Locomotor-respiratory coupling (LRC), phase-locking between breathing and stepping rhythms, occurs in many vertebrates. When quadrupedal mammals gallop, 1∶1 stride per breath coupling is necessitated by pronounced mechanical interactions between locomotion and ventilation. Humans show more flexibility in breathing patterns during locomotion, using LRC ratios of 2∶1, 2.5∶1, 3∶1, or 4∶1 and sometimes no coupling. Previous studies provide conflicting evidence on the mechanical significance of LRC in running humans. Some studies suggest LRC improves breathing efficiency, but others suggest LRC is mechanically insignificant because ‘step-driven flows’ (ventilatory flows attributable to step-induced forces) contribute a negligible fraction of tidal volume. Yet, although step-driven flows are brief, they cause large fluctuations in ventilatory flow. Here we test the hypothesis that running humans use LRC to minimize antagonistic effects of step-driven flows on breathing. We measured locomotor-ventilatory dynamics in 14 subjects running at a self-selected speed (2.6±0.1 ms−1) and compared breathing dynamics in their naturally ‘preferred’ and ‘avoided’ entrainment patterns. Step-driven flows occurred at 1-2X step frequency with peak magnitudes of 0.97±0.45 Ls−1 (mean ±S.D). Step-driven flows varied depending on ventilatory state (high versus low lung volume), suggesting state-dependent changes in compliance and damping of thoraco-abdominal tissues. Subjects naturally preferred LRC patterns that minimized antagonistic interactions and aligned ventilatory transitions with assistive phases of the step. Ventilatory transitions initiated in ‘preferred’ phases within the step cycle occurred 2x faster than those in ‘avoided’ phases. We hypothesize that humans coordinate breathing and locomotion to minimize antagonistic loading of respiratory muscles, reduce work of breathing and minimize rate of fatigue. Future work could address the potential consequences of locomotor-ventilatory interactions for elite endurance athletes and individuals who are overweight or obese, populations in which respiratory muscle fatigue can be limiting.
Citation: Daley MA, Bramble DM, Carrier DR (2013) Impact Loading and Locomotor-Respiratory Coordination Significantly Influence Breathing Dynamics in Running Humans. PLoS ONE 8(8): e70752. https://doi.org/10.1371/journal.pone.0070752
Editor: François Hug, The University of Queensland, Australia
Received: March 4, 2013; Accepted: June 28, 2013; Published: August 12, 2013
Copyright: © 2013 Daley et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: The authors have no support or funding to report.
Competing interests: The authors declare that coauthor David Carrier is a PLOS ONE Editorial Board member. This does not alter the authors' adherence to all the PLOS ONE policies on sharing data and materials.
Effective ventilation is essential for sustained animal locomotion. For many animals, including birds and mammals, this requires integrating movement and breathing so that inspiration and expiration occur during mechanically compatible periods of the locomotor cycle. Several direct mechanical links between locomotion and ventilation necessitate integration , , , . Firstly, sagittal bending of the trunk assists forward progression during locomotion in quadrupeds, and creates a ‘bellows’ effect, altering the pressure and volume of the abdomen and thorax. Additionally, impact loads induce inertial motions of soft-tissues (viscera, adipose), creating a ‘visceral piston’ effect, pulling and pushing on the diaphragm and body wall muscles (abdominals, intercostals) and altering thoraco-abdominal pressures. Finally, many axial muscles of terrestrial vertebrates contribute to both breathing and locomotion , , , , , . Indeed, many ‘ventilatory muscles’ cease respiratory action and become entrained with the locomotor cycle during running in lizards , , birds  and dogs , . As a result of these factors, active inspiration is most compatible with a specific and different phase of the locomotor cycle than active expiration.
Locomotor-respiratory coupling (LRC) refers to phase locking of running and breathing so that the same number of steps occur during each breath, and has been observed in numerous vertebrates, including birds, dogs, hares, horses, wallabies and humans , , , , . LRC is a form of entrainment, in which the two rhythmic activities with different frequencies become phase-locked due to mechanical and neural interactions. LRC has been suggested to have a number of important physiological effects. These include reducing the energy cost of breathing , , , , minimizing conflict in muscles that contribute to both functions , , body stabilization during motion , , and enabling trunk bending and inertial movements of soft-tissues to augment pumping of air in and out of the lungs , .
Whereas most galloping mammals exhibit 1∶1 (strides/breath) coupling, humans demonstrate more flexibility in breathing patterns during locomotion. Humans frequently use LRC ratios of 2∶1, 2.5∶1, 3∶1, or 4∶1 and sometimes lack entrainment altogether, using independent breathing and stepping frequencies , , , , . Variance in LRC pattern is observed both within and across individuals , , , , . This flexibility in human breathing patterns relative to locomotion most likely relates to an upright posture and striding bipedal gait. Unlike quadrupeds, the sagittal bending of the back does not assist forward progression and the forelimbs are not subjected to direct weight-bearing and impact loading. The forces transmitted through the thorax, abdomen and rib cage are consequently smaller, reducing mechanical interactions between locomotion and ventilation. Ground birds, which are also bipedal, exhibit similar flexibility as humans in locomotor-ventilatory coupling patterns , . Bipedal posture may reduce the mechanical and neuromuscular conflicts that constrain quadrupeds to a 1∶1 LRC pattern , .
Does this greater flexibility suggest that breathing and running are mechanically independent in humans? The answer remains controversial. Humans do often exhibit sustained ventilatory entrainment to locomotor forces , , . Furthermore, evidence suggests coupling improves breathing efficiency in some circumstances , , . On the other hand, humans also couple breathing frequency with low impact, non-locomotor movements such as finger and arm tracking , . Thus, large mechanical interactions are not required to induce rhythmic coupling. Theoretically, human LRC during running could arise entirely from neural interactions.
