REC1,
1st supine recovery; REC2, 2nd supine recovery;
MA, mental arithmetics.
AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org
Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
projected on the ceiling using a computer mouse. Throughout the MA
task, participants were disturbed by the rhythmic sound of a metro-
nome and instructed to perform the MA task as quickly as possible
with minimum error. After completing the protocol, participants were
asked to express their subjective perceived stress level (PSL) during
MA on the visual analog scale (0 –10).
Data Analysis
To exclude transient changes in cardiovascular parameters between
consecutive phases of the experimental protocol from analysis, we
extracted segments of 300 beats from the original recordings as
follows: the first segment (REST) started 8 min after the beginning of
recording, the second segment (HUT) started 3 min after tilt, the third
segment (REC1) started 7 min before the MA task, and the fourth
segment (MA) started 2 min after the beginning of this phase. Since
our study focuses on stress responses, the fifth phase (REC2) was not
included in the analysis (Fig. 1).
Univariate Time and Frequency Domain Analysis
HRV analysis. For traditional time domain analysis of HRV, we
computed three most commonly used measures: meanNN, the mean
beat-to-beat interval of normal heartbeats; SDNN, standard deviation
of NN intervals, reflecting the overall variability magnitude; and
RMSSD, the root mean square of successive beat-to-beat differences,
reflecting the average magnitude of changes in RR intervals length
between two consecutive beats, which is regarded a marker of vagal
heart rate control. In the frequency domain, spectral analysis was
performed using fast Fourier transform to obtain spectral powers in
low-frequency (LF HRV: 0.04 – 0.15 Hz) and high-frequency (HF
HRV: 0.15– 0.5 Hz) bands.
BPV analysis. From SBP signals we computed the following linear
measures: mean SBP, mean systolic blood pressure value; SD SBP,
standard deviation of systolic blood pressure values; and RMSSD
SBP, root-mean-square of successive differences of SBP values. In
the frequency domain, spectral analysis was performed using fast
Fourier transform to obtain spectral powers in low-frequency (LF
SBPV: 0.04 – 0.15 Hz) and high-frequency (HF SBPV: 0.15– 0.5 Hz)
bands. In addition, mean DBP for the given time series segment was
calculated.
Model-Based Analysis of Respiratory Sinus Arrhythmia and
Baroreflex
We adopted a multivariate linear autoregressive model to explore
the relationship among RR, SBP, and respiration, focusing on the
effects of respiration and blood pressure changes, respectively, on RR,
yielding indexes of respiratory sinus arrhythmia and baroreflex sen-
sitivity. The model distinguishes the baroreflex effect of SBP on RR
from the hemodynamic influence in the opposite direction (from RR
to SBP). This so-called causal approach extends previous studies on
SBP and RR that considered interactions only in one direction, from
SBP to RR. Recent studies (11, 19) demonstrate that feedforward
influences from RR to SBP are relatively strong and distort the results
of baroreflex analysis from spontaneous RR and SBP oscillations
when neglected. The model also enables quantifying the magnitude of
respiratory sinus arrhythmia by considering changes in RV, because,
as demonstrated in previous studies (6, 29), the respiratory pattern
 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS
significantly influences the magnitude of respiratory sinus arrhythmia,
regardless of autonomic nervous system activity.
Beat-to-beat values of RR, SBP, and RV were detrended using a
time-varying impulse response high-pass filter (37) with a cut-off
frequency of 0.04 Hz and subsequently normalized to zero mean and
unit variance. To study the baroreflex contribution to HRV, power
contribution analysis (10, 28) and parametric transfer function anal-
ysis (12) were performed, using a closed-loop autoregressive model
with two outputs and one external input (2ARX) (3). The model was
defined as:
A1共z兲RR共i兲 ⫽ B1共z兲SBP(i) ⫹ B2共z兲RV(i) ⫹ eRR(i)
A2共z兲SBP共i兲 ⫽ B3共z兲RR(i) ⫹ B4共z兲RV(i) ⫹ eSBP(i)
where A1 and A2 are polynomials that describe the dependence of the
output signals on their own past, while B1, B2, B3, and B4 are
polynomials that describe the effect of the inputs on the output signals.
Variables eRR and eSBP are white noise sources with zero mean and
variances ␴2RR and ␴2SBP, respectively. In this model, RR and SBP
interact in a closed-loop fashion, where the SBP-to-RR arm represents
baroreflex feedback control and the RR-to-SBP arm represents the
