The characterization of different foot actions in running is an helpful method to analyze relationships between kinematic, kinetic, and anatomical parameters of the foot. Various kinematic methods have been used to define two main foot types: the pronated foot combines dorsiflexion, abduction, and eversion, while the supinated foot combines plantar flexion, adduction, and inversion. The footprint measurement has been utilized as a simple and convenient method to characterize the architecture of the foot. Braun et al. (4) used a ratio of the width of the middle third of the footprint to the width of the metatarsal heads area. They defined five categories, from “flat arch,” “intermediate,”“precavus,” and “high arch, first degree” to“high arch, second degree.” Holden et al.(12) recorded footprints in dynamic conditions to analyze the orientation of the geometrical axis of the global foot in relation to the direction of displacement (foot angle). They observed changes in foot angles between right and left feet and between slow and fast running. Cavanagh and Rodgers (7) used static and dynamic footprints to propose a simple measurement of the arch height, called “arch index,” which is an expression of the ratio between the surface of the midfoot area and the total surface of the foot. The authors defined three categories called “high, normal and flat arched groups.” This method was later used by Hamill et al. (11) to compare static arch index with kinematic parameters. It was observed that the arch index was not a sufficient parameter to explain the dynamic variations between subjects and no relevant correlation have yet been established between footprint measurements and dynamic behavior of the foot. In the same way, Nachbauer and Nigg (20) examined the effects of arch height and arch flattening on ground reaction forces, and no significant correlation was established. Foot angle, arch index, and arch height methods assumed that the orientation between rearfoot and forefoot was constant during all phases of foot contact. In fact, Bojsen-Møeller(3) observed that the movement between the rearfoot and the forefoot occurred at the transverse tarsal joint, which consists of the talonavicular and the calcaneocuboid joints. The secondary axis of the calcaneocuboid joint allows then a rotation of the forefoot in relation to the fixed rearfoot, in the horizontal plane. Dahle et al. (8) characterized, by visual assessment, two categories of foot according to the position of the forefoot in relation to the rearfoot. The forefoot was abducted in the “pronated foot,” and it was adducted in the“supinated foot.” These observations suggest that the mobility between the rearfoot and the forefoot may occur in different orientations of each rearfoot and forefoot in the horizontal plane, and may be examined in connection with pronation. Therefore, missing information concerning the functional anatomy and the intrinsic movement of the foot could explain the lack of relationship between footprint measurements and dynamic parameters. Because footprints supply observations of the foot in the horizontal plane, it may be a relevant method to record the rearfoot/forefoot mobility in static and in dynamic conditions.
As a consequence, one may speculate that a footprint method measuring the rearfoot/forefoot mobility could supply additional information to be connected with typical ground reaction forces. The aim of this paper was to measure separately rearfoot and forefoot orientation in static and running condition, by means of a simple footprint method and to find out whether a quantitative footprint distinction could be suggested for the evaluation of lower extremity function during running. The specific purpose was to examine the relationships between the foot orientation and the ground reaction forces.
Thirty-four male students volunteered as subjects for this experiment. They all had training experience in running and reported no injury at the time of the experiment. Before the study, written informed consent was obtained from each subject. Because the footprint method cannot be applied to complete forefoot strikers, two subjects were not taken into account in the present study. Mean age, height, and body weight were, respectively, 26.5 ± 5 yr (mean ± SD), 177.8 ± 6 cm, and 70.0 ± 6.7 kg(N = 32 subjects).
After warm-up, the subjects were asked to run in barefoot condition on a 30-m track of ground sheet (Fig. 1) at a required velocity of 3.8 m·s-1 (from 3.4 to 4.2 m·s-1). The running velocity was controlled and computed by means of two photocells located 2 m before and 2 m after the force platform. To avoid the possibility that velocity influences the ground reaction force parameters, the trial was rejected when the velocity was out of the limits; it was also rejected when visible stride length or audible step frequency adjustment were produced before the measured foot contact. After 15 m, which was the required distance to stabilize the target velocity, the right foot was impregnated with developer by means of a blotter. A force platform, covered with a photosensitive paper was placed at one stride-distance further. This device made it possible to simultaneously record footprint and ground reaction forces. Several trials were needed to obtain from three to five right-foot contacts on the platform in conditions of required velocity and stride length adjustment. The barefoot condition was chosen to cancel the possible alteration of foot behavior due to the shoe, and to reflect the genuine functionality of the foot in running conditions.