The overall ventilatory effect of locomotor-ventilatory interactions likely depends on a number of neural and mechanical factors. The primary mechanical factors in mammalian locomotor-ventilatory interactions are thoracic loading and inertial displacement of the soft-tissues associated with body accelerations , , , , , . In humans, impact loading of the body at footstrike is likely to cause compression of the thorax and downward displacement of the abdominal viscera. Arm swinging is also likely to exert a compression load on the thorax during early to mid-stance . These mechanical actions can alter respiratory flow by changing thoraco-abdominal volumes and pressures. Inertial displacement of the abdominal viscera would tend to pull or push on the respiratory diaphragm through its direct attachment to the liver. The relative magnitude of the effects on the rib cage and diaphragm likely depends on the mass of the abdominal organs, the distribution of other soft-tissues, and the vertical accelerations of the body. Body wall muscles including the intercostals and abdominals could act to resist these effects , . Additionally, the rhythmic activity of intercostal and abdominal muscles is entrained by afferents from the arms and legs, facilitating neural coupling , . The ventilatory outcome of these neuro-mechanical interactions likely depends on the relative timing and magnitude of the mechanical factors (thorax compression and visceral displacement) and the rhythmic activation of body wall muscles.
To address whether locomotor-ventilatory interactions are mechanically significant in running humans, Banzett and colleagues  measured ‘step-driven flow’ in humans running on a treadmill. ‘Step-driven flow’ is ventilatory flow and resulting volumes attributable to locomotor-induced forces, once the primary ventilatory pattern is removed (through either ensemble averaging or filtering – see methods). Banzett and colleagues  found the volumes attributable to step-driven flows to be a negligible fraction of tidal volume, ∼1–2% tidal volume per step. Consequently, they concluded that locomotion has no meaningful mechanical influence on breathing in humans . Following publication of this finding, there has been relatively little research on the mechanical consequences of locomotor-ventilatory interactions in running humans.
Yet, there are reasons to revisit the potential mechanical significance of locomotor-ventilatory interactions in humans. In the study by Banzett and colleagues, the subjects lightly rested their hands on side rails while jogging on the treadmill . Arm swinging contributes to compression loads on the thorax during locomotion, and to control of body and head rotation , . The restriction of natural arm motion may have reduced mechanical interactions between locomotion and ventilation. Furthermore, although the volumes attributed to step-driven flows amounted to only 1–2% of tidal volume per step, Banzett and colleagues measured large transient flows associated with footstrike . These step-driven flows lasted only a small fraction of the ventilatory cycle, but exhibited peak magnitudes of 1 Ls−1. Although brief, step-driven flows reach a large fraction of concurrent ventilatory flow, sometimes resulting in a brief mid-breath reversal in flow (from inspiration to expiration or vice-versa). We suggest that these transient flows have potential to influence breathing dynamics and ventilatory muscle loading without directly driving large ventilatory volumes.
Here we test the hypothesis that humans use LRC to minimize antagonistic effects of step-driven flows on breathing dynamics. Due the natural variability in coupling patterns in running humans , , , , , it is possible to compare breathing dynamics when individuals use their own ‘preferred’ and ‘avoided’ entrainment patterns. We analyze locomotor-ventilatory interactions in fourteen physically fit subjects running on a treadmill at a moderate self-selected speed (2.6±0.1 ms−1) with natural arm motion. Self-selected speed is likely to be close to energetically optimal speed . Running at speeds above preferred speeds would increase both impact loads and ventilatory demand, increasing the potential for antagonistic locomotor-ventilatory interactions. We therefore consider self-selected speed to provide a conservative test of the hypothesis that humans use LRC to minimize antagonistic locomotor-ventilatory dynamics.
Materials and Methods
The Institutional Review Board of University of Utah approved the protocols and informed consent documents, under IRB #12361, and all subjects gave written informed consent.
Subjects and protocol
Fourteen adult subjects volunteered for this study (9 females 5 males; age 36±2 years, mass 65.1±2.8 kg, height 1.72±0.02 m). Seven of the subjects were recreational runners who regularly ran at least 20 miles per week, and seven were physically fit non-runners (Table 1). This sampling was designed to obtain a broad representation of human locomotor-ventilatory patterns, not to statistically compare trained and non-trained runners, which has been previously studied by others , . Participants ran at a steady speed on a treadmill, using natural arm motion. After an initial practice session, subjects were allowed to select their preferred speed, with the instructions to run at a comfortable pace that they could sustain for 30 minutes. They warmed up for 5 minutes and then continued to run for 5 minutes while we recorded running and breathing patterns.
We tracked the step cycle by measuring acceleration at the top of the head using an Endevco microtron accelerometer (model 7290A-10). The subjects wore a lightweight facemask equipped with two-way non-rebreathing valves to measure inspiratory and expiratory flow separately (Hans Rudoph Inc., Kansas City, MO, USA). Each valve was connected by tubing (34.2 mm ID; Hans Rudolph, Inc.) to screen pneumotachs (model 4813, Hans Rudolph Inc.), with a calibrated flow range of 0–800 liters per minute. The pressure ports of the screen pneumotachs were attached to a differential pressure transducer (Omega Engineering, Stamford CT, USA). To prevent motion artifact, the tubing and pneumotachs were secured to a small backpack worn by the subject, fitted with both waist and shoulder straps. Flow was calibrated by recording the output voltage from a known volume of air, using a 3 liter calibration syringe (Hans Rudolph Inc.).
