Differences in lower limb transverse plane joint moments during gait when expressed in two alternative reference frames

Transcription

1 Journal of Biomechanics 40 (2007) Differences in lower limb transverse plane joint moments during gait when expressed in two alternative reference frames Anthony G. Schache a,, Richard Baker a, Christopher L. Vaughan b a Murdoch Childrens Research Institute and Hugh Williamson Gait Laboratory, Royal Childrens Hospital, Flemington Road, Parkville, Melbourne, Victoria 3052, Australia b Department of Human Biology, Faculty of Health Sciences, University of Cape Town, Observatory, Western Cape 7925, South Africa Accepted 5 December 2005 Abstract When comparing previous studies that have measured the three-dimensional moments acting about the lower limb joints (either external moments or opposing internal joint moments) during able-bodied adult gait, significant variation is apparent in the profiles of the reported transverse plane moments. This variation cannot be explained on the basis of adopted convention (i.e. external versus internal joint moment) or inherent variability in gait strategies. The aim of the current study was to determine whether in fact the frame in which moments are expressed has a dominant effect upon transverse plane moments and thus provides a valid explanation for the observed inconsistency in the literature. Kinematic and ground reaction force data were acquired from nine ablebodied adult subjects walking at a self-selected speed. Three-dimensional hip, knee and ankle joint moments during gait were calculated using a standard inverse dynamics approach. In addition to calculating internal joint moments, the components of the external moment occurring in the transverse plane at each of the lower limb joints were calculated to determine their independent effects. All moments were expressed in both the laboratory frame (LF) as well as the anatomical frame (AF) of the distal segment. With the exception of the ankle rotation moment in the foot AF, lower limb transverse plane joint moments during gait were found to display characteristic profiles that were consistent across subjects. Furthermore, lower limb transverse plane joint moments during gait differed when expressed in the distal segment AF compared to the LF. At the hip, the two alternative reference frames produced near reciprocal joint moment profiles. The components of the external moment revealed that the external ground reaction force moment was primarily responsible for this result. Lower limb transverse plane joint moments during gait were therefore found to be highly sensitive to a change in reference frame. These findings indicate that the different transverse plane joint moment profiles during able-bodied adult gait reported in the literature are likely to be explained on this basis. r 2006 Elsevier Ltd. All rights reserved. Keywords: Gait analysis; Inverse dynamics; Hip; Knee; Ankle 1. Introduction Numerous studies have measured the three-dimensional (3D) moments acting about the lower limb joints (either external moments or opposing internal joint moments) during able-bodied adult gait (Allard et al., 1996; Andriacchi and Strickland, 1985; Benedetti et al., Corresponding author. Tel.: +61 (03) address: anthony.schache@mcri.edu.au (A.G. Schache). 1998; Besier et al., 2003; Bowsher and Vaughan, 1995; Bresler and Frankel, 1950; Eng and Winter, 1995; Ramakrishnan et al., 1987). Irrespective of the adopted convention (i.e. external versus internal joint moment), all studies in general report consistent joint moment profiles for the sagittal and frontal planes. However, significant variation does exist when comparing the transverse plane joint moment profiles, particularly for the hip. This was first noted by Allard et al. (1996) and is highlighted when comparing two technical notes, the /$ - see front matter r 2006 Elsevier Ltd. All rights reserved. doi: /j.jbiomech

2 10 A.G. Schache et al. / Journal of Biomechanics 40 (2007) 9 19 first by Eng and Winter (1995) and the second by Bowsher and Vaughan (1995), both of which were published consecutively in Volume 28, Edition No. 6 of the Journal of Biomechanics. These two studies report 3D internal joint moments at the hip during gait. The sagittal and frontal plane hip joint moments are consistent but the transverse plane joint moments are opposite. Eng and Winter (1995) report an external rotator hip joint moment during initial stance followed by an internal rotator hip joint moment during terminal stance. Bowsher and Vaughan (1995) report a profile of similar shape but reversed convention. Whilst at first it may be suggested that this inconsistency is merely a product of inherent