Morphological analysis of the windlass mechanism during running

Abstract

[Purpose] To perform a morphological analysis of the foot windlass mechanism during running and compare it with previous studies. [Participants and Methods] Twenty healthy adults (14 males and six females) participated in this study. A VICON 3D motion-analysis device was used and the analysis used the medial longitudinal arch height model in addition to the full plug-in model. The medial longitudinal arch height, ankle joint internal plantar flexion moment, ankle joint angle, and movement of the center of gravity were measured. The focus was on running cycle synchrony and changes in the medial longitudinal arch height. A one-way analysis of variance for multiple comparisons was performed. [Results] The peak running cycles were 50.5% ± 0.6% for the ankle dorsiflexion angle, 54.4% ± 0.7% for the ankle moment, and 59.5% ± 1.7% for the medial longitudinal arch height, which were significantly different. The medial longitudinal arch showed a reverse windlass phenomenon, with a minimum of 17.1 mm at an approximately 60% running cycle and then a spike-like phenomenon of 29.1 mm at a 98% running cycle. [Conclusion] We did not observe synchronicity of any parameter, such as that observed during walking. The spike phenomenon that occurred during the 98% running cycle was the original windlass mechanism, which caused the push-off phenomenon.

INTRODUCTION

Anatomically, the medial longitudinal arch (MLA) is supported by the calcaneal, talus, navicular, medial cuneiform, and three medial metatarsal bones. The plantar fascia, calcaneonavicular (spring) ligament, medial tarsometatarsal joint, and extrinsic and intrinsic muscles of the foot help maintain the MLA height and shape. The MLA is thought to have passive factors, such as bones and ligaments, and active factors, such as muscles^1,2,3,4,5,6). Traditionally, passive factors have been thought to be predominant¹⁾. In recent years, the active factors are also being reconsidered^2,3,4). Considering the active factors, some groups claim that extrinsic muscles of the foot, such as the tibialis posterior^{6, 7)}, are involved, but other groups claim that intrinsic muscles of the foot, such as the abductor hallucis^2,3,4), are related to the MLA. Research reports on exercises for the intrinsic muscles of the foot include short foot exercises⁸⁾, and other exercises reportedly are related to the MLA include toe spread out^9,10,11), first-toe extension^{11, 12)}, second- to fifth-toes extension^{11, 12)}, lesser-toes exercise¹³⁾, and single-leg swing exercise¹⁴⁾. However, the intrinsic muscles of the foot are anatomically divided into 4 layers of 11 muscles on the plantar surface. It remains unclear which muscles of the intrinsic muscles of the foot are related to the MLA and how they are related to each other.

The dynamic kinematics of the MLA include a windlass mechanism, first described in 1954 by Hicks who stated that it is induced by the proximal pressure exerted by dorsiflexion of the proximal phalanx, resulting in elevation of the corresponding metatarsal row¹⁵⁾. Additionally, there are two mechanisms that support the MLA when load is applied to the lower limbs: a truss mechanism and a beam mechanism. The Truss mechanism supports the MLA through the plantar aponeurosis, which extends from the calcaneus tubercle to the metacarpophalangeal (MP) joint, the tendon sheath of the flexor tendon, and the base of each proximal phalanx^16,17,18,19). On the other hand, the Beam mechanism is generated by the action of the bones, joints, and ligaments¹⁶⁾. The MLA’s kinematic characteristics when weight is applied are determined by the interrelationship between the Beam and Truss mechanisms. When the Truss mechanism is degraded, the Beam mechanism compensates, over-compensation is a typical example of calluses.

The difference between walking and running is that walking usually occurs at 3–5 km/hour, has a double-support phase, joint movements are small and at a constant rhythm, and the center of gravity moves at a constant speed. Running generally occurs at ≥8 km/hour, has only a single-support phase, both legs are not in contact with the ground, and flexion and extension of the knee joint are noticeable. The center of gravity movement repeatedly undergoes deceleration and acceleration^20,21,22). Running is also thought to be an elastic locomotive movement that relies primarily on the storage and release of elastic strain energy in tendons and ligaments to reduce mechanical demands on the lower-limb muscles^{23, 24)}. As with walking, when running, the plantar fascia becomes taut due to dorsiflexion of the MP joint from mid to late stance, and arch elevation reportedly occurs simultaneously^{25, 26)}. Previous studies have not objectively described the running cycle (RC), but their waveform results have indicated that it was approximately 60% of the stance phase^{25, 27)}.

