Vol. 273, Issue 4, R1352-R1360, October 1997
Lack of locomotor-cardiac coupling in trotting dogs
Adam D.
Simmons1,
David R.
Carrier2,
Colleen G.
Farmer1, and
Colin S.
Gregersen2
1 Department of Ecology and
Evolutionary Biology, Brown University, Providence, Rhode Island 02912;
and 2 Department of Biology,
University of Utah, Salt Lake City, Utah 84112
 |
ABSTRACT |
Coupling of locomotor and cardiac cycles
has been suggested to facilitate effective arterial delivery and venous
return during vigorous exercise. In an attempt to document
locomotor-cardiac coupling, we ran five dogs on a motorized treadmill
while monitoring heart activity with surface electrocardiogram
electrodes and locomotor events with high-speed video and an
accelerometer mounted on the dog's back. Analysis of the cardiac and
locomotor frequencies revealed that heart rate was usually slightly
greater than stride frequency. Hence the timing of the cardiac cycles
varied with respect to the phase of the locomotor cycles, and therefore
consistent coupling of the locomotor and cardiac cycles was not
observed in any of the dogs. However, the period of the cardiac cycle
sometimes varied in a rhythmic way that caused brief periods of
transient coupling of the locomotor and cardiac cycles in three of the
five dogs. These brief periods of coupling (5-20 heartbeats)
occurred at approximately the same phase relationship in each of the
three dogs. We hypothesize that the variation in cardiac period and the
resulting transient coupling are a function of locomotor and ventilatory influences on venous return and/or ventricular
ejection. Because venous return and ejection fraction are likely to
vary in an unpredictable manner when animals run in a complex
environment, we suggest that reflex control of heart rate will be
important during locomotion and strict integer coupling of the
locomotor and cardiac cycles is unlikely to evolve.
exercise; cardiac function; cardiac reflex; locomotion
| |
INTRODUCTION |
THE CIRCULATORY SYSTEM plays a central role in aerobic
exercise by linking the sites of gas exchange and energy storage with the mitochondria of working skeletal and cardiac muscle. Several lines
of evidence suggest that the functional capacity of the cardiovascular
system is the principal factor limiting the maximum rate of oxygen
consumption during exercise in many or most mammals (7). The importance
of cardiovascular function to respiration combined with potential
mechanical limitations to arterial delivery and venous return during
exercise has led several physiologists to suggest that coupling of the
cardiac and locomotor cycles might improve locomotor stamina during
vigorous exercise (2, 3, 5, 12, 13, 19-23).
Reports of coupling between the locomotor and cardiac cycles appear as
early as the 1920s. Coleman (12, 13), while working at the zoological
gardens of Regent's Park in London, described a matching of step rate
and heart rate in people and a variety of animals including a lynx, a
badger, and a serval. His observations of heart rate were taken by a
variety of visual methods or by feeling the radial pulse in the wrist.
Consequently, although his results indicate a similarity in rates, they
do not provide a direct demonstration of coupling.
Later, Aulie (2, 3) proposed a mechanical linkage between the locomotor
and cardiac cycles during flapping flight in birds. He postulated that
the increased thoracic pressure during contraction of a bird's
pectoral muscles might push blood back into the heart. For the heart to
work as efficiently as possible, one would expect it to contract when
it is full of blood rather than when it is empty. Therefore, if the
pectoral muscles functioned as a venous pump, the cardiac cycle of
flying birds could be expected to be coupled with the wing stroke.
Although Aulie's birds did not exhibit one-to-one coupling, he
observed that heartbeats were less likely to occur during pectoral
muscle activity than would be expected by pure chance. For much the
same reason, the venous pump that results from skeletal muscle
contractions, acting in conjunction with a series of one-way valves in
the veins of the limbs of running animals, could also be expected to
facilitate coupling of cardiac and locomotor rhythms.
A second type of venous pump, the respiratory pump, might also result
in locomotor-cardiac coupling. In mammals, caudal displacement of the
diaphragm helps produce inspiratory air flow by decreasing intrathoracic pressure. Simultaneously, the pressure in the abdomen increases. The net effect of these pressure changes is an increase in
the pressure differential between the peripheral veins and the right
atrium. Consequently, venous return is enhanced during inspiration.
Because mammals and birds couple their breathing with their locomotor
movements when they run (1, 5, 6, 8-10, 17, 28), it is not
unreasonable to expect that coupling of the heart to the ventilatory
and locomotor cycles would improve cardiac efficiency.
Kirby and collaborators (23) proposed a third possible reason to expect
coupling. They noted that, during normal locomotion, intramuscular
pressure rises to levels that often exceed peak systolic blood
pressures. Therefore, locomotor activity is likely to periodically
occlude blood flow through active muscles during each stride. To allow
efficient transport, these investigators reasoned that the cardiac
cycle should be timed to deliver blood through the locomotor muscles
when they are relaxed rather than contracted. In their initial studies
on humans walking, running, cycling (22, 23), and finger tapping (20),
Kirby et al. found relationships between the rates of these activities
and heart rate. However, their subsequent studies (21) showed that the
relationship in tapping was not statistically significant. Furthermore,
later studies during cycling (14) and hopping (19) showed neither a
consistent phase relationship nor increased metabolic efficiency during
episodes when the locomotor and heart rates were matched. Work by
Baudinette et al. (5) on wallabies hopping on a treadmill also found no
correlation between locomotor and cardiac cycles. Thus, although there
are several reasons to suspect that coupling might be advantageous, it
remains an open question as to whether coupling occurs.
