Development and evaluation of a noninvasive marker cluster technique to assess three-dimensional kinematics of the distal portion of the forelimb in horses

Sarah J. Hobbs Department of Technology, University of Central Lancashire, Preston, UK.

Search for other papers by Sarah J. Hobbs in
Current site
Google Scholar
PubMed
Close
 BEng
,
James Richards Department of Allied Health Professions, University of Central Lancashire, Preston, UK.

Search for other papers by James Richards in
Current site
Google Scholar
PubMed
Close
 PhD
,
Bogdan Matuszewski Department of Technology, University of Central Lancashire, Preston, UK.

Search for other papers by Bogdan Matuszewski in
Current site
Google Scholar
PubMed
Close
 PhD
, and
Charlotte Brigden Myerscough College of Agriculture, Bilsborrow, Preston, UK.

Search for other papers by Charlotte Brigden in
Current site
Google Scholar
PubMed
Close
 MSc

Click on author name to view affiliation information

Abstract

Objective—To develop and evaluate a marker cluster set for measuring sagittal and extrasagittal movement of joints in the distal portion of the forelimb in ponies.

Animals—4 ponies.

Procedures—5 infrared cameras were positioned on a concrete walkway in a frontal-sagittal arc and calibrated. Four segments were defined: hoof, middle phalanx, proximal phalanx, and metacarpus. Rigid clusters with 4 retroreflective markers were placed on each segment. A static trial was recorded with additional anatomic markers on the medial and lateral joint lines. Those anatomic markers were removed, and kinematic data were recorded at 240 Hz during walking. An ensemble mean was computed from the 4 ponies from 5 replicates of the walks. Joint kinematic variables were calculated by use of the calibrated anatomical system technique. The design and error dispersion of each marker were evaluated.

Results—Marker clusters were quasiplanar, but variation in orientation error was reduced because the mean radii were > 10 times the largest error dispersion values. Measurements of sagittal rotations of the distal interphalangeal, proximal interphalangeal, and metacarpophalangeal joints were similar to measurements obtained with bone-fixed triads, but larger discrepancies between the 2 methods were found for extrasagittal rotations.

Conclusions and Clinical Relevance—Development of noninvasive methods for quantifying data pertaining to 3-dimensional motion in horses is important for advancement of clinical analysis. The technique used in the study enabled identification of flexion-extension motions with an acceptable degree of accuracy. Appropriate correction algorithms and improvements to the technique may enable future quantification of extrasagittal motions.

Abstract

Objective—To develop and evaluate a marker cluster set for measuring sagittal and extrasagittal movement of joints in the distal portion of the forelimb in ponies.

Animals—4 ponies.

Procedures—5 infrared cameras were positioned on a concrete walkway in a frontal-sagittal arc and calibrated. Four segments were defined: hoof, middle phalanx, proximal phalanx, and metacarpus. Rigid clusters with 4 retroreflective markers were placed on each segment. A static trial was recorded with additional anatomic markers on the medial and lateral joint lines. Those anatomic markers were removed, and kinematic data were recorded at 240 Hz during walking. An ensemble mean was computed from the 4 ponies from 5 replicates of the walks. Joint kinematic variables were calculated by use of the calibrated anatomical system technique. The design and error dispersion of each marker were evaluated.

Results—Marker clusters were quasiplanar, but variation in orientation error was reduced because the mean radii were > 10 times the largest error dispersion values. Measurements of sagittal rotations of the distal interphalangeal, proximal interphalangeal, and metacarpophalangeal joints were similar to measurements obtained with bone-fixed triads, but larger discrepancies between the 2 methods were found for extrasagittal rotations.

Conclusions and Clinical Relevance—Development of noninvasive methods for quantifying data pertaining to 3-dimensional motion in horses is important for advancement of clinical analysis. The technique used in the study enabled identification of flexion-extension motions with an acceptable degree of accuracy. Appropriate correction algorithms and improvements to the technique may enable future quantification of extrasagittal motions.

Techniques for collecting in vivo kinematic data from horses are numerous, but many models of the distal segment of the forelimb work on the premise that the PIPJ is a rigid joint. However, it is known that motion exists within the PIPJ has been illustrated, and 2 methods1,2 have recently been proposed for obtaining in vivo measurements of 3-D rotations of this joint. In 1 method,1 triads of ultrasonic kinematic markers were attached to underlying bony segments of the distal portion of the forelimb with intracortical pins. Special software was used to record the 3-D coordinates of 3 microphones on each triad at 60 Hz from 4 horses walking in a straight line on a hard surface. Repeatability and consistency of the axes and precision of the method were evaluated, and 3-D rotations of the distal joints were calculated. The range of extrasagittal movements observed in the distal joints and movement within the PIPJ confirmed the need to include 3-D kinematics in equine gait assessment. Those data also yielded important benchmarking information. The use of this method as a gait assessment tool, however, is somewhat restrictive in that the required equipment is highly specialized and fixation of markers is invasive.