variability in gait strategies across subjects, such an explanation does not fit the findings. In contrast to this notion, it has been demonstrated that the transverse plane hip joint moment during the stance phase of gait does in fact display a characteristic profile that is remarkably consistent across subjects (Andriacchi and Strickland, 1985; Benedetti et al., 1998; Eng and Winter, 1995; Ramakrishnan et al., 1987). Furthermore, it is rather difficult to comprehend how a change in gait strategy would have almost no effect upon sagittal and frontal plane hip joint moments yet an apparent opposing effect upon the transverse plane hip joint moment. These factors would therefore suggest that an alternative explanation must exist. It is possible that transverse plane joint moments during gait are highly sensitive to the reference frame in which they are expressed. A comparison of lower limb transverse plane joint moment profiles during gait from previous studies would suggest that this is feasible. Studies that have expressed the net moment vector in the transverse plane of the laboratory frame (LF) (Andriacchi and Strickland, 1985; Bresler and Frankel, 1950; Eng and Winter, 1995; Ramakrishnan et al., 1987) report markedly different lower limb transverse plane joint moment profiles during gait when compared to studies that have expressed the net moment vector in the transverse plane of the distal segment anatomical frame (AF) (Allard et al., 1996; Benedetti et al., 1998; Besier et al., 2003; Bowsher and Vaughan, 1995). Based on this, it was hypothesised that the contrasting transverse plane hip joint moment profiles during gait from Eng and Winter (1995) and Bowsher and Vaughan (1995) could be explained on the basis of differing reference frames. The aim of the current study was to therefore test this hypothesis and compare lower limb transverse plane joint moments during gait when expressed in the LF versus the distal segment AF. If this hypothesis proved to be true, then an additional aim was to analyse the relative effects of each component of the external moment acting in the transverse plane at the lower limb joints in an endeavour to provide a thorough explanation for this finding. 2. Materials and methods 2.1. Subjects Nine able-bodied adults (two males; seven females) with a mean height of (SD 8.5) cm, body mass of 60.0 (SD 11.1) kg and age of 19.8 (SD 2.1) years were voluntarily recruited. Approval was obtained from the Royal Children s Hospital Ethics in Human Research Committee prior to commencement and subjects signed a consent form Instrumentation Kinematic data were acquired using a 3D motion analysis system (VICON, Oxford Metrics, Oxford, England) with six cameras operating at a sampling rate of 120 Hz. The LF was defined such that the positive x direction was in the forward direction, the positive y direction was orientated to the left and the positive z direction was orientated upwards. Two AMTI forceplates (Advanced Mechanical Technology, Inc., Watertown, MA) were used to capture ground reaction force data at a sampling rate of 1080 Hz Frame definitions The 3D pose of the seven body segments of interest (pelvis; left and right thighs; left and right shanks; both feet) were obtained by tracking the trajectories of nonrigid clusters of small spherical retro-reflective markers (25 mm diameter) mounted in a non-colinear fashion as outlined in Table 1. These markers allowed the reconstruction of technical frames (TF) embedded in each of the body segments (Table 2). Given the potential arbitrary relationship between the defined TFs and the anatomy of the underlying bone, a static anatomical landmark calibration trial was performed to reconstruct the various AFs. Respective AF definitions are detailed in Table 3. The knee joint flexion extension was defined using a dynamic optimisation procedure as described by Schache et al. (2005). The vertical (z) of the femur was first defined (knee joint centre (KJC) to hip joint centre (HJC)) and then the mediolateral (y) (or knee joint flexion extension ) was rotated about the vertical (z) through an angle y, whereby y represented the degree of rotation necessary to minimise the variance in the dynamic knee varus-valgus kinematic profile Procedures Anthropometric parameters required for estimating the location of the HJC using the method of Davis et al. (1991) were first measured. Markers were then placed on each subject s pelvis and lower limbs as previously described. The same tester (AS) performed all