Ground reaction data were collected with a PC computer (Victor type M 386) from a force platform (Kistler type 9281B) connected to an amplifier (Kistler type 9861A). The resonant frequency of the force platform was over 200 Hz. The four vertical and four horizontal force channels provided by the platform amplifier were sampled (800-Hz sampling frequency) by a 12-bit analog-to-digital interface card (Metrabyte type DAS8) during the foot contact. Calculation of forces, moment, and path of force application were performed later with a specific software in accordance with the instructions of the force platform manufacturer.
At the end of the measurement session, each subject was asked to stand for 10 s with both feet on the force platform to record his body weight and his static footprint. Horizontal x-ray photographs of the subject's plantar sole were finally recorded, following the method described by Montagne et al.(18). The subject stood, the right foot ahead of the shot, when the first picture was shot of his forefoot. Then, the subject walked one step ahead, without moving the foot in the shot, and a double exposure was made of the rearfoot. This method supplied a complete plantar view of both rearfoot and forefoot.
Three anatomical marks were located on x-ray photographs(Fig. 2a): the rear extremity of the calcaneus (REC), the calcaneocuboid joint (CCJ), and the middle of the second and third metatarsal axis (MM). The distance between REC and CCJ was measured on the x-ray, and then transferred on the static and dynamic footprints (Fig. 2b). The orientation of the rearfoot axis (RFA) was defined by the middle of the footprint area, respectively, measured at 25%, 50%, and 75% of the distance between REC and CCJ. Knowing the distance between REC and CCJ and the orientation of the RFA, it was possible to fully define the location of CCJ on the footprint. The RFA corresponded to the axis of the calcaneus. The MM was located on footprints as the middle of the forefoot area. The axis of the forefoot (FFA) started from the CCJ, ran across the MM, and up to the“spike” corresponding to the middle distance between the second and the third metatarsal head extremities (MH). It was verified that CCJ, MM, and MH were always aligned. The foot was thus represented by two axes: the RFA corresponded to the posterior tarsal area (calcaneus and talus) and the FFA corresponded to the combined midtarsal, metatarsal, and phalangeal areas. The RFA and the FFA intersected at the location of the CCJ.
On the static footprint, the angle between RFA and FFA (αS) was measured. For the dynamic footprint, the direction of running (DOR), parallel to the track, was marked on the sensitive paper and the following angles were determined (Fig. 2c): the angle between the RFA and the FFA (αR), the angle between the RFA and the DOR (αrf), and the angle between the FFA and the DOR (αff). As a convention, a positive arf or aff angle corresponds to a lateral orientation of RFA or FFA in relation to the DOR, and a negative angle would correspond to a medial orientation. A positive αR angle corresponds to a medial rotation of the FFA in relation to the RFA (“closed foot” type), while a negativeαR corresponds to a lateral rotation (“open foot” type). The arch deformation was calculated by subtracting the static from the dynamic rearfoot/forefoot angle (αR - αS).
Mediolateral (Fx), anteroposterior (Fy), vertical (Fz) forces, vertical free moment (Mz), and location of force application point (x,y) data were computed and digitally filtered with a low pass, zero phase lag, fourth order recursive Butterworth filter with a cut-off frequency of 100Hz. A threshold of 50 N on the vertical ground reaction force (Fz) was used to define the beginning and the end of the foot contact. The stance time was scaled for each trial. The force data were further scaled by dividing the data by the individual subject's body weight. The selected variables from the individual force/time curves are presented in Figure 3. The position of the force application point (FAP), usually referred in the literature as“center of pressure,” was related to the location of footprint, by subtracting the calcaneocuboid joint coordinates (on footprint) from the x-y coordinates (in relation to the center of the platform). X and Y coordinates were recorded for four typical locations of the FAP (A, B, C, and D) (seeFig. 4). The total course of FAP on the foot length (yD - yA) was calculated. The orientation of the path during the heel contact(β1) and its change of direction between the rearfoot and the forefoot(β2) were also calculated from the coordinates of A, B, C, and D. The mean value of Mz was calculated at three different phases of foot contact, depending on the location of FAP: 1) during the first phase, the FAP was running through the heel pad, between the two extremities of the calcaneus(REC and CCJ); 2) during the second phase, the FAP was moving forward from the CCJ to the MH. The forefoot supported the braking phase until the FAP stopped its progression; 3) during the third phase, the FAP was moving backward from the metatarsal heads, when the heel had already taken off. The three phases are shown in Figure 4.