The recoil reaction utilizes the stretch reflex of the triceps surae and plantar muscles. Elastic energy is stored from the maximum dorsiflexion position, and this stored elastic energy is released during the swing phase. This reaction converts the movement of muscles and tendons when returning to their original length into propulsive force. This reaction exerts a force greater than that of the concentric contraction, resulting in more efficient running²⁸⁾. One study concluded that the recoil response also occurs during walking²⁹⁾, but in our previous study, the recoil response did not occur during walking³⁰⁾. The present study aim was to calculate the morphological analysis of the foot windlass mechanism during running and compare with those reported in previous studies.

PARTICIPANTS AND METHODS

The participants were 20 healthy persons (14 males and 6 females) who had no problem with toe movements and had not been in a medical institution due to injury near the ankle for the past 6 months (average age: 21.2 ± 0.4 years, average height: 166.4 ± 6.6 cm, and average weight: 59.4 ± 6.1 kg), general university students who participated in recreational exercise but had no particular habit of doing so. This study was performed with the consent of the Ethics Committee for Human Research of Gunma Paz University (Approval number: PAZ14-22). Written informed consent was obtained from each participant.

A three-dimensional motion analyzer Vicon MX (Vicon Motion Systems, Oxford, UK), nine infrared cameras (T10 Vicon Motion System, Nexus1.8.5), and three floor reaction-force meters (AMTI: ADVANCED MECHANICAL TECHNOLOGY INC., Watertown, MA, USA) were used as the measuring instruments, and a full plug-in model was used. The sampling frequency of the camera and ground reaction force was 200 Hz. The starting position for measurement was 3.5-m away from the floor reaction-force meter, and the ending position was also 3.5 m away from the floor reaction-force meter. Free running was performed three times with measurements, and the average value was calculated. The stance phase of running was normalized to 100%, and the internal plantar flexion moment, ankle dorsiflexion angle, and center of gravity movement were measured.

The center of gravity movement was calculated according to the method used in a previous study. The distance from the marker of the knee-joint axis to the center of gravity line was measured³⁰⁾. When the center of gravity was behind the knee axis, the value was negative, and when it was in front of the knee axis, the value was positive. Therefore, a positive value indicated a center of gravity more anterior to the knee axis, and a high value indicated forward, and a low value indicated backward (Fig. 1a).

Fig. 1.

Three-dimensional motion analyzer a) The calculation method is the distance from the marker of the knee joint axis to the center of gravity line. When the center of gravity is behind the knee axis, the value is negative, when the center of gravity line is in front of the knee axis, it is set as a positive value. b) Medial longitudinal arch (MLA) height. The calculation method is the distance between the calcaneus and the navicular bone with the first metatarsal as the bas. The calculation formula is below.

k= −(a • xp + b • yp + c)/(a • vx+b • vy) k=MLA height, a=ye −ys b=xs −xe c=−(a • xs + b • ys) L=√((xe−xs)2 + (ye−ys)2) vx=−ey=−(ye −ys) / L, vy=ex=(xe −xs) / L

The plug-in model alone does not allow for a morphological analysis of the windlass mechanism, and it is necessary to calculate the dynamic height of the MLA. The MLA height was calculated by adding three additional markers^31,32,33). Those markers were first applied to the calcaneus and first metatarsal at a height of 19 mm from the floor and then to the navicular bone. The MLA height was calculated as the distance between the calcaneus and navicular bone with the first metatarsal as the base (Fig. 1b). ICC values ​​for this model method have been reported to be 0.95 (within-day) and 0.94 (inter-day)³¹⁾.

The synchrony values of the ankle-joint angles, maximum ankle internal plantar flexion moment, and minimum MLA height were compared. The Shapiro–Wilk test was performed to assess the normality of the data distribution. After confirming normal distribution, a one-way analysis of variance was performed. Post-hoc multiple comparisons were made by performing Tukey’s HSD test to identify significant differences between groups. The significance level was set at p>0.05. IBM SPSS Statistics ver. 21 (IBM Corp., Armonk, NY, USA) was used to perform all statistical analyses.

The running sessions in this study were then classified into three phases: initial stance (0%–20%), mid-stance (20%–55%), and propulsion (55%–100%), according to the classifications by Welte et al²⁷⁾.

RESULTS

The mean running speed in this study was 4.48 m/s ± 0.22, resulting in a slow running motion.