3 A.G. Schache et al. / Journal of Biomechanics 40 (2007) Table 1 Specific marker locations and orientations Static and dynamic trials LASIS (RASIS) SACR LTH1 (RTH1) LTH2 (RTH2) LTH3 (RTH3) LSH1 (RSH1) LSH2 (RSH2) LANK (RANK) LCAL1 (RCAL1) LMID (RMID) LLATMID (RLATMID) Anterior to left (and right) anterior superior iliac spine (ASIS) lying in plane containing left and right ASIS and the mid-point between both posterior superior iliac spines (PSIS) Posterior to the mid-point between both PSISs lying in plane containing left and right ASISs and the mid-point between both PSISs Distal and anterior aspect of left (and right) thigh Distal and lateral aspect of left (and right) thigh Distal and posterior aspect of left (and right) thigh Proximal end of left (and right) anterior tibia just distal to tibial tubercle Distal end of left (and right) anterior tibia Left (and right) lateral malleolus aligned with bimalleolar Bisection of the distal aspect of the left (and right) posterior calcaneum Left (and right) medial midfoot over the distal and dorsomedial aspect of the navicular Left (and right) lateral midfoot over the dorsal and distal aspect of the cuboid Static anatomical landmark calibration trial only LMFE (RMFE) Most prominent palpable aspect of left (and right) medial femoral epicondyle (MFE) LLFE (RLFE) Most prominent palpable aspect of left (and right) lateral femoral epicondyle (LFE) LTHI ROT (RTHI ROT ) Virtual point, defined as rotated position of LTH2 (or RTH2) marker (see text for further explanation) LMED (RMED) Left (and right) medial malleolus aligned with bimalleolar. LCAL2 (RCAL2) Bisection of the proximal aspect of the left (and right) posterior calcaneum LTOE (RTOE) Dorsal surface of the left (and right) distal forefoot at the midpoint between the 2nd and 3rd metatarsophalangeal joints Table 2 Technical frame definitions Pelvis Mediolateral (y) Vertical (z) Thigh Mediolateral (y) Vertical (z) Shank Vertical (z) Mediolateral (y) Foot Mediolateral (y) Vertical (z) Mid-point between LASIS and RASIS markers In direction from RASIS to LASIS markers Perpendicular to mediolateral (y) in plane containing LASIS, RASIS, SACR markers Mid-point between LTH1 (or RTH1) and LTH3 (or RTH3) markers In direction from LTH3 (or RTH3) to LTH1 (or RTH1) markers Perpendicular to anterior posterior (x) in plane containing LTH1 (or RTH1), LTH2 (or RTH2) and LTH3 (or RTH3) markers LSH2 (or RSH2) marker In direction from LSH2 (or RSH2) marker to LSH1 (or RSH1) marker Perpendicular to vertical (z) in plane containing LSH1 (or RSH1) marker, LSH2 (or RSH2) marker and LANK (or RANK) marker LCAL1 (or RCAL1) marker In direction from LCAL1 (or RCAL1) marker to LLATMID (or RLATMID) marker Perpendicular to anterior posterior (x) in plane containing LCAL1 (or RCAL1) marker, LLATMID (or RLATMID) marker and LMID (or RMID) marker marker placements. Testing commenced with the capture of the static anatomical landmark calibration trial. Calibration markers were then removed and dynamic gait trials were captured. Subjects walked at a self-selected speed ( m/s) through the middle of a walkway with a calibrated field approximately

4 12 A.G. Schache et al. / Journal of Biomechanics 40 (2007) 9 19 Table 3 Anatomical frame definitions Pelvis Mid-point between LASIS and RASIS markers Mediolateral (y) In direction from RASIS to LASIS markers Perpendicular to mediolateral (y) in plane containing LASIS, RASIS, SACR markers Vertical (z) Virtual point LHJC (or RHJC), defined relative to pelvic anatomical frame as per Davis et al. (1991) Femur Vertical (z) Mediolateral (y) Virtual points Tibia (proximal) Vertical (z) Mediolateral (y) Virtual point Tibia (distal) Vertical (z) Mediolateral (y) Virtual point Foot Vertical (z) Mediolateral (y) Virtual point LKJC (or RKJC), defined as mid-point between LLFE (or RLFE) marker and LMFE (or RMFE) marker In direction from LKJC (or RKJC) to LHJC (or RHJC) Perpendicular to vertical (z) in plane containing LKJC (or RKJC), LHJC (or RHJC), and LTHI (RTHI) marker, rotated by angle y about vertical (z), whereby y represents degree of rotation required to minimise variance in dynamic knee varus-valgus kinematic profile LKJC (or RKJC) and LTHI ROT (or RTHI ROT ) as defined above LAJC (or RAJC), defined as mid-point between LANK (or RANK) marker and LMED (or RMED) marker In direction from LAJC (or RAJC) to LKJC (or RKJC) location Perpendicular to vertical (z) and parallel to femur mediolateral (y) when in anatomical landmark calibration configuration LAJC (or RAJC) as defined above LAJC (or RAJC), defined as mid-point between LANK (or RANK) marker and LMED (or RMED) marker In direction from LAJC (or RAJC) to LKJC (or RKJC) location. Perpendicular to vertical (z) in plane containing LKJC (or RKJC), LMED (or RMED) and LANK (or RANK) markers LAJC (or RAJC) as defined above LAJC (or RAJC), defined as mid-point between LANK (or RANK) marker and LMED (or RMED) marker In direction from LCAL1 (or RCAL1) marker to LTOE (or RTOE) marker but rotated in plane containing LCAL1 (or CAL1), LCAL2 (or RCAL2) and LTOE (or RTOE) markers until parallel with floor (horizontal plane of laboratory frame) Perpendicular to anterior posterior (x) in plane containing LCAL1 (or CAL1), LCAL2 (or RCAL2) and LTOE (or RTOE) markers LAJC (or RAJC) as defined above NB: HJC, hip joint centre; KJC, knee joint centre; AJC, ankle joint centre. 