As suggested by Campbell and Machin (5), the multiple measurements were averaged for each individual. Finally, simple linear regression analyses were performed between foot orientations αR,αrf, and αff, arch deformation, ground reaction forces, FAP path, and vertical free moment (N = 32 subjects).
The average running velocity was 3.97 ± 0.1 m·s-1(mean ± SD). The values of selected ground reaction forces variables are presented in Table 1. The values of αS,αR, αrf, and αff were, respectively, 8.4 ± 6.5, 0.7± 7.4, 6.7 ± 6.1, and 6.0 ± 4.6 in the 32 subjects. The angle between RFA and FFA was always different between static (αS) and dynamic (αR) conditions and significantly lower in dynamic condition(αS > αR, P < 0.001) with a mean arch deformation of -7.8° ± 7.8°. It is worth noting that no significant correlation was found between αrf and αff (r = -0.01, P= 0.97). As shown in Table 2, αR was correlated positively with αS, αrf, arch deformation, Fz loading peak (seeFig. 5, upper), anteroposterior rate, mediolateral rate, and Mz during phase 1, and negatively with αff, stance time, heel orientation of FAP (β1), and total course of FAP (yD - yA) (seeFig. 5, lower). The rearfoot orientation αrf was correlated, positively with arch deformation, Fz loading peak, anteroposterior rate, Fx medial peak, mediolateral rate, and Mz-phase 2, and negatively with stance time and total course of FAP (yD - yA). The forefoot orientationαff was correlated positively with β1, β2, and Mz-phase 3.Figure 6
As suggested by Edington et al. (9), the 3.8 m·s-1 running velocity was chosen to make the results comparable to the common velocities used in the literature. It was checked that, within the defined range of velocity (from 3.4 to 4.2 m·s-1), no significant correlation was found between the running velocity and the other measured parameters. The measured ground reaction forces showed typical curves for heel-toe running and were comparable to those found by Cavanagh and Lafortune (6), Munro et al. (19), or Nachbauer and Nigg (20). Mz values were not in all subjects comparable to those observed by Holden and Cavanagh(13), but these authors used shoes specially designed to increase the valgus of the foot. Therefore, the differences may be due to the change between shoe and barefoot conditions. The mean value of footprint angles αrf and αff were of the same order of magnitude as the angle of the global foot observed by Holden et al. (12) in running. The mean stance time was 33 ms shorter than that observed by Munro et al. (19) on subjects running at the same velocity, but these authors used a 15-N threshold and they showed that the stance time they obtained was 15 to 20 ms greater than the stance time observed with the 50-N threshold used in the present experiment. This stance time difference may also due to the barefoot condition, as observed by Bates et al.(1) on subjects running at the same speed in both barefoot and shoe conditions. Data of the FAP cannot be compared to those found by Cavanagh and Lafortune (6) because these authors worked in shoe condition and they only localized the path of the FAP in order to classified rearfoot, midfoot, and forefoot strikers. The present data could therefore provide a basis of comparison for future studies.
It is worth to note that a limitation of this method comes from the fact that it can not be applied to complete forefoot strikers, when the heel doesn't print the paper. However, the method could be applied to three midfoot strikers. Furthermore, because the validity of footprint measurements could be limited by some possible errors on location of anatomical marks due to soft tissue influence, the use of x-rays was thought to be an improved way to connect the footprint with the foot anatomy.
