Download Important influence of respiration on human R

Important influence of respiration
power spectra is largely ignored
on human R-R interval
TROY E. BROWN,
LARRY
A. BEIGHTOL,
JUNKEN
KOH, AND DWAIN
L. ECKBERG
Departments
of Medicine and Physiology, Hunter Holmes McGuire Department of Veterans Affairs
Medical Center and Medical Collegeof Virginia, Richmond, Virginia 23249
heart rate; respiration
rhythmia
rate; tidal volume; respiratory
sinus ar-
orheartratesareused
widely as indexes of the level of autonomic traffic to the
heart. This usage has its origin in a study published by
Katona and Jih (18), who showed that in anesthetized
dogs with constant breathing rates respiration-related
peak-valley R-R interval fluctuations (respiratory sinus
arrhythmia) are related linearly to absolute vagal firing
rates. On the basis of this empirical observation, R-R
interval fluctuations have been promoted as faithful
noninvasive quantitative estimates of human vagal-cardisc nerve traffic. In recent years, the time-domain
method used by Katona and Jih has been supplanted by
frequency-domain methods. Some workers consider
power spectral density or autoregressive approaches to
FLUCTUATIONSOFR-RINTERVALS
2310
be superior to time-domain analyses of R-R intervals because the former approaches exclude random fluctuations and may provide quantitative information on the
level of sympathetic- as well as vagal-cardiac neural outflow (27).
Several studies have shown that respirato ry rate and
tidal volume exert major influ .ences on R-R interval or
heart rate fluctuations (8, 16, 21, 37). This genre of research is important because if fluctuations of R-R intervals are to be taken as valid noninvasive indexes of autonomic neural traffic, then they should reflect such traffic
faithfully and should not be influenced importantly by
respiratory-autonomic
interactions unrelated to net
neural outflow. Our study had two goals. First, we determined the effects of different voluntarily controlled respiratory rates and tidal volumes on R-R interval power
spectra. Second, we surveyed published literature to
learn how workers in this field have dealt with the potential influence of respiration on these measurements. Our
results show that respiration exerts a strong influence on
low-frequency as well as respiratory frequency R-R interval power spectra and that this influence is largely
ignored in published research.
METHODS
Subjects. Nine healthy subjects (8 men and 1 woman;
age 23-32 yr) were studied in recumbency. Six of the
subjects had been studied previously (8), and three were
studied prospectively. The prospective study was identical to the earlier study in all important details (see below). This research was approved by the human investigation committees of the Hunter Holmes McGuire Department of Veterans Affairs Medical Center and the
Medical College of Virginia. All volunteers gave their informed written consent to participate in this research.
Measurements. We recorded the electrocardiogram,
integrated tidal volume [Fleisch pneumotachograph for
the retrospective study and transit-time
ultrasonic
breath analyzer (GHG Medizin-Elektronik,
Zurich,
Switzerland) for the prospective study], and end-tidal
CO, concentration (infrared analyzer) on FM tape.
Respiratory control. Subjects breathed with an auditory
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
BROWN,TROY E., LARRY A. BEIGHTOL,JUNKENKOH,AND
L. ECKBERG.
Important
influence of respiration on human R-R interval power spectra is largely ignored. J. Appl. Physiol. 75(5): 2310-2317,1993.-Frequency-domain
analyses of RR intervals are used widely to estimate levels of autonomic
neural traffic to the human heart. Because respiration
modulates autonomic activity, we determined
for nine healthy subjects the influence of breathing frequency and tidal volume on
R-R interval power spectra (fast-Fourier
transform
method).
We also surveyed published literature
to determine current
practices in this burgeoning
field of scientific inquiry. Supine
subjects breathed at rates of 6, 7.5, 10, 15, 17.1, 20, and 24
breaths/min
and with nominal tidal volumes of 1,000 and 1,500
ml. R-R interval power at respiratory
and low (0.06-0.14 Hz)
frequencies declined significantly
as breathing
frequency increased. R-R interval power at respiratory frequencies was significantly greater at a tidal volume of 1,500 than 1,000 ml. Neither breathing frequency nor tidal volume influenced average
R-R intervals significantly. Our review of studies reporting human R-R interval power spectra showed that 51% of the studies
controlled respiratory
rate, 11% controlled
tidal volume, and
11% controlled both respiratory rate and tidal volume. The major implications
of our analyses are that breathing parameters
strongly influence low-frequency
as well as respiratory
frequency R-R interval power spectra and that this influence is
largely ignored in published research.
DWAIN
RESPIRATORY
MODULATION
OF
AUTONOMIC
OUTFLOW
2311
’ In this article, we use the term respiratory
frequency
to signify R-R
interval
power at respiratory
rates and low frequency
to signify power
between
0.06 and 0.14 Hz. The 0.06- to 0.14-Hz
range was chosen to
encompass
R-R interval
frequencies
that are considered
to reflect sympathetic
outflow
(24). Resting humans also have R-R interval
spectral
power at lower frequencies
(~0.06 Hz), which are thought
to reflect
thermoregulatory
oscillations
(20). We did not evaluate the lowest portion of each subject’s power spectra.