In another method,2 6 infrared camerasa sampling at 240 Hz tracked 13 retroreflective markers positioned on the joint centers, hooves, and MCIII or third metatarsal bone of horses ridden at the trot on asphalt. An extended Kalman filter3 was used to minimize skin displacement, and joint angles of the forelimbs and hind limbs were compared. The sagittal ROM and temporal differences were reported, along with data pertaining to collateromotion (abduction and adduction) of the DIPJ. Because data were recorded at the trot, the ROM (32.2°) reported for the PIPJ in that study cannot be used for comparison with data generated by horses at the walk.

Description of motion in distal segments of the limbsis confounded by soft tissue artifacts, difficulty in locating joint centers, and size of the proximal inter phalangeal segments. The invasive method of measurement used in the first study1 was intended to overcome those problems and reduce error, and precision was reportedly better than that associated with the use of similar photogrammetric methods.4 In the second study,2 a photogrammetric method with a simple marker set mounted on the skin was used; that method compensated for errors by use of new processing methods, but precision was not reported. Despite these advances, a need exists for development of a noninvasive, readily available method for describing the 6 df of the distal segments of the limbs for clinical use in equine gait analysis.

Stereophotogrammetry (motion capture) is used commonly in human movement analysis to measure the instantaneous positions of markers on the skin surface.5 Soft tissue artifacts are the most significant source of error,6 so different models using skin-mounted, wand-mounted, and rigid-plate–mounted markers have been developed and evaluated in an attempt to minimize or compensate for these errors. In addition, soft tissue artifacts have been evaluated by comparing skin-mounted markers with pinor external-fixator–mounted markers or by use of techniques involving roentgen photogrammetry. On the basis of what is presently known from such studies,7 the effects of skin or soft tissue movement can be addressed and data reliability can be improved in human movement analysis.

The calibrated anatomical systems technique (CAST)8 was first proposed to standardize descriptions of movement in the pelvis and lower limb segments in research laboratories and clinical centers. The method involves designating an anatomic frame for each segment by identification of anatomic landmarks and segment-tracking markers or marker clusters. Marker clusters can be attached directly to the skin or mounted on rigid fixtures, depending on anatomic features of the structure being analyzed and the activity and nature of the study.

These criteria allow for variability in marker cluster configuration, position, and orientation on each segment, which may or may not be optimal. Consequently, design criteria and the attachment and positioning of clusters have been addressed in other studies. To minimize error propagation from marker coordinates to bone-embedded frame position and orientation and to identify optimal markercluster position and orientation on the limb, design criteria for optimal reconstruction of 3-D bone movement have been developed.9 In 1 study,10 11 marker cluster sets, including skin-, rigid-plate–, and wand-mounted markers, were compared for estimating tibial rotation in humans. In that study, the marker cluster set considered optimal10 was a constrained rigid shell with a cluster of 4 markers; that finding was in accordance with design criteria specified in the earlier study,9 in which directions were given pertaining to the size, shape, orientation, and position of clusters.9 The aim of the present study was to develop and evaluate a marker cluster set for distal segments of the forelimb that could accurately measure sagittal and extrasagittal joint movement in horses.

Materials and Methods

Animals—Approval from the University of Central Lancashire Animal Projects Committee was obtained for this project. Four ponies without lameness and with mean ± SD age and height of 10 ± 2.9 years and 1.46 ± 0.05 m, respectively, were used in the study. Ponies were shod 21 ± 11 days before testing with standard iron horseshoes on both forefeet. Several days before testing, each pony was trained by having marker clusters attached to the left forelimb and being walked in hand through the indoor concrete walkway. On the test day, ponies were familiarized with the cameras in the same way before markers were attached.

Marker development—Foam imprints of the frontal and lateral aspects of the lower portion of an equine forelimb were produced by pressing deformable foam onto the limb in different positions. Plaster casts of the profiles of each segment were fashioned from the imprints. Rectangular shapes of a 3-mm-thick polymer were heated and molded to the segment profiles. Several shapes were made for each segment to correspond to the different positions. Wands were fashioned with 14-mm-diameter markers attached to aluminium rods of various lengths. Four wands were attached in a cluster array to each individual polymer shape, making up the cluster. Combinations of cluster arrays were captured and tracked on the lower portion of human limbs until a good solution for tracking 16 markers below the carpus resulted (Figure 1).

Figure 1—
Figure 1—

Photograph of the distal portion of a forelimb of a pony instrumented with marker cluster arrays and anatomic markers for data recording during a static trial in a 3-D kinematics study.