The signals were amplified 10x (model P122, Grass Inc., West Warwick RI, USA), sampled at 200 Hz and digitally recorded onto computer using the MP100 data acquisition system with AcqKnowledge software (BIOPAC Systems Inc., Goleta, CA, USA.). Before analysis, data were low pass filtered at 30 Hz using a zero-phase 6th order Butterworth filter in Matlab (butter and filtfilt functions; Matlab release 13). These filter settings were selected because they were found to effectively remove noise while retaining higher frequency signal characteristics such as peaks associated with footstrike impact.
Calculation of average step-driven flow
To enable comparison to previous findings, we calculate average step-driven flow using similar analytical methods of Banzett and colleagues . We calculated an ensemble average (EA) of ventilatory flow relative to the step cycle for each individual, including at least 50 steps per subject . The EA technique samples flow segments distributed equally through the ventilatory cycle, to quantify step-driven flow independent of ventilatory phase. We calculated this EA for the 12 out of 14 subjects who exhibited varying phase of entrainment between running and breathing cycles (Table 1). The resulting EA trace of step-driven flow was integrated over time to obtain the locomotor driven volume (LDV). Expiratory (positive) and inspiratory (negative) volumes were integrated separately to assess the relative magnitude of each. The LDV was normalized by the total concurrent ventilatory volume (Vtot) over the same time periods as the EA flow segments, to obtain a value as a percentage of concurrent ventilatory flow. Note that our calculation of LDV differs slightly from that used by Banzett and colleagues, who reported the fractional contribution to tidal volume per step; whereas we report the total fractional contribution to ventilation. Humans typically use coupling ratios of 2∶1 and higher (e.g. , , , , ), taking multiple steps in each breath cycle; consequently, a per step value does not provide a complete account of the total magnitude of locomotor-driven ventilation.
A potential source of error in the calculated LDV is flow artifact caused by motion of the tubes during running. As mentioned earlier, we minimized this by securing the tubes to a lightweight backpack worn by the subject. To quantify the error associated with any remaining motion, we calculated an ‘artifact LDV’ for 5 subjects. We recorded inspiration and expiration in separate channels, causing each channel to have a baseline of zero during the opposing ventilatory half-cycle. During this period, the only flow measured in that channel was small magnitude, high frequency artifact flow due to tube motion. To quantify this, we calculated a value of LDV using the same method described above, except that the flow segments in the ensemble average were obtained from the baseline of the inactive ventilation channel. This method resulted in an artifact LDV value of 0.7±0.1% of Vtot (mean ± s.e.m.) during control running trials, with a worst case artifact magnitude of 1% Vtot.
Testing for variation in step-driven flow with ventilatory phase
The visceral piston hypothesis predicts that the effect of locomotor forces on flow depends on the phase of the breath cycle in which the step occurs  (Fig. 1C). Deceleration of the body at footstrike likely has two primary mechanical effects: 1) compression of the thorax, and 2) viscoelastic bounce and rebound of the guts , . Contraction of respiratory muscles and inflation of the lungs differentially alters thoracic and abdominal stiffness . Locomotor forces are likely to have an enhanced expiratory effect if they occur when the abdominal muscles are being recruited for forced expiration, because these muscles resist the descent of the guts (Fig. 1C).
(A–B) Typical head acceleration (top) and ventilatory flow (bottom: expiration positive, inspiration negative) during quiet standing (A), and moderate speed treadmill running (B). Note the high frequency oscillations in ventilatory flow during running. (C) Schematic illustration of the ‘visceral-piston’ model for human locomotor-ventilatory interactions. Red arrows indicate muscle actions during inspiration and expiration.
The ensemble average technique cannot reveal this effect of ventilatory phase because it averages step-driven flow across all phases. Therefore, we also compared step-driven ventilation at different phases of the breathing cycle, in bins of 20% of maximum tidal volume (Vmax) during inspiration and expiration. From the ensemble averaged data, we observed that step related oscillations in flow occurred at step frequency and the 1st and 2nd harmonic of step frequency. We high-pass filtered the flow signal to remove the primary breathing pattern, so that only flows associated with the step frequency or higher remained in the signal. We used a zero-phase digital Butterworth filter (‘butter’ and ‘filtfilt’ functions in Matlab version 6.5), with the cutoff frequency and filter order specified using the Matlab function ‘buttord’ to create a filter that lost no more than 1% of the signal greater than or equal to step frequency and attenuated 99% of the signal below half the step frequency. We then calculated an average ± s.e.m of step-driven flow for all steps occurring within each volume bin of the ventilatory cycle.
Analysis of entrainment patterns
We used phase analysis to examine the locomotor-ventilatory coupling patterns. The timing of each inspiratory and expiratory transition was calculated relative to the step cycle. The beginning of footstrike was defined as zero, and the phase of each ventilatory transition was expressed in degrees between 0 and 360 or as a fraction of the step cycle by dividing the phase angle by 360. The step cycle was divided into 20 bins (18 degrees or 5% of the step cycle each), and the frequency of inspiratory and expiratory transitions was calculated for each bin. If breaths and steps occur randomly with respect to each other, the distribution of ventilatory transition events should not significantly differ from a uniform circular distribution. For each subject whose ventilatory transitions significantly differed from a uniform circular distribution (see Statistics), we determined the ‘preferred’ phase for inspiratory and expiratory initiation as the bin in which each of these events occurred most frequently. Likewise, we determined ‘avoided’ phase as the bin in which each transition event occurred least frequently.
