Gait & Posture 40 (2014) 504–509
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Center of pressure trajectory differences between shod and barefoot running James Becker a , Eric Pisciotta a , Stan James b , Louis R. Osternig a , Li-Shan Chou a, * a b
Motion Analysis Laboratory, Department of Human Physiology, University of Oregon, Eugene, OR 97403-1240, USA Slocum Center for Orthopedics and Sports Medicine, Eugene, OR, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 November 2013 Received in revised form 11 June 2014 Accepted 16 June 2014
This study examined differences in center of pressure (COP) trajectories between shod and barefoot running. Ten habitually shod runners ran continuous laps under both shod and barefoot conditions. The COP trajectory was calculated in the global coordinate system but then transformed to the anatomic coordinate system of the foot. The anterior–posterior and medio-lateral positions and excursions of the COP, as well as the most medial location and percent stand at which it occurred were examined. Additionally, external eversion moments and ground reaction forces were assessed. Compared to the shod condition, in the barefoot condition the COP was located more anteriorly early in stance and the COP was located significantly more medially at most time points across stance. There were no differences in external eversion moments during early stance or peak ground reaction forces between conditions. Future studies on mechanical or epidemiological differences between shod and barefoot running may find the COP trajectory an informative parameter to examine. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Center of pressure Running Barefoot running Eversion
1. Introduction The location of the center of pressure (COP) is a commonly used measure for assessing dynamic foot function during running. COP trajectories have been used to examine differences in foot function between high and low arches individuals [1], as a measure for quantifying mediolateral foot stability [2], and as a tool for prescribing and assessing the effects of shoe or orthotic interventions [3]. Analysis of COP trajectories is also informative regarding risks for sustaining common overuse running injuries such as Achilles tendinopathy [4] or exercise related lower leg pain [5]. It is estimated that between 25 and 75% of runners will sustain an overuse injury in any one year period of time [6,7]. Barefoot running has been suggested as a one potential method for reducing the incidence of running related injuries [8–10]. However, to date most studies comparing barefoot and shod running have focused on kinematics [9,11], external ground reaction forces [10], or joint moments [12]. In regards to COP trajectories in particular, it has been suggested that, compared to shod running, the foot position at initial contact during barefoot running results in smaller lever arms for the vertical and mediolateral ground reaction force, thus
* Corresponding author. Tel.: +1 541 346 3391; fax: +1 541 346 2841. E-mail address: [email protected] (L.-S. Chou). http://dx.doi.org/10.1016/j.gaitpost.2014.06.007 0966-6362/ ã 2014 Elsevier B.V. All rights reserved.
reducing the external eversion moment early in stance [9]. Since high amounts or velocities of rearfoot eversion are often cited as contributing factors to several common running injuries [13], this suggests barefoot running may be one method for reducing overuse running injury risk. However, whether the external eversion moment is actually reduced in barefoot compared to shod running is not clear. For instance, Kerrigan et al. [12] reported there were no differences in the net ankle eversion moments between shod and barefoot running. This discrepancy may be partially explained by examining the interplay between the location of the COP and the location of the joint center of rotation. For instance, while the COP may be located more laterally when running shod than when running barefoot, if the joint center of rotation is also located more laterally, this may partially or wholly offset any changes in external eversion moment due to a more lateral COP position [14]. However, to date, there are no descriptions in the literature regarding how COP trajectories differ between shod and barefoot running. Therefore, the purpose of this study was to investigate differences in COP trajectories and resulting external eversion moments when habitually shod runners with a rearfoot strike pattern run barefoot utilizing a midfoot strike pattern. It has been previously reported that having habitually shod individuals run barefoot results in a more plantar flexed ankle at initial contact but with individual specific changes in ankle eversion at contact [11]. Therefore, we hypothesized that compared to shod running, during
J. Becker et al. / Gait & Posture 40 (2014) 504–509
505
Fig. 1. Illustration showing marker placement on the foot with holes cut in the shoe (A), and an illustration showing the dimensions of the loop around which participants ran in the laboratory (B).
barefoot running the COP would be located more anteriorly early in stance phase but that there would not be consistent differences in the COP’s mediolateral positioning nor would there be differences in external eversion moments between conditions. 2. Methods Based on previously reported differences in rearfoot eversion between shod and barefoot running [11], a power analysis suggested eight participants would be required to adequately power this study (effect size 1.55, a = 0.05, b = 0.05). Therefore, ten individuals (sex: four female, six male; age: 41 8.7 years; weekly mileage: 39.5 10.3 miles) participated in this study. Participants were habitually shod runners with no injuries within the preceding six months. Participants were selected from a larger cohort of participants in a separate study on barefoot running and therefore, were only invited to participate if they used a rearfoot strike while running shod but switched to a midfoot strike while running barefoot. This particular criterion was used since many of the proposed benefits of barefoot running are tied to utilizing a midfoot strike and numerous studies examining biomechanical differences between shod and barefoot running have observed this specific change in foot strike pattern [9,15–18]. Since kinematic and kinetic asymmetry may be present even in uninjured runners [19], participants’ left and right feet were analyzed separately. Therefore, the total number of feet included for the statistical analyses was twenty. Thirty nine reflective markers were placed on specific bony landmarks [20]. Foot markers were placed bilaterally on the medial and lateral malleoli, heads of the second metatarsal, the tuberosities of the navicular and the fifth metatarsal, two along the vertical bisections of the posterior calcaneus, and one on the lateral aspect of the calcaneus. For the barefoot condition markers were placed directly on the skin while for the shod condition they were placed on the skin but visible through holes cut into the shoe (Fig. 1A). To ensure consistent placement between conditions marker locations on the skin were marked with a black pen. For both shod and barefoot conditions, prior to placing markers measurements of foot length and width were measured. Participants were allowed five minutes of jogging to warm up. Then they ran continuous laps around a loop (Fig. 1B) in the laboratory at self-selected speeds under shod then barefoot conditions. Marker trajectories were recorded with a 10-camera motion capture system (Motion Analysis Corp., Santa Rosa, CA) while ground reaction forces were measured with three force plates (AMTI, Inc. Watertown, MA) sampling at 200 Hz and
