Skip to main content
NCBI home page
The Journal of Physiology logoLink to The Journal of Physiology
. 2001 Jan 15;530(Pt 2):319–330. doi: 10.1111/j.1469-7793.2001.0319l.x

Respiratory effects of the external and internal intercostal muscles in humans

Theodore A Wilson *, Alexandre Legrand †,‡, Pierre-Alain Gevenois §, André De Troyer †,‡
PMCID: PMC2278403  PMID: 11208979

Abstract

  1. The current conventional view of intercostal muscle actions is based on the theory of Hamberger (1749) and maintains that as a result of the orientation of the muscle fibres, the external intercostals have an inspiratory action on the lung and the internal interosseous intercostals have an expiratory action. Recent studies in dogs, however, have shown that this notion is only approximate.

  2. In the present studies, the respiratory actions of the human external and internal intercostal muscles were evaluated by applying the Maxwell reciprocity theorem. Thus the orientation of the muscle fibres relative to the ribs and the masses of the muscles were first assessed in cadavers. Five healthy individuals were then placed in a computed tomographic scanner to determine the geometry of the ribs and their precise transformation during passive inflation to total lung capacity. The fractional changes in length of lines with the orientation of the muscle fibres were then computed to obtain the mechanical advantages of the muscles. These values were finally multiplied by muscle mass and maximum active stress (3.0 kg cm−2) to evaluate the potential effects of the muscles on the lung.

  3. The external intercostal in the dorsal half of the second interspace was found to have a large inspiratory effect. However, this effect decreases rapidly in the caudal direction, in particular in the ventral portion of the ribcage. As a result, it is reversed into an expiratory effect in the ventral half of the sixth and eighth interspaces.

  4. The internal intercostals in the ventral half of the sixth and eighth interspaces have a large expiratory effect, but this effect decreases dorsally and cranially.

  5. The total pressure generated by all the external intercostals during a maximum contraction would be -15 cmH2O, and that generated by all the internal interosseous intercostals would be +40 cmH2O. These pressure changes are substantially greater than those induced by the parasternal intercostal and triangularis sterni muscles, respectively.

Although it is now well established that the interchondral portions of the internal intercostal muscles (the so-called parasternal intercostals) elevate the ribs and inflate the lung when they contract (De Troyer & Kelly, 1982; De Troyer et al. 1996, 1998), the actions of the external intercostals and the interosseous portion of the internal intercostals remain uncertain, in particular in humans. Indeed, these muscles are inaccessible and cannot be activated selectively.

Recent theoretical studies (Wilson & De Troyer, 1992, 1993), however, have suggested that this question could be answered by an indirect approach, based on the Maxwell reciprocity theorem. When applied to the respiratory system, this standard theorem of mechanics predicts that the respiratory effect of a particular muscle (that is, the change in airway pressure (ΔPao) produced by the muscle during a maximal isolated contraction against a closed airway) is related to the mass (m) of the muscle, the maximal active muscle tension per unit cross-sectional area (σ), and the fractional change in muscle length (ΔL/L) per unit volume increase of the relaxed chest wall (ΔVL)Rel, such that

graphic file with name tjp0530-0319-m1.jpg 1

For a machine, such as a lever, ‘mechanical advantage’ is defined as the ratio of the force delivered at the load to the force applied at the handle. By analogy, the mechanical advantage of a respiratory muscle may therefore be defined as ΔPao/mσ and, according to eqn (1), could be evaluated by measuring (ΔL/(LΔVL))Rel. In other words, a muscle that shortens during passive inflation would have an inspiratory mechanical advantage and would cause a fall in Pao when it contracts. Conversely, a muscle that lengthens during passive inflation would have an expiratory mechanical advantage and would cause a rise in Pao during contraction.

The validity of this equation has been tested experimentally on a number of canine inspiratory muscles, including the parasternal intercostals, and on the triangularis sterni (De Troyer et al. 1996; Legrand et al. 1996, 1997; De Troyer & Legrand, 1998). For all these muscles, a unique relationship between ΔPao/m and (ΔL/(LΔVL))Rel was obtained. Futhermore, the coefficient of proportionality (σ) between these two variables was 3.0 kg cm−2, in close agreement with values of maximal active muscle tension measured in vitro (Close, 1972; Farkas et al. 1985; Farkas, 1991). These observations therefore confirmed eqn (1) in all respects, and this implies that the respiratory effect of any muscle can be estimated simply by measuring its mass and its fractional change in length during passive inflation.

Based on this principle, we have recently evaluated the mechanical advantage and respiratory effect of the external and internal intercostal muscles in dogs (De Troyer et al. 1999). Whereas the conventional view maintains that as a result of the orientation of the muscle fibres, the external intercostals have an inspiratory effect on the lung and the internal intercostals have an expiratory effect (Hamberger, 1749), the muscles were found to show marked topographic differences. Specifically, the canine external intercostals in the dorsal third of the rostral interspaces have a large inspiratory effect, but this effect decreases rapidly toward the costochondral junctions and toward the base of the ribcage. This decrease is such that the muscles in the ventral portion of the caudal interspaces have an expiratory, rather than inspiratory effect. The internal intercostals in the dorsal portion of the caudal interspaces have a large expiratory effect, but this effect decreases ventrally and cranially such that in the most rostral interspaces, it is reversed into an inspiratory effect (De Troyer et al. 1999).

The present studies were undertaken on the basis of the same principle to evaluate the respiratory effects of the external and internal interosseous intercostal muscles in humans. The orientations of the muscle fibres relative to the ribs were first measured in cadavers; the mass of the muscles was also measured. The shape of the ribs and their transformation during passive inflation were subsequently assessed in a group of healthy individuals. From these data, we then computed the fractional changes in length of lines having the orientation of external and internal intercostal muscles. With the values thus computed, the values of muscle mass measured in cadavers, and the value of σ for maximum stress, we finally calculated the maximal ΔPao values for the muscles throughout the ribcage.

METHODS

Experiment 1

The orientations and mass distributions of the external and internal intercostal muscles were assessed in five cadavers selected from the pool of human bodies in the Department of Anatomy of the Brussels School of Medicine. The selection was made on the basis of three criteria: absence of overt malnutrition, absence of obesity or other thoracic deformity, and absence of macroscopic pulmonary emphysema. The five subjects were elderly (> 65-70 years).

