Clinical Biomechanics

Clinical Biomechanics 25 (2010) 1047 1052 Contents lists available at ScienceDirect Clinical Biomechanics journal homepage: www.elsevier.com/locate/clinbiomech Gender differences in lower extremity gait biomechanics during walking using an unstable shoe Benno M. Nigg a,, Karelia E. Tecante G. a, Peter Federolf a, Scott C. Landry a,b a Human Performance Laboratory, Faculty of Kinesiology, University of Calgary, 2500 University Drive N.W., Calgary, Alberta, Canada T2N 1N4 b School of Recreation Management and Kinesiology, Acadia University, 550 Main Street, Wolfville, Nova Scotia, Canada B4P 1C4 article info abstract Article history: Received 27 April 2010 Accepted 21 July 2010 Keywords: MBT Gender differences Gait Kinematics Kinetics Background: In recent years several unstable shoe designs that cause increased instability at the ankle joint have been developed with the aims of training static and dynamic posture and postural control. However, earlier research found significant gender differences in the generation of ankle torque and in the reaction times after a perturbation. Therefore it is possible that men and women are affected differently by the instability that unstable shoes create. The purpose of this study was to investigate if gender differences exist a) during bilateral quiet stance or b) in lower extremity gait kinematics and kinetics when using unstable shoes. Methods: Seventeen females and seventeen males were included in this study. Masai Barefoot Technology shoes were used as test shoes. Center of pressure excursion was recorded during 30 s bilateral quiet stance trials using a force plate. Joint angles, resultant joint moments and joint moment impulses during walking were determined using standard gait analysis methods. Findings: In bipedal stance, female subjects had significantly greater anterior posterior center of pressure excursion than male subjects. In the stance phase of the gait cycle gender differences were found in the ankle joint moments which had not been reported in earlier studies using barefoot or normal shoe conditions. Interpretation: The results suggest that women and men use different strategies to control the ankle joint when standing or walking in unstable shoes. Gender effects should therefore be taken into consideration if functional or therapeutic effects of unstable shoes are assessed. 2010 Elsevier Ltd. All rights reserved. 1. Introduction In recent years, several unstable shoe designs have been developed as functional or therapeutic tools. One example is the shoe design developed by Masai Barefoot Technology (MBT), which was the first unstable shoe produced and commercialized in large quantities. The conceptual aim of most unstable shoe designs is to improve balance and posture by exposing the foot to intrinsic instability of the shoe sole. This constant instability presumably trains the small muscletendon units crossing the ankle joint and consequently improves the wearer's posture and postural control. Several scientific results published in recent years support the general concept that unstable shoes have beneficial effects on postural control: First, center of pressure (COP) excursions are greater while standing in unstable shoes compared to those while standing in a control shoe (Nigg, Hintzen, & Ferber, 2006, Nigg, Emery, & Hiemstra, 2006). Hence, because of the instability created by the shoe, the body sways more while adjusting and maintaining postural control (Nigg, Corresponding author. E-mail address: nigg@ucalgary.ca (B.M. Nigg). Hintzen, & Ferber, 2006, Nigg, Emery, & Hiemstra, 2006, Romkes, 2008). Second, gait patterns and plantar pressure distributions change when walking with unstable shoes compared to regular shoes (Romkes et al., 2006, Stewart et al., 2007). The main differences in gait patterns are an increased dorsiflexion angle at initial contact followed by a continuous plantarflexion movement until push-off which alters the activity of tibialis anterior and gastrocnemius muscles during gait (Romkes et al., 2006). Third, lower subjective pain levels at the knee and lower back were reported when subjects regularly wore unstable shoes (Nigg, Emery, & Hiemstra, 2006; Nigg et al. 2009). These suggest that unstable shoes improve posture or muscle activity facilitating postural control. Finally, model calculations show that strong small muscles that react more quickly to altered posture than