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
signal at rates of 6, 7.5, 10, 15, 17.1, 20, and 24 breaths/
20 log( output /input). Inpu t was the tidal volume in
min. They controlled their tidal volumes at nominal lev- liters, and output was the total absol.ute power of the
els of 1,000 and 1,500 ml according to signals displayed respiratory frequency or low-frequency portion of the Ron a calibrated oscilloscope. Subjects wore a face mask R interval power spectrum.
connected to the pneumotachograph or transit-time ulWe evaluated data for normality with the Kolmotrasonic breath analyzer and a three-way respiratory
gorov-Smirnov test (22). We made statistical comparivalve (Ewald Koegel, San Antonio, TX, for the retrospecsons on normally distributed data with repeated-meative study and Hans Rudolph, Kansas City, MO, for the sures analysis of variance to test for significant individprospective study).
ual differences among mean integrated R-R interval
Experimental protocol. Initial measurements were ob- power at different respiratory rates and tidal volumes.
tained at the lowest ventilation rate, 6 breaths/min with
Then we used Scheffe’s test to identify significantly difa tidal volume of 1,000 ml. The end-tidal CO, concentraferent means. We considered differences significant
tion measured at this level of ventilation was maintained
when P 5 0.05. Two-tailed tests were used.
at all other breathing frequencies and tidal volumes by
Literature reuiew. We searched the National Library of
addition of CO, to the breathing line. The remaining res- Medicine Medline data base for all references between
piratory intervals at the two tidal volumes were used in 1966 and the first quarter of 1991 to “power spectra” or
random sequence. Data for each respiratory rate and ti“power spectrum” and “heart rate” or “R-R interval” in
dal volume were recorded during one continuous 128-s
humans. We examined the articles identified by this
period. Breathing was not controlled between data col- search, references cited in those articles, and other referlection periods.
ences known to us to determine how often respiratory
Data analyses. Electrocardiograms and tidal volumes rate a.nd tidal volume were measured and controlled.2
were digitized (CODAS, Dataq Instruments, Akron, OH) Then we searched the Science Citation Index (Institute
with a Compaq 386/25 microcomputer. Power spectral for Scientific Information, Philadelphia, PA) to deteranalyses were derived from R-R interval time series with
mine how many times each of these references was cited
a custom program developed for use with a software
from 1974 to 1991. We report results from the most influpackage (DADiSP, DSP Development, Cambridge, MA).
ential articles, those cited 210 times, individually.
The technique used for estimation of R-R interval power
was based on the Welch algorithm of averaging periodo- RESULTS
grams (42), which was implemented according to the
Experimental data. Figure 1 depicts the digitized elecmethod of Rabiner et al. (31). A 128-s time series of beatto-beat R-R intervals was fitted to a cubic spline func- trocardiogram, tidal volume, and R-R interval signals
tion, interpolated at 8 Hz to obtain equidistant time in- from one subject during breathing at respiratory rate of
tervals, and divided into seven equal overlapping seg- 6,15, or 24 breaths/min and at a nominal tidal volume of
1,000 ml. The rhythmic R-R interval changes during
ments. Each segment was in turn detrended, Hanning
controlled breathing in this subject were similar to those
filtered, and fast-Fourier transformed to its frequency
representation. The modified periodograms were aver- observed in other subjects. In this subject, R-R interval
aged to produce the spectrum estimate. The method we variability decreased as breathing frequency increased;
used yielded a frequency resolution of 0.0078 Hz and a at the slowest breathing rate, maximum R-R intervals
were longer and minimum R-R intervals were shorter
flat response over the frequencies of interest.
We measured total absolute R-R interval power at low than at the two fastest breathing rates.
Subjects maintained average respiratory rates within
and respiratory frequencies over fixed frequency ranges.
We determined the respiratory frequency power band- 0.3% of targeted levels and tidal volumes at -7% below
width as follows: 1) power spectral analyses were per- targeted levels. End-tidal CO, concentrations were similar (P > 0.05) at all breathing frequencies and tidal volformed on each respiratory signal, 2) the respiratory
umes and ranged between 4.7 and 6%. In all subjects,
bandwidths were determined (95% of peak respiratory
end-tidal CO, was maintained within 0.25% of targeted
power), and 3) the average bandwidth at each respiratory
frequency was calculated (all were 0.08 Hz). The 0.08-Hz levels. Figure 2 shows average R-R intervals at all
bandwidth derived from analysis of respiration was used breathing rates and both tidal volumes. There were no
to integrate R-R intervals. (Thus, R-R interval spectral statistically significant differences among mean R-R inpower was integrated at frequencies to.04 Hz and cen- tervals (which ranged between 0.905 and 1.039 s) at different respiratory rates and tidal volumes. All data were
tered on each respiratory frequency.) A fixed bandwidth,
centered at 0.1 Hz (0.06-0.14 Hz), was also used for low- normally distributed.
Figure 3 shows average R-R interval power spectral
frequency spectral area calculations.’ We calculated
gain-frequency responses to construct Bode plots as waveforms at all respiratory rates and at a nominal tidal
2312
RESPIRATORY
n
MODULATION
OF AUTONOMIC
20
30
OUTFLOW
1.2
0
L
2
1.0
l t 0.8
Qf
I
f~
0.6
0
time
FIG.
R-R
1. Experimental
interval
variability
records from 1 subject. Respiratory
rates
was greatest at slowest breathing
rate.
shown
volume of 1,000 ml. At each breathing interval, R-R interval power spectra were concentrated
at the breathing
frequency. Spectral power declined precipitously
when
the respiratory
rate was increased
from 10 to 15
breaths/min.
Figure 4 depicts average integrated
respiratory
frequency and low-frequency
R-R interval power and gain.
