International Journal of Psychophysiology 101 (2016) 50–58

Contents lists available at ScienceDirect

International Journal of Psychophysiology

journal homepage: www.elsevier.com/locate/ijpsycho

Respiratory hypoalgesia? Breath-holding, but not respiratory phase

modulates nociceptive flexion reflex and pain intensity
Hassan Jafari a,⁎, Karlien Van de Broek a, Léon Plaghki b, Johan W.S. Vlaeyen a,c,
Omer Van den Bergh a, Ilse Van Diest a,⁎⁎
a
University of Leuven, Health Psychology, Leuven, Belgium
b
Institute of Neuroscience (IoNS), Université catholique de Louvain, Brussels, Belgium
c
Department of Clinical Psychological Science, Maastricht University, The Netherlands

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

Article history: Several observations suggest that respiratory phase (inhalation vs. exhalation) and post-inspiratory breath-holds
Received 31 July 2015 could modulate pain and the nociceptive reflex. This experiment aimed to investigate the role of both mecha-
Received in revised form 16 January 2016 nisms. Thirty-two healthy participants received supra-threshold electrocutaneous stimulations to elicit both
Accepted 21 January 2016
the Nociceptive Flexion Reflex (NFR) and pain, either during spontaneous inhalations or exhalations, or during
Available online 22 January 2016
three types of instructed breath-holds: following exhalation, at mid-inhalation and at full-capacity inhalation.
Keywords:
Whether the electrocutaneous stimulus was applied during inhalation or exhalation did not affect the NFR or
Pain pain. Self-reported pain was reduced and the NFR was increased during breath-holding compared to spontane-
Nociceptive flexion reflex ous breathing. Whereas the type of breath-hold did not impact on self-reported pain, breath-holds at full-
Respiration capacity inhalation and following exhalation were associated with a lower NFR amplitude compared to breath-
Breathing holds at mid-inhalation. The present findings confirm that breath-holding can modulate pain (sensitivity) and
Breath-hold suggest that both attentional distraction and changes in vagal activity may underlie the observed effects.
Hypoalgesia © 2016 Elsevier B.V. All rights reserved.

1. Introduction
Breathing techniques involving slow breathing are widely applied as a
key element in many relaxation and meditation exercises, as well as in
strategies to control pain (Bertisch et al., 2009; Brazier et al., 2006;
Busch et al., 2012a; Grant and Rainville, 2009; Kitko, 2007; Mehling
et al., 2005; Miller and Perry, 1990). Both clinical and experimental stud-
ies seem to confirm the potentially analgesic effects of instructed slow
breathing in particular, and various psychological (e.g., expectation, fear
reduction, distraction from pain) and physiological (e.g., baroreceptor
stimulation, vagal activation) factors may contribute to respiration-
induced hypoalgesia.
A series of positive findings stem from clinical studies (Friesner et al.,
2006; Mehling et al., 2005; Park et al., 2013; Yildirim and Sahin, 2004;
but see Downey and Zun, 2009). However, those studies do not allow
for strong conclusions as they often lack necessary control conditions
and do not implement respiratory measures. Recently, experimental
studies have started to investigate the effect of instructed slow breath-
ing on laboratory-induced pain (Arsenault et al., 2013; Busch et al.,
2012a, 2012b; Martin et al., 2012; Zautra et al., 2010; Zunhammer
⁎ Correspondence to: H. Jafari, Health Psychology, University of Leuven (KU Leuven),
Tiensestraat 102, Post Box 3726, 3000 Leuven, Belgium.
⁎⁎ Corresponding author.
E-mail address: Hassan.Jafari@ppw.kuleuven.be (H. Jafari).
et al., 2013). For example, one study found that slow deep breathing in-
creased both thermal pain threshold and tolerance (Chalaye et al.,
2009), suggesting that an increased vagal activity resulting from slow
deep breathing could mediate the analgesic effect. Also, another study
(Zautra et al., 2010) reported on a reduction in self-reported thermal
pain by slow breathing. In a recent study, slow deep breathing was
found to prevent the development of oesophageal pain hypersensitivity.
Vagal blockade with atropine abolished this effect, pointing to a critical
role of the vagus (Botha et al., 2014). In yet another study, slow breath-
ing did not change the Nociceptive Flexion Reflex (NFR), but did reduce
self-reported pain, decreased heart rate, and increased heart rate vari-
ability (HRV) (Martin et al., 2012). Interestingly, the changes in HRV
in the latter study were not correlated with changes in any of the pain
outcomes, which made the authors conclude that efferent cardiac
vagal outflow could not explain the pain-reducing effect of slow deep
breathing. Thus, also experimental studies have documented an effect
of slow deep breathing on pain, but it is still unclear which mechanisms
critically contribute to such effect.
A few experimental studies have also investigated whether the NFR
or pain ratings differ according to whether the painful stimulus was pre-
sented during the inspiratory or expiratory phase of the respiratory
cycle. As vagal outflow to the heart is thought to be higher during expi-
ration compared to inspiration (Appelhans and Luecken, 2006; Eckberg,
2003), potentially anti-nociceptive effects of vagal efferent outflow
could produce a reduced pain (sensitivity) during the expiratory