large muscles reduce joint loading substantially (Nigg, 2005). This theoretical result has been supported by experimental data (Boyer et al., 2008, Nigg, Emery, & Hiemstra, 2006). The instability created by unstable shoe soles affects directly the ankle joint movements and the stability of the ankle joint. Baseline ankle torque, rate of torque generation and peak ankle torque determine the ability to recover from an unexpected disturbance of balance (Robinovitch et al., 2002). However, it was found that women generally have smaller maximum voluntary isometric dorsiflexion 0268-0033/$ see front matter 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.clinbiomech.2010.07.010

1048 B.M. Nigg et al. / Clinical Biomechanics 25 (2010) 1047 1052 strength than men (Thelen et al., 1996). In addition, while the time required to reach a target plantarflexion or dorsiflexion torque does not differ between genders, women have relatively shorter reaction times and need more additional time to reach a target torque than men (Thelen et al., 1996). It is therefore possible that men and women use different strategies to control the instability created by unstable shoes. The purpose of this study was (a) to quantify gender differences in the COP excursion during bipedal stance while wearing unstable shoes and (b) to quantify gender differences in kinematic and kinetic variables describing the lower extremity movement patterns during gait while wearing unstable shoes. The MBT shoe was chosen for this study because it is the unstable shoe design that has been used in most research studies published on unstable shoe designs to date. 2. Methods 2.1. Subjects Thirty-four healthy subjects, 17 females and 17 males, volunteered for this study. The subject statistics are summarized in Table 1. Prior to the study, all subjects gave informed written consent, and the study was approved by the Institutional Review Board. None of the subjects had any orthopedic or neurological condition that could affect their gait. All subjects had owned and regularly worn MBT shoes for at least 6 weeks prior to the study. 2.2. Testing protocol Thirteen reflective markers (spheres with a diameter of 20 mm covered with retro-reflective tape) were placed on the foot, shank, thigh and pelvis to define limb segments for 3D kinetic and kinematic gait analysis. All kinetic and kinematic measurements during walking were performed on the left leg. A standing neutral trial was recorded with five additional markers defining anatomical segment coordinate systems with origins at the ankle, knee and hip joint centers (Silva & Ambrósio, 2002). These additional markers were placed on the medial and lateral malleoli, the medial and lateral epicondyles and the greater trochanter. After recording marker positions for the neutral trial, these additional markers were removed, and the thirteen marker configuration was used for the subsequent motion trials. Subjects completed three standing trials and five walking trials on a 30 m walkway with a force plate (Kistler Instrumente AG, Winterhur, Switzerland) embedded in the middle of the walkway. Force plate data was collected at a sampling rate of 2400 Hz. Kinematic data was collected using an eight camera motion capture system (Motion Analysis Corporation, Santa Rosa CA, USA) with a sampling rate of 240 Hz. EVa Real-Time Software (EVaRT, Motion Analysis Corporation, Santa Rosa CA, USA) was used for real-time motion capture and for post-processing and tracking of the marker data. Center of pressure (COP) excursion was measured during bilateral quiet stance with both feet placed hip-wide apart on the force plate and the subject's gaze focused on a target. For each trial, data were recorded for 30 s, and the middle 15 s were used for further analysis. Table 1 Anthropometric data for female and male subjects. Female (n=17) Males (n=17) Statistics Mean (SD) Mean (SD) t df P values Age (years) 48.1 (12.6) 42.4 (13.6) 1.299 32 0.20 Height (cm)* 163.0 (6.26)* 174.3 (5.4)* 1.753 32 0.00 Weight (kg) 76.5 (15.04) 85.7 (15.7) 5.612 32 0.09 Significant differences between genders (pb0.05) are indicated by an asterisk. During the walking trials, kinematic and kinetic data were collected for one step on the force plate while the subjects walked in their MBT shoes at 5.0±0.5 km/h controlled by infrared timing gates (MULTI-BEAM Series, Banner Engineering Corp., Minneapolis MN, USA). Lower limb kinematics and kinetics were recorded during the stance phase of the walking trials. 