1000 Hz, respectively. Marker trajectories and ground reaction forces were smoothed using second order, zero lag, low pass Butterworth filters with cutoff frequencies of 8 Hz and 50 Hz, respectively [21]. Timing of initial contact and toe off were identified using a 50 N threshold in the vertical ground reaction force [22]. The strike index [22] was used to objectively verify changes in foot strike between shod and barefoot conditions. At each instant during stance phase, a local foot coordinate system (LCS’) defining the orientation of the foot relative to the global laboratory coordinate system (GCS) was established and the COP was calculated in the GCS then transformed to the LCS’ (Fig. 2). The origin of the LCS’ was located at the heel marker, with the x-axis aligning with the vector connecting the heel and toe markers, the y-axis pointing to the left, and the z-axis pointing superiorly. Once expressed in the LCS’, the anterior–posterior and mediolateral COP positions at 10% stance intervals were determined. Additionally, the following COP trajectory related variables were extracted for analysis: COP anterior–posterior and mediolateral excursions, and the most medial location of the COP during stance and the percent stance at which it occurred. Anterior– posterior and mediolateral COP excursions were normalized to foot length and width, measured in the shod and barefoot conditions, respectively. All calculations were performed using custom LabVIEW (National Instruments, Austin TX) software. The location of the ankle joint center, calculated as the midpoint between the medial and lateral malleoli, and the ground reaction forces were also transformed into the LCS’. The external eversion moment from the ground reaction force was then calculated. Both ground reaction forces and external moments were normalized by body mass. The anterior–posterior and mediolateral location of the ankle joint center in the LCS’ was calculated across stance at 10% intervals. Three variables were calculated to describe the external eversion moment: the maximum value during the first 15% of stance, the peak value, and the percent stance where the peak value occurred. Lastly, to aid in explanation of any observed changes in external moments, the ankle joint inversion-eversion angle and the tibia varus angle relative to the room at initial foot contact were calculated. Data collected from eight trials per foot were averaged for each condition. Differences between shod and barefoot conditions in the COP trajectory related variables, kinematic variables, peak ground reaction forces, and the three variables describing the external eversion moments, were compared using paired t -tests, with a value of p < .05 was used to indicate statistical significance. Paired t-tests were also used to compare the anterior–posterior and mediolateral COP and ankle joint center positions at each 10%
506
J. Becker et al. / Gait & Posture 40 (2014) 504–509
Fig. 2. COP trajectories are initially calculated in the global coordinate system (A). A local foot coordinate system was established at each instant during stance (B). The long axis of the foot coordinate system aligns with the longitudinal midline axis of the foot as defined by the heel and toe markers. By rotating the COP trajectory from the global coordinate system to the local coordinate system at each instant, COP trajectories from multiple trials can be compared relative to the anatomic structures of the foot (C) even if the foot contacts occurred on different force plates. The illustration above is for a left foot, showing markers on the calcaneus, 2nd metatarsal, 5th metatarsal, and navicular tuberosity.
of stance. However, to reduce the risk of a Type I error due to multiple comparisons across stance phase, a Bonferroni correction was applied for these comparisons. 3. Results Running speed was not significantly different between shod (3.30 0.41 m/s) and barefoot (3.27 0.42 m/s) conditions (p = 0.198). Mean strike index values confirmed visual observations, confirming that when converting from shod to barefoot conditions participants shifted their initial point of contact anteriorly (p < .001, Table 1). Compared to the shod condition, in the barefoot condition the COP was located significantly more anteriorly at initial contact, 10% and 20% of stance (Fig. 3A) and more medially at all points except initial contact and 20% of stance (Fig. 3B). Mean COP anterior– posterior excursions were smaller in the barefoot condition than the shod condition (p < .001), however there were no differences in COP mediolateral excursion between conditions (p = .753; Table 1).
For the barefoot condition the most medial location of the COP was further medial relative to the long axis of the foot than in the shod condition (p < .001) and occurred later during stance (p < .001; Table 1). There was no differences in tibia varus angle at initial contact (p = .948, Table 1), however, the foot was more everted at initial contact in the barefoot condition than in the shod condition (p = .014, Table 1). There were no differences in peak anterior–posterior, mediolateral, or vertical ground reaction forces between conditions (Fig. 4). However, all time points except initial contact, the ankle joint center was located more posteriorly and more medially in the barefoot condition compared to the shod condition (Fig. 3). The external eversion moment during the first 15% of stance was not different between conditions (p = .274, Table 1), however the peak external eversion moment was higher in the shod condition than the barefoot condition (p = .027, Fig. 4). Timing of the maximal external eversion moment was not different between conditions (p = .349, Table 1).
Table 1 Results for the strike index, COP anterior–posterior and mediolateral positions and excursions, and positioning of the COP at maximal vertical ground reaction force. A negative value for most medial location of COP indicates the COP is medial to the long axis of the foot. Variable
Shod
Strike index values COP anterior–posterior excursion (% foot length) COP mediolateral excursion (% foot width) Most medial COP location (% foot width) Percent stance most medial COP location (%) Tibia varus at foot contact ( ) Rear foot eversion at foot contact ( ) External eversion moment across first 15% stance (Nm/kg) Peak external eversion moment (Nm/kg) Percent stance of peak external eversion moment (Nm/kg)
19.6 ( 7.3) 65.8 ( 11.7) 31.1 ( 18.6) 2.2 ( 7.9) 81.9 ( 21.3) 9.2 ( 5.2) 0.1 ( 6.4) 0.27 ( 0.18) 0.42 ( 0.17) 38.0 ( 18.1)
*
Indicates shod condition is significantly different than barefoot condition with p < .05.
Barefoot 48.59 ( 12.04)* 34.79 ( 22.22)* 29.43 ( 17.46) 13.55 ( 7.63)* 95.5 ( 9.04)* 9.2 ( 4.7) 3.6 (5.2)* 0.22 ( 0.11) 0.33 ( 0.15) 35.3 ( 15.3)
J. Becker et al. / Gait & Posture 40 (2014) 504–509
507
Fig. 3. Average anterior–posterior (A) and mediolateral (B) positions of the COP across stance along with a graphical illustration showing average COP trajectories within a rough outline of the foot (C). The illustration is an example for a right foot showing the heel, 2nd metatarsal, 5th metatarsal, and navicular markers. * indicates SH is significantly different than BF at p < .0045.