Ventilatory transition times
To quantify the time required for transition between ventilatory half cycles (inspiration and expiration), we measured the time between 50% peak flows (T50). The expiratory T50 is the time from 50% peak inspiration to 50% peak expiration. Likewise, the inspiratory T50 is the same measure between 50% peak expiration and 50% peak inspiration. We compared T50 values between ‘preferred’ and ‘avoided’ phase bins, described above. If multiple bins were tied for ‘preferred’ or ‘avoided’, we took the average of the tied bins.
Kuiper’s test of circular uniformity was used to test whether inspiratory and expiratory initiation events were distributed randomly with respect to the step cycle. For subjects with ventilatory transitions distributed non-uniformly relative to the step cycle, a Friedman nonparametric repeated measures ANOVA test was used to compare ventilatory transition times (T50) between preferred and avoided phases of step cycle, with posthoc pairwise comparisons using Dunn’s Multiple Comparisons.
Step-driven oscillations in flow
All subjects exhibited high frequency, step-driven oscillations in flow (Fig. 1B). An inspiratory pulse occurred immediately after footstrike, followed by an expiratory pulse as the body approached peak acceleration (Fig. 2). The associated inspiratory and expiratory volumes amounted to −12.7±4.5% (mean ±S.D) and 10.7±3.2%, respectively, of the concurrent ventilatory volume (Fig. 2; Table 2), with peak flow magnitude averaging 0.97±0.45 Ls−1 across individuals. In some cases, these step-driven flows were large enough to cause transient reversals in flow (Fig. 3A).
(A) Grand mean of step-driven ventilatory flow across subjects (mean ± 95% confidence interval for 12 subjects, see also individual examples in Fig. 5). Expiration is positive. (B) Mean ± SD of step-driven volume as a percentage of total concurrent ventilatory volume (Vtot), during level moderate speed running (N = 12, black dots show data from individuals). Step-driven flow and volume data included here are from subjects with variation in the phase locking between steps and breaths (Table 1).
(A) Variation in the magnitude of step-driven flows is apparent, particularly during low frequency breathing, as shown here during 3∶1 (strides per breath) coupling. Note the reversal in flow in late expiration (asterisk). Dashed vertical lines indicate footstrike events (data from subject 3). (B) Average step-driven flow (means ±95%CI) for a representative subjective at four points in the ventilatory cycle: 1) early expiration, high lung volume (90% Vmax), 2) late expiration, low lung volume (10% Vmax), 3) early inspiration, low lung volume (10% Vmax) and 4) late inspiration, high lung volume (90% Vmax).
Locomotor-ventilatory interactions and entrainment
Although the step-driven inspiratory and expiratory volumes were equal on average (Fig. 2), the effect of locomotor acceleration on flow varied considerably with ventilatory phase (Fig. 3). Across individuals, we observed a significant relationship between tidal volume at footstrike and the net step-driven volume (Fig. 4). Steps occurring in late expiration (low lung volume) tend to have a net inspiratory effect. Similarly, steps occurring in late inspiration (high lung volume) tend to have a net expiratory effect. Thus, the net effect of step-driven flow is synergistic early in the ventilatory half-cycle and antagonistic late in each ventilatory half cycle. Flow reversals occur most often in late expiration during slow, deep breaths and coupling with ratios greater than 2.5∶1 strides per breath (Fig. 3A). However, most runners avoided these flow reversals because they breathed more frequently, closer to a 2∶1 stride per breath rhythm.
We show net step-driven volumes (in milliliters per step) as a function of the tidal volume at the time of footstrike during inspiration (left) and expiration (right), averaged across individuals (mean ± 95% CI, N = 12). At low lung volumes, the inspiratory pulse at footstrike is larger, leading to a net inspiratory effect. At high lung volumes, a larger expiratory pulse occurs, leading to a net expiratory effect.
Runners varied in whether or how tightly they entrained breathing and stepping rhythm (Table 1). Most (11 of 14) subjects exhibited periods of fixed-ratio coupling between stepping and breathing rhythms, with 2∶1 strides per breath as the most common pattern. Two of these subjects exhibited strict phase-locked coupling with a single coupling ratio, whereas the remaining 9 subjects exhibited varied phase or switching among multiple coupling ratios (Table 1). Despite variation in coupling, 13 of 14 subjects had significantly non-uniform distributions of ventilatory transitions relative to the step cycle (inspiratory, expiratory or both; Table 3). Preferred transition phases tended to correspond to regions of the step cycle that mechanically assisted ventilation, or at least, did not impede it (Fig. 5).
The distribution of ventilatory transitions relative to the step cycle is non-uniform (Table 3). Here we show the net bias in ventilatory transitions relative to step cycle (left axis and bars), with transitions to expiration as positive, transition to inspiration as negative. The ‘net bias’ value is the number of expiratory transitions minus the number of inspiratory transitions in each phase bin. Step-driven flow is overlaid for reference (right axis and lines). Data from 3 individuals illustrates typical variation between strongly coupled (A), variably coupled (B) and uncoupled (C) subjects. Although variation exists, the timing of transitions is clearly non-random, and exhibits some correspondence to the step-driven flow.
Although step-driven flows are brief, the phasing of transitions relative to the step cycle does substantially influence breathing dynamics. Whether coupled or not, most runners avoided reversals in flow by using a breathing frequency near a 2∶1 stride per breath ratio and timing ventilatory transitions to coincide favorably with step-driven flow (Fig. 5). Furthermore, the phasing between steps and breath transitions influenced the time required for ventilatory transitions. When runners coordinate the transitions to occur in phases of the step cycle that promote flow, transitions occur 2X more rapidly (Fig. 6). When compared across all uncoupled or variably coupled runners (for which ‘avoided’ transition data were available), ventilatory transitions in ‘preferred’ phase relationships with the step cycle occurred more rapidly than those in ‘avoided’ phases (Fig. 6). The difference between preferred and avoided transition T50 values was statistically significant (p = 0.0002 for Freidman Test, p<0.05 Dunn’s Multiple Comparisons).