2.3. Data analysis COP data were filtered using a zero-lag 4th order low-pass Butterworth filter with a cut-off frequency of 50 Hz. COP excursion was quantified by determining the maximum excursions in the anterior posterior and medial lateral directions. In addition, COP path length was also calculated. Kinematic and kinetic data were filtered using a zero-lag 4th order low-pass Butterworth filter with cut-off frequencies of 12 and 50 Hz, respectively (Nigg et al., 2006a). All data was exported into Kintrak software (Human Performance Laboratory, Calgary AB, Canada) for kinematic and kinetic variable calculations. The method of inverse dynamics was used to obtain resultant internal joint moments using ground reaction force and kinematic data. Peak joint angles and joint moment impulses were calculated for the first and second half of stance. Joint moment impulse was defined as the time integral of the resultant joint moment for each of the three anatomical planes at the ankle, knee and hip joint. The first and second half of stance phase were analyzed separately to reflect upon the functional differences between these two sub-phases, that is weight acceptance and forward propulsion, respectively. For each subject, data for the three standing trials and the five walking trials were averaged. 2.4. Statistical analysis Gender differences in COP excursion, COP path length, peak joint angles and moment impulses were tested for statistical difference using a two-sided Student's t-tests for independent samples. In addition, a univariate ANOVA test using age, weight and height as predictors was performed to find out if these anthropometric variables might have an influence on statistical differences between variables. All statistical tests were performed using the Statistical Package for the Social Sciences version 16.0.1 (SPSS Inc., Chicago IL, USA, 2007). For comparison of the COP variables, the level of statistical significance was set a priori at 0.05. For comparison of the kinematic and kinetic gait variables during the first and second half of stance phase, the level of statistical significance was adjusted using the Bonferroni correction. Throughout the text, significant differences are indicated with an asterisk. 3. Results 3.1. Center of pressure (COP) during bilateral quiet stance COP excursion in the anterior posterior direction was significantly greater for the female group than for the male group (Table 2). Table 2 Mean (1 standard deviation) COP excursion and COP path length (both in mm) for female and male subjects. COP variable Mean (SD) Females Path length [mm] 391.0 (117.5) Medial lateral excursion [mm] 19.0 (16.7) Anterior posterior excursion [mm]* 45.7 (19.0)* Males Path length [mm] 394.6 (110.2) Medial lateral excursion [mm] 17.3 (6.7) Anterior posterior excursion [mm]* 39.2 (11.8)* Significant differences between genders (pb0.05) are indicated by an asterisk.

B.M. Nigg et al. / Clinical Biomechanics 25 (2010) 1047 1052 1049 Similarly, in the medio-lateral direction, females showed greater COP excursion, but the difference between female and male subject groups was not significant. The COP path length was not significantly different between the female group and the male group (Table 2). In addition to the gender effect on the COP movement, body height was found to be a significant covariate (pb0.025) in the anterior posterior COP excursion. 3.2. Walking kinematics Female subjects tended to walk with less dorsiflexion in their ankle joints in the second half of stance than male subjects but these differences were not significant (Table 3). In contrast, movement patterns at the knee joint showed several gender differences. Female subjects generally had less knee flexion from mid-stance to toe off compared to male subjects (Fig. 1a), and the peak extension angle at about 70% of stance phase differed significantly between genders (Table 3). In addition, female subjects had higher knee abduction angles throughout the entire stance phase compared to male subjects (Fig. 1b). At the time of heel strike, the knee was less externally rotated for female compared to male subjects (Fig. 1c). At the hip joint, higher adduction angles were observed for female subjects throughout the entire stance phase compared to male subjects (Fig. 1d). Peak adduction angles differed between groups by 4.4 in the