5 m in length. A single gait trial was captured for each subject, whereby the left and right heels successfully struck the two adjacent force plates in isolation with no evident force plate targeting as observed by the tester Data analysis Coordinate data were filtered using Woltring s general cross-validatory quintic smoothing spline (Woltring, 1986) with a predicted mean-squared error of 15 mm. Kinematic data were calculated using a joint coordinate system convention (Grood and Suntay, 1983). Net internal joint moments were calculated using an inverse dynamics approach with adapted inertial parameters as per De Leva (1996). In addition, the components of the external moment occurring in the transverse plane at each joint were calculated to determine their independent effects. The two components of the external moment calculated were: (a) the moment due to the resultant ground reaction force acting at a distance to the joint centre (external GRF moment) and (b) the free moment of rotation acting about a vertical through the centre of pressure (external Free moment). The component of the external moment due to segment weight and inertia in the transverse plane was considered to be extremely small and thus its independent effect was not calculated. All moments were normalised

5 by dividing by subject s body mass and were expressed in both the LF as well as the AF of the distal segment (i.e. the femoral AF for hip joint moments; the proximal tibial AF for knee joint moments; the foot AF for ankle joint moments). The kinematic and kinetic computations were performed using Bodybuilder software (Oxford Metrics Ltd, Oxford, England). Temporal events defining the gait cycle were identified from the ground reaction force data. Each stride was time normalised to 101 points representing equal intervals from 0% to 100% using Polygon software (Oxford Metrics Ltd, Oxford, England). A.G. Schache et al. / Journal of Biomechanics 40 (2007) Results Lower limb transverse plane internal joint moments during gait are illustrated in Fig. 1. Except for the ankle rotation moment expressed in the foot AF, the patterns were found to be quite consistent across subjects, as indicated by the relatively small standard deviation bands. The transverse plane joint moments expressed in the distal segment AF displayed dramatically different patterns to the equivalent moments expressed in the LF. For the hip joint in particular, the two reference frames produced near reciprocal profiles. In contrast, the particular reference frame had little influence on the profiles for the sagittal plane hip, knee and ankle joint moments as well as the frontal plane hip and knee joint moments. The components of the external moment (external GRF moment and external Free moment) along with their combined effect (external Total moment) in the transverse plane at each of the lower limb joints during gait are illustrated in Figs. 2 and 3. The data highlight three important points. First, for each joint, the external Total moment was virtually equivalent to the reciprocal of the respective internal joint moment expressed in the same reference frame (Fig. 2(c) versus Fig. 1(b); Fig. 3(c) versus Fig. 1(a)). This meant that in the transverse plane the component of the external moment due to segment weight and inertia alone did not have an appreciable effect and thus could be ignored for the purposes of this analysis. Second, the effect of the external Free moment was near identical when expressed in either the LF or the distal segment AF (Fig. 2(b) versus Fig. 3(b)). The external Free moment itself therefore could not be used to explain the observed inconsistencies in the lower limb transverse plane joint moments when expressed in the two different reference frames. Third, in contrast to the external Free moment, the effect of the external GRF moment in the transverse plane was found to be highly sensitive to a change in reference frame (Figs. 2(a) versus Fig. 3(a)). The external GRF moment had differing effects at each of the lower limb joints during gait when expressed in the LF compared to the AF of the distal segment. For example, at the hip joint, the effect of the Fig. 1. Lower limb transverse plane internal joint moments during gait: (a) expressed in the distal segment AF and (b) expressed in the LF. Solid line: right side, dashed line: left side. Shaded area indicates the standard deviation band about the group mean. external GRF moment in the femoral AF was opposite to its effect in the LF. The observed differences in the lower limb transverse plane joint moments when expressed in the two reference frames (Fig. 1(a) and (b)) could therefore be explained with consideration to the effect of the external GRF moment. 