mechanical feedforward influences from RR to SBP. RV was modeled
as an independent autoregressive process.
Model orders were selected in the range 4 to 12, minimizing the
Akaike information criterion (1). Model parameters were estimated
using least squares method and validated using residual analysis and
goodness of fit measures. Model parameters were transformed into the
frequency domain by using the Fourier transform and the following
measures were estimated for each data segment: power contribution
from SBP to RR, expressed as a percentage of total RR power
(PSBP¡RR); power contribution from RV to RR, expressed as a
percentage of total RR power (PRV¡RR); gain of the causal transfer
function from SBP to RR, sampled at the LF peak of the SBP
spectrum (GSBP¡RR); and gain of the transfer function from RV to
RR, sampled at the HF peak of the R spectrum (GRV¡RR).
Statistics
Nonparametric tests were used to take into account the non-
Gaussian distribution of the assessed measures. Between-phase dif-
ferences were analyzed by the Friedman test followed by the Conover
test for post hoc comparison (effect of HUT: phase 2 vs. phase 1;
effect of MA: phase 4 vs. 3; and between stressors comparison: phase
2 vs. phase 4). Sex differences were tested with the Mann-Whitney
U-test. Correlations between PSL and cardiovascular or respiratory
variables were quantified by the Spearman correlation coefficient.
P ⬍ 0.05 was considered statistically significant.
RESULTS
HRV and SBPV
Baseline resting heart rate was 66 beats/min on average and
tended to be higher in women (for meanNN: P ⫽ 0.057) (Fig.
2). During both challenges (HUT and MA), mean RR interval
(meanNN) significantly decreased (heart rate increased) (P ⬍
0.001 in both HUT and MA compared with preceding resting
phase), and this decrease was significantly more prominent
during HUT than during MA (P ⬍ 0.001 for HUT vs. MA
comparison). The magnitude of meanNN decrease (heart rate
increase) in response to HUT was higher in men (P ⫽ 0.003)
resulting in nonsignificant difference in meanNN during HUT
(P ⫽ 0.582). No significant sex difference in the response
magnitude of heart rate to MA was observed (P ⫽ 0.489).
During both HUT and MA, all time domain HRV (Fig. 3,
top) measures significantly decreased (for all comparisons P ⬍
0.001). The decrease was more pronounced during HUT, and
R763
(1)
(2)
Fig. 2. Mean beat-to-beat interval of normal heartbeats (meanNN; a reciprocal
value of heart rate) during 4 phases of study protocol. Bars correspond to
median values; black and gray bars correspond to men and women, respec-
tively. REST, supine rest; HUT, head-up tilt to 45°; REC1, 1st supine
recovery; REC2, 2nd supine recovery; MA, mental arithmetics. *Significant
effect of stress (HUT vs. REST or MA vs. REC1), §Significant difference
between 2 challenges (MA vs. HUT). #Significant sex difference.
the differences between HUT and MA were all statistically
significant (for SDNN: P ⫽ 0.012; for RMSSD: P ⬍ 0.001).
While similar effects were observed for power HF in the
frequency domain of HRV (decrease during HUT: P ⬍ 0.001;
decrease during MA: P ⬍ 0.001; HUT vs. MA: P ⬍ 0.001), the
LF power decreased significantly only during MA compared
with the preceding phase (MA: P ⬍ 0.001; HUT: P ⫽ 0.070;
Fig. 3, bottom). No significant sex differences in time and
frequency domain HRV measures (P ⫽ 0.122– 0.934) or their
response magnitude from baseline to HUT (P ⫽ 0.216 – 0.975)
or from REC1 to MA were found (P ⫽ 0.406 – 0.547).
SBP and DBP at baseline were 121 and 70 mmHg on
average, respectively. During HUT, SBP significantly de-
creased (P ⬍ 0.001) while no significant change in DBP was
found (P ⫽ 0.999) (Fig. 4). Both mean SBP and mean DBP
significantly increased during MA compared with preceding
supine rest (P ⬍ 0.001 for both). No significant sex differences
in mean SBP values were observed during protocol (P ⫽
0.088 – 0.289). Mean DBP values observed during baseline,
HUT, and MA, on the other hand, were all significantly higher
in men (P ⫽ 0.030, 0.013, and 0.020, respectively). No signif-
icant sex differences in the response magnitude of mean SBP
or DBP changes to the HUT or MA in comparison with the
preceding resting period were found (P ⫽ 0.236 – 0.553).
All SBPV measures (Fig. 4) significantly increased during
HUT and decreased during MA (for all comparisons: P ⬍
0.001). When comparing both sexes, we found higher beat-to-
beat SBPV (RMSSD SBP: P ⫽ 0.006; HF SBPV: P ⫽ 0.002)
during orthostasis in men. Accordingly, the response magni-
tude of RMSSD SBP and HF SBPV from supine rest to HUT
was higher in men (P ⫽ 0.038 and P ⬍ 0.001, respectively).