Power at the respiratory
frequency (Fig. 4A) declined
significantly
as respiratory
rate increased (P < 0.001); a
major reduction occurred between respiratory
rates of
7.5 and 15 breaths/min
at both tidal volumes (P < 0.05).
Respiratory
frequency power was significantly
greater at
tidal volumes of 1,500 than 1,000 ml (by an average of
1.50 .b
1000
1500
0
0
ul
z
0
4
8
15 (B),
and 24 breaths/min
12
0
2.5
5
7.5
(s)
are 6 (A),
(C).
17%; range, 4-33s; P < 0.05). There was a highly significant linear correlation
between R-R interval power at
respiratory
frequencies and R-R interval standard deviations (r = 0.87 at 1,000 ml and 0.93 at 1,500 ml).
Power at low frequencies (Fig. 4C) also declined signifi-
ml
ml
1.25
L
a,
c
'CY
I
fx
c
1.00
0.75
F/i
E
0.50
5
respiratory
10
15
rate
2. Mean (+-SE) of R-R interval
tidal volume. There were no statistically
mean R-R intervals.
FIG.
20
25
(breaths/min)
at each respiratory
rate and
significant
differences
among
//-31
2
FIG. 3. Average
R-R interval
power spectra at respiratory
rates of 6,
7.5, 10, 15, 17.1,>0,
and 24 breaths/m&
and nominal
tidal volume
of
1,000 ml. Respiratory
frequency
power declines as respiratory
rate increases.
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
cll
RESPIRATORY
MODULATION
OF
AUTONOMIC
OUTFLOW
2313
-25
n
5
5
25
15
respiratory
rate
(breaths/min)
cantly as respiratory rate increased (P < 0.001). Increases of low-frequency power at rates of 510 breaths/
min resulted from overlap of respiratory frequency and
low-frequency bands (Fig. 3). Low-frequency power at
the two tidal volumes was similar (P > 0.05).
The gain-frequency response (Bode plot) of the respiratory frequency component of R-R interval power spectra documented significantly decreasing system gain
with increasing respiratory rate (P < 0.001; Fig. 4B). A
nonlinear segmented model fit (SAS Institute, Cary, NC)
analysis suggested that the gain-respiratory rate response relationship was best described by a plateau region between 6 and 10 breaths/min (0.10-0.17 Hz) for
both tidal volumes. The plateau region was succeededby
roll-offs of 10.6 dB/decade for the l,OOO-ml tidal volume
and 14.7 dB/decade for the 1,500-ml tidal volume. Gain
of respiratory frequency R-R interval power spectra at
the lowest breathing frequencies (6-10 breaths/min) was
significantly greater than that at all other breathing frequencies (15-24 breaths/min) for both tidal volumes
(P < 0.05). Respiratory frequency system gain was significantly greater at tidal volumes of 1,500 than 1,000 ml (by
an average of 7%; range, 4-13%; P < 0.001).
The gain-frequency response of the low-frequency
component of the R-R interval power spectra also documented significantly decreasing system gain with increasing respiratory rate (P < 0.001; Fig. 40). Gain of
low-frequency R-R interval power spectra at the two lowest breathing frequencies (6 and 7.5 breaths/min) was
significantly greater than that at the highest breathing
frequencies (lo-24 breathsimin) for both tidal volumes
(P < 0.001). Gain at 10 breaths/min also was significantly greater than that at higher breathing frequencies
(15-24 breaths/min; P < 0.05). Low-frequency system
gains at the two tidal volumes were nearly identical
(P > 0.05).
Literature
review. Our medical literature search identified 147 articles that reported use of R-R interval or
heart rate power spectral analysis in humans.3 Table 1
summarizes these articles. Respiratory rate and tidal volume were controlled in only 11% of studies in which control was possible.
Our Science Citation Index search indicated that 23 of
the 147 articles were cited 210 times. Table 2 summarizes these articles. There were 133 citations of the article
by Pagani et al. (27) and 233 citations of the remaining
articles. Four articles (6, 19, 29, 33) dealt with methods
and modeling and so were excluded. Thirteen of the remaining 19 articles (68.4%) included measurements of
respiratory rate. Three articles (4, 25, 34) involved ambulatory subjects and thus were excluded. Four of the
remaining 16 studies (25%) included measurements of
tidal volume. Control of respiratory rate and tidal volume was impossible for studies that involved ambulatory
subjects and infants and for those that involved psychological treatments. Of the remaining eight studies in
which respiratory rate and tidal volume could have been
controlled, respiratory rate was controlled in six (75%)
and tidal volume was controlled in one (12.5%).
DISCUSSION
We studied the influence of respiratory rate and tidal
volume on human R-R interval power spectra and sur3 A complete
list of the articles uncovered
by our search can be obtained from the National
Auxiliary
Publications
Service, c/o Microfiche Publications,
PO Box 3513, Grand
Central
Station,
NY 10017
(NAPS
No. 05059).
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
FIG. 4. Effects
of respiratory
rate on R-R
interval
power spectra
at nominal
tidal volumes of 1,000 and 1,500 ml. Data are means +
SE. A: respiratory
frequency
power
significantly
declined
as respiratory
rate increased
(P < 0.001). Power was greater at tidal volumes
of 1,500 than 1,000 ml (P < 0.05). B: respiratory
frequency
gain also significantly
declined
with
increasing
respiratory
rate (P < 0.001). * There
was significant
reduction
in gain from lower (610 breathsimin)
to higher breathing
rates (15
24 breaths/min;
P < 0.05). Gain was significantly higher (P < 0.001) at 1,500- than l,OOOml tidal
volume.