4. Discussion There are two major findings from this study. First, with the exception of the ankle rotation moment in the foot AF, lower limb transverse plane joint moments during gait displayed characteristic profiles that were consistent across subjects. This is in accordance with previous studies (Andriacchi and Strickland, 1985; Benedetti et al., 1998; Eng and Winter, 1995; Ramakrishnan et al., 1987). Second, lower limb transverse plane joint moments during gait differed when expressed

6 14 A.G. Schache et al. / Journal of Biomechanics 40 (2007) 9 19 Fig. 2. The components of the external moment in the transverse plane, expressed in the LF, at each of the lower limb joints during gait: (a) the resultant ground reaction force acting at a distance to the joint centre (external GRF moment); (b) the free moment of rotation acting about a vertical through the centre of pressure (external Free moment); (c) the external GRF and Free moments combined excluding the component of the external moment due to segment weight and inertia (external Total moment). The external Free moment and the external GRF moment can be seen to have complimentary effects at the ankle joint but opposing effects at the hip and knee joints. Solid line: right side, Dashed line: left side. Shaded area indicates the standard deviation band about the group mean. in the distal segment AF compared to the LF. This was not the case for the sagittal plane hip, knee and ankle joint moments as well as the frontal plane hip and knee joint moments. The ankle rotation moment in the foot AF displayed the greatest degree of variability across subjects (Fig. 1(a)), consistent with a previous study (Hunt and Smith, 2001). This moment is a product of forces acting about the defined vertical (z) of the foot AF. As this was aligned with the bisection of the posterior calcaneus (Table 3 and Fig. 4), the ankle rotation moment in the foot AF was therefore highly sensitive to the orientation of the posterior calcaneus in the frontal plane (Hunt and Smith, 2001). Given that both static posterior calcaneal alignment and dynamic inversion eversion movements during gait can vary considerably across subjects, it is not surprising that this moment was variable across subjects. As the ankle rotation moment in the foot AF is dependent upon the specific foot model, this moment can also be expected to vary somewhat across studies. Numerous studies have measured lower limb transverse plane joint moments during able-bodied adult gait (Allard et al., 1996; Andriacchi and Strickland, 1985; Benedetti et al., 1998; Besier et al., 2003; Bowsher and Vaughan, 1995; Bresler and Frankel, 1950; Eng and Winter, 1995; Ramakrishnan et al., 1987). Andriacchi and Strickland (1985), Bresler and Frankel (1950), Eng and Winter (1995) and Ramakrishnan et al. (1987) all expressed the net moment vector in the LF. The transverse plane joint moment profiles from these studies are consistent with those from the current study when expressed in the LF. Allard et al. (1996), Benedetti et al. (1998), Besier et al. (2003) and Bowsher and Vaughan (1995) all expressed the net moment vector in the distal segment AF. The transverse plane joint moment profiles from these studies are consistent with those from the current study when expressed in the

7 A.G. Schache et al. / Journal of Biomechanics 40 (2007) Fig. 3. The components of the external moment in the transverse plane, expressed in the distal segment AF, at each of the lower limb joints during gait: (a) the resultant ground reaction force acting at a distance to the joint centre (external GRF moment); (b) the free moment of rotation acting about a vertical through the centre of pressure (external Free moment); (c) the external GRF and Free moments combined excluding the component of the external moment due to segment weight and inertia (external Total moment). The external Free moment and the external GRF moment can be seen to have complimentary effects at all lower limb joints. Solid line: right side, dashed line: left side. Shaded area indicates the standard deviation band about the group mean. distal segment AF. This suggests that the conflicting transverse plane joint moment profiles reported in the literature are