4. Discussion The purpose of this study was to examine differences in COP trajectories and resulting external moments when habitually shod runners using a rearfoot strike pattern ran barefoot using a midfoot strike pattern. Compared to the shod condition, the COP was located more medially in the barefoot condition from 30% of stance through toe off. In addition to the position, the overall shape of the COP trajectory was different between conditions. The COP trajectory observed during the shod condition was characterized by a large initial medial displacement up through 20% of stance, after which it essentially tracked straight up the long axis of the foot with little additional mediolateral movement (Fig. 3). This trajectory is similar to those previously reported for shod runners utilizing a rearfoot strike [14,23]. The COP trajectory in the barefoot condition demonstrated a similar initial medial displacement during the first 20% of stance. However, it then continued to move medially throughout the remainder of stance (Fig. 3). The continued medial COP progression observed in the barefoot condition may reflect the fact that participants were running without any of the supporting structures associated with the shoe. Previous studies have shown that as support is removed from the medial side of the shoe, pressures increase under the medial aspect of the foot [24,25]. Higher medial pressures suggest the COP will be located more medially as well. While the specific type of running shoe worn by participants in the current study was not controlled, this pattern of higher medial pressures when medial support is removed has been observed when participants switched from motion control to neutral cushioning shoes [24], and again from neutral cushioning shoes to light weight racing flats [25] or barefoot conditions [24]. Given these findings it is possible this pattern of a more medially oriented COP trajectory would have been observed no matter what type of shoes participants wore; it would simply be a matter of degree of the difference between shod and barefoot conditions. For instance, the change may have been greater in participants who wore a
motion control shoe than those who routinely wore a lightweight trainer. This study also examined how the external eversion moments differed between shod and barefoot running. It has been suggested that the foot orientation at contact during barefoot running reduces the moment arms from the vertical and mediolateral ground reaction forces, thereby reducing the tendency for the rearfoot to evert [9]. Since high amounts or velocities of rearfoot eversion are often cited as contributing factors to numerous common running injuries [13], a running strategy which reduces the eversion moment arm might be one way to reduce overuse injury risk. However, the results of the current study do not support the hypothesis that barefoot running results in smaller external eversion moments. This may be explained by examining the interplay between the locations of the COP and center of rotation. As discussed by Dixon [14] the external eversion moment depends on the location the COP and the joint center of rotation, both of which are influenced by the foot orientation. In the current study participants demonstrated a more everted foot at initial contact in the barefoot condition than the shod condition, resulting in a more medially located COP. All other variables being equal, the medial shift of the COP would result in a shorter lever arm and thus a reduced external eversion moment. However, the different leg and foot orientation resulted in a medial shift of the center of rotation, assumed in this study to be the ankle joint center (Fig. 4). Since both the ground reaction forces and the external eversion moment over the initial 15% of stance were not different between conditions (Fig. 4), the medial shift of the center of rotation suggests there were similar eversion moment arms between the shod and barefoot conditions. Therefore, even though the COP was located more laterally during the shod condition, there were no differences in the external eversion moment during early stance. However, what, if anything, this means from an injury perspective still needs clarification, as even if there are no differences in external eversion moments, research has indicated
508
J. Becker et al. / Gait & Posture 40 (2014) 504–509
Fig. 4. Anterior–posterior (A), mediolateral (B), and vertical (D) ground reaction forces. Also shown are the trajectories of the ankle joint center (C), expressed in the local foot coordinate system, and the external inversion-eversion (E) moments across stance. C shows a right foot with the heel, 2nd metatarsal, 5th metatarsal, and navicular markers. * indicates SH is significantly different than BF at p < .05.
that barefoot running increase the injury risk from other variables such as external vertical loading rates [5]. An examination of the COP trajectories presented in Fig. 3 shows that the COP was initially located outside the illustrated base of support. A similar finding was reported by Cavanagh and Lafortune [22], who also used a force platform to measure COP trajectories relative to the shoe outline. They [22] suggested two possible explanations for this finding. First, there is likely some scuffing of the foot at initial contact where the forces are high enough to suggest that contact has been made but the placement of the foot has not been finalized. Second, the outline of the shoe was drawn at a stationary point, when in fact there is an abductory twist of the foot on the force plate during stance phase. These conditions likely apply to the current study as well. While the COP locations were referenced to the LCS’ at every moment, the illustrated outline of the shoe presented in Figs. 3 and 4 were taken from a single instant during midstance. Thus, any twisting movements of the foot during stance could result in the COP appearing outside the illustrated base of support. Additionally, the outline of the shoe shown in Figs. 3 and 4 is from a representative sample participants’ foot size while the COP trajectories are mean paths across all participants. Thus, on an individual basis, the COP could be closer or further to the actual outline of the base of support. One factor to consider when interpreting the results of the current study is that many studies examining COP locations or trajectories relative to foot anatomical axes do so using pressure sensing insoles [3] or pressure platforms [5,14,26] instead of force platforms. Therefore, it is important to note that both the COP anterior–posterior and mediolateral excursions and positions observed in the current study were similar to those previously reported when using pressure platforms, for both shod [23] and barefoot [5,26] running in healthy, uninjured runners. It should