If coupling does enhance cardiovascular efficiency, we would expect it
to be most important during vigorous exercise when metabolic demand is
elevated. In an attempt to document locomotor-cardiac coupling in
running mammals, we monitored heart activity and stride events in dogs
as they trotted on a treadmill. If coupling exists, it is likely to be
most apparent in species that have evolved high levels of locomotor
endurance. As a family, canids are known to cover great distances while
foraging.
| |
MATERIALS AND METHODS |
Experimental animals.
Five dogs of mixed breed were used as subjects in this investigation.
Their weights ranged from 17.6 to 30.0 kg, and their mean weight was
24.2 kg. They were purchased at ~1 yr of age from a USDA-licensed
animal dealer. All five were easily trained to walk and run on a
motorized treadmill and were comfortable trotting over a range of
speeds.
Data collection.
To monitor the timing of the heartbeat, we recorded an
electrocardiogram (ECG) signal with surface electrodes instrumented to
an AC amplifier (P511, Grass Instruments, West Warwick, RI). Sites for
the electrodes were shaved. Each electrode was held in place by a belt
of masking tape wrapped around the trunk, and a generous dollop of
electrode gel provided the electrical connection to the subject. One
electrode was positioned on the dorsal midline just caudal to the tips
of the scapulae, and another was placed on the skin over the middle of
the sternum. A third electrode served as a ground and was positioned on
the dorsal midline in the lumbar region. Leads from the electrodes
traveled in a shielded cable to a high-impedance probe (F-HIP511G,
Grass Instruments) that was taped to the side of the treadmill. The ECG
signal was amplified 500 or 1,000 times, filtered below 10 Hz and above
300 Hz, and recorded on a Macintosh computer at 500 samples/s with the
use of an Acknowledge data acquisition board and software (Biopac
Systems, Goleta, CA).
To monitor stride frequency, we recorded the vertical
accelerations of the trunk with an accelerometer (model 104, Omega, Stamford, CT) that was attached to the dog's back. To correlate the
signal from the accelerometer with the phases of limb support, we
recorded the trials with a high-speed video system (Peak Performance Technologies, Englewood, CO) at 120 fields/s. The accelerometer and
video recordings were synchronized with a circuit that triggered the
beginning of data acquisition and simultaneously illuminated a
light-emitting diode in the view of the video camera.
Experimental protocol and analysis.
Recordings were made from the dogs while they ran at a range of
trotting speeds. Specifically, we started each subject running at its
slowest trotting speed, just above the speed of the walk-to-trot transition. The dog was run at this speed for 3 min before the beginning of recording to provide a warm up and to allow the dog to
approach steady state. We then recorded cardiac and locomotor cycles
for 3 min. The speed of the treadmill was then increased by
0.2-0.7 m/s (depending on the dog). The dog was given a 1-min acclimation period at the new speed, and then data were recorded for 3 min. Data were collected in this way for four to five speeds that
covered the entire range of trotting speeds for the dog. Immediately
after the 3-min recording period at the highest speed, the treadmill
was slowed to the lowest speed, and the full protocol was repeated
twice so that 9 min of data were collected at each speed. Thus each dog
was run for a total of 51 or 63 min. In three of the dogs, this
hour-long protocol was repeated on 2 or 3 different days.
We analyzed the raw recordings to determine the phase relationship
between heartbeats and stride. We used Igor software (WaveMetrics, Lake
Oswego, OR) to determine timing of the peak of the R wave in the ECG
signal and the timing of the zero crossing of the accelerometer signal.
These times were then used to calculate the phase angle (0-360°) at which each heartbeat began in the locomotor cycle. We defined the beginning of the heart cycle as the timing of the peak
in the R wave. We also identified the timing of the peak of the T wave
for each cardiac cycle. This allowed us to determine the duration of
the systole (time from R wave to T wave) and the duration of diastole
(time from T wave to R wave). This made it possible to sort out whether
variation in cardiac cycle duration during locomotion was due to
changes in systole, diastole, or both.
Test of the hypothesis of coupling.
Because each heartbeat can be assumed to influence the timing of the
heartbeat that follows it, individual heartbeat times cannot be assumed
to be independent for purposes of statistical analysis. We therefore
designed and carried out a series of randomization tests (25) that
exploit the independence of individual runs (trials) performed by each
dog. On the null hypothesis that there is no coupling between the
cardiac cycle and the locomotor cycle, cardiac cycles from different
runs that are at different running speeds and separated in time by
minutes can certainly be assumed to be independent of each other, even
if individual heartbeats within a given run are not.