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org

Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
R764 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS

Fig. 3. Time and frequency domain heart rate variability indexes during four phases of study protocol. REST, supine rest; HUT, head-up tilt to 45°; REC1, 1st
supine recovery; MA, mental arithmetics; SDNN, standard deviation of normal RR intervals; RMSSD, root-mean-square of successive beat-to-beat difference
in RR interval length; LF_HRV and HF_HRV, spectral power of RR intervals oscillations in low (0.04 – 0.15 Hz)- and high-frequency (0.15– 0.5 Hz) bands,
respectively. Bars correspond to median values; black and gray bars correspond to men and women, respectively. *Significant effect of stress (HUT vs. REST
or MA vs. REC1), §Significant difference between 2 challenges (MA vs. HUT).

Respiration Measures ventilation in men overall (P ⬍ 0.001 during whole protocol).

Minute ventilation significantly and similarly increased dur-
ing both HUT and MA compared with baseline (P ⬍ 0.001 for
both HUT vs. REST and MA vs. REC1; P ⫽ 0.495 for HUT
vs. MA). Orthostatic challenge was accompanied by an in-
crease of tidal volume (P ⬍ 0.001) and modest but significant
decrease of respiratory rate (P ⫽ 0.003). During MA, the
respiratory rate increased (P ⬍ 0.001) and tidal volume mar-
ginally significantly increased (P ⫽ 0.049) (Fig. 5). While tidal
volume was significantly higher in men compared with women
throughout the entire protocol (P ⫽ 0.001– 0.004), no signifi-
cant sex differences in respiratory rate were observed (P ⫽
0.246 – 0.869). This resulted in a significantly higher minute
When analyzing the response magnitude as the difference
between orthostasis or MA and the preceding resting phase, the
increase of tidal volume and minute ventilation during HUT
was significantly higher in men (orthostatic response: P ⫽
0.009 and 0.003 for tidal volume and minute ventilation,
respectively; MA response: P ⫽ 0.484 and 0.126 for tidal
volume and minute ventilation, respectively).
Model-Based Analysis
Power contribution from SBP to RR, expressed as a percent-
age of total RR power (PSBP¡RR), significantly increased
during HUT (P ⬍ 0.001) but not during MA (P ⫽ 0.631; HUT

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org

Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS R765

Fig. 4. Mean systolic (mean SBP) and diastolic blood pressure (mean DBP) values and linear systolic blood pressure variability indexes during study protocol.
REST, supine rest; HUT, head-up tilt to 45°; REC1, 1st supine recovery; MA, mental arithmetics; SD SBP, standard deviation of beat-to-beat systolic blood
pressure values; RMSSD SBP, root-mean-square of successive differences in systolic blood pressure values; LF_SBPV and HF_SBPV, spectral power of systolic
blood pressure oscillations in low (0.04 – 0.15 Hz) and high frequency (0.15– 0.5 Hz) bands, respectively. Bars correspond to median values; black and gray bars
correspond to men and women, respectively. *Significant effect of stress (HUT vs. REST or MA vs. REC1), §Significant difference between 2 challenges (MA
vs. HUT), #Significant sex difference.

vs. MA: P ⬍ 0.001) (Fig. 6). Gain of the causal transfer
function from SBP to RR (GSBP¡RR) was significantly lower
during MA compared with HUT (P ⫽ 0.003). No significant
sex differences in SBP to RR interconnection measures were
observed. However, when the magnitude of response was
calculated as the difference between MA and preceding rest,
the decrease in GSBP¡RR was significantly more expressed in
women (P ⫽ 0.048).
Power contribution from respiration to RR signal, expressed
as a percentage of total RR power (PRV¡RR), significantly and
similarly (HUT vs. MA: P ⫽ 0.378) decreased during HUT
(P ⫽ 0.001) and MA (P ⬍ 0.001). Although significant during
both HUT and MA (P ⬍ 0.001 for both), the decrease in the
causal transfer function gain from respiration to RR (GRV¡RR)
was more prominent during orthostatic challenge (HUT vs.
MA: P ⫽ 0.020). A significantly higher PRV¡RR was found in
women at REC1 phase (P ⫽ 0.008).
Correlation Between PSL During MA Task and
Cardiovascular and Respiratory Measures
The PSL during MA was significantly higher in women
compared with men (P ⫽ 0.034). When pooling all partici-
pants, significant negative correlations between PSL and sev-
eral HRV measures (SDNN: rho ⫽ ⫺0.367, P ⫽ 0.013;
RMSSD: rho ⫽ ⫺0.327, P ⫽ 0.028; power HF: rho ⫽
⫺0.310, P ⫽ 0.038) and SD SBP (rho ⫽ ⫺0.308, P ⫽ 0.039)
obtained during MA were found. When analyzed separately by
sex, significant correlations between PSL and HRV or BPV
measures in the MA phase were found in men only in HRV
measures (SDNN: rho ⫽ ⫺0.602, P ⫽ 0.006; RMSSD: rho ⫽
⫺0.640, P ⫽ 0.003; power HF: rho ⫽ ⫺0.610, P ⫽ 0.006).
DISCUSSION
In this study, we determined HRV and BPV during ortho-
static and mental stress, focusing on the effect of respiration.
Analysis of RV indicates distinct differences between both
challenges. Respiratory sinus arrhythmia magnitude demon-
strates the decrease in cardiac parasympathetic activity during
both HUT and MA. Considering bidirectional interconnection
between heart rate and blood pressure, causal analysis shows
that baroreflex function is preserved during orthostasis, but it
decreases during mental stress. While sex differences were not
observed during baseline conditions, cardiovascular response