first half of stance phase and by 6.0 in the second half of stance phase, respectively (Table 3). None of the anthropometric variables were found to have a significant effect on the peak joint angles in gait. 3.3. Walking kinetics Joint moments as determined from inverse dynamic calculations showed some gender differences at the ankle and knee joints but not at the hip joint. Compared to males, females had smaller ankle plantarflexion moments from about 20% to about 90% of stance phase (Fig. 2a). Ankle plantarflexion moment impulse during the first and second half of stance phase was significantly smaller for female subjects than those for male subjects (Table 4). In addition, female subjects showed smaller ankle adduction moments during push-off (Fig. 2b) and 34% smaller ankle adduction moment impulse in the second half of stance phase compared to male subjects (Table 4). At the knee joint, female subjects had smaller extension moments and 61% smaller extension moment impulses during the first half of stance phase than male subjects (Fig 2c; Table 4). When including the anthropometric variables as covariates in the statistical analysis, weight was found to be the only variable that significantly influenced the moment impulses (p b0.025), with the exception of ankle abduction moments which also showed a significant gender effect. 4. Discussion 4.1. Gender differences during bilateral quiet stance in unstable shoes Female subjects showed larger COP excursions than male subjects. Previous studies had found no significant differences in postural sway between genders (Hageman et al., 1995, Røgind et al., 2003) or even reported less sway for female subjects compared to male subjects (Era et al., 2006). Thus, the differences in anterior posterior sway observed in this study are most likely caused by the unstable shoe design and not by general gender differences or anthroprometric factors affecting sway. When standing in unstable shoes, female subjects presumably experience greater instability than male subjects, particularly in the anterior posterior direction. The physiological and neuromuscular differences between women and men reported by Thelen et al. (1996) seem to influence the ability to compensate for the instability imposed by the unstable shoes. Hence, potential small muscle training or rehabilitative effects would likely be more pronounced in women than in men. However, the gender difference in COP excursion found in this study (9% in medial lateral and 17% in anterior posterior direction) is relatively small compared to the difference in COP excursion when standing in MBT shoes as compared to standing in normal shoes of 105% and 52% as reported by Nigg, Hintzen & Ferber (2006). 4.2. Gender differences in walking kinematics Generally, the joint movement patterns observed in this study are in agreement with those reported in previous studies investigating gender differences using three-dimensional gait analysis (Boyer et al., 2008, Cho et al., 2004, Chumanov et al., 2008, Ferber et al., 2003, Kerrigan et al., 1998). Specifically, good agreement between our results and previously published results was found for the peak hip adduction angle (Boyer et al., 2008, Cho et al., 2004, Chumanov et al., 2008, Ferber et al., 2003, Kerrigan et al., 1998. Some previous studies Table 3 Mean (1 standard deviation) peak angles (degrees) during stance phase of gait for males and females. Females Males Statistics Stance phase Stance phase 1st half of stance phase 2nd half of stance phase 1st half 2nd half 1st half 2nd half t df P values t df P values Ankle dorsiflexion 0.4 12.3 1.7 15.5 0.87 32 0.390 2.024 32 0.051 (5.3) (4.67) (3.1) (4.4) Ankle adduction 5.5 5.9 7.2 5.3 1.26 32 0.219 0.318 32 0.752 (4.7) (4.5) (3.2) (5.1) Ankle inversion 6.3 6.2 4.3 7.1 1.87 32 0.071 0.619 32 0.54 (3.5) (5.0) (3.0) (3.5) Knee extension 22.5 1.5 27.4 7.5 2.18 32 0.037 2.596 32 0.014* (5.6) (7.0) (7.3) (6.5) Knee adduction 1.9 5.3 5.4 2.8 2.75 32 0.010* 1.975 32 0.057 (3.5) (3.8) (3.9) (3.5) Knee internal rotation 3.5 4.3 9.3 3.9 3.32 32 0.002* 0.324 32 0.748 (4.1) (2.9) (5.89) (3.7) Hip flexion 37.7 19.1 39.0 18.8 0.41 32 0.682 0.139 32 0.891 (7.3) (5.1) (10.8) (8.3) Hip adduction 9.8 2.1 5.4 8.1 3.20 32 0.003* 5.090 32 b0.001* (3.6) (3.4) (4.4) (3.5) Hip internal rotation 6.9 2.7 5.7 2.6 0.72 32 0.475 0.062 32 0.951 (5.2) (3.6) (4.3) (4.4) Significant differences (pb0.025) are indicated with an asterisk.

