BIOMECHANICAL INFLUENCE OF RUNNING SHOES WITH MIDSOLE HOLLOW STRUCTURE ON LOWER LIMBS DURING RUNNING

1. Introduction

It is clear that running shoes play a central role in running as an intermediary between the foot and the ground, which affects the biomechanical performance of human running² and the risk of injury.³ Running shoes with advanced footwear technology can significantly improve sports performance; for example, running shoes with thicker midsoles and mechanical structures can significantly reduce the work of the ankle joint,⁴ thereby reducing the workload on calf muscles.

Besides the training load,⁵ running-related sports injuries are closely related to the runner’s biomechanical movement pattern.⁶ Injury is most likely the moment your foot strikes the ground while running. There are generally three types of running landings: (1) rearfoot strike; (2) midfoot/forefoot strike; and (3) forefoot strike. Sprinters often land on the forefoot, but more than 93.67% of contemporary shoe marathon runners fall on the rearfoot.^7,8 The rearfoot-strike running style produces a typical vertical ground reaction force (GRF) force–time curve with two peaks (impact peak and active peak), as depicted in Fig. 1. The impact peak (Fz1) is passively borne by the runner. The runner who strikes the rearfoot must repeatedly deal with the instantaneous impact of the vertical GRF within the first 50ms (time to Fz1, tFz1) of support, which is about 1.5–3 times the sudden impact force of the body weight.⁹

Research on long-distance running shows it can lead to joint overload and a variety of other running-related injuries.¹⁰ The transient impacts associated with rearfoot strike running are sudden forces with high rates and loads that propagate rapidly upward through the body and may result in a high incidence of running injuries, such as plantar fasciitis and knee injuries.¹⁰ A meta-analysis¹¹ revealed that runners with a history of stress fractures exhibited higher vertical GRF loading rates. Human heel pad,¹² synovial fluid, knee flexion amplitude, joint positioning,¹³ and muscle activity¹⁴ can also cushion transient impact, but it is difficult to change due to biomechanical or morphological constraints. The midsole properties of the shoes can effectively attenuate the impact force transmitted to the human body. Specific properties of the midsole system can alter an individual’s gait pattern,⁹ and these changes are caused by sensory mechanisms at the foot–ground interface during foot strike, which the body perceives as a source of potential injury. Recreational and professional athletes are protected by cushioning systems in technical footwear to mitigate these transient forces. Modern running shoes aim to allow rearfoot strike to run comfortably and with less injury by using elastic materials or structures in the rearfoot to absorb some of the transient forces and spread the impulse over more time,¹⁵ i.e., reduce the loading rate (LR).

Reducing the first peak impact force of running shoes, the Fz1 loading rate, prolonging the Fz1 appearance time, reducing joint work, and lowering body energy consumption have always been important concerns for shoe designers and manufacturers. Running shoes with a midfoot hollow structure have demonstrated a significant reduction in ankle joint work.⁴ The main possible explanation is that the hollow structure with deformation provides more energy rebound during the stretching phase.

This research stems from an identified gap in the existing literature that provides limited insight into how midsole structural modifications directly affect lower limb biomechanics. As such, the specific hypotheses that will be tested in this study are enumerated as follows:

In sum, this study is based on the presumption that different midsole hollow structures have measurable effects on the biomechanics of the lower limbs in runners.

2. Materials and Methods

2.1. Participants

A priori power analysis was calculated by G*power (version 3.1.9.7; Heinrich Heine University, Germany) with a power of 0.8, indicating that a minimum number of 15 participants were required. Therefore, we recruited 21 male (32.33±6.27 years, 66.14±3.98kg, height 175.29±4.35cm, 4.69±1.47 training years, 61.43±26.91km per week) marathon enthusiasts with running habits to participate in the test, wearing running shoes (size US8.5–9.5). Subject inclusion criteria were as follows: (1) no lower extremity musculoskeletal injury; (2) rearfoot strikers while running; and (3) running proficiently on the treadmill. The study was performed in accordance with the ethical standards of the Declaration of Helsinki. Ethics approval was obtained from the Li Ning Institutional Ethics Committee. All subjects knew the purpose of this experimental study and signed informed consent.

2.2. Shoe conditions

In this study, two pairs of shoes (Hollow shoe1 and Hollow shoe2, US men size 9.0) were used. The midsole of the Hollow shoe1 (ARHQ262, Li-Ning, Beijing, China) running shoe is a hollow structure composed of upper and lower arched carbon plates. The upper carbon plate is a structure that protrudes upward at the arch of the foot, and the lower carbon plate is a structure that protrudes downward, see Fig. 2. Hollow shoe2 (ARRT017, Li-Ning, Beijing, China) has the same structure as Hollow shoe1. The difference is that, while maintaining the shape of the carbon plate, the thickness of the rearfoot and the thickness of the forefoot are increased, making the hollow structure higher and the span larger. Measurement parameters for the two hollow shoes are listed in Table 1. According to the F1833-97 (2017) standard of ASTM, all test samples were tried on for 8km of running to avoid the instability of the performance of the midsole structure and cushioning of the new shoes.