Rearfoot and Forefoot Alignment
The lack of relationship between αrf and αff suggests that rearfoot and forefoot acted independently. Furthermore, the significant correlations observed between αR and αrf, and between αR andαff showed that 1) the more the foot was “open,” the more the heel was oriented close to the DOR, while the forefoot was laterally rotated and 2) the more the foot was “closed,” the more the forefoot was oriented close to the DOR, while the rearfoot was laterally rotated. Mobility between the rearfoot and the forefoot around the transverse tarsal joint was previously observed by several authors. Bojsen-Møeller(3) described movements about the secondary axis of the calcaneocuboid joint, whose direction allowed adduction and abduction of the rearfoot in relation to the fixed forefoot. Tardieu et al.(24), Ker et al. (14), Nachbauer and Nigg (20), and Scott and Winter(22) observed changes in height of the medial arch by flattening between the calcaneus and the midtarsal region. Stacoff et al.(23) used frontal plane kinematics to observe a torsion between the rearfoot and the forefoot about the longitudinal axis of the foot. As a consequence, the mobility between the rearfoot and the forefoot about the transverse tarsal joint seems to occur in three dimensions and could be observed 1) by frontal plane kinematics, as a calcaneal in/eversion and forefoot torsion; 2) by sagittal plane kinematics, as an arch flattening; and 3) in the present study, from the footprint as a lateral/medial rotation of the forefoot in relation to the rearfoot in horizontal plane. Consequently, future foot measurements may be encouraged to take into account this two-part constitution of the foot.
It was observed that the angle between the RFA and the FFA was always different between the static (αS) and the dynamic (αR) measurements. This suggests that a lateral rotation of the forefoot occurred in relation to the rearfoot. Dahle et al. (7) described that, in the “pronated foot” the forefoot was abducted in relation to the rearfoot, and in the “supinated foot,” the forefoot was adducted in relation to the rearfoot. More recently, Freychat et al.(10) observed on seven subjects, that αR was correlated with the rearfoot angle (angle of eversion/inversion of the calcaneus in relation to the leg). However, the pronation and supination phenomenon was only partly connected to respectively open and closed foot actions (r2 = 0.6). Nevertheless, the angle between the rearfoot and the forefoot on footprints (αR) could supply a quantitative measurement connected to the range of pronation and supination of the foot.
Bates et al. (1) observed that the foot gets into pronation during the first phase of foot contact and after, into supination during the propulsion phase. Holden et al. (12) observed from footprints that, when running on a waxed surface, the foot gets first into abduction during the first third of contact, and after, into adduction during the last third of contact. The concept of horizontal rearfoot/forefoot mobility completes these previous observations: during the first pronation phase, the lateral rotation of the forefoot relative to the rearfoot (open foot) forced the foot into abduction, while during the last supination phase, the medial rotation (closed foot) replaced the foot close to the direction of running.
Foot Alignment and Dynamic Behavior
Significant relationships were observed between the orientation of both the RFA and the FFA determined on the footprint (αR, αrf, andαff), and the path of force application point (β1, β2, and yD- yA), suggesting that these footprint measurements could be connected with functional parameters. These results are different from those of Nachbauer and Nigg (20), where no significant relationships were observed between arch height, arch flattening, and ground reaction forces. In fact, contrary to arch height, which can be considered as an architectural factor, the rearfoot/forefoot angle could be connected to pronation, which is a functional factor, and may better influence the dynamics of foot contact and the ground reaction forces.
On the one hand, a “closed foot” behavior was related to both a lateral orientation of the rearfoot and a lateral direction of the FAP on the heel area (β1). Because of the lateral orientation of the rearfoot, the rounded surface of the heel could roll laterally on the ground, and this may explain the larger mediolateral force rate and vertical free moment Mz produced during the first and second phase of foot contact. The lateral orientation of the rearfoot may also favour the inversion of the heel and the supination of the foot. As described by Mann et al. (16) and Nigg et al. (21), the inversion of the heel may be transferred to the external rotation of the leg. Bojsen-Møeller(2) observed in the inverted foot that the dorsiflexion of the toes occurred mainly about the second to the fifth metatarsophalangeal joint. The short lever arm between the rear extremity of the calcaneus and the axis of toe flexion was called “low gear.” Due to this short lever arm, the muscular action during the flexion-extension of the foot could then be faster and the transfer between braking and propulsion phase may occur more quickly. As a consequence, greater anteroposterior rate and shorter course of the FAP may explain the reduction of stance time in “closed foot” behavior.