likely to be explained on the basis of alternative reference frames rather than inherent variability in gait strategies. For a given lower limb joint (hip, knee or ankle) and reference frame (LF or distal segment AF), an analysis of the relative effects of each of the components of the external moment in the transverse plane makes it possible to determine which are dominant. As segment weight and inertia were not found to have an appreciable effect, the dominant components were the external GRF and Free moments. In general, the magnitude of the external GRF moment exceeded that of the external Free moment (Figs. 2 and 3). The addition of these two moments (external Total moment) was able to provide a reasonable reflection of the net effect of the external moment in the transverse plane during stance. From this one can deduce the expected internal joint moments, which should then match the measured data from the inverse dynamics solution. Inspection of the data demonstrated this to be the case. For a particular lower limb joint and reference frame, the profile of the stance phase internal joint moment (Fig. 1) was virtually equivalent to the reciprocal of the respective external Total moment (Figs. 2(c) and 3(c)). Due to the relative dominance of the external GRF moment, its effect can be used to explain the observed differences in the lower limb transverse plane joint moments during gait when expressed in alternative reference frames (Fig. 1(a) and (b)). The external GRF moment had quite different effects at all lower limb joints when expressed in the LF compared to the distal segment AF, whereas the action of the external Free moment remained unaffected. To explain the sensitivity of the external GRF moment to a change in reference

8 16 A.G. Schache et al. / Journal of Biomechanics 40 (2007) 9 19 Fig. 4. Coronal view of the foot AF. Point A depicts the bi-malleolar, point B depicts the bisection of the posterior calcaneus and AJC indicates the ankle joint centre. Note the dependency of the foot AF vertical (z) on the alignment of the posterior calcaneus which medially inclines the foot AF transverse plane with respect to the LF for an everted rearfoot. frame, an appreciation of the orientation of the GRF vector with respect to the relevant joint centre for a given lower limb joint and reference frame is required. At the hip, the GRF vector was medial to the HJC during stance, irrespective of the reference frame, as both conventions calculated an abductor internal hip joint moment. In the LF transverse plane, the GRF vector was antero-medial to the HJC pointing posteromedially during initial stance whilst postero-medial to the HJC pointing antero-medially during terminal stance (Figs. 5 and 6(a)). The external GRF moment would therefore be expected to induce an internal rotator external hip joint moment during initial stance and an external rotator external hip joint moment during terminal stance, which is consistent with Fig. 2(a). Alternatively, in the femoral AF transverse plane, the GRF vector was postero-medial to the HJC pointing antero-medially during initial stance whilst antero-lateral to the HJC pointing postero-medially during terminal stance (Figs. 5 and 6(b)). The external GRF moment would therefore be expected to induce an external rotator external hip joint moment during initial stance and an internal rotator external hip joint moment during terminal stance, which is consistent with Fig. 3(a). At the knee, the GRF vector was slightly medial to the KJC during stance, irrespective of the reference frame, as both conventions calculated an abductor internal knee joint moment. In the LF transverse plane, the GRF vector was antero-medial to the KJC pointing postero-medially during initial stance whilst posteromedial to the KJC pointing anteriorly during terminal stance (Figs. 5 and 6(a)). However, at both of these times, the moment arm of the GRF vector and the magnitude of its projection in the LF transverse plane were both considerably small (Fig. 6(a)). The external GRF moment would therefore be expected to induce a small internal rotator external knee joint moment during initial stance and a small external rotator external knee joint moment during terminal stance, which is consistent with Fig. 2(a). Note though that the effect the external GRF moment in the LF transverse plane at the knee joint was relatively small (Fig. 2(a)) in comparison to the corresponding external Free moment (Fig. 2(b)). Thus, at the knee joint in the LF transverse Fig. 5. Orientation of the resultant GRF vector in the LF sagittal plane for the left lower limb during (a) initial stance and (b) terminal stance. The GRF vector is indicated by the thick solid line.