http://dx.doi.org/10.1016/j.ijpsycho.2016.01.005
0167-8760/© 2016 Elsevier B.V. All rights reserved.
 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58
phase of the respiratory cycle. As slower breathing is typically accom-
plished by lengthening the expiratory phase (thus decreasing the inhala-
tion/exhalation ratio), such respiratory gating of vagal outflow could be
involved in the anti-nociceptive effects of slow breathing. However, the
literature does not report consistent findings of respiratory phase on
pain, as one study has reported on a zero-finding (Martin et al., 2012), an-
other on a positive finding (Iwabe et al., 2014), and still another on an ef-
fect in the opposite direction (Arsenault et al., 2013). Interestingly, these
studies have looked at the effect of respiratory phase on pain during
instructed (paced) breathing. It remains thus unknown whether respira-
tory phase influences pain (sensitivity) during spontaneous breathing.
Until now, the main focus of most intervention studies with breath-
ing manipulations was on breathing frequency (instructions to breathe
slower). Studies typically do not manipulate or measure changes in
breathing depth (volume) or respiratory pauses, although an increase
in both components is known to accompany voluntary attempts to
slow down respiratory rate. Several studies documented an increase
in inspiratory flow and volume as a reaction to experimental pain
(Duranti et al., 1991; Hotta et al., 2009; Hotta et al., 2006; Kato et al.,
2001; Sarton et al., 1997). Phenomenologically, acute pain typically trig-
gers a reaction similar to the respiratory component of the startle reflex,
that is, an inspiratory gasp followed by a post-inspiratory breath-hold
(Van Diest et al., 2005). In addition, intolerable pain was accompanied
by repeated breath-holds (Tanii et al., 1973) and cold pain increased in-
spiratory pause duration (breath-hold) (Boiten, 1998). It is conceivable
that these respiratory responses to pain are functional in pain reduction.
Moreover, breath-holding could stimulate the anti-nociceptive effects of
baroreceptor stimulation (Dworkin et al., 1994; Dworkin et al., 1979)
and concomitant increases in vagal activation (Bruehl and Chung, 2004;
Triedman and Saul, 1994). Consistent with this, the Valsalva Manoeuvre
(VM, a forceful attempted exhalation against a closed airway) has been
found to decrease acute pain (Agarwal et al., 2005).
The aims of this study were three-fold: (1) to investigate whether
pain and nociception differ between spontaneous breathing versus
breath-holding, (2) to study whether post-inspiratory breath-holding
is instrumental in reducing pain, and (3) to explore whether pain and
nociception differ between the inspiratory and expiratory phase during
spontaneous breathing.
2. Methods
2.1. Participants
Fifty participants were recruited either through the Experiment
Management System website of the Faculty of Psychology and Educa-
tional Sciences at the University of Leuven, or through advertisements
and flyers. Prior to the experiment, people who expressed interest in
participating received an email with information on the experiment
and the exclusion criteria. The information was enclosed in a document
explaining that the experiment aimed at investigating the influence of
respiration on sensitivity to pain. It was clearly stated that supra-
threshold electrical stimulations would be applied during the experi-
ment, and that participants were requested to refrain from medication
during 24 h prior to the experiment, abstain from caffeine, nicotine
and alcohol, and to avoid exertion during 4 h prior to the experiment.
Finally, the following exclusion criteria were specified: cardiovascular
or blood circulation disorders, respiratory disorders, neurological disor-
ders, severe acute pain, pacemaker or any other electronic medical im-
plant, injury or trauma to lower extremities (hip, thigh, knee, ankle
and foot), hearing and visual impairment, psychiatric disorders, recent
psychological or stressful trauma, regular medication intake (except
contraceptives), pregnancy and, finally, body mass index over 35. Each
participant was reimbursed with either 25 Euro or two course credits
(only for student participants) depending on their preference. All par-
ticipants provided their informed consent. The experiment was ap-
proved by the Medical Ethics Committee of the University of Leuven
51
and conducted according to the guidelines laid down in the Declaration
of Helsinki. Of the 50 participants, five were excluded because they met
at least one exclusion criterion. Among the people that were invited, six
did not show up without providing any specific reason. From the re-
maining 39 healthy people, seven were excluded because no proper
NFR could be obtained with stimulus intensities remaining below the
participant's pain tolerance threshold. Thirty-two participants (22 fe-
males, 10 males), aged between 18 and 30 (M = 20.7, SD = 2.5), com-
pleted the experiment.
2.2. Instruments and measurements
The experiment was programmed using Affect 4.0 software (Spruyt
et al., 2010), including psychophysiological recordings (respiration and
EMG), pain ratings and stimuli presentations (breath holding task,
electrocutaneous stimulation).
2.2.1. Respiratory recording and breathing–holding task
A pneumograph chest belt (respiratory belt Philips and Bird Compa-
ny, US) was used to record respiratory activity. This device is sensitive to
air pressure changes inside a tube caused by breathing-related expan-
sions of the chest. Because our main interest was to investigate changes
in respiration with respect to a within-subject manipulation, no calibra-
tion procedure to transform the recorded signal into absolute volumes
was performed. The belt was fixed around the subjects' upper abdomen
adjacent to the lower thoracic rib region, and DC-coupled to a differen-
tial aneroid pressure transducer (Coulbourn V72-25B, Coulbourn In-
struments, Allentown, Pennsylvania). The signal was sampled and
stored at 1000 Hz.
The breath-holding task comprised three types of instructed breath-
holds, corresponding to voluntary breath-holds of 4 s at three different
levels of Maximum Inspiratory Thoracic Expansion (MITE): at 50% of
MITE (mid-inhalation breath-hold), at 80% of MITE (full-capacity inha-
lation breath-hold) and post-expiration (exhalation breath-hold). Par-
ticipants' MITE was assessed prior to the experimental procedure. To
this end, participants performed three maximal inhalations. The peak
value of the inhalation with the greatest amplitude served as an approx-
imation of the participant's maximal inspiratory capacity. The ampli-
tudes representing 50% and 80% of the participant's MITE were
calculated accordingly. During the instructed breath-holding task, the
amplitudes representing exhalation, 50% and 80% of MITE were
displayed with horizontal lines on a monitor in front of the participant.
Also, the respiratory signal (pneumographic chest belt) was displayed
on the monitor, providing the participant with continuous feedback of
his or her ongoing respiratory activity. Participants were instructed to
keep the respiratory signal by means of a breath-hold at a specific hor-
izontal target line for 4 s whenever such instruction appeared on the
monitor. The task comprised 10 mid-inhalation, 10 full-inhalation and
10 exhalation breath-holds instructed in a random order with 15 to
25 millisecond time interval.
2.2.2. Electrocutaneous stimulation and pain rating
To elicit pain and the NFR, a constant current stimulator (Digitimer
DS5, U.K.) generated electrocutaneous stimuli. Each stimulation
consisted of a volley of ten 1 ms rectangular pulses with 1 ms inter-
pulse interval (total duration = 20 ms). A bar shaped bipolar stimulat-
ing electrode with two round electrodes (8 mm diameter, 30 mm inter-
electrode distance) was fixed well with a Velcro strap over the
retromalleolar pathway of the sural nerve of the left leg.
Electrocutaneous stimuli were triggered manually by the experimenter.
A computerized online Numerical Rating Scale (NRS) (McMahon
et al., 2013) was used for the pain ratings, which ranged from 0 (no
pain) to 100 (worst possible pain). Pain tolerance was defined as a rat-
ing of 90 on this NRS, and was determined for each participant by ad-
ministering stimuli with incremental steps of 2 mA up until a rating of
90 on the NRS was reached.
52 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58
2.2.3. Nociceptive flexion reflex
The NFR is a polysynaptic spinal withdrawal reflex that is elicited fol-
lowing the activation of nociceptive A-delta afferent. It is considered a
measure of sensitivity to pain at the spinal level that can be modulated
by higher brain structures (Miller et al., 1979; Rhudy et al., 2005). Based
on the observed EMG response, the intensity of stimulation required to
elicit the NFR is used as an objective index of nociceptive threshold, and
the electromyography response amplitude measures the NFR (Rhudy
and France, 2007; Sandrini et al., 2005; Skljarevski and Ramadan,
2002). For the NFR recording, two 8 mm Ag/AgCl electrodes (2 cm cen-
ter to center distance) were placed over the left biceps femoris muscle,
10 cm above the popliteal fossa, with a reference electrode attached
over the antero-lateral aspect of the tibial bone plateau (see Fig. 1). To
avoid recording of motion artefact and noise, the disposable recording
electrodes were integrated in an adhesive cloth that stuck firmly to
the skin, preventing their displacement. To elicit the NFR,
electrocutaneous stimulation was applied over the sural nerve at the
back of the lateral malleolus of the same side leg. NFR was defined as bi-
ceps femoris electromyographic (EMG) activity in the 85–180 millisec-
ond period after stimulus.
The NFR threshold (NFRT) was conceptualized as the stimulation re-
quired us to elicit a stable series of responses. The NFRT was determined
for each participant using an up–down staircase method (4–2–1–
0.5 mA staircase procedure). The stimulation level was increased with
4 mA once the NFR could be observed. Then, the stimulus intensity
was decreased in steps of 2 mA until the NFR was no longer detected,
and increased and decreased again for two more times with steps of
1 mA and 0.5 mA, respectively. The average of the two last ascending
stimulations capable of eliciting the NFR was taken as NFRT. Participants
rated the pain generated by each stimulus in this procedure using the
digital NRS. For safety purposes, when no NFR could be observed with
intensities that were rated lower than 90 on the NRS and/or lower
than 35 mA, the experiment was discontinued. The intensity of the
electrocutaneous stimulation delivered for the remainder of the study
for eliciting the NFR was set to 120% of each participant's NFRT. This pro-
cedure of NFR measurement and testing is frequently used in recent
studies and the standards are well established (Rhudy et al., 2007;
Fig. 1. Spinal nociceptive flexion reflex (NFR) electrode placement. The recording electrodes were placed over the left biceps femoris muscle 10 cm above the popliteal fossa, with a
reference electrode attached over the antero-lateral aspect of tibial bone plateau. To elicit NFR activity, repeated electro-cutaneous stimulation was applied over the sural nerve at the
back of the lateral malleolus of the same side leg.
Rhudy et al., 2005; Rhudy and France, 2007; Sandrini et al., 2005; Sarton
et al., 1997; Skljarevski and Ramadan, 2002).
The EMG signal was recorded at 1000 Hz, using an Isolated
Bioamplifier Model V75-04 (Coulbourn Instruments, Allentown, Penn-
sylvania) with a band pass filter of 8–1000 Hz. EMG recordings started
200 ms (baseline) prior until 800 ms following the electrical stimulus.
2.3. Experimental design and procedure
The experiment took place at the Health Psychology Psychophysiol-
ogy Laboratories of the University of Leuven. Participants were seated in
an adjustable dentist's chair in a semi-reclined position with a knee flex-
ion of approximately 140 degrees and a neutral position of the ankle.
The participants read and signed the informed consent form. Following
this, participants were fitted with the electrodes and the electrodes'
wires were taped to the skin. Participants were asked to lay relaxed
and to avoid any unnecessary voluntary movement or contraction of
the muscles. Next, the participant's pain tolerance and NFR threshold
were determined (see Sections 2.2.2 and 2.2.3). Following this, partici-
pants were instructed how to perform the breath-holding task and
practiced it for 2 min.
The actual experiment consisted of a spontaneous breathing and a
breath-holding block for each participant (see Fig. 2). Participants
were counterbalanced across the two possible order sequences. Partici-
pants rated their pain level on the NRS scale that appeared on the mon-
itor following each pain stimulus. The NFR and pain ratings were
recorded with respect to each of the electrical stimuli.
During the breath-holding task, the experimenter monitored the
participants' performance on the breath-holding tasks and adminis-
tered 18 electrical stimuli. Six stimuli for each type of breath-hold
were administered. Thus, for each type of breath-hold, four randomly
chosen breath-holds were without any pain stimulation (see Fig. 3).
During spontaneous breathing, participants received no breathing
instructions and could not see their own respiratory signal on the mon-
itor in the participant's room. In the experimenter room, the experi-
menter monitored the participants' spontaneous breathing pattern
 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58 53