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org

Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
R766 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS

Fig. 5. Ventilation characteristics during 4 phases of study protocol. REST, supine rest; HUT, head-up tilt to 45°; REC1, 1st supine recovery; MA, mental
arithmetics. Bars correspond to median values; black and gray bars correspond to men and women, respectively. *Significant effect of stress (HUT vs. REST
or MA vs. REC1). §Significant difference between 2 challenges (MA vs. HUT), #Significant sex difference.

to orthostatic stress and respiratory response to mental stress
was more prominent in men compared with women.
Comparing the Effects of Orthostatic and Cognitive
Challenges on the Cardiovascular System
During both HUT and MA, all HRV measures decreased,
except LF power in HRV, which only decreased during HUT.
These findings are in accordance with previous studies on HUT
(32, 40) and MA (5, 16) and are generally interpreted as a
decrease of the parasympathetic activity to the heart.
The dominant influence of respiration on short-term HRV
(respiratory sinus arrhythmia) (6), and the ventilatory changes
observed during HUT, increase of tidal volume and a decrease
in respiratory rate, would increase the magnitude of respiratory
sinus arrhythmia (15). Despite this, HF power and HRV
magnitude in general decreased during HUT, indicating de-
creased parasympathetic activity during orthostasis. This was
confirmed by attenuated gain from respiratory signal to heart
rate oscillations; the unit magnitude of tidal volume evoked
lower respiratory sinus arrhythmia during HUT compared with
preceding supine rest.
On the contrary, the respiratory changes observed during
MA, a marked increase in breathing rate and only a slight
increase in tidal volume, would theoretically decrease respira-
tory sinus arrhythmia magnitude and HRV variability. There-
fore, in this case, the interpretation of decreased HF power as
parasympathetic withdrawal should be taken with caution. We
found that the gain from respiration to heart rate was also
decreased during MA compared with the preceding supine
phase. In addition, the relative contribution of respiration to
HRV was decreased, confirming that cardiac parasympathetic
activity is also decreased during cognitive load.
The decrease in HRV and the decrease in gain from respi-
ration to heart rate were more prominent during HUT com-
pared with MA task, indicating a more pronounced parasym-
pathetic withdrawal during orthostatic challenge. The mecha-
nisms behind parasympathetic inhibition during the two
challenges are rather different. During HUT, blood pooling in
the lower part of the body leads to a decreased venous return
and a decrease of arterial blood pressure sensed by high-
pressure baroreceptors. The responses include cardiac para-
sympathetic withdrawal and increased sympathetic activity to
the heart and vasculature. In contrast, during cognitive load,
parasympathetic withdrawal and sympathetic activation are
presumably initiated by projections from the amygdala and
occur without major changes in venous return (24, 33, 42). As
a result, the SBP and DBP significantly increased during MA.
This blood pressure response corresponding to autonomic
changes was overridden by a decreased venous return during
HUT resulting in a decrease of both SBP and DBP compared
with preceding supine rest.
In accordance with previous studies (26, 31), SBP variability
increased during HUT. This increase was related to both
overall and beat-to-beat variability. Previously, it was inter-
preted as an effect of an increased sympathetic activity to the
blood vessels (39). However, the respiratory changes observed
in our study, the increase in tidal volume and the mild decrease
in the breathing rate, could also explain this phenomenon. The
importance of respiration on short-term BPV measures was
also evident during MA. BPV was expected to increase due to
sympathetic activation, when in fact it decreased, possibly due
to the increase in breathing rate and the decrease in tidal
volume. Alternatively, the decrease in BPV during MA could
be related to the lower feedforward transfer of oscillations from
heart rate to SBP related to the decreased HRV in this phase as
demonstrated in several previous studies on causal SBP to RR
oscillations transfer analysis (11, 20).
The increase in the relative contribution of SBP to RR
(PSBP¡RR) during HUT illustrates the cardiovascular control
system’s heightened effort to maintain blood pressure and
perfusion during orthostatic challenge. Despite decreased para-
sympathetic activity, the gain from SBP to RR oscillations, i.e.,
causal analysis-derived baroreflex sensitivity, was not changed
during HUT. In contrast, the gain from SBP to RR decreased
during MA when blood pressure is relatively stable and para-
sympathetic activity decreased.
Respiratory Pattern Changes
In agreement with previous studies, we found an increase in
minute ventilation during both challenges (14, 17). The pattern
to achieve this was different: while during HUT the effect of
increased tidal volume was prevailing, the increase in breath-