We divided the stride cycle into 24 bins and assigned the cardiac
cycles from each run to bins on the basis of their start times. On the
null hypothesis, there should be no tendency for cardiac cycles to
occur disproportionately at any particular point in the cycle, so we
expect them to be distributed evenly (on average) over the 24 bins. We
therefore used the goodness-of-fit
2 statistic
2 = 
(O 
E)2/E as our measure of
association between the locomotor and cardiac cycles [here O is
the observed number of cardiac cycles in a given bin, and E is the
expected (average) number (total/24)]. A sample in which the
cardiac cycles were uniformly distributed among the bins (i.e., no
coupling of cardiac cycle to stride) would give a
2 value near zero, whereas a
sample in which all of the cardiac cycles occurred in one bin (i.e.,
highly coupled) would produce a very large value of
2. We calculated this statistic
for the pooled data from all runs of each dog individually and for the
pooled data from all five dogs together.
These observed values of 2
cannot be tested for significance by comparison to the
2 distribution because we
cannot assume that individual heartbeats are temporally independent of
each other. So to generate the sampling distribution of
2 under an appropriately
conservative null hypothesis, we added a random phase shift to the
entire sequence of heartbeats from each run and repeated the
2 calculation. Within each
simulated run, the sequence of heartbeats remains exactly the same as
in the actual data, but the relationship to the stride cycle is
randomized. On the null hypothesis, this should have no effect on the
average value of 2. But if
there really is a tendency for heartbeats to occur at particular points
in the stride cycle, then the randomization process will distribute
heartbeats more evenly over bins than in the original data, thereby
reducing the calculated 2 in
most cases. For each test, we performed 1,000 such randomizations and
interpreted the proportion of randomized
2 values that were greater than
the observed value of 2 as an
apparent level of statistical significance. Tests were carried out for
all of the recorded runs of each dog separately and then for all of the
runs from all of the dogs together.
| |
RESULTS |
Initial attempts to record an ECG from surface electrodes placed on the
lateral body wall just caudal to the forelimb resulted in very noisy
signals during locomotion. Often the QRS waves were not distinguishable
from the noise. Fortunately, the technique of placing the electrodes on
the dorsal and ventral midlines produced clean and repeatable signals
in which the QRS and T waves were clearly visible (Fig.
1).
View larger version (15K):
[in this window]
[in a new window]
|
Fig. 1.
Sample electrocardiogram (ECG) (A)
and accelerometer signal (B) from
dog trotting at 2.2 m/s. QRS and T waves are marked in the first
cardiac cycle shown. In accelerometer trace, each peak represents
vertical acceleration that occurs during 1 step. Consequently, 2 peaks
represent 1 locomotor cycle. Notice that length of cardiac cycle is
slightly less than length of locomotor cycle.
|
|
In all of the recordings, the period of the cardiac and locomotor
cycles differed, with the heart beating slightly faster than the
locomotor cycle (Fig. 2). This small
difference in frequency was observed in all five dogs and was
maintained over the full range of trotting speeds. The frequencies did
not differ by an integral amount; hence the cardiac cycle drifted with
respect to the phase of the locomotor cycle. This was true in all five dogs and at all speeds that were studied. Consequently, none of the
dogs exhibited phase locking of the locomotor and cardiac cycles (Figs.
3 and 4). If coupling were
to occur at one cardiac cycle per stride, the vast majority of
heartbeats would occur over a single narrow range of phase angles in
Figs. 3 and 4. Similarly, if there were coupling at two cardiac cycles
per stride there would be two peaks of phase angle in which most of the
heartbeats occurred. However, heartbeats occurred at all phases of the
locomotor cycle.
View larger version (17K):
[in this window]
[in a new window]
|
Fig. 2.
Histogram comparing frequency of the heart cycle (open bars) to
frequency of locomotor cycle (solid bars) over a range of speeds in 4 dogs
(A-D)
during trotting. Each bar represents mean and SD of mean frequencies
from 5-9 1-min samples.
|
|
View larger version (50K):
[in this window]
[in a new window]
View larger version (68K):
[in this window]
[in a new window]
|
Fig. 3.
Histograms from 4 dogs
(A-D)
showing number of heartbeats that began in different phase angles of
locomotor cycle. In each case, locomotor cycle begins (zero phase
angle) at end of support by right hindlimb. Running speed (m/s) is
listed in top left corner of each graph. Bottom graph from
A shows data from 2 3-min trials in
which we varied speed of treadmill continuously from 2.0 to 3.5 m/s.
Although heartbeats begin at all phases of locomotor cycle, in 3 of the
dogs shown here, there are speeds in which a few more heartbeats were
found to occur at phase angles of 60-120 and 260-320°:
A at all speeds,
B at 2.2 and 2.8 m/s, and
C at 2.2 m/s.
|
|
View larger version (55K):
[in this window]
[in a new window]
|
Fig. 4.