C: low-frequency
power
significantly
declined
as respiratory
rate increased (P < 0.001). Power at 2 tidal volumes
was similar.
D: low-frequency
gain also significantly decreased
as respiratory
rate increased
(P < 0.001). ** Gain at 2 lower breathing
frequencies
(6 and 7.5 breathsimin)
was greater
than at all other breathing
frequencies
(lo-24
breaths/min;
P < 0.001). * Additionally,
lobreaths/min
respiratory
rate gain was significantly
greater
than higher
respiratory
rates
(15-24 breathsjmin;
P < 0.05). Values for 2 tidal volumes
were similar.
2314
RESPIRATORY
MODULATION
TABLE
1. Review of publications reporting R-R interval
power spectra in humans
n
Articles
reviewed
Possibility
of respiratory
Measured
respiratory
Possibility
of tidal volume
Measured
tidal volume
Possibility
of respiratory
Controlled
respiratory
Possibility
of tidal volume
Controlled
tidal volume
Controlled
respiratory
rate measurement
rate
measurement
rate control
rate
control
rate and tidal
n, no. of publications;
Percentages,
in which possibility
of measurement
volume
Percentages
147
119
61
90
20
65
24
65
7
7
percentage
of published
or control was realized.
51
22
37
11
11
studies
AUTONOMIC
OUTFLOW
power spectra at fewer respiratory rates (5 vs. 7 in our
study), at more (6 vs. 2 in our study) and different (4OO1,400 vs. 1,000 and 1,500 ml in our study) tidal volumes,
and at different experimentally maintained average endtidal CO, concentrations (3.5-4.6 vs. 4.7-6% in our
study).
Other studies also document a dependence of R-R interval (or heart rate) fluctuations on respiratory parameters. The study most similar to ours is that of Hirsch and
Bishop (16), who characterized heart rate fluctuations in
the time domain as respiration-related peak-valley R-R
intervals (the longest minus the shortest R-R interval
during each breath). Hirsch and Bishop derived Bode
plots very similar to ours with a plateau region, a corner
frequency at 0.11 Hz, and a linear roll-off of 21 dB/decade. They also showed that peak-valley R-R interval
changes and the steepness of the roll-off portion are directly proportional to tidal volume. Our more limited
data (Fig. 4B) are concordant with the last observation.
Hirsch and Bishop did not indicate if the relationship
between tidal volume and respiratory R-R interval
changes was statistically significant. However, their Fig.
4 [and data published by Eckoldt and Schubert (lo)] suggests that tidal volume may be an important determinant
of such changes: in one subject, sixfold increases of tidal
volume led to approximately threefold increases of slowbreath peak-valley R-R interval intercepts.
Saul et al. (35), who used frequency-domain methods
similar to ours but a broad-band respiratory frequency
input, also documented a critical dependence of R-R interval power spectra on breathing-rate. Although the
2. Influential R-R interval power spectral
research on humans
TABLE
Respiratory
Measured
Authors
Pagani et al. (27)
Saul et al. (34)
Gordon
et al. (13)
Myers et al. (25)
Porges et al. (29)
Rohmert
et al. (32)
Hyndman
and Gregory
(17)
Baselli et al. (2)
De Boer et al. (7)
Eckberg
et al. (9)
Walter
and Porges (41)
Gordon
et al. (12)
Rompelman
et ai. (33)
Lishneretal=(23)
~~~n~~~~PEgreY(ll)
Kit-ey et al’ tlg)
Nugentand*Finley(26)
Vybiral et al. (40)
Bigger
et al. (4)
~~~~~~~“,$~)
Shannonet g,. (38)
No. of
Citations
Parameters
Controlled
Rate
Volume
Rate
Volume
133
33
30
30
25
25
24
23
23
22
22
18
17
16
+
+
+
x
+
+
x
x
x
x
+
+
x
x
x
x
X
X
+
+
+
+
x
+
+
+
x
-
+
X
X
x
+
X
X
x
-
ti
+
-
--
x-
x-
12
12
12
11
11
10
10
x
+
+
x
+
x
x
-
x
x
+
x
x
x
+
x
x
x
x
x
-
-
Reference
no. is given in parentheses.
+, Parameter
measured
or
controlled;
-, parameter
not measured
or controlled;
X, research design prevented
measurement
or control
of narameter
or article was
devoted to methods
or modeling.
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
veyed the medical literature to learn how scientists who
employ power spectral analyses in their research have
dealt with respiration. We reached three major conclusions. First, respiratory rate and tidal volume strongly
influence low-frequency as well as respiratory frequency
R-R interval power. Second, because the breathing parameters we studied did not alter average R-R intervals,
variations of respiration distribute autonomic outflow
within the respiratory cycle but do not alter net levels of
autonomic traffic. Thus, vagal activity during the various
phases of the respiratory cycle varied considerably,
whereas the overall tonic level remained constant. Third,
most of the scientists who performed frequency-domain
R-R interval analyses neither controlled nor accounted
for the strong influence of respiration on their results.
Respiratory influences on R-R interval power spectra at
respiratory frequencies. Breathing rates of - 10 breaths/
min or less yielded maximum R-R interval power.
Breathing rates of >lO breaths/min yielded reduced levels of R-R interval power in an inverse relationship to
breathing rate (Figs. 3 and 4, A and B). However, over the
range we studied, breathing rate did not significantly
alter average R-R intervals (Fig. 2). Our study effectively
doubles the published literature on R-R interval power
spectra in subjects who breathed at multiple experimentally controlled breathing rates and tidal volumes; we are
aware of only 0ne similar article (37).