Citation: American Journal of Veterinary Research 67, 9; 10.2460/ajvr.67.9.1511

Data collectionmdash;Five infrared camerasa were positioned on a concrete walkway in a frontal-sagittal arc and calibrated. Four segments were defined for 1 side only: hoof (PIII), PII, PI, and McIII. Rigid clusters with 4 retroreflective markers were placed on each segment of the left forelimb. Marker clusters were attached with double-sided tape on the underside and secured with additional insulation tape over the top of the rigid plate for all but the hoof cluster. Insulation tape was tightened over the top of the polymer plate with a tension similar to what would be exerted with an exercise bandage to fix the plate firmly to each segment; this procedure was performed by the same person for each pony for consistency. Additional anatomic markers were positioned at medial and lateral locations on the proximal portion of McIII, at the site of attachment of the proximal collateral ligaments of the MPJ and PIPJ, and at the coronary band and distal border of the hoof in line with the PIPJ and MPJ markers. A static trial was recorded with each pony standing with limbs squarely positioned, after which the markers were removed.

Each pony was led in hand at a walk through the calibrated space, and kinematic data from the rigid clusters were recorded at 240 Hz. Data from a minimum of 10 trials were collected to ensure that sufficient stance-phase data were recorded within the optimum volume of the calibrated space. Completed successful trials were digitized with a commercially available software programa and exported as a C3-D file.b Kinematic data were filtered with a low-pass, fourthorder Butterworth filter with a cutoff frequency of 10 Hz. For each pony, an ensemble mean was computed from 5 replicates of the walks.

An anatomical frame or local SCS was defined for each segment with respect to the calibrated LCS8 in 3-D movement analysis software.b Measurement of the segment pose was determined by use of a least squares procedure9 in which the orientation matrix T and translation vector O were computed at point Pi from the noncolinear markers on each cluster point Ai, according to the following equation:

article image

where Pi is the coordinate of the ith cluster point in the calibrated LCS, Ai is the coordinate of the ith cluster point in the calibrated SCS, and m is the number of target markers on the segment (m > 2).

The SCS was defined as the y-axis as cranial-caudal, x-axis as medial-lateral, and z-axis as proximal-distal. Joint rotations were calculated from the static trial by use of the Cardan sequence xyz, where flexion (rotation of the distal segment towards the palmar surface of the proximal segment) was defined as positive rotation about the x-axis, abduction or lateromotion (rotation of the distal segment in the lateral direction) was defined as positive rotation about the y-axis, and external or lateral rotation (rotation of the distal segment about the long axis away from midline) was defined as positive rotation about the z-axis.

Hoof angles were calculated relative to the LCS, so values of 0 were observed for pitch when the longitudinal axis of the hoof segment was aligned with the dorsal wall in the sagittal plane, for eversion-inversion when the hoof segment was horizontal in the frontal plane, and for axial rotation when the hoof segment was aligned with the x- and y-axes in the transverse plane. Pitch rotation was defined as rotation of the hoof about the x-axis, with palmar rotation of the heel about the toe designated as positive.11 Outward rotation or eversion of the hoof so that the sole faced laterally was defined as positive rotation about the y-axis, and inward rotation was defined as inversion. External or lateral rotation of the hoof about the long axis away from midline was defined as positive rotation about the z-axis. Sagittal rotation of McIII was calculated relative to the LCS, with backward rotation of the proximal end defined as positive.

Impact was determined from inspection of the vertical hoof displacement and velocity curves.12 Hoof stabilization was defined as the point where pitch velocity stabilized. Midstance was defined as the point where McIII reached a vertical position in the sagittal plane. Heel-off was defined as the point where pitch velocity resumed, and TO was also identified from inspection of the vertical hoof displacement and velocity curves.12

Statistical analysis—Mean ± SD angle-time diagrams of the MPJ, PIPJ, and DIPJ were plotted from an ensemble mean for each pony. The percentage stance to stride and joint events and the maximum, minimum, and ROM values for the MPJ, PIPJ, and DIPJ were found, and coefficient of variability (CV) and values for SD divided by mean times 100 were calculated.

Each cluster was evaluated according to 6 described design criteria9 by use of a random sample of 5 trials. The mean radius (r) of the cluster was calculated. Experimental errors were estimated with a computer programc developed by the authors (rather than with the uncorrelated isotropic model9) to find the principal axes (axes along which data fluctuate the most) and the corresponding error dispersion around these axes (σ1, σ2, and σ3). To calculate dispersion values, the positions of detected markers from all captured frames were coregistered by use of the rigid transformation. It was assumed that correspondence between positions of markers in different frames was given by the rigid 3-D transformation, so deformation of the plate to which markers in the cluster were attached was neglected. For each frame, the rigid transformation was calculated by use of the singular value decomposition method.13 For each marker in the cluster, the autocorrelation matrix was estimated. This was followed by eigen decomposition of the matrix to determine the principal axes and the corresponding error dispersion values. The largest error dispersion value (σmax = max [σ1, σ2, and σ3]) estimated from a minimum of 150 frames of data was used to calculate the r/σmax value to satisfy design criteria.9