In each subject, the ribcage and intercostal muscles were exposed on both sides of the chest from the first to the tenth rib by deflection of the skin and underlying muscle layers. The length of the second intercostal space was measured bilaterally from the angle of the ribs dorsally to the costochondral junctions ventrally and muscle bundles situated in the centres of the dorsal half and the ventral half of the interspace were selected. For each bundle, a protractor was then laid along the upper edge of the third rib, and the acute angle between the rib and the muscle fibres of the external intercostal was measured. The process was repeated for the fourth, sixth and eighth interspaces, after which the external intercostal muscles in the dorsal and ventral halves of the four interspaces studied were harvested on both sides of the chest and weighed. The obtuse angles between the upper edge of ribs 3, 5, 7 and 9 and the direction of the muscle fibres of the internal intercostal were subsequently measured, and the muscles in the dorsal and ventral portions of each interspace were harvested as well. The parasternal intercostal muscles in interspaces 2, 3, 4 and 5 were also harvested bilaterally in each subject.

Experiment 2

Five healthy subjects (three men, 29-45 years old) were then studied to assess the detailed shape of the bony ribs and to determine their transformation during passive inflation. The subjects were respiratory physicians and gave informed consent to the procedures; these conformed with the Declaration of Helsinki and were approved by the Ethics Committee of the Brussels School of Medicine. The subjects were all non-smokers and had normal pulmonary function tests, and their inspiratory capacity in the supine position averaged 3.5 l (range: 2.1-4.9 l). Three subjects had previously participated in many respiratory experiments and were highly trained in relaxing their respiratory muscles at different lung volumes, but two subjects had little prior experience as respiratory subjects. Before the study, these two subjects therefore underwent several practice sessions with pairs of respiratory magnetometers (Norman H. Peterson, Boston, MA, USA) placed on the abdomen and ribcage, during which they were coached to relax their respiratory muscles. At the time of the study, the five subjects were consequently able to produce consistent relaxation curves of the chest wall from resting end-expiration (functional residual capacity, FRC) to total lung capacity (TLC).

On the day of the study, the subject was placed in a computed tomographic (CT) scanner (Somaton Plus 4A, Siemens Medical System, Erlangen, Germany), and all restrictive garments were removed. Once positioned, the subject hyperventilated for 10-20 s and held his or her breath at FRC for 25 s, at which time a spiral data acquisition starting at the cranial border of the manubrium sterni and extending 10 cm caudally was performed. The scanning parameters were: 140 kV, 206 mA, 1.0 s per revolution scanning time, 3 mm collimation, and 4 mm s−1 table feed. The subject repeated this manoeuvre two more times and two more data acquisitions were performed, each extending 10 cm from the caudal margin of the previous one. After this procedure was completed, the subject breathed in up to TLC and relaxed the respiratory muscles with the glottis closed. A second set of spiral data acquisition was performed at this stage.

Transverse CT scans at FRC and TLC were reconstructed at 2.5 mm intervals using a 360 deg linear interpolation algorithm and a standard kernel. In each subject, a total of 111 successive transverse images were thus obtained at each lung volume, and on these reconstructed images, points on the caudal edge of the even-numbered ribs and on the cranial edge of the odd-numbered ribs were identified on both sides of the chest. The x (dorsoventral) and y (lateral) co-ordinates of these points were recorded with a cursor; these co-ordinates, plus the slice number of the image, thus provided the three-dimensional co-ordinates of points on the edges of the ribs making up interspaces 2, 4, 6 and 8. Typically, the co-ordinates of 40-50 points were obtained for each of the shorter bony ribs (2, 3, 8 and 9) and 70-80 points were obtained for the longer ribs 4-7. As shown in Fig. 1, the data points for all ribs in each subject were then aligned in a fixed co-ordinate system in which the z axis was oriented along the rostrocaudal direction, the x-z plane corresponded with the sagittal midplane, and the x-y plane corresponded with the transverse plane.

Figure 1. Example of rib data obtained from the CT scans.

Figure 1

Open in a new tab

Lateral (left panel) and ventrodorsal (right panel) views of the data points for ribs 2-9 in the left hemithorax of one subject at FRC. These points correspond to the caudal edge of the even-numbered ribs and the cranial edge of the odd-numbered ribs; they indicate, therefore, the boundaries of interspaces 2, 4, 6 and 8. The x axis lies in the sagittal mid-plane, the y axis lies in the coronal plane, and the z axis is oriented in the rostrocaudal direction.

Computation of the changes in muscle length

The data points thus obtained at FRC and TLC do not provide any information on the position of the same material points at the two lung volumes. Therefore, they do not allow the changes in intercostal muscle length to be calculated directly. The problem of identifying the same material points in the ribs at the two lung volumes was resolved by determining the precise transformation that was required to align the image of each rib at TLC with its image at FRC. This transformation was made in two stages.

First, the orientation of the ribs at FRC and TLC was assessed by using the method previously described by Wilson et al. (1987). Thus, a plane of the form:

graphic file with name tjp0530-0319-m2.jpg 2

was fitted to each rib at the two lung volumes, and the coefficients A, B and C were determined by a linear regression analysis. Consequently, the data for each rib were transformed to a new co-ordinate system defined by the mutually perpendicular ξ, η and ζ axes, as shown in Fig. 2. The ξ axis is the line of intersection between the plane of the rib and the sagittal midplane and forms an angle with the x axis; this angle, denoted α, is conventionally referred to as the ‘pump-handle’ angle of the rib. The η axis also lies in the plane of the rib but is oriented laterally, while the ζ axis is perpendicular to the plane of the rib. The η axis forms with the y axis the angle β, which is conventionally called the ‘bucket-handle’ angle of the rib. Once the coefficients A, B and C in eqn (2) were determined, the angles α and β for each rib at FRC and TLC were thus derived from the relationships

Figure 2. Reference co-ordinate system used to assess the orientation of the ribs.

Figure 2

Open in a new tab

A plane (shaded area) was fitted to the data points in the cranial or caudal edge of the ribs, and its orientation relative to the transverse (x-y) plane was described by the angles α (pump-handle angle) and β (bucket-handle angle). By convention, for the rib shown in this figure, the values of α and β are negative.

graphic file with name tjp0530-0319-m3.jpg 3

By convention, angle α was therefore negative when the plane of the rib sloped ventrally and caudally, and angle β was negative when the plane of the rib was oriented laterally and caudally.

The location of the rib in the plane at TLC, however, may be different from that in the plane at FRC. Therefore, in a second stage, the data obtained for each rib in the ξ-η plane were fitted by an ellipse of the form:

graphic file with name tjp0530-0319-m4.jpg 4

where ξ and η are the transformed x and y axes in the plane of the rib, ξo and ηo are the co-ordinates of the centre of the ellipse, and R1 and R2 are the values of the principal radii of the ellipse (Fig. 3). In the fit to the data at FRC, ηo was set equal to zero, and the values of ξo, R1 and R2 that gave the best fit were determined. In the fit to the data at TLC, the values of R1 and R2 were held at the FRC values, and the values of ξo and ηo that gave the best fit were determined. That is, the shape of the ellipse was held constant, but its position in the plane was adjusted.