Fig. 2. The overall design of study comprised two breathing tasks of spontaneous breathing and breath holding. Each task consisted of different phases.

and triggered the electrical stimulation manually. Every participant re-
ceived 20 electrical stimuli with a 15–25 s interstimulus time interval.
During breath-holding task the experimenter monitored the partic-
ipants' performance on the breath-holding tasks and administered the
electrical stimuli at the targeted breath-holds. The NFR and pain ratings
were recorded with respect to each of the electrical stimuli.
2.4. Data reduction and analysis
Respiratory and EMG recordings were synchronized using
AcqKnowledge software (V4.2, BIOPAC Systems Inc., USA ), allowing
to check the exact location of the electrical stimulation within the respi-
ratory cycle. The raw EMG signal rectified by a root-mean-square trans-
formation per 5 ms bin. The integral of the rectified EMG was calculated
for the 85–180 ms time interval following the onset of the
electrocutaneous stimulus, as a measure of NFR amplitude (see Fig. 4).
For the spontaneous breathing task, the on- and offset of each respi-
ratory cycle were visually determined using AcqKnowledge software.
The peak within each respiratory cycle was determined automatically
by the software as the time of the highest value of respiration amplitude
between onset and end times. To determine whether an
electrocutaneous stimulus occurred during in- or exhalation, and
when it occurred relative to inspiratory peak, a stimulus-to-peak time
ratio (SP time ratio) was calculated for each electrical stimulus. For ex-
ample, if the inspiratory peak (indicating the switch from inspiration
to expiration) occurred at 2 s, and the electrical stimulus was delivered
at 1.7 s, then the SP ratio is 1.7/2 = 0.85. Thus, SP time ratios smaller
than 1 indicate that the electrical stimulus occurred during inspiration,