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org

Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS R767

Fig. 6. Model-based indexes of respiratory sinus arrhythmia and baroreflex during four phases of study protocol. REST, supine rest; HUT, head-up tilt to 45°;
REC1, 1st supine recovery; MA, mental arithmetics; PSBP¡RR and PRV¡RR, power contribution from systolic blood pressure to RR intervals or from respiratory
volume signal to RR intervals, respectively, both expressed as a percentage of total RR power; GSBP¡RR and GRV¡RR, gain of the causal transfer function from
systolic blood pressure to RR or from respiratory volume signal to RR, respectively. Bars correspond to median values; black and gray bars correspond to men
and women, respectively. *Significant effect of stress (HUT vs. REST or MA vs. REC1), §Significant difference between 2 challenges (MA vs. HUT).
#Significant sex difference.

ing rate was more dominant during MA. We attribute this
difference to the minimization of breathing effort to increase
the minute ventilation after a change from supine to upright
position; it is more effective to increase tidal volume than
breathing rate.
Sex Differences
At rest in the supine position, no significant sex differences
in mean HR, HRV, and BPV were found. Although this is in
accordance with several previous studies on HRV, a recent
meta-analysis demonstrated significantly higher mean heart
rate and lower overall magnitude of HRV together with an
increased HF power in women (23). The difference between
age groups analyzed in our study (adolescents and young
adults) and those included in the meta-analysis performed
aggregated across all age groups could be responsible for this
discrepancy. For example, in our previous study on the rela-
tively large cohort of adolescents (n ⫽ 206; age range: 15–19
yr), a paradoxically decreased magnitude of respiratory sinus
arrhythmia in women was consistently found (38).
Mean DBP values were higher in men compared with
women. Given the similar HR in both sexes, the lower decay of
blood pressure curve during diastole (run-off effect) could
indicate either decreased peripheral vascular resistance and/or
arterial compliance in men (41). This observation warrants
further investigation.
In response to HUT, the heart rate increase was more
prominent in men. The heart rate increase compensates the

AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org

Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
R768 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS
decrease in blood pressure associated with a decreased venous
return during a changed body position. Higher capacity of the
venous bed in the lower part of the body of men, leading to
reduced venous return compared with women (8, 27), could
require more prominent HR response. The alternative expla-
nation, decreased sensitivity of baroreflex, was not supported
by our data, showing no significant difference in the model-
based baroreflex sensitivity estimation (GSBP¡RR). In contrast
to our observations, few previous studies, comparing cardio-
vascular response to HUT between the subgroups of men and
women, reported no significant differences (34) or a more
prominent heart rate increase in women only at a tilt angle of
60° (36). Attenuated heart rate increase in young women could
be one of the factors contributing to the higher incidence of
orthostatic hypotension and fainting (21).
Given the comparable cardiac baroreflex characteristics in
men and women, higher beat-to-beat SBP variability observed
in men during passive orthostasis could be explained by the
mechanical effect of the respiration. Since the increase in tidal
volume (and hence minute ventilation) was higher in men,
intrathoracic pressure changes and subsequent venous return
oscillations are expected to be more prominent, resulting in
augmented respiration-related SBP oscillations.
Cardiovascular autonomic control differs between sexes, but
is often ignored in the analysis of resting cardiovascular vari-
ability and responses to various challenges. In summary, our
data suggest that the influence of sex should be considered
when performing cardiovascular variability analysis, in partic-
ular stress response tests. The age of our participants ranged
between 13 and 23 yr (interquartile range: 17.0 –20.3 yr),
corresponding to adolescents and young adults according to
World Health Organization classification. We assume that a
broader range of pubertal development including earlier stages
in men in contrast to completed pubertal development in the
majority of girls could be partially responsible for the observed
sex-related differences. A relation between pubertal develop-
ment (e.g., using Tanner score) and cardiovascular/respiratory
response to stress requires further study.
PSL During Mental Arithmetic Task and Autonomic
Nervous System Response
Despite significantly higher PSLs during MA in women, the
changes in basic cardiovascular (HR, SBP, and DBP) and
respiratory measures (tidal volume, respiratory rate, and min-
ute ventilation) in response to MA were not significantly
different between sexes. This illustrates that subjective percep-
tion of stress is not necessarily correlated with the level of
autonomic nervous system response. Only cardiac baroreflex
sensitivity responded more prominently (higher parasympa-
thetic withdrawal) in women compared with men. A study
comparing the response to mental load related to Stroop color
word test between sexes reported a more prominent increase in
blood pressure in men but no difference in respiratory sinus
arrhythmia magnitude (35). The sex differences in mental
stress response could be test and population specific, requiring
further study.
The PSL score correlated negatively with HRV measures
and SD SBP (indicating a higher decrease in parasympathetic
activity in subjects with higher PSL) only in men. The lack of
correlations in women deserves further investigation.
AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org
Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
Perspectives and Significance
Historically, spontaneous cardiovascular oscillations, HRV
and BPV, were analyzed separately. It was believed that
short-term variability mostly reflects autonomic nervous sys-
tem activity. This triggered a search for the best measures
reflecting parasympathetic or sympathetic activity from HRV
or BPV during various challenges including orthostasis and
cognitive load. However, by ignoring the complex relation-
ships underpinning these oscillations, the results of previous
studies may have been distorted. The bidirectional interaction
between heart rate and blood pressure should be considered
when baroreflex function is analyzed from the cardiovascular
oscillations. Additionally, an important driving force of respi-
ration should not be neglected.
From the physiological point of view, our study demon-
strated that even after adjusting for breathing pattern changes,
parasympathetic activity directed to the sinoatrial node, medi-
ating respiratory sinus arrhythmia, is significantly reduced
during both challenges. Considering the cardiac branch of
baroreflex (cardiac chronotropic response to blood pressure
change), we found well-preserved baroreflex sensitivity during
both stress conditions (although it was slightly lower during a
cognitive load). Decreased baroreflex sensitivity reported in
previous studies during both challenges may therefore be an
artifactual finding originating in the noncausal approach to
baroreflex function analysis. To correctly interpret cardiovas-
cular oscillations, simultaneous recording of several parame-
ters, including respiration, is necessary.
Conclusion
The respiratory response to orthostatic stress is distinctly
different from that to mental stress, raising tidal volume and