Results

Marker analysis—Each cluster contained 4 markers (Figure 2; Appendix 1). The mean radius of the cluster was found to be > 10 times the largest error dispersion value. Values for each marker were calculated from a random sample of trials from ponies 1, 2, and 3, where a similar number of frames was used in the analysis (Table 1). The 3-D isotropy index was found to be close to 0 for all clusters, so they were evaluated as quasiplanar clusters.9 Values for the 2D isotropy index were > 0.5 for the PII and PIII clusters but were < 0.5 for the McIII and PI clusters. Marker clusters were attached to the skin in locations where slippage was expected to be minimal. Principal axes for the McIII and PI clusters in the SCS obtained from the static trial were oriented in the proximal-distal direction, and axes for the PII and PIII clusters were oriented in the medial-lateral and proximal-distal directions.

Table 1—

Values for largest error dispersion value (σmax) and mean radius of the cluster divided by the largest error dispersion value (r/σmax) for individual markers in each marker cluster in 5 random trials from ponies 1, 2, and 3.

SegmentParameterTrial
12345
McIIIσmax1.011.392.501.561.14
r/σmax37.927.615.424.233.8
PIσmax1.021.551.371.241.15
r/σmax40.925.329.833.734.3
PIIσmax1.051.531.051.220.98
r/σmax27.719.628.324.430.2
PIIIσmax0.530.570.440.590.60
r/σmax64.662.181.260.957.9

σ = Dispersion value. r = Radius.

Figure 2—
Figure 2—

Photograph of marker clusters used in 4 ponies in a 3-D kinematics study of forelimb movement. Clusters were noninvasively affixed to the skin overlying distal segments of the forelimb. Marker clusters were affixed to McIII (a), PI (b), PII (c), and PIII (d).X,Y, and Z represent the axes around which movement was measured.

Citation: American Journal of Veterinary Research 67, 9; 10.2460/ajvr.67.9.1511

Kinematics—Movement patterns recorded for the distal forelimb joints were summarized (Figures 3 and 4). Angle data were expressed with respect to the limb position at the moment of impact. Temporal data and flexion-extension data for the MPJ, PIPJ, and DIPJ for important gait events were summarized (Tables 2 and 3).

Table 2—

Percentage of time spent in stance phase (SD) during flexion-extension movement in the MPJ, PIPJ, and DIPJ at stride and joint events for all ponies and for each of the 4 ponies separately.

Event% Stance
All poniesPony 1Pony 2Pony 3Pony 4
Stride
HS6.1 (1.8)6.0 (1.8)5.1 (1.7)7.2 (1.4)6.1 (2.1)
Mid39.6 (4.3)45.9 (0.3)39.1 (2.2)35.7 (2.8)37.7 (0.8)
HO78.2 (2.2)78.3 (2.4)78.4 (2.5)76.8 (2.6)79.2 (1.0)
TO100 (0)100 (0)100 (0)100 (0)100 (0)
Joint
MPJ-ME48.6 (11.7)56.2 (8.5)44.0 (14.6)42.2 (10.9)52.1 (8.9)
PIPJ-MF11.7 (6.8)18.0 (9.3)5.9 (3.8)10.6 (1.4)11.2 (4.6)
PIPJ-ME86.2 (8.1)86.2 (8.7)79.1 (8.4)87.5 (6.5)92.1 (3.8)
DIPJ-MF20.7 (4.1)24.4 (0.8)21.9 (2.1)15.1 (3.0)20.6 (3.4)
DIPJ-ME87.6 (6.5)88.3 (5.4)89.3 (3.1)88.6 (10.5)84.2 (5.7)

Mid = Midstance. MPJ-ME = From impact to maximal extension. PIPJ-ME = From peak flexion to maximal extension. PIPJ-MF = From impact to peak flexion. DIPJ-MF = From impact to maximal flexion. DIPJ-ME = From maximal flexion to maximal extension.

Table 3—

Range of motion (SD; % CV) for flexion-extension movement in the MPJ, PIPJ, and DIPJ at joint events for all ponies and for each of the 4 ponies separately. Negative values are observed when peak flexion is less than flexion at the point of impact. Data represent number of degrees.

EventROM
All poniesPony 1Pony 2Pony 3Pony 4
Joint
MPJ-ME−25.5 (4.1; 16)−28.7 (1.8; 6)−25.3 (5.1; 20)−26.2 (1.8; 7)−21.6 (3.6; 17)
PIPJ-MF0.40 (2.5; 633)1.69 (0.6; 38)1.9 (1.5; 78)−1.74 (0.8; 48)−0.3 (4.0; 1,424)
PIPJ-ME−15.6 (3.4; 22)−10.7 (0.4; 4)−17.1 (1.9; 11)−17.5 (1.1; 6)−16.9 (3.0; 18)
DIPJ-MF7.4 (3.3; 44)6.7 (0.9; 14)12.3 (2.4; 19)6.7 (3.1; 46)5.0 (1.7; 34)
DIPJ-ME−28.5 (8.6; 30)−23.4 (1.0; 4)−39.1 (4.0; 10)21.4 (4.5; 21)−33.0 (8.6; 26)

See Table 2 for key.