Figure 3. Description of the arc of the rib at FRC.

Figure 3

Open in a new tab

An ellipse was fitted to the data points in the ξ-η plane, and it was described by its dorsoventral radius (R1), its lateral radius (R2), and its centre in the sagittal midplane (ξo).

A representative example of the fits of planes and ellipses to the data points for one rib is shown in Fig. 4. In Fig. 4A, the data for FRC and TLC are shown viewed along the η axes of the planes of the rib at the two lung volumes. In most cases, points near the neck of the rib lay above the plane of the bulk of the data and in many cases, a few points near the ventral end of the rib lay below the plane; these points were omitted from the fits. However, the data corresponding to the middle arc of the rib were well-fitted by a plane. The root mean square (r.m.s.) deviation of these data from the plane was similar for all ribs and both lung volumes and averaged 0.08 cm. In Fig. 4B, the data points for this rib are viewed along the two ζ axes corresponding to FRC and TLC. In all cases, the data projected onto the plane deviated from the ellipse at the dorsal end, and in some cases, particularly for the second rib, the data deviated from the ellipse at the ventral end; these points were omitted from the fits of the ellipses. The r.m.s. distance in the plane between the points that were fitted and the ellipse, averaged over all ribs, was 0.09 cm at FRC and 0.11 cm at TLC.

Figure 4. Example of plane and ellipse fitting of the ribs.

Figure 4

Open in a new tab

In A, the data points for the left sixth rib in one subject at FRC (•) and TLC (•) have been fitted to two planes (viewed along the η axes); the twelve points on the left (corresponding to the neck of the rib) and the six points on the right (corresponding to the ventral end of the bony rib) were omitted from the fits. In B, the same data points (viewed along the ζ axes) have been fitted to the arc of an ellipse (continuous lines). The centres of the ellipse at FRC and TLC are marked by the closed and open squares, respectively.

With these fits, each rib at FRC was therefore characterized by the values of the parameters A, α and β that describe the plane of the rib and the parameters ξo, R1 and R2 that describe the position and shape of the rib in that plane. The position of the rib at TLC was similarly characterized by the values A, α, β, ξo and ηo, and these results were then used to compute the fractional changes in muscle length by the following procedure. In each subject, on both the right and left sides, two points situated at the centres of the dorsal and ventral halves of the ellipse representing the cranial edge of the third rib at FRC were selected. For each point, the tangent to the ellipse was computed, and the angles between this tangent and a series of lines extending from this point to the ellipse representing the caudal edge of the second rib were computed as well. The two points on the second rib for which the computed angles matched the angles measured for the external and internal intercostal muscles in cadavers (Experiment 1) were identified, and the distances between the selected point and these two points were calculated. The distances between these points in their positions at TLC were also calculated, and the fractional changes in length were obtained from the differences between the distances at FRC and TLC. The procedure was then repeated for the fourth, sixth and eighth interspaces. As noted above, the data at the dorsal and ventral ends of the ribs were omitted from the fits. However, the points at which muscle lengths were calculated were within the range of the data that was fitted.

We were concerned that the angles between the ribs and the intercostal muscle fibres at FRC might be different from those measured in cadavers. The latter had some degree of rigor mortis. More importantly, they were significantly older than our healthy subjects, and it is well known that lung recoil pressure in man decreases with increasing age, leading to an increase in FRC (Turner et al. 1968; Gibson et al. 1976). Therefore, in agreement with our previous measurements of the angles between the sternum and the costal cartilages (De Troyer et al. 1998), the ribs in cadavers may have been in a more cranial position, such that both the acute angles between the ribs and the external intercostals and the obtuse angles between the ribs and the internal intercostals would be greater. We have evaluated the effect of this factor on the computed changes in muscle length during passive inflation by assuming that the ribs in cadavers had the same orientation as in our healthy subjects at TLC. With this very large difference in rib orientation, the angles between the ribs and the muscle fibres at FRC were only about 5 deg smaller than in cadavers, and the difference in the fractional changes in muscle length in the dorsal region of the interspaces was less than 1 %. In the ventral region, where the length changes are greater (see Results), the difference was larger but still less than 3 %. It was concluded, therefore, that the error due to the potential difference in rib and muscle orientation does not fundamentally question the validity of our observations.

Statistical analysis

The values of angles and muscle masses obtained in a given site on the right and left sides of the chest were averaged for each cadaver, and they were then averaged for the subject group. The values of the parameters describing the orientation and shape of the left and right ribs and the values of the fractional changes in muscle length were similarly averaged first for each subject and then for the subject group. All these data are presented as means ±s.e.m. Statistical comparisons between angles, masses, and changes in muscle length in the dorsal and ventral portions of the different interspaces were made by analysis of variance (ANOVA) with repeated measures, and multiple comparison testing of the mean values was performed, when appropriate, using Student-Newman-Keuls tests. The criterion for statistical significance was taken as P < 0.05.

RESULTS

Muscle orientation

The angles between the ribs and the intercostal muscle fibres were relatively uniform throughout the ribcage, as shown in Fig. 5. The acute angle between the rib and the external intercostal, however, was greater in the ventral half of the second interspace than in the ventral half of the fourth and sixth interspaces (P < 0.01) and in the dorsal half of the second to eighth interspaces (P < 0.05). Moreover, the obtuse angle between the rib and the internal intercostal in the dorsal portion of the second interspace was smaller than that in the dorsal portion of the fourth and sixth interspaces (P < 0.05).

Figure 5. Orientations of the external and internal intercostal muscles in cadavers.

Figure 5

Open in a new tab

The values shown in A are the acute angles between the ribs and the external intercostal fibres in the centres of the dorsal and ventral halves of interspaces 2, 4, 6 and 8; the values shown in B are the obtuse angles between the ribs and the internal intercostal fibres. Data are means ±s.e.m. obtained from five subjects.

Muscle mass

Figure 6 shows the values of bilateral external and internal intercostal muscle mass in the different areas of the ribcage. External intercostal muscle mass decreased gradually from the second to the eighth interspace (P < 0.001) and decreased markedly in each interspace from the dorsal to the ventral portion (P < 0.01). In contrast, the mass of internal intercostal muscle was larger in the ventral than in the dorsal portion of the ribcage (P < 0.02) and did not show any clear-cut difference between the different interspaces.

Figure 6. Mass of the external (A) and internal (B) intercostal muscles in cadavers.

Figure 6

Open in a new tab

Data are means ±s.e.m. obtained from five cadavers on both sides of the sternum. Same symbols as in Fig. 5. The values of parasternal intercostal muscle mass in these subjects are also shown (inset).