respiratory rate, respectively. When analyzing cardiovascular
oscillations, respiration must be included in the interpretation
of heart rate and blood pressure phenomena related to respira-
tory sinus arrhythmia and spontaneous baroreflex sensitivity.
GRANTS
This study was supported by Grants APVV-0235-12 and VEGA 1/0117/17,
BioMed (ITMS 26220220187), and Grant UK/83/2016.
DISCLOSURES
No conflicts of interest, financial or otherwise, are declared by the authors.
AUTHOR CONTRIBUTIONS
M.J. and M.B. conceived and designed research; M.J., B.C., Z.T., J.K.,
Z.L., and M.B. performed experiments; M.J., F.E.-H., B.C., Z.T., J.K., Z.L.,
and M.B. analyzed data; M.J., F.E.-H., B.C., and M.B. interpreted results of
experiments; M.J. and J.K. prepared figures; M.J., F.E.-H., and M.B. drafted
manuscript; M.J., F.E.-H., and M.B. edited and revised manuscript; M.J.,
F.E.-H., B.C., Z.T., J.K., Z.L., and M.B. approved final version of manuscript.
REFERENCES
1. Akaike H. A new look at the statistical model identification. IEEE Trans
Automat Contr 19: 716 –723, 1974. doi:10.1109/TAC.1974.1100705.
2. Bahjaoui-Bouhaddi M, Henriet MT, Cappelle S, Dumoulin G, Reg-
nard J. Active standing and passive tilting similarly reduce the slope of
spontaneous baroreflex in healthy subjects. Physiol Res 47: 227–235,
1998.
3. Baselli G, Porta A, Rimoldi O, Pagani M, Cerutti S. Spectral decom-
position in multichannel recordings based on multivariate parametric
identification. IEEE Trans Biomed Eng 44: 1092–1101, 1997. doi:10.
1109/10.641336.
 RESPIRATION AND CARDIOVASCULAR RESPONSE TO STRESS
4. Bernardi L, Rossi M, Ricordi L. Clinical assessment of respiratory sinus
arrhythmia by computerized analysis of RR interval and respiration. G Ital
Cardiol 22: 517–529, 1992.
5. Berntson GG, Cacioppo JT, Binkley PF, Uchino BN, Quigley KS,
Fieldstone A. Autonomic cardiac control. III. Psychological stress and
cardiac response in autonomic space as revealed by pharmacological
blockades. Psychophysiology 31: 599 –608, 1994. doi:10.1111/j.1469-
8986.1994.tb02352.x.
6. Brown TE, Beightol LA, Koh J, Eckberg DL. Important influence of
respiration on human R-R interval power spectra is largely ignored. J Appl
Physiol (1985) 75: 2310 –2317, 1993. doi:10.1152/jappl.1993.75.5.2310.
7. Cole TJ, Bellizzi MC, Flegal KM, Dietz WH. Establishing a standard
definition for child overweight and obesity worldwide: international sur-
vey. BMJ 320: 1240 –1243, 2000. doi:10.1136/bmj.320.7244.1240.
8. Convertino VA. Gender differences in autonomic functions associated
with blood pressure regulation. Am J Physiol Regul Integr Comp Physiol
275: R1909 –R1920, 1998.
9. De Groote A, Paiva M, Verbandt Y. Mathematical assessment of
qualitative diagnostic calibration for respiratory inductive plethysmogra-
phy. J Appl Physiol (1985) 90: 1025–1030, 2001. doi:10.1152/jappl.2001.
90.3.1025.
10. El-Hamad F, Lambert E, Abbott D, Baumert M. Relation between QT
interval variability and muscle sympathetic nerve activity in normal
subjects. Am J Physiol Heart Circ Physiol 309: H1218 –H1224, 2015.
doi:10.1152/ajpheart.00230.2015.
11. Faes L, Nollo G, Porta A. Mechanisms of causal interaction between
short-term RR interval and systolic arterial pressure oscillations during
orthostatic challenge. J Appl Physiol (1985) 114: 1657–1667, 2013.
doi:10.1152/japplphysiol.01172.2012.
12. Faes L, Porta A, Cucino R, Cerutti S, Antolini R, Nollo G. Causal
transfer function analysis to describe closed loop interactions between
cardiovascular and cardiorespiratory variability signals. Biol Cybern 90:
390 –399, 2004. doi:10.1007/s00422-004-0488-0.
13. Fauvel JP, Cerutti C, Quelin P, Laville M, Gustin MP, Paultre CZ,
Ducher M. Mental stress-induced increase in blood pressure is not related
to baroreflex sensitivity in middle-aged healthy men. Hypertension 35:
887–891, 2000. doi:10.1161/01.HYP.35.4.887.
14. Grassmann M, Vlemincx E, von Leupoldt A, Mittelstädt JM, Van den
Bergh O. Respiratory changes in response to cognitive load: a systematic
review. Neural Plast 2016: 8146809, 2016. doi:10.1155/2016/8146809.
15. Grossman P, Taylor EW. Toward understanding respiratory sinus ar-
rhythmia: relations to cardiac vagal tone, evolution and biobehavioral
functions. Biol Psychol 74: 263–285, 2007. doi:10.1016/j.biopsycho.2005.
11.014.
16. Guasti L, Simoni C, Mainardi L, Crespi C, Cimpanelli M, Klersy C,
Gaudio G, Grandi AM, Cerutti S, Venco A. Global link between heart
rate and blood pressure oscillations at rest and during mental arousal in
normotensive and hypertensive subjects. Auton Neurosci 120: 80 –87,
2005. doi:10.1016/j.autneu.2005.02.008.
17. Chang AT, Boots RJ, Brown MG, Paratz JD, Hodges PW. Ventilatory
changes following head-up tilt and standing in healthy subjects. Eur J Appl
Physiol 95: 409 –417, 2005. doi:10.1007/s00421-005-0019-2.
18. Iacoviello M, Forleo C, Guida P, Sorrentino S, D’Andria V, Rodio M,
D’Alonzo L, Favale S. Independent role of reduced arterial baroreflex
sensitivity during head-up tilt testing in predicting vasovagal syncope
recurrence. Europace 12: 1149 –1155, 2010. doi:10.1093/europace/
euq149.
19. Javorka M, Czippelova B, Turianikova Z, Lazarova Z, Tonhajzerova
I, Faes L. Causal analysis of short-term cardiovascular variability: state-
dependent contribution of feedback and feedforward mechanisms. Med
Biol Eng Comput 55: 179 –190, 2017. doi:10.1007/s11517-016-1492-y.
20. Javorka M, Krohova J, Czippelova B, Turianikova Z, Lazarova Z,
Javorka K, Faes L. Basic cardiovascular variability signals: mutual