also be pointed out that the COP measures from the shod condition represent conditions at the shoe-ground interface, not at the shoefoot interface. As long as an individual is running shod, this is true regardless of whether a force platform or a pressure platform is used to measure COP trajectories. Previous research examining how well measurements at the ground-shoe interface correspond to measurements at the shoe-foot interface suggest that there is good agreement in the anterior–posterior directions but only marginal agreement in the mediolateral directions [27,28], a difference attributed to the fact that the force plates compute the COP location with consideration of the mediolateral shearing forces while the pressure insoles used to compute the COP at the shoe-foot interface only considers the vertical forces [28]. Whether it is the conditions at the ground-shoe interface or the shoe-foot interface which is more important for performance and injury concerns is a matter for future studies. One limitation to this study was that all participants were habitually shod runners participating in an acute bout of barefoot running. It is unknown if the differences in COP trajectories would have remained had the participants been given more time to adapt to barefoot running. Similarly, it is unknown whether these differences would have been observed had habitually barefoot runners been used as participants. It should be noted that participants were included in this study only if they naturally transitioned from a rearfoot strike when shod to a midfoot strike when barefoot. Thus, the differences in COP trajectories reported in this study may reflect the type of foot contact used by the participants rather than being inherent differences between running shod or barefoot and while this combination of foot strike patterns is the most commonly reported [9,15–18], additional research is required to clarify how the COP trajectory might change if this particular foot strike combination is not used. Finally, it should be noted that the order of the shod and barefoot
J. Becker et al. / Gait & Posture 40 (2014) 504–509
trials was not randomized, and therefore any order effects, if they exist, cannot be ruled out. In conclusion, this study used a force plate to examine differences in COP trajectories and resulting external eversion moments when habitually shod runners utilizing a rearfoot strike pattern run barefoot utilizing a midfoot strike pattern. The results suggest, that when these two foot strike patterns are used, there are differences in the COP trajectory between shod and barefoot running. These differences may have implications for both performance and injury risk factors. Therefore, future studies on mechanical and epidemiological differences between shod and barefoot running may find the COP trajectory to be a useful to examine. Conflict of interest The authors declare no conflict of interest. References [1] Williams DS, McClay IS, Hamill J. Arch structure and injury patterns in runners. Clin Biomech 2001;16:341–7. [2] Fuller E. Center of pressure and its theoretical relationship to foot pathology. J Am Podiatr Med Assoc 1999;89:278–91. [3] Nigg BM, Stergiou P, Cole GK, Stefanyshyn D, Mundermann A, Humble N. Effect of shoe inserts on kinematics, center of pressure, and leg joint moments during running. Med Sci Sports Exerc 2003;35:314–9. [4] Van Ginckel A, Thijs Y, Hesar NGZ, Mahieu N, De Clercq D, Roosen P, Witvrouw E. Intrinsic gait-related risk factors for Achilles tendinopathy in novice runners: a prospective study. Gait Posture 2009;29:387–91. [5] Willems TM, De Clercq D, Delbaere K, Vanderstraeten G, De Cock A, Witvrouw E. A prospective study of gait related risk factors for exercise-related lower leg pain. Gait Posture 2006;23:91–8. [6] Van Gent RN, Siem D, van Middelkoop M, van Os AG, Bierma-Zeinstra SMA, Koes BW. Incidence and determinants of lower extremity running injuries in long distance runners: a systematic review. Br J Sports Med 2007;41:469–80 [discussion 480]. [7] Taunton JE, Ryan MB, Clement DB, McKenzie DC, Lloyd-Smith DR, Zumbo BD. A retrospective case-control analysis of running injuries. Br J Sports Med 2002;36:95–101. [8] Robbins SE, Hanna AM. Running injury reduction through barefoot adaptation. Med Sci Sports Exerc 1987;19:148–56. [9] Altman A, Davis IS. Barefoot running biomechanics and implications for running injuries. Curr Sports Med Rep 2012;11:244–50. [10] Lieberman DE, Venkadesan M, Werbel WA, Daoud AI, D'Andrea S, Davis IS. Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature 2010;463:531–5 [Nature Publishing Group].
509
[11] De Wit B, De Clercq D, Aerts P. Biomechanical analysis of the stance phase during barefoot and shod running. J Biomech 2000;33:269–78. [12] Kerrigan DC, Franz JR, Keenan GS, Dicharry J, Della Croce U, Wilder RP. The effect of running shoes on lower extremity joint torques. PM&R 2009;1:1058– 63 [Elsevier Inc.]. [13] Hreljac A. Etiology, prevention, and early intervention of overuse injuries in runners: a Biomechanical perspective. Phys Med Rehabil Clin N Am 2005;16:651–67. [14] Dixon S. Application of center-of-pressure data to indicate rearfoot inversioneversion in shod running. J Am Podiatr Med Assoc 2006;96:305–12. [15] Divert C, Mornieux G, Baur H, Mayer F, Belli A. Mechanical comparison of barefoot and shod running. Int J Sports Med 2005;26:593–8. [16] Squadrone R, Gallozzi C. Biomechanical and physiological comparison of barefoot and two shod conditions in experienced barefoot runners. J Sports Med Phys Fitness 2009;49:6–13. [17] Bonacci J, Saunders PU, Hicks A, Rantalainen T, Vicenzino BGT, Spratford W. Running in a minimalist and lightweight shoe is not the same as running barefoot: a biomechanical study. Br J Sports Med 2013;47:387–92. [18] Hamill J, Russell EM, Gruber AH, Miller R. Impact characteristics in shod and barefoot running. Footwear Sci 2011;3:33–40. [19] Zifchock RA, Davis I, Higginson J, McCaw S, Royer T. Side-to-side differences in overuse running injury susceptibility: a retrospective study. Human Mov Sci 2008;27:888–902 [Elsevier B.V.]. [20] Hahn M, Chou L-S. Age related reduction in sagittal plan center of mass motion during obstacle crossing. J Biomech 2004;37:837–44. [21] Edwards WB, Gillette JC, Thomas JM, Derrick TR. Internal femoral forces and moments during running: Implications for stress fracture development. Clin Biomech 2008;23:1269–78. [22] Cavanagh PR, Lafortune MA. Ground reaction forces in distance running. J Biomech 1980;13:397–406. [23] Willems TM, Witvrouw E, De Cock A, De Clercq D. Gait related risk factors for exercise related lower leg pain during shod running. Med Sci Sports Exerc 2007;39:330–9. [24] Molloy JM, Christie DS, Teyhen DS, Yeykal NS, Tragord BS, Neal MS, et al. Effect of running shoe type on the distribution and magnitdue of plantar pressures in individuals with low- or high-arched feet. J Am Podiatr Med Assoc 2009;99:330–8. [25] Wiegerinck JI, Boyd J, Yoder JC, Abbey AN, Nunley JA, Queen RM. Differences in plantar loading between training shoes and racing flats at a self-selected running speed. Gait Posture 2009;29:514–9. [26] De Cock A, Vanrenterghem J, Willems T, Witvrouw E, De Clercq D. The trajectory of the centre of pressure during barefoot running as a potential measure for foot function. Gait Posture 2008;27:669–75. [27] Chesnin KJ, Selby-Silverstein L, Besser MP. Comparison of an in-shoe pressure measurement device to a force plate: concurrent validity of center of pressure measurements. Gait Posture 2000;12:128–33. [28] Debbi EM, Wolf A, Goryachev Y, Yizhar Z, Luger E, Debi R. In-shoe center of pressure: indirect force plate vs. direct insole measurement. Foot 2012;22:269–75 [Elsevier Ltd].