The total mass of external intercostal for the four interspaces studied thus averaged 52.5 ± 3.6 g, whereas the total mass of internal intercostal was only 34.4 ± 5.0 g (P < 0.05). Both values were significantly greater (P < 0.05) than the mass of parasternal intercostal muscle (Fig. 6, inset). Indeed, in agreement with our previous study (De Troyer et al. 1998), bilateral parasternal muscle mass was 4-5 g in interspaces two to four and 2.7 g in interspace five. As a result, the total mass of parasternal intercostal amounted to only 16.5 ± 2.5 g.

Shape, orientation and transformation of the ribs

The values of the angles α and β at FRC and TLC are displayed in Fig. 7, and the values of the other parameters describing the orientation and shape of the ribs at the two lung volumes are summarized in Table 1. The angle α at FRC increased (became more negative) gradually from the second to the ninth rib (P < 0.001), thus confirming that the ribs are slanted more and more caudally in the dorsoventral direction when going from the top to the base of the ribcage. However, whereas we had anticipated that all ribs at FRC also sloped caudally in the lateral direction, the values of β were negative for ribs two to five but positive for ribs seven to nine. In other words, the lateral slope of the lower ribs at FRC was cranial. This progressive change in sign of the angle β, in fact, was already apparent in the data shown in Fig. 1 (left panel); whereas the middle arcs of ribs 2-5 in this particular subject were slightly concave upward, the middle arcs of ribs 6 and 7 were nearly straight and those of ribs 8 and 9 were concave downward.

Figure 7. Orientation of the plane of the ribs.

Figure 7

Open in a new tab

The values shown in A are the pump-handle angles (α) of the planes of the ribs at FRC (•) and TLC (•), and those shown in B are the bucket-handle angles (β). Angle α is negative when the plane of the rib slopes ventrally and caudally, and angle β is negative when the plane of the rib slopes laterally and caudally. The differences between the values of α and β at the two lung volumes thus correspond, respectively, to the pump-handle and bucket-handle rotations of the ribs during passive inflation. Data are means ±s.e.m. obtained from five healthy subjects.

Table 1.

Parameters describing the orientation and shape of the ribs

Rib no. A (cm) α (deg) β (deg) ξ0(cm) R1 (cm) R2 (cm)
At FRC
2 27.1 (1.2) −40.3 (1.6) −23.1 (4.5) 3.6 (1.0) 7.2 (0.2) 10.6 (0.4)
3 25.1 (1.1) −37.8 (2.1) −21.2 (4.7) 3.1 (0.8) 9.2 (0.2) 11.6 (0.3)
4 19.8 (1.2) −40.3 (2.3) −16.3 (4.7) 2.5 (1.0) 11.7 (0.3) 12.7 (0.4)
5 16.6 (0.7) −41.9 (2.4) −9.4 (3.9) 1.5 (0.8) 12.8 (0.4) 12.6 (0.4)
6 9.6 (0.6) −44.6 (2.8) −0.3 (4.8) −0.4 (0.7) 14.0 (0.3) 12.8 (0.5)
7 5.1 (0.6) −46.1 (2.7) +8.8 (3.7) −2.2 (0.8) 14.7 (0.3) 13.0 (0.6)
8 −0.9 (1.1) −50.7 (2.1) +10.8 (3.0) −2.6 (1.3) 15.8 (0.6) 13.3 (0.6)
9 −2.7 (1.1) −50.7 (1.7) +10.3 (2.9) −3.6 (1.1) 14.4 (0.7) 13.4 (0.6)
Rib no. A (cm) α (deg) β (deg) ξ0(cm) η0 (cm)
At TLC
2 26.4 (0.9) −26.0 (1.2) −9.4 (3.6) 3.2 (0.7) −0.5 (0.1)
3 24.7 (0.9) −26.4 (1.7) −7.9 (4.1) 2.7 (0.8) −0.5 (0.1)
4 20.5 (0.8) −29.6 (1.9) −6.2 (3.9) 2.9 (0.8) −0.1 (0.1)
5 17.4 (0.6) −32.3 (2.5) −0.5 (3.6) 2.2 (0.8) 0.1 (0.0)
6 11.2 (0.6) −35.2 (2.9) +6.6 (5.1) 1.2 (0.7) 0.6 (0.1)
7 6.4 (0.7) −38.2 (2.7) +15.4 (3.9) −0.7 (0.8) 0.9 (0.3)
8 1.1 (1.3) −42.8 (1.6) +17.0 (2.8) −0.5 (1.1) 1.2 (0.3)
9 −1.5 (1.6) −44.7 (1.9) +16.6 (3.6) −2.1 (1.3) 1.3 (0.3)
Open in a new tab

Values are means (S. E. M.)of data obtained from five subjects on both sides of the chest.

All the ribs showed a decrease in both α and β with passive inflation to TLC. However, the amplitude of the change in α decreased progressively from 14.3 ± 0.6 deg for rib 2 to 6.0 ± 0.6 deg for rib 9 (P < 0.001). The amplitude of the change in β similarly decreased from 13.7 ± 1.3 deg for rib 2 to 6.3 ± 1.4 deg for rib 9 (P < 0.001).

Computed changes in muscle length

The computed fractional changes in external intercostal muscle length are shown in Fig. 8A. Passive inflation from FRC to TLC induced a clear-cut shortening of the muscle in the second interspace. However, this fractional muscle shortening decreased continuously and rapidly toward the eighth interspace (P < 0.001), in particular in the ventral portion of the ribcage. As a result, the external intercostal muscle in the ventral half of the eighth interspace showed a large fractional lengthening.

Figure 8. Computed fractional changes in external (A) and internal (B) intercostal muscle length.

Figure 8

Open in a new tab

Data are the computed values during passive inflation from FRC to TLC and are expressed as percentage changes relative to the muscle length at FRC (LFRC); negative changes in length represent muscle shortening, and positive changes in length represent muscle lengthening. Same symbols as in Fig. 5.

The internal intercostal muscle in the dorsal portion of the ribcage lengthened in all interspaces (Fig. 8B) and showed a prominent rostrocaudal gradient as well (P < 0.001). Thus, the fractional muscle lengthening decreased gradually from 30 % in the second interspace to only 11 % in the eighth interspace. However, whereas the fractional shortening of the external intercostals decreased from the dorsal to the ventral aspect of the ribcage, the fractional lengthening of the internal intercostals did not show any uniform dorsoventral gradient. Specifically, the muscle lengthening in the second interspace decreased from the dorsal to the ventral portion, whereas in the sixth and eighth interspaces, the muscle lengthening increased considerably. In the ventral portion of the ribcage, therefore, the fractional lengthening of the internal intercostal increased markedly and continuously from the second to the eighth interspace (P < 0.001).