Selman et al. (37), in an excellent but rarely cited
study, obtained results that differed from ours qualitatively and quantitatively. First, they found an inverse
relationship between tidal volume and system gain. We
found a direct relationship (Fig. 4B). Second, they obtained a Bode plot described by a distinct “M shape”
such that system gains were much higher at breathing
rates of 10 and 16 than at 12 breaths/min. In our study
(Figs. 3 and 4), there was only a hint of an M shape.
Third, they did not define a convincing roll-off of system
gain at higher breathing frequencies (probably because
they studied only 5 breathing frequencies). We found a
steady decline of R-R interval power at breathing rates
above -10 breaths/min. Fourth, they did not report lowfrequency power separately. The methods used by Selman et al. differed from ours in several respects. They
studied subjects whose ages spanned a wider range (2652 vs. 23-32 yr in our study). Their subjects were seated
and ours were supine. They measured R-R interval
OF
RESPIRATORY
MODULATION
AUTONOMIC
OUTFLOW
2315
interval power, however, there were only trivial changes
of mean R-R intervals (Fig. 2). This observation
is related to one made by Grossman
et al. (14), who found
that in ,&blocked subjects respiration-induced
changes of
peak-valley
R-R intervals
were not associated
with
changes of mean R-R intervals. We suggest that within
subjects different breathing frequencies and depths distribute vagal firing within the respiratory
cycle but do
not alter the net level of vagal outflow.
Literature review. We identified and reviewed 147 articles that reported measurements
of human R-R interval
(or heart rate) power spectra to determine the prevailing
practices of workers
in this field. Only about one-half of
the studies involved measurement
of breathing rate and
only about one-third involved control of breathing rate.
We also determined the number of times each of the 147
articles had been cited to identify the most influential
studies in this field. Authors of the most heavily cited
studies (Table 2) were more likely to recognize the importance of measurement
and control of respiratory
rate
than those of less cited studies. However, very few (only
11%) of the studies involved control of tidal volume.
Pagani et al. (27), authors of the most heavily cited
article in this field, discount the importance of control of
respiration
for frequency-domain
analyses of R-R intervals. They argue that voluntary control of breathing increases power at respiratory
frequencies (and net vagal
outflow)
and decreases power at low frequencies. This
conclusion was based on a comparison
of R-R interval
power (as delimited by an autoregressive
algorithm)
at
respiratory
frequencies
during spontaneous
breathing
(at an average rate of 14.4 breaths/min)
and during controlled breathing
(at an average rate of 19.8 breaths/
min). The approach of Pagani et al. may have been
flawed by their comparison of R-R interval power spectra
at different breathing rates. However, two other groups
compared respiration-related
R-R interval fluctuations
during controlled and spontaneous breathing at the same
breathing rates and concluded that voluntary control of
breathing
does not alter R-R interval
fluctuations.
Hirsch and Bishop (16) found that the plateau frequency
intercept in 17 subjects was similar during controlled and
spontaneous breathing. Eckberg et al. (9) found that respiration-related
R-R interval fluctuations
(and muscle
sympathetic
nerve traffic) in eight subjects were nearly
superimposable
during
controlled
and spontaneous
breathing.
Potential limitations. First, use of fast-Fourier
transform methods for power spectral analysis requires that
signals be stationary.
The mathematical
concept of stationarity
may not be applicable to the highly variable
signals that can be obtained from conscious human volunteers, and we did not test our data for stationarity.
However,
our subjects maintained exquisite control of
breathing
frequency
during data collection
periods.
Moreover, our results (Fig. 4) are very similar to those of
Hirsch and Bishop (16), who used time-domain
methods.
Second, in our analysis, actual tidal volumes used by our
subjects (-930 and 1,400 ml) were larger than the average tidal volumes studied by others. However, Bendixen
et al. (3) showed that almost one-half of the healthy
women they studied had resting tidal volumes as high as
the smaller tidal volume we used.
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
shapes of the transfer
function of Saul et al. and our
Bode plot are similar, there may be quantitative
differences. We found (Fig. 4B) that the plateau height and the
roll-off slope are influenced significantly
by tidal volume.
In the study of Saul et al., tidal volumes and respiratory
rates covaried; therefore, the separate influence of tidal
volume on the height of the plateau and the slope of the
roll-off could not be determined.
Respiratory
influences on R-R interval power spectra at
low frequencies. A major rationale for study of low-frequency R-R interval fluctuations
is that they reflect levels of sympathetic
traffic to the heart. Although sympathetic nerve traffic also fluctuates
at respiratory
frequencies (9, 30), R-R interval responses to such rapid
changes of sympathetic
activity are small because of the
sluggishness
of adrenergically
mediated sinus node responses. Our results (Fig. 4, C and D) indicate that slow
breathing
provokes
major increases of low-frequency
(0.06-0.14 Hz) R-R interval power. In literature involving R-R interval (or heart rate) power spectra, the terms
respiratory
frequency and high frequency are used interchangeably. Surprisingly,
this literature does not anticipate that during periods of slow breathing respiratory
frequency
and low frequency
overlap. Obviously,
as
breathing rate slows, the possibility that fluctuations
of
sympathetic
nerve traffic contribute to respiratory
frequency R-R interval changes increases (however, probably not as a step function). Under such circumstances,
the view that changes of sympathetic outflow do not contribute to respiratory
frequency power becomes increasingly untenable. It is highly unlikely that the greatly augmented low-frequency
R-R interval power we recorded at
slow breathing rates (Figs. 3 and 4C) reflects increased
sympathetic
neural outflow. Seals et al. (36) found that,
although rapid breathing redistributes
directly measured
human (muscle, not cardiac) sympathetic
nerve activity
within the various phases of the respiratory
cycle, it does
not alter the net tonic level of sympathetic
outflow
within the overall cycle.