Fig. 3. A sample of experimental diagram for 30 breath-holds and 18 electro-cutaneous stimuli in one participant. A computerized analogue visual biofeedback of the respiratory signal
allowed the participants to perform the instructed breath-holds. The amplitudes representing full-inhalation, mid-inhalation and exhalation breath holding were displayed with
horizontal lines on a monitor. The programme instructed the participants to perform in random order 10 breath-holds at mid-inhalation, full-inhalation and exhalation.
54 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58

Fig. 4. The scatterplot of electro-cutaneous stimuli relative to the inspiratory peak time during spontaneous breathing for 32 participants. The respiration curve is only a symbolic cycle for a
better representation. Each dot represents one stimulation. Stimulations which occurred at a SP-ratio smaller than 1 occurred during inspiration, those at a SP-ratio bigger than 1 are
occurred during expiration.

whereas SP time ratios bigger than 1 indicate that the stimulus occurred baseline EMG activity during the 150 ms interval prior to the
during expiration (see Fig. 5). electrocutaneous stimulation differed among the three types of
Because instructed breath-holds could impact on the leg's muscle instructed breath-holds (exhalation, mid inhalation and full inhalation)
tone, particularly during the full-inspiration task, we checked whether (see Fig. 4). To this end, and as the assumption of sphericity (Mauchly's

Fig. 5. An illustration of the time windows for electromyography processing for (A) Baseline activity prior to electro-cutaneous stimuli, and (B) NFR 85 ms after electro-cutaneous stimuli.
This graph represents six random trials of NFR after root-mean-squared processing for one of the subjects.
 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58 55

Table 1
Analysis of the estimates, standard error, deviance and significant level of the three models of breathing task, breath-holding and spontaneous breathing (SB = spontaneous breathing; BH
= breath-hold).

Null model versus alternative model Condition Response

Pain NFR

Coefficient/mean Deviance P Coefficient/mean Deviance P

Breathing task
Null model 38.7 (3.15) 9279.18 b.001 5.9 (0.46) 6171.8 .017
Task SB 43.3 (3.21) 9121.43 5.6 (0.48) 6166.2
BH −9.6 (.73) 0.5 (0.21)

Spontaneous breathing (inspiration vs. expiration)

Null model 41.9 (3.57) 4491.6 .34 5.9 (.44) 3131.8 .45
Resp. Phase Inspiration 42.3 (3.60) 4490.7 6.01 (.46) 3131.3
Expiration −0.76 (.83) −0.21 (.28)

Breath-holding
Null model 33.3 (3.41) 4324.29 .57 6.15 (.58) 2979.2 .010
Depth of BH Exhalation 33.9 (3.48) 4323.17 6.01 (.63) 2970.1
Mid-inhalation −0.52 (1.18) +0.71 (.36)
Full-inhalation −1.27 (1.21) −0.33 (.37)

The coefficients indicate the grand mean for the null model and mean for first condition in alternative model, and the difference from the mean for the second or third condition. Parameter
estimate standard errors listed in parentheses. SB: spontaneous breathing, BH: breath-holding, Resp: respiration.