2.3. Procedure

After the subjects arrived at the scene, their height, weight, foot length, foot width, and other information were measured. Twenty-two reflective markers with a diameter of 9mm were placed: two markers on the anterior superior iliac spine, two markers on the posterior superior iliac spine, four markers on the right thigh, two markers on the right knee joint, four markers on the right shank, two markers on the right ankle joint, and six markers on the right foot, see Figs. 3(a) and 3(b). Electrode sheets with a diameter of 2cm containing silver chloride conductive silica gel were placed on the right shank tibialis anterior (TA), medial of gastrocnemius (MG), lateral of gastrocnemius (LG). According to SENIAM (Surface EMG for Non-Invasive Assessment of Muscles) and Noraxon MR3.16 recommendations for electrode pad placement, each muscle group of the subject’s body was positioned.¹⁶ When affixing the electrode sheets, they were placed on the muscle belly along the muscle fiber direction with a distance of 2cm between the centers of the electrodes¹⁷ (see Figs. 3(c) and 3(d)).

Before the start of the experiment, the subjects wore their own shoes and performed a 5-min warm-up on a level treadmill at a personal pace. Before the test, all subjects wore hollow shoes and ran on the treadmill at a comfortable speed for 15min to adapt to the new structure. The subjects randomly put on the Hollow shoe1/2 and ran steadily on the 3D force plate-treadmill (Fully Instrumented Treadmill, BERTEC, Columbus, USA) at a speed of $3.30\pm 0.02$m/s for 1min, then began to continuously and synchronously collect the kinematics (200Hz, 9 Vicon_T40S camera, VICON, Oxford Metrics Ltd., Oxford, UK), surface EMG (2000Hz, Ultium EMG, NORAXON, Scottsdale, USA), and GRF data (1000Hz, BERTEC, Columbus, USA) of the right lower limb for 10 gait cycles. Then, the Hollow shoe2 was replaced with an interval of 5min between the two shoes. Ten subjects wore Hollow shoe1 first, and the other 11 subjects wore Hollow shoe2 first. The collected data included lower limb joint angle, RoM, moment, work, surface electromyography (sEMG), and GRF.

2.4. Data processing

The marker track and GRF data were collected synchronously by VICON NEXUS software (Oxford Metrics Ltd., Oxford, UK), and the sEMG data were synchronously collected by MR3.16 software (Noraxon USA Inc., Scottsdale, USA). After naming the marker points with the GRF data, they are imported into the Visual3D software (C-Motion Inc., Germantown, USA) to calculate the required indicators, and the sEMG activation characteristics are calculated by the MR3.16 software. Kinematic and GRF data were filtered using Butterworth fourth-order low-pass filters with cut-off frequencies of 12Hz and 35Hz, respectively. To facilitate statistical analysis among different subjects, all force-related parameters were normalized to subject body weight. The sEMG amplitude signature data were normalized using the maximum normalization method.¹⁸ Landing and take-off are calculated through the threshold of 30N vertical GRF, and the critical point from the braking phase to the push phase is calculated through the anterior–posterior GRF. Joint angles, RoM, moment, and work were calculated by Visual3D, with moments and work determined through the inverse dynamic model method. Joint zero degree was defined as joint angles in static standing. Knee extension and ankle dorsiflexion were collectively defined as positive values. The sEMG data were subjected to Butterworth fourth-order band-pass filtering and full-wave rectification at 20–500Hz. RMS envelopes with a window length of 50ms were calculated to reproduce muscle activation patterns.

2.5. Data analysis

In this study, statistical calculations were performed on the mean values of the collected data from 10 right-foot running cycle, and descriptive statistics expressed as mean±standard deviation (Mean±SD). For comparative statistics, paired t-tests were performed on kinematics, kinetic, and sEMG data for the two shoe conditions that fit the normal distribution assumptions. Otherwise, nonparametric tests (Wilcoxon signed-rank tests) were used. Since the above analysis method only performs statistical analysis of 0D parameters for key variables, for the analysis of 1D parameters of kinematic, dynamic, and sEMG activation characteristics, statistical parameter mapping (SPM) technology was used, and SPM technology is more suitable for forming probabilistic conclusions about 1D biomechanical trajectories.¹⁹ Traditional statistical methods often require a priori determination of discrete events for analysis, which can introduce bias. Conversely, SPM allows for an unbiased exploration of data, as all points in the trajectory are treated equally. SPM analysis was calculated using SPM1D v0.4 for Matlab (www.spm1d.org, Patakyetal.2015). The 0D data were statistically processed by SPSS 20.0 statistical software, and the significance level of 0D and 1D statistical tests was set at 0.05.