In the initial stance (0%–20% RC), the center of gravity movement stagnates and the vertical component of ground reaction force becomes slightly disrupted. The ankle joint goes from dorsiflexion to slight plantar flexion before dorsiflexion. The ankle internal plantar flexion moment was similar to the ankle-joint angle, with a slight internal dorsiflexion moment being shown before the internal plantar flexion moment. The MLA height also showed a slightly low value before becoming a slightly high value. We observed a slightly reversed motion during the early stance phase of running, as described above. We also did not intentionally have the participants land their forefoot on the ground (plantar flexion) or rear foot on the ground (dorsiflexion), but 17 participants landed in dorsiflexion and 3 landed in plantar flexion.

In the mid-stance (20%–55% RC), there was also an increase in the ground reaction force, ankle dorsiflexion angle, and ankle internal plantar flexion moment as well as a decrease in MLA height. These phenomena initially appeared to be synchronized. However, the maximum peak RC values for each were 50.5% ± 0.6% for the ankle dorsiflexion angle and 54.4% ± 0.7% for the moment, and the minimum peak RC value for MLA height was 59.5% ± 1.7% in the next phase (the propulsion phase), (N=20, p<0.05, effect size=0.90, power=0.99: Fig. 2, Table 1).

Fig. 2.

Medial longitudinal arch (MLA) height (mm).

VGRF: vertical component of ground reaction force; AA: ankle dorsiflexion angle (°); GM: center of gravity movement (mm); AM: ankle joint internal plantar flexion moment (Nm); MH: MLA height (mm).

Table 1. Comparison between peak running cycle values.

	AAPC**	AMPC**	MHPC**
Peak (%)	50.5 ± 0.6%	54.4 ± 0.7	59.5 ± 1.7

Mean ± SD. **p<0.01.

AAPC: Ankle angle peak running cycle (%); AMPC: The internal plantar flexion Ankle moment peak running cycle (%); MHPC: MLA height minimum peak cycle (%).

Statistical analysis was performed by comparing each change amount with ANOVA. Post hoc multiple comparisons were performed using Tukey’s HSD test, all of which showed significant differences (N=20, p<0.05, effect size=0.90, power=0.99).

In the propulsion phase (55%–100% RC), the ground reaction force decreased, and the ankle-joint dorsiflexion angle and ankle-joint internal plantar flexion moment also decreased. The MLA height reached a minimum of 17.1 mm at 59.5% ± 1.7%. Subsequently, as the swing phase approached, it rose upward, and a spike-like phenomenon appeared, reaching 29.1 mm at 98% of the propulsion phase. This spike phenomenon occurred in 8 of the 20 participants.

DISCUSSION

In this study, we measured the morphological changes in MLA height rather than the tension of the plantar fasciitis to analyze the windlass mechanism. Furthermore, raw data measured during real running rather than data from computer models or programs were used in the analysis. First, regarding the synchronicity of each parameter, the maximum peak RC values (local maximum values) of the ankle dorsiflexion range of motion and ankle joint internal plantar flexion moment were 50.5% ± 0.6 and 54.4% ± 0.7, respectively. The minimum peak RC value (local minimum value) of the MLA height was 59.5% ± 1.7. All peak RC value periods were significantly different. Regarding the center of gravity movement, a previous walking study showed a peak value of forward movement at approximately 50% walking cycle, and synchronicity with other parameters was observed³⁰⁾. On the other hand, during running, the forward movement stagnated until around 20% RC, after which it only moved linearly forward, and there was no peak value itself. The synchrony observed in walking in a previous study³⁰⁾ for ground reaction force, dorsiflexion range of motion, ankle moment, maximum peak value of center of mass shift, and minimum peak value of MLA height was not observed in running. Therefore, running does not show any reflex phenomena, such as is observed in walking.

Next, when divided into each phase, in the initial stance (RC: 0%–20%), the ground reaction force was slightly unstable, the ankle-joint angle changed from plantar flexion to dorsiflexion, the ankle-joint internal plantar flexion moment changed from dorsiflexion to plantar flexion moment, and the MLA height showed a slightly low value followed by a slightly high value, resulting in slight reverse movements occurring within a short period of time. Furthermore, the center of gravity movement stagnated. These characteristics are thought to be due to the shock-absorbing effect.

In the mid-stance phase (20%–55% RC), The ground reaction force increased, the ankle joint dorsiflexion angle, and the ankle joint internal plantar flexion moment also increased, and the MLA height decreased. At first glance, this is a phenomenon similar to that of walking. However, the peak RC values of each factor above showed asynchrony. Synchrony similar to that of the walking motion³⁰⁾ could not be achieved. In particular, the MLA height reached a minimum value (local minimum value) during the next phase (the propulsion phase).