Histograms for all analyzed trials from each of 5 dogs
(A-E)
showing number of heartbeats that began in different phase angles of
locomotor cycle. In these graphs, all running speeds are lumped
together. In each case, locomotor cycle begins (zero phase angle) at
end of support by right hindlimb. F
shows data from all trials from
A-E:
65 trials and 27,603 heartbeats inclusive.
|
|
Results of the randomization test clearly indicate that tight coupling
does not occur in these dogs (Table 1). If
the heartbeats were uniformly distributed among the 24 bins of the
stride cycle, the randomization test would give a value near zero. If,
however, all of the heartbeats had occurred in a single bin, which
would have indicated very strong coupling, the test statistic would have been >80,000 in every case. The largest
2 value observed was 171, which
was 0.14% of the value that would have occurred if there had been
tight coupling.
View this table:
[in this window]
[in a new window]
|
Table 1.
2 Values for observed relationship between
timing of cardiac cycle relative to locomotor cycle and for
randomization test
|
|
Although none of the dogs exhibited tight coupling of the locomotor and
cardiac cycles, in three of the five dogs we observed a tendency toward
slightly more heartbeats at ~80-120 and 260-300° of the
locomotor cycle (Figs. 3 and 4). This weak relationship was found to be
statistically significant in three of the five dogs (Table 1 and Fig.
4). A significant effect was also found when the test was performed on
all of the data from all five dogs (Table 1). This pattern was most
apparent at intermediate speeds and during the 3-min recording periods
early in the recording session. The pattern was rarely observed late in
the recording sessions. We also observed the pattern in one dog when we
continuously varied the belt speed of the treadmill from 2.0 to 3.5 m/s
(Fig. 3). The timing was such that the QRS waves were slightly more likely to occur just as the diagonal forelimbs and hindlimbs were touching down to begin the support phase of a step in the trotting cycle. Conversely, the QRS waves were slightly less likely to occur
right before the limbs were about to come off the ground, leading into
the flight phase.
To gain a better understanding of what this pattern of slightly more
heartbeats occurring at the beginning of the support phase might mean,
we compared the period of each cardiac cycle to the phase of the
locomotor cycle in which it occurred (Fig. 5). Because more heartbeats occurred at
~100 and 280° of the locomotor cycle, we expected those
heartbeats to be of shorter duration. However, the phase of the
locomotor cycle that had the most heartbeats also tended to have
cardiac cycles of slightly longer duration. Examination of variation in
the duration of systole and diastole showed that changes in the period
of the cardiac cycles were entirely due to changes in the length of
diastole (Fig. 6). The duration of diastole
increased as the period of the cardiac cycle increased, but the length
of systole remained constant. This was further illustrated by plotting
diastole duration against the phase of the locomotor cycle in which the
heartbeat occurred (Fig. 7). Thus the two
phases in the locomotor cycle that tend to have a few more heartbeats
have cardiac cycles of longer duration because of a longer diastole.
View larger version (31K):
[in this window]
[in a new window]
|
Fig. 5.
Histograms plotting average length of cardiac cycles that occurred at
different phase angles of locomotor cycle. Data are from speeds at
which 3 of the dogs exhibited a few more heartbeats at phase angles of
60-120 and 260-320°. Sample sizes:
A, 1,755 heartbeats;
B, 738 heartbeats;
C, 1,122 heartbeats;
D, 840 heartbeats.
|
|
View larger version (13K):
[in this window]
[in a new window]
|
Fig. 6.
Graphs of systole (A) and diastole
duration (B) vs. heartbeat duration
for 1,262 heartbeats from a dog trotting at 2.8 m/s. Least-squared
regression to the power equation yielded
y = 0.018x + 0.151 (r2 = 0.038) for
systole duration and y = 0.980x 0.141 (r2 = 0.988) for
diastole.
|
|
View larger version (45K):
[in this window]
[in a new window]
|
Fig. 7.
Histograms plotting average length of diastole vs. phase angle at which
heartbeats began. Data are presented for 2 dogs. Running speeds were
3.5 (A) and 2.8 (B) m/s. Sample sizes were 1,755 (A) and 738 (B) heartbeats.
|
|
The correlation of the longest heartbeats also being the most frequent
was a function of the heart rate normally being faster than the stride
rate. Because the heart cycle was of shorter duration than the stride
cycle, the cardiac cycle drifted in time relative to the stride cycle,
and it appeared to drift through the stride cycle. Those heartbeats
that were of longer duration were closer in length to the duration of
the stride cycle, and so the drift through the stride cycle was slowed.
If these longer heartbeats were very close to the length of the stride
cycle, the two functions appeared to become coupled for a brief period
of time. This resulted in a greater number of heartbeats occurring at
any phase angle of the stride in which there was a tendency for the
heartbeats to be of slightly longer duration. The heartbeats that
occurred at the beginning of the support phase tended to be longer than heartbeats that occurred at other times in the stride, and this greater
length caused the cardiac and locomotor cycles to become transiently
coupled. Transient coupling slowed the drift of the heart cycle through
the stride cycle and resulted in a few more heartbeats occurring at the
beginning of the support phase.