2.6. Mechanical testing protocol

To evaluate the cushioning and rebound properties of both pairs of shoes, mechanical shock tests were also performed. The impact test was carried out on the rearfoot and forefoot of running shoes, respectively, referring to the F1614-99 standard revised and released in 2006 by ASTM < Standard Test Method for Shock Attenuating Properties of Materials Systems for Athletic Footwear > and the F1976-06 standard revised and released in 2006 < Standard Test Method for Impact Attenuation Properties of Athletic Shoes Using an Impact Test >. To evaluate the cushioning and rebound performance of the midfoot structure, the impact test was also conducted on the maximum convex position of the structure according to these standards. This test method is limited to 1D force in the vertical direction. In real running, the contact between the foot, shoe and ground produces a 3D force, with variations in force loading during different stance phases. Additionally, different landing techniques and individual runners exhibit varied interactions among feet, shoes and the ground. Although limited, this simplified test method reflects the midsole’s mechanical energy storage and rebound performance in directions related to the runner’s spring-mass behavior.²⁰

3. Results

For the test results, one subject landed on the forefoot, one subject landed on the whole palm, and the kinematic data of one subject were lost in the test results. The above three subjects were excluded from the statistical results. Finally, 18 subjects were included in this study for statistical processing.

Mechanical tests showed (Table 2) that Hollow shoe2 had lower impact peaks at the rearfoot, midfoot, and forefoot compared to Hollow shoe1. The rearfoot and midfoot deformation of Hollow shoe2 was greater than that of Hollow shoe1, and the midfoot energy return of Hollow shoe2 was slightly higher by 3% compared to Hollow shoe1.

The tFz1 of Hollow shoe2 is significantly longer than Hollow shoe1 ($p<0.01$), and the maximum loading rate ($p<0.01$) and average loading rate ($p<0.05$) of Fz1 of Hollow shoe2 are significantly lower than Hollow shoe1 at 13–14% during the touchdown phase (Table 3 and Fig. 4 (left)).

The contact time ($p<0.01$), cushioning time ($p<0.01$), and push time ($p<0.01$) of Hollow shoe2 are significantly longer than Hollow shoe1. The step size of Hollow shoe2 is significantly longer than that of Hollow shoe1 ($p<0.01$) (Table 3).

Hollow shoe2 ankle and knee joint work, medial–lateral direction moment and ankle introversion and introversion RoM did not find significant differences (Table 3). TA muscle activation profile in Hollow shoe1 was significantly lower than Hollow shoe1 during 68–72% of the push phase (Fig. 4 (right)). No significant differences were found in activation profiles in other muscles.

4. Discussion

The main finding of this study is that the Hollow shoe2, with a thicker midsole and a larger hollow structure, can significantly increase the time to peak impact force, reduce the maximum and average loading rate, and increase the breaking time, push time, and contact time. Additionally, a significant reduction in the activation characteristics of the TA muscle during the push phase was observed.

Possible explanations for these results are as follows: At the initial stage of landing during the braking phase, Hollow shoe2 can provide more midsole deformation to cushion the initial impact (with the rearfoot deformation occurring not only vertically but also potentially forward), thus delaying Fz1 significantly. This delay provides more reaction time to the muscles around the ankle and knee joints to stabilize the joints. Contrary to expectations, Fz1 showed an increasing trend, although it was not significant. Similar studies on softer soles have found an increased tFz1 without significant changes in Fz1.²¹ The thicker midsole also attenuates impact but might reduce plantar sensationst.²² The prolonged Fz1 time resulted in a significant decrease in the maximum and average loading rates, which can positively affect running, as increased loading rates have been associated with injury in some studies.^23,24 The maximum loading rate usually occurs before Fz1,²⁵ and both maximum and average loading rates are crucial indicators for evaluating the cushioning of running shoes. Studies have shown²⁶ that stress fractures are significantly related to the instantaneous and average vertical GRF loading rates, and reducing the vertical GRF loading rate can lower the risk of stress fractures. This is especially beneficial for novices, who are more susceptible to high load-rate-related injuries; hence, the Hollow shoe2 may be more effective in reducing such injuries. Thicker running shoes²⁷ have been claimed to reduce impact-related injuries. Mechanical test results show that Hollow shoe2 midsole provides more deformation, with a corresponding increase in the time to reach maximum deformation, resulting in a significant increase in Hollow shoe2 braking time (consistent with mechanical test results). In this case, although there is no difference in energy return between Hollow2 and Hollow1, Hollow shoe2 stores more mechanical energy in the rearfoot and midfoot. Studies on advanced running shoes with thicker midsoles have shown²⁸ that the increased joint negative work may utilize greater midsole energy storage compared to the minimal shoe. Similarly, the increased ankle negative work of Hollow shoe2 indicates that the energy stored in the midsole may be greater than that in Hollow shoe1, even though the difference is not significant.