In the propulsion phase (56%–100% RC), the ground reaction force decreased, and the ankle-joint dorsiflexion angle and ankle joint internal plantar flexion moment also decreased, MLA height was the minimum peak value 17.1 mm at 59.5% RC. This situation may be due to the reverse windlass phenomenon. Hicks refers to this phenomenon as the reverse windlass effect, explained that it occurs when the MLA height decreases, but the strength of the toe muscles compensates¹⁵⁾. Subsequently, as the swing phase approached, the MLA height increased, and a spike-like windlass mechanism of 29.1 mm at 98% RC appeared in eight participants. Seven of the eight who had spike-like phenomena (approximately 41%) had their heel strike (RC0%) in dorsiflexion, and the remaining one (approximately 33%) had their heel strike land in plantar flexion. No relationship was found between the phenomenon and heel strike. Similar findings have been reported in previous studies²⁵⁾. Focusing on running speed, the spike phenomenon positive group had an average of 4.79 m/s ± 0.12, while the spike phenomenon negative group had an average of 4.32 m/s ± 0.25, suggesting that there may be a relationship between running speed and the spike phenomenon, which is an issue for future research.

In previous studies, no clear period was stated, but the windlass mechanism was found to have a RC of approximately 60% from the figures of their results^{25, 27)}. However, in this study, it was 98%. In previous studies, the tension of the plantar fascia was calculated by defining it as a windlass mechanism, but in this study, the change in MLA height was directly analyzed morphologically and defined as such, which is thought to explain why the results differ. The windlass mechanism refers to the winding up phenomenon of the MLA, or the morphological change in the MLA, so we believe that the calculation method used in this study is appropriate. Furthermore, the spike windlass phenomenon occurs in some participants but not in others. This phenomenon occurs in only 0.03 to 0.04 s, and the influence of the stretch reflex due to external information is also quite conceivable. Therefore, it would be difficult to reproduce this phenomenon in models or simulations. From the above, we believe that the reverse windlass mechanism works at approximately 60% RC during the propulsive phase and the MLA height then increases and that the spike-like phenomenon that occurs at 98% RC is the windlass mechanism and push-off phenomenon.

However, we cannot exclude the possibility that these small spikes are artifacts caused by sudden skin movements, especially since the method for calculating MLA height in this study did not use a filter. However, it is impossible to overlook the fact that many participants had sudden movements that induced stretch reflexes in the same period. This was also the case in our previous report on walking³⁰⁾. There are other limitations to this study as well. We could not clarify that this spike windlass mechanism did not appear in all participants. Although the mechanism needs to be analyzed in the future, it may be due to the slow running speed. The data for this study was slightly biased towards males because there were fewer females. Furthermore, we could not measure elderly people and the need to study the spike windlass phenomenon in elderly people.

In conclusion, when comparing the running motion in this study with that in our previous walking study, it is natural that the ground reaction force changed from a bimodal to a unimodal waveform, but it is important to also note the ankle joint dorsiflexion angle and moment, change in MLA height, and asynchrony of the center of gravity movement. From a developmental perspective, walking is a reflex action that develops as motor learning in the brain and is, perfectly efficient. It is likely that this reflex action creates the synchrony of walking movements. However, this perspective suggests that running is not as structured as walking in the brain as a reflex.

Push-off occurs during the pre-swing late in the stance phase of walking²⁹⁾; however, the push-off phenomenon has been considered from the perspective of running not walking^{21, 22)}. Running and walking are completely different movements both kinematically and developmentally, so it is misleading to consider them as similar movements. Therefore, although there is a push-off reaction in running, it does not necessarily mean that a similar phenomenon occurs in walking.

In addition, the recoil reaction is defined as “an explosive force that exerts a force greater than muscle strength”, Therefore, we classify the recoil reaction and push-off phenomenon as similar, but not synonymous, with the recoil reaction, which has a much greater force than the push-off phenomenon. Ker et al.¹⁹⁾ reported that the recoil reaction energy is approximately 17% of the total mechanical energy consumed when running at 4.5 ms−1, and Stearne et al.³⁴⁾ reported that the recoil reaction energy is only 8% when running at 2.7 ms−1. In other words, while running, there is more of a push-off effect than a recoil reaction. Additionally, in this study, the waveform was only a slight spike phenomenon. The second windlass phenomenon that we analyzed in walking is much larger³⁰⁾. It is questionable whether or not such a movement can be shown in a human running motion. Some elite athletes might have a reaction called a “recoil reaction”, but we believe that this expression should be used with caution.

Conflict of interest

Authors declare no conflicts of interest associated with this manuscript.