Figure 8 compares the timing of systole and
diastole with the vertical acceleration of the dog's trunk to
illustrate the phase relationship of the locomotor cycle in which
heartbeats are most likely to occur (Fig.
8A) and the phase relationship in
which heartbeats are least likely to occur (Fig.
8B). Several points warrant mention.
First, the cardiac cycles that occurred during the phase of the
locomotor cycle that had more heartbeats had periods that were
approximately the same length as that of the locomotor cycle. When this
occurred, the two cycles were "transiently coupled" for periods
of 5-20 cycles. Second, the timing of systole in these heartbeats
coincided with the support phase and the peak vertical accelerations,
whereas diastole began at the beginning of the suspension phase and
ended a step and one-half later at the start of the next support phase.
In contrast, during the phase of the locomotor cycle in which
heartbeats were least likely to occur, the periods of the cardiac
cycles were of shorter duration than the locomotor cycles. This
resulted in rapid drift of the cardiac cycle through the locomotor
cycle. Systole during these heartbeats coincided with the suspension
phase and the lowest accelerations, and diastole began early in the
support phase and ended less than a step later at the start of the next
support phase.
View larger version (20K):
[in this window]
[in a new window]
|
Fig. 8.
Timing and duration of systole and diastole plotted relative to ECG
signal and vertical acceleration of trunk for heartbeats that occur
most frequently and are the longest
(A) and for heartbeats that occur
least frequently and are the shortest
(B). In accelerometer trace
(bottom), each large peak, lasting slightly less than 0.2 s, represents
time in which feet are in contact with ground during a single step.
Thus the most frequent heartbeats occur at beginning of support
phase.
|
|
Changes in the length of the cardiac cycle as a result of variation in
the length of diastole were readily apparent in our dogs during rest.
Figure 9 plots an ECG signal relative to a
recording of inspiratory and expiratory airflow in a resting dog. The
heart cycle slowed down during each expiration and then increased in frequency during each inspiration. The longest cardiac periods during
expiration were approximately twice as long as the shortest cardiac
periods during inspiration. This variation resulted from changes in the
length of diastole.
View larger version (19K):
[in this window]
[in a new window]
|
Fig. 9.
Relationship between period of cardiac cycle and phase of lung
ventilation in a resting dog. A: ECG.
B: lung ventilation recorded at nares
with mask-mounted screen pneumotach. Expiration (Expir) occurred when
trace was above baseline, and inspiration (Inspir) occurred when trace
was below baseline. Notice that shortest cardiac cycles occur during
inspiration and that variation in cycle time is primarily because of
differences in length of diastole (time from T wave to R wave).
|
|
| |
DISCUSSION |
Lack of coupling in trotting dogs.
If coupling between the cardiac and locomotor cycles existed,
contraction of the heart would be restricted to one or more phases of
the locomotor cycle, presumably phases that maximized efficient gas
transport. We found no evidence of a consistent temporal correlation
between cardiac cycle and stride. In our recordings, the cardiac cycle
was almost always shorter than the locomotor cycle. Hence the ECG
signal drifted in time relative to the stride. This was true for all
five dogs throughout the full range of speeds in the trot. Thus, under
the conditions of this study, the locomotor and cardiac cycles were not
phase locked. This finding is consistent with a number of studies that
have looked for but not observed locomotor-cardiac coupling (2, 3, 5,
14, 19, 21). If coupling is an important phenomenon, it should occur
consistently, and it would likely be most apparent when dogs run at a
steady pace in uncomplicated terrain, as was the case in this study.
However, because running speed was controlled by us, it is possible
that the dogs did not run at a speed that was conducive to coupling, as
might be the case when they run in the natural world. We believe this
is unlikely because we monitored four or five speeds for each subject
within their trotting range, and we checked each of these speeds three
times over the course of an hour of running, during which time the
level of exertion was constantly increasing. Additionally, in three of
the dogs, the entire protocol was repeated two or three times on
different days. Consequently, this study indicates that the locomotor
and cardiac cycles are not tightly coupled in trotting dogs.
Influence of stride and/or ventilation on the cardiac cycle
during trotting.
Although our results indicate that trotting dogs do not couple their
heart cycle with their stride, the cardiac cycle was at times
influenced by some aspect of locomotion or possibly by ventilation that
was coupled to the stride. In three of the five dogs, there was a
repeatable and statistically significant trend in which a few more
heartbeats coincided with the beginning of the support phase of the
locomotor cycle. The period of these cardiac cycles was slightly longer
as a result of a longer diastole, and this resulted in brief periods in
which the heart became transiently coupled to the stride. Because
diastole was what changed in a regular way relative to the phase of the
locomotor cycle, we suspect that the observed pattern was a function of
stride-related variation in the time it takes to fill the heart with
blood [within limits, changes in the duration of diastole reflect
changes in the rate of ventricular filling (27)]. At rest, the
length of diastole was strongly influenced by lung ventilation (Fig.