On the other hand, in the “open foot” behavior, the lateral orientation of the forefoot was confirmed by the change in FAP orientation between the rearfoot and the forefoot areas (β2). Because of an alignment of the rearfoot in relation to the DOR, the rounded surface of the heel rolled forwards and inward on the ground and favor the eversion of the foot, which could be transferred to the internal rotation of the leg. Bojsen-Møeller (2) observed in everted foot that the dorsiflexion of the toes occurred about the first to the second metatarsophalangeal joint. The long lever arm between the rear extremity of the calcaneus and the axis of toe flexion was called “high gear.” An “open foot” behavior was correlated to an extension of the FAP course (yD - yA). Both longer lever arm and longer course of the FAP may increase the stance time in “open foot” behavior.
Mechanism of Vertical Force Reduction
The reduction of Fz loading peak (P < 0.001) related to a decrease ofαR suggests that the medial/lateral rotation of the forefoot could be associated to a stiffness control mechanism of the arch of the foot. According to Ker et al. (14), the elastic property of the arch of the foot depends on the stretching of the plantar aponeurosis and ligaments. The lateral rotation of the forefoot increases the distance between the calcaneal and the metatarsal insertions of the plantar aponeurosis(Fig. 5a), and therefore could contribute to a better utilization of elastic structure of the arch and to a reduction of vertical loading, Bojsen-Møeller (3) and Lorenzton(15) suggested that such eversion of the foot may improve the flexibility of the arch and the shock absorption mechanism by unlocking the calcaneocuboid joint and increasing the internal leg rotation. More recently, Messier et al. (17) observed that greater eversion of the heel was associated to higher shock absorption necessity in obese subjects.
Furthermore, the arch deformation observed between static (αS) and dynamic (αR) loading conditions could reflect the ability of the arch to stretch its elastic structures. As a consequence, the “open foot” behavior, related to a greater arch deformation between αS and αR, may be considered as a “flexible-spring” foot(Fig. 5a), while the “closed foot” behavior, related to a smaller deformation, may be considered as a“stiffer-spring” foot (Fig. 5b). In“open foot,” lower vertical loading and longer stance time may be more suitable for long distance running; but as suggested by Nigg et al.(21), the transfer of more eversion to internal leg rotation may increase the risk of knee injuries. Conversely, higher vertical forces and shorter stance time suggest that the “closed foot” behavior may be more suitable for fast running but could also increase the risk of injuries because of greater vertical loading and lateral instability. However, further investigations should be performed in order to confirm whether the “open foot” and “closed foot” behaviors could be associated with different lower limb stiffness.
The footprint method described in this paper has the advantage of using very simple material and resources and could be applied to provide quantitative measurements for characterization of the dynamic behavior of the foot. During running, the “open” and “closed foot” behaviors, defined by means of footprint, were characterized by specific spatial orientation of the rearfoot and the forefoot as well as specific dynamic behavior. Each lateral and medial rotation of the forefoot in relation to the rearfoot may be related to typical pronation and supination of the foot. The results also reveal the ability of the foot to change its horizontal bending from static to dynamic conditions and it is suggested that “open foot” and “closed foot” behaviors could also be related to different foot flexibilities.
1. Bates, B. T, L. R. Osterning, B. Mason, AND S. L. James. Lower extremity function during the support phase of running
. In:Biomechanics VI-B
, E. Asmussen and K. Jorgensen (Eds.). Baltimore: University Park Press, 1978, pp. 30-39.
2. Bojsen-Møeller, F. The human foot: a two speed construction. In: Biomechanics VI-A
, E. Asmussen and K. Jorgensen(Eds.). Baltimore: University Park Press, 1978, pp. 261-266.
3. Bojsen-Møeller, F. Calcaneocuboid joint and stability of the longitudinal arch of the foot at high and low gear push off.J. Anat.