Figure 3—
Figure 3—

Mean curves of the joint angles for flexion (Fl)–extension (E), abduction (Ab)–adduction (Ad), and internal (In)–external (Ex) rotation in the MPJ, PIPJ, and DIPJ plotted against time (expressed as percentage of stance phase) for 4 ponies. Angle pertains to limb position at the moment of impact. Black = Pony 1. Gray = Pony 2. Red = Pony 3. Green = Pony 4.

Citation: American Journal of Veterinary Research 67, 9; 10.2460/ajvr.67.9.1511

Figure 4—
Figure 4—

Mean curves of rotation of the hoof and McIII relative to the lab during stance phase in the same 4 ponies as in Figure 3; data were generated from 5 replicates of the walks. See Figure 3 for key.

Citation: American Journal of Veterinary Research 67, 9; 10.2460/ajvr.67.9.1511

MPJ kinematics—Maximal extension of the MPJ was found after midstance at 48.6 ± 11.7%. The ROM from impact to maximal extension was 28.7 ± 1.8° to 21.6 ± 3.6° for ponies 1 and 4, respectively, and was followed by flexion through the remainder of the stance phase. After midstance, a pattern of external and then internal rotation was observed in ponies 1, 2, and 3. Several degrees of abduction and adduction were also recorded for each pony, but patterns varied among individuals.

PIPJ kinematics—During landing, a slight reduction in flexion of −1.11 ± 1.27° was accompanied by abduction or a reduction in adduction of 1.26 ± 1.01° and by slight axial rotation in the external direction.

This was followed by a flexion peak, which reached a maximum soon after HS. Slight adduction (abduction for pony 2) and internal rotation followed as the joint extended. Maximum extension occurred close to HO, with internal rotation and an abduction peak occurring at the same time or soon afterwards, except for in pony 4, in which external rotation and an adduction peak occurred. During breakover, rapid flexion and adduction occurred, except for in pony 4, in which rapid flexion and abduction occurred.

DIPJ kinematics—The DIPJ flexed during early stance to a maximum value at 20.5 ± 7.1% of the stance phase. This was followed by extension, which reached a maximum at 87.6 ± 6.5% of the stance phase. Maximum ROM from peak to peak was 51.4 ± 5.8° for pony 2, and the minimum was 28.2 ± 6.5° for pony 3. Slight extrasagittal rotations were observed with peaks or inflections between HS and HO, but the direction of rotation varied among ponies, and no typical pattern was ascertained.

Hoof kinematics—Three ponies landed heel first. Ponies 4 and 2, respectively, had 8.3 ± 3.0° and 6.1 ± 3.1° of forward rotation before HS, whereas pony 3 landed toe first with −1.8 ± 3.7° of rotation before HS. Slight extrasagittal movements of the hoof were observed from impact to HS, but the directions of rotations differed among ponies. For all ponies, pitch resumed at HO with an ROM from −86.3 ± 4.6° to −76.6 ± 1.7° for ponies 4 and 1, respectively, with inversion and external rotation to TO.

Discussion

We evaluated a noninvasive method for collecting 3-D kinematic data from the distal portion of equine forelimbs. The design of each marker cluster pattern was constrained by segment size, movement of that segment during gait, distance to neighboring clusters, and success in tracking segment movement. Once sufficient data were collected, the design of each cluster could be analyzed. Dispersion errors for the McIII clusters were the largest (σmax = 1.5 ± 0.6 mm), with values decreasing in more distal segments of the limb. The increase in displacement and velocity of more proximal segments over the stance phase may have influenced error propagation, but more notably, the 2-dimensional isotropy index for the McIII and PI marker clusters was < 0.5, which increased the dispersion error.9 Improvements in marker cluster configurations, camera set-up, and calibration procedures may help reduce the value of σmax. Variation in the orientation error was reduced, as values for r/σmax were > 10 for all marker clusters despite the 3-D isotropy index being low. All clusters were found to be quasiplanar, and although this was not a critical factor,9 further work is needed to address the effects of cluster shape and size on alignment of principal axes.

Range of motion and temporal patterns for MPJ movement were comparable with values reported in earlier studies.1,14 Maximum MPJ extension occurred later in the stance phase than has been reported but with an equivalent ROM from impact to maximum extension. However, mean values for ROM in abduction-adduction and internal-external rotation were larger, which was attributed primarily to soft tissue artifact. Extrasagittal motion has been reported1,14 and suggested to occur at impact as a result of global limb adduction and inclination of the hoof at the point of ground contact.1 After midstance, internal rotation during flexion of the MPJ was observed in ponies 1, 2, and 3. This finding has previously been reported1,14 and may be explained by external rotation of McIII occurring while the hoof is still in contact with the ground. This rotation would induce changes in the 3-D balance of the distal segments, particularly as tension in the superficial flexor tendon and suspensory ligament decreased.