9). Inspiration facilitates venous return to the right side of the
heart (15) while decreasing stroke volume from the left side of the
heart (18), and, in our dogs at rest, inspiration was associated with cardiac cycles that were as much as 50% shorter than those during expiration. Such changes in heart rate are thought to be mediated through stretch receptors in and around the heart that elicit an
increase in heart rate through the Bainbridge and baroreceptor reflexes
(4, 16). Consequently, both ventilation coupled to the locomotor cycle
and locomotor forces that influence vascular dynamics could possibly
lead to phases of the stride in which the atria fill relatively rapidly
because of increases in venous return. This would cause stretch
reflex-mediated increases in heart rate that could be expected to
result in the observed relationship between cardiac cycle duration and
phase of the locomotor cycle.
Although the observed variation in diastole during trotting appears to
result from a locomotor influence, sorting out whether this modulation
of cardiac function results from changes in preload or afterload is
difficult. Preload, or the rate of ventricular filling, could possibly
be affected by the locomotor cycle in at least two ways. First,
pressure applied to veins by contracting skeletal muscles, in
combination with one-way venous valves, acts to pump blood toward the
heart. These oscillating skeletal muscle pumps presumably produce
variations in the rate of venous return that would be expected to
correlate with the specific phases of the locomotor cycle. Second,
locomotor- and/or respiratory-induced changes in thoracic and
abdominal pressures may vary the rate of venous return. However,
changes in thoracic pressure can independently influence both
ventricular filling and emptying (24). Consequently, this mechanism
could either vary the rate of venous return and filling of the heart or
vary the degree to which the ventricles empty (ejection fraction) and
therefore the time required to fill the heart during the next cardiac
cycle.
In those cardiac cycles of the shortest duration, diastole began early
in the support phase of a step (Fig.
8B). These diastoles began during
the phase of the locomotor cycle when there was peak loading of the
skeletal muscles responsible for support. Thus these relatively short
diastoles could potentially have been a result of enhanced venous
return from the skeletal muscle pumps. Additionally, the short
diastoles also corresponded to peak inspiratory airflow (9, 11) and
therefore could possibly reflect enhanced venous return resulting from
subatmospheric intrathoracic pressures aspiring blood to the heart from
the vena cava. These mechanisms are not mutually exclusive, and both
could play a role in enhancing venous return and thereby decreasing the
duration of the cardiac cycle.
On the other hand, the cardiac cycles with the longest diastoles may
either be a consequence of reduced venous return or greater ejection
fraction, requiring a longer time for filling during the next cardiac
cycle. Emptying of the heart could be facilitated by increased
contractility of the heart or decreased afterload. In our dogs, systole
of the cardiac cycles of longest duration began early in support and
therefore completely overlapped the time in a step when thoracic
pressure would have been elevated (Fig. 8). Elevation in thoracic
pressure during locomotion can have a significant effect on vascular
dynamics; for example, during galloping in horses thoracic pressures
undergo transient elevations of 20 mmHg during each stride (26). Such
elevations in thoracic pressures are known to reduce afterload and
facilitate blood flow to the extremities (reviewed in Ref. 18).
Consequently, the longer cardiac cycles observed in the dogs may
possibly have been a consequence of a greater ejection fraction, which
necessitated a longer period of filling during the next cardiac cycle
and led to a longer diastole.
The question remains as to whether the pattern of longer cardiac cycles
occurring at the beginning of limb support and the resulting transient
coupling is of any functional significance. It is possible that this
pattern reflects homeostatic reflexes of the heart that do not confer
any improved energetic efficiency in blood transport during locomotion.
Alternatively, if the increased thoracic pressure during support aids
pumping of blood to the periphery, then the longer cardiac cycles that
are associated with the beginning of support may represent an
energetically advantageous way of pumping blood. Simultaneous
recordings of the cardiac cycle (ECG) and cardiac output in trotting
dogs may help resolve this question.
This influence of locomotion on the cardiac cycle (i.e., longer cardiac
cycles at the beginning of limb support) was observed relatively
rarely. It was most apparent at the beginning of the exercise protocol.
The absence of an interaction that occurs all the time may be due to
variation in the physiological state of the running dogs. If the
interaction is a function of venous return, then factors such as blood
volume, blood viscosity, and sympathetic stimulation, which are
important determinants of venous return, may play a role in whether the
interaction occurs. For example, the heart is sensitive to changes in
right atrial pressures over a discrete range; if factors such as a high
blood volume or high sympathetic stimulation to the veins elevate the
preload beyond this range, then the heart might be insensitive to
changes in venous return caused by variation in the respiratory or
locomotor pumps, and the interaction would be less apparent.
Would strict locomotor-cardiac coupling be advantageous?
As outlined above, there are a number of reasons why locomotor-cardiac
coupling might enhance the efficiency of blood delivery, and there are
several reports of coupling in the literature. Consequently, the
absence of consistent coupling in trotting dogs came as a surprise.
However, one factor that may make strict coupling between the cardiac
and locomotor systems impractical, and perhaps even disadvantageous, is
the variation in venous return from the skeletal muscle pumps and
respiratory pump that can be expected to occur on a stride-by-stride
basis when animals run in the natural world. If the amount of blood
returning to the heart varies from stride to stride, then a strict
coupling of locomotor and cardiac cycles would result in variable
stroke volumes and possibly ineffective pumping by the heart.