directed interactions explored in the information domain. Physiol Meas
38: 877–894, 2017. doi:10.1088/1361-6579/aa5b77.
21. Joyner MJ, Wallin BG, Charkoudian N. Sex differences and blood
pressure regulation in humans. Exp Physiol 101: 349 –355, 2016. doi:10.
1113/EP085146.
22. Kabir MM, Beig MI, Baumert M, Trombini M, Mastorci F, Sgoifo A,
Walker FR, Day TA, Nalivaiko E. Respiratory pattern in awake rats:
effects of motor activity and of alerting stimuli. Physiol Behav 101: 22–31,
2010. doi:10.1016/j.physbeh.2010.04.004.
AJP-Regul Integr Comp Physiol • doi:10.1152/ajpregu.00430.2017 • www.ajpregu.org
Downloaded from www.physiology.org/journal/ajpregu by ${individualUser.givenNames} ${individualUser.surname} (163.015.154.053) on August 5, 2018.
Copyright © 2018 American Physiological Society. All rights reserved.
R769
23. Koenig J, Thayer JF. Sex differences in healthy human heart rate
variability: A meta-analysis. Neurosci Biobehav Rev 64: 288 –310, 2016.
doi:10.1016/j.neubiorev.2016.03.007.
24. Kuipers NT, Sauder CL, Carter JR, Ray CA. Neurovascular responses
to mental stress in the supine and upright postures. J Appl Physiol (1985)
104: 1129 –1136, 2008. doi:10.1152/japplphysiol.01285.2007.
25. Lackner HK, Papousek I, Batzel JJ, Roessler A, Scharfetter H,
Hinghofer-Szalkay H. Phase synchronization of hemodynamic variables
and respiration during mental challenge. Int J Psychophysiol 79: 401–409,
2011. doi:10.1016/j.ijpsycho.2011.01.001.
26. Laitinen T, Niskanen L, Geelen G, Länsimies E, Hartikainen J. Age
dependency of cardiovascular autonomic responses to head-up tilt in
healthy subjects. J Appl Physiol (1985) 96: 2333–2340, 2004. doi:10.
1152/japplphysiol.00444.2003.
27. Patel K, Rössler A, Lackner HK, Trozic I, Laing C, Lorr D, Green
DA, Hinghofer-Szalkay H, Goswami N. Effect of postural changes on
cardiovascular parameters across gender. Medicine (Baltimore) 95: e4149,
2016. doi:10.1097/MD.0000000000004149.
28. Porta A, Baselli G, Caiani E, Malliani A, Lombardi F, Cerutti S.
Quantifying electrocardiogram RT-RR variability interactions. Med Biol
Eng Comput 36: 27–34, 1998. doi:10.1007/BF02522854.
29. Porta A, Bassani T, Bari V, Pinna GD, Maestri R, Guzzetti S.
Accounting for respiration is necessary to reliably infer Granger causality
from cardiovascular variability series. IEEE Trans Biomed Eng 59: 832–
841, 2012. doi:10.1109/TBME.2011.2180379.
30. Porta A, Marchi A, Bari V, De Maria B, Esler M, Lambert E,
Baumert M. Assessing the strength of cardiac and sympathetic baroreflex
controls via transfer entropy during orthostatic challenge. Philos Trans A
Math Phys Eng Sci 375: 20160290, 2017. doi:10.1098/rsta.2016.0290.
31. Reulecke S, Charleston-Villalobos S, Voss A, González-Camarena R,
González-Hermosillo J, Gaitán-González MJ, Hernández-Pacheco G,
Schroeder R, Aljama-Corrales T. Orthostatic stress causes immediately
increased blood pressure variability in women with vasovagal syncope.
Comput Methods Programs Biomed 127: 185–196, 2016. doi:10.1016/j.
cmpb.2015.12.005.
32. Reyes del Paso GA, Langewitz W, Mulder LJ, van Roon A, Duschek S.
The utility of low frequency heart rate variability as an index of sympathetic
cardiac tone: a review with emphasis on a reanalysis of previous studies.
Psychophysiology 50: 477–487, 2013. doi:10.1111/psyp.12027.
33. Sant’anna ID, de Sousa EB, de Moraes AV, Loures DL, Mesquita ET,
da Nóbrega AC. Cardiac function during mental stress: cholinergic
modulation with pyridostigmine in healthy subjects. Clin Sci (Lond) 105:
161–165, 2003. doi:10.1042/CS20030064.
34. Sarafian D, Miles-Chan JL. The influence of gender and anthropometry
on haemodynamic status at rest and in response to graded incremental
head-up tilt in young, healthy adults. Front Physiol 7: 656, 2017. doi:10.
3389/fphys.2016.00656.
35. Satish P, Muralikrishnan K, Balasubramanian K, Shanmugapriya;
Shanmugapriya. Heart rate variability changes during stroop color and
word test among genders. Indian J Physiol Pharmacol 59: 9 –15, 2015.
36. Shoemaker JK, Hogeman CS, Khan M, Kimmerly DS, Sinoway LI.
Gender affects sympathetic and hemodynamic response to postural stress.
Am J Physiol Heart Circ Physiol 281: H2028 –H2035, 2001. doi:10.1152/
ajpheart.2001.281.5.H2028.
37. Tarvainen MP, Ranta-Aho PO, Karjalainen PA. An advanced detrend-
ing method with application to HRV analysis. IEEE Trans Biomed Eng 49:
172–175, 2002. doi:10.1109/10.979357.
38. Tonhajzerová I, Javorka K, Javorka M, Petrásková M. Cardiovascular
autonomic nervous system tests: reference values in young people (15-19
years) and influence of age and gender. Clin Physiol Funct Imaging 22:
398 –403, 2002. doi:10.1046/j.1475-097X.2002.00449.x.
39. Verheyden B, Beckers F, Couckuyt K, Liu J, Aubert AE. Respiratory
modulation of cardiovascular rhythms before and after short-duration
human spaceflight. Acta Physiol (Oxf) 191: 297–308, 2007. doi:10.1111/
j.1748-1716.2007.01744.x.
40. Wejer D, Graff B, Makowiec D, Budrejko S, Struzik ZR. Complexity
of cardiovascular rhythms during head-up tilt test by entropy of patterns.
Physiol Meas 38: 819 –832, 2017. doi:10.1088/1361-6579/aa64a8.
41. Westerhof N, Lankhaar JW, Westerhof BE. The arterial Windkessel. Med Biol
Eng Comput 47: 131–141, 2009. doi:10.1007/s11517-008-0359-2.
42. Widjaja D, Montalto A, Vlemincx E, Marinazzo D, Van Huffel S, Faes
L. Cardiorespiratory information dynamics during mental arithmetic and
sustained attention. PLoS One 10: e0129112, 2015. doi:10.1371/journal.
pone.0129112.