(A) Examples of ventilatory transitions timed with preferred and assistive phases of the step cycle, facilitating rapid transitions, and (B) timed with an avoided and antagonistic phases of the step cycle. Dashed vertical lines indicate zero-crossings of ventilatory transitions. The transition durations (T50) are calculated between the black dots indicating the times of 50% peak flows. (C) Transition times averaged across individuals (mean ± s.e.m), comparing breaths at ‘avoided’ and ‘preferred’ phase relationships. Preferred refers to the most used phase bin (e.g., see Fig. 5), and ‘avoided’ refers to the least used phase bin that was represented in the data.
The goal of this study was to examine whether the mechanical interaction between running and breathing in humans is large enough to be physiologically important. We measured the ventilatory flows and volumes attributable to step-driven flows during running at a moderate self-selected speed. The timing of step-driven flows is consistent with the visceral piston hypothesis for locomotor-ventilatory interactions in humans , , , , . Step-driven ventilation averaged around 10–12% of total ventilatory volume with flow magnitudes around 1 Ls−1. We found that step-driven flow dynamics varied depending on ventilatory state (high versus low lung volume), suggesting phase-dependent changes in compliance and damping of thoraco-abdominal tissues. We also discovered that the timing of impact loading relative to the ventilatory cycle significantly influenced breathing dynamics. Ventilatory transitions initiated in preferred (assistive) phases of the step cycle occurred 2x faster than those in avoided (antagonistic) phases. These findings suggest a physiologically significant mechanical interaction exists between locomotion and ventilation in humans.
The step-driven ventilatory volumes we measured appear to be larger on average than those reported previously . Here, step-driven volumes amounted to 10–12% of total ventilation, which amounts to 2.5–3.0% tidal volume per step during a 2∶1 step per breath rhythm. Banzett and colleagues reported a value of 1–2% tidal volume per step. Nonetheless, the timing of peak step-driven flow was consistent with Banzett and colleagues. The difference in magnitude could be explained by a difference in arm motion between the two studies. In the earlier study, the subjects lightly rested their hands on side rails, whereas in the current study the subjects were allowed to move their arms naturally while running. Both step-induced thoracic loading and inertial displacement of abdominal viscera likely contribute to locomotor-ventilatory interactions while running , , , , . Minimizing the effects of arm loading on thoracic compression during running could reduce both the mechanical and neural interactions between locomotion and ventilation , , , . Nonetheless, we do find a similar pattern, and although the total step-driven volumes are small on a per step basis, they do have a significant mechanical influence on breathing dynamics.
What is the physiological significance of locomotor-ventilatory entrainment in humans?
The frequency of step related flows is too high relative to breath frequency to allow direct coupling to drive ventilation (Figs. 1, 2). Yet, our data reveal that runners prefer to time ventilatory transitions to periods in the step cycle that assist rather than impede flow (Fig. 5), and this timing significantly reduces the time required for transitions (Fig. 6). These data suggest that step-driven flows have potential to influence the work of respiratory muscles, because they significantly influence breathing dynamics, particularly at ventilatory transitions (expiration to inspiration and vice versa). Appropriate coordination of stepping and breathing rhythm may act to minimize antagonistic loading of the respiratory muscles caused by motions of the abdomen and chest wall. Although locomotor-ventilatory coupling is more flexible in humans than in quadrupedal mammals , , , , , , coupling in humans may reduce conflicting demands placed on the diaphragm and body wall muscles (abdominal muscles and intercostals).
The 2∶1 LRC pattern preferred by human runners may also reflect optimization to minimize antagonistic loading of respiratory muscles. Humans prefer a 2∶1 ratio across a wide range of sustainable running speeds (e.g. , , , , ). A recent study by O’Holloran and colleagues found the highest flexibility in LRC patterns at preferred stride frequency, when energy cost is minimal. At stride frequencies above or below preferred, energy demand sharply increases and subjects exhibit a stronger preference for 2∶1 coupling . Here, we find evidence that a 2∶1 pattern could minimize work of respiratory muscles by allowing one footstrike to assist the ventilatory transition and the second to occur at intermediate lung volumes (see Fig. 6A). At intermediate lung volumes, the net effect of step-driven flows is either assistive or neutral (Fig. 4). However, near the end of each ventilatory half-cycle, approaching the extrema of lung volume, step-driven flows become antagonistic (Fig. 4). Higher coupling ratios require footstrikes near the end of each ventilatory half-cycle, leading to antagonistic loading of the respiratory system and transient reversals in flow (Fig. 3). Thus, a 2∶1 stride per breath rhythm might be more strongly preferred during intense endurance running, when fatigue of respiratory muscles could be limiting.