Movement patterns were detected in the PIPJ in flexion-extension, abduction-adduction, and internalexternal rotation. The PIPJ movements in flexionextension were similar to those cited previously,14 with values for ROM being within 5° of those cited in the earlier study. The variability in ROM for maximum flexion both within and between subjects was greater than has previously been reported,1 particularly for pony 4, but an SD of 11° for an approximate ROM of 10° at the trot has been reported.15 Maximum flexion of the PIPJ was observed slightly earlier in the stance phase than has been previously reported for horses at the walk1,14 but occurred before maximum flexion of the DIPJ, a finding that was in agreement with findings from the earlier studies. A slight peak in abduction from impact to HS was observed that was consistent with previously published data,14 but little movement after HS except for slight abduction during breakover was detected in that study. In ponies in the present study, several degrees of rotation were observed but there was no distinct pattern. Internal rotation at HS followed by external rotation through the remainder of the stance phase was observed in horses in the previous study,14 whereas internal rotation after HS that reached a minimum after midstance and was followed by external rotation was recorded in the present study. The ROMs for abduction-adduction and internalexternal rotations in this study were also greater than those that have been reported.

Maximum flexion of the DIPJ in the present study was found to be within 1% of the results for flexion reported from previous work,1 but maximum extension occurred more than 10% earlier in the stance phase. In addition, a smaller ROM was recorded between the points of maximum flexion and maximum extension. Underestimation of DIPJ extension and slight overestimation of PIPJ extension may result from proximity of the PII cluster to the coronet band, where skin movement may be more influenced by hoof movement. Distinct patterns of abduction at impact, internal rotation after HS, and adduction and external rotation during breakover have been reported,14 but no such patterns were detected in the present study.

Forward rotation before HS for 3 ponies was similar to values previously reported,14 but the ROM during breakover was much greater. Values obtained in the present study were comparable with those from an earlier study,11 in which pitch rotation followed the inclination of McIII before TO to ensure hoof clearance from the ground in the swing phase. External rotation and inversion (or a lateral rocking motion) were observed in the present study, and breakover occurred over the lateral aspect of the toe. Magnitudes were greater for all axes than those that have been recorded with invasive measurement methods,14 but because those results relate to the hoof segment with respect to the laboratory, differences may not be attributable to soft tissue artifacts. Differences in gait patterns with respect to the hoof may therefore be surmised to be related to differences in experimental methods, special shoes used in the previous studies,1,14 or individual differences among horses or ponies.

Findings of the study reported here indicate that flexion-extension patterns for joints of the distal portion of the forelimb may be recorded noninvasively with an acceptable degree of accuracy. Sagittal movements of the PIPJ in trotting horses recorded from skin markers on the center of rotation have been reported,15–18 and the ROM in sound horses at the trot has been reported17 to be as high as 55.5 ± 8.1° during flexion-extension. In one of those studies,15 it was reported that the data were broadly representative of the movements and functions of the proximal interphalangeal joint, but investigators acknowledged that skin displacement artifacts had not been investigated for marker locations on the PI and that the data may not represent motion of the underlying bone.

Large discrepancies exist among data from previous work for extrasagittal joint motion and are most evident in regard to axial rotation. Soft tissue artifacts likely contribute to error, particularly in the axial direction, because the skin can slide more easily around the long axis of the underlying bones. Soft tissue artifacts from markers positioned on joint centers, which have been used in 3-D studies,2 were large in earlier 2-D studies.19 Although methods of compensating for those types of errors have been proposed,20 it was suggested that these large artifacts may counteract rigid body theory and that invasive markers are necessary.1 Use of the calibrated anatomical systems technique requires that tracking markers be attached to each bony segment, and segments are identified from an anatomic frame. Because the frame is reproduced with the pony standing squarely, better definition of joint centers and segment endpoints can be achieved. The tracking markers in each cluster are used to track segment motion. Therefore, errors resulting from soft tissue artifacts during motion relate to movement between markers and underlying bone rather than movement of an anatomic marker away from a segment end point or center of rotation.