9 A.G. Schache et al. / Journal of Biomechanics 40 (2007) Fig. 6. Orientation of the resultant GRF vector projected into (a) the LF transverse plane, (b) the femoral AF transverse plane and (c) the tibial (proximal) AF transverse plane during initial stance and terminal stance for the left lower limb. The projected GRF vector is indicated by the thick solid line. Note that LF (HJC) and LF (KJC) represent the translated LF such that the origin is located at the HJC and KJC respectively. plane, the resultant profile of the external Total moment was not simply a reflection of the external GRF moment. Alternatively, in the tibial AF transverse plane, the GRF vector was antero-medial to the KJC pointing poster-medially during initial stance (Figs. 5 and 6(c)). Not only was its moment arm quite small at this time but the magnitude of the projected GRF vector in the tibial AF transverse plane was also small because the vertical (z) of the tibial AF was closely aligned with the GRF vector in the sagittal plane for a short period following heel strike. As stance progressed, the GRF vector continued to point antero-medially but the magnitude of its projection into the transverse plane of the tibial AF increased (Fig. 6(c)). The external GRF moment would therefore be expected to have little effect during initial stance but induce an increasing internal rotator external knee joint moment during terminal stance, which is consistent with Fig. 3(a). At the ankle in the LF transverse plane, the GRF vector was close to the AJC in the medio-lateral direction during initial stance but shifted lateral to it through mid to late stance phase, as can be determined from Fig. 7(b). During initial stance, the GRF vector was pointing postero-medially but was close to the AJC (Fig. 8(a)). By terminal stance, the GRF vector was pointing antero-medially and had shifted anteriorly along the foot to the first metatarso-phalangeal joint, increasing its LF transverse plane moment arm about Fig. 7. Invertor-evertor internal ankle joint moment during gait expressed in (a) the foot AF and (b) the LF. The orientation of the projected GRF vector in the frontal plane of both the foot AF and LF during mid stance for the left foot is also displayed. the AJC (Fig. 8(b)). The external GRF moment would therefore be expected to have little effect during initial stance but induce an increasing internal rotator

10 18 A.G. Schache et al. / Journal of Biomechanics 40 (2007) 9 19 Fig. 9. Orientation of the resultant GRF vector projected into the foot AF transverse plane at the AJC during (a) initial stance, (b) mid stance for the left foot (varus alignment) and (c) mid stance for the left foot (valgus alignment). The projected GRF vector is indicated by the thick solid line. Fig. 8. Orientation of the resultant GRF vector projected into the LF transverse plane at the AJC during (a) initial stance and (b) terminal stance for the left foot. The projected GRF vector is indicated by the thick solid line. (adductor) external ankle joint moment during terminal stance, which is consistent with Fig. 2(a). At the ankle in the foot AF transverse plane, the GRF vector was slightly lateral to the AJC during initial stance and slightly medial to the AJC during terminal stance, as can be determined from Fig. 7(a). During initial stance, the GRF vector was antero-lateral to the AJC pointing postero-laterally but with a small moment arm (Fig. 9(a)). The external GRF moment would therefore be expected to have little effect during initial stance, which is consistent with Fig. 3(a). By mid stance, the GRF vector had shifted anteriorly along the foot to the first metatarso-phalangeal joint, increasing its foot AF transverse plane arm about the AJC (Fig. 9(b) and (c)). However, the orientation of the GRF vector at this time was variable across subjects and depended upon the frontal plane posterior calcaneal alignment. For a varus alignment, the GRF vector was pointing posteromedially (Fig. 9(b)) whereas for a valgus alignment, the GRF vector was pointing postero-laterally (Fig. 9(c)). From mid stance onwards, the