We hypothesize that human runners benefit from locomotor-ventilatory entrainment by reducing the work of ventilatory muscles, and minimizing fatigue of respiratory muscles that are critical to endurance aerobic activity. Appropriate tuning of locomotor-ventilatory interactions likely minimizes antagonistic loading of ventilatory muscles and allows inertial displacement of the guts to passively assist the action of respiratory muscles. These mechanical interactions likely have the greatest impact on diaphragm performance because the abdominal viscera directly attach to this muscle. Work of breathing increases as a squared function of ventilatory demand, and breathing may account for up to 10–15% of energy demand in intense exercise , . Although the properties of the mammalian diaphragm appear to confer considerable resistance to fatigue , some evidence suggests that it can be vulnerable to fatigue during prolonged or intense exercise in both sedentary and fit individuals , , , , , , . Declines in respiratory function following marathon and ultra-marathon competitions also suggest respiratory muscle fatigue , , . Respiratory muscle fatigue may be a limiting factor in human endurance activity, and locomotor-ventilatory coupling has potential to minimize fatigue, especially during activities that involve impact loading with each footstrike, such as walking and running.
Unfortunately, it remains challenging to directly test this hypothesis. Ventilatory muscles are a relatively small fraction of the metabolically active tissue in the body, making it difficult to measure changes in respiratory muscle work using standard respirometry techniques. Though technically challenging, it may be possible to use recordings of muscle activity (e.g., ) to examine the response of ventilatory muscles to manipulations of visceral load and locomotor-ventilatory coordination during running.
In future work, it would be interesting to investigate whether antagonistic locomotor-ventilatory interactions could explain, in part, why obese individuals experience ‘breathlessness’ and rapid fatigue during locomotion. Obese individuals face a dual problem of increased energy cost and impaired respiratory function during locomotion. Obese individuals incur a 10–25% higher metabolic energy cost of walking per kilogram body mass compared to people with healthy weight , ; meaning that carrying fat costs more than carrying lean weight. The source of this added cost remains controversial. External work does not explain the added cost, because the gait of obese individuals involves similar total mechanical work on their body center of mass . Increased internal work associated bouncing soft tissues and increased costs for stabilizing the body and joints may contribute to increased locomotor costs due to excess fat mass , , , , , . Furthermore, excess fat is associated with reduced operating lung volumes and decreased respiratory compliance, factors that increase the work of breathing , . Soft-tissue bouncing induced by impact loads during walking and running may further increase the work of breathing and the rate of respiratory muscle fatigue in overweight and obese individuals. Whereas in lean individuals a large fraction of soft-tissue mass is concentrated intra-abdominally, obesity results in a large fraction of soft-tissue mass distributed external to the body wall muscles. Thus, high adiposity will increase soft-tissue bouncing and reduce the ability to actively tune soft-tissue dynamics through abdominal muscle contraction. Larger, uncontrolled soft-tissue motions may result in antagonistic locomotor-ventilatory interactions and higher mechanical work of breathing during exercise. We predict that the problems of obesity are exacerbated by antagonistic locomotor-ventilatory interactions during walking and running.
Biomechanics studies often focus on musculoskeletal and biomechanical factors as limits to performance. For example, leg muscle strength is widely thought to limit top running speed during sprinting , . However, during endurance locomotion, respiratory muscles, not leg muscles, may limit maximum exercise intensity and duration . We suggest that when assessing the physiological importance of locomotor-respiratory coupling in humans, it may be short sighted to place too much emphasis on the comparatively small volumes attributed to it. Entrainment might serve a number of beneficial physiological functions – i.e., reducing the work of ventilatory muscles, preventing respiratory muscle fatigue, and improving respiratory efficiency through enhanced gas mixing, transport and exchange. Consequently, the precise volume driven by locomotion might be less important than the associated intrapulmonary dynamics. Humans are well adapted for economic walking and endurance running , , , . Multiple derived characters suggest aerobic endurance as a key evolutionary pressure in the human lineage, including skeletal morphology, developed tendon springs and enhanced heat dissipation , , , . We will not be surprised, therefore, if the unusual locomotor-respiratory coupling patterns exhibited by human runners proves to be another example of evolutionary adaptation in support of the exceptional endurance running capabilities of humans , .
We thank Eric Stakebake for assistance in the experiments, Franz Goller for loan of equipment and the late Farish Jenkins Jr. for helping to make the project possible. We also thank Joanne Gordon for her comments on the manuscript draft.
Conceived and designed the experiments: MAD DRC DMB. Performed the experiments: MAD. Analyzed the data: MAD. Wrote the paper: MAD. Interpreted the data and revised the manuscript: DMB DRC.
- 1. Alexander RM (1989) On the synchronization of breathing with running in wallabies (Macropus Spp) and horses (Equus-Caballus). Journal of Zoology 218: 69–85.
- 2. Bramble DM, Carrier DR (1983) Running and breathing in mammals. Science 219: 251–256.
- 3. Boggs DF (2002) Interactions between locomotion and ventilation in tetrapods. Comparative Biochemistry and Physiology A-Molecular and Integrative Physiology 133: 269–288.
- 4. Giuliodori MJ, Lujan HL, Briggs WS, DiCarlo SE (2009) A model of locomotor-respiratory coupling in quadrupeds. Advances in Physiology Education 33: 315–318.
- 5. Carrier DR (1990) Activity of the hypaxial muscles during walking in the lizard. Journal of Experimental Biology 152: 453–470.
- 6. Carrier DR (1991) Conflict in the hypaxial musculoskeletal system – Documenting an evolutionary constraint. American Zoologist 31: 644–654.
- 7. Codd JR, Boggs DF, Perry SF, Carrier DR (2005) Activity of three muscles associated with the uncinate processes of the giant canada goose Branta canadensis maximus. Journal of Experimental Biology 208: 849–857.
- 8. Carrier DR (1996) Function of the intercostal muscles in trotting dogs: Ventilation or locomotion? Journal of Experimental Biology 199: 1455–1465.
- 9. Deban SM, Carrier DR (2002) Hypaxial muscle activity during running and breathing in dogs. Journal of Experimental Biology 205: 1953–1967.