Although discrepancies exist between findings in the present study and other studies involving 3-D kinematics in horses, our results provide a basis for further development of noninvasive quantification of 3-D movement of the distal portion of the equine forelimb. Because data were obtained from a small sample population, typical patterns that have been described in the present study and in previous studies may not necessarily represent typical patterns that would be observed in clinically normal horses and ponies. Real differences have been found between the 3-D kinematics of trotting horses,21,22 particularly for abduction and adduction movement, and it has been proposed that further investigations should include a larger population of horses. Our observation of differences in extrasaggital movement among subjects was consistent with findings pertaining to gait in humans,23 in which movement in the frontal plane of the distal portion of the limb is inconsistent among runners. The authors of that study23 hypothesized that frontal-plane kinematics are influenced by anatomy of the foot and ankle, and it is possible that factors such as breed, hoof balance, fitness, and conformation likewise influence equine gait patterns. Distinct differences in extrasagittal movement, but not in flexion-extension movement, were observed in pony 4, which had been shod more recently than the other ponies. Interestingly, adduction and axial rotation angles towards terminal stance were amplified, but interpretation of those data was beyond the scope of this study.

Further work is required to develop improved techniques for collecting 3-D kinematic data with skinbased marker clusters, quantifying marker cluster movement with respect to the bones in the distal limb segments, and developing correction algorithms to improve data reliability. Quantification of skin displacement on the tibia and third metatarsal bone24 and on the radius25 has been described. In those studies, the external segment surface movement was modelled by means of a truncated Fourier series. In another study,26 bone-fixed marker data were compared with skinmarker data and corrected skin-marker data, incorporating correction algorithms published in the earlier study.24 Good agreement was found for flexion-extension movement for both the swing and stance phases, and much closer agreement was observed after correction for skin displacement for abduction-adduction in the swing phase. However, it was concluded that internal-external rotation and medial-lateral translation were not measured to an acceptable degree of accuracy, highlighting the difficulties in quantification of extrasagittal movements noninvasively.

Simultaneous measurement of segment movement from bone-pin marker clusters and skin marker clusters is required to assess soft tissue artifacts with the calibrated anatomical systems technique because the characteristics of those errors will be different from those associated with individual skin markers as a result of their properties of inertia. Gait, marker location, segment length, and the pressure applied by the insulating tape may influence underlying movement and must be taken into consideration when developing compensation or correction algorithms. Further work may include development of alternative marker cluster configurations to reduce inertial characteristics and allow for quantification of faster gaits. Alternative methods of marker cluster attachment that would reduce soft tissue artifact should also be investigated. Finally, pose estimation may also be improved by use of the multiple anatomical landmark calibration technique.27

Removal of or compensation for soft tissue artifacts will improve the reliability of data derived from distal segments of the limb; with those improvements, the calibrated anatomical systems technique may represent an important advancement in noninvasive assessment of gain in the distal portions of the limb in horses. Detailed 3-D analysis of gait would then be possible in elite sports horses and clinical patients, which cannot be evaluated with invasive methods. The technique was able to identify flexion-extension motions with an acceptable degree of accuracy, and correction algorithms and technical improvements may allow quantification of extrasagittal motions.

ABBREVIATIONS

PIPJ

Proximal interphalangeal joint

3-D

3-dimensional

McIII

Third metacarpal bone

ROM

Range of motion

DIPJ

Distal interphalangeal joint

PIII

Distal or third phalanx

PII

Middle or second phalanx

PI

Proximal or first phalanx

MPJ

Metacarpophalangeal joint

SCS

Segment coordinate system

LCS

Laboratory coordinate system

TO

Toe-off

HS

Hoof stabilization

HO

Heel-off

a.

ProReflex, Qualysis Medical AB, Goteburg, Sweden.

b.

C-Motion Inc, Rockville, Md.

c.

Matlab, The Mathworks Inc, Natick, Mass.

References

  • 1

    Chateau H, Degueurce C, Denoix JM. Evaluation of three-dimensional kinematics of the distal portion of the forelimb in horses walking in a straight line. Am J Vet Res 2004;65:447455.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 2

    Johnston C, Kendall A & Spannar I, et al. The 3-D kinematics of the distal limb in the ridden horse, in Proceedings. 5th Int Conf Equine Locomotion 2004;2529.

    • Search Google Scholar
    • Export Citation
  • 3

    Cerveri P, Pedotti A, Ferrigno G. Robust recovery of human motion from video using Kalman filters and virtual humans. Hum Mov Sci 2003;22:377404.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 4

    Degueurce C, Chateau H & Pasqui-Boutard V, et al. Concrete use of the joint coordinate system for quantification of articular rotations in the digital joints of the horse. J Vet Res 2000;31:297311.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 5

    Cappozzo A, Della Croce U & Leardini A, et al. Human movement analysis using photogrammetry. Part 1: theoretical background. Gait Posture 2005;21:186196.

    • Search Google Scholar
    • Export Citation
  • 6

    Andriacchi TP, Alexander EJ. Studies of human motion: past, present and future. J Biomech 2000;33:12171224.