4. Braun, S., L. Basquin, AND C. Mery. Considérations sur l'assise plantaire normale: étude statistique. Rev. Rhumatol.
5. Campbell, M. J. and D. Machin. Correlation and linear regression. In: Medical Statistics: A Commonsense Approach
, M. J. Campbell and D. Machin (Eds.). Chichester, West Sussex, U.K.: John Wiley & Sons, Ltd., 1990, p. 91.
6. Cavanagh, P. R. and M. A. Lafortune. Ground reaction forces in distance running
. J. Biomech.
7. Cavanagh, P. R. AND M. M. Rodgers. The arch index: a useful measure from footprints. J. Biomech.
8. Dahle, L. K, M. Mueller, A. Delitto, AND J. E. Diamond. Visual assessment of foot type and relationship of foot type to lower extremity injury. J. Orthop. Sports Phys. Ther.
9. Edington, C. J., E. C. Frederick, AND P. R. Cavanagh. Rearfoot motion in distance running
. In: Biomechanics of Distance Running
, Champaign, IL: Human Kinetics Books, 1990, pp. 135-164.
10. Freychat, P. F., A. Belli, J. P. Carret, AND J. R. Lacour. Rearfoot and forefoot alignment during running
. XIVth Congress of the International Society of Biomechanics, Paris, 1993, pp. 424-425.
11. Hamill, J., B. T. Bates, K. M. Knutzen, and G. M. Kirkpatrick. Relationship between selected static and dynamic lower extremity measures. Clin. Biomech.
12. Holden, J. P., P. R. Cavanagh, K. R. Williams, and K. N. Bednarski. Foot angles during walking and running
. In: Biomechanics IX-A
, D. A. Winter, R. W. Norman, R. P. Wells, K. C. Hayes, and A. E. Patla(Eds.). Champaign, IL: Human Kinetics Publishers, 1985, pp. 417-422.
13. Holden, J. P. and P. R. Cavanagh. The free moment of ground reaction in distance running
and its changes with pronation
. J. Biomech.
14. Ker, R. F., M. B. Bennett, S. R. Bibby, R. C. Kester, and R. McN. Alexander. The spring in the arch of the human foot.Nature
15. Lorenzton, R. Causes of injuries: intrinsic factors. In: The Olympic Book of Sports Medicine
, Vol. 1, A. Dirix, H. G. Knuttgen, and K. Tittel (Eds.). Cambridge, MA: Blackwell Scientific Publications, 1988, pp. 376-390.
16. Mann, R. A, D. E. Baxter, AND L. D. Lutter. Running
symposium. Foot Ankle
17. Messier, S. P., A. B. Davies, D. T. Moore, S. E. Davis, R. J. Pack, and S. C. Kazmar. Severe obesity: effects on foot mechanics during walking. Foot Ankle
18. Montagne, J., A. Chevrot, AND J. M. Galmiche. Techniques radiographiques: repères et mesures. In: Atlas de Radiologie du Pied.
Paris: Masson, 1980, pp. 44-46.
19. Munro, C. F., D. I. Miller, AND A. J. Fuglevand. Ground reaction forces in running
: a reexamination. J. Biomech.
20. Nachbauer, W. AND B. M. Nigg. Effects of arch height of the foot on ground reaction forces in running
. Med. Sci. Sports Exerc.
21. Nigg, B. M., G. K. Cole, AND W. Nachbauer. Effects of arch height of the foot on angular motion of the lower extremities during running
. J. Biomech.
22. Scott, S. H. AND D. A. Winter. Biomechanical model of the human foot: kinematics and kinetics during the stance phase of walking.J. Biomech.
23. Stacoff, A., X. Kaelin, E. Stuessi, AND B. Segesser. The torsion of the foot in running
. Int. J. Sport. Biomech. 5: 375-389, 1989.
24. Tardieu, C., C. Tabary, AND G. Tardieu. Mesure strobo-photogrammétrique des angles tibio-calcanéens et de voûte plantaire. J. Biophys. Biomec.
Keywords:©1996The American College of Sports Medicine
FOOTPRINT; FOOT ALIGNMENT; PRONATION; FORCE PLATFORM; RUNNING