Venous return during locomotion can be expected to vary because animals
move through a complex, three-dimensional landscape. Running animals
are constantly forced to change direction and speed, avoid obstacles,
and adjust to the angle of the terrain. Their speed and direction of
travel are also influenced by conspecifics, predators, and prey. Thus,
in a natural setting, stride rate, intramuscular pressures in each
limb, and locomotor influences on thoracic pressure are all likely to
vary with each step. Under these circumstances, venous return from the
skeletal muscle and respiratory pumps and ejection fraction from the
ventricles will likely be as variable as the strides the animals take.
Effective cardiovascular function during periods when external
variables are strongly influencing vascular dynamics may require that
the cardiac pump follow the lead of the skeletal muscle and respiratory pumps. For example, it would likely be advantageous for heart rate to
respond if venous return suddenly slowed or increased. Such a change in
the cardiac cycle would be accomplished through the reflexes that
control the rate and contractility of the heart (stretch of the sinus
node, Bainbridge reflex, baroreceptor reflex, and Frank-Starling
reflex). These reflexes enable modulation of both heart rate and stroke
volume so that the cardiac pump maintains homeostasis. Hence the reflex
capacity to vary heart rate and stroke volume may enable the cardiac
pump to compensate for fluctuations in the circulation of blood caused
by locomotion. We suspect that this more subtle and complex coupling of
cardiac function to the locomotor system is likely to be more effective
in the production of efficient circulation than would be possible with
a strict entrainment of the cardiac and locomotor cycles.
Perspectives
The results of this investigation suggest that dogs do not phase lock
their locomotor and cardiac cycles during trotting. We hypothesize that
variation in venous return and ventricular ejection during running in a
three-dimensional, complex environment may make strict
locomotor-cardiac coupling impractical. Although there does not appear
to be tight coupling, the locomotor and/or ventilatory cycles
appear to influence the cardiac cycle. Our data showed a repeatable
pattern of cardiac cycles of longer duration occurring at the beginning
of the support phase of a step rather than at other times in the step
cycle. Because the length of diastole is what varies relative to the
phase of the step cycle, we believe that this pattern is a function of
the time required to fill the heart. It is unclear whether the pattern
reflects variation in the rate of venous return or if it reflects
greater emptying of the ventricles in the previous cardiac cycle. In
either case, these data suggest that stretch receptors in and around
the heart are modulating heart rate on a beat-by-beat basis. Variations in venous return and ejection fractions that are likely to result from
running in a complex environment may make the reflexes that control the
rate and contractility of the heart more important to homeostasis
during locomotion than during rest.
| |
ACKNOWLEDGEMENTS |
We owe special thanks to John Seger, who devised and ran the
statistical test reported in this study, without whose assistance we
could not have done this analysis. We are grateful to a number of
people who helped work with the dogs: Peter Nassar, Natalie Silverton,
Chris Gaydos, Erika Mitchell, and Christine Huo. Dennis Bramble and Lee
Kirby provided comments critical to our thinking on this subject.
| |
FOOTNOTES |
This research was supported by The National Science Foundation Grants
IBN-9258243 and IBN-9306466.
Address for reprint requests: D. R. Carrier, Dept. of Biology, 201 Biology Bldg., Univ. of Utah, Salt Lake City, UT 84112.
Received 4 March 1996; accepted in final form 12 June 1997.
| |
REFERENCES |
1.
Attenburrow, D. P.
Time relationship between the respiratory cycle and limb cycle in the horse.
Equine Vet. J.
14:
69-72,
1982[Medline].
2.
Aulie, A.
Co-ordination between the activity of the heart and the flight muscles during flight in small birds.
Comp. Biochem. Physiol. A Physiol.
38A:
91-97,
1970.
3.
Aulie, A.
Co-ordination between the activity of the heart and the pectoral muscles during flight in the pigeon.
Comp. Biochem. Physiol. A Physiol.
41A:
43-48,
1971.
4.
Bainbridge, F. A.
The influence of venous filling upon the rate of the heart.
J. Physiol. (Lond.)
50:
65-84,
1915.
5.
Baudinette, R. V.,
B. J. Gannon,
W. B. Runciman,
and
S. Wells.
Do cardiorespiratory frequencies show entrainment with hopping in the Tammar wallaby?
J. Exp. Biol.
129:
251-263,
1987[Abstract/Free Full Text].
6.
Berger, M.,
O. Z. Roy,
and
J. S. Hart.
The co-ordination between respiration and wing beats in birds.
Z. V. Gl. Physiol.
66:
190-200,
1970.
7.
Blomqvist, C. G.,
and
B. Saltin.
Cardiovascular adaptations to physical training.
Annu. Rev. Physiol.
45:
169-189,
1983[Medline].
8.
Bramble, D. M.,
and
D. R. Carrier.
Running and breathing in mammals.
Science
219:
251-256,
1983[Abstract/Free Full Text].
9.