Computed respiratory effects

According to eqn (1), the mechanical advantage of a respiratory muscle can be evaluated by measuring (ΔL/(LΔVL))Rel, and the maximum ΔPao that this muscle can produce can be calculated by multiplying mechanical advantage by muscle mass and maximum active stress. The mechanical advantages of the external and internal intercostal muscles in the five subjects studied were thus obtained by dividing the average fractional changes in muscle length (shown in Fig. 8) by the mean inspiratory capacity, and these values were then multiplied by the values of muscle mass measured in cadavers. The maximum active muscle stress was assumed to be similar to the dog (i.e. ♦ 3.0 kg cm−2).

The results of these computations are summarized in Fig. 9. As anticipated from the data reported in Figs 6A and 8A, the maximum ΔPao for the external intercostals decreased gradually from -1.19 cmH2O in the dorsal half of the second interspace to -0.13 cmH2O in the dorsal half of the eighth interspace and -0.50 cmH2O in the ventral half of the second interspace. The inspiratory effect was even reversed into a small expiratory effect in the ventral half of the sixth (+0.15 cmH2O) and eighth (+0.32 cmH2O) interspaces. On the other hand, the ΔPao values for the internal intercostals were positive in all areas, and the largest values were obtained for the ventral half of the caudal interspaces (Fig. 9B). The value for the ventral half of the eighth interspace amounted to +2.00 cmH2O.

Figure 9. Computed respiratory effects of the external (A) and internal (B) intercostal muscles.

Figure 9

Open in a new tab

Same symbols as in Fig. 5. These data are the computed maximal changes in airway pressure (ΔPao) that the muscles can generate when contracting alone against a closed airway. A negative ΔPao indicates an inspiratory effect, and a positive ΔPao indicates an expiratory effect.

DISCUSSION

The current studies provide the first comprehensive analysis of the mechanics of the external and internal interosseous intercostal muscles in man, and the most important results can be summarized as follows. (1) The external intercostals in many areas of the ribcage shorten during passive inflation whereas the internal intercostals lengthen. Therefore, the external intercostals, by and large, have an inspiratory mechanical advantage and the internal intercostals have an expiratory mechanical advantage, in agreement with the theory of Hamberger (1749). (2) However, the magnitudes of the mechanical advantages vary throughout the ribcage, such that the inspiratory advantage of the external intercostals is greatest in the rostral interspaces and the expiratory advantage of the internal intercostals is greatest in the ventral portion of the caudal interspaces. (3) The mass of the external intercostal muscle is also greater in the dorsal half of the ribcage, and the mass of the internal intercostal muscle is greater in the ventral half. As a result, the external intercostals in the dorsal aspect of the rostral segments have a significant inspiratory effect on the lung, and the internal intercostals in the ventral aspect of the caudal segments have a large expiratory effect. These results will be successively discussed, and the similarities and differences with the canine intercostal muscles will be emphasized. However, because these results rely on the observed motion of the ribs during passive inflation, the orientation and transformation of the ribs will be considered first.

Orientation and transformation of the ribs

From lateral roentgenograms of the chest, Sharp et al. (1986) have previously measured the angles between ribs 4-7 at FRC and TLC and the rostrocaudal axis of the body in twenty healthy individuals. Wilson et al. (1987) have also examined the three-dimensional orientation of ribs 3-7 at FRC and TLC in two normal subjects by using the dynamic spatial reconstructor (DSR). The subjects in both studies maintained TLC actively, whereas our subjects were specifically coached to relax their respiratory muscles against a closed glottis. In addition, the subjects studied by Sharp et al. (1986) held their arms at right angles to the sagittal midplane and those studied by Wilson et al. (1987) had the arms raised above the head, whereas our subjects were studied with the arms resting along the sides of the body. One would expect that maintaining TLC actively (Sharp et al. 1986) and raising the arms would raise the ribs, and indeed, the angles reported by Sharp et al. (1986) and Wilson et al. (1987) suggested that the ribs in these studies were closer to the transverse (x-y) plane of the body than in the current study. Yet the main observations reported in these studies were similar in all respects to the current findings (Fig. 7). Specifically, the pump-handle angle of the ribs at FRC, relative to the transverse plane of the body, increased progressively with increasing rib number, and the pump-handle rotation of the ribs between FRC and TLC gradually decreased. With increasing rib number, the bucket-handle angle of the ribs at FRC and their bucket-handle rotation from FRC to TLC decreased gradually as well.

As the pump-handle rotation of the ribs from FRC to TLC (Δα) represents the component of the rib rotation vector that lies perpendicular to the sagittal mid-plane and the bucket-handle rotation of the ribs (Δβ) represents the component of the rib rotation vector that lies in the sagittal midplane, the ratio Δα/Δβ is the tangent of the dorsal angle between the axis of rib rotation and the midplane. For any given rib, the magnitudes of Δα and Δβ in our subjects were about the same, so all the ratios Δα/Δβ were ♦1.00. Consequently, the angle between the axis of rotation of all the ribs and the mid-plane must be ♦45 deg, and this is different from the conventional idea that this angle decreases with increasing rib number. It must be emphasized, however, that this idea is based on the anatomical description by von Hayek (1953) of the costovertebral joints and the orientation of the neck of the ribs, whereas our values correspond to the (effective) axis of rib rotation. Passive inflation might be associated with a rostrocaudal displacement of the costotransverse joint, or of the head of the rib, or both. In that case, the axis of rib rotation and the orientation of the rib neck would not coincide. It may be difficult to envisage a rostrocaudal displacement of the head of the rib during inspiration, but gliding movements of the costotransverse joints in ribs 5-10 have been previously observed during radiological studies (von Hayek, 1953). Saumarez (1986), in his detailed analysis of the kinematics of the human ribcage, has also concluded that the costotransverse facets in ribs 2-6 glide over each other during breathing.

Changes in muscle length

According to the theory of Hamberger (1749), all the external intercostals should shorten as the relaxed respiratory system is inflated and the ribs rotate cranially, whereas all the internal interosseous intercostals should lengthen. Sharp et al. (1986) have previously calculated that in normal humans, the external intercostal in the middle portion of the fourth interspace shortens during inspiration and that the underlying internal intercostal lengthens. Similarly, passive inflation in our subjects was found to induce shortening of the external intercostals in many areas of the ribcage and lengthening of the internal intercostals throughout the cage (Fig. 8). Therefore, in agreement with the theory of Hamberger (1749), the conclusion must be drawn that the orientation of the muscle fibres is a major determinant of the mechanics of the human intercostal muscles. However, the fractional changes in muscle length were also found to be distributed along dorsoventral and rostrocaudal gradients, thus indicating that other determinants operate as well.