test) was not violated (p N .05), a general Linear Model (GLM) repeated
measure ANOVA with an alpha threshold of .05 was used for the data
analysis of the baseline EMG recording (SPSS Statistics version 21).
MLwiN software version 2.30 was used for the multilevel modelling
of subjective pain ratings and the NFR. The likelihood ratio test was used
for parameter estimation, and for the testing of two competing models
(null model versus alternative model; Rasbash et al., 2005). The longitu-
dinal multilevel model was designed such that the repeated measure-
ments of painful stimuli (level 1) were nested within participants
(level 2). To allow for the comparison of the models' coefficients, we
used a full information maximum likelihood approach. To compare
the relative fit of two competing models, the deviance (−2*Log-likeli-
hood) for each model was computed and tested using one-sided Chi
square test with a threshold of .05. The multilevel modelling was per-
formed in three distinct phases, testing the effects of (1) an instructed
breath-holding task, (2) different types of breath-holding, and (3) respi-
ratory phase during spontaneous breathing. For the effect of the
instructed breath-holding task, we compared the null model (uncondi-
tional model) including all observations within all participants with
the task model (alternative model). The alternative model included
‘breathing task’ as a variable, dummy coded for spontaneous breathing
versus instructed breath-holding, and centred on spontaneous breath-
ing. To test the effect of the type of breath-holding, the null model com-
prising all data of the breath holding task within all participants was
compared with a depth model (alternative model) that included three
categories of breath-holding: exhalation, mid-inhalation and full- inha-
lation, with the first category as a reference. Finally, to test the effect of
respiratory phase during spontaneous breathing, we compared the null
model including all observations during spontaneous breathing with a
phase model (alternative model) that also included a phase variable,
which was dummy coded into two phases of inspiration and expiration,
centred on inspiration. For each of models the unstandardized coeffi-
cient, deviance and Chi-square probability level were reported. The un-
standardized coefficients can be interpreted as the grand mean in the
null model, the mean in centred category of alternative model, and
the mean difference from centred category in other condition(s).
3. Results
3.1. Baseline EMG activity
The muscle's baseline activity prior to the presentation of the
electrocutaneous stimulus did not differ between the three types of
breath-holding (main effect: p = .932).
3.2. Instructed breath-holding task
Performing instructed breath-holds significantly altered self-
reported pain (Chi21 = 157.75, p b .0001) and NFR (Chi21 = 5.6, p =
.17) compared to the null model (unconditional intercept). In this
model, self-reported pain was lower (9.6 points on the NRS scale) dur-
ing breath-holding compared to spontaneous breathing. The NFR in-
creased 0.5 mV·s from spontaneous breathing to breath-holding, but
this effect was not significant (see Table 1).
3.3. Spontaneous breathing
The phase model of spontaneous breathing did not improve the pre-
diction of subjective pain levels (Chi21 = .9, p = .34) or the NFR
(Chi21 = .5, p = .45) above the null model. The decreases in pain
(B = −.76 points on the NRS) and the NFR (B = −.21 mV·s) during ex-
piration relative to inspiration were not significant (see Table 1).
3.4. Breath-holding model
The breath-holding model did not add to the null model for self-
reported pain (Chi22 = 1.12, p = 0.57), whereas it did for the NFR
(Chi22 = 9.1, p = .01). Pairwise comparisons of the three different
types of breath-holds showed a significant decrease in the NFR ampli-
tude after full-inhalation (B = − 1.04 mV·s, Chi21 = 8.72, p = .003)
and exhalation (B = −0.71 mV·s, Chi21 = 3.86, p = .049) compared
to mid-inhalation.
4. Discussion
The objectives of the present study were three-fold: To explore
whether self-reported pain and nociception as measured by the NFR
(1) vary between spontaneous breathing versus instructed breath-
holding, (2) are reduced by post-inspiratory breath-holds, and (3) differ
between the inspiratory versus expiratory phase during spontaneous
breathing.
Compared to non-manipulated, spontaneous breathing, the breath-
ing manipulation applied in the present study (breath-holding) signifi-
cantly reduced self-reported pain, whereas it increased the NFR. The
decrease in self-reported pain can likely be understood from a higher at-
tentional involvement and predictability in the breath-holding com-
pared to the spontaneous breathing task. As participants had to
comply with the instructed depths and timings of the breath-holds, it
56 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58
is plausible to assume that the task required substantial attentional re-
sources. Several studies have shown that distraction results in lower
pain ratings (Roy et al., 2011; Van Damme et al., 2008; Verhoeven
et al., 2011). More generally, any guided breathing task is likely capable
of drawing attention away from the subjective pain experience and
dampening the pain experience, especially when the instructed breath-
ing pattern is still relatively untrained and attentional demands are
high. Besides having been distracted, participants may also have experi-
enced the painful stimulus as relatively more predictable during the
breath-hold compared to the spontaneous breathing task, resulting in
overall lower levels of anxiety during instructed compared to spontane-
ous breathing (Carlsson et al., 2006). Both phenomena could produce
opposite effects for nociception as studies have shown that the nocicep-
tive withdrawal reflex is higher when participants concentrate on the
stimuli, and prediction or distraction usually do not decrease the reflex
(Bjerre et al., 2011). Thus, the present findings highlight a potentially
important influence of attention and distraction on pain in any
instructed breathing task, particularly in studies that aim to investigate
mechanisms underlying respiratory hypoalgesia.
An alternative explanation for an increased NFR during instructed
breath-holds relates to a potential subthreshold facilitation of alpha mo-
toneurons. For example, contractions of upper limb muscles are known
to facilitate reflexes elicited in lower limb muscles without changes in