- 10. Hodges PW, Gandevia SC (2000) Changes in intra-abdominal pressure during postural and respiratory activation of the human diaphragm. Journal of Applied Physiology 89: 967–976.
- 11. Simons RS (1999) Running, breathing and visceral motion in the domestic rabbit (Oryctolagus cuniculus): Testing visceral displacement hypotheses. Journal of Experimental Biology 202: 563–577.
- 12. Baudinette RV, Gannon BJ, Runciman WB, Wells S, Love JB (1987) Do cardiorespiratory frequencies show entrainment with hopping in the tammar wallaby? Journal of Experimental Biology 129: 251–263.
- 13. Funk GD, Sholomenko GN, Valenzuela IJ, Steeves JD, Milsom WK (1993) Coordination of wing beat and respiration in canada geese during free flight. Journal of Experimental Biology 175: 317–323.
- 14. Bernasconi P, Kohl J (1993) Analysis of coordination between breathing and exercise rhythms in man. Journal of Physiology-London 471: 693–706.
- 15. Bonsignore MR, Morici G, Abate P, Romano S, Bonsignore G (1998) Ventilation and entrainment of breathing during cycling and running in triathletes. Medicine and Science in Sports and Exercise 30: 239–245.
- 16. Funk GD, Valenzuela IJ, Milsom WK (1997) Energetic consequences of coordinating wingbeat and respiratory rhythms in birds. Journal of Experimental Biology 200: 915–920.
- 17. Nassar PN, Jackson AC, Carrier DR (2001) Entraining the natural frequencies of running and breathing in guinea fowl (Numida meleagris). Journal of Experimental Biology 204: 1641–1651.
- 18. Simons RS (1996) Lung morphology of cursorial and non-cursorial mammals: Lagomorphs as a case study for a pneumatic stabilization hypothesis. Journal of Morphology 230: 299–316.
- 19. Bramble DM (1989) Axial-appendicular dynamics and the integration of breathing and gait in mammals. American Zoologist 29: 171–186.
- 20. Banzett RB, Mead J, Reid MB, Topulos GP (1992) Locomotion in men has no appreciable mechanical effect on breathing. Journal of Applied Physiology 72: 1922–1926.
- 21. Persegol L, Jordan M, Viala D (1991) Evidence for the entrainment of breathing by locomotor pattern in human. Journal De Physiologie 85: 38–43.
- 22. O'Halloran J, Hamill J, McDermott WJ, Remelius JG, Van Emmerik REA (2012) Locomotor-respiratory coupling patterns and oxygen consumption during walking above and below preferred stride frequency. European Journal of Applied Physiology 112: 929–940.
- 23. McDermott W, Emmerik RA, Hamill J (2003) Running training and adaptive strategies of locomotor-respiratory coordination. European Journal of Applied Physiology 89: 435–444.
- 24. Bernasconi P, Burki P, Buhrer A, Koller EA, Kohl J (1995) Running training and coordination between breathing and running rhythms during aerobic and anaerobic donditions in humans. European Journal of Applied Physiology and Occupational Physiology 70: 387–393.
- 25. Ebert D, Rassler B, Hefter H (2000) Coordination between breathing and forearm movements during sinusoidal tracking. European Journal of Applied Physiology 81: 288–296.
- 26. Rassler B, Ebert D, Waurick S, Junghans R (1996) Coordination between breathing and finger tracking in man. Journal of Motor Behavior 28: 48–56.
- 27. Bramble DM, Jenkins FA (1993) Mammalian locomotor-respiratory integration – Implications for diaphragmatic and pulmonary design. Science 262: 235–240.
- 28. Lee HT, Banzett RB (1997) Mechanical links between locomotion and breathing: Can you breathe with your legs? News in Physiological Sciences 12: 273–278.
- 29. Pontzer H, Holloway JH, Raichlen DA, Lieberman DE (2009) Control and function of arm swing in human walking and running. Journal of Experimental Biology 212: 523–534.
- 30. Loring SH (1992) Action of human respiratory muscles inferred from finite element analysis of rib cage. Journal of Applied Physiology 72: 1461–1465.
- 31. Loring SH, De Troyer A (1985) Actions of respiratory muscles. In: Macklem PT, editor. The Thorax. New York: M. Dekker. 327–349.
- 32. Giraudin A, Cabirol-Pol M-J, Simmers J, Morin D (2008) Intercostal and abdominal respiratory motoneurons in the neonatal rat spinal cord: Spatiotemporal organization and responses to limb afferent stimulation. Journal of Neurophysiology 99: 2626–2640.
- 33. Giraudin A, Le Bon-Jego M, Cabirol M-J, Simmers J, Morin D (2012) Spinal and pontine relay pathways mediating respiratory rhythm entrainment by limb proprioceptive inputs in the neonatal rat. Journal of Neuroscience 32: 11841–11853.
- 34. Pontzer H, Raichlen D, Sockol M (2009) The metabolic cost of walking in humans, chimpanzees, and early hominins. Journal of Human Evolution 56: 43–54.
- 35. Steudel-Numbers KL, Wall-Scheffler CM (2009) Optimal running speed and the evolution of hominin hunting strategies. Journal of Human Evolution 56: 355–360.
- 36. DeTroyer A (1991) Respiratory Muscles. In: West JB, editor. The Lung: Scientific Foundations. New York: Raven Press, Ltd. 869–883.
- 37. Fabre N, Perrey S, Arbez L, Rouillon J-D (2007) Neuro-mechanical and chemical influences on locomotor respiratory coupling in humans. Respiratory Physiology & Neurobiology 155: 128–136.