As we have recently pointed out (De Troyer et al. 1999), the effects of passive inflation on the length of the intercostal muscles in a given interspace depend on the position of the muscle around the arc of the ribs. Indeed, the ribs are curved, rather than straight, and for two parallel curved ribs that rotate by equal amounts around parallel axes, the change in length of muscles connected to the ribs is proportional to only one component of rib rotation, namely the component normal to the surface of the ribs. Because the axis of rib rotation lies at an angle of ♦45 deg from the sagittal mid-plane, the normal component of rib rotation is greatest in the dorsal part of the ribcage and decreases gradually as one moves ventrally toward the costochondral junctions. With passive inflation, therefore, the external intercostal in a given segment would shorten most in the dorsal aspect and least in the ventral aspect; and the internal intercostal would lengthen most in the dorsal aspect of the ribcage and least in the ventral aspect. However, the changes in muscle length in a given interspace also depend on the relative displacements of the two ribs making up the interspace. If passive inflation causes greater cranial displacement of the rib above than the rib below, the interspace expands and both the external and the internal intercostal tend to lengthen. Conversely, if the cranial displacement of the rib below is greater, the interspace is compressed and both muscles tend to shorten.

In the dorsal region of the ribcage, the normal component of rib rotation is large and rib displacements are small. Consequently, rib rotation is the major determinant of the changes in muscle length, leading to a shortening of the external intercostals and a lengthening of the internal intercostals. The large rostrocaudal reduction in the magnitude of rib rotation (Fig. 7) also leads to progressive reductions in external shortening and internal lengthening in this region of the ribcage (Fig. 8). On the other hand, in the ventral region, the normal component of rib rotation is small and rib displacements are large, such that the changes in muscle length should be primarily determined by differences in rib displacement. The radii of rib 3 are considerably larger than those of rib 2 (Table 1). With passive inflation, therefore, the cranial displacement of rib 3 is bigger and the second interspace is compressed, causing marked shortening of the external intercostal throughout the segment and reducing the lengthening of the internal intercostal. In contrast, the radii of ribs 6-9 are nearly equal (Fig. 7 and Table 1) but the magnitude of rib rotation, in particular α, decreases continuously with increasing rib number. Consequently, the cranial displacement of the ribs decreases caudally and the lower interspaces are expanded; hence, the shortening of the external intercostals is reversed into a muscle lengthening and the lengthening of the internal intercostals is accentuated.

Mechanical advantage

The topographic distribution of mechanical advantage among the external intercostals in humans is qualitatively similar to that observed in the dog (De Troyer et al. 1999). In both species, the inspiratory mechanical advantage of the muscle is greatest in the rostral interspaces and is reversed into an expiratory mechanical advantage in the ventral portion of the caudal interspaces. On the other hand, the topographic distribution of mechanical advantage among the human internal intercostals differs from the dog in two respects. First, the expiratory mechanical advantage of the muscle in the dorsal portion of the ribcage decreases markedly from the second to the eighth interspace in humans, whereas in the dog, it gradually increases. Secondly, even though the expiratory mechanical advantage is greatest in the caudal segments in both species, it peaks in the ventral portion in humans and in the dorsal portion in the dog.

These differences are probably the result of the species differences in ribcage shape and rib displacements. In humans, the ribcage at FRC is larger along its lateral than dorsoventral axis, whereas in the dog, as in most quadrupeds, it is larger along its dorsoventral axis. Therefore, the normal component of rib rotation in the dorsal aspect of the ribcage is larger in humans than in the dog and plays a larger role in determining the mechanical advantage of the intercostal muscles in this region. In addition, the ratio Δα/Δβ is about twice as large in humans as in the dog (Margulies et al. 1989), thus implying that the axis of rib rotation lies much closer to the midplane in the dog. This difference not only increases further the importance of the normal component of rib rotation in the dorsal part of the human ribcage but it also accentuates the expansion of the ventral portion of the caudal interspaces. The lengthening of the internal and external intercostal muscles in this portion should therefore be larger than in the dog. Finally, the ribs in humans are slanted caudally (Fig. 1, left panel), whereas the ribs in the dog are not (see Fig. 1, right, in Margulies et al. 1989); as we have previously discussed in relation to the parasternal intercostals and triangularis sterni (De Troyer et al. 1998), the caudal slope of the human ribs should also augment the lengthening of the internal intercostals in the ventral portion of the ribcage.

The mechanical advantages of the intercostal muscles in humans and dogs show some quantitative differences as well. Making such comparisons requires that the size of the subjects be considered. Thus, one would expect that in any particular species, the fractional change in length of a given respiratory muscle over the inspiratory capacity would be independent of subject's size. Consequently, the mechanical advantage of the muscle would be inversely proportional to inspiratory capacity. The dogs studied in our previous investigation (De Troyer et al. 1999) had body masses between 15 and 25 kg and an inspiratory capacity of 1-1.5 l, and with a 1 l passive inflation, the external intercostal in the dorsal part of the second interspace shortened by 10.0 % and the internal intercostal in the eighth interspace lengthened by 15.0 %. On the other hand, when the subjects of this study increased lung volume by 3.5 l from FRC to TLC, the external intercostal in the dorsal part of the second interspace shortened by 11 % and the internal intercostal in the ventral part of the eighth interspace lengthened by 42 %. In other words, the inspiratory mechanical advantage of the external intercostal muscles is about three times smaller in humans than in the dog, whereas the expiratory mechanical advantage of the internal intercostal muscles, notwithstanding the difference in dorsoventral gradient, is about the same in the two species.

Respiratory effect

As a result of the distributions of mechanical advantage and muscle mass (Fig. 6), the respiratory effects of the muscles showed definite gradients in the dorsoventral direction. For example, the computed ΔPao for the external intercostal situated in the dorsal half of the second interspace (Fig. 9) was two times more negative than that computed for the muscle in the ventral half of the same interspace. The computed ΔPao values for the internal intercostals in the ventral half of the sixth and eighth interspaces were also three to five times more positive than those computed for the muscles in the dorsal half.