external GRF moment would therefore be expected to induce an internal rotator (adductor) external ankle joint moment for a varus alignment and an external rotator (abductor) external ankle joint moment for a valgus alignment. This is consistent with Fig. 3(a), which from mid stance onwards displays a high degree inter-subject variability. It is important to note that the findings of the current study relate specifically to able-bodied adult gait. Whilst the particular reference frame was shown to have little influence on the resulting profiles for the sagittal plane hip, knee and ankle joint moments as well as the frontal plane hip and knee joint moments during able-bodied adult gait, this is unlikely to be the case for pathological gait. For example, children with diplegic cerebral palsy typically walk with increased hip and knee flexion as well as increased hip internal rotation (Gage et al., 1995). In such circumstances, the limbs are no longer approximately aligned with the GRF vector in the sagittal plane and differences between reference frames may become apparent for sagittal and frontal plane joint moments during gait. Further research is required to explore this issue. Acknowledgments This project was financially supported by a Health Professional Research Training Fellowship from the Australian National Health and Medical Research Council (Grant ID: ). We would also like to acknowledge the assistance of Mr Bill Reid from the Educational Resource Centre at the Royal Children s Hospital, Melbourne, Australia for the preparation of the figures presented in this study. References Allard, P., Lachance, R., Aissaoui, R., Duhaime, M., Simultaneous bilateral 3-D able-bodied gait. Human Movement Science 15, Andriacchi, T.P., Strickland, A.B., Gait analysis as a tool to assess joint kinetics. In: Berme, N., Engin, A.E., Correia da Silva, K.M. (Eds.), Biomechanics of Normal and Pathological Human Articulating Joints. Martinus Nijhoff Publishers, Dordrecht, pp Benedetti, M.G., Catani, F., Leardini, A., Pignotti, E., Giannini, S., Data management in gait analysis for clinical applications. Clinical Biomechanics 13,

11 A.G. Schache et al. / Journal of Biomechanics 40 (2007) Besier, T.F., Sturnieks, D.L., Alderson, J.A., Lloyd, D.G., Repeatability of gait data using a functional hip joint centre and a mean helical knee. Journal of Biomechanics 36, Bowsher, K.A., Vaughan, C.L., Effect of foot-progression angle on hip joint moments during gait. Journal of Biomechanics 28, Bresler, B., Frankel, J.P., The forces and moments in the leg during level walking. Transactions of the American Society of Mechanical Engineering 72, Davis, R.B., Ounpuu, S., Tyburski, D., Gage, J.R., A gait analysis data collection and reduction technique. Human Movement Science 10, De Leva, P., Adjustments to Zatsiorsky-Seluyanov s segment inertia parameters. Journal of Biomechanics 29, Eng, J.J., Winter, D.A., Kinetic analysis of the lower limbs during walking: what information can be gained from a three dimensional model? Journal of Biomechanics 28, Gage, J.R., DeLuca, P.A., Renshaw, T.S., Gait analysis: Principles and applications. Journal of Bone and Joint Surgery 77A, Grood, E.S., Suntay, W.J., A joint coordinate system for the clinical description of three-dimensional motions: application to the knee. Journal of Biomechanical Engineering 105, Hunt, A.E., Smith, R.M., Interpretation of ankle joint moments during the stance phase of walking: a comparison of two orthogonal axes systems. Journal of Applied Biomechanics 17, Ramakrishnan, K.K., Kadaba, M.P., Wootten, M.E., Lower extremity joint moments and ground reaction torque in adult gait. In: Stein, J.L. (Ed.), Biomechanics of Normal and Prosthetic Gait. The American Society of Mechanical Engineers, New York, pp Schache, A.G., Baker, R., Lamoreux, L.W., Defining the knee joint flexion extension for purposes of quantitative gait analysis: an evaluation of methods. Gait and Posture (in press). Woltring, H.J., A fortran package for generalised, crossvalidatory spline smoothing and differentiation. Advances in Engineering Software 8,