R-R interval power spectra as noninvasive indexes of
vagal-cardiac nerve traffic. Kollai and Mizsei (21) argued
persuasively
that respiratory
rate and tidal volume
should be factored into estimates of human vagal-cardiac
nerve traffic. They compared peak-valley
R-R interval
changes during uncontrolled
breathing and R-R interval
shortening provoked by large-dose intravenous
atropine
(the current “gold standard” for quantification
of resting
vagal-cardiac
nerve traffic in humans).
They found a
weak but significant (r = 0.61) linear correlation between
peak-valley R-R intervals and atropine-induced
R-R interval shortening. However, they also found that when
respiratory
interval and tidal volume were factoredin
as
coregressors this correlation was much stronger (r = 0.93).
Our study raises new questions about the validity of
R-R interval power spectra as quantitative
indexes of
vagal-cardiac
neural traffic within individual subjects.
Our data suggest that even with explicit knowledge (and
control) of respiration, respiratory
sinus arrhythmia
may
not reflect tonic vagal-cardiac
neural traffic. Our subjects varied the magnitude of their respiratory
frequency
R-R interval power -IO-fold
simply by changing their
breathing rates (Fig. 3). Despite huge changes of R-R
OF
2316
RESPIRATORY
MODULATION
NOTE
ADDED
IN
PROOF
We call attention to the article by Novak et al. (%a) that was
published while our study was in press. Novak et al. plotted
Wigner distributions
of R-R intervals, tidal volumes, and systolic and diastolic pressures during continuously
varying periods of slow to rapid breathing. Although the subjects in this
study did not control their tidal volumes, their R-R interval
Wigner distributions
bear a striking similarity
to the average
data we depict in Fig. 3.
REFERENCES
F., G. BASELLI, AND S. CERUTTI. AR identification and
spectral estimate applied to the R-R interval measurements. Int. J.
Bio-Med.
Comput.
16: 201-215. 1985.
1. BARTOLI,
G., S. CERUTTI, S. CIVARDI, F. LOMBARDI, A. MALLIANI,
M. PAGANI, AND G. RIZZO. Heart rate variability signal
processing: a quantitative approach as an aid to diagnosis in cardiovascular pathologies. Int. J. Bio-Med.
Comput.
20: 51-70, 1987.
BENDIXEN,
H. H., G. M. SMITH, AND J. MEAD. Pattern of ventilation in young adults. J. Appl. Physiol. 19: 195-198, 1964.
BIGGER, J. T., JR., M. T. LA ROVERE, R. C. STEINMAN,
J. L. FLEISS,
J. N. ROTTMAN, L. M. ROLNITZKY, AND P. J. SCHWARTZ. Comparison of baroreflex sensitivity and heart period variability after myocardial infarction. J. Am. Coil. Cardiol. 14: 1511-1518, 1989.
DAVIS, J. N., AND D. STAGG. Interrelationships
of the volume and
time components of individual breaths in resting man. J. Physiol.
Lond. 245: 481-498, 1975.
DE BOER, R. W., J. M. KAREMAKER,
AND J. STRACKEE. Description of heart-rate variability data in accordance with a physiological model for the genesis of heartbeats. Psychophysiology
22: l47155, 1985.
DE BOER, R. W., J. M. KAREMAKER,
AND J. STRACKEE. Relationships between short-term blood-pressure fluctuations and heartrate variability in resting subjects. I. A spectral analysis approach.
Med. Biol. Eng. Comput. 23: 352-358,
1985.
ECKBERG,
D. L. Human sinus arrhythmia as an index of vagal
cardiac outflow. J. Appl. Physiol. 54: 961-966, 1983.
ECKBERG, D. L., C. NERHED, AND B. G. WALLIN.
Respiratory modulation of muscle sympathetic and vagal cardiac outflow in man. J.
Physiol. Lond. 365: 181-196, 1985.
ECKOLDT,
K., AND E. SCHUBERT. Zum Einfluss der Atemtiefe auf
die Sinusarrhythmie des Herzens. Acta Biol. Med. Ger. 34: 767-771,
1975.
GIDDENS, D. P., AND R. I. KITNEY. Neonatal heart rate variability
and its relation to respiration. J. Theor. Biol. 113: 759-780, 1985.
GORDON,
D., D. P. SOUTHALL, D. H. KELLY, A. WILSON, S. AKSELROD, J. RICHARDS,
B. KENET, R. KENET, R. J. COHEN, AND
D. C. SHANNON. Analysis of heart rate and respiratory patterns in
sudden infant death syndrome victims and control infants. Pediatr.
Res. 20: 680-684,
1986.
GORDON,
D., R. J. COHEN, D. KELLY, S. AKSELROD, AND D. C.
SHANNON.
Sudden infant death syndrome: abnormalities in short
term fluctuations in heart rate and respiratory activity. Pediatr.
Res. 18: 921-926, 1984.
GROSSMAN,
P., J. KAREMAKER,
AND W. WIELING.