The distributions of respiratory effect showed even stronger gradients in the rostrocaudal direction. As the fibres of the external or internal intercostal muscle in a given interspace are parallel to each other, the forces they develop on the ribs are additive. Therefore, the ΔPao values generated by the areas of external or internal intercostal muscle in the dorsal and ventral halves of a given segment must be additive as well, and the results of such additions in the different interspaces are shown in Fig. 10. Whereas the total ΔPao for the external intercostal in the second interspace amounts to -1.69 cmH2O, the total ΔPao for the external intercostal in the sixth interspace is only -0.12 cmH2O and that for the external intercostal in the eighth interspace amounts to +0.19 cmH2O. In contrast, the total ΔPao for the internal intercostals decreases gradually from +2.39 cmH2O in the eighth interspace to +0.85 cmH2O in the second interspace. It should be concluded, therefore, that if all the muscles were contracting maximally, the internal intercostals would have a clear-cut expiratory effect in all segments of the ribcage whereas the external intercostals would have an inspiratory effect only in the upper six interspaces.

Figure 10. Net respiratory effect of the external and internal intercostal muscles in different segments.

Figure 10

Open in a new tab

These data were obtained by adding the changes in airway pressure (ΔPao) computed for the muscle areas in the dorsal and ventral halves of each interspace (data shown in Fig. 9).

We have previously shown that the maximal ΔPao values produced by the canine external and internal intercostal muscles in different segments are essentially additive (Legrand et al. 1998), and there is no reason to believe that this principle does not apply to humans as well. Adding the ΔPao values thus calculated for the four interspaces studied (shown in Fig. 10) and multiplying by two for the odd-numbered interspaces yields a total value of -4.30 cmH2O for the external intercostals and +12.70 cmH2O for the internal intercostals. If, as in the dog (Legrand & De Troyer, 1999), only the external intercostals with an inspiratory mechanical advantage contract during inspiration, then the total ΔPao for the external intercostals increases to -5.30 cmH2O, and yet these estimates are probably low for two reasons. Firstly, as we have previously pointed out (De Troyer et al. 1998), it is most likely that the relationships between lung volume and external or internal intercostal muscle length are curvilinear. Thus, for a given volume increase, the changes in muscle length, in particular for the external intercostals, should be larger near FRC than near TLC. Consequently, the computed mechanical advantages were probably underestimated. Secondly, even though the cadavers were carefully selected and did not show any evidence of overt undernutrition, their external and internal intercostal mucles may have been somewhat atrophied. If one assumes that the mechanical advantages of the muscles are 1.5 times greater than the values herein computed and that muscle masses in young healthy individuals are 2 times greater than in cadavers, then the total ΔPao values for the external and internal intercostals in all interspaces would be about -15 and +40 cmH2O, respectively.

These values are substantially greater than those for the parasternal intercostals and triangularis sterni. Indeed, using similar techniques to those in the current studies, we have recently estimated that the total ΔPao values for all the parasternal intercostals and the entire triangularis sterni in humans were only -3 and +3 cmH2O, respectively (De Troyer et al. 1998). The greater inspiratory effect of the human external intercostals vs. the parasternal intercostals is largely related to the difference in muscle mass. As shown in Fig. 6, the total mass of external intercostal for the four interspaces studied was 52 g. Multiplying this value by two (for the odd-numbered interspaces) thus yields a total value of 104 g, whereas the total parasternal muscle mass was only 16.5 g. The total mass of internal intercostal muscle (69 g) is similarly greater than the triangularis sterni muscle mass (10 g) by a factor of ♦7. In addition, the computed mechanical advantages for the internal intercostals in the second and sixth interspaces average +4.6 and +6.0 % l−1, respectively, whereas the computed values for the triangularis sterni in the same interspaces are only +1.6 and +3.8 % l−1. The greater expiratory mechanical avantage of the internal intercostals, combined with the greater muscle mass, yields a total ΔPao that is ♦15 times greater than that of the triangularis sterni.

The difference between the total ΔPao values calculated for the human external and parasternal intercostals is opposite to that seen in the dog. In dogs with body masses of 15-25 kg, the mass of the parasternal intercostal muscle in a given interspace is about 10-11 g (Legrand et al. 1996; De Troyer et al. 1996), whereas the mass of external intercostal muscle, including in the rostral interspaces, is only 5 g (De Troyer et al. 1999). Furthermore, the inspiratory mechanical advantage of both muscles decreases caudally in the dog but the advantage of the parasternal intercostals decreases more slowly than that of the external intercostals. As a result, the canine parasternal intercostals have, on average, a greater mechanical advantage than the external intercostals, and this difference, combined with the greater muscle mass, yields a greater inspiratory effect. Whereas the parasternal intercostals in the dog play a more important role than the external intercostals during breathing (De Troyer, 1991; De Troyer & Wilson, 2000), one would therefore predict that in humans, the external intercostals predominate.

Respiratory activity

Our recent electromyographic studies of the canine parasternal intercostals have shown that in the dog, the distribution of activity among these muscles during breathing is closely matched to the distribution of inspiratory effect, such that the muscle areas with the greatest effect are also those which receive the greatest neural drive during inspiration (De Troyer & Legrand, 1995; De Troyer et al. 1996; Legrand et al. 1996). A similar matching was also demonstrated for the canine external and internal interosseous intercostals (Legrand & De Troyer, 1999). If one assumes that this principle applies to humans, the distribution of respiratory effect observed in the current study would therefore lead to the prediction that the external intercostals in man would contract during inspiration, in particular in the dorsal portion of the rostral segments. On the other hand, the internal intercostals would contract during expiration and display greater activity in the ventrolateral aspect of the caudal interspaces.

In agreement with these predictions, the extensive electrical recordings from the intercostal muscles in normal humans performed by Taylor (1960) have established that the external intercostals are active only during inspiration and that the internal interosseous intercostals are active only during expiration. Also, expiratory EMG activity during resting breathing was detected only in the internal intercostals of the ventrolateral aspect of the seventh through tenth interspaces (see Fig. 7 in Taylor, 1960). However, inspiratory EMG activity in this condition was limited to parasternal intercostals. No activity was recorded from the external intercostals, which is contradictory to our prediction. Taylor himself has pointed out, however, that the dorsal aspect of the rostral segments of the ribcage was hardly investigated, and Campbell (1970), referring to some experiments in man by Sears & Newsom Davis, stated that in fact, the external intercostals in this region do contract during resting inspiration. How the neural drive to these muscles compares with that to the parasternal intercostals, however, remains unknown.

Acknowledgments

This study was supported in part by a grant (HL-45545) from the National Institutes of Health (USA).