  • 7

    Leardini A, Chiari L & Della Croce U, et al. Human movement analysis using stereophotogrammetry. Part 3. Soft tissue artifact assessment and compensation. Gait Posture 2005;21:212225.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 8

    Cappozzo A, Catani F & Della Croce U, et al. Position and orientation in space of bones during movement: anatomical frame definition and determination. Clin Biomech 1995;10:171178.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 9

    Cappozzo A, Cappello A. Surface-marker cluster design criteria for 3-D bone movement reconstruction. IEEE Trans Biomed Eng 1997;44:165174.

    • Search Google Scholar
    • Export Citation
  • 10

    Manal K, McClay I & Stanhope S, et al. Comparison of surface mounted markers and attachment methods in estimating tibial rotations during walking: an in vivo study. Gait Posture 2000;11:3845.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 11

    Fredricson I, Drevemo S. A new method of investigating equine locomotion. Equine Vet J 1971;3:137140.

  • 12

    Mickelborough J, van der Linden ML & Richards J, et al. Validity and reliability of a kinematic protocol for determining foot contact events. Gait Posture 2000;11:3237.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 13

    Umeyama S. Least-squares estimation of transformation parameters between two point patterns. IEEE Trans Pattern Anal Mach Intell 1991;13:376380.

  • 14

    Chateau H, Degueurce C, Denoix JM. Effects of 6degree elevation of the heels on 3-D kinematics of the distal portion of the forelimb in the walking horse. Equine Vet J 2004;36:649654.

    • Search Google Scholar
    • Export Citation
  • 15

    Clayton HM, Singleton WH & Lanovas JL, et al. Sagittal plane kinematics and kinetics of the pastern joint during the stance phase of the trot. Vet Comp Orthop Traumatol 2002;15:1517.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 16

    Drevemo S, Johnston C & Roepstorff L, et al. Nerve block and intra-articular anasthesia of the forelimb in the sound horse. Equine Vet J Suppl 1999;30:266269.

    • Search Google Scholar
    • Export Citation
  • 17

    Johnston C, Gottlieb-Vedi M, Drevemo S. The kinematics and fatigue in the standard bred trotter. Equine Vet J Suppl 1999;30:249253.

  • 18

    Roepstorff L, Johnston C, Drevemo S. The effect of shoeing on kinetics and kinematics during the stance phase. Equine Vet J Suppl 1999;30:279285.

    • Search Google Scholar
    • Export Citation
  • 19

    Crevier-Denoix N, Roosen C, Dardillat C. Effects of heel and toe elevation upon the digital joint angles in the standing horse. Equine Vet J Suppl 2001;33:7478.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 20

    van Weeren PR, van den Bogert AJ, Barneveld A. Correction models for skin displacement in equine kinematic gait analysis. J Equine Vet Sci Suppl 1992;12:178192.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 21

    Clayton HM, Sha D & Stick JA, et al. Three-dimensional carpal kinematics of trotting horses. Equine Vet J 2004;36:671676.

  • 22

    van Weeren PR, van den Bogert AJ & Back W, et al. Kinematics of the standardbred trotter measured at 6, 7, 8 and 9 m/s on a treadmill, before and after 5 months of prerace training. Acta Anat (Basel) 1993;146:154161.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 23

    De Wit B, De Clercq D, Aerts P. Biomechanical analysis of the stance phase during barefoot and shod running. J Biomech 2000;33:269278.

  • 24

    Lanovaz JL, Khumsap S, Clayton HM. Quantification of 3-D skin displacement artifacts on the equine tibia and third metatarsus. Equine Comp Exerc Physiol 2004;1 141150.

    • Crossref
    • Search Google Scholar
    • Export Citation
  • 25

    Sha DH, Mullineaux DR, Clayton HM. Three-dimensional analysis of patterns of skin displacement over the equine radius. Equine Vet J 2004;36:665670.

    • Search Google Scholar
    • Export Citation
  • 26

    Khumsap S, Lanovaz JL, Clayton HM. Verification of skinbased markers for 3-dimensional kinematic analysis of the equine tarsal joint. Equine Vet J 2004;36:655658.

    • Search Google Scholar
    • Export Citation
  • 27

    Cappello A, Cappozzo A & La Palombara PF, et al. Multiple anatomical landmark calibration for optimal bone pose estimation. Hum Mov Sci 1997;16:259274.

    • Crossref
    • Search Google Scholar
    • Export Citation

Appendix 1

Appendix 1

Location and physical characteristics of marker clusters attached noninvasively to the skin of 4 ponies during collection of 3-D kinematic data from joints of the distal portion of the forelimb.

TrialPolymer plate mass (g)Polymer plate dimensions (mm)Location on forelimb
McIII20.756 × 35 × 3Dorsal aspect of McIII towards the proximal end
PI22.038 × 35 × 3Lateral aspect of PI over the extensor branch of the suspensory ligament
PII12.227 × 19 × 3Dorsal aspect of PII above the coronary band
All Time Past Year Past 30 Days
Abstract Views 67 0 0
Full Text Views 591 435 38
PDF Downloads 109 62 4
Advertisement