Contents lists available at ScienceDirect
Gait & Posture journal homepage: www.elsevier.com/locate/gaitpost
Center of pressure trajectory differences between shod and barefoot running James Becker a , Eric Pisciotta a , Stan James b , Louis R. Osternig a , Li-Shan Chou a, * a b
Motion Analysis Laboratory, Department of Human Physiology, University of Oregon, Eugene, OR 97403-1240, USA Slocum Center for Orthopedics and Sports Medicine, Eugene, OR, USA
A R T I C L E I N F O
A B S T R A C T
Article history: Received 21 November 2013 Received in revised form 11 June 2014 Accepted 16 June 2014
This study examined differences in center of pressure (COP) trajectories between shod and barefoot running. Ten habitually shod runners ran continuous laps under both shod and barefoot conditions. The COP trajectory was calculated in the global coordinate system but then transformed to the anatomic coordinate system of the foot. The anterior–posterior and medio-lateral positions and excursions of the COP, as well as the most medial location and percent stand at which it occurred were examined. Additionally, external eversion moments and ground reaction forces were assessed. Compared to the shod condition, in the barefoot condition the COP was located more anteriorly early in stance and the COP was located significantly more medially at most time points across stance. There were no differences in external eversion moments during early stance or peak ground reaction forces between conditions. Future studies on mechanical or epidemiological differences between shod and barefoot running may find the COP trajectory an informative parameter to examine. ã 2014 Elsevier B.V. All rights reserved.
Keywords: Center of pressure Running Barefoot running Eversion
1. Introduction The location of the center of pressure (COP) is a commonly used measure for assessing dynamic foot function during running. COP trajectories have been used to examine differences in foot function between high and low arches individuals [1], as a measure for quantifying mediolateral foot stability [2], and as a tool for prescribing and assessing the effects of shoe or orthotic interventions [3]. Analysis of COP trajectories is also informative regarding risks for sustaining common overuse running injuries such as Achilles tendinopathy [4] or exercise related lower leg pain [5]. It is estimated that between 25 and 75% of runners will sustain an overuse injury in any one year period of time [6,7]. Barefoot running has been suggested as a one potential method for reducing the incidence of running related injuries [8–10]. However, to date most studies comparing barefoot and shod running have focused on kinematics [9,11], external ground reaction forces [10], or joint moments [12]. In regards to COP trajectories in particular, it has been suggested that, compared to shod running, the foot position at initial contact during barefoot running results in smaller lever arms for the vertical and mediolateral ground reaction force, thus
* Corresponding author. Tel.: +1 541 346 3391; fax: +1 541 346 2841. E-mail address: [email protected] (L.-S. Chou). http://dx.doi.org/10.1016/j.gaitpost.2014.06.007 0966-6362/ ã 2014 Elsevier B.V. All rights reserved.
reducing the external eversion moment early in stance [9]. Since high amounts or velocities of rearfoot eversion are often cited as contributing factors to several common running injuries [13], this suggests barefoot running may be one method for reducing overuse running injury risk. However, whether the external eversion moment is actually reduced in barefoot compared to shod running is not clear. For instance, Kerrigan et al. [12] reported there were no differences in the net ankle eversion moments between shod and barefoot running. This discrepancy may be partially explained by examining the interplay between the location of the COP and the location of the joint center of rotation. For instance, while the COP may be located more laterally when running shod than when running barefoot, if the joint center of rotation is also located more laterally, this may partially or wholly offset any changes in external eversion moment due to a more lateral COP position [14]. However, to date, there are no descriptions in the literature regarding how COP trajectories differ between shod and barefoot running. Therefore, the purpose of this study was to investigate differences in COP trajectories and resulting external eversion moments when habitually shod runners with a rearfoot strike pattern run barefoot utilizing a midfoot strike pattern. It has been previously reported that having habitually shod individuals run barefoot results in a more plantar flexed ankle at initial contact but with individual specific changes in ankle eversion at contact [11]. Therefore, we hypothesized that compared to shod running, during
J. Becker et al. / Gait & Posture 40 (2014) 504–509
505
Fig. 1. Illustration showing marker placement on the foot with holes cut in the shoe (A), and an illustration showing the dimensions of the loop around which participants ran in the laboratory (B).
barefoot running the COP would be located more anteriorly early in stance phase but that there would not be consistent differences in the COP’s mediolateral positioning nor would there be differences in external eversion moments between conditions. 2. Methods Based on previously reported differences in rearfoot eversion between shod and barefoot running [11], a power analysis suggested eight participants would be required to adequately power this study (effect size 1.55, a = 0.05, b = 0.05). Therefore, ten individuals (sex: four female, six male; age: 41 8.7 years; weekly mileage: 39.5 10.3 miles) participated in this study. Participants were habitually shod runners with no injuries within the preceding six months. Participants were selected from a larger cohort of participants in a separate study on barefoot running and therefore, were only invited to participate if they used a rearfoot strike while running shod but switched to a midfoot strike while running barefoot. This particular criterion was used since many of the proposed benefits of barefoot running are tied to utilizing a midfoot strike and numerous studies examining biomechanical differences between shod and barefoot running have observed this specific change in foot strike pattern [9,15–18]. Since kinematic and kinetic asymmetry may be present even in uninjured runners [19], participants’ left and right feet were analyzed separately. Therefore, the total number of feet included for the statistical analyses was twenty. Thirty nine reflective markers were placed on specific bony landmarks [20]. Foot markers were placed bilaterally on the medial and lateral malleoli, heads of the second metatarsal, the tuberosities of the navicular and the fifth metatarsal, two along the vertical bisections of the posterior calcaneus, and one on the lateral aspect of the calcaneus. For the barefoot condition markers were placed directly on the skin while for the shod condition they were placed on the skin but visible through holes cut into the shoe (Fig. 1A). To ensure consistent placement between conditions marker locations on the skin were marked with a black pen. For both shod and barefoot conditions, prior to placing markers measurements of foot length and width were measured. Participants were allowed five minutes of jogging to warm up. Then they ran continuous laps around a loop (Fig. 1B) in the laboratory at self-selected speeds under shod then barefoot conditions. Marker trajectories were recorded with a 10-camera motion capture system (Motion Analysis Corp., Santa Rosa, CA) while ground reaction forces were measured with three force plates (AMTI, Inc. Watertown, MA) sampling at 200 Hz and