References

  1. Campbell EJM, Newsom Davis J. The intercostal muscles and other muscles of the rib cage. In: Campbell EJM, Agostoni E, Newsom Davis J, editors. The Respiratory Muscles. Mechanics and Neural Control. 2. Philadelphia: W. B. Saunders; 1970. pp. 161–174. [Google Scholar]
  2. Close RI. Dynamic properties of mammalian skeletal muscles. Physiological Reviews. 1972;52:129–197. doi: 10.1152/physrev.1972.52.1.129. [DOI] [PubMed] [Google Scholar]
  3. De Troyer A. Inspiratory elevation of the ribs in the dog: primary role of the parasternals. Journal of Applied Physiology. 1991;70:1447–1455. doi: 10.1152/jappl.1991.70.4.1447. [DOI] [PubMed] [Google Scholar]
  4. De Troyer A, Kelly S. Chest wall mechanics in dogs with acute diaphragm paralysis. Journal of Applied Physiology. 1982;53:373–379. doi: 10.1152/jappl.1982.53.2.373. [DOI] [PubMed] [Google Scholar]
  5. De Troyer A, Legrand A. Inhomogeneous activation of the parasternal intercostals during breathing. Journal of Applied Physiology. 1995;79:55–62. doi: 10.1152/jappl.1995.79.1.55. [DOI] [PubMed] [Google Scholar]
  6. De Troyer A, Legrand A. Mechanical advantage of the canine triangularis sterni. Journal of Applied Physiology. 1998;84:562–568. doi: 10.1152/jappl.1998.84.2.562. [DOI] [PubMed] [Google Scholar]
  7. De Troyer A, Legrand A, Gevenois PA, Wilson TA. Mechanical advantage of the human parasternal intercostal and triangularis sterni muscles. The Journal of Physiology. 1998;513:915–925. doi: 10.1111/j.1469-7793.1998.915ba.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. De Troyer A, Legrand A, Wilson TA. Rostrocaudal gradient of mechanical advantage in the parasternal intercostal muscles of the dog. The Journal of Physiology. 1996;495:239–246. doi: 10.1113/jphysiol.1996.sp021588. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. De Troyer A, Legrand A, Wilson TA. Respiratory mechanical advantage of the canine external and internal intercostal muscles. The Journal of Physiology. 1999;518:283–289. doi: 10.1111/j.1469-7793.1999.0283r.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. De Troyer A, Wilson TA. The canine parasternal and external intercostal muscles drive the ribs differently. The Journal of Physiology. 2000;523:799–806. doi: 10.1111/j.1469-7793.2000.00799.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  11. Farkas GA. Mechanical properties of respiratory muscles in primates. Respiration Physiology. 1991;86:41–50. doi: 10.1016/0034-5687(91)90038-k. [DOI] [PubMed] [Google Scholar]
  12. Farkas GA, Decramer M, Rochester DF, De Troyer A. Contractile properties of intercostal muscles and their functional significance. Journal of Applied Physiology. 1985;59:528–535. doi: 10.1152/jappl.1985.59.2.528. [DOI] [PubMed] [Google Scholar]
  13. Gibson GJ, Pride NB, O'Cain C, Quagliato R. Sex and age differences in pulmonary mechanics in normal nonsmoking subjects. Journal of Applied Physiology. 1976;41:20–25. doi: 10.1152/jappl.1976.41.1.20. [DOI] [PubMed] [Google Scholar]
  14. Hamberger GE. De Respirationis Mechanismo et usu Genuino. Iena; 1749. [Google Scholar]
  15. Legrand A, Brancatisano A, Decramer M, De Troyer A. Rostrocaudal gradient of electrical activation in the parasternal intercostal muscles of the dog. The Journal of Physiology. 1996;495:247–254. doi: 10.1113/jphysiol.1996.sp021589. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Legrand A, De Troyer A. Spatial distribution of external and internal intercostal activity in dogs. The Journal of Physiology. 1999;518:291–300. doi: 10.1111/j.1469-7793.1999.0291r.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
  17. Legrand A, Ninane V, De Troyer A. Mechanical advantage of sternomastoid and scalene muscles in dogs. Journal of Applied Physiology. 1997;82:1517–1522. doi: 10.1152/jappl.1997.82.5.1517. [DOI] [PubMed] [Google Scholar]
  18. Legrand A, Wilson TA, De Troyer A. Mediolateral gradient of mechanical advantage in the canine parasternal intercostals. Journal of Applied Physiology. 1996;80:2097–2101. doi: 10.1152/jappl.1996.80.6.2097. [DOI] [PubMed] [Google Scholar]
  19. Legrand A, Wilson TA, De Troyer A. Rib cage muscle interaction in airway pressure generation. Journal of Applied Physiology. 1998;85:198–203. doi: 10.1152/jappl.1998.85.1.198. [DOI] [PubMed] [Google Scholar]
  20. Margulies SS, Rodarte JR, Hoffman EA. Geometry and kinematics of dog ribs. Journal of Applied Physiology. 1989;67:707–712. doi: 10.1152/jappl.1989.67.2.707. [DOI] [PubMed] [Google Scholar]
  21. Saumarez RC. An analysis of possible movements of human upper rib cage. Journal of Applied Physiology. 1986;60:678–689. doi: 10.1152/jappl.1986.60.2.678. [DOI] [PubMed] [Google Scholar]
  22. Sharp JT, Beard GAT, Sunga M, Kim TW, Modh A, Lind J, Walsh J. The rib cage in normal and emphysematous subjects: a roentgenographic approach. Journal of Applied Physiology. 1986;61:2050–2059. doi: 10.1152/jappl.1986.61.6.2050. [DOI] [PubMed] [Google Scholar]
  23. Taylor A. The contribution of the intercostal muscles to the effort of respiration in man. The Journal of Physiology. 1960;151:390–402. doi: 10.1113/jphysiol.1960.sp006446. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Turner JM, Mead J, Wohl ME. Elasticity of human lungs in relation to age. Journal of Applied Physiology. 1968;25:664–671. doi: 10.1152/jappl.1968.25.6.664. [DOI] [PubMed] [Google Scholar]
  25. von Hayek H. Die Menschliche Lunge. Berlin, Germany: Springer-Verlag; 1953. [Google Scholar]
  26. Wilson TA, De Troyer A. Effect of respiratory muscle tension on lung volume. Journal of Applied Physiology. 1992;73:2283–2288. doi: 10.1152/jappl.1992.73.6.2283. [DOI] [PubMed] [Google Scholar]
  27. Wilson TA, De Troyer A. Respiratory effect of the intercostal muscles in the dog. Journal of Applied Physiology. 1993;75:2636–2645. doi: 10.1152/jappl.1993.75.6.2636. [DOI] [PubMed] [Google Scholar]
  28. Wilson TA, Rehder K, Krayer S, Hoffman EA, Whitney CG, Rodarte JR. Geometry and respiratory displacement of human ribs. Journal of Applied Physiology. 1987;62:1872–1877. doi: 10.1152/jappl.1987.62.5.1872. [DOI] [PubMed] [Google Scholar]

Articles from The Journal of Physiology are provided here courtesy of The Physiological Society

RESOURCES