Prediction of
tonic parasympathetic cardiac control using respiratory sinus arrhythmia: the need for respiratory control. Psychophysiology
28:
201-216,
1991.
GROSSMAN,
P., AND K. WIENTJES.
Respiratory sinus arrhythmia
and parasympathetic cardiac control: some basic issues concerning
quantification, applications and implications. In: Cardiorespiratory
and Cardiosomatic
Psychophysiology,
edited by P. Grossman,
K. H. L. Janssen, and D. Vaitl. New York: Plenum, 1986, p. 117138.
HIRSCH,
J. A., AND B. BISHOP. Respiratory sinus arrhythmia in
humans: how breathing pattern modulates heart rate. Am. J. PhysioZ. 241 (Heart Circ. Physiol. 10): H620-H629, 1981.
HYNDMAN,
B. W., AND J. R. GREGORY. Spectral analysis of sinus
arrhythmia during mental loading. Ergonomics
18: 255-270, 1975.
KATONA, P. G., AND F. JIH. Respiratory sinus arrhythmia: noninvasive measure of parasympathetic cardiac control. J. Appl. Physiol. 39: 801-805,
1975.
KITNEY,
R., D. LINKENS, A. SELMAN, AND A. MCDONALD.
The
interaction between heart rate and respiration. Part II. Nonlinear
analysis based on computer modelling. Automedica
Lond. 4: 141-
2. BASELLI,
M. MERRI,
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
We thank Paul Grossman, Abhijit Patwardhan, and J. Andrew Taylor for their critical reviews of this manuscript.
This research was supported by National Heart, Lung, and Blood
Institute Grant HL-‘22296, a grant from the Department of Veterans
Affairs, and National Aeronautics and Space Administration Grants
NAS9-16046 and NAG9-412.
Present addresses: J. Koh, Dept. of Anesthesia, Kure National Hospital, Kure, Hiroshima 737, Japan; T. E. Brown, Krug Life Sciences,
1290 Hercules, Suite 120, Houston, TX 77058.
Address for reprint requests: D. L. Eckberg, Cardiovascular Physiology, Hunter Holmes McGuire Dept. of VA Medical Center, 1201 Broad
Rock Blvd., Richmond, VA 23249.
Received 25 January 1993; accepted in final form 25 May 1993.
OUTFLOW
19.
153,1982.
20. KITNEY,
R. I. Proceedings: entrainment of the human RR interval
by thermal stimuli. J. Physiol. Lond. 252: 37P-38P,
1975.
21. KOLLAI, M., AND G. MIZSEI. Respiratory sinus arrhythmia is a limited measure of cardiac parasympathetic control in man. J. Physiol.
Lond. 424: 329-342,199O.
22. LILLIEFORS,
H. W. On
the Kolmogorov-Smirnov test for normality
with mean and variance unknown. J. Am. Stat. Assoc. 62: 399-402,
1967.
23. LISHNER, M., S. AKSELROD,
V. MOR AVI, 0. Oz, M. DIVON, AND M.
RAVID. Spectral analysis of heart rate fluctuations. A non-invasive,
sensitive method for the early diagnosis of autonomic neuropathy
in diabetes mellitus. J. Auton. New. Syst. 19: 119-125, 1987.
24. MALLIANI,
A., M. PAGANI, F. LOMBARDI, AND S. CERUTTI. Cardio-
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
Implications of this study. Frequency-domain
analyses
of R-R intervals represent sophisticated attempts to estimate, noninvasively,
levels of efferent autonomic
neural traffic to the human heart. Our analysis shows
that respiration influences both respiratory frequency
and low-frequency R-R interval power spectra. We propose that, since the influence of respiration is so strong
and pervasive, interpretation
of R-R interval power
spectra must be grounded on understanding of the respiratory patterns that shaped them.
It may be inappropriate
to use R-R interval power
spectra obtained during brief recording sessions to gauge
levels of efferent autonomic neural traffic if breathing
rate and tidal volume are neither measured nor controlled. During brief recording sessions, breathing patterns of healthy subjects are highly variable (5,39), and a
substantial percentage of breaths may occur at frequencies <9 and >20 breaths/min (15). Although the majority
of healthy supine young subjects have respiratory power
in the roll-off region, where small changes of breathing
rate are likely to provoke large changes of respiratory
frequency R-R interval power, some of the subjects have
substantial respiratory power in the plateau region (40
breaths/min;
J. A. Taylor, T. E. Brown, and D. L. Eckberg, unpublished data), where breathing rate is likely to
sharply augment low-frequency R-R interval power.
In conclusion, we studied the effects of respiratory rate
and tidal volume on R-R interval power spectra. Our results indicate that respiratory parameters strongly influence both low-frequency and respiratory frequency R-R
interval power spectra. The major implication
of these
results is that respiration must be controlled if R-R interval power spectra are to be interpretable.
Our survey
of published articles in this field shows that the great
majority do not meet this minimum condition.
OF AUTONOMIC
RESPIRATORY
MODULATION
AUTONOMIC
OUTFLOW
2317
32. ROHMERT,
W., W. LAURIG,
U. PHILIPP,
AND H. LUCZAK. Heart
rate variability
and work-load
measurement.
Ergonomics
16: 33-44,
1973.
O., A. J. R. M. COENEN, AND R. I. KITNEY. Measure33. ROMPELMAN,
ment of heart-rate
variability.
Part I. Comparative
study of heartrate variability
analysis methods.
Med. Biol. Eng. Comput.
15: 233239, 1977.