In the push phase, the compressed and deformed midsole gradually recovers to generate energy return. Due to the greater deformation of the midsole than the Hollow shoe1, the energy return time of Hollow shoe2 is significantly longer than Hollow shoe1 in the push phase. The midsole of Hollow shoe2 returns more energy than Hollow shoe1, which reduces the activation characteristics of TA during the push phase (Fig. 4 (right)), and decreases the positive work of the ankle and knee joints, although the difference is not significant (but there is a decreasing trend). Studies have indicated²⁹ that footwear structures can store and return sufficient energy at the correct timing and location to enhance runners’ performance. In this study, the lack of significant reduction in ankle joint work for Hollow shoe2 compared to Hollow shoe1 might be due to its failure to improve the timing or location of energy return significantly. Another possible reason is that the subjects did not fully adapt to the shoe structure. Despite a 15-min period, it may be relatively short. A 6-month follow-up study on shoe adaptation³⁰ showed that subjects’ lower limb kinematic parameters underwent adaptive changes. A different result might be observed if the subjects were adequately acclimated to the shoes. Previous studies⁴ have shown that Hollow shoe1 can significantly reduce the positive work of ankle joint compared to traditional shoes with a whole piece of midsole material. Based on this, we infer that although Hollow shoe2 does not show a significant reduction in ankle positive work compared with Hollow shoe1, it likely reduces ankle work compared to traditional shoes and should produce similar results. Therefore, shoes with this type of midsole structure may have lower energy consumption than traditional shoes, as they significantly reduce the positive work of ankle joints and the activation characteristics of certain lower limb muscle groups. Further research is required to evaluate energy consumption while running in these shoes.

The above results combine to increase the contact time, push time, and flight time (although the difference is not significant) of the Hollow shoe2, ultimately leading to a significant increase in step length. McMahon et al.³¹ showed that running on compliant surfaces resulted in longer step length compared to hard surfaces because the bending angle of the hip joint is greater, and also found longer ground contact times. During the test in this study, the subject’s subjective feedback indicated that Hollow shoe2 was more compliant than Hollow shoe1 during the stance phase. Compliance here can be understood as the smoothness of the process from landing to take-off. The longer step length observed with Hollow Shoe2 supports this point. The contact time for Hollow shoe2 is longer than Hollow shoe1, aligning with the trend observed by Chambon et al.³² with increased midsole thickness. This study hypothesized that ground contact time would be longer due to a thicker midsole, consistent with previous studies. However, the result was contrary to expectations. The ground contact time in this study (232–239ms) is shorter than the other two studies (250–265ms),^27,32 likely due to differences between treadmill and overground running. Riley et al.³³ showed that runners tend to shorten stride length and increase stride rate while running on a treadmill. Regarding muscle activity, Hollow shoe2 showed temporarily lower activation in the TA during the propulsion phase compared to Hollow shoe1, but the RMS average or SPM analysis of other muscle groups did not show significant differences. This likely results from the structural similarity between the two shoes. When running conditions are similar, the body requires only minor adjustments in muscle activation to maintain movement, whereas significantly different conditions necessitate large adjustments.³⁴ This reflects an optimization strategy where the neuromuscular system adapts to conserve energy.³⁵

There are some limitations to this study. Unlike outdoor running, this research involved treadmill test in a laboratory environment, which may influence running kinematics, kinetics, and muscle activation.³⁶ Real outdoor running could be affected by various surfaces such as asphalt roads, dirt paths, or plastic tracks, as well as different running speeds, which might yield different results. Additionally, the study’s subjects were all adult men with running experience, which may not be representative of women, the elderly, or beginners in running.

5. Conclusions

Running shoes featuring a hollow midsole complex structure present a promising avenue for enhancing the design of athletic footwear. Within the bounds of this study, it has been observed that running shoes incorporating larger hollow structures in their midsoles could be more effective in mitigating damage associated with landing impact forces. Additionally, these structures may potentially contribute to decreasing the workload of lower limb muscles during running.

ORCID

Zongxiang Hu  https://orcid.org/0009-0002-0141-2683

Huanhuan Meng  https://orcid.org/0009-0003-4818-6019

Fengqin Fu  https://orcid.org/0009-0002-3489-6637

Haojie Huang  https://orcid.org/0009-0001-5122-2274

Tongtong Guo  https://orcid.org/0000-0001-9552-308X

Zhaowei Chu  https://orcid.org/0000-0003-3927-3666

Shengwei Jia  https://orcid.org/0000-0002-4559-8646

Fan Yang  https://orcid.org/0000-0001-7240-1936