1050 B.M. Nigg et al. / Clinical Biomechanics 25 (2010) 1047 1052 Fig. 2. Mean joint moments (and standard error of all trials) during the stance phase of walking for those joint moments with significant gender effects. Curves for female subjects are shown as solid lines, and curves for male subjects are shown as dotted lines. Fig. 1. Mean joint angles (and standard error of all trials) during the stance phase of walking for those joint angles with significant gender effects. Curves for female subjects are shown as solid lines, and curves for male subjects are shown as dotted lines. have reported differences in hip flexion angles (Cho et al., 2004, Kerrigan et al., 1998) or hip internal rotation angles (Cho et al., 2004, Chumanov et al., 2008), which were not observed in this study. One possible reason for the discrepancy between results might be subtle changes in gait characteristics with age. The mean age of the subjects in this study was in the forties, while the subjects in those studies that reported differences in hip angles were in their twenties. Boyer et al. (2008), studied an older subject group and also did not find differences in hip internal rotation. For the knee joint, most previous studies reported smaller flexion angles for female subjects during most of the stance phase (Boyer et al., 2008, Kerrigan et al., 1998) which agrees with the findings in this study, however, Chumanov et al., 2008 reported greater flexion angles for female subjects compared to male subjects. Higher knee abduction angles for female subjects compared to male subjects during running were reported by Ferber et al. (2003), but there are no reports of a gender difference in this variable for walking. Gender differences for ankle joint angles have been mentioned previously specifically for plantarflexion at heel strike and toe off (Boyer et al., 2008, Kerrigan et al., 1998). In this study, gender differences in peak ankle angles during the first and second half of stance phase were not significant. All gender differences in joint movement patterns during walking with MBT shoes found in this study are in agreement with the results of earlier studies that tested barefoot or stable shoe conditions. Hence, walking in MBT shoes does not appear to affect the walking kinematics of females differently than those of males. 4.3. Gender difference in joint moments Gender differences in the joint kinetic variables reported in the literature are not consistent. Boyer et al. (2008) reported differences in hip joint moments in all three anatomical planes between female and male subjects, while other studies only reported gender differences in hip adductor and internal rotation torques (Cho et al., 2004) or no gender differences in hip joint moments in all anatomical planes (Kerrigan et al., 2000). While in this study we observed a trend indicating smaller internal hip rotation moments in the first half of the stance phase for female subjects compared to male subjects, our results are in general agreement with Kerrigan et al. (2000). At the knee joint, higher extension moments and extension moment impulses in the first half of the stance phase were found in

B.M. Nigg et al. / Clinical Biomechanics 25 (2010) 1047 1052 1051 Table 4 Mean (1 standard deviation) angular impulses (Nm s) during stance phase of gait for males and females. Females Males Stance phase Stance phase 1st half of stance phase 2nd half of stance phase 1st half 2nd half 1st half 2nd half t df P values t df P values Ankle dorsiflexion 4.0 22.1 6.2 28.1 2.82 27.23 0.009* 3.94 32 b0.001* (1.7) (4.0) (2.7) (4.8) Ankle adduction 0.4 2.7 0.5 4.0 0.70 32 0.489 2.84 32 0.008* (0.5) (1.5) (0.5) (1.3) Ankle inversion 1.2 1.6 1.5 1.8 0.64 32 0.008* 0.30 32 0.53 (1.3) (2.2) (1.3) (1.5) Knee extension 3.5 5.6 8.9 3.8 3.76 32 0.001* 1.29 32 0.205 (3.5) (3.9) (4.8) (4.2) Knee adduction 6.6 5.3 8.56 6.7 1.71 32 0.098 1.46 32 0.155 (2.9) (2.5) (3.7) (3.2) Knee internal rotation 0.6 2.7 0.6 3.3 0.27 32 0.792 1.81 32 0.080 (0.5) (0.89) (0.6) (1.2) Hip flexion 10.2 5.3 12.4 3.7 1.01 32 0.321 0.92 32 0.365 (5.0) (4.1) (7.6) (5.6) Hip adduction 15.2 15.6 13.6 15.9 0.86 32 0.398 0.18 32 0.858 (5.8) (5.2) (5.0) (4.4) Hip internal rotation 1.4 2.4 2.3 1.8 2.15 32 0.040 1.67 32 0.105 (1.1) (1.2) (1.3) (0.9) Significant differences (pb0.025) are indicated with an asterisk. some previous studies (Kerrigan et al., 1998, 2000) while other studies (Boyer et al., 2008) did not report gender differences for kinetic gait variables at the knee joint. The gender differences in knee joint moments found in this study appear to be caused by the weight difference between women and men rather than an intrinsic gender effect in gait kinetics. There are no previous reports of significant gender differences for kinetic gait variables at the ankle joint during walking. However, this study showed very clear differences in ankle plantarflexion moments and abduction moments between genders. Particularly during the second half of the stance phase (push-off), the MBT shoe appears to cause higher ankle plantarflexion moments and ankle abduction moments in female subjects compared to male subjects. No differences in these variables have been reported for walking in MBT shoes compared to normal control shoes (Nigg, Hintzen & Ferber, 2006). However, in running with unstable shoes a reduction of the ankle plantarflexion, dorsiflexion and inversion moments of comparable magnitude was found (Boyer & Andriacchi, 2009). Similar to the quiet stance situation, the ratio of reaction time to additional time to achieve a target plantarflexion torque in women compared to men may be disadvantageous for their ability to compensate for the instability imposed by the MBT shoes (Thelen et al., 1996). The fact that differences in ankle moments but not in ankle kinematics between genders were apparent suggests a gender difference in the center of pressure motion during the stance phase of walking. In general, these results imply that women and men use different strategies to control their center of mass motion when walking in MBT shoes. 4.4. Clinical implications Gender differences in the effectiveness of unstable shoes as a therapeutic treatment should be taken into consideration, for example, when setting the position of the pivot point of the sole on. Wearing unstable shoes generally has substantial effects on lower extremity function during quiet bilateral stance and these effects appear to be stronger in women than in men. During ambulation the unstable shoes affect the kinematics of the ankle joint movement and they seem to reduce joint moments (Nigg, Emery, & Hiemstra, 2006). However, these changes, as well as the gender differences observed in this study are probably too small to be of clinical relevance. In general, the MBT shoe is a potential therapeutic intervention for the training of neuromuscular control without posing additional loads on joints, yet it may be more effective in women than in men. 5. Conclusions When wearing unstable shoes gender differences were observed a) in the anterior posterior sway during quiet bipedal stance and b) in the moments controlling the ankle joint movement during the stance phase of the gait cycle. These differences suggest that men and women are affected differently by the instability created by unstable shoes and that they might use different strategies when compensating for this instability. Therefore, gender differences and particularly differences in how women and men control the ankle joint, should be taken into consideration when investigating functional or therapeutic effects of unstable shoe designs. Conflict of interest The company MBT has sponsored several research projects conducted by the authors. MBT has not in any way influenced the study design, the measurement procedures, the data analysis or the interpretation of the results. Acknowledgements This study combines the results of several previous studies sponsored by the company Masai Barefoot Technology MBT, Switzerland. Additional funding was provided by the Da Vinci Foundation in Calgary, Canada. The help of Jonathan Melvin and Mac Kim in data collection and data analysis are gratefully acknowledged. References Boyer, K.A., Andriacchi, T.P., 2009. Changes in running kinematics and kinetics in response to a rocker shoe intervention. Clin. Biomech. 24, 872 876. Boyer, K.A., Beaupre, G.S., Andriacchi, T.P., 2008. Gender differences exist in the hip joint moments of healthy older walkers. J. Biomech. 41, 3360 3365. Cho, S.H., Park, J.M., Kwon, O.Y., 2004. Gender differences in three dimensional gait analysis data from 98 healthy Korean adults. Clin. Biomech. 19, 145 152. Chumanov, E.S., Wall-Scheffler, C., Heiderscheit, B.C., 2008. Gender differences in walking and running on level and inclined surfaces. Clin. Biomech. 23, 1260 1268.