1000 Hz, respectively. Marker trajectories and ground reaction forces were smoothed using second order, zero lag, low pass Butterworth filters with cutoff frequencies of 8 Hz and 50 Hz, respectively [21]. Timing of initial contact and toe off were identified using a 50 N threshold in the vertical ground reaction force [22]. The strike index [22] was used to objectively verify changes in foot strike between shod and barefoot conditions. At each instant during stance phase, a local foot coordinate system (LCS’) defining the orientation of the foot relative to the global laboratory coordinate system (GCS) was established and the COP was calculated in the GCS then transformed to the LCS’ (Fig. 2). The origin of the LCS’ was located at the heel marker, with the x-axis aligning with the vector connecting the heel and toe markers, the y-axis pointing to the left, and the z-axis pointing superiorly. Once expressed in the LCS’, the anterior–posterior and mediolateral COP positions at 10% stance intervals were determined. Additionally, the following COP trajectory related variables were extracted for analysis: COP anterior–posterior and mediolateral excursions, and the most medial location of the COP during stance and the percent stance at which it occurred. Anterior– posterior and mediolateral COP excursions were normalized to foot length and width, measured in the shod and barefoot conditions, respectively. All calculations were performed using custom LabVIEW (National Instruments, Austin TX) software. The location of the ankle joint center, calculated as the midpoint between the medial and lateral malleoli, and the ground reaction forces were also transformed into the LCS’. The external eversion moment from the ground reaction force was then calculated. Both ground reaction forces and external moments were normalized by body mass. The anterior–posterior and mediolateral location of the ankle joint center in the LCS’ was calculated across stance at 10% intervals. Three variables were calculated to describe the external eversion moment: the maximum value during the first 15% of stance, the peak value, and the percent stance where the peak value occurred. Lastly, to aid in explanation of any observed changes in external moments, the ankle joint inversion-eversion angle and the tibia varus angle relative to the room at initial foot contact were calculated. Data collected from eight trials per foot were averaged for each condition. Differences between shod and barefoot conditions in the COP trajectory related variables, kinematic variables, peak ground reaction forces, and the three variables describing the external eversion moments, were compared using paired t -tests, with a value of p < .05 was used to indicate statistical significance. Paired t-tests were also used to compare the anterior–posterior and mediolateral COP and ankle joint center positions at each 10%
506
J. Becker et al. / Gait & Posture 40 (2014) 504–509
Fig. 2. COP trajectories are initially calculated in the global coordinate system (A). A local foot coordinate system was established at each instant during stance (B). The long axis of the foot coordinate system aligns with the longitudinal midline axis of the foot as defined by the heel and toe markers. By rotating the COP trajectory from the global coordinate system to the local coordinate system at each instant, COP trajectories from multiple trials can be compared relative to the anatomic structures of the foot (C) even if the foot contacts occurred on different force plates. The illustration above is for a left foot, showing markers on the calcaneus, 2nd metatarsal, 5th metatarsal, and navicular tuberosity.
of stance. However, to reduce the risk of a Type I error due to multiple comparisons across stance phase, a Bonferroni correction was applied for these comparisons. 3. Results Running speed was not significantly different between shod (3.30 0.41 m/s) and barefoot (3.27 0.42 m/s) conditions (p = 0.198). Mean strike index values confirmed visual observations, confirming that when converting from shod to barefoot conditions participants shifted their initial point of contact anteriorly (p < .001, Table 1). Compared to the shod condition, in the barefoot condition the COP was located significantly more anteriorly at initial contact, 10% and 20% of stance (Fig. 3A) and more medially at all points except initial contact and 20% of stance (Fig. 3B). Mean COP anterior– posterior excursions were smaller in the barefoot condition than the shod condition (p < .001), however there were no differences in COP mediolateral excursion between conditions (p = .753; Table 1).
For the barefoot condition the most medial location of the COP was further medial relative to the long axis of the foot than in the shod condition (p < .001) and occurred later during stance (p < .001; Table 1). There was no differences in tibia varus angle at initial contact (p = .948, Table 1), however, the foot was more everted at initial contact in the barefoot condition than in the shod condition (p = .014, Table 1). There were no differences in peak anterior–posterior, mediolateral, or vertical ground reaction forces between conditions (Fig. 4). However, all time points except initial contact, the ankle joint center was located more posteriorly and more medially in the barefoot condition compared to the shod condition (Fig. 3). The external eversion moment during the first 15% of stance was not different between conditions (p = .274, Table 1), however the peak external eversion moment was higher in the shod condition than the barefoot condition (p = .027, Fig. 4). Timing of the maximal external eversion moment was not different between conditions (p = .349, Table 1).
Table 1 Results for the strike index, COP anterior–posterior and mediolateral positions and excursions, and positioning of the COP at maximal vertical ground reaction force. A negative value for most medial location of COP indicates the COP is medial to the long axis of the foot. Variable
Shod
Strike index values COP anterior–posterior excursion (% foot length) COP mediolateral excursion (% foot width) Most medial COP location (% foot width) Percent stance most medial COP location (%) Tibia varus at foot contact ( ) Rear foot eversion at foot contact ( ) External eversion moment across first 15% stance (Nm/kg) Peak external eversion moment (Nm/kg) Percent stance of peak external eversion moment (Nm/kg)
19.6 ( 7.3) 65.8 ( 11.7) 31.1 ( 18.6) 2.2 ( 7.9) 81.9 ( 21.3) 9.2 ( 5.2) 0.1 ( 6.4) 0.27 ( 0.18) 0.42 ( 0.17) 38.0 ( 18.1)
*
Indicates shod condition is significantly different than barefoot condition with p < .05.
Barefoot 48.59 ( 12.04)* 34.79 ( 22.22)* 29.43 ( 17.46) 13.55 ( 7.63)* 95.5 ( 9.04)* 9.2 ( 4.7) 3.6 (5.2)* 0.22 ( 0.11) 0.33 ( 0.15) 35.3 ( 15.3)
J. Becker et al. / Gait & Posture 40 (2014) 504–509
507
Fig. 3. Average anterior–posterior (A) and mediolateral (B) positions of the COP across stance along with a graphical illustration showing average COP trajectories within a rough outline of the foot (C). The illustration is an example for a right foot showing the heel, 2nd metatarsal, 5th metatarsal, and navicular markers. * indicates SH is significantly different than BF at p < .0045.