34. SAUL, J. P., Y. ARAI, R. D. BERGER, L. S. LILLY, W. S. COLUCCI,
AND R. J. COHEN. Assessment
of autonomic
regulation
in chronic
congestive
heart failure by heart rate spectral analysis. Am. J. Cardial. 61: 1292-1299,
1988.
M. H. CHEN, AND R. J. COHEN.
35. SAUL, J. P., R. D. BERGER,
Transfer
function
analysis of autonomic
regulation.
II. Respiratory
sinus arrhythmia.
Am. J. Physiol.
256 (Heart
Circ. Physiol.
25):
H153-H161,
1989.
Influence
of
36. SEALS, D. R., N. 0. SUWARNO, AND J. A. DEMPSEY.
lung volume
on sympathetic
nerve discharge
in normal
humans.
Circ. Res. 67: 130-141,
1990.
A., A. MCDONALD,
R. KITNEY,
AND D. LINKENS.
The
37. SELMAN,
interaction
between heart rate and respiration.
Part I. Experimental studies in man. Automedica
Land. 4: 131-139,
1982.
D. C., D. W. CARLEY, AND H. BENSON. Aging of modula38* SHANNON,
tion of heart rate. Am. J. Physiol.
253 (Heart
Circ. Physiol.
22):
H874-H877,1987.
TOBIN, M. J., M. J. MADOR, S. M. GUENTHER,
R. F. LODATO, AND
3g*
M. A. SACKNER. Variability
of resting respiratory
drive and timing
in healthy
subjects. J. Appl. Physiol. 65: 309-317,
1988.
4.
VYBIRAL, T., R. J. BRYG, M. E. MADDENS,
AND W. E. BODEN. Ef’ fect of passive tilt on sympathetic
and parasympathetic
components of heart rate variability
in normal subjects. Am. J. Cardiol.
63: 1117-1120,
1989.
41 WALTER,
G. F., AND S. W. PORGES. Heart
rate and respiratory
’ responses
as a function
of task difficulty:
the use of discriminant
analysis in the selection
of psychologically
sensitive
physiological
responses.
Psychophysiology
13: 563-571,
1976.
42. WELCH, P. D. The use of fast Fourier
transform
for the estimation
of power spectra:
a method
based on time averaging
over short,
modified
periodograms.
IEEE Trans. Audio Electroaccoust.
15: 7073, 1967.
Downloaded from http://jap.physiology.org/ by 10.220.33.2 on October 1, 2016
vascular
neural regulation
explored
in the frequency
domain.
Circulation 84: 482-492,
1991.
25. MYERS, G. A., G. J. MARTIN,
N. M. MAGID, P. S. BARNETT, J. W.
SCHAAD, J. S. WEISS, M. LESCH, AND D. H. SINGER. Power spectral
analysis of heart rate variability
in sudden cardiac death: comparison to other methods.
IEEE Trans. Biomed. Eng. 33: 1149-1156,
1986.
%a.NOvAK,
V., P. NOVAK, J. DE CHAMPLAIN,
A. R. LE BLANC, R. MARTIN, AND R. NADEAU.
Influence
of respiration
on heart rate and
blood pressure
fluctuations.
J. Appl. Physiol. 74: 617-626,
1993.
26. NUGENT, S. T., AND J. P. FINLEY. Spectral analysis of periodic and
normal
breathing
in infants.
IEEE Trans. Biomed. Eng. 30: 672675, 1983.
27. PAGANI, M., F. LOMBARDI,
S. GUZZETTI,
0. RIMOLDI,
R. FURLAN,
P. PIZZINELLI,
G. SANDRONE,
G. MALFATTO,
S. DELL’ORTO,
E.
PICCALUGA,
M. TURIEL, G. BASELLI,
S. CERUTTI,
AND A. MALLIANI. Power spectral analysis of heart rate and arterial
pressure
variabilities
as a marker
of sympatho-vagal
interaction
in man and
conscious
dog. Circ. Res. 59: 178-193,
1986.
28. PAGANI, M., F. LOMBARDI,
S. GUZZETTI,
G. SANDRONE,
0. RIMOLDI, G. MALFATTO,
S. CERUTTI, AND A. MALLIANI.
Power spectral density of heart rate variability
as an index of sympatho-vagal
interaction
in normal
and hypertensive
subjects. J. Hypertens.
2,
Suppl. 3: 383-385,
1984.
29. PORGES, S. W., R. E. BOHRER, M. N. CHEUNG, F. DRASGOW, P. M.
MCCABE,
AND G. KEREN. New time-series
statistic
for detecting
rhythmic
co-occurrence
in the frequency
domain: the weighted
coherence and its application
to psychophysiological
research.
Psychol. Bull. 88: 580-587,
1980.
30. PORTER, T. R., D. L. ECKBERG, J. M. FRITSCH, R. F. REA, L. A.
BEIGHTOL, J. F. SCHMEDTJE,
JR., AND P. K. MOHANTY.
Autonomic
pathophysiology
in heart failure patients.
Sympathetic-cholinergic
interrelations.
J. Clin. Invest. 85: 1362-1371,
1990.
31. RABINER,
L. R., R. W. SCHAFER, AND D. DLUGOS. Periodogram
method
for power spectrum
estimation.
In: Programs
for Digital
Signal Processing,
edited by Digital
Signal Processing
Committee,
IEEE Acoustics,
Speech, and Signal Processing
Sot. New York:
IEEE, 1979, p. 2.1-l-2.1-10.
OF