1052 B.M. Nigg et al. / Clinical Biomechanics 25 (2010) 1047 1052 Era, P., Sainio, P., Koskinen, S., Haavisto, P., Vaara, M., Aromaa, A., 2006. Postural balance in a random sample of 7979 subjects aged 30 years and over. Gerontology 52, 204 213. Ferber, R., Davis, I.M., Williams III, D.S., 2003. Gender differences in lower extremity mechanics during running. Clin. Biomech. 18, 350 357. Hageman, P.A., Leibowitz, J.M., Blanke, D., 1995. Age and gender effects on postural control measures. Arch. Phys. Med. Rehabil. 76, 961 965. Kerrigan, D.C., Todd, M.K., Della Croce, U., 1998. Gender differences in joint biomechanics during walking: normative study in young adults. Am. J. Phys. Med. Rehab. 77, 2 7. Kerrigan, D.C., Riley, P.O., Nieto, T.J., Della Croce, U., 2000. Knee joint torques: a comparison between women and men during barefoot walking. Arch. Phys. Med. Rehabil. 81, 1162 1165. Nigg, B., 2005. Der MBT-Schuh und seine biomechanische/therapeutische Wirkungsweise. Schuh-Orthopädie 3, 77 78. Nigg, B., Hintzen, S., Ferber, R., 2006a. Effect of an unstable shoe construction on lower extremity gait characteristics. Clin. Biomech. 21, 82 88. Nigg, B.M., Emery, C., Hiemstra, L.A., 2006b. Unstable shoe construction and reduction of pain in osteoarthritis patients. Med. Sci. Sports Exerc. 38, 1701 1708. Nigg, B.M., Davis, E.B.S., Lindsay, D., Emery, C.B.S., 2009. The effectiveness of an unstable sandal on low back pain and golf performance. Clin. J. Sport Med. 19 (6), 464 470. Robinovitch, S.N., Heller, B., Lui, A., Cortez, J., 2002. Effect of strength and speed of torque development on balance recovery with the ankle strategy. J. Neurophysiol. 88, 613 620. Røgind, H., Lykkegaard, J.J., Bliddal, H., Danneskiold-Samsøe, B., 2003. Postural sway in normal subjects aged 20 70 years. Clin. Physiol. Funct. Imaging 23, 171 176. Romkes, J., 2008. Static postural control with the MBT shoe. Schweiz. Z. Sportmed. Sporttraumatologie. 56, 61 65. Romkes, J., Rudmann, C., Brunner, R., 2006. Changes in gait and EMG when walking with the Masai Barefoot Technique. Clin. Biomech. 21, 75 81. Silva, M.P.T., Ambrósio, J.A.C., 2002. Kinematic data consistency in the inverse dynamic analysis of biomechanical systems. Multibody Sys. Dyn. 8, 219 239. Stewart, L., Gibson, J.N.A., Thomson, C.E., 2007. In-shoe pressure distribution in unstable (MBT) shoes and flat-bottomed training shoes: a comparative study. Gait Posture 25, 648 651. Thelen, D.G., Schultz, A.B., Alexander, N.B., 1996. Effects of age on rapid ankle torque development. J. Gerontol. A Biol. Sci. Med. Sci. 51, M226 M232.