4. Discussion The purpose of this study was to examine differences in COP trajectories and resulting external moments when habitually shod runners using a rearfoot strike pattern ran barefoot using a midfoot strike pattern. Compared to the shod condition, the COP was located more medially in the barefoot condition from 30% of stance through toe off. In addition to the position, the overall shape of the COP trajectory was different between conditions. The COP trajectory observed during the shod condition was characterized by a large initial medial displacement up through 20% of stance, after which it essentially tracked straight up the long axis of the foot with little additional mediolateral movement (Fig. 3). This trajectory is similar to those previously reported for shod runners utilizing a rearfoot strike [14,23]. The COP trajectory in the barefoot condition demonstrated a similar initial medial displacement during the first 20% of stance. However, it then continued to move medially throughout the remainder of stance (Fig. 3). The continued medial COP progression observed in the barefoot condition may reflect the fact that participants were running without any of the supporting structures associated with the shoe. Previous studies have shown that as support is removed from the medial side of the shoe, pressures increase under the medial aspect of the foot [24,25]. Higher medial pressures suggest the COP will be located more medially as well. While the specific type of running shoe worn by participants in the current study was not controlled, this pattern of higher medial pressures when medial support is removed has been observed when participants switched from motion control to neutral cushioning shoes [24], and again from neutral cushioning shoes to light weight racing flats [25] or barefoot conditions [24]. Given these findings it is possible this pattern of a more medially oriented COP trajectory would have been observed no matter what type of shoes participants wore; it would simply be a matter of degree of the difference between shod and barefoot conditions. For instance, the change may have been greater in participants who wore a
motion control shoe than those who routinely wore a lightweight trainer. This study also examined how the external eversion moments differed between shod and barefoot running. It has been suggested that the foot orientation at contact during barefoot running reduces the moment arms from the vertical and mediolateral ground reaction forces, thereby reducing the tendency for the rearfoot to evert [9]. Since high amounts or velocities of rearfoot eversion are often cited as contributing factors to numerous common running injuries [13], a running strategy which reduces the eversion moment arm might be one way to reduce overuse injury risk. However, the results of the current study do not support the hypothesis that barefoot running results in smaller external eversion moments. This may be explained by examining the interplay between the locations of the COP and center of rotation. As discussed by Dixon [14] the external eversion moment depends on the location the COP and the joint center of rotation, both of which are influenced by the foot orientation. In the current study participants demonstrated a more everted foot at initial contact in the barefoot condition than the shod condition, resulting in a more medially located COP. All other variables being equal, the medial shift of the COP would result in a shorter lever arm and thus a reduced external eversion moment. However, the different leg and foot orientation resulted in a medial shift of the center of rotation, assumed in this study to be the ankle joint center (Fig. 4). Since both the ground reaction forces and the external eversion moment over the initial 15% of stance were not different between conditions (Fig. 4), the medial shift of the center of rotation suggests there were similar eversion moment arms between the shod and barefoot conditions. Therefore, even though the COP was located more laterally during the shod condition, there were no differences in the external eversion moment during early stance. However, what, if anything, this means from an injury perspective still needs clarification, as even if there are no differences in external eversion moments, research has indicated
508
J. Becker et al. / Gait & Posture 40 (2014) 504–509
Fig. 4. Anterior–posterior (A), mediolateral (B), and vertical (D) ground reaction forces. Also shown are the trajectories of the ankle joint center (C), expressed in the local foot coordinate system, and the external inversion-eversion (E) moments across stance. C shows a right foot with the heel, 2nd metatarsal, 5th metatarsal, and navicular markers. * indicates SH is significantly different than BF at p < .05.
that barefoot running increase the injury risk from other variables such as external vertical loading rates [5]. An examination of the COP trajectories presented in Fig. 3 shows that the COP was initially located outside the illustrated base of support. A similar finding was reported by Cavanagh and Lafortune [22], who also used a force platform to measure COP trajectories relative to the shoe outline. They [22] suggested two possible explanations for this finding. First, there is likely some scuffing of the foot at initial contact where the forces are high enough to suggest that contact has been made but the placement of the foot has not been finalized. Second, the outline of the shoe was drawn at a stationary point, when in fact there is an abductory twist of the foot on the force plate during stance phase. These conditions likely apply to the current study as well. While the COP locations were referenced to the LCS’ at every moment, the illustrated outline of the shoe presented in Figs. 3 and 4 were taken from a single instant during midstance. Thus, any twisting movements of the foot during stance could result in the COP appearing outside the illustrated base of support. Additionally, the outline of the shoe shown in Figs. 3 and 4 is from a representative sample participants’ foot size while the COP trajectories are mean paths across all participants. Thus, on an individual basis, the COP could be closer or further to the actual outline of the base of support. One factor to consider when interpreting the results of the current study is that many studies examining COP locations or trajectories relative to foot anatomical axes do so using pressure sensing insoles [3] or pressure platforms [5,14,26] instead of force platforms. Therefore, it is important to note that both the COP anterior–posterior and mediolateral excursions and positions observed in the current study were similar to those previously reported when using pressure platforms, for both shod [23] and barefoot [5,26] running in healthy, uninjured runners. It should