baseline EMG activity in these muscles (Delwaide and Toulouse,
1981). Likewise, muscle contractions associated with instructed, volun-
tary breath-holds may facilitate the NFR (Bishop et al., 1970;
Schmidt-Vanderheyden and Koepchen, 1970; Schmidt-Vanderheyden
et al., 1970).
As pain has been described to concurrently increase tidal volume,
mean inspiratory flow and the occurrence of spontaneous breath-
holds (Boiten, 1998), it is plausible that breath-holding after a deep in-
halation could be functional in reducing the impact of painful stimula-
tion. We hypothesised that breath-holds at different levels of MITE
(full-inhalation vs. mid-inhalation vs. exhalation) would impact on
self-reported pain and the nociceptive reflex. More specifically, we ex-
pected a decrease in the nociceptive reflex when the breath-hold oc-
curred at a high level of inspiratory capacity compared to lower levels.
This hypothesis was only partially confirmed, as full-inhalation
breath-holds were associated with a reduced nociceptive reflex com-
pared to a breath-hold at 50% of MITE, but not compared to breath-
holds following exhalation. Potentially, participants may have generat-
ed an expiratory pressure against a closed glottis during the full-
inhalation breath-hold, very much like the Valsalva Manoeuvre. This
would yield a sudden and high intrathoracic pressure following the
end of inspiration, resulting in a higher stroke volume and blood pres-
sure (Ghione, 1996; Novak, 2011; Vogel et al., 2005), and in concomi-
tant baroreceptor stimulation producing antinociceptive effects
(Edwards et al., 2002; Edwards et al., 2001). To further explore this pos-
sible mechanism, future studies could consider investigating whether
participants close their glottis during an instructed breath-hold at full-
inspiration as applied in the present study. Post-expiratory pauses
may have anti-nociceptive effects due to higher activation of barorecep-
tors. This is conceivable as blood pressure is typically maximal towards
the end of expiration (Sin et al., 2010). To test these speculations, fur-
ther studies are needed that apply beat-to-beat variations in blood
pressure.
In contrast to the above reported effects on the NFR, the different
types of breath-holds did not produce differential effects in self-
reported pain. Possibly, the immediate nociceptive effects of breath-
holds as reflected in the NFR are too small to result in a variable pain
perception, especially at a time that participants are being distracted
from pain by the breath-holding task. As such, applying longer-lasting
and more variable pain stimuli may facilitate observing effects on self-
reported pain.
To the best of our knowledge, the present study is the first to explore
the association between respiratory phase and pain during spontaneous
breathing. Thus, unlike other studies that have looked at the effect of re-
spiratory phase on pain during paced breathing, the present study did
not apply any kind of breathing manipulation or feedback that could
generate a distraction effect, and/or mask or override central mecha-
nisms of respiratory gating (Eckberg, 2003). We hypothesised that
pain and nociception would be lower during spontaneous exhalations
compared to inhalations. As blood pressure is higher during expiration
compared to inspiration, baroreceptors are stimulated more strongly
during expiration, which in turn could activate descending pain inhibi-
tory mechanisms (Berntson et al., 1993; Bruehl and Chung, 2004;
Dworkin et al., 1979, 1994; Triedman and Saul, 1994). Although differ-
ences between inhalation and exhalation were in the hypothesised di-
rection, our findings indicate that pain and nociception were not
significantly reduced during spontaneous exhalations compared to in-
halations in the present study. Thus, our findings seem to suggest that
either there is no pain reduction during exhalation, or that spontaneous
respiration in the natural range is a too weak stimulus for changes in au-
tonomic discharge patterns that could modulate nociception and pain.
Alternatively, our null-finding may result from the followed procedure.
In the present study, we manually applied only 20 electrocutaneous
stimuli to test the effect of respiratory phase. Post-hoc processing of
data indicated that the pain stimuli were not distributed in a homoge-
nous way across the respiratory cycle, and occurred mainly during late
inspiration and early and mid-expiration (see Fig. 4). Such unequal dis-
tribution may have limited the power to detect a respiratory phase ef-
fect in the present study. Therefore, we recommend future studies to
investigate respiratory phase effects with a higher number of stimuli
that are randomly distributed across the respiratory cycle curve. In ad-
dition, concurrently measuring cardiac inter-beat intervals and the
beat-to-beat blood pressure waveform would allow to test whether
lower pain and nociception occur specifically at times of maximal
blood pressure and longest inter beat intervals.
4.1. Limitations
Some limitations of the present study should be acknowledged. First,
the findings may not be generalized to other types of pain and other
populations. Second, as the low pass filter applied on the EMG recording
equalized the sampling frequency of the signal (1000 Hz), a potential
risk of aliasing is present in the EMG recordings. Finally, as the present
study did not include beat-to-beat blood pressure, and intrathoracic
pressure recordings, any interpretation in terms of antinociceptive ef-
fects of baroreceptor activation remains speculative at this stage.
5. Conclusion
Both self-reported pain and the NFR are affected by an instructed
breath-holding task. Compared to spontaneous breathing, instructed
breath-holds were associated with lower pain ratings, but with a rela-
tively potentiated NFR. Both effects could be due to distraction from
pain and and/or lower anxiety during breath-holding task.
Our findings further confirm that the depth of breath-holding affects
spinal nociceptive processes: NFR activity was significantly less during
breath-holds at full-range inhalation and following exhalation, com-
pared to breath-holds at mid-range inhalation. However, further re-
search should investigate the different processes that may underlie
these effects.
Pain ratings and the NFR were not affected by respiratory phase dur-
ing spontaneous breathing in the present study. Also, pain ratings were
not affected by different types of breath-holding manoeuvres.