- 38. Aaron EA, Seow KC, Johnson BD, Dempsey JA (1992) Oxygen cost of exercise hyperpnea: implications for performance. Journal of Applied Physiology 72: 1818–1825.
- 39. Dominelli PB, Sheel AW (2012) Experimental approaches to the study of the mechanics of breathing during exercise. Respiratory Physiology & Neurobiology 180: 147–161.
- 40. Dempsey JA, Adams L, Ainsworth DM, Fregori RF, Gallagher CG, et al.. (1996) Airway, lung, and respiratory muscle function during exercise. In: Shepard JT, editor. Handbook of Physiology, Sect 12, Exercise: Regulation and Integration of Multiple Systems: Oxford Univ. Press. 448–514.
- 41. Johnson BD, Aaron EA, Babcock MA, Dempsey JA (1996) Respiratory muscle fatigue during exercise: Implications for performance. Medicine and Science in Sports and Exercise 28: 1129–1137.
- 42. Babcock MA, Pegelow DF, Johnson BD, Dempsey JA (1996) Aerobic fitness effects on exercise-induced low-frequency diaphragm fatigue. Journal of Applied Physiology 81: 2156–2164.
- 43. Mador MJ, Magalang UJ, Rodis A, Kufel TJ (1993) Diaphragmatic fatigue after exercise in healthy-human subjects. American Review of Respiratory Disease 148: 1571–1575.
- 44. Johnson BD, Babcock MA, Suman OE, Dempsey JA (1993) Exercise-induced diaphragmatic fatigue in healthy humans. Journal of Physiology-London 460: 385–405.
- 45. Babcock MA, Pegelow DF, Harms CA, Dempsey JA (2002) Effects of respiratory muscle unloading on exercise-induced diaphragm fatigue. Journal of Applied Physiology 93: 201–206.
- 46. Romer LM, McConnell AK, Jones DA (2002) Inspiratory muscle fatigue in trained cyclists: Effects of inspiratory muscle training. Medicine & Science in Sports & Exercise 34: 785–792.
- 47. Romer LM, Polkey MI (2008) Exercise-induced respiratory muscle fatigue: Implications for performance. Journal of Applied Physiology 104: 879–888.
- 48. Warren G, Cureton KJ, Sparling PB (1989) Does lung function limit performance in a 24-hour ultramarathon? Respiration Physiology 78: 253–2264.
- 49. Mahler DA, Loke JD (1981) Lung function after marathon running at warm and cold ambient temperatures. American Review Respiratory Disease 124: 154–157.
- 50. Loke JD, Virgulto JA (1982) Respiratory muscle fatigue after marathon running. Journal Applied Physiology 52: 821–824.
- 51. Saunders SW, Rath D, Hodges PW (2004) Postural and respiratory activation of the trunk muscles changes with mode and speed of locomotion. Gait & Posture 20: 280–290.
- 52. Browning RC, Baker EA, Herron JA, Kram R (2006) Effects of obesity and sex on the energetic cost and preferred speed of walking. Journal of Applied Physiology 100: 390–398.
- 53. Peyrot N, Thivel D, Isacco L, Morin J-B, Duche P, et al. (2009) Do mechanical gait parameters explain the higher metabolic cost of walking in obese adolescents? Journal of Applied Physiology 106: 1763–1770.
- 54. Browning RC, McGowan CP, Kram R (2009) Obesity does not increase external mechanical work per kilogram body mass during walking. Journal of Biomechanics 42: 2273–2278.
- 55. Zelik KE, Kuo AD (2010) Human walking isn't all hard work: Evidence of soft tissue contributions to energy dissipation and return. Journal of Experimental Biology 213: 4257–4264.
- 56. DeVita P, Helseth J, Hortobagyi T (2007) Muscles do more positive than negative work in human locomotion. Journal of Experimental Biology 210: 3361–3373.
- 57. Daley MA, Usherwood JR (2010) Two explanations for the compliant running paradox: reduced work of bouncing viscera and increased stability in uneven terrain. Biology Letters 6: 418–421.
- 58. Soo CH, Donelan JM (2010) Mechanics and energetics of step-to-step transitions isolated from human walking. Journal of Experimental Biology 213: 4265–4271.
- 59. Dreher M, Kabitz H-J (2012) Impact of obesity on exercise performance and pulmonary rehabilitation. Respirology 17: 899–907.
- 60. Villiot-Danger J-C, Villiot-Danger E, Borel J-C, Pépin J-L, Wuyam B, et al. (2011) Respiratory muscle endurance training in obese patients. International Journal of Obesity 35: 692–699.
- 61. Usherwood JR, Wilson AM (2006) Accounting for elite indoor 200 m sprint results. Biology Letters 2: 47–50.
- 62. Weyand PG, Sternlight DB, Bellizzi MJ, Wright S (2000) Faster top running speeds are achieved with greater ground forces not more rapid leg movements. Journal of Applied Physiology 89: 1991–1999.
- 63. Carrier DR (1984) The energetic paradox of human running and Hominid evolution. Current Anthropology 25: 483–495.
- 64. Bramble DM, Lieberman DE (2004) Endurance running and the evolution of Homo. Nature 432: 345–352.
- 65. Liebenberg L (2008) The relevance of persistence hunting to human evolution. Journal of Human Evolution 55: 1156–1159.
- 66. Raichlen DA, Armstrong H, Lieberman DE (2011) Calcaneus length determines running economy: implications for endurance running performance in modern humans and Neandertals. Journal of Human Evolution 60: 299–308.