also be pointed out that the COP measures from the shod condition represent conditions at the shoe-ground interface, not at the shoefoot interface. As long as an individual is running shod, this is true regardless of whether a force platform or a pressure platform is used to measure COP trajectories. Previous research examining how well measurements at the ground-shoe interface correspond to measurements at the shoe-foot interface suggest that there is good agreement in the anterior–posterior directions but only marginal agreement in the mediolateral directions [27,28], a difference attributed to the fact that the force plates compute the COP location with consideration of the mediolateral shearing forces while the pressure insoles used to compute the COP at the shoe-foot interface only considers the vertical forces [28]. Whether it is the conditions at the ground-shoe interface or the shoe-foot interface which is more important for performance and injury concerns is a matter for future studies. One limitation to this study was that all participants were habitually shod runners participating in an acute bout of barefoot running. It is unknown if the differences in COP trajectories would have remained had the participants been given more time to adapt to barefoot running. Similarly, it is unknown whether these differences would have been observed had habitually barefoot runners been used as participants. It should be noted that participants were included in this study only if they naturally transitioned from a rearfoot strike when shod to a midfoot strike when barefoot. Thus, the differences in COP trajectories reported in this study may reflect the type of foot contact used by the participants rather than being inherent differences between running shod or barefoot and while this combination of foot strike patterns is the most commonly reported [9,15–18], additional research is required to clarify how the COP trajectory might change if this particular foot strike combination is not used. Finally, it should be noted that the order of the shod and barefoot
J. Becker et al. / Gait & Posture 40 (2014) 504–509
trials was not randomized, and therefore any order effects, if they exist, cannot be ruled out. In conclusion, this study used a force plate to examine differences in COP trajectories and resulting external eversion moments when habitually shod runners utilizing a rearfoot strike pattern run barefoot utilizing a midfoot strike pattern. The results suggest, that when these two foot strike patterns are used, there are differences in the COP trajectory between shod and barefoot running. These differences may have implications for both performance and injury risk factors. Therefore, future studies on mechanical and epidemiological differences between shod and barefoot running may find the COP trajectory to be a useful to examine. Conflict of interest The authors declare no conflict of interest. References [1] Williams DS, McClay IS, Hamill J. Arch structure and injury patterns in runners. Clin Biomech 2001;16:341–7. [2] Fuller E. Center of pressure and its theoretical relationship to foot pathology. J Am Podiatr Med Assoc 1999;89:278–91. [3] Nigg BM, Stergiou P, Cole GK, Stefanyshyn D, Mundermann A, Humble N. Effect of shoe inserts on kinematics, center of pressure, and leg joint moments during running. Med Sci Sports Exerc 2003;35:314–9. [4] Van Ginckel A, Thijs Y, Hesar NGZ, Mahieu N, De Clercq D, Roosen P, Witvrouw E. Intrinsic gait-related risk factors for Achilles tendinopathy in novice runners: a prospective study. Gait Posture 2009;29:387–91. [5] Willems TM, De Clercq D, Delbaere K, Vanderstraeten G, De Cock A, Witvrouw E. A prospective study of gait related risk factors for exercise-related lower leg pain. Gait Posture 2006;23:91–8. [6] Van Gent RN, Siem D, van Middelkoop M, van Os AG, Bierma-Zeinstra SMA, Koes BW. Incidence and determinants of lower extremity running injuries in long distance runners: a systematic review. Br J Sports Med 2007;41:469–80 [discussion 480]. [7] Taunton JE, Ryan MB, Clement DB, McKenzie DC, Lloyd-Smith DR, Zumbo BD. A retrospective case-control analysis of running injuries. Br J Sports Med 2002;36:95–101. [8] Robbins SE, Hanna AM. Running injury reduction through barefoot adaptation. Med Sci Sports Exerc 1987;19:148–56. [9] Altman A, Davis IS. Barefoot running biomechanics and implications for running injuries. Curr Sports Med Rep 2012;11:244–50. [10] Lieberman DE, Venkadesan M, Werbel WA, Daoud AI, D'Andrea S, Davis IS. Foot strike patterns and collision forces in habitually barefoot versus shod runners. Nature 2010;463:531–5 [Nature Publishing Group].
509
[11] De Wit B, De Clercq D, Aerts P. Biomechanical analysis of the stance phase during barefoot and shod running. J Biomech 2000;33:269–78. [12] Kerrigan DC, Franz JR, Keenan GS, Dicharry J, Della Croce U, Wilder RP. The effect of running shoes on lower extremity joint torques. PM&R 2009;1:1058– 63 [Elsevier Inc.]. [13] Hreljac A. Etiology, prevention, and early intervention of overuse injuries in runners: a Biomechanical perspective. Phys Med Rehabil Clin N Am 2005;16:651–67. [14] Dixon S. Application of center-of-pressure data to indicate rearfoot inversioneversion in shod running. J Am Podiatr Med Assoc 2006;96:305–12. [15] Divert C, Mornieux G, Baur H, Mayer F, Belli A. Mechanical comparison of barefoot and shod running. Int J Sports Med 2005;26:593–8. [16] Squadrone R, Gallozzi C. Biomechanical and physiological comparison of barefoot and two shod conditions in experienced barefoot runners. J Sports Med Phys Fitness 2009;49:6–13. [17] Bonacci J, Saunders PU, Hicks A, Rantalainen T, Vicenzino BGT, Spratford W. Running in a minimalist and lightweight shoe is not the same as running barefoot: a biomechanical study. Br J Sports Med 2013;47:387–92. [18] Hamill J, Russell EM, Gruber AH, Miller R. Impact characteristics in shod and barefoot running. Footwear Sci 2011;3:33–40. [19] Zifchock RA, Davis I, Higginson J, McCaw S, Royer T. Side-to-side differences in overuse running injury susceptibility: a retrospective study. Human Mov Sci 2008;27:888–902 [Elsevier B.V.]. [20] Hahn M, Chou L-S. Age related reduction in sagittal plan center of mass motion during obstacle crossing. J Biomech 2004;37:837–44. [21] Edwards WB, Gillette JC, Thomas JM, Derrick TR. Internal femoral forces and moments during running: Implications for stress fracture development. Clin Biomech 2008;23:1269–78. [22] Cavanagh PR, Lafortune MA. Ground reaction forces in distance running. J Biomech 1980;13:397–406. [23] Willems TM, Witvrouw E, De Cock A, De Clercq D. Gait related risk factors for exercise related lower leg pain during shod running. Med Sci Sports Exerc 2007;39:330–9. [24] Molloy JM, Christie DS, Teyhen DS, Yeykal NS, Tragord BS, Neal MS, et al. Effect of running shoe type on the distribution and magnitdue of plantar pressures in individuals with low- or high-arched feet. J Am Podiatr Med Assoc 2009;99:330–8. [25] Wiegerinck JI, Boyd J, Yoder JC, Abbey AN, Nunley JA, Queen RM. Differences in plantar loading between training shoes and racing flats at a self-selected running speed. Gait Posture 2009;29:514–9. [26] De Cock A, Vanrenterghem J, Willems T, Witvrouw E, De Clercq D. The trajectory of the centre of pressure during barefoot running as a potential measure for foot function. Gait Posture 2008;27:669–75. [27] Chesnin KJ, Selby-Silverstein L, Besser MP. Comparison of an in-shoe pressure measurement device to a force plate: concurrent validity of center of pressure measurements. Gait Posture 2000;12:128–33. [28] Debbi EM, Wolf A, Goryachev Y, Yizhar Z, Luger E, Debi R. In-shoe center of pressure: indirect force plate vs. direct insole measurement. Foot 2012;22:269–75 [Elsevier Ltd].