Conflict of interest statement
None of the authors have any conflicts of interests related to this
manuscript.
 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58
Acknowledgements
This study was supported by the Odysseus Grant “The Psychology of
Pain and Disability Research Programme” and projects G.0543.09N and
G.0715.10N funded by the Research Foundation — Flanders, Belgium
(F.W.O. Vlaanderen).
References
Agarwal, A., Sinha, P.K., Tandon, M., Dhiraaj, S., Singh, U., 2005. Evaluating the efficacy of
the valsalva maneuver on venous cannulation pain: a prospective, randomized study.
Anesth. Analg. 101, 1230–1232. http://dx.doi.org/10.1213/01.ane.0000167270.
15047.49 table of contents.
Appelhans, B.M.B.M., Luecken, L.J.L.J.L.J., 2006. Heart rate variability as an index of regulat-
ed emotional responding. Rev. Gen. Psychol. 10, 229–240. http://dx.doi.org/10.1037/
1089-2680.10.3.229.
Arsenault, M., Ladouceur, A., Lehmann, A., Rainville, P., Piché, M., 2013. Pain modulation
induced by respiration: phase and frequency effects. Neuroscience 252, 1–11.
http://dx.doi.org/10.1016/j.neuroscience.2013.07.048.
Berntson, G.G., Cacioppo, J.T., Quigley, K.S., 1993. Respiratory sinus arrhythmia: autonom-
ic origins, physiological mechanisms, and psychophysiological implications. Psycho-
physiology 30, 183–196. http://dx.doi.org/10.1111/j.1469-8986.1993.tb01731.x.
Bertisch, S.M., Wee, C.C., Phillips, R.S., McCarthy, E.P., 2009. Alternative mind–body thera-
pies used by adults with medical conditions. J. Psychosom. Res. 66, 511–519. http://
dx.doi.org/10.1016/j.jpsychores.2008.12.003.
Bishop, B., Johnston, R.M., Walsh, W., 1970. Changes in alpha and gamma motoneuron ex-
citability with respiration. Arch. Phys. Med. Rehabil. 51, 383–390.
Bjerre, L., Andersen, A.T., Hagelskjær, M.T., Ge, N., Mørch, C.D., Andersen, O.K., 2011. Dy-
namic tuning of human withdrawal reflex receptive fields during cognitive attention
and distraction tasks. Eur. J. Pain 15, 816–821. http://dx.doi.org/10.1016/j.ejpain.
2011.01.015.
Boiten, F.A., 1998. The effects of emotional behaviour on components of the respiratory
cycle. Biol. Psychol. 49, 29–51. http://dx.doi.org/10.1016/S0301-0511(98)00025-8.
Botha, C., Farmer, A.D., Nilsson, M., Brock, C., Gavrila, A.D., Drewes, A.M., Knowles, C.H.,
Aziz, Q., 2014. Preliminary report: modulation of parasympathetic nervous system
tone influences oesophageal pain hypersensitivity. Gut 1–7. http://dx.doi.org/10.
1136/gutjnl-2013-306698.
Brazier, A., Mulkins, A., Verhoef, M., 2006. Evaluating a yogic breathing and meditation in-
tervention for individuals living with HIV/AIDS. Am. J. Health Promot. 20, 192–195.
Bruehl, S., Chung, O.Y., 2004. Interactions between the cardiovascular and pain regulatory
systems: an updated review of mechanisms and possible alterations in chronic pain.
Neurosci. Biobehav. Rev. 28, 395–414. http://dx.doi.org/10.1016/j.neubiorev.2004.06.
004.
Busch, V., Magerl, W., Kern, U., Haas, J., Hajak, G., Eichhammer, P., 2012a. The effect of
deep and slow breathing on pain perception, autonomic activity, and mood process-
ing — an experimental study. Pain Med. 13, 215–228. http://dx.doi.org/10.1111/j.
1526-4637.2011.01243.x.
Busch, V., Zeman, F., Heckel, A., Menne, F., Ellrich, J., Eichhammer, P., 2012b. The effect of
transcutaneous vagus nerve stimulation on pain perception — an experimental study.
Brain Stimul. 1–8. http://dx.doi.org/10.1016/j.brs.2012.04.006.
Carlsson, K., Andersson, J., Petrovic, P., Petersson, K.M., Ohman, A., Ingvar, M., 2006. Pre-
dictability modulates the affective and sensory-discriminative neural processing of
pain. NeuroImage 32, 1804–1814. http://dx.doi.org/10.1016/j.neuroimage.2006.05.
027.
Chalaye, P., Goffaux, P., Lafrenaye, S., Marchand, S., 2009. Respiratory effects on experi-
mental heat pain and cardiac activity. Pain Med. 10, 1334–1340. http://dx.doi.org/
10.1111/j.1526-4637.2009.00681.x.
Delwaide, P.J., Toulouse, P., 1981. Facilitation of monosynaptic reflexes by voluntary con-
traction of muscle in remote parts of the body. Mechanisms involved in the
Jendrassik Manoeuvre. Brain 104, 701–709.
Downey, L.V.A., Zun, L.S., 2009. The effects of deep breathing training on pain manage-
ment in the emergency department. South. Med. J. 102, 688–692. http://dx.doi.org/
10.1097/SMJ.0b013e3181a93fc5.
Duranti, R., Pantaleo, T., Bellini, F., Bongianni, F., Scano, G., 1991. Respiratory responses in-
duced by the activation of somatic nociceptive afferents in humans. J. Appl. Physiol.
71, 2440–2448.
Dworkin, B.R., Filewich, R.J., Miller, N.E., Craigmyle, N., Pickering, T.G., 1979. Baroreceptor
activation reduces reactivity to noxious stimulation: implications for hypertension.
Science 205, 1299–1301.
Dworkin, B.R., Elbert, T., Rau, H., Birbaumer, N., Pauli, P., Droste, C., Brunia, C.H., 1994. Cen-
tral effects of baroreceptor activation in humans: attenuation of skeletal reflexes and
pain perception. Proc. Natl. Acad. Sci. U. S. A. 91, 6329–6333.
Eckberg, D.L., 2003. The human respiratory gate. J. Physiol. 548, 339–352. http://dx.doi.
org/10.1113/jphysiol.2002.037192.
Edwards, L., Ring, C., McIntyre, D., Carroll, D., 2001. Modulation of the human nociceptive
flexion reflex across the cardiac cycle. Psychophysiology 38, 712–718.
Edwards, L., McIntyre, D., Carroll, D., Ring, C., Martin, U., 2002. The human nociceptive
flexion reflex threshold is higher during systole than diastole. Psychophysiology 39,
678–681 doi:10.1017.S0048577202011770.
Friesner, S.A., Curry, D.M., Moddeman, G.R., 2006. Comparison of two pain-management
strategies during chest tube removal: relaxation exercise with opioids and opioids
alone. Heart Lung 35, 269–276. http://dx.doi.org/10.1016/j.hrtlng.2005.10.005.
57
Ghione, S., 1996. Hypertension-associated hypalgesia. Evidence in experimental animals
and humans, pathophysiological mechanisms, and potential clinical consequences.
Hypertension 28, 494–504.
Grant, J.A., Rainville, P., 2009. Pain sensitivity and analgesic effects of mindful states in Zen
meditators: a cross-sectional study. Psychosom. Med. 71, 106–114. http://dx.doi.org/
10.1097/PSY.0b013e31818f52ee.
Hotta, N., Sato, K., Sun, Z., Katayama, K., Akima, H., Kondo, T., Ishida, K., 2006. Ventilatory
and circulatory responses at the onset of exercise after eccentric exercise. Eur. J. Appl.
Physiol. 97, 598–606. http://dx.doi.org/10.1007/s00421-006-0212-y.
Hotta, N., Yamamoto, K., Katayama, K., Ishida, K., 2009. The respiratory response to pas-
sive and active arm movements is enhanced in delayed onset muscle soreness. Eur.