Bramble, D. M.,
and
F. A. Jenkins.
Mammalian locomotor-respiratory integration: implications for diaphragmatic and pulmonary design.
Science
262:
235-240,
1994.
10.
Butler, P. J.
Respiration during flight and diving in birds.
In: Exogenous and Endogenous Influences in Metabolic and Neural Control, edited by A. D. F. Addink,
and N. Spronk. Oxford, UK: Pergamon, 1982, p. 103-114.
11.
Carrier, D. R.
Function of the intercostal muscles in trotting dogs: ventilation or locomotion?
J. Exp. Biol.
199:
1455-1465,
1996[Abstract].
12.
Coleman, W. M.
The psychological significance of bodily rhythms.
J. Comp. Psychol.
1:
213-220,
1921.
13.
Coleman, W. M.
On the correlation of the rate of heart beat, breathing, bodily movement and sensory stimuli.
J. Physiol. (Lond.)
54:
213-217,
1921.
14.
Donville, J. E.,
R. L. Kirby,
T. J. Doherty,
S. K. Gupta,
B. J. Eastwood,
and
D. A. Macleod.
Effect of cardiac-locomotor coupling on the metabolic efficiency of pedaling.
Can. J. Appl. Physiol.
18:
379-391,
1993[Medline].
15.
Guyton, A. C.,
and
J. E. Hall.
Medical Physiology. Philadelphia, PA: Saunders, 1994, p. 1148.
16.
Hakumaki, M. O. K.
Seventy years of the Bainbridge reflex.
Acta Physiol. Scand.
130:
177-185,
1987[Medline].
17.
Hornicke, H.,
R. M. Meixner,
and
U. Pollmann.
Respiration in exercising horses.
In: Equine Exercise Physiology, edited by D. H. Snow,
S. G. B. Persson,
and R. J. Rose. Cambridge, UK: Burlington, 1983, p. 7-16.
18.
Janicki, J. S.,
D. D. Sheriff,
J. L. Robotham,
and
R. A. Wise.
Cardiac output during exercise: contributions of the cardiac, circulatory, and respiratory systems.
In: Handbook of Physiology. Exercise: Regulation and Integration of Multiple Systems. Bethesda, MD: Am. Physiol. Soc., 1996, sect. 12, chapt. 15, p. 649-704.
19.
Kirby, R. L.,
S. M. Atkinson,
J. E. Donville,
M. F. Urdang,
D. A. Stanley,
S. K. Gupta,
and
D. A. Macleod.
Failure of accentuated vertical body movements to induce cardiac-locomotor coupling.
J. Appl. Physiol.
72:
1368-1374,
1992[Abstract/Free Full Text].
20.
Kirby, R. L.,
S. E. Carr,
and
D. A. Macleod.
Cardiac-locomotor coupling while finger tapping.
Percept. Mot. Skills
71:
1099-1104,
1990[Medline].
21.
Kirby, R. L.,
S. K. Gupta,
S. E. Carr,
and
D. A. Macleod.
Cardiac-locomotor coupling while finger tapping: part II. A cross-over control study.
Percept. Mot. Skills
73:
831-834,
1991[Medline].
22.
Kirby, R. L.,
D. A. Macleod,
and
A. E. Marble.
Coupling between cardiac and locomotor rhythms: the phase lag between heart beats and pedal thrusts.
Angiology
40:
620-625,
1989.
23.
Kirby, R. L.,
S. T. Nugent,
R. W. Marlow,
D. A. Macleod,
and
A. E. Marble.
Coupling of cardiac and locomotor rhythms.
J. Appl. Physiol.
66:
323-329,
1989[Abstract/Free Full Text].
24.
Peters, J.,
C. Fraser,
R. S. Stuart,
W. Baumgartner,
and
J. L. Robotham.
Negative intrathoracic pressure decreases independently left ventricular filling and emptying.
Am. J. Physiol.
257 (Heart Circ. Physiol. 26):
H120-H131,
1989[Abstract/Free Full Text].
25.
Sokal, R. R.,
and
F. J. Rohlf.
Biometry. New York: Freeman, 1981.
26.
Sinha, A. K.,
R. D. Gleed,
T. S. Hakim,
A. Dobson,
and
K. J. Shannon.
Pulmonary capillary pressure during exercise in horses.
J. Appl. Physiol.
80:
1792-1798,
1996[Abstract/Free Full Text].
27.
Weisdorf, D.,
and
D. H. Spodick.
Duration of diastole versus cycle length as correlates of left ventricular ejection time.
Br. Heart J.
38:
282-284,
1976[Abstract/Free Full Text].
28.
Young, I. S.,
R. McN. Alexander,
A. J. Woakes,
P. J. Butler,
and
L. Anderson.
The synchronization of ventilation and locomotion in horses (Equus caballus).
J. Exp. Biol.
166:
19-31,
1992[Abstract/Free Full Text].
AJP Regul Integr Compar Physiol 273(4):R1352-R1360
0363-6119/97 $5.00
Copyright © 1997 the American Physiological Society
Top
Abstract
Introduction