J. Appl. Physiol. 105, 483–491. http://dx.doi.org/10.1007/s00421-008-0926-0.
Iwabe, T., Ozaki, I., Hashizume, A., 2014. The respiratory cycle modulates brain potentials,
sympathetic activity, and subjective pain sensation induced by noxious stimulation.
Neurosci. Res. 1–13. http://dx.doi.org/10.1016/j.neures.2014.03.003.
Kato, Y., Kowalski, C.J., Stohler, C.S., 2001. Habituation of the early pain-specific respirato-
ry response in sustained pain. Pain 91, 57–63. http://dx.doi.org/10.1016/S0304-
3959(00)00419-X.
Kitko, J., 2007. Rhythmic breathing as a nursing intervention. Holist. Nurs. Pract. 21,
85–88. http://dx.doi.org/10.1097/01.HNP.0000262023.27572.65.
Martin, S.L., Kerr, K.L., Bartley, E.J., Kuhn, B.L., Palit, S., Terry, E.L., Delventura, J.L., Rhudy,
J.L., 2012. Respiration-induced hypoalgesia: exploration of potential mechanisms.
J. Pain 13, 755–763. http://dx.doi.org/10.1016/j.jpain.2012.05.001.
McMahon, S., Koltzenburg, M., Tracey, I., Turk, D., 2013. Wall & Melzack's Textbook of
Pain: Expert Consult, 6. edition. ed. Saunders, Philadelphia.
Mehling, W.E., Hamel, K.A., Acree, M., Byl, N., Hecht, F.M., 2005. Randomized, controlled
trial of breath therapy for patients with chronic low-back pain. Altern. Ther. Health
Med. 11, 44–52.
Miller, K.M., Perry, P.A., 1990. Relaxation technique and postoperative pain in patients un-
dergoing cardiac surgery. Heart Lung 19, 136–146.
Miller, J., Boureau, F., Albe-Fessard, D., 1979. Supraspinal influences on nociceptive flexion
reflex and pain sensation in man. Brain Res. 179, 61–68.
Novak, P., 2011. Assessment of sympathetic index from the Valsalva maneuver. Neurolo-
gy 76, 2010–2016. http://dx.doi.org/10.1212/WNL.0b013e31821e5563.
Park, E., Oh, H., Kim, T., 2013. The effects of relaxation breathing on procedural pain and
anxiety during burn care. Burns 1–6. http://dx.doi.org/10.1016/j.burns.2013.01.006.
Rasbash, J., Charlton, C., Browne, W., Healy, M., Cameron, B., 2005. MLwiN Version 2.02.
Centre for Multilevel Modelling, University of Bristol.
Rhudy, J.L.J., France, C.C.R., 2007. Defining the nociceptive flexion reflex (NFR) threshold
in human participants: a comparison of different scoring criteria. Pain 128,
244–253. http://dx.doi.org/10.1016/j.pain.2006.09.024.
Rhudy, J.L., Williams, A.E., McCabe, K.M., Nguyen, M.A.T.V., Rambo, P., 2005. Affective
modulation of nociception at spinal and supraspinal levels. Psychophysiology 42,
579–587. http://dx.doi.org/10.1111/j.1469-8986.2005.00313.x.
Rhudy, J.L., McCabe, K.M., Williams, A.E., 2007. Affective modulation of autonomic reac-
tions to noxious stimulation. Int. J. Psychophysiol. 63, 105–109. http://dx.doi.org/10.
1016/j.ijpsycho.2006.09.001.
Roy, M., Lebuis, A., Peretz, I., Rainville, P., 2011. The modulation of pain by attention and
emotion: a dissociation of perceptual and spinal nociceptive processes. Eur. J. Pain
15, 641.e1-10. http://dx.doi.org/10.1016/j.ejpain.2010.11.013.
Sandrini, G., Serrao, M., Rossi, P., Romaniello, A., Cruccu, G., Willer, J.C., 2005. The lower
limb flexion reflex in humans. Prog. Neurobiol. 77, 353–395. http://dx.doi.org/10.
1016/j.pneurobio.2005.11.003.
Sarton, E., Dahan, A., Teppema, L., Berkenbosch, A., van den Elsen, M., van Kleef, J., 1997.
Influence of acute pain induced by activation of cutaneous nociceptors on ventilatory
control. Anesthesiology.
Schmidt-Vanderheyden, W., Koepchen, H.P., 1970. Investigations into the fluctuations of
proprioceptive reflexes in man. II. Fluctuations of the patellar tendon reflex and
their relation to vegetative rhythms during controlled respiration. Pflugers Arch.
317, 72–83.
Schmidt-Vanderheyden, W., Heinich, L., Koepchen, H.P., 1970. Investigations into the fluc-
tuations of proprioceptive reflexes in man. I. Fluctuations of the patellar tendon reflex
and their relation to the vegetative rhythms during spontaneous respiration. Pflugers
Arch. 317, 56–71.
Sin, P.Y.W., Galletly, D.C., Tzeng, Y.C., 2010. Influence of breathing frequency on the pat-
tern of respiratory sinus arrhythmia and blood pressure: old questions revisited.
Am. J. Physiol. Heart Circ. Physiol. 298, H1588–H1599. http://dx.doi.org/10.1152/
ajpheart.00036.2010.
Skljarevski, V., Ramadan, N.M., 2002. The nociceptive flexion reflex in humans — review
article. Pain 96, 3–8.
Spruyt, A., Clarysse, J., Vansteenwegen, D., Baeyens, F., Hermans, D., 2010. Affect 4.0: a free
software package for implementing psychological and psychophysiological experi-
ments. Exp. Psychol. 57, 36–45. http://dx.doi.org/10.1027/1618-3169/a000005.
Tanii, K., Sadoyama, T., Sanjo, Y., Kogi, K., 1973. Appearance of effort-depending changes
in static local fatigue. J. Hum. Ergol. (Tokyo) 2, 31–45.
Triedman, J.K., Saul, J.P., 1994. Blood pressure modulation by central venous pressure and
respiration. Buffering effects of the heart rate reflexes. Circulation 89, 169–179.
http://dx.doi.org/10.1161/01.CIR.89.1.169.
Van Damme, S., Crombez, G., Van Nieuwenborgh-De Wever, K., Goubert, L., 2008. Is dis-
traction less effective when pain is threatening? An experimental investigation
with the cold pressor task. Eur. J. Pain 12, 60–67. http://dx.doi.org/10.1016/j.ejpain.
2007.03.001.
Van Diest, I., Perlman, G., Bradley, M.M., Laplante, M., Lang, P.J., 2005. Gasp! You startled
me! Respiratory patterns to startling probes. Psychophysiology 42.
Verhoeven, K., Van Damme, S., Eccleston, C., Van Ryckeghem, D.M.L., Legrain, V., Crombez,
G., 2011. Distraction from pain and executive functioning: an experimental
58 H. Jafari et al. / International Journal of Psychophysiology 101 (2016) 50–58

investigation of the role of inhibition, task switching and working memory. Eur.
J. Pain 15, 866–873. http://dx.doi.org/10.1016/j.ejpain.2011.01.009.
Vogel, E.R., Sandroni, P., Low, P.A., 2005. Blood pressure recovery from Valsalva maneuver
in patients with autonomic failure. Neurology 65, 1533–1537. http://dx.doi.org/10.
1212/01.wnl.0000184504.13173.ef.
Yildirim, G., Sahin, N.H., 2004. The effect of breathing and skin stimulation techniques on
labour pain perception of Turkish women. Pain Res. Manag. 9, 183–187.
Zautra, A.J., Fasman, R., Davis, M.C., Craig, A.D.B., 2010. The effects of slow breathing on af-
fective responses to pain stimuli: an experimental study. Pain 149, 12–18. http://dx.
doi.org/10.1016/j.pain.2009.10.001.
Zunhammer, M., Eichhammer, P., Busch, V., 2013. Do cardiorespiratory variables predict
the antinociceptive effects of deep and slow breathing? Pain Med. 14, 843–854.
http://dx.doi.org/10.1111/pme.12085.